diff --git a/.editorconfig b/.editorconfig index c90b171f55..0722ac73c8 100644 --- a/.editorconfig +++ b/.editorconfig @@ -52,3 +52,11 @@ insert_final_newline = unset [vendor/miniaudio/miniaudio.h] trim_trailing_whitespace = unset insert_final_newline = unset + +[tools/server/webui/**] +indent_style = unset +indent_size = unset +end_of_line = unset +charset = unset +trim_trailing_whitespace = unset +insert_final_newline = unset diff --git a/.github/workflows/server.yml b/.github/workflows/server.yml index f6da488576..1ea1300c2e 100644 --- a/.github/workflows/server.yml +++ b/.github/workflows/server.yml @@ -76,51 +76,206 @@ jobs: run: | pip install -r tools/server/tests/requirements.txt - # Setup nodejs (to be used for verifying bundled index.html) - - uses: actions/setup-node@v4 + webui-setup: + name: WebUI Setup + runs-on: ubuntu-latest + steps: + - name: Checkout code + uses: actions/checkout@v4 with: - node-version: '22.11.0' + fetch-depth: 0 + ref: ${{ github.event.inputs.sha || github.event.pull_request.head.sha || github.sha || github.head_ref || github.ref_name }} - - name: WebUI - Install dependencies - id: webui_lint + - name: Setup Node.js + uses: actions/setup-node@v4 + with: + node-version: "22" + cache: "npm" + cache-dependency-path: "tools/server/webui/package-lock.json" + + - name: Cache node_modules + uses: actions/cache@v4 + id: cache-node-modules + with: + path: tools/server/webui/node_modules + key: ${{ runner.os }}-node-modules-${{ hashFiles('tools/server/webui/package-lock.json') }} + restore-keys: | + ${{ runner.os }}-node-modules- + + - name: Install dependencies + if: steps.cache-node-modules.outputs.cache-hit != 'true' + run: npm ci + working-directory: tools/server/webui + + webui-check: + needs: webui-setup + name: WebUI Check + runs-on: ubuntu-latest + steps: + - name: Checkout code + uses: actions/checkout@v4 + with: + fetch-depth: 0 + ref: ${{ github.event.inputs.sha || github.event.pull_request.head.sha || github.sha || github.head_ref || github.ref_name }} + + - name: Setup Node.js + uses: actions/setup-node@v4 + with: + node-version: "22" + + - name: Restore node_modules cache + uses: actions/cache@v4 + with: + path: tools/server/webui/node_modules + key: ${{ runner.os }}-node-modules-${{ hashFiles('tools/server/webui/package-lock.json') }} + restore-keys: | + ${{ runner.os }}-node-modules- + + - name: Run type checking + run: npm run check + working-directory: tools/server/webui + + - name: Run linting + run: npm run lint + working-directory: tools/server/webui + + webui-build: + needs: webui-check + name: WebUI Build + runs-on: ubuntu-latest + steps: + - name: Checkout code + uses: actions/checkout@v4 + with: + fetch-depth: 0 + ref: ${{ github.event.inputs.sha || github.event.pull_request.head.sha || github.sha || github.head_ref || github.ref_name }} + + - name: Setup Node.js + uses: actions/setup-node@v4 + with: + node-version: "22" + + - name: Restore node_modules cache + uses: actions/cache@v4 + with: + path: tools/server/webui/node_modules + key: ${{ runner.os }}-node-modules-${{ hashFiles('tools/server/webui/package-lock.json') }} + restore-keys: | + ${{ runner.os }}-node-modules- + + - name: Build application + run: npm run build + working-directory: tools/server/webui + + webui-tests: + needs: webui-build + name: Run WebUI tests + permissions: + contents: read + + runs-on: ubuntu-latest + + steps: + - name: Checkout code + uses: actions/checkout@v4 + + - name: Setup Node.js + uses: actions/setup-node@v4 + with: + node-version: "22" + + - name: Restore node_modules cache + uses: actions/cache@v4 + with: + path: tools/server/webui/node_modules + key: ${{ runner.os }}-node-modules-${{ hashFiles('tools/server/webui/package-lock.json') }} + restore-keys: | + ${{ runner.os }}-node-modules- + + - name: Install Playwright browsers + run: npx playwright install --with-deps + working-directory: tools/server/webui + + - name: Build Storybook + run: npm run build-storybook + working-directory: tools/server/webui + + - name: Run Client tests + run: npm run test:client + working-directory: tools/server/webui + + - name: Run Server tests + run: npm run test:server + working-directory: tools/server/webui + + - name: Run UI tests + run: npm run test:ui + working-directory: tools/server/webui + + - name: Run E2E tests + run: npm run test:e2e + working-directory: tools/server/webui + + server-build: + needs: [webui-tests] + runs-on: ubuntu-latest + + strategy: + matrix: + sanitizer: [ADDRESS, UNDEFINED] # THREAD is broken + build_type: [RelWithDebInfo] + include: + - build_type: Release + sanitizer: "" + fail-fast: false # While -DLLAMA_SANITIZE_THREAD=ON is broken + + steps: + - name: Dependencies + id: depends run: | - cd tools/server/webui - npm ci + sudo apt-get update + sudo apt-get -y install \ + build-essential \ + xxd \ + git \ + cmake \ + curl \ + wget \ + language-pack-en \ + libcurl4-openssl-dev - - name: WebUI - Check code format - id: webui_format + - name: Clone + id: checkout + uses: actions/checkout@v4 + with: + fetch-depth: 0 + ref: ${{ github.event.inputs.sha || github.event.pull_request.head.sha || github.sha || github.head_ref || github.ref_name }} + + - name: Python setup + id: setup_python + uses: actions/setup-python@v5 + with: + python-version: '3.11' + + - name: Tests dependencies + id: test_dependencies run: | - git config --global --add safe.directory $(realpath .) - cd tools/server/webui - git status + pip install -r tools/server/tests/requirements.txt - npm run format - git status - modified_files="$(git status -s)" - echo "Modified files: ${modified_files}" - if [ -n "${modified_files}" ]; then - echo "Files do not follow coding style. To fix: npm run format" - echo "${modified_files}" - exit 1 - fi + - name: Setup Node.js for WebUI + uses: actions/setup-node@v4 + with: + node-version: "22" + cache: "npm" + cache-dependency-path: "tools/server/webui/package-lock.json" - - name: Verify bundled index.html - id: verify_server_index_html - run: | - git config --global --add safe.directory $(realpath .) - cd tools/server/webui - git status + - name: Install WebUI dependencies + run: npm ci + working-directory: tools/server/webui - npm run build - git status - modified_files="$(git status -s)" - echo "Modified files: ${modified_files}" - if [ -n "${modified_files}" ]; then - echo "Repository is dirty or server/webui is not built as expected" - echo "Hint: You may need to follow Web UI build guide in server/README.md" - echo "${modified_files}" - exit 1 - fi + - name: Build WebUI + run: npm run build + working-directory: tools/server/webui - name: Build (no OpenMP) id: cmake_build_no_openmp diff --git a/.gitignore b/.gitignore index 595831accb..e1ecdb7a81 100644 --- a/.gitignore +++ b/.gitignore @@ -148,3 +148,7 @@ poetry.toml /run-vim.sh /run-chat.sh .ccache/ + +# Code Workspace +*.code-workspace + diff --git a/.windsurf/rules/css-architecture.md b/.windsurf/rules/css-architecture.md new file mode 100644 index 0000000000..10a183585c --- /dev/null +++ b/.windsurf/rules/css-architecture.md @@ -0,0 +1,7 @@ +--- +trigger: manual +--- + +#### Tailwind & CSS + +- We are using Tailwind v4 which uses oklch colors so we now want to refer to the CSS vars directly, without wrapping it with any color function like `hsla/hsl`, `rgba` etc. diff --git a/.windsurf/rules/sveltekit-architecture.md b/.windsurf/rules/sveltekit-architecture.md new file mode 100644 index 0000000000..c2c0aa07ce --- /dev/null +++ b/.windsurf/rules/sveltekit-architecture.md @@ -0,0 +1,48 @@ +--- +trigger: manual +--- + +# Coding rules + +## Svelte & SvelteKit + +### Services vs Stores Separation Pattern + +#### `lib/services/` - Pure Business Logic + +- **Purpose**: Stateless business logic and external communication +- **Contains**: + - API calls to external services (ApiService) + - Pure business logic functions (ChatService, etc.) +- **Rules**: + - NO Svelte runes ($state, $derived, $effect) + - NO reactive state management + - Pure functions and classes only + - Can import types but not stores + - Focus on "how" - implementation details + +#### `lib/stores/` - Reactive State Management + +- **Purpose**: Svelte-specific reactive state with runes +- **Contains**: + - Reactive state classes with $state, $derived, $effect + - Database operations (DatabaseStore) + - UI-focused state management + - Store orchestration logic +- **Rules**: + - USE Svelte runes for reactivity + - Import and use services for business logic + - NO direct database operations + - NO direct API calls (use services) + - Focus on "what" - reactive state for UI + +#### Enforcement + +- Services should be testable without Svelte +- Stores should leverage Svelte's reactivity system +- Clear separation: services handle data, stores handle state +- Services can be reused across multiple stores + +#### Misc + +- Always use `let` for $derived state variables diff --git a/.windsurf/rules/tests.md b/.windsurf/rules/tests.md new file mode 100644 index 0000000000..e388fa9544 --- /dev/null +++ b/.windsurf/rules/tests.md @@ -0,0 +1,9 @@ +--- +trigger: manual +--- + +# Automated Tests + +## General rules + +- NEVER include any test code in the production code - we should always have it in a separate dedicated files diff --git a/.windsurf/rules/typescript-architecture.md b/.windsurf/rules/typescript-architecture.md new file mode 100644 index 0000000000..a61ff6b98f --- /dev/null +++ b/.windsurf/rules/typescript-architecture.md @@ -0,0 +1,7 @@ +--- +trigger: manual +--- + +## TypeScript + +- Add JSDocs for functions diff --git a/ci/run.sh b/ci/run.sh index ba099680cd..8e20b4e24d 100755 --- a/ci/run.sh +++ b/ci/run.sh @@ -45,7 +45,7 @@ SRC=`pwd` CMAKE_EXTRA="-DLLAMA_FATAL_WARNINGS=ON -DLLAMA_CURL=ON" if [ ! -z ${GG_BUILD_METAL} ]; then - CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_METAL=ON -DGGML_METAL_USE_BF16=ON" + CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_METAL=ON" fi if [ ! -z ${GG_BUILD_CUDA} ]; then diff --git a/convert_hf_to_gguf.py b/convert_hf_to_gguf.py index 15bb1cf233..85cb9b2142 100755 --- a/convert_hf_to_gguf.py +++ b/convert_hf_to_gguf.py @@ -2393,7 +2393,10 @@ class SmolVLMModel(MmprojModel): return [] # skip other tensors -@ModelBase.register("Llama4ForConditionalGeneration") +@ModelBase.register( + "Llama4ForConditionalGeneration", + "Llama4ForCausalLM", +) class Llama4Model(LlamaModel): model_arch = gguf.MODEL_ARCH.LLAMA4 undo_permute = False @@ -2411,6 +2414,10 @@ class Llama4Model(LlamaModel): super().set_gguf_parameters() self.gguf_writer.add_interleave_moe_layer_step(self.hparams["interleave_moe_layer_step"]) self.gguf_writer.add_expert_feed_forward_length(self.hparams["intermediate_size_moe"]) + if "layer_types" in self.hparams: + if all(lt == "full_attention" for lt in self.hparams["layer_types"]): + # all layers are full attention (for MobileLLM), disable swa + self.gguf_writer.add_sliding_window(0) def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None): if name.startswith("language_model."): diff --git a/ggml/CMakeLists.txt b/ggml/CMakeLists.txt index d06464f5eb..f4ccf273b7 100644 --- a/ggml/CMakeLists.txt +++ b/ggml/CMakeLists.txt @@ -190,7 +190,6 @@ option(GGML_WEBGPU "ggml: use WebGPU" option(GGML_WEBGPU_DEBUG "ggml: enable WebGPU debug output" OFF) option(GGML_ZDNN "ggml: use zDNN" OFF) option(GGML_METAL "ggml: use Metal" ${GGML_METAL_DEFAULT}) -option(GGML_METAL_USE_BF16 "ggml: use bfloat if available" OFF) option(GGML_METAL_NDEBUG "ggml: disable Metal debugging" OFF) option(GGML_METAL_SHADER_DEBUG "ggml: compile Metal with -fno-fast-math" OFF) option(GGML_METAL_EMBED_LIBRARY "ggml: embed Metal library" ${GGML_METAL}) diff --git a/ggml/include/ggml-metal.h b/ggml/include/ggml-metal.h index 1163438bc2..433838f0d6 100644 --- a/ggml/include/ggml-metal.h +++ b/ggml/include/ggml-metal.h @@ -39,6 +39,7 @@ extern "C" { // user-code should use only these functions // +// TODO: remove in the future GGML_BACKEND_API ggml_backend_t ggml_backend_metal_init(void); GGML_BACKEND_API bool ggml_backend_is_metal(ggml_backend_t backend); diff --git a/ggml/include/ggml.h b/ggml/include/ggml.h index f10f41b4be..e661ff42bc 100644 --- a/ggml/include/ggml.h +++ b/ggml/include/ggml.h @@ -284,19 +284,19 @@ __host__ __device__ constexpr inline void ggml_unused_vars_impl(Args&&...) noexc // GGML_TENSOR_LOCALS(size_t, nb1, src1, nb); // #define GGML_TENSOR_LOCALS_1(type, prefix, pointer, array) \ - const type prefix##0 = (pointer)->array[0]; \ + const type prefix##0 = (pointer) ? (pointer)->array[0] : 0; \ GGML_UNUSED(prefix##0); #define GGML_TENSOR_LOCALS_2(type, prefix, pointer, array) \ GGML_TENSOR_LOCALS_1 (type, prefix, pointer, array) \ - const type prefix##1 = (pointer)->array[1]; \ + const type prefix##1 = (pointer) ? (pointer)->array[1] : 0; \ GGML_UNUSED(prefix##1); #define GGML_TENSOR_LOCALS_3(type, prefix, pointer, array) \ GGML_TENSOR_LOCALS_2 (type, prefix, pointer, array) \ - const type prefix##2 = (pointer)->array[2]; \ + const type prefix##2 = (pointer) ? (pointer)->array[2] : 0; \ GGML_UNUSED(prefix##2); #define GGML_TENSOR_LOCALS(type, prefix, pointer, array) \ GGML_TENSOR_LOCALS_3 (type, prefix, pointer, array) \ - const type prefix##3 = (pointer)->array[3]; \ + const type prefix##3 = (pointer) ? (pointer)->array[3] : 0; \ GGML_UNUSED(prefix##3); #define GGML_TENSOR_UNARY_OP_LOCALS \ diff --git a/ggml/src/ggml-cann/ggml-cann.cpp b/ggml/src/ggml-cann/ggml-cann.cpp index 56d82b4af3..cbeafa6bc9 100755 --- a/ggml/src/ggml-cann/ggml-cann.cpp +++ b/ggml/src/ggml-cann/ggml-cann.cpp @@ -1728,7 +1728,6 @@ static bool ggml_cann_compute_forward(ggml_backend_cann_context& ctx, ggml_cann_get_rows(ctx, dst); break; case GGML_OP_SET_ROWS: - std::cout << "lcg GGML_OP_SET_ROWS"<< std::endl; ggml_cann_set_rows(ctx, dst); break; case GGML_OP_DUP: diff --git a/ggml/src/ggml-cuda/common.cuh b/ggml/src/ggml-cuda/common.cuh index b0feea3623..045c6d3006 100644 --- a/ggml/src/ggml-cuda/common.cuh +++ b/ggml/src/ggml-cuda/common.cuh @@ -75,6 +75,8 @@ #define GGML_CUDA_CC_IS_RDNA4(cc) (cc >= GGML_CUDA_CC_RDNA4) #define GGML_CUDA_CC_IS_GCN(cc) (cc > GGML_CUDA_CC_OFFSET_AMD && cc < GGML_CUDA_CC_CDNA1) #define GGML_CUDA_CC_IS_CDNA(cc) (cc >= GGML_CUDA_CC_CDNA1 && cc < GGML_CUDA_CC_RDNA1) +#define GGML_CUDA_CC_IS_CDNA1(cc) (cc >= GGML_CUDA_CC_CDNA1 && cc < GGML_CUDA_CC_CDNA2) +#define GGML_CUDA_CC_IS_CDNA2(cc) (cc >= GGML_CUDA_CC_CDNA2 && cc < GGML_CUDA_CC_CDNA3) #define GGML_CUDA_CC_IS_CDNA3(cc) (cc >= GGML_CUDA_CC_CDNA3 && cc < GGML_CUDA_CC_RDNA1) // Moore Threads @@ -325,6 +327,20 @@ static constexpr __device__ int ggml_cuda_get_physical_warp_size() { #endif // defined(GGML_USE_HIP) && (defined(__GFX9__) || defined(__GFX8__)) } +// Maximum number of bytes that can be copied in a single instruction. +static constexpr __device__ int ggml_cuda_get_max_cpy_bytes() { +#ifdef GGML_USE_HIP + return 16; +#else +#if __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA + return 16; +#else + return 8; +#endif // __CUDA_ARCH__ >= GGML_CUDA_CC_VOLTA +#endif // GGML_USE_HIP +} + + [[noreturn]] static __device__ void no_device_code( const char * file_name, const int line, const char * function_name, const int arch, const char * arch_list) { diff --git a/ggml/src/ggml-cuda/fattn-common.cuh b/ggml/src/ggml-cuda/fattn-common.cuh index b69f57d659..142a3a88d1 100644 --- a/ggml/src/ggml-cuda/fattn-common.cuh +++ b/ggml/src/ggml-cuda/fattn-common.cuh @@ -647,9 +647,7 @@ static __global__ void flash_attn_stream_k_fixup( } template // D == head size -#if !defined(GGML_USE_HIP) __launch_bounds__(D, 1) -#endif // !(defined(GGML_USE_HIP) static __global__ void flash_attn_combine_results( const float * __restrict__ VKQ_parts, const float2 * __restrict__ VKQ_meta, @@ -692,10 +690,7 @@ static __global__ void flash_attn_combine_results( float VKQ_numerator = 0.0f; float VKQ_denominator = 0.0f; for (int l = 0; l < parallel_blocks; ++l) { - const float diff = meta[l].x - kqmax; - float KQ_max_scale = expf(diff); - const uint32_t ftz_mask = 0xFFFFFFFF * (diff > SOFTMAX_FTZ_THRESHOLD); - *((uint32_t *) &KQ_max_scale) &= ftz_mask; + const float KQ_max_scale = expf(meta[l].x - kqmax); VKQ_numerator += KQ_max_scale * VKQ_parts[l*D + tid]; VKQ_denominator += KQ_max_scale * meta[l].y; @@ -836,11 +831,10 @@ void launch_fattn( CUDA_CHECK(cudaGetLastError()); } - int parallel_blocks = 1; - const dim3 block_dim(warp_size, nwarps, 1); int max_blocks_per_sm = 1; // Max. number of active blocks limited by occupancy. CUDA_CHECK(cudaOccupancyMaxActiveBlocksPerMultiprocessor(&max_blocks_per_sm, fattn_kernel, block_dim.x * block_dim.y * block_dim.z, nbytes_shared)); + int parallel_blocks = max_blocks_per_sm; dim3 blocks_num; if (stream_k) { @@ -862,9 +856,6 @@ void launch_fattn( GGML_ASSERT(K->ne[1] % KQ_row_granularity == 0); const int ntiles_KQ = K->ne[1] / KQ_row_granularity; // Max. number of parallel blocks limited by tensor size. - // parallel_blocks should be at least large enough to achieve max. occupancy for a single wave: - parallel_blocks = std::max((nsm * max_blocks_per_sm) / ntiles_total, 1); - // parallel_blocks must not be larger than what the tensor size allows: parallel_blocks = std::min(parallel_blocks, ntiles_KQ); diff --git a/ggml/src/ggml-cuda/fattn-tile.cu b/ggml/src/ggml-cuda/fattn-tile.cu index c6a399ce5d..a2d9951ea5 100644 --- a/ggml/src/ggml-cuda/fattn-tile.cu +++ b/ggml/src/ggml-cuda/fattn-tile.cu @@ -2,20 +2,30 @@ #include "fattn-common.cuh" #include "fattn-tile.cuh" -#define FATTN_TILE_NTHREADS 256 +// kq_stride == number of KQ rows to process per iteration +// kq_nbatch == number of K columns to load in parallel for KQ calculation static int fattn_tile_get_kq_stride_host(const int D, const int ncols, const int cc, const int warp_size) { if (GGML_CUDA_CC_IS_AMD(cc)) { + if (GGML_CUDA_CC_IS_RDNA(cc)) { + switch (D) { + case 64: + return 128; + case 128: + case 256: + return ncols <= 16 ? 128 : 64; + default: + GGML_ABORT("fatal error"); + return -1; + } + } switch (D) { case 64: - return 64; + return ncols == 32 ? 128 : 64; case 128: + return ncols == 32 ? 64 : 32; case 256: - if (GGML_CUDA_CC_IS_GCN(cc) || GGML_CUDA_CC_IS_CDNA(cc)) { - return ncols <= 16 ? 64 : 32; - } else { - return 64; - } + return 32; default: GGML_ABORT("fatal error"); return -1; @@ -49,24 +59,28 @@ static int fattn_tile_get_kq_stride_host(const int D, const int ncols, const int static constexpr __device__ int fattn_tile_get_kq_stride_device(int D, int ncols, int warp_size) { #ifdef GGML_USE_HIP +#ifdef RDNA switch (D) { case 64: - return 64; + return 128; case 128: -#if defined(GCN) || defined(CDNA) - return ncols <= 16 ? 64 : 32; -#else - return 64; -#endif // defined(GCN) || defined(CDNA) case 256: -#if defined(GCN) || defined(CDNA) - return ncols <= 16 ? 64 : 32; -#else - return 64; -#endif // defined(GCN) || defined(CDNA) + return ncols <= 16 ? 128 : 64; default: return -1; } +#else + switch (D) { + case 64: + return ncols == 32 ? 128 : 64; + case 128: + return ncols == 32 ? 64 : 32; + case 256: + return 32; + default: + return -1; + } +#endif // RDNA #else #ifdef FAST_FP16_AVAILABLE switch (D) { @@ -100,17 +114,8 @@ static constexpr __device__ int fattn_tile_get_kq_nbatch_device(int D, int ncols case 64: return 64; case 128: -#if defined(GCN) || defined(CDNA) - return ncols <= 16 ? 64 : 128; -#else - return 64; -#endif // defined(GCN) || defined(CDNA) case 256: -#if defined(GCN) || defined(CDNA) - return ncols <= 16 ? 64 : 128; -#else - return ncols <= 16 ? 64 : 256; -#endif // defined(GCN) || defined(CDNA) + return 128; default: return -1; } @@ -120,9 +125,8 @@ static constexpr __device__ int fattn_tile_get_kq_nbatch_device(int D, int ncols case 64: return 64; case 128: - return ncols <= 16 ? 128 : 64; case 256: - return ncols <= 16 ? 64 : 128; + return 128; default: return -1; } @@ -142,12 +146,27 @@ static constexpr __device__ int fattn_tile_get_kq_nbatch_device(int D, int ncols GGML_UNUSED_VARS(ncols, warp_size); } -template // D == head size -#ifdef GGML_USE_HIP -__launch_bounds__(FATTN_TILE_NTHREADS, 1) +static int fattn_tile_get_nthreads_host(const int cc, const int ncols) { + return 256; + GGML_UNUSED_VARS(cc, ncols); +} + +static constexpr __device__ int fattn_tile_get_nthreads_device(int ncols) { + return 256; + GGML_UNUSED(ncols); +} + +static constexpr __device__ int fattn_tile_get_occupancy_device(int ncols) { +#ifdef RDNA + return 3; #else -__launch_bounds__(FATTN_TILE_NTHREADS, 2) -#endif // GGML_USE_HIP + return ncols <= 16 ? 3 : 2; +#endif // RDNA + GGML_UNUSED(ncols); +} + +template // D == head size +__launch_bounds__(fattn_tile_get_nthreads_device(ncols), fattn_tile_get_occupancy_device(ncols)) static __global__ void flash_attn_tile( const char * __restrict__ Q, const char * __restrict__ K, @@ -193,7 +212,7 @@ static __global__ void flash_attn_tile( } constexpr int warp_size = 32; - constexpr int nwarps = FATTN_TILE_NTHREADS / warp_size; + constexpr int nwarps = fattn_tile_get_nthreads_device(ncols) / warp_size; constexpr int kq_stride = fattn_tile_get_kq_stride_device(D, ncols, warp_size); static_assert(kq_stride % warp_size == 0, "kq_stride not divisable by warp_size."); constexpr int kq_nbatch = fattn_tile_get_kq_nbatch_device(D, ncols, warp_size); @@ -206,90 +225,126 @@ static __global__ void flash_attn_tile( const int sequence = blockIdx.z / ne02; const int head = blockIdx.z - sequence*ne02; const int gqa_ratio = ne02 / ne12; // With grouped query attention there are > 1 Q matrices per K, V matrix. - const float2 * Q_f2 = (const float2 *) (Q + nb03* sequence + nb02* head + nb01*ic0); - const half2 * K_h2 = (const half2 *) (K + nb13* sequence + nb12*(head / gqa_ratio)); - const half2 * V_h2 = (const half2 *) (V + nb13* sequence + nb12*(head / gqa_ratio)); // K and V have same shape - const half * maskh = (const half *) (mask + nb33*(sequence % ne33) + nb31*ic0); - const float * sinksf = (const float *) (sinks); + const float * Q_f = (const float *) (Q + nb03* sequence + nb02* head + nb01*ic0); + const half2 * K_h2 = (const half2 *) (K + nb13* sequence + nb12*(head / gqa_ratio)); + const half2 * V_h2 = (const half2 *) (V + nb13* sequence + nb12*(head / gqa_ratio)); // K and V have same shape + const half * maskh = (const half *) (mask + nb33*(sequence % ne33) + nb31*ic0); + const float * sinksf = (const float *) (sinks); const int stride_KV2 = nb11 / sizeof(half2); const float slope = get_alibi_slope(max_bias, head, n_head_log2, m0, m1); -#if defined(GGML_USE_HIP) - constexpr int cpy_nb = 16; -#else - constexpr int cpy_nb = 8; -#endif // defined(GGML_USE_HIP) && defined(GCN) + constexpr int cpy_nb = ggml_cuda_get_max_cpy_bytes(); constexpr int cpy_ne = cpy_nb / 4; - __shared__ float KQ[ncols][kq_stride]; + constexpr int cpw = ncols/nwarps; // cols per warp + + // softmax_iter_j == number of KQ columns for which to calculate softmax in parallel. + // KQ is originall 2D but uses a Z-shaped memory pattern for larger reads/writes. #ifdef FAST_FP16_AVAILABLE + constexpr int softmax_iter_j = cpw < 2*cpy_ne ? cpw : 2*cpy_ne; + + __shared__ half KQ[ncols/softmax_iter_j][kq_stride][softmax_iter_j]; __shared__ half2 Q_tmp[ncols][D/2]; - __shared__ half2 KV_tmp_h2[kq_stride * (kq_nbatch/2 + cpy_ne)]; // Padded to avoid memory bank conflicts. - half2 VKQ[ncols/nwarps][D/(2*warp_size)] = {{{0.0f, 0.0f}}}; + __shared__ half2 KV_tmp[kq_stride * (kq_nbatch/2 + cpy_ne)]; // Padded to avoid memory bank conflicts. + half2 VKQ[cpw][D/(2*warp_size)] = {{{0.0f, 0.0f}}}; #else + constexpr int softmax_iter_j = cpw < 1*cpy_ne ? cpw : 1*cpy_ne; + + __shared__ float KQ[ncols/softmax_iter_j][kq_stride][softmax_iter_j]; __shared__ float Q_tmp[ncols][D]; - __shared__ float KV_tmp_f[kq_stride * (kq_nbatch + cpy_ne)]; // Padded to avoid memory bank conflicts. - float2 * KV_tmp_f2 = (float2 *) KV_tmp_f; - float2 VKQ[ncols/nwarps][D/(2*warp_size)] = {{{0.0f, 0.0f}}}; + __shared__ float KV_tmp[kq_stride * (kq_nbatch + cpy_ne)]; // Padded to avoid memory bank conflicts. + float2 VKQ[cpw][D/(2*warp_size)] = {{{0.0f, 0.0f}}}; #endif // FAST_FP16_AVAILABLE + static_assert(cpw % softmax_iter_j == 0, "bad softmax_iter_j"); - - float kqmax[ncols/nwarps]; + float KQ_max[cpw]; #pragma unroll for (int j0 = 0; j0 < ncols; j0 += nwarps) { - kqmax[j0/nwarps] = -FLT_MAX/2.0f; + KQ_max[j0/nwarps] = -FLT_MAX/2.0f; } - float kqsum[ncols/nwarps] = {0.0f}; + float KQ_sum[cpw] = {0.0f}; + + // Load Q data, convert to FP16 if fast. +#pragma unroll + for (int j0 = 0; j0 < cpw; ++j0) { + const int j = j0 + threadIdx.y*cpw; + + constexpr int cpy_ne_D = cpy_ne < D/warp_size ? cpy_ne : D/warp_size; #pragma unroll - for (int j0 = 0; j0 < ncols; j0 += nwarps) { - const int j = j0 + threadIdx.y; + for (int i0 = 0; i0 < D; i0 += warp_size*cpy_ne_D) { + float tmp_f[cpy_ne_D] = {0.0f}; + if (ic0 + j < ne01) { + ggml_cuda_memcpy_1(tmp_f, &Q_f[j*(nb01/sizeof(float)) + i0 + threadIdx.x*cpy_ne_D]); + } #pragma unroll - for (int i0 = 0; i0 < D/2; i0 += warp_size) { - const float2 tmp = ic0 + j < ne01 ? Q_f2[j*(nb01/sizeof(float2)) + i0 + threadIdx.x] : make_float2(0.0f, 0.0f); + for (int i1 = 0; i1 < cpy_ne_D; ++i1) { + tmp_f[i1] *= scale; + } + #ifdef FAST_FP16_AVAILABLE - Q_tmp[j][i0 + threadIdx.x] = make_half2(tmp.x * scale, tmp.y * scale); + half2 tmp_h2[cpy_ne_D/2]; +#pragma unroll + for (int i1 = 0; i1 < cpy_ne_D; i1 += 2) { + tmp_h2[i1/2] = make_half2(tmp_f[i1 + 0], tmp_f[i1 + 1]); + } + ggml_cuda_memcpy_1(&Q_tmp[j][i0/2 + threadIdx.x*(cpy_ne_D/2)], tmp_h2); #else - Q_tmp[j][2*i0 + threadIdx.x] = tmp.x * scale; - Q_tmp[j][2*i0 + warp_size + threadIdx.x] = tmp.y * scale; + ggml_cuda_memcpy_1 (&Q_tmp[j][i0 + threadIdx.x* cpy_ne_D], tmp_f); #endif // FAST_FP16_AVAILABLE } } __syncthreads(); + // Main loop over KV cache: const int k_VKQ_max = KV_max ? KV_max[sequence*gridDim.x + blockIdx.x] : ne11; for (int k_VKQ_0 = blockIdx.y*kq_stride; k_VKQ_0 < k_VKQ_max; k_VKQ_0 += gridDim.y*kq_stride) { // Calculate KQ tile and keep track of new maximum KQ values: - float kqmax_new[ncols/nwarps]; + float KQ_max_new[cpw]; #pragma unroll - for (int j = 0; j < ncols/nwarps; ++j) { - kqmax_new[j] = kqmax[j]; + for (int j = 0; j < cpw; ++j) { + KQ_max_new[j] = KQ_max[j]; } - float sum[kq_stride/warp_size][ncols/nwarps] = {{0.0f}}; + float KQ_acc[kq_stride/warp_size][cpw] = {{0.0f}}; // Accumulators for KQ matrix multiplication. + // KQ = K @ Q matrix multiplication: #pragma unroll for (int k_KQ_0 = 0; k_KQ_0 < D; k_KQ_0 += kq_nbatch) { #pragma unroll for (int i_KQ_0 = 0; i_KQ_0 < kq_stride; i_KQ_0 += nwarps) { const int i_KQ = i_KQ_0 + threadIdx.y; -#pragma unroll - for (int k_KQ_1 = 0; k_KQ_1 < kq_nbatch/2; k_KQ_1 += warp_size) { - const half2 tmp_h2 = K_h2[int64_t(k_VKQ_0 + i_KQ)*stride_KV2 + k_KQ_0/2 + k_KQ_1 + threadIdx.x]; #ifdef FAST_FP16_AVAILABLE - KV_tmp_h2[i_KQ*(kq_nbatch/2 + cpy_ne) + k_KQ_1 + threadIdx.x] = tmp_h2; -#else - const float2 tmp_f2 = __half22float2(tmp_h2); - KV_tmp_f[i_KQ*(kq_nbatch + cpy_ne) + 2*k_KQ_1 + threadIdx.x] = tmp_f2.x; - KV_tmp_f[i_KQ*(kq_nbatch + cpy_ne) + 2*k_KQ_1 + warp_size + threadIdx.x] = tmp_f2.y; -#endif // FAST_FP16_AVAILABLE + constexpr int cpy_ne_kqnb = cpy_ne < kq_nbatch/(2*warp_size) ? cpy_ne : kq_nbatch/(2*warp_size); +#pragma unroll + for (int k_KQ_1 = 0; k_KQ_1 < kq_nbatch/2; k_KQ_1 += warp_size*cpy_ne_kqnb) { + ggml_cuda_memcpy_1( + &KV_tmp[i_KQ*(kq_nbatch/2 + cpy_ne) + k_KQ_1 + threadIdx.x*cpy_ne_kqnb], + &K_h2[int64_t(k_VKQ_0 + i_KQ)*stride_KV2 + k_KQ_0/2 + k_KQ_1 + threadIdx.x*cpy_ne_kqnb]); } +#else + constexpr int cpy_ne_kqnb = cpy_ne < kq_nbatch/warp_size ? cpy_ne : kq_nbatch/warp_size; +#pragma unroll + for (int k_KQ_1 = 0; k_KQ_1 < kq_nbatch; k_KQ_1 += warp_size*cpy_ne_kqnb) { + half2 tmp_h2[cpy_ne_kqnb/2]; + ggml_cuda_memcpy_1( + tmp_h2, &K_h2[int64_t(k_VKQ_0 + i_KQ)*stride_KV2 + k_KQ_0/2 + k_KQ_1/2 + threadIdx.x*(cpy_ne_kqnb/2)]); + + float2 tmp_f2[cpy_ne_kqnb/2]; +#pragma unroll + for (int k_KQ_2 = 0; k_KQ_2 < cpy_ne_kqnb/2; ++k_KQ_2) { + tmp_f2[k_KQ_2] = __half22float2(tmp_h2[k_KQ_2]); + } + ggml_cuda_memcpy_1( + &KV_tmp[i_KQ*(kq_nbatch + cpy_ne) + k_KQ_1 + threadIdx.x*cpy_ne_kqnb], tmp_f2); + } +#endif // FAST_FP16_AVAILABLE } __syncthreads(); @@ -298,12 +353,12 @@ static __global__ void flash_attn_tile( #pragma unroll for (int k_KQ_1 = 0; k_KQ_1 < kq_nbatch/2; k_KQ_1 += cpy_ne) { half2 K_k[kq_stride/warp_size][cpy_ne]; - half2 Q_k[ncols/nwarps][cpy_ne]; + half2 Q_k[cpw][cpy_ne]; #else #pragma unroll for (int k_KQ_1 = 0; k_KQ_1 < kq_nbatch; k_KQ_1 += cpy_ne) { float K_k[kq_stride/warp_size][cpy_ne]; - float Q_k[ncols/nwarps][cpy_ne]; + float Q_k[cpw][cpy_ne]; #endif // FAST_FP16_AVAILABLE #pragma unroll @@ -311,29 +366,29 @@ static __global__ void flash_attn_tile( const int i_KQ = i_KQ_0 + threadIdx.x; #ifdef FAST_FP16_AVAILABLE - ggml_cuda_memcpy_1(&K_k[i_KQ_0/warp_size], &KV_tmp_h2[i_KQ*(kq_nbatch/2 + cpy_ne) + k_KQ_1]); + ggml_cuda_memcpy_1(&K_k[i_KQ_0/warp_size], &KV_tmp[i_KQ*(kq_nbatch/2 + cpy_ne) + k_KQ_1]); #else - ggml_cuda_memcpy_1(&K_k[i_KQ_0/warp_size], &KV_tmp_f [i_KQ*(kq_nbatch + cpy_ne) + k_KQ_1]); + ggml_cuda_memcpy_1(&K_k[i_KQ_0/warp_size], &KV_tmp[i_KQ*(kq_nbatch + cpy_ne) + k_KQ_1]); #endif // FAST_FP16_AVAILABLE } #pragma unroll - for (int j_KQ_0 = 0; j_KQ_0 < ncols; j_KQ_0 += nwarps) { - const int j_KQ = j_KQ_0 + threadIdx.y; + for (int j_KQ_0 = 0; j_KQ_0 < cpw; ++j_KQ_0) { + const int j_KQ = j_KQ_0 + threadIdx.y*cpw; #ifdef FAST_FP16_AVAILABLE - ggml_cuda_memcpy_1(&Q_k[j_KQ_0/nwarps], &Q_tmp[j_KQ][k_KQ_0/2 + k_KQ_1]); + ggml_cuda_memcpy_1(&Q_k[j_KQ_0], &Q_tmp[j_KQ][k_KQ_0/2 + k_KQ_1]); #else - ggml_cuda_memcpy_1(&Q_k[j_KQ_0/nwarps], &Q_tmp[j_KQ][k_KQ_0 + k_KQ_1]); + ggml_cuda_memcpy_1(&Q_k[j_KQ_0], &Q_tmp[j_KQ][k_KQ_0 + k_KQ_1]); #endif // FAST_FP16_AVAILABLE } #pragma unroll for (int i_KQ_0 = 0; i_KQ_0 < kq_stride; i_KQ_0 += warp_size) { #pragma unroll - for (int j_KQ_0 = 0; j_KQ_0 < ncols; j_KQ_0 += nwarps) { + for (int j_KQ_0 = 0; j_KQ_0 < cpw; ++j_KQ_0) { #pragma unroll for (int k = 0; k < cpy_ne; ++k) { - ggml_cuda_mad(sum[i_KQ_0/warp_size][j_KQ_0/nwarps], K_k[i_KQ_0/warp_size][k], Q_k[j_KQ_0/nwarps][k]); + ggml_cuda_mad(KQ_acc[i_KQ_0/warp_size][j_KQ_0], K_k[i_KQ_0/warp_size][k], Q_k[j_KQ_0][k]); } } } @@ -344,104 +399,77 @@ static __global__ void flash_attn_tile( } } + // Apply logit softcap, mask, update KQ_max: #pragma unroll for (int i_KQ_0 = 0; i_KQ_0 < kq_stride; i_KQ_0 += warp_size) { const int i_KQ = i_KQ_0 + threadIdx.x; #pragma unroll - for (int j_KQ_0 = 0; j_KQ_0 < ncols; j_KQ_0 += nwarps) { - const int j_KQ = j_KQ_0 + threadIdx.y; + for (int j_KQ_0 = 0; j_KQ_0 < cpw; ++j_KQ_0) { + const int j_KQ = j_KQ_0 + threadIdx.y*cpw; if (use_logit_softcap) { - sum[i_KQ_0/warp_size][j_KQ_0/nwarps] = logit_softcap * tanhf(sum[i_KQ_0/warp_size][j_KQ_0/nwarps]); + KQ_acc[i_KQ_0/warp_size][j_KQ_0] = logit_softcap * tanhf(KQ_acc[i_KQ_0/warp_size][j_KQ_0]); } - sum[i_KQ_0/warp_size][j_KQ_0/nwarps] += mask ? slope*__half2float(maskh[j_KQ*ne11 + k_VKQ_0 + i_KQ]) : 0.0f; + KQ_acc[i_KQ_0/warp_size][j_KQ_0] += mask ? slope*__half2float(maskh[j_KQ*ne11 + k_VKQ_0 + i_KQ]) : 0.0f; - kqmax_new[j_KQ_0/nwarps] = fmaxf(kqmax_new[j_KQ_0/nwarps], sum[i_KQ_0/warp_size][j_KQ_0/nwarps]); - - KQ[j_KQ][i_KQ] = sum[i_KQ_0/warp_size][j_KQ_0/nwarps]; + KQ_max_new[j_KQ_0] = fmaxf(KQ_max_new[j_KQ_0], KQ_acc[i_KQ_0/warp_size][j_KQ_0]); } } __syncthreads(); + // Calculate KQ softmax, write to shared KQ buffer, re-scale VKQ accumulators: #pragma unroll - for (int j0 = 0; j0 < ncols; j0 += nwarps) { - const int j = j0 + threadIdx.y; - - kqmax_new[j0/nwarps] = warp_reduce_max(kqmax_new[j0/nwarps]); - const float KQ_max_scale = expf(kqmax[j0/nwarps] - kqmax_new[j0/nwarps]); - kqmax[j0/nwarps] = kqmax_new[j0/nwarps]; - - float kqsum_add = 0.0f; - if (kq_stride % (4*warp_size) == 0 && cpy_ne % 4 == 0) { -#pragma unroll - for (int i0 = 0; i0 < kq_stride; i0 += 4*warp_size) { - const int i = i0 + 4*threadIdx.x; - - float4 val = *(const float4 *) &KQ[j][i]; - val.x = expf(val.x - kqmax[j0/nwarps]); - val.y = expf(val.y - kqmax[j0/nwarps]); - val.z = expf(val.z - kqmax[j0/nwarps]); - val.w = expf(val.w - kqmax[j0/nwarps]); - kqsum_add += val.x + val.y + val.z + val.w; - + for (int j0 = 0; j0 < cpw; j0 += softmax_iter_j) { #ifdef FAST_FP16_AVAILABLE - const half2 tmp[2] = {make_half2(val.x, val.y), make_half2(val.z, val.w)}; - ggml_cuda_memcpy_1(&KQ[j][i/2], &tmp); + half tmp[kq_stride/warp_size][softmax_iter_j]; #else - ggml_cuda_memcpy_1(&KQ[j][i], &val); + float tmp[kq_stride/warp_size][softmax_iter_j]; #endif // FAST_FP16_AVAILABLE - } - } else if (kq_stride % (2*warp_size) == 0 && cpy_ne % 2 == 0) { -#pragma unroll - for (int i0 = 0; i0 < kq_stride; i0 += 2*warp_size) { - const int i = i0 + 2*threadIdx.x; - float2 val = *(const float2 *) &KQ[j][i]; - val.x = expf(val.x - kqmax[j0/nwarps]); - val.y = expf(val.y - kqmax[j0/nwarps]); - kqsum_add += val.x + val.y; -#ifdef FAST_FP16_AVAILABLE - const half2 tmp = make_half2(val.x, val.y); - ggml_cuda_memcpy_1(&KQ[j][i/2], &tmp); -#else - ggml_cuda_memcpy_1(&KQ[j][i], &val); -#endif // FAST_FP16_AVAILABLE - } - } else { +#pragma unroll + for (int j1 = 0; j1 < softmax_iter_j; ++j1) { + KQ_max_new[j0+j1] = warp_reduce_max(KQ_max_new[j0+j1]); + const float KQ_max_scale = expf(KQ_max[j0+j1] - KQ_max_new[j0+j1]); + KQ_max[j0+j1] = KQ_max_new[j0+j1]; + + float KQ_sum_add = 0.0f; +#pragma unroll for (int i0 = 0; i0 < kq_stride; i0 += warp_size) { - const int i = i0 + threadIdx.x; - - const float diff = KQ[j][i] - kqmax[j0/nwarps]; - const float val = expf(diff); - kqsum_add += val; -#ifdef FAST_FP16_AVAILABLE - ((half *) KQ[j])[i] = val; -#else - KQ[j][i] = val; -#endif // FAST_FP16_AVAILABLE + const float val = expf(KQ_acc[i0/warp_size][j0+j1] - KQ_max[j0+j1]); + KQ_sum_add += val; + tmp[i0/warp_size][j1] = val; } - } - kqsum[j0/nwarps] = kqsum[j0/nwarps]*KQ_max_scale + kqsum_add; + KQ_sum[j0+j1] = KQ_sum[j0+j1]*KQ_max_scale + KQ_sum_add; #ifdef FAST_FP16_AVAILABLE - const half2 KQ_max_scale_h2 = make_half2(KQ_max_scale, KQ_max_scale); + const half2 KQ_max_scale_h2 = make_half2(KQ_max_scale, KQ_max_scale); #pragma unroll - for (int i0 = 0; i0 < D/2; i0 += warp_size) { - VKQ[j0/nwarps][i0/warp_size] *= KQ_max_scale_h2; - } + for (int i0 = 0; i0 < D/2; i0 += warp_size) { + VKQ[j0+j1][i0/warp_size] *= KQ_max_scale_h2; + } #else #pragma unroll - for (int i0 = 0; i0 < D/2; i0 += warp_size) { - VKQ[j0/nwarps][i0/warp_size].x *= KQ_max_scale; - VKQ[j0/nwarps][i0/warp_size].y *= KQ_max_scale; - } + for (int i0 = 0; i0 < D/2; i0 += warp_size) { + VKQ[j0+j1][i0/warp_size].x *= KQ_max_scale; + VKQ[j0+j1][i0/warp_size].y *= KQ_max_scale; + } #endif // FAST_FP16_AVAILABLE + } + +#pragma unroll + for (int i0 = 0; i0 < kq_stride; i0 += warp_size) { + const int i = i0 + threadIdx.x; + + ggml_cuda_memcpy_1( + KQ[j0/softmax_iter_j + threadIdx.y*(cpw/softmax_iter_j)][i], tmp[i0/warp_size]); + } } - constexpr int V_cols_per_iter = kq_stride*kq_nbatch / D; + // VKQ = V @ KQ matrix multiplication: + constexpr int V_cols_per_iter = kq_stride*kq_nbatch / D; // Number of V columns that fit in SRAM for K. static_assert(kq_stride % V_cols_per_iter == 0, "bad V_cols_per_iter"); #pragma unroll for (int k0 = 0; k0 < kq_stride; k0 += V_cols_per_iter) { @@ -449,65 +477,96 @@ static __global__ void flash_attn_tile( for (int k1 = 0; k1 < V_cols_per_iter; k1 += nwarps) { const int k_tile = k1 + threadIdx.y; -#pragma unroll - for (int i0 = 0; i0 < D/2; i0 += warp_size) { - const int i = i0 + threadIdx.x; - - const half2 tmp = V_h2[int64_t(k_VKQ_0 + k0 + k_tile)*stride_KV2 + i]; #ifdef FAST_FP16_AVAILABLE - KV_tmp_h2[k_tile*(D/2) + i] = tmp; -#else - KV_tmp_f2[k_tile*(D/2) + i] = __half22float2(tmp); -#endif // FAST_FP16_AVAILABLE + constexpr int cpy_ne_D = cpy_ne < D/(2*warp_size) ? cpy_ne : D/(2*warp_size); +#pragma unroll + for (int i0 = 0; i0 < D/2; i0 += warp_size*cpy_ne_D) { + ggml_cuda_memcpy_1( + &KV_tmp[k_tile*(D/2) + i0 + threadIdx.x*cpy_ne_D], + &V_h2[int64_t(k_VKQ_0 + k0 + k_tile)*stride_KV2 + i0 + threadIdx.x*cpy_ne_D]); } +#else + constexpr int cpy_ne_D = cpy_ne < D/warp_size ? cpy_ne : D/warp_size; +#pragma unroll + for (int i0 = 0; i0 < D; i0 += warp_size*cpy_ne_D) { + half2 tmp_h2[cpy_ne_D/2]; + ggml_cuda_memcpy_1( + tmp_h2, &V_h2[int64_t(k_VKQ_0 + k0 + k_tile)*stride_KV2 + i0/2 + threadIdx.x*(cpy_ne_D/2)]); + + float2 tmp_f2[cpy_ne_D/2]; +#pragma unroll + for (int i1 = 0; i1 < cpy_ne_D/2; ++i1) { + tmp_f2[i1] = __half22float2(tmp_h2[i1]); + } + ggml_cuda_memcpy_1( + &KV_tmp[k_tile*D + i0 + threadIdx.x*cpy_ne_D], tmp_f2); + } +#endif // FAST_FP16_AVAILABLE } __syncthreads(); +#ifdef FAST_FP16_AVAILABLE #pragma unroll for (int k1 = 0; k1 < V_cols_per_iter; ++k1) { -#ifdef FAST_FP16_AVAILABLE half2 V_k[(D/2)/warp_size]; - half2 KQ_k[ncols/nwarps]; -#else - float2 V_k[(D/2)/warp_size]; - float KQ_k[ncols/nwarps]; -#endif // FAST_FP16_AVAILABLE + half2 KQ_k[cpw]; + constexpr int cpy_ne_D = cpy_ne/2 < (D/2)/warp_size ? cpy_ne/2 : (D/2)/warp_size; #pragma unroll - for (int i0 = 0; i0 < D/2; i0 += warp_size) { - const int i = i0 + threadIdx.x; - -#ifdef FAST_FP16_AVAILABLE - V_k[i0/warp_size] = KV_tmp_h2[k1*(D/2) + i]; -#else - V_k[i0/warp_size] = KV_tmp_f2[k1*(D/2) + i]; -#endif // FAST_FP16_AVAILABLE + for (int i0 = 0; i0 < D/2; i0 += warp_size*cpy_ne_D) { + ggml_cuda_memcpy_1(&V_k[i0/warp_size], &KV_tmp[k1*(D/2) + i0 + threadIdx.x*cpy_ne_D]); } #pragma unroll - for (int j0 = 0; j0 < ncols; j0 += nwarps) { - const int j = j0 + threadIdx.y; + for (int j0 = 0; j0 < cpw; j0 += softmax_iter_j) { + const int j = j0/softmax_iter_j + threadIdx.y*(cpw/softmax_iter_j); -#ifdef FAST_FP16_AVAILABLE - KQ_k[j0/nwarps] = __half2half2(((const half *)KQ[j])[k0 + k1]); -#else - KQ_k[j0/nwarps] = KQ[j][k0 + k1]; -#endif // FAST_FP16_AVAILABLE + half tmp[softmax_iter_j]; + ggml_cuda_memcpy_1( + &tmp, KQ[j][k0 + k1]); +#pragma unroll + for (int j1 = 0; j1 < softmax_iter_j; ++j1) { + KQ_k[j0+j1] = __half2half2(tmp[j1]); + } } #pragma unroll for (int i0 = 0; i0 < D/2; i0 += warp_size) { #pragma unroll - for (int j0 = 0; j0 < ncols; j0 += nwarps) { -#ifdef FAST_FP16_AVAILABLE - VKQ[j0/nwarps][i0/warp_size] += V_k[i0/warp_size] *KQ_k[j0/nwarps]; -#else - VKQ[j0/nwarps][i0/warp_size].x += V_k[i0/warp_size].x*KQ_k[j0/nwarps]; - VKQ[j0/nwarps][i0/warp_size].y += V_k[i0/warp_size].y*KQ_k[j0/nwarps]; -#endif // FAST_FP16_AVAILABLE + for (int j0 = 0; j0 < cpw; ++j0) { + VKQ[j0][i0/warp_size] += V_k[i0/warp_size]*KQ_k[j0]; } } } +#else +#pragma unroll + for (int k1 = 0; k1 < V_cols_per_iter; ++k1) { + float2 V_k[(D/2)/warp_size]; + float KQ_k[cpw]; + + constexpr int cpy_ne_D = cpy_ne < D/warp_size ? cpy_ne : D/warp_size; +#pragma unroll + for (int i0 = 0; i0 < D; i0 += warp_size*cpy_ne_D) { + ggml_cuda_memcpy_1(&V_k[i0/(2*warp_size)], &KV_tmp[k1*D + i0 + threadIdx.x*cpy_ne_D]); + } +#pragma unroll + for (int j0 = 0; j0 < cpw; j0 += softmax_iter_j) { + const int j = j0/softmax_iter_j + threadIdx.y*(cpw/softmax_iter_j); + + ggml_cuda_memcpy_1( + &KQ_k[j0], KQ[j][k0 + k1]); + } + +#pragma unroll + for (int i0 = 0; i0 < D/2; i0 += warp_size) { +#pragma unroll + for (int j0 = 0; j0 < cpw; ++j0) { + VKQ[j0][i0/warp_size].x += V_k[i0/warp_size].x*KQ_k[j0]; + VKQ[j0][i0/warp_size].y += V_k[i0/warp_size].y*KQ_k[j0]; + } + } + } +#endif // FAST_FP16_AVAILABLE __syncthreads(); } @@ -519,69 +578,92 @@ static __global__ void flash_attn_tile( const float sink = sinksf[head]; #pragma unroll - for (int j0 = 0; j0 < ncols; j0 += nwarps) { - float kqmax_new_j = fmaxf(kqmax[j0/nwarps], sink); - kqmax_new_j = warp_reduce_max(kqmax_new_j); + for (int j0 = 0; j0 < cpw; ++j0) { + float KQ_max_new_j = fmaxf(KQ_max[j0], sink); + KQ_max_new_j = warp_reduce_max(KQ_max_new_j); - const float KQ_max_scale = expf(kqmax[j0/nwarps] - kqmax_new_j); - kqmax[j0/nwarps] = kqmax_new_j; + const float KQ_max_scale = expf(KQ_max[j0] - KQ_max_new_j); + KQ_max[j0] = KQ_max_new_j; - const float val = expf(sink - kqmax[j0/nwarps]); - kqsum[j0/nwarps] = kqsum[j0/nwarps] * KQ_max_scale; + const float val = expf(sink - KQ_max[j0]); + KQ_sum[j0] = KQ_sum[j0] * KQ_max_scale; if (threadIdx.x == 0) { - kqsum[j0/nwarps] += val; + KQ_sum[j0] += val; } #ifdef FAST_FP16_AVAILABLE const half2 KQ_max_scale_h2 = make_half2(KQ_max_scale, KQ_max_scale); #pragma unroll for (int i0 = 0; i0 < D/2; i0 += warp_size) { - VKQ[j0/nwarps][i0/warp_size] *= KQ_max_scale_h2; + VKQ[j0][i0/warp_size] *= KQ_max_scale_h2; } #else #pragma unroll for (int i0 = 0; i0 < D/2; i0 += warp_size) { - VKQ[j0/nwarps][i0/warp_size].x *= KQ_max_scale; - VKQ[j0/nwarps][i0/warp_size].y *= KQ_max_scale; + VKQ[j0][i0/warp_size].x *= KQ_max_scale; + VKQ[j0][i0/warp_size].y *= KQ_max_scale; } #endif // FAST_FP16_AVAILABLE } } - float2 * dst2 = (float2 *) dst; - #pragma unroll - for (int j_VKQ_0 = 0; j_VKQ_0 < ncols; j_VKQ_0 += nwarps) { - const int j_VKQ = j_VKQ_0 + threadIdx.y; + for (int j_VKQ_0 = 0; j_VKQ_0 < cpw; ++j_VKQ_0) { + KQ_sum[j_VKQ_0] = warp_reduce_sum(KQ_sum[j_VKQ_0]); + } + if (gridDim.y == 1) { +#pragma unroll + for (int j_VKQ_0 = 0; j_VKQ_0 < cpw; ++j_VKQ_0) { +#ifdef FAST_FP16_AVAILABLE + const half2 KQ_sum_j_inv = make_half2(1.0f/KQ_sum[j_VKQ_0], 1.0f/KQ_sum[j_VKQ_0]); +#pragma unroll + for (int i = 0; i < (D/2)/warp_size; ++i) { + VKQ[j_VKQ_0][i] *= KQ_sum_j_inv; + } +#else + const float KQ_sum_j_inv = 1.0f/KQ_sum[j_VKQ_0]; +#pragma unroll + for (int i = 0; i < (D/2)/warp_size; ++i) { + VKQ[j_VKQ_0][i].x *= KQ_sum_j_inv; + VKQ[j_VKQ_0][i].y *= KQ_sum_j_inv; + } +#endif // FAST_FP16_AVAILABLE + } + } + + // Write back results: +#pragma unroll + for (int j_VKQ_0 = 0; j_VKQ_0 < cpw; ++j_VKQ_0) { + const int j_VKQ = j_VKQ_0 + threadIdx.y*cpw; if (ic0 + j_VKQ >= ne01) { return; } - float kqsum_j = kqsum[j_VKQ_0/nwarps]; - kqsum_j = warp_reduce_sum(kqsum_j); - const int j_dst_unrolled = ((sequence*ne01 + ic0 + j_VKQ)*ne02 + head)*gridDim.y + blockIdx.y; -#pragma unroll - for (int i00 = 0; i00 < D/2; i00 += warp_size) { - const int i0 = i00 + threadIdx.x; - #ifdef FAST_FP16_AVAILABLE - float2 dst_val = __half22float2(VKQ[j_VKQ_0/nwarps][i0/warp_size]); + constexpr int cpy_ne_D = cpy_ne/2 < (D/2)/warp_size ? cpy_ne/2 : (D/2)/warp_size; +#pragma unroll + for (int i0 = 0; i0 < D/2; i0 += warp_size*cpy_ne_D) { + float2 tmp[cpy_ne_D]; +#pragma unroll + for (int i1 = 0; i1 < cpy_ne_D; ++i1) { + tmp[i1] = __half22float2(VKQ[j_VKQ_0][i0/warp_size + i1]); + } + ggml_cuda_memcpy_1(&dst[j_dst_unrolled*D + 2*i0 + threadIdx.x*(2*cpy_ne_D)], tmp); + } #else - float2 dst_val = VKQ[j_VKQ_0/nwarps][i0/warp_size]; + constexpr int cpy_ne_D = cpy_ne < D/warp_size ? cpy_ne : D/warp_size; +#pragma unroll + for (int i0 = 0; i0 < D; i0 += warp_size*cpy_ne_D) { + ggml_cuda_memcpy_1( + &dst[j_dst_unrolled*D + i0 + threadIdx.x*cpy_ne_D], &VKQ[j_VKQ_0][i0/(2*warp_size)]); + } #endif // FAST_FP16_AVAILABLE - if (gridDim.y == 1) { - dst_val.x /= kqsum_j; - dst_val.y /= kqsum_j; - } - dst2[j_dst_unrolled*(D/2) + i0] = dst_val; - } - if (gridDim.y != 1 && threadIdx.x == 0) { - dst_meta[j_dst_unrolled] = make_float2(kqmax[j_VKQ_0/nwarps], kqsum_j); + dst_meta[j_dst_unrolled] = make_float2(KQ_max[j_VKQ_0], KQ_sum[j_VKQ_0]); } } #else @@ -602,15 +684,29 @@ template static void launch_fattn_tile_switch_ncols(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { const ggml_tensor * Q = dst->src[0]; - const int id = ggml_cuda_get_device(); - const int cc = ggml_cuda_info().devices[id].cc; - const int warp_size = 32; - const int nwarps = FATTN_TILE_NTHREADS / warp_size; + const int id = ggml_cuda_get_device(); + const int cc = ggml_cuda_info().devices[id].cc; + const int warp_size = 32; constexpr size_t nbytes_shared = 0; +#ifdef GGML_USE_HIP + if constexpr (D <= 128) { + if (Q->ne[1] > 32) { + constexpr int cols_per_block = 64; + const int nwarps = fattn_tile_get_nthreads_host(cc, cols_per_block) / warp_size; + fattn_kernel_t fattn_kernel = flash_attn_tile; + const int kq_stride = fattn_tile_get_kq_stride_host(D, cols_per_block, cc, warp_size); + launch_fattn + (ctx, dst, fattn_kernel, nwarps, nbytes_shared, kq_stride, true, true, false, warp_size); + return; + } + } +#endif // GGML_USE_HIP + if (Q->ne[1] > 16) { constexpr int cols_per_block = 32; + const int nwarps = fattn_tile_get_nthreads_host(cc, cols_per_block) / warp_size; fattn_kernel_t fattn_kernel = flash_attn_tile; const int kq_stride = fattn_tile_get_kq_stride_host(D, cols_per_block, cc, warp_size); launch_fattn @@ -619,6 +715,7 @@ static void launch_fattn_tile_switch_ncols(ggml_backend_cuda_context & ctx, ggml } constexpr int cols_per_block = 16; + const int nwarps = fattn_tile_get_nthreads_host(cc, cols_per_block) / warp_size; fattn_kernel_t fattn_kernel = flash_attn_tile; const int kq_stride = fattn_tile_get_kq_stride_host(D, cols_per_block, cc, warp_size); launch_fattn diff --git a/ggml/src/ggml-cuda/vendors/hip.h b/ggml/src/ggml-cuda/vendors/hip.h index 12bbee4556..37386afcd4 100644 --- a/ggml/src/ggml-cuda/vendors/hip.h +++ b/ggml/src/ggml-cuda/vendors/hip.h @@ -158,41 +158,41 @@ #define __CUDA_ARCH__ 1300 -#if defined(__gfx803__) || defined(__gfx900__) || defined(__gfx906__) -#define GCN -#endif - #if defined(__gfx900__) || defined(__gfx906__) #define GCN5 -#endif +#endif // defined(__gfx900__) || defined(__gfx906__) #if defined(__gfx803__) #define GCN4 -#endif +#endif // defined(__gfx803__) -#if defined(__gfx908__) || defined(__gfx90a__) || defined(__gfx942__) -#define CDNA // For the entire family -#endif +#if defined(GCN5) || defined(GCN4) +#define GCN +#endif // defined(GCN5) || defined(GCN4) #if defined(__gfx942__) #define CDNA3 -#endif +#endif // defined(__gfx942__) #if defined(__gfx90a__) #define CDNA2 -#endif +#endif // defined(__gfx90a__) #if defined(__gfx908__) #define CDNA1 -#endif +#endif // defined(__gfx908__) + +#if defined(CDNA3) || defined(CDNA2) || defined(CDNA1) +#define CDNA // For the entire family +#endif // defined(CDNA3) || defined(CDNA2) || defined(CDNA1) #if defined(__GFX12__) #define RDNA4 -#endif +#endif // defined(__GFX12__) #if defined(__GFX11__) #define RDNA3 -#endif +#endif // defined(__GFX11__) #if defined(__gfx1030__) || defined(__gfx1031__) || defined(__gfx1032__) || defined(__gfx1033__) || \ defined(__gfx1034__) || defined(__gfx1035__) || defined(__gfx1036__) || defined(__gfx1037__) @@ -201,7 +201,11 @@ #if defined(__gfx1010__) || defined(__gfx1012__) #define RDNA1 -#endif +#endif // defined(__gfx1010__) || defined(__gfx1012__) + +#if defined(RDNA4) || defined(RDNA3) || defined(RDNA2) || defined(RDNA1) +#define RDNA // For the entire family +#endif // defined(RDNA4) || defined(RDNA3) || defined(RDNA2) || defined(RDNA1) #ifndef __has_builtin #define __has_builtin(x) 0 diff --git a/ggml/src/ggml-metal/CMakeLists.txt b/ggml/src/ggml-metal/CMakeLists.txt index 65c131b621..63418fe143 100644 --- a/ggml/src/ggml-metal/CMakeLists.txt +++ b/ggml/src/ggml-metal/CMakeLists.txt @@ -5,8 +5,12 @@ find_library(METALKIT_FRAMEWORK MetalKit REQUIRED) message(STATUS "Metal framework found") ggml_add_backend_library(ggml-metal - ggml-metal.m + ggml-metal.cpp + ggml-metal-device.m + ggml-metal-device.cpp ggml-metal-common.cpp + ggml-metal-context.m + ggml-metal-ops.cpp ) target_link_libraries(ggml-metal PRIVATE @@ -19,10 +23,6 @@ if (GGML_METAL_NDEBUG) add_compile_definitions(GGML_METAL_NDEBUG) endif() -if (GGML_METAL_USE_BF16) - add_compile_definitions(GGML_METAL_USE_BF16) -endif() - # copy metal files to bin directory configure_file(../ggml-common.h ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-common.h COPYONLY) configure_file(ggml-metal.metal ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.metal COPYONLY) diff --git a/ggml/src/ggml-metal/ggml-metal-common.cpp b/ggml/src/ggml-metal/ggml-metal-common.cpp index cb39e5b2ab..34d27b6324 100644 --- a/ggml/src/ggml-metal/ggml-metal-common.cpp +++ b/ggml/src/ggml-metal/ggml-metal-common.cpp @@ -22,7 +22,7 @@ struct ggml_mem_ranges { int debug = 0; }; -struct ggml_mem_ranges * ggml_mem_ranges_init(int debug) { +ggml_mem_ranges_t ggml_mem_ranges_init(int debug) { auto * res = new ggml_mem_ranges; res->ranges.reserve(256); @@ -31,15 +31,15 @@ struct ggml_mem_ranges * ggml_mem_ranges_init(int debug) { return res; } -void ggml_mem_ranges_free(ggml_mem_ranges * mrs) { +void ggml_mem_ranges_free(ggml_mem_ranges_t mrs) { delete mrs; } -void ggml_mem_ranges_reset(ggml_mem_ranges * mrs) { +void ggml_mem_ranges_reset(ggml_mem_ranges_t mrs) { mrs->ranges.clear(); } -static bool ggml_mem_ranges_add(ggml_mem_ranges * mrs, ggml_mem_range mr) { +static bool ggml_mem_ranges_add(ggml_mem_ranges_t mrs, ggml_mem_range mr) { mrs->ranges.push_back(mr); return true; @@ -87,7 +87,7 @@ static ggml_mem_range ggml_mem_range_from_tensor_dst(const ggml_tensor * tensor) return ggml_mem_range_from_tensor(tensor, MEM_RANGE_TYPE_DST); } -static bool ggml_mem_ranges_add_src(ggml_mem_ranges * mrs, const ggml_tensor * tensor) { +static bool ggml_mem_ranges_add_src(ggml_mem_ranges_t mrs, const ggml_tensor * tensor) { GGML_ASSERT(tensor); ggml_mem_range mr = ggml_mem_range_from_tensor_src(tensor); @@ -99,7 +99,7 @@ static bool ggml_mem_ranges_add_src(ggml_mem_ranges * mrs, const ggml_tensor * t return ggml_mem_ranges_add(mrs, mr); } -static bool ggml_mem_ranges_add_dst(ggml_mem_ranges * mrs, const ggml_tensor * tensor) { +static bool ggml_mem_ranges_add_dst(ggml_mem_ranges_t mrs, const ggml_tensor * tensor) { GGML_ASSERT(tensor); ggml_mem_range mr = ggml_mem_range_from_tensor_dst(tensor); @@ -111,7 +111,7 @@ static bool ggml_mem_ranges_add_dst(ggml_mem_ranges * mrs, const ggml_tensor * t return ggml_mem_ranges_add(mrs, mr); } -bool ggml_mem_ranges_add(ggml_mem_ranges * mrs, const ggml_tensor * tensor) { +bool ggml_mem_ranges_add(ggml_mem_ranges_t mrs, const ggml_tensor * tensor) { for (int i = 0; i < GGML_MAX_DIMS; i++) { if (tensor->src[i]) { ggml_mem_ranges_add_src(mrs, tensor->src[i]); @@ -121,7 +121,7 @@ bool ggml_mem_ranges_add(ggml_mem_ranges * mrs, const ggml_tensor * tensor) { return ggml_mem_ranges_add_dst(mrs, tensor); } -static bool ggml_mem_ranges_check(const ggml_mem_ranges * mrs, ggml_mem_range mr) { +static bool ggml_mem_ranges_check(ggml_mem_ranges_t mrs, ggml_mem_range mr) { for (size_t i = 0; i < mrs->ranges.size(); i++) { const auto & cmp = mrs->ranges[i]; @@ -152,7 +152,7 @@ static bool ggml_mem_ranges_check(const ggml_mem_ranges * mrs, ggml_mem_range mr return true; } -static bool ggml_mem_ranges_check_src(const ggml_mem_ranges * mrs, const ggml_tensor * tensor) { +static bool ggml_mem_ranges_check_src(ggml_mem_ranges_t mrs, const ggml_tensor * tensor) { GGML_ASSERT(tensor); ggml_mem_range mr = ggml_mem_range_from_tensor_src(tensor); @@ -162,7 +162,7 @@ static bool ggml_mem_ranges_check_src(const ggml_mem_ranges * mrs, const ggml_te return res; } -static bool ggml_mem_ranges_check_dst(const ggml_mem_ranges * mrs, const ggml_tensor * tensor) { +static bool ggml_mem_ranges_check_dst(ggml_mem_ranges_t mrs, const ggml_tensor * tensor) { GGML_ASSERT(tensor); ggml_mem_range mr = ggml_mem_range_from_tensor_dst(tensor); @@ -172,7 +172,7 @@ static bool ggml_mem_ranges_check_dst(const ggml_mem_ranges * mrs, const ggml_te return res; } -bool ggml_mem_ranges_check(const ggml_mem_ranges * mrs, const ggml_tensor * tensor) { +bool ggml_mem_ranges_check(ggml_mem_ranges_t mrs, const ggml_tensor * tensor) { for (int i = 0; i < GGML_MAX_DIMS; i++) { if (tensor->src[i]) { if (!ggml_mem_ranges_check_src(mrs, tensor->src[i])) { @@ -222,7 +222,7 @@ struct node_info { static std::vector ggml_metal_graph_optimize_reorder(const std::vector & nodes) { // helper to add node src and dst ranges - const auto & h_add = [](ggml_mem_ranges * mrs, const node_info & node) { + const auto & h_add = [](ggml_mem_ranges_t mrs, const node_info & node) { for (int i = 0; i < GGML_MAX_SRC; i++) { if (node.node->src[i]) { if (!ggml_mem_ranges_add_src(mrs, node.node->src[i])) { @@ -246,7 +246,7 @@ static std::vector ggml_metal_graph_optimize_reorder(const std::vectorsrc[i]) { if (!ggml_mem_ranges_check_src(mrs, node.node->src[i])) { @@ -301,10 +301,10 @@ static std::vector ggml_metal_graph_optimize_reorder(const std::vector used(n, false); // the memory ranges for the set of currently concurrent nodes - ggml_mem_ranges * mrs0 = ggml_mem_ranges_init(0); + ggml_mem_ranges_t mrs0 = ggml_mem_ranges_init(0); // the memory ranges for the set of nodes that haven't been processed yet, when looking forward for a node to reorder - ggml_mem_ranges * mrs1 = ggml_mem_ranges_init(0); + ggml_mem_ranges_t mrs1 = ggml_mem_ranges_init(0); for (int i0 = 0; i0 < n; i0++) { if (used[i0]) { @@ -375,7 +375,7 @@ static std::vector ggml_metal_graph_optimize_reorder(const std::vectorn_nodes; diff --git a/ggml/src/ggml-metal/ggml-metal-common.h b/ggml/src/ggml-metal/ggml-metal-common.h index c1402895b9..3acbc6ae17 100644 --- a/ggml/src/ggml-metal/ggml-metal-common.h +++ b/ggml/src/ggml-metal/ggml-metal-common.h @@ -25,27 +25,27 @@ enum ggml_mem_range_type { // can be added to the set without violating the constraints (i.e. if it can be executed concurrently with the // tasks already in the set) // -struct ggml_mem_ranges; +typedef struct ggml_mem_ranges * ggml_mem_ranges_t; -struct ggml_mem_ranges * ggml_mem_ranges_init(int debug); -void ggml_mem_ranges_free(struct ggml_mem_ranges * mrs); +ggml_mem_ranges_t ggml_mem_ranges_init(int debug); +void ggml_mem_ranges_free(ggml_mem_ranges_t mrs); // remove all ranges from the set -void ggml_mem_ranges_reset(struct ggml_mem_ranges * mrs); +void ggml_mem_ranges_reset(ggml_mem_ranges_t mrs); // add src or dst ranges to track -bool ggml_mem_ranges_add(struct ggml_mem_ranges * mrs, const struct ggml_tensor * tensor); +bool ggml_mem_ranges_add(ggml_mem_ranges_t mrs, const struct ggml_tensor * tensor); // return false if: // - new src range overlaps with any existing dst range // - new dst range overlaps with any existing range (src or dst) -bool ggml_mem_ranges_check(const struct ggml_mem_ranges * mrs, const struct ggml_tensor * tensor); +bool ggml_mem_ranges_check(ggml_mem_ranges_t mrs, const struct ggml_tensor * tensor); // reorder the nodes in the graph to improve concurrency, while respecting fusion // // note: this implementation is generic and not specific to metal // if it proves to work well, we can start using it for other backends in the future -void ggml_metal_graph_optimize(struct ggml_cgraph * gf); +void ggml_graph_optimize(struct ggml_cgraph * gf); #ifdef __cplusplus } diff --git a/ggml/src/ggml-metal/ggml-metal-context.h b/ggml/src/ggml-metal/ggml-metal-context.h new file mode 100644 index 0000000000..ec2b686b73 --- /dev/null +++ b/ggml/src/ggml-metal/ggml-metal-context.h @@ -0,0 +1,33 @@ +#pragma once + +#include "ggml-metal-device.h" + +#ifdef __cplusplus +extern "C" { +#endif + +// +// backend context +// + +typedef struct ggml_metal * ggml_metal_t; + +ggml_metal_t ggml_metal_init(ggml_metal_device_t dev); +void ggml_metal_free(ggml_metal_t ctx); + +void ggml_metal_synchronize(ggml_metal_t ctx); + +void ggml_metal_set_tensor_async(ggml_metal_t ctx, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size); +void ggml_metal_get_tensor_async(ggml_metal_t ctx, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size); + +enum ggml_status ggml_metal_graph_compute (ggml_metal_t ctx, struct ggml_cgraph * gf); +void ggml_metal_graph_optimize(ggml_metal_t ctx, struct ggml_cgraph * gf); + +void ggml_metal_set_n_cb (ggml_metal_t ctx, int n_cb); +void ggml_metal_set_abort_callback (ggml_metal_t ctx, ggml_abort_callback abort_callback, void * user_data); +bool ggml_metal_supports_family (ggml_metal_t ctx, int family); +void ggml_metal_capture_next_compute(ggml_metal_t ctx); + +#ifdef __cplusplus +} +#endif diff --git a/ggml/src/ggml-metal/ggml-metal-context.m b/ggml/src/ggml-metal/ggml-metal-context.m new file mode 100644 index 0000000000..af9ff21436 --- /dev/null +++ b/ggml/src/ggml-metal/ggml-metal-context.m @@ -0,0 +1,575 @@ +#import "ggml-metal-context.h" + +#import "ggml-impl.h" +#import "ggml-backend-impl.h" + +#import "ggml-metal-impl.h" +#import "ggml-metal-common.h" +#import "ggml-metal-ops.h" + +#import + +#import + +#undef MIN +#undef MAX +#define MIN(a, b) ((a) < (b) ? (a) : (b)) +#define MAX(a, b) ((a) > (b) ? (a) : (b)) + +// max number of MTLCommandBuffer used to submit a graph for processing +#define GGML_METAL_MAX_COMMAND_BUFFERS 8 + +struct ggml_metal_command_buffer { + id obj; +}; + +struct ggml_metal { + id device; + id queue; // currently a pointer to the device queue, but might become separate queue [TAG_QUEUE_PER_BACKEND] + + ggml_metal_device_t dev; + ggml_metal_library_t lib; + + dispatch_queue_t d_queue; + + // additional, inference-time compiled pipelines + ggml_metal_pipelines_t pipelines_ext; + + bool use_bfloat; + bool use_fusion; + bool use_concurrency; + bool use_graph_optimize; + + int debug_graph; + int debug_fusion; + + // how many times a given op was fused + uint64_t fuse_cnt[GGML_OP_COUNT]; + + // capture state + bool capture_next_compute; + bool capture_started; + + id capture_scope; + + // command buffer state + int n_cb; // number of extra threads used to submit the command buffers + int n_nodes_0; // number of nodes submitted by the main thread + int n_nodes_1; // remaining number of nodes submitted by the n_cb threads + int n_nodes_per_cb; + + struct ggml_cgraph * gf; + + // the callback given to the thread pool + void (^encode_async)(size_t ith); + + // n_cb command buffers + 1 used by the main thread + struct ggml_metal_command_buffer cmd_bufs[GGML_METAL_MAX_COMMAND_BUFFERS + 1]; + + // extra command buffers for things like getting, setting and copying tensors + NSMutableArray * cmd_bufs_ext; + + // the last command buffer queued into the Metal queue with operations relevant to the current Metal backend + id cmd_buf_last; + + // abort ggml_metal_graph_compute if callback returns true + ggml_abort_callback abort_callback; + void * abort_callback_data; +}; + +ggml_metal_t ggml_metal_init(ggml_metal_device_t dev) { + GGML_LOG_INFO("%s: allocating\n", __func__); + +#if TARGET_OS_OSX && !GGML_METAL_NDEBUG + // Show all the Metal device instances in the system + NSArray * devices = MTLCopyAllDevices(); + for (id device in devices) { + GGML_LOG_INFO("%s: found device: %s\n", __func__, [[device name] UTF8String]); + } + [devices release]; // since it was created by a *Copy* C method +#endif + + // init context + ggml_metal_t res = calloc(1, sizeof(struct ggml_metal)); + + res->device = ggml_metal_device_get_obj(dev); + + GGML_LOG_INFO("%s: picking default device: %s\n", __func__, [[res->device name] UTF8String]); + + // TODO: would it be better to have one queue for the backend and one queue for the device? + // the graph encoders and async ops would use the backend queue while the sync ops would use the device queue? + //res->queue = [device newCommandQueue]; [TAG_QUEUE_PER_BACKEND] + res->queue = ggml_metal_device_get_queue(dev); + if (res->queue == nil) { + GGML_LOG_ERROR("%s: error: failed to create command queue\n", __func__); + return NULL; + } + + res->dev = dev; + res->lib = ggml_metal_device_get_library(dev); + if (res->lib == NULL) { + GGML_LOG_WARN("%s: the device does not have a precompiled Metal library - this is unexpected\n", __func__); + GGML_LOG_WARN("%s: will try to compile it on the fly\n", __func__); + + res->lib = ggml_metal_library_init(dev); + if (res->lib == NULL) { + GGML_LOG_ERROR("%s: error: failed to initialize the Metal library\n", __func__); + + free(res); + + return NULL; + } + } + + const struct ggml_metal_device_props * props_dev = ggml_metal_device_get_props(dev); + + res->d_queue = dispatch_queue_create("ggml-metal", DISPATCH_QUEUE_CONCURRENT); + + res->use_bfloat = props_dev->has_bfloat; + res->use_fusion = getenv("GGML_METAL_FUSION_DISABLE") == nil; + res->use_concurrency = getenv("GGML_METAL_CONCURRENCY_DISABLE") == nil; + + { + const char * val = getenv("GGML_METAL_GRAPH_DEBUG"); + res->debug_graph = val ? atoi(val) : 0; + } + + { + const char * val = getenv("GGML_METAL_FUSION_DEBUG"); + res->debug_fusion = val ? atoi(val) : 0; + } + + res->use_graph_optimize = true; + + if (getenv("GGML_METAL_GRAPH_OPTIMIZE_DISABLE") != NULL) { + res->use_graph_optimize = false; + } + + memset(res->fuse_cnt, 0, sizeof(res->fuse_cnt)); + + GGML_LOG_INFO("%s: use bfloat = %s\n", __func__, res->use_bfloat ? "true" : "false"); + GGML_LOG_INFO("%s: use fusion = %s\n", __func__, res->use_fusion ? "true" : "false"); + GGML_LOG_INFO("%s: use concurrency = %s\n", __func__, res->use_concurrency ? "true" : "false"); + GGML_LOG_INFO("%s: use graph optimize = %s\n", __func__, res->use_graph_optimize ? "true" : "false"); + + res->capture_next_compute = false; + res->capture_started = false; + res->capture_scope = nil; + + res->gf = nil; + res->encode_async = nil; + for (int i = 0; i < GGML_METAL_MAX_COMMAND_BUFFERS; ++i) { + res->cmd_bufs[i].obj = nil; + } + + res->cmd_bufs_ext = [[NSMutableArray alloc] init]; + + res->cmd_buf_last = nil; + + res->pipelines_ext = ggml_metal_pipelines_init(); + + return res; +} + +void ggml_metal_free(ggml_metal_t ctx) { + GGML_LOG_INFO("%s: deallocating\n", __func__); + + for (int i = 0; i < GGML_METAL_MAX_COMMAND_BUFFERS; ++i) { + if (ctx->cmd_bufs[i].obj) { + [ctx->cmd_bufs[i].obj release]; + } + } + + for (int i = 0; i < (int) ctx->cmd_bufs_ext.count; ++i) { + if (ctx->cmd_bufs_ext[i]) { + [ctx->cmd_bufs_ext[i] release]; + } + } + + [ctx->cmd_bufs_ext removeAllObjects]; + [ctx->cmd_bufs_ext release]; + + if (ctx->pipelines_ext) { + ggml_metal_pipelines_free(ctx->pipelines_ext); + ctx->pipelines_ext = nil; + } + + if (ctx->debug_fusion > 0) { + GGML_LOG_DEBUG("%s: fusion stats:\n", __func__); + for (int i = 0; i < GGML_OP_COUNT; i++) { + if (ctx->fuse_cnt[i] == 0) { + continue; + } + + // note: cannot use ggml_log here + GGML_LOG_DEBUG("%s: - %s: %" PRIu64 "\n", __func__, ggml_op_name((enum ggml_op) i), ctx->fuse_cnt[i]); + } + } + + Block_release(ctx->encode_async); + + //[ctx->queue release]; // [TAG_QUEUE_PER_BACKEND] + + dispatch_release(ctx->d_queue); + + free(ctx); +} + +void ggml_metal_synchronize(ggml_metal_t ctx) { + // wait for any backend operations to finish + if (ctx->cmd_buf_last) { + [ctx->cmd_buf_last waitUntilCompleted]; + ctx->cmd_buf_last = nil; + } + + // release any completed command buffers + if (ctx->cmd_bufs_ext.count > 0) { + for (size_t i = 0; i < ctx->cmd_bufs_ext.count; ++i) { + id cmd_buf = ctx->cmd_bufs_ext[i]; + + MTLCommandBufferStatus status = [cmd_buf status]; + if (status != MTLCommandBufferStatusCompleted) { + GGML_LOG_ERROR("%s: error: command buffer %d failed with status %d\n", __func__, (int) i, (int) status); + if (status == MTLCommandBufferStatusError) { + GGML_LOG_ERROR("error: %s\n", [[cmd_buf error].localizedDescription UTF8String]); + } + GGML_ABORT("fatal error"); + } + + [cmd_buf release]; + } + + [ctx->cmd_bufs_ext removeAllObjects]; + } +} + +static struct ggml_metal_buffer_id ggml_metal_get_buffer_id(const struct ggml_tensor * t) { + if (!t) { + return (struct ggml_metal_buffer_id) { nil, 0 }; + } + + ggml_backend_buffer_t buffer = t->view_src ? t->view_src->buffer : t->buffer; + + return ggml_metal_buffer_get_id(buffer->context, t); +} + +void ggml_metal_set_tensor_async(ggml_metal_t ctx, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { + @autoreleasepool { + // wrap the source data into a Metal buffer + id buf_src = [ctx->device newBufferWithBytes:data + length:size + options:MTLResourceStorageModeShared]; + + struct ggml_metal_buffer_id bid_dst = ggml_metal_get_buffer_id(tensor); + if (bid_dst.metal == nil) { + GGML_ABORT("%s: failed to find buffer for tensor '%s'\n", __func__, tensor->name); + } + + bid_dst.offs += offset; + + // queue the copy operation into the queue of the Metal context + // this will be queued at the end, after any currently ongoing GPU operations + id cmd_buf = [ctx->queue commandBufferWithUnretainedReferences]; + id encoder = [cmd_buf blitCommandEncoder]; + + [encoder copyFromBuffer:buf_src + sourceOffset:0 + toBuffer:bid_dst.metal + destinationOffset:bid_dst.offs + size:size]; + + [encoder endEncoding]; + [cmd_buf commit]; + + // do not wait here for completion + //[cmd_buf waitUntilCompleted]; + + // instead, remember a reference to the command buffer and wait for it later if needed + [ctx->cmd_bufs_ext addObject:cmd_buf]; + ctx->cmd_buf_last = cmd_buf; + + [cmd_buf retain]; + } +} + +void ggml_metal_get_tensor_async(ggml_metal_t ctx, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { + @autoreleasepool { + id buf_dst = [ctx->device newBufferWithBytesNoCopy:data + length:size + options:MTLResourceStorageModeShared + deallocator:nil]; + + struct ggml_metal_buffer_id bid_src = ggml_metal_get_buffer_id(tensor); + if (bid_src.metal == nil) { + GGML_ABORT("%s: failed to find buffer for tensor '%s'\n", __func__, tensor->name); + } + + bid_src.offs += offset; + + // queue the copy operation into the queue of the Metal context + // this will be queued at the end, after any currently ongoing GPU operations + id cmd_buf = [ctx->queue commandBufferWithUnretainedReferences]; + id encoder = [cmd_buf blitCommandEncoder]; + + [encoder copyFromBuffer:bid_src.metal + sourceOffset:bid_src.offs + toBuffer:buf_dst + destinationOffset:0 + size:size]; + + [encoder endEncoding]; + [cmd_buf commit]; + + // do not wait here for completion + //[cmd_buf waitUntilCompleted]; + + // instead, remember a reference to the command buffer and wait for it later if needed + [ctx->cmd_bufs_ext addObject:cmd_buf]; + ctx->cmd_buf_last = cmd_buf; + + [cmd_buf retain]; + } +} + +enum ggml_status ggml_metal_graph_compute(ggml_metal_t ctx, struct ggml_cgraph * gf) { + // number of nodes encoded by the main thread (empirically determined) + const int n_main = 64; + + // number of threads in addition to the main thread + const int n_cb = ctx->n_cb; + + // submit the ggml compute graph to the GPU by creating command buffers and encoding the ops in them + // the first n_nodes_0 are encoded and submitted for processing directly by the calling thread + // while these nodes are processing, we start n_cb threads to enqueue the rest of the nodes + // each thread creates it's own command buffer and enqueues the ops in parallel + // + // tests on M1 Pro and M2 Ultra using LLaMA models, show that optimal values for n_cb are 1 or 2 + + @autoreleasepool { + ctx->gf = gf; + + ctx->n_nodes_0 = MIN(n_main, gf->n_nodes); + ctx->n_nodes_1 = gf->n_nodes - ctx->n_nodes_0; + + ctx->n_nodes_per_cb = (ctx->n_nodes_1 + ctx->n_cb - 1) / ctx->n_cb; + + const bool use_capture = ctx->capture_next_compute; + if (use_capture) { + ctx->capture_next_compute = false; + + // make sure all previous computations have finished before starting the capture + if (ctx->cmd_buf_last) { + [ctx->cmd_buf_last waitUntilCompleted]; + ctx->cmd_buf_last = nil; + } + + if (!ctx->capture_started) { + // create capture scope + ctx->capture_scope = [[MTLCaptureManager sharedCaptureManager] newCaptureScopeWithDevice:ctx->device]; + + MTLCaptureDescriptor * descriptor = [MTLCaptureDescriptor new]; + descriptor.captureObject = ctx->capture_scope; + descriptor.destination = MTLCaptureDestinationGPUTraceDocument; + descriptor.outputURL = [NSURL fileURLWithPath:[NSString stringWithFormat:@"/tmp/perf-metal.gputrace"]]; + + NSError * error = nil; + if (![[MTLCaptureManager sharedCaptureManager] startCaptureWithDescriptor:descriptor error:&error]) { + GGML_LOG_ERROR("%s: error: unable to start capture '%s'\n", __func__, [[error localizedDescription] UTF8String]); + } else { + [ctx->capture_scope beginScope]; + ctx->capture_started = true; + } + } + } + + // the main thread commits the first few commands immediately + // cmd_buf[n_cb] + { + id cmd_buf = [ctx->queue commandBufferWithUnretainedReferences]; + [cmd_buf retain]; + + if (ctx->cmd_bufs[n_cb].obj) { + [ctx->cmd_bufs[n_cb].obj release]; + } + ctx->cmd_bufs[n_cb].obj = cmd_buf; + + [cmd_buf enqueue]; + + ctx->encode_async(n_cb); + } + + // remember the command buffer for the next iteration + ctx->cmd_buf_last = ctx->cmd_bufs[n_cb].obj; + + // prepare the rest of the command buffers asynchronously (optional) + // cmd_buf[0.. n_cb) + for (int cb_idx = 0; cb_idx < n_cb; ++cb_idx) { + id cmd_buf = [ctx->queue commandBufferWithUnretainedReferences]; + [cmd_buf retain]; + + if (ctx->cmd_bufs[cb_idx].obj) { + [ctx->cmd_bufs[cb_idx].obj release]; + } + ctx->cmd_bufs[cb_idx].obj = cmd_buf; + + // always enqueue the first two command buffers + // enqueue all of the command buffers if we don't need to abort + if (cb_idx < 2 || ctx->abort_callback == NULL) { + [cmd_buf enqueue]; + + // update the pointer to the last queued command buffer + // this is needed to implement synchronize() + ctx->cmd_buf_last = cmd_buf; + } + } + + dispatch_apply(n_cb, ctx->d_queue, ctx->encode_async); + + // for debugging: block until graph is computed + //[ctx->cmd_buf_last waitUntilCompleted]; + + // enter here only when capturing in order to wait for all computation to finish + // otherwise, we leave the graph to compute asynchronously + if (!use_capture && ctx->capture_started) { + // wait for completion and check status of each command buffer + // needed to detect if the device ran out-of-memory for example (#1881) + { + id cmd_buf = ctx->cmd_bufs[n_cb].obj; + [cmd_buf waitUntilCompleted]; + + MTLCommandBufferStatus status = [cmd_buf status]; + if (status != MTLCommandBufferStatusCompleted) { + GGML_LOG_INFO("%s: command buffer %d failed with status %lu\n", __func__, n_cb, status); + if (status == MTLCommandBufferStatusError) { + GGML_LOG_INFO("error: %s\n", [[cmd_buf error].localizedDescription UTF8String]); + } + + return GGML_STATUS_FAILED; + } + } + + for (int i = 0; i < n_cb; ++i) { + id cmd_buf = ctx->cmd_bufs[i].obj; + [cmd_buf waitUntilCompleted]; + + MTLCommandBufferStatus status = [cmd_buf status]; + if (status != MTLCommandBufferStatusCompleted) { + GGML_LOG_INFO("%s: command buffer %d failed with status %lu\n", __func__, i, status); + if (status == MTLCommandBufferStatusError) { + GGML_LOG_INFO("error: %s\n", [[cmd_buf error].localizedDescription UTF8String]); + } + + return GGML_STATUS_FAILED; + } + + id next_buffer = (i + 1 < n_cb ? ctx->cmd_bufs[i + 1].obj : nil); + if (!next_buffer) { + continue; + } + + const bool next_queued = ([next_buffer status] != MTLCommandBufferStatusNotEnqueued); + if (next_queued) { + continue; + } + + if (ctx->abort_callback && ctx->abort_callback(ctx->abort_callback_data)) { + GGML_LOG_INFO("%s: command buffer %d aborted", __func__, i); + return GGML_STATUS_ABORTED; + } + + [next_buffer commit]; + } + + [ctx->capture_scope endScope]; + [[MTLCaptureManager sharedCaptureManager] stopCapture]; + } + } + + return GGML_STATUS_SUCCESS; +} + +void ggml_metal_graph_optimize(ggml_metal_t ctx, struct ggml_cgraph * gf) { + //const int64_t t_start = ggml_time_us(); + + if (ctx->use_graph_optimize) { + ggml_graph_optimize(gf); + } + + //printf("%s: graph optimize took %.3f ms\n", __func__, (ggml_time_us() - t_start) / 1000.0); +} + +void ggml_metal_set_n_cb(ggml_metal_t ctx, int n_cb) { + if (ctx->n_cb != n_cb) { + ctx->n_cb = MIN(n_cb, GGML_METAL_MAX_COMMAND_BUFFERS); + + if (ctx->n_cb > 2) { + GGML_LOG_WARN("%s: n_cb = %d, using n_cb > 2 is not recommended and can degrade the performance in some cases\n", __func__, n_cb); + } + } + + if (ctx->encode_async) { + Block_release(ctx->encode_async); + } + + ctx->encode_async = Block_copy(^(size_t iter) { + const int cb_idx = iter; + const int n_cb_l = ctx->n_cb; + + const int n_nodes_0 = ctx->n_nodes_0; + const int n_nodes_1 = ctx->n_nodes_1; + + const int n_nodes_per_cb = ctx->n_nodes_per_cb; + + int idx_start = 0; + int idx_end = n_nodes_0; + + if (cb_idx < n_cb_l) { + idx_start = n_nodes_0 + ( (cb_idx + 0) * n_nodes_per_cb); + idx_end = n_nodes_0 + (MIN((cb_idx == n_cb_l - 1) ? n_nodes_1 : (cb_idx + 1) * n_nodes_per_cb, n_nodes_1)); + } + + id cmd_buf = ctx->cmd_bufs[cb_idx].obj; + + ggml_metal_op_t ctx_op = ggml_metal_op_init( + ctx->dev, + cmd_buf, + ctx->gf, + idx_start, + idx_end, + ctx->use_fusion, + ctx->use_concurrency, + ctx->capture_next_compute, + ctx->debug_graph, + ctx->debug_fusion); + + for (int idx = idx_start; idx < idx_end;) { + const int res = ggml_metal_op_encode(ctx_op, idx); + if (res == 0) { + break; + } + + idx += res; + } + + ggml_metal_op_free(ctx_op); + + if (cb_idx < 2 || ctx->abort_callback == NULL) { + [cmd_buf commit]; + } + }); +} + +void ggml_metal_set_abort_callback(ggml_metal_t ctx, ggml_abort_callback abort_callback, void * user_data) { + ctx->abort_callback = abort_callback; + ctx->abort_callback_data = user_data; +} + +bool ggml_metal_supports_family(ggml_metal_t ctx, int family) { + GGML_ASSERT(ctx->device != nil); + + return [ctx->device supportsFamily:(MTLGPUFamilyApple1 + family - 1)]; +} + +void ggml_metal_capture_next_compute(ggml_metal_t ctx) { + ctx->capture_next_compute = true; +} diff --git a/ggml/src/ggml-metal/ggml-metal-device.cpp b/ggml/src/ggml-metal/ggml-metal-device.cpp new file mode 100644 index 0000000000..bada84cef3 --- /dev/null +++ b/ggml/src/ggml-metal/ggml-metal-device.cpp @@ -0,0 +1,1371 @@ +#include "ggml-metal-device.h" + +#include "ggml-metal-impl.h" + +#include "ggml-impl.h" + +#include +#include +#include +#include + +struct ggml_metal_device_deleter { + void operator()(ggml_metal_device_t ctx) { + ggml_metal_device_free(ctx); + } +}; + +typedef std::unique_ptr ggml_metal_device_ptr; + +ggml_metal_device_t ggml_metal_device_get(void) { + static ggml_metal_device_ptr ctx { ggml_metal_device_init() }; + + return ctx.get(); +} + +struct ggml_metal_pipelines { + std::unordered_map data; +}; + +ggml_metal_pipelines_t ggml_metal_pipelines_init(void) { + ggml_metal_pipelines_t res = new ggml_metal_pipelines(); + + return res; +} + +void ggml_metal_pipelines_free(ggml_metal_pipelines_t ppls) { + if (!ppls) { + return; + } + + for (auto it = ppls->data.begin(); it != ppls->data.end(); ++it) { + ggml_metal_pipeline_free(it->second); + } + + delete ppls; +} + +void ggml_metal_pipelines_add(ggml_metal_pipelines_t ppls, const char * name, ggml_metal_pipeline_t pipeline) { + ppls->data[name] = pipeline; +} + +ggml_metal_pipeline_t ggml_metal_pipelines_get(ggml_metal_pipelines_t ppls, const char * name) { + if (ppls->data.find(name) == ppls->data.end()) { + return nullptr; + } + + return ppls->data[name]; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_base(ggml_metal_library_t lib, ggml_op op) { + char base[256]; + char name[256]; + + const char * op_str = "undefined"; + switch (op) { + case GGML_OP_ADD_ID: op_str = "add_id"; break; + case GGML_OP_CONCAT: op_str = "concat"; break; + default: GGML_ABORT("fatal error"); + }; + + snprintf(base, 256, "kernel_%s", op_str); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_cpy(ggml_metal_library_t lib, ggml_type tsrc, ggml_type tdst) { + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_cpy_%s_%s", ggml_type_name(tsrc), ggml_type_name(tdst)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_pool_2d(ggml_metal_library_t lib, const ggml_tensor * op, ggml_op_pool op_pool) { + GGML_ASSERT(ggml_is_contiguous(op->src[0])); + GGML_ASSERT(op->src[0]->type == GGML_TYPE_F32 && op->src[0]->type == op->type); + + const char * pool_str = "undefined"; + switch (op_pool) { + case GGML_OP_POOL_AVG: pool_str = "avg"; break; + case GGML_OP_POOL_MAX: pool_str = "max"; break; + default: GGML_ASSERT(false && "not implemented"); + }; + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_pool_2d_%s_%s", pool_str, ggml_type_name(op->src[0]->type)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_get_rows(ggml_metal_library_t lib, ggml_type tsrc) { + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_get_rows_%s", ggml_type_name(tsrc)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_set_rows(ggml_metal_library_t lib, ggml_type tdst) { + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_set_rows_%s", ggml_type_name(tdst)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_repeat(ggml_metal_library_t lib, ggml_type tsrc) { + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_repeat_%s", ggml_type_name(tsrc)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_unary(ggml_metal_library_t lib, const ggml_tensor * op) { + GGML_ASSERT(ggml_is_contiguous(op->src[0])); + + char base[256]; + char name[256]; + + const int64_t n = ggml_nelements(op); + + const char * op_str = "undefined"; + switch (op->op) { + case GGML_OP_SCALE: op_str = "scale"; break; + case GGML_OP_CLAMP: op_str = "clamp"; break; + case GGML_OP_SQR: op_str = "sqr"; break; + case GGML_OP_SQRT: op_str = "sqrt"; break; + case GGML_OP_SIN: op_str = "sin"; break; + case GGML_OP_COS: op_str = "cos"; break; + case GGML_OP_LOG: op_str = "log"; break; + case GGML_OP_LEAKY_RELU: op_str = "leaky_relu"; break; + case GGML_OP_UNARY: + switch (ggml_get_unary_op(op)) { + case GGML_UNARY_OP_TANH: op_str = "tanh"; break; + case GGML_UNARY_OP_RELU: op_str = "relu"; break; + case GGML_UNARY_OP_SIGMOID: op_str = "sigmoid"; break; + case GGML_UNARY_OP_GELU: op_str = "gelu"; break; + case GGML_UNARY_OP_GELU_ERF: op_str = "gelu_erf"; break; + case GGML_UNARY_OP_GELU_QUICK: op_str = "gelu_quick"; break; + case GGML_UNARY_OP_SILU: op_str = "silu"; break; + case GGML_UNARY_OP_ELU: op_str = "elu"; break; + case GGML_UNARY_OP_NEG: op_str = "neg"; break; + case GGML_UNARY_OP_ABS: op_str = "abs"; break; + case GGML_UNARY_OP_SGN: op_str = "sgn"; break; + case GGML_UNARY_OP_STEP: op_str = "step"; break; + case GGML_UNARY_OP_HARDSWISH: op_str = "hardswish"; break; + case GGML_UNARY_OP_HARDSIGMOID: op_str = "hardsigmoid"; break; + case GGML_UNARY_OP_EXP: op_str = "exp"; break; + default: GGML_ABORT("fatal error"); + } break; + default: GGML_ABORT("fatal error"); + }; + + const char * suffix = ""; + if (n % 4 == 0) { + suffix = "_4"; + } + + snprintf(base, 256, "kernel_%s_%s%s", op_str, ggml_type_name(op->src[0]->type), suffix); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_glu(ggml_metal_library_t lib, const ggml_tensor * op) { + GGML_ASSERT(ggml_is_contiguous_1(op->src[0])); + + char base[256]; + char name[256]; + + const char * op_str = "undefined"; + switch (op->op) { + case GGML_OP_GLU: + switch (ggml_get_glu_op(op)) { + case GGML_GLU_OP_REGLU: op_str = "reglu"; break; + case GGML_GLU_OP_GEGLU: op_str = "geglu"; break; + case GGML_GLU_OP_SWIGLU: op_str = "swiglu"; break; + case GGML_GLU_OP_SWIGLU_OAI: op_str = "swiglu_oai"; break; + case GGML_GLU_OP_GEGLU_ERF: op_str = "geglu_erf"; break; + case GGML_GLU_OP_GEGLU_QUICK: op_str = "geglu_quick"; break; + default: GGML_ABORT("fatal error"); + } break; + default: GGML_ABORT("fatal error"); + }; + + snprintf(base, 256, "kernel_%s_%s", op_str, ggml_type_name(op->src[0]->type)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_sum_rows(ggml_metal_library_t lib, const ggml_tensor * op) { + GGML_ASSERT(op->src[0]->nb[0] == ggml_type_size(op->src[0]->type)); + + char base[256]; + char name[256]; + + const char * op_str = "undefined"; + switch (op->op) { + case GGML_OP_SUM_ROWS: + op_str = "sum_rows"; break; + case GGML_OP_MEAN: + op_str = "mean"; break; + default: GGML_ABORT("fatal error"); + }; + + snprintf(base, 256, "kernel_%s_%s", op_str, ggml_type_name(op->src[0]->type)); + + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + ggml_metal_pipeline_set_smem(res, 32*sizeof(float)); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_soft_max(ggml_metal_library_t lib, const ggml_tensor * op) { + GGML_ASSERT(!op->src[1] || op->src[1]->type == GGML_TYPE_F16 || op->src[1]->type == GGML_TYPE_F32); + + char base[256]; + char name[256]; + + const char * suffix = ""; + + if (op->src[0]->ne[0] % 4 == 0) { + suffix = "_4"; + } + + const ggml_type tsrc1 = op->src[1] ? op->src[1]->type : GGML_TYPE_F32; + + snprintf(base, 256, "kernel_soft_max_%s%s", ggml_type_name(tsrc1), suffix); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + ggml_metal_pipeline_set_smem(res, 32*sizeof(float)); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_ssm_conv(ggml_metal_library_t lib, const ggml_tensor * op) { + GGML_ASSERT(op->src[0]->type == GGML_TYPE_F32); + GGML_ASSERT(op->src[1]->type == GGML_TYPE_F32); + + GGML_ASSERT(ggml_is_contiguous(op->src[0])); + GGML_ASSERT(ggml_is_contiguous(op->src[1])); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_ssm_conv_%s_%s", ggml_type_name(op->src[0]->type), ggml_type_name(op->src[1]->type)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_ssm_scan(ggml_metal_library_t lib, const ggml_tensor * op) { + char base[256]; + char name[256]; + + if (op->src[3]->ne[0] == 1) { + snprintf(base, 256, "kernel_ssm_scan_group_%s", ggml_type_name(op->src[0]->type)); + } else { + snprintf(base, 256, "kernel_ssm_scan_%s", ggml_type_name(op->src[0]->type)); + } + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + ggml_metal_pipeline_set_smem(res, 32*sizeof(float)); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_rwkv(ggml_metal_library_t lib, const ggml_tensor * op) { + char base[256]; + char name[256]; + + const int64_t C = op->ne[0]; + const int64_t H = op->src[0]->ne[1]; + + switch (op->op) { + case GGML_OP_RWKV_WKV6: + { + GGML_ASSERT(op->src[5]->type == GGML_TYPE_F32); + GGML_ASSERT(C % H == 0); + GGML_ASSERT(C / H == 64); + + snprintf(base, 256, "kernel_rwkv_wkv6_%s", ggml_type_name(op->src[0]->type)); + } break; + case GGML_OP_RWKV_WKV7: + { + GGML_ASSERT(op->src[6]->type == GGML_TYPE_F32); + GGML_ASSERT(C % H == 0); + GGML_ASSERT(C / H == 64); + + snprintf(base, 256, "kernel_rwkv_wkv7_%s", ggml_type_name(op->src[0]->type)); + } break; + default: + GGML_ABORT("fatal error"); + } + + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_mul_mv_ext(ggml_metal_library_t lib, ggml_type tsrc0, ggml_type tsrc1, int r1ptg) { + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_mul_mv_ext_%s_%s_r1_%d", ggml_type_name(tsrc0), ggml_type_name(tsrc1), r1ptg); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_mul_mm(ggml_metal_library_t lib, ggml_type tsrc0, ggml_type tsrc1) { + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_mul_mm_%s_%s", ggml_type_name(tsrc0), ggml_type_name(tsrc1)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + ggml_metal_pipeline_set_smem(res, 8192); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_mul_mv(ggml_metal_library_t lib, const ggml_tensor * op) { + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + + char base[256]; + char name[256]; + + int nsg = 0; // number of simdgroups + int nr0 = 0; // number of src0 rows per simdgroup + int nr1 = 1; // number of src1 rows per threadgroup + + size_t smem = 0; // shared memory + + const ggml_type tsrc0 = op->src[0]->type; + const ggml_type tsrc1 = op->src[1]->type; + + const char * suffix = ""; + + // use custom matrix x vector kernel + switch (tsrc0) { + case GGML_TYPE_F32: + case GGML_TYPE_F16: + case GGML_TYPE_BF16: + { + if (ne00 == 4) { + nsg = 1; + nr0 = 32; + nr1 = 4; + suffix = "_c4"; + } else if (ne00 % 4 == 0) { + nsg = N_SG_F; + nr0 = N_R0_F; + nr1 = 1; + smem = 32*sizeof(float)*N_R0_F; + suffix = "_4"; + } else { + nsg = N_SG_F; + nr0 = N_R0_F; + nr1 = 1; + smem = 32*sizeof(float)*N_R0_F; + } + } break; + case GGML_TYPE_Q4_0: + { + nsg = N_SG_Q4_0; + nr0 = N_R0_Q4_0; + } break; + case GGML_TYPE_Q4_1: + { + nsg = N_SG_Q4_1; + nr0 = N_R0_Q4_1; + } break; + case GGML_TYPE_Q5_0: + { + nsg = N_SG_Q5_0; + nr0 = N_R0_Q5_0; + } break; + case GGML_TYPE_Q5_1: + { + nsg = N_SG_Q5_1; + nr0 = N_R0_Q5_1; + } break; + case GGML_TYPE_Q8_0: + { + nsg = N_SG_Q8_0; + nr0 = N_R0_Q8_0; + smem = 32*sizeof(float)*N_R0_Q8_0; + } break; + case GGML_TYPE_MXFP4: + { + nsg = N_SG_MXFP4; + nr0 = N_R0_MXFP4; + smem = 32*sizeof(float); + } break; + case GGML_TYPE_Q2_K: + { + nsg = N_SG_Q2_K; + nr0 = N_R0_Q2_K; + } break; + case GGML_TYPE_Q3_K: + { + nsg = N_SG_Q3_K; + nr0 = N_R0_Q3_K; + } break; + case GGML_TYPE_Q4_K: + { + nsg = N_SG_Q4_K; + nr0 = N_R0_Q4_K; + } break; + case GGML_TYPE_Q5_K: + { + nsg = N_SG_Q5_K; + nr0 = N_R0_Q5_K; + } break; + case GGML_TYPE_Q6_K: + { + nsg = N_SG_Q6_K; + nr0 = N_R0_Q6_K; + } break; + case GGML_TYPE_IQ2_XXS: + { + nsg = N_SG_IQ2_XXS; + nr0 = N_R0_IQ2_XXS; + smem = 256*8+128; + } break; + case GGML_TYPE_IQ2_XS: + { + nsg = N_SG_IQ2_XS; + nr0 = N_R0_IQ2_XS; + smem = 512*8+128; + } break; + case GGML_TYPE_IQ3_XXS: + { + nsg = N_SG_IQ3_XXS; + nr0 = N_R0_IQ3_XXS; + smem = 256*4+128; + } break; + case GGML_TYPE_IQ3_S: + { + nsg = N_SG_IQ3_S; + nr0 = N_R0_IQ3_S; + smem = 512*4; + } break; + case GGML_TYPE_IQ2_S: + { + nsg = N_SG_IQ2_S; + nr0 = N_R0_IQ2_S; + } break; + case GGML_TYPE_IQ1_S: + { + nsg = N_SG_IQ1_S; + nr0 = N_R0_IQ1_S; + } break; + case GGML_TYPE_IQ1_M: + { + nsg = N_SG_IQ1_M; + nr0 = N_R0_IQ1_M; + } break; + case GGML_TYPE_IQ4_NL: + { + nsg = N_SG_IQ4_NL; + nr0 = N_R0_IQ4_NL; + smem = 32*sizeof(float); + } break; + case GGML_TYPE_IQ4_XS: + { + nsg = N_SG_IQ4_XS; + nr0 = N_R0_IQ4_XS; + smem = 32*sizeof(float); + } break; + default: + { + GGML_LOG_ERROR("Asserting on type %d\n", (int) tsrc0); + GGML_ABORT("not implemented"); + } + }; + + snprintf(base, 256, "kernel_mul_mv_%s_%s%s", ggml_type_name(tsrc0), ggml_type_name(tsrc1), suffix); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + ggml_metal_cv_t cv = ggml_metal_cv_init(); + + ggml_metal_cv_set_int16(cv, nsg, FC_MUL_MV + 0); + + res = ggml_metal_library_compile_pipeline(lib, base, name, cv); + + ggml_metal_cv_free(cv); + + ggml_metal_pipeline_set_nr0 (res, nr0); + ggml_metal_pipeline_set_nr1 (res, nr1); + ggml_metal_pipeline_set_nsg (res, nsg); + ggml_metal_pipeline_set_smem(res, smem); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_mul_mm_id_map0(ggml_metal_library_t lib, int ne02, int ne20) { + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_mul_mm_id_map0_ne20_%d", ne20); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + const size_t smem = (size_t) ne02*ne20*sizeof(uint16_t); + + ggml_metal_pipeline_set_smem(res, smem); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_mul_mm_id(ggml_metal_library_t lib, ggml_type tsrc0, ggml_type tsrc1) { + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_mul_mm_id_%s_%s", ggml_type_name(tsrc0), ggml_type_name(tsrc1)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + ggml_metal_pipeline_set_smem(res, 8192); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_mul_mv_id(ggml_metal_library_t lib, const ggml_tensor * op) { + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + + char base[256]; + char name[256]; + + int nsg = 0; // number of simdgroups + int nr0 = 0; // number of src0 rows per simdgroup + int nr1 = 1; // number of src1 rows per threadgroup + + size_t smem = 0; // shared memory + + const ggml_type tsrc0 = op->src[0]->type; + const ggml_type tsrc1 = op->src[1]->type; + + const char * suffix = ""; + + // use custom matrix x vector kernel + switch (tsrc0) { + case GGML_TYPE_F32: + case GGML_TYPE_F16: + case GGML_TYPE_BF16: + { + if (ne00 % 4 == 0) { + nsg = N_SG_F; + nr0 = N_R0_F; + nr1 = 1; + smem = 32*sizeof(float)*N_R0_F; + suffix = "_4"; + } else { + nsg = N_SG_F; + nr0 = N_R0_F; + nr1 = 1; + smem = 32*sizeof(float)*N_R0_F; + } + } break; + case GGML_TYPE_Q4_0: + { + nsg = N_SG_Q4_0; + nr0 = N_R0_Q4_0; + } break; + case GGML_TYPE_Q4_1: + { + nsg = N_SG_Q4_1; + nr0 = N_R0_Q4_1; + } break; + case GGML_TYPE_Q5_0: + { + nsg = N_SG_Q5_0; + nr0 = N_R0_Q5_0; + } break; + case GGML_TYPE_Q5_1: + { + nsg = N_SG_Q5_1; + nr0 = N_R0_Q5_1; + } break; + case GGML_TYPE_Q8_0: + { + nsg = N_SG_Q8_0; + nr0 = N_R0_Q8_0; + smem = 32*sizeof(float)*N_R0_Q8_0; + } break; + case GGML_TYPE_MXFP4: + { + nsg = N_SG_MXFP4; + nr0 = N_R0_MXFP4; + smem = 32*sizeof(float); + } break; + case GGML_TYPE_Q2_K: + { + nsg = N_SG_Q2_K; + nr0 = N_R0_Q2_K; + } break; + case GGML_TYPE_Q3_K: + { + nsg = N_SG_Q3_K; + nr0 = N_R0_Q3_K; + } break; + case GGML_TYPE_Q4_K: + { + nsg = N_SG_Q4_K; + nr0 = N_R0_Q4_K; + } break; + case GGML_TYPE_Q5_K: + { + nsg = N_SG_Q5_K; + nr0 = N_R0_Q5_K; + } break; + case GGML_TYPE_Q6_K: + { + nsg = N_SG_Q6_K; + nr0 = N_R0_Q6_K; + } break; + case GGML_TYPE_IQ2_XXS: + { + nsg = N_SG_IQ2_XXS; + nr0 = N_R0_IQ2_XXS; + smem = 256*8+128; + } break; + case GGML_TYPE_IQ2_XS: + { + nsg = N_SG_IQ2_XS; + nr0 = N_R0_IQ2_XS; + smem = 512*8+128; + } break; + case GGML_TYPE_IQ3_XXS: + { + nsg = N_SG_IQ3_XXS; + nr0 = N_R0_IQ3_XXS; + smem = 256*4+128; + } break; + case GGML_TYPE_IQ3_S: + { + nsg = N_SG_IQ3_S; + nr0 = N_R0_IQ3_S; + smem = 512*4; + } break; + case GGML_TYPE_IQ2_S: + { + nsg = N_SG_IQ2_S; + nr0 = N_R0_IQ2_S; + } break; + case GGML_TYPE_IQ1_S: + { + nsg = N_SG_IQ1_S; + nr0 = N_R0_IQ1_S; + } break; + case GGML_TYPE_IQ1_M: + { + nsg = N_SG_IQ1_M; + nr0 = N_R0_IQ1_M; + } break; + case GGML_TYPE_IQ4_NL: + { + nsg = N_SG_IQ4_NL; + nr0 = N_R0_IQ4_NL; + smem = 32*sizeof(float); + } break; + case GGML_TYPE_IQ4_XS: + { + nsg = N_SG_IQ4_XS; + nr0 = N_R0_IQ4_XS; + smem = 32*sizeof(float); + } break; + default: + { + GGML_LOG_ERROR("Asserting on type %d\n", (int)op->src[2]->type); + GGML_ABORT("not implemented"); + } + }; + + snprintf(base, 256, "kernel_mul_mv_id_%s_%s%s", ggml_type_name(tsrc0), ggml_type_name(tsrc1), suffix); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + ggml_metal_cv_t cv = ggml_metal_cv_init(); + + ggml_metal_cv_set_int16(cv, nsg, FC_MUL_MV + 0); + + res = ggml_metal_library_compile_pipeline(lib, base, name, cv); + + ggml_metal_cv_free(cv); + + ggml_metal_pipeline_set_nr0 (res, nr0); + ggml_metal_pipeline_set_nr1 (res, nr1); + ggml_metal_pipeline_set_nsg (res, nsg); + ggml_metal_pipeline_set_smem(res, smem); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_argmax(ggml_metal_library_t lib, const ggml_tensor * op) { + GGML_ASSERT(op->src[0]->type == GGML_TYPE_F32); + GGML_ASSERT(ggml_is_contiguous_1(op->src[0])); + GGML_ASSERT(op->src[0]->nb[0] == ggml_type_size(op->src[0]->type)); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_argmax_%s", ggml_type_name(op->src[0]->type)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + ggml_metal_pipeline_set_smem(res, 32*(sizeof(float) + sizeof(int32_t))); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_argsort(ggml_metal_library_t lib, const ggml_tensor * op) { + assert(op->op == GGML_OP_ARGSORT); + + char base[256]; + char name[256]; + + ggml_sort_order order = (ggml_sort_order) op->op_params[0]; + + const char * order_str = "undefined"; + switch (order) { + case GGML_SORT_ORDER_ASC: order_str = "asc"; break; + case GGML_SORT_ORDER_DESC: order_str = "desc"; break; + default: GGML_ABORT("fatal error"); + }; + + snprintf(base, 256, "kernel_argsort_%s_%s_%s", ggml_type_name(op->src[0]->type), ggml_type_name(op->type), order_str); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_flash_attn_ext( + ggml_metal_library_t lib, + const ggml_tensor * op, + bool has_mask, + bool has_sinks, + bool has_bias, + bool has_scap, + int32_t nsg) { + assert(op->op == GGML_OP_FLASH_ATTN_EXT); + + char base[256]; + char name[256]; + + const int32_t dk = (int32_t) op->src[1]->ne[0]; + const int32_t dv = (int32_t) op->src[2]->ne[0]; + + const int32_t ns10 = op->src[1]->nb[1]/op->src[1]->nb[0]; + const int32_t ns20 = op->src[2]->nb[1]/op->src[2]->nb[0]; + + snprintf(base, 256, "kernel_%s_%s_dk%d_dv%d", + "flash_attn_ext", + ggml_type_name(op->src[1]->type), + dk, + dv); + + snprintf(name, 256, "kernel_%s_%s_dk%d_dv%d_mask=%d_sinks=%d_bias=%d_scap=%d_ns10=%d_ns20=%d_nsg=%d", + "flash_attn_ext", + ggml_type_name(op->src[1]->type), + dk, + dv, + has_mask, + has_sinks, + has_bias, + has_scap, + ns10, + ns20, + nsg); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + ggml_metal_cv_t cv = ggml_metal_cv_init(); + + ggml_metal_cv_set_bool(cv, has_mask, FC_FLASH_ATTN_EXT + 0); + ggml_metal_cv_set_bool(cv, has_sinks, FC_FLASH_ATTN_EXT + 1); + ggml_metal_cv_set_bool(cv, has_bias, FC_FLASH_ATTN_EXT + 2); + ggml_metal_cv_set_bool(cv, has_scap, FC_FLASH_ATTN_EXT + 3); + + ggml_metal_cv_set_int32(cv, ns10, FC_FLASH_ATTN_EXT + 20); + ggml_metal_cv_set_int32(cv, ns20, FC_FLASH_ATTN_EXT + 21); + ggml_metal_cv_set_int32(cv, nsg, FC_FLASH_ATTN_EXT + 22); + + res = ggml_metal_library_compile_pipeline(lib, base, name, cv); + + ggml_metal_cv_free(cv); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_flash_attn_ext_vec( + ggml_metal_library_t lib, + const ggml_tensor * op, + bool has_mask, + bool has_sinks, + bool has_bias, + bool has_scap, + int32_t nsg, + int32_t nwg) { + assert(op->op == GGML_OP_FLASH_ATTN_EXT); + + char base[256]; + char name[256]; + + const int32_t dk = (int32_t) op->src[1]->ne[0]; + const int32_t dv = (int32_t) op->src[2]->ne[0]; + + const int32_t ns10 = op->src[1]->nb[1]/op->src[1]->nb[0]; + const int32_t ns20 = op->src[2]->nb[1]/op->src[2]->nb[0]; + + snprintf(base, 256, "kernel_%s_%s_dk%d_dv%d", + "flash_attn_ext_vec", + ggml_type_name(op->src[1]->type), + dk, + dv); + + snprintf(name, 256, "kernel_%s_%s_dk%d_dv%d_mask=%d_sink=%d_bias=%d_softcap=%d_ns10=%d_ns20=%d_nsg=%d_nwg=%d", + "flash_attn_ext_vec", + ggml_type_name(op->src[1]->type), + dk, + dv, + has_mask, + has_sinks, + has_bias, + has_scap, + ns10, + ns20, + nsg, nwg); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + ggml_metal_cv_t cv = ggml_metal_cv_init(); + + ggml_metal_cv_set_bool(cv, has_mask, FC_FLASH_ATTN_EXT_VEC + 0); + ggml_metal_cv_set_bool(cv, has_sinks, FC_FLASH_ATTN_EXT_VEC + 1); + ggml_metal_cv_set_bool(cv, has_bias, FC_FLASH_ATTN_EXT_VEC + 2); + ggml_metal_cv_set_bool(cv, has_scap, FC_FLASH_ATTN_EXT_VEC + 3); + + ggml_metal_cv_set_int32(cv, ns10, FC_FLASH_ATTN_EXT_VEC + 20); + ggml_metal_cv_set_int32(cv, ns20, FC_FLASH_ATTN_EXT_VEC + 21); + ggml_metal_cv_set_int32(cv, nsg, FC_FLASH_ATTN_EXT_VEC + 22); + ggml_metal_cv_set_int32(cv, nwg, FC_FLASH_ATTN_EXT_VEC + 23); + + res = ggml_metal_library_compile_pipeline(lib, base, name, cv); + + ggml_metal_cv_free(cv); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_flash_attn_ext_vec_reduce( + ggml_metal_library_t lib, + const ggml_tensor * op, + int32_t dv, + int32_t nwg) { + assert(op->op == GGML_OP_FLASH_ATTN_EXT); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_flash_attn_ext_vec_reduce"); + snprintf(name, 256, "kernel_flash_attn_ext_vec_reduce_dv=%d_nwg=%d", dv, nwg); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + ggml_metal_cv_t cv = ggml_metal_cv_init(); + + ggml_metal_cv_set_int32(cv, dv, FC_FLASH_ATTN_EXT_VEC_REDUCE + 0); + ggml_metal_cv_set_int32(cv, nwg, FC_FLASH_ATTN_EXT_VEC_REDUCE + 1); + + res = ggml_metal_library_compile_pipeline(lib, base, name, cv); + + ggml_metal_cv_free(cv); + + return res; + + GGML_UNUSED(op); +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_bin( + ggml_metal_library_t lib, + ggml_op op, + int32_t n_fuse, + bool row) { + char base[256]; + char name[256]; + + const char * op_str = "undefined"; + switch (op) { + case GGML_OP_ADD: op_str = "add"; break; + case GGML_OP_SUB: op_str = "sub"; break; + case GGML_OP_MUL: op_str = "mul"; break; + case GGML_OP_DIV: op_str = "div"; break; + default: GGML_ABORT("fatal error"); + }; + + if (row) { + snprintf(base, 256, "kernel_%s_row_c4_fuse_%d", op_str, n_fuse); + } else { + snprintf(base, 256, "kernel_%s_fuse_%d", op_str, n_fuse); + } + + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_rms_norm(ggml_metal_library_t lib, const ggml_tensor * op, int32_t n_fuse) { + assert(op->op == GGML_OP_RMS_NORM); + + GGML_ASSERT(op->src[0]->ne[0] % 4 == 0); + GGML_ASSERT(ggml_is_contiguous_rows(op->src[0])); + + char base[256]; + char name[256]; + + switch (n_fuse) { + case 1: snprintf(base, 256, "kernel_rms_norm_f32"); break; + case 2: snprintf(base, 256, "kernel_rms_norm_mul_f32"); break; + case 3: snprintf(base, 256, "kernel_rms_norm_mul_add_f32"); break; + default: GGML_ABORT("fatal error"); + } + + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + ggml_metal_pipeline_set_smem(res, 32*sizeof(float)); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_l2_norm(ggml_metal_library_t lib, const ggml_tensor * op) { + assert(op->op == GGML_OP_L2_NORM); + + GGML_ASSERT(op->src[0]->ne[0] % 4 == 0); + GGML_ASSERT(ggml_is_contiguous_1(op->src[0])); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_l2_norm_f32"); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + ggml_metal_pipeline_set_smem(res, 32*sizeof(float)); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_group_norm(ggml_metal_library_t lib, const ggml_tensor * op) { + assert(op->op == GGML_OP_GROUP_NORM); + + GGML_ASSERT(ggml_is_contiguous(op->src[0])); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_group_norm_f32"); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + ggml_metal_pipeline_set_smem(res, 32*sizeof(float)); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_norm(ggml_metal_library_t lib, const ggml_tensor * op) { + assert(op->op == GGML_OP_NORM); + + GGML_ASSERT(op->src[0]->ne[0] % 4 == 0); + GGML_ASSERT(ggml_is_contiguous_1(op->src[0])); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_norm_f32"); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + ggml_metal_pipeline_set_smem(res, 32*sizeof(float)); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_rope(ggml_metal_library_t lib, const ggml_tensor * op) { + assert(op->op == GGML_OP_ROPE); + + char base[256]; + char name[256]; + + const int mode = ((const int32_t *) op->op_params)[2]; + + const bool is_neox = mode & GGML_ROPE_TYPE_NEOX; + const bool is_mrope = mode & GGML_ROPE_TYPE_MROPE; + const bool is_vision = mode == GGML_ROPE_TYPE_VISION; + + if (is_neox) { + snprintf(base, 256, "kernel_rope_neox_%s", ggml_type_name(op->src[0]->type)); + } else if (is_mrope && !is_vision) { + GGML_ASSERT(op->src[1]->ne[0]*4 >= op->src[0]->ne[2]); // need at least 4 pos per token + snprintf(base, 256, "kernel_rope_multi_%s", ggml_type_name(op->src[0]->type)); + } else if (is_vision) { + GGML_ASSERT(op->src[1]->ne[0]*4 >= op->src[0]->ne[2]); // need at least 4 pos per token + snprintf(base, 256, "kernel_rope_vision_%s", ggml_type_name(op->src[0]->type)); + } else { + snprintf(base, 256, "kernel_rope_norm_%s", ggml_type_name(op->src[0]->type)); + } + + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_im2col(ggml_metal_library_t lib, const ggml_tensor * op) { + assert(op->op == GGML_OP_IM2COL); + + GGML_ASSERT(ggml_is_contiguous(op->src[1])); + GGML_ASSERT(op->src[1]->type == GGML_TYPE_F32); + GGML_ASSERT(op->type == GGML_TYPE_F16 || op->type == GGML_TYPE_F32); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_im2col_ext_%s", ggml_type_name(op->type)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_conv_transpose_1d(ggml_metal_library_t lib, const ggml_tensor * op) { + assert(op->op == GGML_OP_CONV_TRANSPOSE_1D); + + GGML_ASSERT(ggml_is_contiguous(op->src[0])); + GGML_ASSERT(ggml_is_contiguous(op->src[1])); + GGML_ASSERT(op->src[0]->type == GGML_TYPE_F16 || op->src[0]->type == GGML_TYPE_F32); + GGML_ASSERT(op->src[1]->type == GGML_TYPE_F32); + GGML_ASSERT(op->type == GGML_TYPE_F32); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_conv_transpose_1d_%s_%s", ggml_type_name(op->src[0]->type), ggml_type_name(op->src[1]->type)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_upscale(ggml_metal_library_t lib, const ggml_tensor * op) { + assert(op->op == GGML_OP_UPSCALE); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_upscale_%s", ggml_type_name(op->src[0]->type)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_pad(ggml_metal_library_t lib, const ggml_tensor * op) { + assert(op->op == GGML_OP_PAD); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_pad_%s", ggml_type_name(op->src[0]->type)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_pad_reflect_1d(ggml_metal_library_t lib, const ggml_tensor * op) { + assert(op->op == GGML_OP_PAD_REFLECT_1D); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_pad_reflect_1d_%s", ggml_type_name(op->src[0]->type)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_arange(ggml_metal_library_t lib, const ggml_tensor * op) { + assert(op->op == GGML_OP_ARANGE); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_arange_%s", ggml_type_name(op->type)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_timestep_embedding(ggml_metal_library_t lib, const ggml_tensor * op) { + assert(op->op == GGML_OP_TIMESTEP_EMBEDDING); + + char base[256]; + char name[256]; + + snprintf(base, 256, "kernel_timestep_embedding_%s", ggml_type_name(op->src[0]->type)); + snprintf(name, 256, "%s", base); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + return res; + } + + res = ggml_metal_library_compile_pipeline(lib, base, name, nullptr); + + return res; +} + diff --git a/ggml/src/ggml-metal/ggml-metal-device.h b/ggml/src/ggml-metal/ggml-metal-device.h new file mode 100644 index 0000000000..4a3a819fc0 --- /dev/null +++ b/ggml/src/ggml-metal/ggml-metal-device.h @@ -0,0 +1,227 @@ +#pragma once + +#include "ggml.h" + +#ifdef __cplusplus +extern "C" { +#endif + +struct ggml_metal_buffer_id { + void * metal; // id + size_t offs; +}; + +typedef struct ggml_metal_device * ggml_metal_device_t; + +// +// MTLFunctionConstantValues wrapper +// + +typedef struct ggml_metal_cv * ggml_metal_cv_t; + +ggml_metal_cv_t ggml_metal_cv_init(void); +void ggml_metal_cv_free(ggml_metal_cv_t cv); + +void ggml_metal_cv_set_int16(ggml_metal_cv_t cv, int16_t value, int32_t idx); +void ggml_metal_cv_set_int32(ggml_metal_cv_t cv, int32_t value, int32_t idx); +void ggml_metal_cv_set_bool (ggml_metal_cv_t cv, bool value, int32_t idx); + +// +// MTLComputePipelineState wrapper +// + +typedef struct ggml_metal_pipeline * ggml_metal_pipeline_t; + +ggml_metal_pipeline_t ggml_metal_pipeline_init(void); +void ggml_metal_pipeline_free(ggml_metal_pipeline_t pipeline); + +void ggml_metal_pipeline_set_nsg(ggml_metal_pipeline_t pipeline, int nsg); +int ggml_metal_pipeline_get_nsg(ggml_metal_pipeline_t pipeline); + +void ggml_metal_pipeline_set_nr0(ggml_metal_pipeline_t pipeline, int nr0); +int ggml_metal_pipeline_get_nr0(ggml_metal_pipeline_t pipeline); + +void ggml_metal_pipeline_set_nr1(ggml_metal_pipeline_t pipeline, int nr1); +int ggml_metal_pipeline_get_nr1(ggml_metal_pipeline_t pipeline); + +void ggml_metal_pipeline_set_smem(ggml_metal_pipeline_t pipeline, size_t smem); +size_t ggml_metal_pipeline_get_smem(ggml_metal_pipeline_t pipeline); + +int ggml_metal_pipeline_max_theads_per_threadgroup(ggml_metal_pipeline_t pipeline); + +// a collection of pipelines +typedef struct ggml_metal_pipelines * ggml_metal_pipelines_t; + +ggml_metal_pipelines_t ggml_metal_pipelines_init(void); +void ggml_metal_pipelines_free(ggml_metal_pipelines_t ppls); + +void ggml_metal_pipelines_add(ggml_metal_pipelines_t ppls, const char * name, ggml_metal_pipeline_t pipeline); +ggml_metal_pipeline_t ggml_metal_pipelines_get(ggml_metal_pipelines_t ppls, const char * name); + +// +// MTLCommandBuffer wrapper +// + +typedef void * ggml_metal_cmd_buf_t; + +// +// MTLComputeCommandEncoder wrapper +// + +typedef struct ggml_metal_encoder * ggml_metal_encoder_t; + +ggml_metal_encoder_t ggml_metal_encoder_init(ggml_metal_cmd_buf_t cmd_buf_raw, bool concurrent); +void ggml_metal_encoder_free(ggml_metal_encoder_t encoder); + +void ggml_metal_encoder_debug_group_push(ggml_metal_encoder_t encoder, const char * name); +void ggml_metal_encoder_debug_group_pop (ggml_metal_encoder_t encoder); + +void ggml_metal_encoder_set_pipeline(ggml_metal_encoder_t encoder, ggml_metal_pipeline_t pipeline); + +void ggml_metal_encoder_set_bytes (ggml_metal_encoder_t encoder, void * data, size_t size, int idx); +void ggml_metal_encoder_set_buffer(ggml_metal_encoder_t encoder, struct ggml_metal_buffer_id buffer, int idx); + +void ggml_metal_encoder_set_threadgroup_memory_size(ggml_metal_encoder_t encoder, size_t size, int idx); + +void ggml_metal_encoder_dispatch_threadgroups(ggml_metal_encoder_t encoder, int tg0, int tg1, int tg2, int tptg0, int tptg1, int tptg2); + +void ggml_metal_encoder_memory_barrier(ggml_metal_encoder_t encoder); + +void ggml_metal_encoder_end_encoding(ggml_metal_encoder_t encoder); + +// +// MTLLibrary wrapper +// + +typedef struct ggml_metal_library * ggml_metal_library_t; + +ggml_metal_library_t ggml_metal_library_init(ggml_metal_device_t dev); +void ggml_metal_library_free(ggml_metal_library_t lib); + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline (ggml_metal_library_t lib, const char * name); +ggml_metal_pipeline_t ggml_metal_library_compile_pipeline(ggml_metal_library_t lib, const char * base, const char * name, ggml_metal_cv_t cv); + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_base (ggml_metal_library_t lib, enum ggml_op op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_cpy (ggml_metal_library_t lib, enum ggml_type tsrc, enum ggml_type tdst); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_pool_2d (ggml_metal_library_t lib, const struct ggml_tensor * op, enum ggml_op_pool op_pool); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_get_rows (ggml_metal_library_t lib, enum ggml_type tsrc); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_set_rows (ggml_metal_library_t lib, enum ggml_type tdst); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_repeat (ggml_metal_library_t lib, enum ggml_type tsrc); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_unary (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_glu (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_sum_rows (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_soft_max (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_ssm_conv (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_ssm_scan (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_rwkv (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_mul_mv_ext (ggml_metal_library_t lib, enum ggml_type tsrc0, enum ggml_type tsrc1, int r1ptg); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_mul_mm (ggml_metal_library_t lib, enum ggml_type tsrc0, enum ggml_type tsrc1); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_mul_mv (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_mul_mm_id_map0 (ggml_metal_library_t lib, int ne02, int ne20); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_mul_mm_id (ggml_metal_library_t lib, enum ggml_type tsrc0, enum ggml_type tsrc1); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_mul_mv_id (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_argmax (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_argsort (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_bin (ggml_metal_library_t lib, enum ggml_op op, int32_t n_fuse, bool row); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_rms_norm (ggml_metal_library_t lib, const struct ggml_tensor * op, int32_t n_fuse); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_l2_norm (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_group_norm (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_norm (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_rope (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_im2col (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_conv_transpose_1d (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_upscale (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_pad (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_pad_reflect_1d (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_arange (ggml_metal_library_t lib, const struct ggml_tensor * op); +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_timestep_embedding(ggml_metal_library_t lib, const struct ggml_tensor * op); + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_flash_attn_ext( + ggml_metal_library_t lib, + const struct ggml_tensor * op, + bool has_mask, + bool has_sinks, + bool has_bias, + bool has_scap, + int32_t nsg); + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_flash_attn_ext_vec( + ggml_metal_library_t lib, + const struct ggml_tensor * op, + bool has_mask, + bool has_sinks, + bool has_bias, + bool has_scap, + int32_t nsg, + int32_t nwg); + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline_flash_attn_ext_vec_reduce( + ggml_metal_library_t lib, + const struct ggml_tensor * op, + int32_t dv, + int32_t nwg); + +// +// device +// + +struct ggml_metal_device_props { + char name[128]; + + size_t max_buffer_size; + size_t max_working_set_size; + size_t max_theadgroup_memory_size; + + bool has_simdgroup_reduction; + bool has_simdgroup_mm; + bool has_unified_memory; + bool has_bfloat; + bool use_residency_sets; + bool use_shared_buffers; + + bool supports_gpu_family_apple7; +}; + +ggml_metal_device_t ggml_metal_device_init(void); +void ggml_metal_device_free(ggml_metal_device_t dev); + +// return a singleton that is automatically destroyed when the program exits +ggml_metal_device_t ggml_metal_device_get(void); + +void * ggml_metal_device_get_obj (ggml_metal_device_t dev); // id +void * ggml_metal_device_get_queue(ggml_metal_device_t dev); // id + +ggml_metal_library_t ggml_metal_device_get_library(ggml_metal_device_t dev); + +void ggml_metal_device_get_memory(ggml_metal_device_t dev, size_t * free, size_t * total); +bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_tensor * op); + +const struct ggml_metal_device_props * ggml_metal_device_get_props(ggml_metal_device_t dev); + +// +// device buffers +// + +typedef struct ggml_metal_buffer * ggml_metal_buffer_t; + +ggml_metal_buffer_t ggml_metal_buffer_init(ggml_metal_device_t dev, size_t size, bool shared); +ggml_metal_buffer_t ggml_metal_buffer_map (ggml_metal_device_t dev, void * ptr, size_t size, size_t max_tensor_size); + +void ggml_metal_buffer_free (ggml_metal_buffer_t buf); +void * ggml_metal_buffer_get_base (ggml_metal_buffer_t buf); +bool ggml_metal_buffer_is_shared(ggml_metal_buffer_t buf); + +void ggml_metal_buffer_memset_tensor(ggml_metal_buffer_t buf, struct ggml_tensor * tensor, uint8_t value, size_t offset, size_t size); +void ggml_metal_buffer_set_tensor (ggml_metal_buffer_t buf, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size); +void ggml_metal_buffer_get_tensor (ggml_metal_buffer_t buf, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size); +void ggml_metal_buffer_clear (ggml_metal_buffer_t buf, uint8_t value); + +// finds the Metal buffer that contains the tensor data on the GPU device +// the assumption is that there is 1-to-1 mapping between the host and device memory buffers, so we can find the +// Metal buffer based on the host memory pointer +// +struct ggml_metal_buffer_id ggml_metal_buffer_get_id(ggml_metal_buffer_t buf, const struct ggml_tensor * t); + +#ifdef __cplusplus +} +#endif diff --git a/ggml/src/ggml-metal/ggml-metal-device.m b/ggml/src/ggml-metal/ggml-metal-device.m new file mode 100644 index 0000000000..67f71ace2b --- /dev/null +++ b/ggml/src/ggml-metal/ggml-metal-device.m @@ -0,0 +1,1303 @@ +#import "ggml-metal-device.h" + +#import "ggml-impl.h" +#import "ggml-threading.h" + +#include + +#include + +#ifndef TARGET_OS_VISION +#define TARGET_OS_VISION 0 +#endif + +// create residency sets only on macOS >= 15.0 +#if !TARGET_CPU_X86_64 && TARGET_OS_OSX && __MAC_OS_X_VERSION_MAX_ALLOWED >= 150000 || \ + TARGET_OS_IOS && __IPHONE_OS_VERSION_MAX_ALLOWED >= 180000 || \ + TARGET_OS_TV && __TV_OS_VERSION_MAX_ALLOWED >= 180000 || \ + TARGET_OS_VISION && __VISION_OS_VERSION_MAX_ALLOWED >= 200000 +#define GGML_METAL_HAS_RESIDENCY_SETS 1 +#endif + +// overload of MTLGPUFamilyMetal3 (not available in some environments) +static const NSInteger MTLGPUFamilyMetal3_GGML = 5001; + +#if !GGML_METAL_EMBED_LIBRARY +// Here to assist with NSBundle Path Hack +@interface GGMLMetalClass : NSObject +@end +@implementation GGMLMetalClass +@end +#endif + +// +// MTLFunctionConstantValues wrapper +// + +struct ggml_metal_cv { + MTLFunctionConstantValues * obj; +}; + +ggml_metal_cv_t ggml_metal_cv_init(void) { + ggml_metal_cv_t res = calloc(1, sizeof(struct ggml_metal_cv)); + + res->obj = [[MTLFunctionConstantValues alloc] init]; + + return res; +} + +void ggml_metal_cv_free(ggml_metal_cv_t cv) { + [cv->obj release]; + free(cv); +} + +void ggml_metal_cv_set_int16(ggml_metal_cv_t cv, int16_t value, int32_t idx) { + [cv->obj setConstantValue:&value type:MTLDataTypeShort atIndex:idx]; +} + +void ggml_metal_cv_set_int32(ggml_metal_cv_t cv, int32_t value, int32_t idx) { + [cv->obj setConstantValue:&value type:MTLDataTypeInt atIndex:idx]; +} + +void ggml_metal_cv_set_bool(ggml_metal_cv_t cv, bool value, int32_t idx) { + [cv->obj setConstantValue:&value type:MTLDataTypeBool atIndex:idx]; +} + +// +// MTLComputePipelineState wrapper +// + +struct ggml_metal_pipeline { + id obj; + + // suggested dispatch sizes + int nsg; + + int nr0; + int nr1; + + size_t smem; +}; + +ggml_metal_pipeline_t ggml_metal_pipeline_init(void) { + ggml_metal_pipeline_t res = calloc(1, sizeof(struct ggml_metal_pipeline)); + + *res = (struct ggml_metal_pipeline) { + /*.obj =*/ nil, + /*.nsg =*/ 0, + /*.nr0 =*/ 0, + /*.nr1 =*/ 0, + /*.smem =*/ 0, + }; + + return res; +} + +void ggml_metal_pipeline_free(ggml_metal_pipeline_t pipeline) { + [pipeline->obj release]; + + free(pipeline); +} + +void ggml_metal_pipeline_set_nsg(ggml_metal_pipeline_t pipeline, int nsg) { + pipeline->nsg = nsg; +} + +int ggml_metal_pipeline_get_nsg(ggml_metal_pipeline_t pipeline) { + return pipeline->nsg; +} + +void ggml_metal_pipeline_set_nr0(ggml_metal_pipeline_t pipeline, int nr0) { + pipeline->nr0 = nr0; +} + +int ggml_metal_pipeline_get_nr0(ggml_metal_pipeline_t pipeline) { + return pipeline->nr0; +} + +void ggml_metal_pipeline_set_nr1(ggml_metal_pipeline_t pipeline, int nr1) { + pipeline->nr1 = nr1; +} + +int ggml_metal_pipeline_get_nr1(ggml_metal_pipeline_t pipeline) { + return pipeline->nr1; +} + +void ggml_metal_pipeline_set_smem(ggml_metal_pipeline_t pipeline, size_t smem) { + pipeline->smem = smem; +} + +size_t ggml_metal_pipeline_get_smem(ggml_metal_pipeline_t pipeline) { + return pipeline->smem; +} + +int ggml_metal_pipeline_max_theads_per_threadgroup(ggml_metal_pipeline_t pipeline) { + return pipeline->obj.maxTotalThreadsPerThreadgroup; +} + +struct ggml_metal_library { + id obj; + id device; + + ggml_metal_pipelines_t pipelines; // cache of compiled pipelines +}; + +ggml_metal_library_t ggml_metal_library_init(ggml_metal_device_t dev) { + id library = nil; + id device = ggml_metal_device_get_obj(dev); + + // load library + // + // - first check if the library is embedded + // - then check if the library is in the bundle + // - if not found, load the source and compile it + // - if that fails, return NULL + // + // TODO: move to a function + { + const int64_t t_start = ggml_time_us(); + + NSError * error = nil; + NSString * src = nil; + +#if GGML_METAL_EMBED_LIBRARY + GGML_LOG_INFO("%s: using embedded metal library\n", __func__); + + extern const char ggml_metallib_start[]; + extern const char ggml_metallib_end[]; + + src = [[NSString alloc] initWithBytes:ggml_metallib_start length:(ggml_metallib_end-ggml_metallib_start) encoding:NSUTF8StringEncoding]; +#else + +#ifdef SWIFT_PACKAGE + NSBundle * bundle = SWIFTPM_MODULE_BUNDLE; +#else + NSBundle * bundle = [NSBundle bundleForClass:[GGMLMetalClass class]]; +#endif + + NSString * path_lib = [bundle pathForResource:@"default" ofType:@"metallib"]; + if (path_lib == nil) { + // Try to find the resource in the directory where the current binary located. + NSString * bin_cur = [[NSProcessInfo processInfo] arguments][0]; + NSString * bin_dir = [bin_cur stringByDeletingLastPathComponent]; + + NSString * path_lib_default = [NSString pathWithComponents:@[bin_dir, @"default.metallib"]]; + if ([[NSFileManager defaultManager] isReadableFileAtPath:path_lib_default]) { + GGML_LOG_INFO("%s: found '%s'\n", __func__, [path_lib_default UTF8String]); + + NSDictionary * atts = [[NSFileManager defaultManager] attributesOfItemAtPath:path_lib_default error:&error]; + if (atts && atts[NSFileType] == NSFileTypeSymbolicLink) { + // Optionally, if this is a symlink, try to resolve it. + path_lib_default = [[NSFileManager defaultManager] destinationOfSymbolicLinkAtPath:path_lib_default error:&error]; + if (path_lib_default && [path_lib_default length] > 0 && ![[path_lib_default substringToIndex:1] isEqualToString:@"/"]) { + // It is a relative path, adding the binary directory as directory prefix. + path_lib_default = [NSString pathWithComponents:@[bin_dir, path_lib_default]]; + } + if (!path_lib_default || ![[NSFileManager defaultManager] isReadableFileAtPath:path_lib_default]) { + // Link to the resource could not be resolved. + path_lib_default = nil; + } else { + GGML_LOG_INFO("%s: symlink resolved '%s'\n", __func__, [path_lib_default UTF8String]); + } + } + } else { + // The resource couldn't be found in the binary's directory. + path_lib_default = nil; + } + + path_lib = path_lib_default; + } + + if (path_lib != nil) { + // pre-compiled library found + NSURL * libURL = [NSURL fileURLWithPath:path_lib]; + GGML_LOG_INFO("%s: loading '%s'\n", __func__, [path_lib UTF8String]); + + library = [device newLibraryWithURL:libURL error:&error]; + if (error) { + GGML_LOG_ERROR("%s: error: %s\n", __func__, [[error description] UTF8String]); + return nil; + } + } else { + GGML_LOG_INFO("%s: default.metallib not found, loading from source\n", __func__); + + NSString * path_source; + NSString * path_resource = [[NSProcessInfo processInfo].environment objectForKey:@"GGML_METAL_PATH_RESOURCES"]; + + GGML_LOG_INFO("%s: GGML_METAL_PATH_RESOURCES = %s\n", __func__, path_resource ? [path_resource UTF8String] : "nil"); + + if (path_resource) { + path_source = [path_resource stringByAppendingPathComponent:@"ggml-metal.metal"]; + } else { + path_source = [bundle pathForResource:@"ggml-metal" ofType:@"metal"]; + } + + if (path_source == nil) { + GGML_LOG_WARN("%s: error: could not use bundle path to find ggml-metal.metal, falling back to trying cwd\n", __func__); + path_source = @"ggml-metal.metal"; + } + + GGML_LOG_INFO("%s: loading '%s'\n", __func__, [path_source UTF8String]); + + src = [NSString stringWithContentsOfFile:path_source encoding:NSUTF8StringEncoding error:&error]; + if (error) { + GGML_LOG_ERROR("%s: error: %s\n", __func__, [[error description] UTF8String]); + return nil; + } + } +#endif + + if (!library) { + @autoreleasepool { + // dictionary of preprocessor macros + NSMutableDictionary * prep = [NSMutableDictionary dictionary]; + + if (ggml_metal_device_get_props(dev)->has_bfloat) { + [prep setObject:@"1" forKey:@"GGML_METAL_HAS_BF16"]; + } + +#if GGML_METAL_EMBED_LIBRARY + [prep setObject:@"1" forKey:@"GGML_METAL_EMBED_LIBRARY"]; +#endif + + MTLCompileOptions * options = [MTLCompileOptions new]; + options.preprocessorMacros = prep; + + //[options setFastMathEnabled:false]; + + library = [device newLibraryWithSource:src options:options error:&error]; + if (error) { + GGML_LOG_ERROR("%s: error: %s\n", __func__, [[error description] UTF8String]); + return nil; + } + +#if !__has_feature(objc_arc) + [options release]; +#endif + } + } + +#if GGML_METAL_EMBED_LIBRARY + [src release]; +#endif // GGML_METAL_EMBED_LIBRARY + + GGML_LOG_INFO("%s: loaded in %.3f sec\n", __func__, (ggml_time_us() - t_start) / 1e6); + } + + ggml_metal_library_t res = calloc(1, sizeof(struct ggml_metal_library)); + + res->obj = library; + res->device = device; + res->pipelines = ggml_metal_pipelines_init(); + + return res; +} + +void ggml_metal_library_free(ggml_metal_library_t lib) { + if (!lib) { + return; + } + + if (lib->obj) { + [lib->obj release]; + } + + ggml_metal_pipelines_free(lib->pipelines); + + free(lib); +} + +ggml_metal_pipeline_t ggml_metal_library_get_pipeline(ggml_metal_library_t lib, const char * name) { + return ggml_metal_pipelines_get(lib->pipelines, name); +} + +ggml_metal_pipeline_t ggml_metal_library_compile_pipeline(ggml_metal_library_t lib, const char * base, const char * name, ggml_metal_cv_t cv) { + // note: the pipelines are cached in the library per device, so they are shared across all metal contexts + ggml_critical_section_start(); + + ggml_metal_pipeline_t res = ggml_metal_library_get_pipeline(lib, name); + if (res) { + ggml_critical_section_end(); + + return res; + } + + res = ggml_metal_pipeline_init(); + + @autoreleasepool { + NSError * error = nil; + + NSString * base_func = [NSString stringWithUTF8String:base]; + + GGML_LOG_DEBUG("%s: compiling pipeline: base = '%s', name = '%s'\n", __func__, base, name); + + id mtl_function; + if (!cv) { + mtl_function = [lib->obj newFunctionWithName:base_func]; + } else { + mtl_function = [lib->obj newFunctionWithName:base_func constantValues:cv->obj error:&error]; + } + if (!mtl_function) { + ggml_critical_section_end(); + + GGML_LOG_ERROR("%s: error: failed to compile pipeline: base = '%s', name = '%s'\n", __func__, base, name); + if (error) { + GGML_LOG_ERROR("%s: error: %s\n", __func__, [[error description] UTF8String]); + } + + return nil; + } + + res->obj = [lib->device newComputePipelineStateWithFunction:mtl_function error:&error]; + + ggml_metal_pipelines_add(lib->pipelines, name, res); + + [mtl_function release]; + + GGML_LOG_DEBUG("%s: loaded %-40s %16p | th_max = %4d | th_width = %4d\n", __func__, name, (void *) res->obj, + (int) res->obj.maxTotalThreadsPerThreadgroup, + (int) res->obj.threadExecutionWidth); + } + + ggml_critical_section_end(); + + return res; +} + +// +// MTLComputeCommandEncoder wrapper +// + +struct ggml_metal_encoder { + id obj; +}; + +ggml_metal_encoder_t ggml_metal_encoder_init(ggml_metal_cmd_buf_t cmd_buf_raw, bool concurrent) { + ggml_metal_encoder_t res = calloc(1, sizeof(struct ggml_metal_encoder)); + + id cmd_buf = (id) cmd_buf_raw; + + if (concurrent) { + res->obj = [cmd_buf computeCommandEncoderWithDispatchType: MTLDispatchTypeConcurrent]; + } else { + res->obj = [cmd_buf computeCommandEncoder]; + } + + [res->obj retain]; + + return res; +} + +void ggml_metal_encoder_free(ggml_metal_encoder_t encoder) { + [encoder->obj release]; + free(encoder); +} + +void ggml_metal_encoder_debug_group_push(ggml_metal_encoder_t encoder, const char * name) { + [encoder->obj pushDebugGroup:[NSString stringWithCString:name encoding:NSUTF8StringEncoding]]; +} + +void ggml_metal_encoder_debug_group_pop (ggml_metal_encoder_t encoder) { + [encoder->obj popDebugGroup]; +} + +void ggml_metal_encoder_set_pipeline(ggml_metal_encoder_t encoder, ggml_metal_pipeline_t pipeline) { + [encoder->obj setComputePipelineState:pipeline->obj]; +} + +void ggml_metal_encoder_set_bytes(ggml_metal_encoder_t encoder, void * data, size_t size, int idx) { + [encoder->obj setBytes:data length:size atIndex:idx]; +} + +void ggml_metal_encoder_set_buffer(ggml_metal_encoder_t encoder, struct ggml_metal_buffer_id buffer, int idx) { + [encoder->obj setBuffer:buffer.metal offset:buffer.offs atIndex:idx]; +} + +void ggml_metal_encoder_set_threadgroup_memory_size(ggml_metal_encoder_t encoder, size_t size, int idx) { + [encoder->obj setThreadgroupMemoryLength:size atIndex:idx]; +} + +void ggml_metal_encoder_dispatch_threadgroups(ggml_metal_encoder_t encoder, int tg0, int tg1, int tg2, int tptg0, int tptg1, int tptg2) { + [encoder->obj dispatchThreadgroups:MTLSizeMake(tg0, tg1, tg2) threadsPerThreadgroup:MTLSizeMake(tptg0, tptg1, tptg2)]; +} + +void ggml_metal_encoder_memory_barrier(ggml_metal_encoder_t encoder) { + [encoder->obj memoryBarrierWithScope:MTLBarrierScopeBuffers]; +} + +void ggml_metal_encoder_end_encoding(ggml_metal_encoder_t encoder) { + [encoder->obj endEncoding]; +} + +struct ggml_metal_device { + id mtl_device; + + // a single global queue shared by all Metal backends + // technically not needed for devices with unified memory, but enables discrete GPUs support + // ref: https://github.com/ggml-org/llama.cpp/pull/15906 + id mtl_queue; + + ggml_metal_library_t library; + + struct ggml_metal_device_props props; +}; + +ggml_metal_device_t ggml_metal_device_init(void) { + ggml_metal_device_t dev = calloc(1, sizeof(struct ggml_metal_device)); + + assert(dev != NULL); + + if (dev->mtl_device == nil) { + dev->mtl_device = MTLCreateSystemDefaultDevice(); + + if (dev->mtl_device) { + dev->mtl_queue = [dev->mtl_device newCommandQueue]; + if (dev->mtl_queue == nil) { + GGML_LOG_ERROR("%s: error: failed to create command queue\n", __func__); + } + + dev->props.has_simdgroup_reduction = [dev->mtl_device supportsFamily:MTLGPUFamilyApple7]; + dev->props.has_simdgroup_reduction |= [dev->mtl_device supportsFamily:MTLGPUFamilyMetal3_GGML]; + + dev->props.has_simdgroup_mm = [dev->mtl_device supportsFamily:MTLGPUFamilyApple7]; + dev->props.has_unified_memory = dev->mtl_device.hasUnifiedMemory; + + dev->props.has_bfloat = [dev->mtl_device supportsFamily:MTLGPUFamilyMetal3_GGML]; + dev->props.has_bfloat |= [dev->mtl_device supportsFamily:MTLGPUFamilyApple6]; + + dev->props.use_residency_sets = true; +#if defined(GGML_METAL_HAS_RESIDENCY_SETS) + dev->props.use_residency_sets = getenv("GGML_METAL_NO_RESIDENCY") == nil; +#endif + + dev->props.use_shared_buffers = dev->props.has_unified_memory; + + if (getenv("GGML_METAL_SHARED_BUFFERS_DISABLE") != NULL) { + dev->props.use_shared_buffers = false; + } + + dev->props.supports_gpu_family_apple7 = [dev->mtl_device supportsFamily:MTLGPUFamilyApple7]; + + dev->props.max_buffer_size = dev->mtl_device.maxBufferLength; + dev->props.max_working_set_size = dev->mtl_device.recommendedMaxWorkingSetSize; + dev->props.max_theadgroup_memory_size = dev->mtl_device.maxThreadgroupMemoryLength; + + strncpy(dev->props.name, [[dev->mtl_device name] UTF8String], sizeof(dev->props.name) - 1); + + dev->library = ggml_metal_library_init(dev); + if (!dev->library) { + GGML_LOG_ERROR("%s: error: failed to create library\n", __func__); + } + + // -------------------------------------------------- + + // print MTL GPU family: + GGML_LOG_INFO("%s: GPU name: %s\n", __func__, dev->props.name); + + // determine max supported GPU family + // https://developer.apple.com/metal/Metal-Shading-Language-Specification.pdf + // https://developer.apple.com/metal/Metal-Feature-Set-Tables.pdf + { + for (int i = MTLGPUFamilyApple1 + 20; i >= MTLGPUFamilyApple1; --i) { + if ([dev->mtl_device supportsFamily:i]) { + GGML_LOG_INFO("%s: GPU family: MTLGPUFamilyApple%d (%d)\n", __func__, i - (int) MTLGPUFamilyApple1 + 1, i); + break; + } + } + + for (int i = MTLGPUFamilyCommon1 + 5; i >= MTLGPUFamilyCommon1; --i) { + if ([dev->mtl_device supportsFamily:i]) { + GGML_LOG_INFO("%s: GPU family: MTLGPUFamilyCommon%d (%d)\n", __func__, i - (int) MTLGPUFamilyCommon1 + 1, i); + break; + } + } + + for (int i = MTLGPUFamilyMetal3_GGML + 5; i >= MTLGPUFamilyMetal3_GGML; --i) { + if ([dev->mtl_device supportsFamily:i]) { + GGML_LOG_INFO("%s: GPU family: MTLGPUFamilyMetal%d (%d)\n", __func__, i - (int) MTLGPUFamilyMetal3_GGML + 3, i); + break; + } + } + } + + GGML_LOG_INFO("%s: simdgroup reduction = %s\n", __func__, dev->props.has_simdgroup_reduction ? "true" : "false"); + GGML_LOG_INFO("%s: simdgroup matrix mul. = %s\n", __func__, dev->props.has_simdgroup_mm ? "true" : "false"); + GGML_LOG_INFO("%s: has unified memory = %s\n", __func__, dev->props.has_unified_memory ? "true" : "false"); + GGML_LOG_INFO("%s: has bfloat = %s\n", __func__, dev->props.has_bfloat ? "true" : "false"); + GGML_LOG_INFO("%s: use residency sets = %s\n", __func__, dev->props.use_residency_sets ? "true" : "false"); + GGML_LOG_INFO("%s: use shared buffers = %s\n", __func__, dev->props.use_shared_buffers ? "true" : "false"); + +#if TARGET_OS_OSX || (TARGET_OS_IOS && __clang_major__ >= 15) + if (@available(macOS 10.12, iOS 16.0, *)) { + GGML_LOG_INFO("%s: recommendedMaxWorkingSetSize = %8.2f MB\n", __func__, dev->props.max_working_set_size / 1e6); + } +#endif + } + } + + return dev; +} + +void ggml_metal_device_free(ggml_metal_device_t dev) { + assert(dev != NULL); + + ggml_metal_library_free(dev->library); + dev->library = NULL; + + if (dev->mtl_queue) { + [dev->mtl_queue release]; + dev->mtl_queue = nil; + } + + if (dev->mtl_device) { + [dev->mtl_device release]; + dev->mtl_device = nil; + } + + free(dev); +} + +void * ggml_metal_device_get_obj(ggml_metal_device_t dev) { + return dev->mtl_device; +} + +void * ggml_metal_device_get_queue(ggml_metal_device_t dev) { + return dev->mtl_queue; +} + +ggml_metal_library_t ggml_metal_device_get_library(ggml_metal_device_t dev) { + return dev->library; +} + +void ggml_metal_device_get_memory(ggml_metal_device_t dev, size_t * free, size_t * total) { + if (@available(macOS 10.12, iOS 16.0, *)) { + *total = dev->mtl_device.recommendedMaxWorkingSetSize; + *free = *total - dev->mtl_device.currentAllocatedSize; + } else { + *free = 0; + *total = 0; + } +} + +bool ggml_metal_device_supports_op(ggml_metal_device_t dev, const struct ggml_tensor * op) { + const bool has_simdgroup_mm = dev->props.has_simdgroup_mm; + const bool has_simdgroup_reduction = dev->props.has_simdgroup_reduction; + const bool has_bfloat = dev->props.has_bfloat; + + if (!has_bfloat) { + if (op->type == GGML_TYPE_BF16) { + return false; + } + + for (size_t i = 0, n = 3; i < n; ++i) { + if (op->src[i] != NULL && op->src[i]->type == GGML_TYPE_BF16) { + return false; + } + } + } + + switch (op->op) { + case GGML_OP_UNARY: + switch (ggml_get_unary_op(op)) { + case GGML_UNARY_OP_TANH: + case GGML_UNARY_OP_RELU: + case GGML_UNARY_OP_SIGMOID: + case GGML_UNARY_OP_GELU: + case GGML_UNARY_OP_GELU_ERF: + case GGML_UNARY_OP_GELU_QUICK: + case GGML_UNARY_OP_SILU: + case GGML_UNARY_OP_ELU: + case GGML_UNARY_OP_NEG: + case GGML_UNARY_OP_ABS: + case GGML_UNARY_OP_SGN: + case GGML_UNARY_OP_STEP: + case GGML_UNARY_OP_HARDSWISH: + case GGML_UNARY_OP_HARDSIGMOID: + case GGML_UNARY_OP_EXP: + return ggml_is_contiguous(op->src[0]) && op->src[0]->type == GGML_TYPE_F32; + default: + return false; + } + case GGML_OP_GLU: + switch (ggml_get_glu_op(op)) { + case GGML_GLU_OP_REGLU: + case GGML_GLU_OP_GEGLU: + case GGML_GLU_OP_SWIGLU: + case GGML_GLU_OP_SWIGLU_OAI: + case GGML_GLU_OP_GEGLU_ERF: + case GGML_GLU_OP_GEGLU_QUICK: + return ggml_is_contiguous_1(op->src[0]) && op->src[0]->type == GGML_TYPE_F32; + default: + return false; + } + case GGML_OP_NONE: + case GGML_OP_RESHAPE: + case GGML_OP_VIEW: + case GGML_OP_TRANSPOSE: + case GGML_OP_PERMUTE: + case GGML_OP_CONCAT: + return true; + case GGML_OP_ADD: + case GGML_OP_SUB: + case GGML_OP_MUL: + case GGML_OP_DIV: + case GGML_OP_ADD_ID: + return op->src[0]->type == GGML_TYPE_F32; + case GGML_OP_ACC: + case GGML_OP_REPEAT: + case GGML_OP_SCALE: + case GGML_OP_CONV_TRANSPOSE_1D: + return true; + case GGML_OP_CLAMP: + return op->src[0]->type == GGML_TYPE_F32; + case GGML_OP_SQR: + case GGML_OP_SQRT: + case GGML_OP_SIN: + case GGML_OP_COS: + case GGML_OP_LOG: + return ggml_is_contiguous(op->src[0]) && op->src[0]->type == GGML_TYPE_F32; + case GGML_OP_SUM_ROWS: + case GGML_OP_MEAN: + case GGML_OP_SOFT_MAX: + case GGML_OP_GROUP_NORM: + return has_simdgroup_reduction && ggml_is_contiguous_rows(op->src[0]); + case GGML_OP_RMS_NORM: + case GGML_OP_L2_NORM: + return has_simdgroup_reduction && (op->ne[0] % 4 == 0 && ggml_is_contiguous_1(op->src[0])); + case GGML_OP_ARGMAX: + return has_simdgroup_reduction; + case GGML_OP_NORM: + return has_simdgroup_reduction && (op->ne[0] % 4 == 0 && ggml_is_contiguous_1(op->src[0])); + case GGML_OP_ROPE: + return true; + case GGML_OP_IM2COL: + return ggml_is_contiguous(op->src[1]) && op->src[1]->type == GGML_TYPE_F32 && (op->type == GGML_TYPE_F16 || op->type == GGML_TYPE_F32); + case GGML_OP_POOL_1D: + return false; + case GGML_OP_UPSCALE: + return op->src[0]->type == GGML_TYPE_F32 && op->op_params[0] == GGML_SCALE_MODE_NEAREST; + case GGML_OP_POOL_2D: + return op->src[0]->type == GGML_TYPE_F32; + case GGML_OP_PAD: + return (ggml_get_op_params_i32(op, 0) == 0) && (ggml_get_op_params_i32(op, 2) == 0) && + (ggml_get_op_params_i32(op, 4) == 0) && (ggml_get_op_params_i32(op, 6) == 0); + case GGML_OP_PAD_REFLECT_1D: + case GGML_OP_TIMESTEP_EMBEDDING: + case GGML_OP_ARGSORT: + case GGML_OP_LEAKY_RELU: + return op->src[0]->type == GGML_TYPE_F32; + case GGML_OP_ARANGE: + return true; + case GGML_OP_FLASH_ATTN_EXT: + // for new head sizes, add checks here + if (op->src[0]->ne[0] != 40 && + op->src[0]->ne[0] != 64 && + op->src[0]->ne[0] != 80 && + op->src[0]->ne[0] != 96 && + op->src[0]->ne[0] != 112 && + op->src[0]->ne[0] != 128 && + op->src[0]->ne[0] != 192 && + op->src[0]->ne[0] != 256) { + return false; + } + if (op->src[0]->ne[0] == 576) { + // DeepSeek sizes + // TODO: disabled for now, until optmized + return false; + } + if (op->src[1]->type != op->src[2]->type) { + return false; + } + return has_simdgroup_mm; // TODO: over-restricted for vec-kernels + case GGML_OP_SSM_CONV: + case GGML_OP_SSM_SCAN: + return has_simdgroup_reduction; + case GGML_OP_RWKV_WKV6: + case GGML_OP_RWKV_WKV7: + return true; + case GGML_OP_MUL_MAT: + case GGML_OP_MUL_MAT_ID: + return has_simdgroup_reduction && + (op->src[0]->type != GGML_TYPE_F32 || op->src[1]->type == GGML_TYPE_F32); + case GGML_OP_CPY: + case GGML_OP_DUP: + case GGML_OP_CONT: + { + switch (op->src[0]->type) { + case GGML_TYPE_F32: + switch (op->type) { + case GGML_TYPE_F32: + case GGML_TYPE_F16: + case GGML_TYPE_BF16: + case GGML_TYPE_Q8_0: + case GGML_TYPE_Q4_0: + case GGML_TYPE_Q4_1: + case GGML_TYPE_Q5_0: + case GGML_TYPE_Q5_1: + case GGML_TYPE_IQ4_NL: + case GGML_TYPE_I32: + return true; + default: + return false; + } + case GGML_TYPE_F16: + switch (op->type) { + case GGML_TYPE_F32: + case GGML_TYPE_F16: + return true; + default: + return false; + } + case GGML_TYPE_BF16: + switch (op->type) { + case GGML_TYPE_F32: + case GGML_TYPE_BF16: + return true; + default: + return false; + } + case GGML_TYPE_Q4_0: + case GGML_TYPE_Q4_1: + case GGML_TYPE_Q5_0: + case GGML_TYPE_Q5_1: + case GGML_TYPE_Q8_0: + switch (op->type) { + case GGML_TYPE_F32: + case GGML_TYPE_F16: + return true; + default: + return false; + } + case GGML_TYPE_I32: + return op->type == GGML_TYPE_F32; + default: + return false; + }; + } + case GGML_OP_GET_ROWS: + { + return op->ne[3] == 1; + } + case GGML_OP_SET_ROWS: + { + if (op->src[0]->type != GGML_TYPE_F32) { + return false; + } + + switch (op->type) { + case GGML_TYPE_F32: + case GGML_TYPE_F16: + case GGML_TYPE_BF16: + case GGML_TYPE_Q8_0: + case GGML_TYPE_Q4_0: + case GGML_TYPE_Q4_1: + case GGML_TYPE_Q5_0: + case GGML_TYPE_Q5_1: + case GGML_TYPE_IQ4_NL: + return true; + default: + return false; + }; + } + default: + return false; + } +} + +const struct ggml_metal_device_props * ggml_metal_device_get_props(ggml_metal_device_t dev) { + return &dev->props; +} + +// +// device buffers +// + +// max memory buffers that can be mapped to the device +#define GGML_METAL_MAX_BUFFERS 64 + +struct ggml_metal_buffer_wrapper { + void * data; + size_t size; + + id metal; +}; + +struct ggml_metal_buffer { + void * all_data; // TODO: https://github.com/ggml-org/llama.cpp/pull/15985 + size_t all_size; + + // if false, the Metal buffer data is allocated in private GPU memory and is not shared with the host + bool is_shared; + bool owned; + + // multiple buffers are used only to avoid the maximum buffer size limitation when using mmap + int n_buffers; + struct ggml_metal_buffer_wrapper buffers[GGML_METAL_MAX_BUFFERS]; + + bool use_residency_sets; + + // optional MTLResidencySet + // note: cannot use explicity "id" here because it is not available on certain OSes + id rset; + + // pointers to global device objects + id device; + id queue; +}; + +static void ggml_metal_log_allocated_size(id device, size_t size_aligned) { +#ifndef GGML_METAL_NDEBUG +#if TARGET_OS_OSX || (TARGET_OS_IOS && __clang_major__ >= 15) + if (@available(macOS 10.12, iOS 16.0, *)) { + GGML_LOG_DEBUG("%s: allocated buffer, size = %8.2f MiB, (%8.2f / %8.2f)\n", + __func__, + size_aligned / 1024.0 / 1024.0, + device.currentAllocatedSize / 1024.0 / 1024.0, + device.recommendedMaxWorkingSetSize / 1024.0 / 1024.0); + + if (device.currentAllocatedSize > device.recommendedMaxWorkingSetSize) { + GGML_LOG_WARN("%s: warning: current allocated size is greater than the recommended max working set size\n", __func__); + } + } else { + GGML_LOG_INFO("%s: allocated buffer, size = %8.2f MiB, (%8.2f)\n", + __func__, + size_aligned / 1024.0 / 1024.0, + device.currentAllocatedSize / 1024.0 / 1024.0); + } +#endif +#endif + GGML_UNUSED(device); + GGML_UNUSED(size_aligned); +} + +// rset init +static bool ggml_metal_buffer_rset_init(ggml_metal_buffer_t buf) { + buf->rset = nil; + + if (!buf->use_residency_sets) { + return true; + } + +#if defined(GGML_METAL_HAS_RESIDENCY_SETS) + if (@available(macOS 15.0, iOS 18.0, tvOS 18.0, visionOS 2.0, *)) { + MTLResidencySetDescriptor * desc = [[MTLResidencySetDescriptor alloc] init]; + desc.label = @"ggml_metal"; + desc.initialCapacity = buf->n_buffers; + + NSError * error; + buf->rset = [buf->device newResidencySetWithDescriptor:desc error:&error]; + if (error) { + GGML_LOG_ERROR("%s: error: %s\n", __func__, [[error description] UTF8String]); + [desc release]; + return false; + } + + [desc release]; + + for (int i = 0; i < buf->n_buffers; i++) { + [buf->rset addAllocation:buf->buffers[i].metal]; + } + + [buf->rset commit]; + [buf->rset requestResidency]; + + return true; + } +#endif + + return true; +} + +// rset free +static void ggml_metal_buffer_rset_free(ggml_metal_buffer_t buf) { +#if defined(GGML_METAL_HAS_RESIDENCY_SETS) + if (@available(macOS 15.0, iOS 18.0, tvOS 18.0, visionOS 2.0, *)) { + if (buf->rset) { + [buf->rset endResidency]; + [buf->rset removeAllAllocations]; + [buf->rset release]; + } + } +#else + GGML_UNUSED(buf); +#endif +} + +static void * ggml_metal_host_malloc(size_t n) { + void * data = NULL; + +#if TARGET_OS_OSX + kern_return_t err = vm_allocate((vm_map_t) mach_task_self(), (void *) &data, n, VM_FLAGS_ANYWHERE); + if (err != KERN_SUCCESS) { + GGML_LOG_ERROR("%s: error: vm_allocate failed\n", __func__); + return NULL; + } +#else + const int result = posix_memalign((void **) &data, sysconf(_SC_PAGESIZE), n); + if (result != 0) { + GGML_LOG_ERROR("%s: error: posix_memalign failed\n", __func__); + return NULL; + } +#endif + + return data; +} + +ggml_metal_buffer_t ggml_metal_buffer_init(ggml_metal_device_t dev, size_t size, bool shared) { + ggml_metal_buffer_t res = calloc(1, sizeof(struct ggml_metal_buffer)); + + const size_t size_page = sysconf(_SC_PAGESIZE); + + size_t size_aligned = size; + if ((size_aligned % size_page) != 0) { + size_aligned += (size_page - (size_aligned % size_page)); + } + + const struct ggml_metal_device_props * props_dev = ggml_metal_device_get_props(dev); + + shared = shared && props_dev->use_shared_buffers; + + // allocate shared buffer if the device supports it and it is required by the buffer type + if (shared) { + res->all_data = ggml_metal_host_malloc(size_aligned); + res->is_shared = true; + res->owned = true; + } else { + // dummy, non-NULL value - we'll populate this after creating the Metal buffer below + res->all_data = (void *) 0x000000400ULL; + res->is_shared = false; + } + res->all_size = size_aligned; + + res->device = ggml_metal_device_get_obj(dev); + res->queue = ggml_metal_device_get_queue(dev); + + res->n_buffers = 1; + + if (res->all_data != NULL) { + res->buffers[0].size = size; + res->buffers[0].metal = nil; + + if (size_aligned > 0) { + if (props_dev->use_shared_buffers &&shared) { + res->buffers[0].metal = [res->device newBufferWithBytesNoCopy:res->all_data + length:size_aligned + options:MTLResourceStorageModeShared + deallocator:nil]; + } else { + res->buffers[0].metal = [res->device newBufferWithLength:size_aligned options:MTLResourceStorageModePrivate]; + + res->all_data = (void *) (res->buffers[0].metal.gpuAddress); + } + } + + res->buffers[0].data = res->all_data; + } + + if (size_aligned > 0 && (res->all_data == NULL || res->buffers[0].metal == nil)) { + GGML_LOG_ERROR("%s: error: failed to allocate buffer, size = %8.2f MiB\n", __func__, size_aligned / 1024.0 / 1024.0); + free(res); + return NULL; + } + + res->use_residency_sets = props_dev->use_residency_sets; + + if (!ggml_metal_buffer_rset_init(res)) { + GGML_LOG_ERROR("%s: error: failed to initialize residency set\n", __func__); + free(res); + return NULL; + } + + //ggml_metal_log_allocated_size(device, size_aligned); + + return res; +} + +ggml_metal_buffer_t ggml_metal_buffer_map(ggml_metal_device_t dev, void * ptr, size_t size, size_t max_tensor_size) { + ggml_metal_buffer_t res = calloc(1, sizeof(struct ggml_metal_buffer)); + + res->all_data = ptr; + res->all_size = size; + + res->is_shared = true; + res->owned = false; + + res->n_buffers = 0; + + const size_t size_page = sysconf(_SC_PAGESIZE); + + // page-align the data ptr + { + const uintptr_t offs = (uintptr_t) ptr % size_page; + ptr = (void *) ((char *) ptr - offs); + size += offs; + } + + size_t size_aligned = size; + if ((size_aligned % size_page) != 0) { + size_aligned += (size_page - (size_aligned % size_page)); + } + + res->device = ggml_metal_device_get_obj(dev); + res->queue = ggml_metal_device_get_queue(dev); + + const struct ggml_metal_device_props * props_dev = ggml_metal_device_get_props(dev); + + // the buffer fits into the max buffer size allowed by the device + if (size_aligned <= props_dev->max_buffer_size) { + res->buffers[res->n_buffers].data = ptr; + res->buffers[res->n_buffers].size = size; + res->buffers[res->n_buffers].metal = nil; + + if (size_aligned > 0) { + res->buffers[res->n_buffers].metal = [res->device newBufferWithBytesNoCopy:ptr length:size_aligned options:MTLResourceStorageModeShared deallocator:nil]; + + if (res->buffers[res->n_buffers].metal == nil) { + GGML_LOG_ERROR("%s: error: failed to allocate buffer, size = %8.2f MiB\n", __func__, size_aligned / 1024.0 / 1024.0); + free(res); + return NULL; + } + } + + ggml_metal_log_allocated_size(res->device, size_aligned); + + ++res->n_buffers; + } else { + // this overlap between the views will guarantee that the tensor with the maximum size will fully fit into + // one of the views + const size_t size_ovlp = ((max_tensor_size + size_page - 1) / size_page + 1) * size_page; // round-up 2 pages just in case + const size_t size_step = props_dev->max_buffer_size - size_ovlp; + const size_t size_view = props_dev->max_buffer_size; + + for (size_t i = 0; i < size; i += size_step) { + const size_t size_step_aligned = (i + size_view <= size) ? size_view : (size_aligned - i); + + res->buffers[res->n_buffers].data = (void *) ((uint8_t *) ptr + i); + res->buffers[res->n_buffers].size = size_step_aligned; + res->buffers[res->n_buffers].metal = nil; + + if (size_step_aligned > 0) { + res->buffers[res->n_buffers].metal = [res->device newBufferWithBytesNoCopy:(void *) ((uint8_t *) ptr + i) length:size_step_aligned options:MTLResourceStorageModeShared deallocator:nil]; + + if (res->buffers[res->n_buffers].metal == nil) { + GGML_LOG_ERROR("%s: error: failed to allocate buffer, size = %8.2f MiB\n", __func__, size_step_aligned / 1024.0 / 1024.0); + free(res); + return NULL; + } + } + + ggml_metal_log_allocated_size(res->device, size_step_aligned); + + if (i + size_step < size) { + GGML_LOG_INFO("\n"); + } + + ++res->n_buffers; + } + } + + res->use_residency_sets = props_dev->use_residency_sets; + + if (!ggml_metal_buffer_rset_init(res)) { + GGML_LOG_ERROR("%s: error: failed to initialize residency set\n", __func__); + free(res); + return NULL; + } + + return res; +} + +void ggml_metal_buffer_free(ggml_metal_buffer_t buf) { + for (int i = 0; i < buf->n_buffers; i++) { + [buf->buffers[i].metal release]; + } + + ggml_metal_buffer_rset_free(buf); + + if (buf->is_shared && buf->owned) { +#if TARGET_OS_OSX + vm_deallocate((vm_map_t)mach_task_self(), (vm_address_t)buf->all_data, buf->all_size); +#else + free(buf->all_data); +#endif + } + + free(buf); +} + +void * ggml_metal_buffer_get_base(ggml_metal_buffer_t buf) { + return buf->all_data; +} + +bool ggml_metal_buffer_is_shared(ggml_metal_buffer_t buf) { + return buf->is_shared; +} + +void ggml_metal_buffer_memset_tensor(ggml_metal_buffer_t buf, struct ggml_tensor * tensor, uint8_t value, size_t offset, size_t size) { + if (buf->is_shared) { + memset((char *)tensor->data + offset, value, size); + return; + } + + @autoreleasepool { + // dst + struct ggml_metal_buffer_id bid_dst = ggml_metal_buffer_get_id(buf, tensor); + bid_dst.offs += offset; + + id queue = buf->queue; + id cmd_buf = [queue commandBufferWithUnretainedReferences]; + + { + id encoder = [cmd_buf blitCommandEncoder]; + + [encoder fillBuffer:bid_dst.metal + range:NSMakeRange(bid_dst.offs, bid_dst.offs + size) + value:value]; + + [encoder endEncoding]; + } + + [cmd_buf commit]; + [cmd_buf waitUntilCompleted]; + } +} + +void ggml_metal_buffer_set_tensor(ggml_metal_buffer_t buf, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { + if (buf->is_shared) { + memcpy((char *)tensor->data + offset, data, size); + return; + } + + @autoreleasepool { + // src + void * data_ptr = (void *)(uintptr_t) data; // "const cast" the src data + id buf_src = [buf->device newBufferWithBytesNoCopy:data_ptr + length:size + options:MTLResourceStorageModeShared + deallocator:nil]; + + // dst + struct ggml_metal_buffer_id bid_dst = ggml_metal_buffer_get_id(buf, tensor); + bid_dst.offs += offset; + + // note: for experimentation purposes, here we use a semaphore to wait for the copy to complete + // this is alternative to waitUntilCompleted, which should be faster, but don't seem to make much difference + dispatch_semaphore_t completion_semaphore = dispatch_semaphore_create(0); + + id queue = buf->queue; + id cmd_buf = [queue commandBufferWithUnretainedReferences]; + + { + id encoder = [cmd_buf blitCommandEncoder]; + + [encoder copyFromBuffer:buf_src + sourceOffset:0 + toBuffer:bid_dst.metal + destinationOffset:bid_dst.offs + size:size]; + + [encoder endEncoding]; + } + + [cmd_buf addCompletedHandler:^(id cb) { + // TODO: can check for errors here + GGML_UNUSED(cb); + + dispatch_semaphore_signal(completion_semaphore); + }]; + + [cmd_buf commit]; + + dispatch_semaphore_wait(completion_semaphore, DISPATCH_TIME_FOREVER); + dispatch_release(completion_semaphore); + + //[cmd_buf waitUntilCompleted]; + } +} + +void ggml_metal_buffer_get_tensor(ggml_metal_buffer_t buf, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { + if (buf->is_shared) { + memcpy(data, (const char *)tensor->data + offset, size); + return; + } + + @autoreleasepool { + // src + struct ggml_metal_buffer_id bid_src = ggml_metal_buffer_get_id(buf, tensor); + bid_src.offs += offset; + + // dst + id buf_dst = [buf->device newBufferWithBytesNoCopy:data + length:size + options:MTLResourceStorageModeShared + deallocator:nil]; + + id queue = buf->queue; + id cmd_buf = [queue commandBufferWithUnretainedReferences]; + + { + id encoder = [cmd_buf blitCommandEncoder]; + + [encoder copyFromBuffer:bid_src.metal + sourceOffset:bid_src.offs + toBuffer:buf_dst + destinationOffset:0 + size:size]; + + [encoder endEncoding]; + } + + [cmd_buf commit]; + [cmd_buf waitUntilCompleted]; + } +} + +void ggml_metal_buffer_clear(ggml_metal_buffer_t buf, uint8_t value) { + if (buf->is_shared) { + memset(buf->all_data, value, buf->all_size); + return; + } + + @autoreleasepool { + id queue = buf->queue; + id cmd_buf = [queue commandBufferWithUnretainedReferences]; + + { + id encoder = [cmd_buf blitCommandEncoder]; + + [encoder fillBuffer:buf->buffers[0].metal + range:NSMakeRange(0, buf->buffers[0].size) + value:value]; + + [encoder endEncoding]; + } + + [cmd_buf commit]; + [cmd_buf waitUntilCompleted]; + } +} + +struct ggml_metal_buffer_id ggml_metal_buffer_get_id(ggml_metal_buffer_t buf, const struct ggml_tensor * t) { + struct ggml_metal_buffer_id res = { nil, 0 }; + + const int64_t tsize = ggml_nbytes(t); + + // find the view that contains the tensor fully + for (int i = 0; i < buf->n_buffers; ++i) { + const int64_t ioffs = (int64_t) t->data - (int64_t) buf->buffers[i].data; + + //GGML_LOG_INFO("ioffs = %10ld, tsize = %10ld, sum = %10ld, buf->buffers[%d].size = %10ld\n", ioffs, tsize, ioffs + tsize, i, buf->buffers[i].size); + if (ioffs >= 0 && ioffs + tsize <= (int64_t) buf->buffers[i].size) { + res.metal = buf->buffers[i].metal; + res.offs = (size_t) ioffs; + + //GGML_LOG_INFO("%s: tensor '%16s', offs = %8ld\n", __func__, t->name, *offs); + + return res; + } + } + + GGML_LOG_ERROR("%s: error: tensor '%s' buffer is nil\n", __func__, t->name); + + return res; +} diff --git a/ggml/src/ggml-metal/ggml-metal-impl.h b/ggml/src/ggml-metal/ggml-metal-impl.h index 651943fa92..b257296540 100644 --- a/ggml/src/ggml-metal/ggml-metal-impl.h +++ b/ggml/src/ggml-metal/ggml-metal-impl.h @@ -8,6 +8,9 @@ // // TODO: for optimal performance, become function of the device and work size +#define N_R0_F 2 +#define N_SG_F 4 + #define N_R0_Q4_0 4 #define N_SG_Q4_0 2 @@ -72,6 +75,7 @@ #define FC_FLASH_ATTN_EXT 100 #define FC_FLASH_ATTN_EXT_VEC 200 #define FC_FLASH_ATTN_EXT_VEC_REDUCE 300 +#define FC_MUL_MV 400 // kernel argument structs // @@ -165,6 +169,16 @@ typedef struct { uint64_t nb3; } ggml_metal_kargs_repeat; +typedef struct { + float scale; + float bias; +} ggml_metal_kargs_scale; + +typedef struct { + float min; + float max; +} ggml_metal_kargs_clamp; + typedef struct { int64_t ne00; int64_t ne01; @@ -453,7 +467,7 @@ typedef struct { uint64_t nb00; uint64_t nb01; uint64_t nb02; - int32_t n_groups; + int32_t ngrp; float eps; } ggml_metal_kargs_group_norm; @@ -506,14 +520,6 @@ typedef struct { uint64_t nb01; uint64_t nb02; uint64_t nb03; - int64_t ne10; - int64_t ne11; - int64_t ne12; - int64_t ne13; - uint64_t nb10; - uint64_t nb11; - uint64_t nb12; - uint64_t nb13; int64_t ne0; int64_t ne1; int64_t ne2; @@ -547,12 +553,6 @@ typedef struct { int32_t n_head_log2; } ggml_metal_kargs_soft_max; -typedef struct { - int64_t ne00; - int64_t ne01; - int n_past; -} ggml_metal_kargs_diag_mask_inf; - typedef struct { int64_t ne00; int64_t ne01; @@ -579,7 +579,7 @@ typedef struct { int64_t n_group; int64_t n_seq_tokens; int64_t n_seqs; - int64_t s_off; + uint64_t s_off; uint64_t nb01; uint64_t nb02; uint64_t nb03; @@ -719,7 +719,12 @@ typedef struct { int64_t IW; int64_t OH; int64_t OW; - int64_t parallel_elements; + int64_t np; } ggml_metal_kargs_pool_2d; +typedef struct { + int64_t ne00; + uint64_t nb01; +} ggml_metal_kargs_argmax; + #endif // GGML_METAL_IMPL diff --git a/ggml/src/ggml-metal/ggml-metal-ops.cpp b/ggml/src/ggml-metal/ggml-metal-ops.cpp new file mode 100644 index 0000000000..a28206e786 --- /dev/null +++ b/ggml/src/ggml-metal/ggml-metal-ops.cpp @@ -0,0 +1,3194 @@ +#include "ggml-metal-ops.h" + +#include "ggml.h" +#include "ggml-impl.h" +#include "ggml-backend-impl.h" + +#include "ggml-metal-impl.h" +#include "ggml-metal-common.h" +#include "ggml-metal-device.h" + +#include +#include + +static ggml_metal_buffer_id ggml_metal_get_buffer_id(const ggml_tensor * t) { + if (!t) { + return { nullptr, 0 }; + } + + ggml_backend_buffer_t buffer = t->view_src ? t->view_src->buffer : t->buffer; + + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t) buffer->context; + + return ggml_metal_buffer_get_id(ctx, t); +} + +struct ggml_metal_op { + ggml_metal_device_t dev; + ggml_metal_library_t lib; + ggml_metal_encoder_t enc; + ggml_mem_ranges_t mem_ranges; + + ggml_cgraph * gf; + + int idx_start; + int idx_end; + + bool use_fusion; + bool use_concurrency; + bool use_capture; + + int debug_graph; + int debug_fusion; +}; + +ggml_metal_op_t ggml_metal_op_init( + ggml_metal_device_t dev, + ggml_metal_cmd_buf_t cmd_buf, + ggml_cgraph * gf, + int idx_start, + int idx_end, + bool use_fusion, + bool use_concurrency, + bool use_capture, + int debug_graph, + int debug_fusion) { + ggml_metal_op_t res = new ggml_metal_op(); + + *res = { + /*.dev =*/ dev, + /*.lib =*/ ggml_metal_device_get_library(dev), + /*.enc =*/ ggml_metal_encoder_init(cmd_buf, use_concurrency), + /*.mem_ranges =*/ ggml_mem_ranges_init(debug_graph), + /*.gf =*/ gf, + /*.idx_start =*/ idx_start, + /*.idx_end =*/ idx_end, + /*.use_fusion =*/ use_fusion, + /*.use_concurrency =*/ use_concurrency, + /*.use_capture =*/ use_capture, + /*.debug_graph =*/ debug_graph, + /*.debug_fusion =*/ debug_fusion, + }; + + return res; +} + +void ggml_metal_op_free(ggml_metal_op_t ctx) { + ggml_metal_encoder_end_encoding(ctx->enc); + ggml_metal_encoder_free(ctx->enc); + ggml_mem_ranges_free(ctx->mem_ranges); + + delete ctx; +} + +static bool ggml_metal_op_concurrency_reset(ggml_metal_op_t ctx) { + if (!ctx->mem_ranges) { + return true; + } + + ggml_metal_encoder_memory_barrier(ctx->enc); + + ggml_mem_ranges_reset(ctx->mem_ranges); + + return true; +} + +static bool ggml_metal_op_concurrency_check(ggml_metal_op_t ctx, const ggml_tensor * node) { + if (!ctx->mem_ranges) { + return false; + } + + return ggml_mem_ranges_check(ctx->mem_ranges, node); +} + +static bool ggml_metal_op_concurrency_add(ggml_metal_op_t ctx, const ggml_tensor * node) { + if (!ctx->mem_ranges) { + return true; + } + + return ggml_mem_ranges_add(ctx->mem_ranges, node); +} + +static int ggml_metal_op_encode_impl(ggml_metal_op_t ctx, int idx) { + struct ggml_cgraph * gf = ctx->gf; + + struct ggml_tensor ** nodes = ggml_graph_nodes(gf) + idx; + struct ggml_tensor * node = nodes[0]; + + //GGML_LOG_INFO("%s: encoding node %3d, op = %8s\n", __func__, idx, ggml_op_name(node->op)); + + if (ggml_is_empty(node)) { + return 1; + } + + switch (node->op) { + case GGML_OP_NONE: + case GGML_OP_RESHAPE: + case GGML_OP_VIEW: + case GGML_OP_TRANSPOSE: + case GGML_OP_PERMUTE: + { + // noop -> next node + } return 1; + default: + { + } break; + } + + if (!ggml_metal_device_supports_op(ctx->dev, node)) { + GGML_LOG_ERROR("%s: error: unsupported op '%s'\n", __func__, ggml_op_desc(node)); + GGML_ABORT("unsupported op"); + } + + int n_fuse = 1; + + // check if the current node can run concurrently with other nodes before it + // the condition is that: + // - the current node cannot write to any previous src or dst ranges + // - the current node cannot read from any previous dst ranges + // + // if the condition is not satisfied, we put a memory barrier and clear all ranges + // otherwise, we add the new ranges to the encoding context and process the node concurrently + // + { + const bool is_concurrent = ggml_metal_op_concurrency_check(ctx, node); + + if (!is_concurrent) { + ggml_metal_op_concurrency_reset(ctx); + } + + if (ctx->debug_graph > 0) { + GGML_LOG_DEBUG("%s: node[%5d] - %-12s %s\n", __func__, idx, ggml_op_name(node->op), is_concurrent ? "(concurrent)" : ""); + } + if (ctx->debug_graph > 1) { + GGML_TENSOR_LOCALS( int64_t, ne0, node->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, node->src[0], nb); + GGML_TENSOR_LOCALS( int64_t, ne1, node->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, node->src[1], nb); + GGML_TENSOR_LOCALS( int64_t, ne, node, ne); + GGML_TENSOR_LOCALS(uint64_t, nb, node, nb); + + if (node->src[0]) { + GGML_LOG_DEBUG("%s: src0 - %4s [%5lld, %5lld, %5lld, %5lld] [%5lld, %5lld, %5lld, %5lld], %d, %s\n", __func__, ggml_type_name(node->src[0]->type), ne00, ne01, ne02, ne03, nb00, nb01, nb02, nb03, + ggml_is_contiguous(node->src[0]), node->src[0]->name); + } + if (node->src[1]) { + GGML_LOG_DEBUG("%s: src1 - %4s [%5lld, %5lld, %5lld, %5lld] [%5lld, %5lld, %5lld, %5lld], %d, %s\n", __func__, ggml_type_name(node->src[1]->type), ne10, ne11, ne12, ne13, nb10, nb11, nb12, nb13, + ggml_is_contiguous(node->src[1]), node->src[1]->name); + } + if (node) { + GGML_LOG_DEBUG("%s: node - %4s [%5lld, %5lld, %5lld, %5lld] [%5lld, %5lld, %5lld, %5lld], 1, %s\n", __func__, ggml_type_name(node->type), ne0, ne1, ne2, ne3, nb0, nb1, nb2, nb3, + node->name); + } + } + } + + switch (node->op) { + case GGML_OP_CONCAT: + { + n_fuse = ggml_metal_op_concat(ctx, idx); + } break; + case GGML_OP_ADD: + case GGML_OP_SUB: + case GGML_OP_MUL: + case GGML_OP_DIV: + { + n_fuse = ggml_metal_op_bin(ctx, idx); + } break; + case GGML_OP_ADD_ID: + { + n_fuse = ggml_metal_op_add_id(ctx, idx); + } break; + case GGML_OP_REPEAT: + { + n_fuse = ggml_metal_op_repeat(ctx, idx); + } break; + case GGML_OP_ACC: + { + n_fuse = ggml_metal_op_acc(ctx, idx); + } break; + case GGML_OP_SCALE: + { + n_fuse = ggml_metal_op_scale(ctx, idx); + } break; + case GGML_OP_CLAMP: + { + n_fuse = ggml_metal_op_clamp(ctx, idx); + } break; + case GGML_OP_SQR: + case GGML_OP_SQRT: + case GGML_OP_SIN: + case GGML_OP_COS: + case GGML_OP_LOG: + case GGML_OP_UNARY: + { + n_fuse = ggml_metal_op_unary(ctx, idx); + } break; + case GGML_OP_GLU: + { + n_fuse = ggml_metal_op_glu(ctx, idx); + } break; + case GGML_OP_SUM_ROWS: + case GGML_OP_MEAN: + { + n_fuse = ggml_metal_op_sum_rows(ctx, idx); + } break; + case GGML_OP_SOFT_MAX: + { + n_fuse = ggml_metal_op_soft_max(ctx, idx); + } break; + case GGML_OP_SSM_CONV: + { + n_fuse = ggml_metal_op_ssm_conv(ctx, idx); + } break; + case GGML_OP_SSM_SCAN: + { + n_fuse = ggml_metal_op_ssm_scan(ctx, idx); + } break; + case GGML_OP_RWKV_WKV6: + case GGML_OP_RWKV_WKV7: + { + n_fuse = ggml_metal_op_rwkv(ctx, idx); + } break; + case GGML_OP_MUL_MAT: + { + n_fuse = ggml_metal_op_mul_mat(ctx, idx); + } break; + case GGML_OP_MUL_MAT_ID: + { + n_fuse = ggml_metal_op_mul_mat_id(ctx, idx); + } break; + case GGML_OP_GET_ROWS: + { + n_fuse = ggml_metal_op_get_rows(ctx, idx); + } break; + case GGML_OP_SET_ROWS: + { + n_fuse = ggml_metal_op_set_rows(ctx, idx); + } break; + case GGML_OP_RMS_NORM: + { + n_fuse = ggml_metal_op_rms_norm(ctx, idx); + } break; + case GGML_OP_L2_NORM: + { + n_fuse = ggml_metal_op_l2_norm(ctx, idx); + } break; + case GGML_OP_GROUP_NORM: + { + n_fuse = ggml_metal_op_group_norm(ctx, idx); + } break; + case GGML_OP_NORM: + { + n_fuse = ggml_metal_op_norm(ctx, idx); + } break; + case GGML_OP_ROPE: + { + n_fuse = ggml_metal_op_rope(ctx, idx); + } break; + case GGML_OP_IM2COL: + { + n_fuse = ggml_metal_op_im2col(ctx, idx); + } break; + case GGML_OP_CONV_TRANSPOSE_1D: + { + n_fuse = ggml_metal_op_conv_transpose_1d(ctx, idx); + } break; + case GGML_OP_UPSCALE: + { + n_fuse = ggml_metal_op_upscale(ctx, idx); + } break; + case GGML_OP_PAD: + { + n_fuse = ggml_metal_op_pad(ctx, idx); + } break; + case GGML_OP_PAD_REFLECT_1D: + { + n_fuse = ggml_metal_op_pad_reflect_1d(ctx, idx); + } break; + case GGML_OP_ARANGE: + { + n_fuse = ggml_metal_op_arange(ctx, idx); + } break; + case GGML_OP_TIMESTEP_EMBEDDING: + { + n_fuse = ggml_metal_op_timestep_embedding(ctx, idx); + } break; + case GGML_OP_ARGSORT: + { + n_fuse = ggml_metal_op_argsort(ctx, idx); + } break; + case GGML_OP_LEAKY_RELU: + { + n_fuse = ggml_metal_op_leaky_relu(ctx, idx); + } break; + case GGML_OP_FLASH_ATTN_EXT: + { + n_fuse = ggml_metal_op_flash_attn_ext(ctx, idx); + } break; + case GGML_OP_DUP: + case GGML_OP_CPY: + case GGML_OP_CONT: + { + n_fuse = ggml_metal_op_cpy(ctx, idx); + } break; + case GGML_OP_POOL_2D: + { + n_fuse = ggml_metal_op_pool_2d(ctx, idx); + } break; + case GGML_OP_ARGMAX: + { + n_fuse = ggml_metal_op_argmax(ctx, idx); + } break; + default: + { + GGML_LOG_ERROR("%s: error: node %3d, op = %8s not implemented\n", __func__, idx, ggml_op_name(node->op)); + GGML_ABORT("fatal error"); + } + } + + if (ctx->debug_graph > 0) { + if (n_fuse > 1) { + GGML_LOG_DEBUG("%s: fuse %d ops\n", __func__, n_fuse); + } + } + + // update the mem ranges in the encoding context + for (int i = 0; i < n_fuse; ++i) { + if (!ggml_metal_op_concurrency_add(ctx, nodes[i])) { + ggml_metal_op_concurrency_reset(ctx); + } + } + + return n_fuse; +} + +int ggml_metal_op_encode(ggml_metal_op_t ctx, int idx) { + if (ctx->use_capture) { + ggml_metal_encoder_debug_group_push(ctx->enc, ggml_op_desc(ggml_graph_node(ctx->gf, idx))); + } + + int res = ggml_metal_op_encode_impl(ctx, idx); + if (idx + res > ctx->idx_end) { + GGML_ABORT("fusion error: nodes spanning multiple encoders have been fused. this indicates a bug in the fusion logic %s", + "https://github.com/ggml-org/llama.cpp/pull/14849"); + } + + if (ctx->use_capture) { + ggml_metal_encoder_debug_group_pop(ctx->enc); + } + + return res; +} + +int ggml_metal_op_concat(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint64_t, nb, op, nb); + + const int32_t dim = ((const int32_t *) op->op_params)[0]; + + ggml_metal_kargs_concat args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.ne13 =*/ ne13, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + /*.dim =*/ dim, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_base(lib, GGML_OP_CONCAT); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 3); + + const int nth = std::min(1024, ne0); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne1, ne2, ne3, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_repeat(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_repeat(lib, op->type); + + ggml_metal_kargs_repeat args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + }; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + const int nth = std::min(ggml_metal_pipeline_max_theads_per_threadgroup(pipeline), ne0); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne1, ne2, ne3, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_acc(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + GGML_ASSERT(op->src[0]->type == GGML_TYPE_F32); + GGML_ASSERT(op->src[1]->type == GGML_TYPE_F32); + GGML_ASSERT(op->type == GGML_TYPE_F32); + + GGML_ASSERT(ggml_is_contiguous(op->src[0])); + GGML_ASSERT(ggml_is_contiguous(op->src[1])); + + const size_t pnb1 = ((const int32_t *) op->op_params)[0]; + const size_t pnb2 = ((const int32_t *) op->op_params)[1]; + const size_t pnb3 = ((const int32_t *) op->op_params)[2]; + const size_t offs = ((const int32_t *) op->op_params)[3]; + + const bool inplace = (bool) ((const int32_t *) op->op_params)[4]; + + if (!inplace) { + // run a separete kernel to cpy src->dst + // not sure how to avoid this + // TODO: make a simpler cpy_bytes kernel + + //const id pipeline = ctx->pipelines[GGML_METAL_PIPELINE_TYPE_CPY_F32_F32].obj; + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_cpy(lib, op->src[0]->type, op->type); + + ggml_metal_kargs_cpy args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + }; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + const int nth = std::min(ggml_metal_pipeline_max_theads_per_threadgroup(pipeline), ne00); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne01, ne02, ne03, nth, 1, 1); + + ggml_metal_op_concurrency_reset(ctx); + } + + ggml_metal_kargs_bin args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ pnb1, + /*.nb02 =*/ pnb2, + /*.nb03 =*/ pnb3, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.ne13 =*/ ne13, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ pnb1, + /*.nb2 =*/ pnb2, + /*.nb3 =*/ pnb3, + /*.offs =*/ offs, + /*.o1 =*/ { 0 }, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_bin(lib, GGML_OP_ADD, 1, false); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 3); + + const int nth = std::min(ggml_metal_pipeline_max_theads_per_threadgroup(pipeline), ne00); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne11, ne12, ne13, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_scale(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + float scale; + float bias; + memcpy(&scale, ((const int32_t *) op->op_params) + 0, sizeof(float)); + memcpy(&bias, ((const int32_t *) op->op_params) + 1, sizeof(float)); + + ggml_metal_kargs_scale args = { + /*.scale =*/ scale, + /*.bias =*/ bias, + }; + + int64_t n = ggml_nelements(op); + + if (n % 4 == 0) { + n /= 4; + } + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_unary(lib, op); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_dispatch_threadgroups(enc, n, 1, 1, 1, 1, 1); + + return 1; +} + +int ggml_metal_op_clamp(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + float min; + float max; + memcpy(&min, ((const int32_t *) op->op_params) + 0, sizeof(float)); + memcpy(&max, ((const int32_t *) op->op_params) + 1, sizeof(float)); + + ggml_metal_kargs_clamp args = { + /*.min =*/ min, + /*.max =*/ max, + }; + + int64_t n = ggml_nelements(op); + + if (n % 4 == 0) { + n /= 4; + } + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_unary(lib, op); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_dispatch_threadgroups(enc, n, 1, 1, 1, 1, 1); + + return 1; +} + +int ggml_metal_op_unary(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + int64_t n = ggml_nelements(op); + + if (n % 4 == 0) { + n /= 4; + } + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_unary(lib, op); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 1); + + ggml_metal_encoder_dispatch_threadgroups(enc, n, 1, 1, 1, 1, 1); + + return 1; +} + +int ggml_metal_op_glu(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + if (op->src[1]) { + GGML_ASSERT(ggml_are_same_shape(op->src[0], op->src[1])); + } + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_glu(lib, op); + + const int32_t swp = ggml_get_op_params_i32(op, 1); + const float alpha = ggml_get_op_params_f32(op, 2); + const float limit = ggml_get_op_params_f32(op, 3); + + const int32_t i00 = swp ? ne0 : 0; + const int32_t i10 = swp ? 0 : ne0; + + ggml_metal_kargs_glu args = { + /*.ne00 =*/ ne00, + /*.nb01 =*/ nb01, + /*.ne10 =*/ op->src[1] ? ne10 : ne00, + /*.nb11 =*/ op->src[1] ? nb11 : nb01, + /*.ne0 =*/ ne0, + /*.nb1 =*/ nb1, + /*.i00 =*/ op->src[1] ? 0 : i00, + /*.i10 =*/ op->src[1] ? 0 : i10, + /*.alpha=*/ alpha, + /*.limit=*/ limit + }; + + const int64_t nrows = ggml_nrows(op->src[0]); + + const int32_t nth = std::min(ggml_metal_pipeline_max_theads_per_threadgroup(pipeline), ne00/2); + + //[encoder setComputePipelineState:pipeline]; + //[encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; + //if (src1) { + // [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; + //} else { + // [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + //} + //[encoder setBuffer:id_dst offset:offs_dst atIndex:2]; + //[encoder setBytes:&args length:sizeof(args) atIndex:3]; + + //[encoder dispatchThreadgroups:MTLSizeMake(nrows, 1, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + if (op->src[1]) { + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + } else { + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 2); + } + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 3); + + ggml_metal_encoder_dispatch_threadgroups(enc, nrows, 1, 1, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_sum_rows(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + ggml_metal_kargs_sum_rows args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_sum_rows(lib, op); + + int nth = 32; // SIMD width + + while (nth < ne00 && nth < ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)) { + nth *= 2; + } + + nth = std::min(nth, ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)); + nth = std::min(nth, ne00); + + const size_t smem = ggml_metal_pipeline_get_smem(pipeline); + + //[encoder setComputePipelineState:pipeline]; + //[encoder setBytes:&args length:sizeof(args) atIndex:0]; + //[encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; + //[encoder setBuffer:id_dst offset:offs_dst atIndex:2]; + //[encoder setThreadgroupMemoryLength:32*sizeof(float) atIndex:0]; + + //[encoder dispatchThreadgroups:MTLSizeMake(ne01, ne02, ne03) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne01, ne02, ne03, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_get_rows(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_get_rows(lib, op->src[0]->type); + + ggml_metal_kargs_get_rows args = { + /*.ne00 =*/ ne00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.ne10 =*/ ne10, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + }; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 3); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne10, ne11, ne12, 32, 1, 1); + + return 1; +} + +int ggml_metal_op_set_rows(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_set_rows(lib, op->type); + + const int32_t nk0 = ne0/ggml_blck_size(op->type); + + int nth = 32; // SIMD width + + while (nth < nk0 && nth < ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)) { + nth *= 2; + } + + int nrptg = 1; + if (nth > nk0) { + nrptg = (nth + nk0 - 1)/nk0; + nth = nk0; + + if (nrptg*nth > ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)) { + nrptg--; + } + } + + nth = std::min(nth, nk0); + + ggml_metal_kargs_set_rows args = { + /*.nk0 =*/ nk0, + /*.ne01 =*/ ne01, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + }; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 3); + + ggml_metal_encoder_dispatch_threadgroups(enc, (ne01 + nrptg - 1)/nrptg, ne02, ne03, nth, nrptg, 1); + + return 1; +} + +int ggml_metal_op_soft_max(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne2, op->src[2], ne); + GGML_TENSOR_LOCALS(uint64_t, nb2, op->src[2], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + float scale; + float max_bias; + + memcpy(&scale, ((const int32_t *) op->op_params) + 0, sizeof(scale)); + memcpy(&max_bias, ((const int32_t *) op->op_params) + 1, sizeof(max_bias)); + + const uint32_t n_head = op->src[0]->ne[2]; + const int32_t n_head_log2 = 1u << (uint32_t) floorf(log2f((float) n_head)); + + const float m0 = powf(2.0f, -(max_bias ) / n_head_log2); + const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2); + + // softmax + + ggml_metal_kargs_soft_max args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.ne13 =*/ ne13, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + /*.scale =*/ scale, + /*.max_bias =*/ max_bias, + /*.m0 =*/ m0, + /*.m1 =*/ m1, + /*.n_head_log2 =*/ n_head_log2, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_soft_max(lib, op); + + int nth = 32; // SIMD width + + if (ne00%4 == 0) { + while (nth < ne00/4 && nth*ne01*ne02*ne03 < 256) { + nth *= 2; + } + } else { + while (nth < ne00 && nth*ne01*ne02*ne03 < 256) { + nth *= 2; + } + } + + const size_t smem = ggml_metal_pipeline_get_smem(pipeline); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes(enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[0]), 1); + if (op->src[1]) { + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[1]), 2); + } else { + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[0]), 2); + } + if (op->src[2]) { + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[2]), 3); + } else { + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[0]), 3); + } + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op), 4); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne01, ne02, ne03, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_ssm_conv(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + ggml_metal_kargs_ssm_conv args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_ssm_conv(lib, op); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes(enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op), 3); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne01, ne1, ne02, 1, 1, 1); + + return 1; +} + +int ggml_metal_op_ssm_scan(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne2, op->src[2], ne); + GGML_TENSOR_LOCALS(uint64_t, nb2, op->src[2], nb); + GGML_TENSOR_LOCALS( int32_t, ne3, op->src[3], ne); + GGML_TENSOR_LOCALS(uint64_t, nb3, op->src[3], nb); + GGML_TENSOR_LOCALS( int32_t, ne4, op->src[4], ne); + GGML_TENSOR_LOCALS(uint64_t, nb4, op->src[4], nb); + GGML_TENSOR_LOCALS( int32_t, ne5, op->src[5], ne); + GGML_TENSOR_LOCALS(uint64_t, nb5, op->src[5], nb); + GGML_TENSOR_LOCALS( int32_t, ne6, op->src[6], ne); + GGML_TENSOR_LOCALS(uint64_t, nb6, op->src[6], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + const ggml_tensor * src3 = op->src[3]; + const ggml_tensor * src4 = op->src[4]; + const ggml_tensor * src5 = op->src[5]; + const ggml_tensor * src6 = op->src[6]; + + GGML_ASSERT(src3); + GGML_ASSERT(src4); + GGML_ASSERT(src5); + GGML_ASSERT(src6); + + const int64_t d_state = ne00; + const int64_t d_inner = ne01; + const int64_t n_head = ne02; + const int64_t n_group = ne41; + const int64_t n_seq_tokens = ne12; + const int64_t n_seqs = ne13; + + ggml_metal_kargs_ssm_scan args = { + /*.d_state =*/ d_state, + /*.d_inner =*/ d_inner, + /*.n_head =*/ n_head, + /*.n_group =*/ n_group, + /*.n_seq_tokens =*/ n_seq_tokens, + /*.n_seqs =*/ n_seqs, + /*.s_off =*/ ggml_nelements(op->src[1]) * sizeof(float), + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.nb21 =*/ nb21, + /*.nb22 =*/ nb22, + /*.nb31 =*/ nb31, + /*.nb41 =*/ nb41, + /*.nb42 =*/ nb42, + /*.nb43 =*/ nb43, + /*.nb51 =*/ nb51, + /*.nb52 =*/ nb52, + /*.nb53 =*/ nb53, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_ssm_scan(lib, op); + + const size_t sms = ggml_metal_pipeline_get_smem(pipeline); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[2]), 3); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[3]), 4); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[4]), 5); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[5]), 6); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[6]), 7); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 8); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, sms, 0); + + if (ne30 == 1) { + // Mamba-2 + ggml_metal_encoder_dispatch_threadgroups(enc, d_inner, n_head, n_seqs, d_state, 1, 1); + } else { + GGML_ASSERT(d_inner == 1); + ggml_metal_encoder_dispatch_threadgroups(enc, n_head, n_seqs, 1, d_state, 1, 1); + } + + return 1; +} + +int ggml_metal_op_rwkv(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + const int64_t B = op->op == GGML_OP_RWKV_WKV6 ? op->src[5]->ne[1] : op->src[6]->ne[1]; + const int64_t T = op->src[0]->ne[2]; + const int64_t C = op->ne[0]; + const int64_t H = op->src[0]->ne[1]; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_rwkv(lib, op); + + int ida = 0; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), ida++); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), ida++); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[2]), ida++); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[3]), ida++); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[4]), ida++); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[5]), ida++); + if (op->op == GGML_OP_RWKV_WKV7) { + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[6]), ida++); + } + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), ida++); + ggml_metal_encoder_set_bytes (enc, (void *) &B, sizeof(B), ida++); + ggml_metal_encoder_set_bytes (enc, (void *) &T, sizeof(T), ida++); + ggml_metal_encoder_set_bytes (enc, (void *) &C, sizeof(C), ida++); + ggml_metal_encoder_set_bytes (enc, (void *) &H, sizeof(H), ida++); + + ggml_metal_encoder_dispatch_threadgroups(enc, B * H, 1, 1, C/H, 1, 1); + + return 1; +} + +int ggml_metal_op_cpy(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_cpy(lib, op->src[0]->type, op->type); + + GGML_ASSERT(ne00 % ggml_blck_size(op->src[0]->type) == 0); + + // TODO: support + //const int32_t nk00 = ne00/ggml_blck_size(op->type); + const int32_t nk00 = ne00; + + int nth = 32; // SIMD width + + while (nth < nk00 && nth < ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)) { + nth *= 2; + } + + nth = std::min(nth, ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)); + + // when rows are small, we can batch them together in a single threadgroup + int nrptg = 1; + + // TODO: relax this constraint in the future + if (ggml_blck_size(op->src[0]->type) == 1 && ggml_blck_size(op->type) == 1) { + if (nth > nk00) { + nrptg = (nth + nk00 - 1)/nk00; + nth = nk00; + + if (nrptg*nth > ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)) { + nrptg--; + } + } + } + + nth = std::min(nth, nk00); + + ggml_metal_kargs_cpy args = { + /*.ne00 =*/ nk00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + }; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne01, ne02, ne03, nth, nrptg, 1); + + return 1; +} + +int ggml_metal_op_pool_2d(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + const int32_t * opts = op->op_params; + ggml_op_pool op_pool = (ggml_op_pool) opts[0]; + + const int32_t k0 = opts[1]; + const int32_t k1 = opts[2]; + const int32_t s0 = opts[3]; + const int32_t s1 = opts[4]; + const int32_t p0 = opts[5]; + const int32_t p1 = opts[6]; + + const int64_t IH = op->src[0]->ne[1]; + const int64_t IW = op->src[0]->ne[0]; + + const int64_t N = op->ne[3]; + const int64_t OC = op->ne[2]; + const int64_t OH = op->ne[1]; + const int64_t OW = op->ne[0]; + + const int64_t np = N * OC * OH * OW; + + ggml_metal_kargs_pool_2d args_pool_2d = { + /* .k0 = */ k0, + /* .k1 = */ k1, + /* .s0 = */ s0, + /* .s1 = */ s1, + /* .p0 = */ p0, + /* .p1 = */ p1, + /* .IH = */ IH, + /* .IW = */ IW, + /* .OH = */ OH, + /* .OW = */ OW, + /* .np = */ np + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_pool_2d(lib, op, op_pool); + + const int nth = std::min(ggml_metal_pipeline_max_theads_per_threadgroup(pipeline), (int) np); + const int ntg = (np + nth - 1) / nth; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args_pool_2d, sizeof(args_pool_2d), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_dispatch_threadgroups(enc, ntg, 1, 1, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_mul_mat(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + const ggml_metal_device_props * props_dev = ggml_metal_device_get_props(ctx->dev); + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + GGML_ASSERT(ne00 == ne10); + + GGML_ASSERT(ne12 % ne02 == 0); + GGML_ASSERT(ne13 % ne03 == 0); + + const int16_t r2 = ne12/ne02; + const int16_t r3 = ne13/ne03; + + // find the break-even point where the matrix-matrix kernel becomes more efficient compared + // to the matrix-vector kernel + const int ne11_mm_min = 8; + + // first try to use small-batch mat-mv kernels + // these should be efficient for BS [2, ~8] + if (op->src[1]->type == GGML_TYPE_F32 && (ne00%128 == 0) && + ( + ( + ( + op->src[0]->type == GGML_TYPE_F32 || // TODO: helper function + op->src[0]->type == GGML_TYPE_F16 || + op->src[0]->type == GGML_TYPE_Q4_0 || + op->src[0]->type == GGML_TYPE_Q4_1 || + op->src[0]->type == GGML_TYPE_Q5_0 || + op->src[0]->type == GGML_TYPE_Q5_1 || + op->src[0]->type == GGML_TYPE_Q8_0 || + op->src[0]->type == GGML_TYPE_MXFP4 || + op->src[0]->type == GGML_TYPE_IQ4_NL || + false) && (ne11 >= 2 && ne11 <= 8) + ) || + ( + ( + op->src[0]->type == GGML_TYPE_Q4_K || + op->src[0]->type == GGML_TYPE_Q5_K || + op->src[0]->type == GGML_TYPE_Q6_K || + false) && (ne11 >= 4 && ne11 <= 8) + ) + ) + ) { + // TODO: determine the optimal parameters based on grid utilization + // I still don't know why we should not always use the maximum available threads: + // + // nsg = pipeline.maxTotalThreadsPerThreadgroup / 32 + // + // my current hypothesis is that the work grid is not evenly divisible for different nsg + // values and there can be some tail effects when nsg is high. need to confirm this + // + const int nsg = 2; // num simdgroups per threadgroup + + // num threads along row per simdgroup + int16_t nxpsg = 0; + if (ne00 % 256 == 0 && ne11 < 3) { + nxpsg = 16; + } else if (ne00 % 128 == 0) { + nxpsg = 8; + } else { + nxpsg = 4; + } + + const int16_t nypsg = 32/nxpsg; // num threads along col per simdgroup (i.e. a simdgroup processes that many src0 rows at a time) + const int16_t r0ptg = nypsg*nsg; // num src0 rows per threadgroup + int16_t r1ptg = 4; // num src1 rows per threadgroup + + // note: not sure how optimal are those across all different hardware. there might be someting cleverer + switch (ne11) { + case 2: + r1ptg = 2; break; + case 3: + case 6: + r1ptg = 3; break; + case 4: + case 7: + case 8: + r1ptg = 4; break; + case 5: + r1ptg = 5; break; + default: + GGML_ABORT("unsupported ne11"); + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_mul_mv_ext(lib, op->src[0]->type, op->src[1]->type, r1ptg); + + ggml_metal_kargs_mul_mv_ext args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.r2 =*/ r2, + /*.r3 =*/ r3, + /*.nsg =*/ nsg, + /*.nxpsg =*/ nxpsg, + /*.r1ptg =*/ r1ptg, + }; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 3); + + ggml_metal_encoder_dispatch_threadgroups(enc, ((ne01 + r0ptg - 1)/r0ptg), ((ne11 + r1ptg - 1)/r1ptg), ne12*ne13, 32, nsg, 1); + } else if ( + !ggml_is_transposed(op->src[0]) && + !ggml_is_transposed(op->src[1]) && + // for now the matrix-matrix multiplication kernel only works on A14+/M1+ SoCs + // AMD GPU and older A-chips will reuse matrix-vector multiplication kernel + props_dev->has_simdgroup_mm && + op->src[1]->type == GGML_TYPE_F32 && + ne00 % 32 == 0 && ne00 >= 64 && + (ne11 > ne11_mm_min || (ggml_is_quantized(op->src[0]->type) && ne12 > 1))) { + //printf("matrix: ne00 = %6d, ne01 = %6d, ne02 = %6d, ne11 = %6d, ne12 = %6d\n", ne00, ne01, ne02, ne11, ne12); + + // some Metal matrix data types require aligned pointers + // ref: https://developer.apple.com/metal/Metal-Shading-Language-Specification.pdf (Table 2.5) + switch (op->src[0]->type) { + case GGML_TYPE_F32: GGML_ASSERT(nb01 % 16 == 0); break; + case GGML_TYPE_F16: GGML_ASSERT(nb01 % 8 == 0); break; + case GGML_TYPE_BF16: GGML_ASSERT(nb01 % 8 == 0); break; + default: break; + } + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_mul_mm(lib, op->src[0]->type, op->src[1]->type); + + ggml_metal_kargs_mul_mm args = { + /*.ne00 =*/ ne00, + /*.ne02 =*/ ne02, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne12 =*/ ne12, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.r2 =*/ r2, + /*.r3 =*/ r3, + }; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 3); + + const size_t smem = ggml_metal_pipeline_get_smem(pipeline); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + ggml_metal_encoder_dispatch_threadgroups(enc, ((ne11 + 31)/32), ((ne01 + 63)/64), ne12*ne13, 128, 1, 1); + } else { + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_mul_mv(lib, op); + + ggml_metal_kargs_mul_mv args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.r2 =*/ r2, + /*.r3 =*/ r3, + }; + + const int nr0 = ggml_metal_pipeline_get_nr0(pipeline); + const int nr1 = ggml_metal_pipeline_get_nr1(pipeline); + const int nsg = ggml_metal_pipeline_get_nsg(pipeline); + + const size_t smem = ggml_metal_pipeline_get_smem(pipeline); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 3); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + if (op->src[0]->type == GGML_TYPE_F32 || + op->src[0]->type == GGML_TYPE_F16 || + op->src[0]->type == GGML_TYPE_BF16 || + op->src[0]->type == GGML_TYPE_Q8_0) { + ggml_metal_encoder_dispatch_threadgroups(enc, ((ne01 + nr0 - 1)/(nr0)), ((ne11 + nr1 - 1)/nr1), ne12*ne13, 32, nsg, 1); + } else { + ggml_metal_encoder_dispatch_threadgroups(enc, ((ne01 + nr0*nsg - 1)/(nr0*nsg)), ((ne11 + nr1 - 1)/nr1), ne12*ne13, 32, nsg, 1); + } + } + + return 1; +} + +size_t ggml_metal_op_mul_mat_id_extra_tpe(const ggml_tensor * op) { + assert(op->op == GGML_OP_MUL_MAT_ID); + + const int64_t ne02 = op->src[0]->ne[2]; // n_expert + + return ggml_type_size(GGML_TYPE_I32)*ne02; +} + +size_t ggml_metal_op_mul_mat_id_extra_ids(const ggml_tensor * op) { + assert(op->op == GGML_OP_MUL_MAT_ID); + + const int64_t ne02 = op->src[0]->ne[2]; // n_expert + const int64_t ne21 = op->src[2]->ne[1]; // n_token + + return ggml_type_size(GGML_TYPE_I32)*ne02*ne21; +} + +int ggml_metal_op_mul_mat_id(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + const ggml_metal_device_props * props_dev = ggml_metal_device_get_props(ctx->dev); + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne2, op->src[2], ne); + GGML_TENSOR_LOCALS(uint64_t, nb2, op->src[2], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + // src2 = ids + GGML_ASSERT(op->src[2]->type == GGML_TYPE_I32); + + GGML_ASSERT(!ggml_is_transposed(op->src[0])); + GGML_ASSERT(!ggml_is_transposed(op->src[1])); + + GGML_ASSERT(op->src[1]->type == GGML_TYPE_F32); + + GGML_ASSERT(ne03 == 1); + GGML_ASSERT(ne13 == 1); + + ggml_metal_buffer_id bid_src0 = ggml_metal_get_buffer_id(op->src[0]); + ggml_metal_buffer_id bid_src1 = ggml_metal_get_buffer_id(op->src[1]); + ggml_metal_buffer_id bid_src2 = ggml_metal_get_buffer_id(op->src[2]); + ggml_metal_buffer_id bid_dst = ggml_metal_get_buffer_id(op); + + const uint32_t r2 = 1; + const uint32_t r3 = 1; + + // find the break-even point where the matrix-matrix kernel becomes more efficient compared + // to the matrix-vector kernel + // ne20 = n_used_experts + // ne21 = n_rows (batch size) + const int ne21_mm_id_min = 32; + + if (props_dev->has_simdgroup_mm && + ne00 % 32 == 0 && ne00 >= 64 && + (ne21 >= ne21_mm_id_min)) { + GGML_ASSERT(ne00 % 4 == 0); + + // some Metal matrix data types require aligned pointers + // ref: https://developer.apple.com/metal/Metal-Shading-Language-Specification.pdf (Table 2.5) + switch (op->src[0]->type) { + case GGML_TYPE_F32: GGML_ASSERT(nb01 % 16 == 0); break; + case GGML_TYPE_F16: GGML_ASSERT(nb01 % 8 == 0); break; + case GGML_TYPE_BF16: GGML_ASSERT(nb01 % 8 == 0); break; + default: break; + } + + // extra buffers for intermediate id mapping + ggml_metal_buffer_id bid_tpe = bid_dst; + bid_tpe.offs += ggml_nbytes(op); + + ggml_metal_buffer_id bid_ids = bid_tpe; + bid_ids.offs += ggml_metal_op_mul_mat_id_extra_tpe(op); + + { + ggml_metal_kargs_mul_mm_id_map0 args = { + ne02, + ne10, + ne11, // n_expert_used (bcast) + nb11, + nb12, + ne21, // n_tokens + ne20, // n_expert_used + nb21, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_mul_mm_id_map0(lib, ne02, ne20); + + const size_t smem = ggml_metal_pipeline_get_smem(pipeline); + + GGML_ASSERT(ne02 <= ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)); + + GGML_ASSERT(smem <= props_dev->max_theadgroup_memory_size); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, bid_src2, 1); + ggml_metal_encoder_set_buffer (enc, bid_tpe, 2); + ggml_metal_encoder_set_buffer (enc, bid_ids, 3); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + ggml_metal_encoder_dispatch_threadgroups(enc, 1, 1, 1, ne02, 1, 1); + } + + // this barrier is always needed because the next kernel has to wait for the id maps to be computed + ggml_metal_op_concurrency_reset(ctx); + + { + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_mul_mm_id(lib, op->src[0]->type, GGML_TYPE_F16); + + ggml_metal_kargs_mul_mm_id args = { + /*.ne00 =*/ ne00, + /*.ne02 =*/ ne02, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne11 =*/ ne11, // n_expert_used (bcast) + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne20 =*/ ne20, // n_expert_used + /*.ne21 =*/ ne21, // n_tokens + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.r2 =*/ r2, + /*.r3 =*/ r3, + }; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, bid_src0, 1); + ggml_metal_encoder_set_buffer (enc, bid_src1, 2); + ggml_metal_encoder_set_buffer (enc, bid_tpe, 3); + ggml_metal_encoder_set_buffer (enc, bid_ids, 4); + ggml_metal_encoder_set_buffer (enc, bid_dst, 5); + + const size_t smem = ggml_metal_pipeline_get_smem(pipeline); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + ggml_metal_encoder_dispatch_threadgroups(enc, (ne21 + 31)/32, (ne01 + 63)/64, ne02, 128, 1, 1); + } + } else { + ggml_metal_kargs_mul_mv_id args = { + /*.nei0 =*/ ne20, + /*.nei1 =*/ ne21, + /*.nbi1 =*/ nb21, + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.ne13 =*/ ne13, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.nb1 =*/ nb1, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_mul_mv_id(lib, op); + + const int nr0 = ggml_metal_pipeline_get_nr0(pipeline); + const int nr1 = ggml_metal_pipeline_get_nr1(pipeline); + const int nsg = ggml_metal_pipeline_get_nsg(pipeline); + + const size_t smem = ggml_metal_pipeline_get_smem(pipeline); + + if (ggml_is_quantized(op->src[0]->type)) { + GGML_ASSERT(ne00 >= nsg*nr0); + } + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes(enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer(enc, bid_src0, 1); + ggml_metal_encoder_set_buffer(enc, bid_src1, 2); + ggml_metal_encoder_set_buffer(enc, bid_dst, 3); + ggml_metal_encoder_set_buffer(enc, bid_src2, 4); + + const int64_t _ne1 = 1; + const int64_t ne123 = ne20*ne21; + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + if (op->src[0]->type == GGML_TYPE_F32 || + op->src[0]->type == GGML_TYPE_F16 || + op->src[0]->type == GGML_TYPE_BF16 || + op->src[0]->type == GGML_TYPE_Q8_0) { + ggml_metal_encoder_dispatch_threadgroups(enc, (ne01 + nr0 - 1)/(nr0), (_ne1 + nr1 - 1)/nr1, ne123, 32, nsg, 1); + } else { + ggml_metal_encoder_dispatch_threadgroups(enc, (ne01 + nr0*nsg - 1)/(nr0*nsg), (_ne1 + nr1 - 1)/nr1, ne123, 32, nsg, 1); + } + } + + return 1; +} + +int ggml_metal_op_add_id(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne2, op->src[2], ne); + GGML_TENSOR_LOCALS(uint64_t, nb2, op->src[2], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + + GGML_ASSERT(op->src[0]->type == GGML_TYPE_F32); + GGML_ASSERT(op->src[1]->type == GGML_TYPE_F32); + GGML_ASSERT(op->src[2]->type == GGML_TYPE_I32); + GGML_ASSERT(op->type == GGML_TYPE_F32); + + GGML_ASSERT(ggml_is_contiguous_rows(op->src[0])); + + ggml_metal_kargs_add_id args = { + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb11 =*/ nb11, + /*.nb21 =*/ nb21, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_base(lib, GGML_OP_ADD_ID); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[2]), 3); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 4); + + const int nth = std::min(ggml_metal_pipeline_max_theads_per_threadgroup(pipeline), ne00); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne01, ne02, 1, nth, 1, 1); + + return 1; +} + +bool ggml_metal_op_flash_attn_ext_use_vec(const ggml_tensor * op) { + assert(op->op == GGML_OP_FLASH_ATTN_EXT); + + const int64_t ne00 = op->src[0]->ne[0]; // head size + const int64_t ne01 = op->src[0]->ne[1]; // batch size + + // use vec kernel if the batch size is small and if the head size is supported + return (ne01 < 20) && (ne00 % 32 == 0); +} + +size_t ggml_metal_op_flash_attn_ext_extra_tmp(const ggml_tensor * op) { + assert(op->op == GGML_OP_FLASH_ATTN_EXT); + + const int64_t nwg = 32; + + const int64_t ne01 = op->src[0]->ne[1]; + const int64_t ne02 = op->src[0]->ne[2]; + const int64_t ne03 = op->src[0]->ne[3]; + const int64_t ne20 = op->src[2]->ne[0]; + + // temp buffer for writing the results from each workgroup + // - ne20: the size of the Value head + // - + 2: the S and M values for each intermediate result + return ggml_type_size(GGML_TYPE_F32)*(ne01*ne02*ne03*nwg*(ne20 + 2)); +} + +int ggml_metal_op_flash_attn_ext(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + const ggml_metal_device_props * props_dev = ggml_metal_device_get_props(ctx->dev); + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne2, op->src[2], ne); + GGML_TENSOR_LOCALS(uint64_t, nb2, op->src[2], nb); + GGML_TENSOR_LOCALS( int32_t, ne3, op->src[3], ne); + GGML_TENSOR_LOCALS(uint64_t, nb3, op->src[3], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS( int32_t, nb, op, nb); + + GGML_ASSERT(ne00 % 4 == 0); + GGML_ASSERT(ne11 % 32 == 0); + + GGML_ASSERT(op->src[0]->type == GGML_TYPE_F32); + GGML_ASSERT(op->src[1]->type == op->src[2]->type); + + //GGML_ASSERT(ggml_are_same_shape (src1, src2)); + GGML_ASSERT(ne11 == ne21); + GGML_ASSERT(ne12 == ne22); + + GGML_ASSERT(!op->src[3] || op->src[3]->type == GGML_TYPE_F16); + GGML_ASSERT(!op->src[3] || op->src[3]->ne[1] >= GGML_PAD(op->src[0]->ne[1], 8) && + "the Flash-Attention Metal kernel requires the mask to be padded to 8 and at least n_queries big"); + + float scale; + float max_bias; + float logit_softcap; + + memcpy(&scale, ((const int32_t *) op->op_params) + 0, sizeof(scale)); + memcpy(&max_bias, ((const int32_t *) op->op_params) + 1, sizeof(max_bias)); + memcpy(&logit_softcap, ((const int32_t *) op->op_params) + 2, sizeof(logit_softcap)); + + if (logit_softcap != 0.0f) { + scale /= logit_softcap; + } + + const bool has_mask = op->src[3] != NULL; + const bool has_sinks = op->src[4] != NULL; + const bool has_bias = max_bias != 0.0f; + const bool has_scap = logit_softcap != 0.0f; + + const uint32_t n_head = op->src[0]->ne[2]; + const int32_t n_head_log2 = 1u << (uint32_t) floorf(log2f((float) n_head)); + + const float m0 = powf(2.0f, -(max_bias ) / n_head_log2); + const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2); + + GGML_ASSERT(ne01 < 65536); + + if (!ggml_metal_op_flash_attn_ext_use_vec(op)) { + // half8x8 kernel + const int64_t nqptg = 8; // queries per threadgroup !! sync with kernel template arguments !! + const int64_t ncpsg = 64; // cache values per simdgroup !! sync with kernel template arguments !! + + GGML_ASSERT(nqptg <= 32); + GGML_ASSERT(nqptg % 8 == 0); + GGML_ASSERT(ncpsg % 32 == 0); + + const int is_q = ggml_is_quantized(op->src[1]->type) ? 1 : 0; + + // 2*(2*ncpsg) + // ncpsg soft_max values + ncpsg mask values + // + // 16*32*(nsg) + // the shared memory needed for the simdgroups to load the KV cache + // each thread loads (dequantizes) 16 head elements, there are 32 threads in th SG + // +#define FATTN_SMEM(nsg) (GGML_PAD((nqptg*(ne00 + 2*GGML_PAD(ne20, 64) + 2*(2*ncpsg)) + is_q*(16*32*(nsg)))*(sizeof(float)/2), 16)) + + //int64_t nsgmax = 4; + // + //if (is_q) { + // nsgmax = 2; + // while (true) { + // const size_t smem = FATTN_SMEM(nsgmax); + // if (smem > props_dev->max_theadgroup_memory_size) { + // break; + // } + // nsgmax *= 2; + // } + // nsgmax /= 2; + //} + + // simdgroups per threadgroup (a.k.a. warps) + //nsg = ne01 <= nqptg ? MAX(4, MIN(nsgmax, MIN(ne11/ncpsg, (int64_t) pipeline.maxTotalThreadsPerThreadgroup/32))) : 4; + int32_t nsg = 4; + + const size_t smem = FATTN_SMEM(nsg); + + ggml_metal_kargs_flash_attn_ext args = { + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne11 =*/ ne11, + /*.ne_12_2 =*/ ne12, + /*.ne_12_3 =*/ ne13, + /*.ns10 =*/ int32_t(nb11/nb10), + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ns20 =*/ int32_t(nb21/nb20), + /*.nb21 =*/ nb21, + /*.nb22 =*/ nb22, + /*.nb23 =*/ nb23, + /*.ne32 =*/ ne32, + /*.ne33 =*/ ne33, + /*.nb31 =*/ nb31, + /*.nb32 =*/ nb32, + /*.nb33 =*/ nb33, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.scale =*/ scale, + /*.max_bias =*/ max_bias, + /*.m0 =*/ m0, + /*.m1 =*/ m1, + /*.n_head_log2 =*/ n_head_log2, + /*.logit_softcap =*/ logit_softcap, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_flash_attn_ext(lib, op, has_mask, has_sinks, has_bias, has_scap, nsg); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[2]), 3); + if (op->src[3]) { + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[3]), 4); + } else { + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[0]), 4); + } + if (op->src[4]) { + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[4]), 5); + } else { + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[0]), 5); + } + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 6); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + ggml_metal_encoder_dispatch_threadgroups(enc, (ne01 + nqptg - 1)/nqptg, ne02, ne03, 32, nsg, 1); +#undef FATTN_SMEM + } else { + // half4x4 kernel + const int64_t nqptg = 1; // queries per threadgroup !! sync with kernel template arguments !! + const int64_t ncpsg = 32; // cache values per simdgroup !! sync with kernel template arguments !! + const int64_t nkpsg = 1*ncpsg; + + GGML_ASSERT(nqptg <= 32); + GGML_ASSERT(nqptg % 1 == 0); + GGML_ASSERT(ncpsg % 32 == 0); + + // ne00 + 2*ncpsg*(nsg) + // for each query, we load it as f16 in shared memory (ne00) + // and store the soft_max values and the mask + // + // ne20*(nsg) + // each simdgroup has a full f32 head vector in shared mem to accumulate results + // +#define FATTN_SMEM(nsg) (GGML_PAD((nqptg*(GGML_PAD(ne00, 128) + 4*ncpsg*(nsg)) + 2*GGML_PAD(ne20, 128)*(nsg))*(sizeof(float)/2), 16)) + + int64_t nsgmax = 2; + while (true) { + const size_t smem = FATTN_SMEM(nsgmax); + // avoid using more than half of the threadgroup memory - can cause slow downs especially for large head sizes + if (smem > props_dev->max_theadgroup_memory_size/2) { + break; + } + nsgmax *= 2; + } + nsgmax /= 2; + + // simdgroups per threadgroup (a.k.a. warps) + //const int64_t nsgt = MAX(2, MIN(nsgmax, MIN((ne11 + nkpsg - 1)/(nkpsg), (int64_t) pipeline.maxTotalThreadsPerThreadgroup/32))); + const int64_t nsgt = MAX(2, MIN(nsgmax, MIN((ne11 + nkpsg - 1)/(nkpsg), (int64_t) 1024/32))); + + int64_t nsg = 1; + while (nsg <= nsgt) { + nsg *= 2; + } + nsg /= 2; + + // workgroups + // each workgroup handles nsg*nkpsg cache values + int32_t nwg = 1; + if (false) { + // for small KV caches, we could launch a single workgroup and write the results directly to dst/ + // however, this does not lead to significant improvement, so disabled + nwg = 1; + nsg = 4; + } else { + nwg = 32; + nsg = 1; + while (2*nwg*nsg*nkpsg < ne11 && nsg < 4) { + nsg *= 2; + } + } + + ggml_metal_kargs_flash_attn_ext_vec args = { + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne11 =*/ ne11, + /*.ne_12_2 =*/ ne12, + /*.ne_12_3 =*/ ne13, + /*.ns10 =*/ int32_t(nb11/nb10), + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ns20 =*/ int32_t(nb21/nb20), + /*.nb21 =*/ nb21, + /*.nb22 =*/ nb22, + /*.nb23 =*/ nb23, + /*.ne32 =*/ ne32, + /*.ne33 =*/ ne33, + /*.nb31 =*/ nb31, + /*.nb32 =*/ nb32, + /*.nb33 =*/ nb33, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.scale =*/ scale, + /*.max_bias =*/ max_bias, + /*.m0 =*/ m0, + /*.m1 =*/ m1, + /*.n_head_log2 =*/ n_head_log2, + /*.logit_softcap =*/ logit_softcap, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_flash_attn_ext_vec(lib, op, has_mask, has_sinks, has_bias, has_scap, nsg, nwg); + + GGML_ASSERT(nsg*32 <= ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[2]), 3); + if (op->src[3]) { + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[3]), 4); + } else { + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[0]), 4); + } + if (op->src[4]) { + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[4]), 5); + } else { + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op->src[0]), 5); + } + + const size_t smem = FATTN_SMEM(nsg); + + //printf("smem: %zu, max: %zu, nsg = %d, nsgmax = %d\n", smem, props_dev->max_theadgroup_memory_size, (int) nsg, (int) nsgmax); + GGML_ASSERT(smem <= props_dev->max_theadgroup_memory_size); + + if (nwg == 1) { + // using 1 workgroup -> write the result directly into dst + ggml_metal_encoder_set_buffer(enc, ggml_metal_get_buffer_id(op), 6); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + ggml_metal_encoder_dispatch_threadgroups(enc, (ne01 + nqptg - 1)/nqptg, ne02, ne03*nwg, 32, nsg, 1); + } else { + // sanity checks + GGML_ASSERT(ne01*ne02*ne03 == ne1*ne2*ne3); + GGML_ASSERT((uint64_t)ne1*ne2*ne3 <= (1u << 31)); + + ggml_metal_buffer_id bid_dst = ggml_metal_get_buffer_id(op); + + // write the results from each workgroup into a temp buffer + ggml_metal_buffer_id bid_tmp = bid_dst; + bid_tmp.offs += ggml_nbytes(op); + ggml_metal_encoder_set_buffer(enc, bid_tmp, 6); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + ggml_metal_encoder_dispatch_threadgroups(enc, (ne01 + nqptg - 1)/nqptg, ne02, ne03*nwg, 32, nsg, 1); + + // sync the 2 kernels + ggml_metal_op_concurrency_reset(ctx); + + // reduce the results from the workgroups + { + const int32_t nrows = ne1*ne2*ne3; + + ggml_metal_kargs_flash_attn_ext_vec_reduce args0 = { + nrows, + }; + + ggml_metal_pipeline_t pipeline0 = ggml_metal_library_get_pipeline_flash_attn_ext_vec_reduce(lib, op, ne20, nwg); + + ggml_metal_encoder_set_pipeline(enc, pipeline0); + ggml_metal_encoder_set_bytes (enc, &args0, sizeof(args0), 0); + ggml_metal_encoder_set_buffer (enc, bid_tmp, 1); + ggml_metal_encoder_set_buffer (enc, bid_dst, 2); + + ggml_metal_encoder_dispatch_threadgroups(enc, nrows, 1, 1, 32*nwg, 1, 1); + } + } +#undef FATTN_SMEM + } + + return 1; +} + +int ggml_metal_op_bin(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_tensor ** ops = ggml_graph_nodes(gf) + idx; + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + const int idx_end = ctx->idx_end; + + const bool use_fusion = ctx->use_fusion; + + const int debug_fusion = ctx->debug_fusion; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint64_t, nb, op, nb); + + GGML_ASSERT(op->src[0]->type == GGML_TYPE_F32); + GGML_ASSERT(op->src[1]->type == GGML_TYPE_F32); + + GGML_ASSERT(ggml_is_contiguous_rows(op->src[0])); + GGML_ASSERT(ggml_is_contiguous_rows(op->src[1])); + + bool bcast_row = false; + + ggml_metal_buffer_id bid_src0 = ggml_metal_get_buffer_id(op->src[0]); + ggml_metal_buffer_id bid_src1 = ggml_metal_get_buffer_id(op->src[1]); + ggml_metal_buffer_id bid_dst = ggml_metal_get_buffer_id(op); + + ggml_metal_kargs_bin args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne10 =*/ ne10, + /*.ne11 =*/ ne11, + /*.ne12 =*/ ne12, + /*.ne13 =*/ ne13, + /*.nb10 =*/ nb10, + /*.nb11 =*/ nb11, + /*.nb12 =*/ nb12, + /*.nb13 =*/ nb13, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + /*.offs =*/ 0, + /*.o1 =*/ { bid_src1.offs }, + }; + + ggml_op fops[8]; + + int n_fuse = 1; + + // c[0] = add(a, b[0]) + // c[1] = add(c[0], b[1]) + // c[2] = add(c[1], b[2]) + // ... + if (use_fusion) { + fops[0] = GGML_OP_ADD; + fops[1] = GGML_OP_ADD; + fops[2] = GGML_OP_ADD; + fops[3] = GGML_OP_ADD; + fops[4] = GGML_OP_ADD; + fops[5] = GGML_OP_ADD; + fops[6] = GGML_OP_ADD; + fops[7] = GGML_OP_ADD; + + // note: in metal, we sometimes encode the graph in parallel so we have to avoid fusing ops + // across splits. idx_end indicates the last node in the current split + for (n_fuse = 0; n_fuse <= 6 && idx + n_fuse + 1 < idx_end; ++n_fuse) { + if (!ggml_can_fuse(gf, idx + n_fuse, fops + n_fuse, 2)) { + break; + } + + if (ops[n_fuse] != ops[n_fuse + 1]->src[0]) { + break; + } + + // b[0] === b[1] === ... + if (!ggml_are_same_layout(ops[n_fuse]->src[1], ops[n_fuse + 1]->src[1])) { + break; + } + + // only fuse ops if src1 is in the same Metal buffer + ggml_metal_buffer_id bid_fuse = ggml_metal_get_buffer_id(ops[n_fuse + 1]->src[1]); + if (bid_fuse.metal != bid_src1.metal) { + break; + } + + //ctx->fuse_cnt[ops[n_fuse + 1]->op]++; + + args.o1[n_fuse + 1] = bid_fuse.offs; + } + + ++n_fuse; + + if (debug_fusion > 1 && n_fuse > 1) { + GGML_LOG_DEBUG("%s: fuse: ADD x %d\n", __func__, n_fuse); + } + } + + // the offsets of src1 and all fused buffers are relative to the start of the src1 buffer + bid_src1.offs = 0; + + ggml_metal_pipeline_t pipeline = nullptr; + + if (ggml_nelements(op->src[1]) == ne10 && ggml_is_contiguous(op->src[1]) && ne00 % 4 == 0 && ne10 % 4 == 0) { + GGML_ASSERT(ggml_is_contiguous(op->src[0])); + + // src1 is a row + GGML_ASSERT(ne11 == 1); + + pipeline = ggml_metal_library_get_pipeline_bin(lib, op->op, n_fuse, true); + + bcast_row = true; + } else { + pipeline = ggml_metal_library_get_pipeline_bin(lib, op->op, n_fuse, false); + } + + if (n_fuse > 1) { + bid_dst = ggml_metal_get_buffer_id(ops[n_fuse - 1]); + + for (int i = 1; i < n_fuse; ++i) { + if (!ggml_metal_op_concurrency_check(ctx, ops[i])) { + ggml_metal_op_concurrency_reset(ctx); + + break; + } + } + } + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, bid_src0, 1); + ggml_metal_encoder_set_buffer (enc, bid_src1, 2); + ggml_metal_encoder_set_buffer (enc, bid_dst, 3); + + if (bcast_row) { + const int64_t n = ggml_nelements(op)/4; + + ggml_metal_encoder_dispatch_threadgroups(enc, n, 1, 1, 1, 1, 1); + } else { + int nth = 32; + + while (16*nth < ne0 && nth < ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)) { + nth *= 2; + } + + ggml_metal_encoder_dispatch_threadgroups(enc, ne01, ne02, ne03, nth, 1, 1); + } + + return n_fuse; +} + +int ggml_metal_op_rms_norm(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + const int idx_end = ctx->idx_end; + + const bool use_fusion = ctx->use_fusion; + + const int debug_fusion = ctx->debug_fusion; + + ggml_tensor ** ops = ggml_graph_nodes(gf) + idx; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + float eps; + memcpy(&eps, op->op_params, sizeof(float)); + + ggml_metal_buffer_id bid_src0 = ggml_metal_get_buffer_id(op->src[0]); + ggml_metal_buffer_id bid_dst = ggml_metal_get_buffer_id(op); + + ggml_metal_kargs_rms_norm args = { + /*.ne00 =*/ ne00, + /*.ne00_4 =*/ ne00/4, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + /*.eps =*/ eps, + /*.nef1 =*/ { ne01 }, + /*.nef2 =*/ { ne02 }, + /*.nef3 =*/ { ne03 }, + /*.nbf1 =*/ { nb01 }, + /*.nbf2 =*/ { nb02 }, + /*.nbf3 =*/ { nb03 }, + }; + + ggml_op fops[8]; + + int n_fuse = 1; + + ggml_metal_buffer_id bid_fuse[2] = { bid_src0, bid_src0 }; + + // d[0] = rms_norm(a) + // d[1] = mul(d[0], b) + // d[2] = add(d[1], c) + if (use_fusion) { + fops[0] = GGML_OP_RMS_NORM; + fops[1] = GGML_OP_MUL; + fops[2] = GGML_OP_ADD; + + for (n_fuse = 0; n_fuse <= 1 && idx + n_fuse + 1 < idx_end; ++n_fuse) { + if (!ggml_can_fuse(gf, idx + n_fuse, fops + n_fuse, 2)) { + break; + } + + if (ops[n_fuse] != ops[n_fuse + 1]->src[0]) { + break; + } + + if (ops[n_fuse + 1]->src[1]->ne[0] != op->ne[0]) { + break; + } + + if (!ggml_is_contiguous_rows(ops[n_fuse + 1]->src[1])) { + break; + } + + if (ops[n_fuse + 1]->type != GGML_TYPE_F32) { + break; + } + + //ctx->fuse_cnt[ops[n_fuse + 1]->op]++; + + bid_fuse[n_fuse] = ggml_metal_get_buffer_id(ops[n_fuse + 1]->src[1]); + + args.nef1[n_fuse + 1] = ops[n_fuse + 1]->src[1]->ne[1]; + args.nef2[n_fuse + 1] = ops[n_fuse + 1]->src[1]->ne[2]; + args.nef3[n_fuse + 1] = ops[n_fuse + 1]->src[1]->ne[3]; + + args.nbf1[n_fuse + 1] = ops[n_fuse + 1]->src[1]->nb[1]; + args.nbf2[n_fuse + 1] = ops[n_fuse + 1]->src[1]->nb[2]; + args.nbf3[n_fuse + 1] = ops[n_fuse + 1]->src[1]->nb[3]; + } + + ++n_fuse; + + if (debug_fusion > 1 && n_fuse > 1) { + if (n_fuse == 2) { + GGML_LOG_DEBUG("%s: fuse: RMS_NORM + MUL\n", __func__); + } + if (n_fuse == 3) { + GGML_LOG_DEBUG("%s: fuse: RMS_NORM + MUL + ADD\n", __func__); + } + } + } + + if (n_fuse > 1) { + bid_dst = ggml_metal_get_buffer_id(ops[n_fuse - 1]); + + for (int i = 1; i < n_fuse; ++i) { + if (!ggml_metal_op_concurrency_check(ctx, ops[i])) { + ggml_metal_op_concurrency_reset(ctx); + + break; + } + } + } + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_rms_norm(lib, op, n_fuse); + + int nth = 32; // SIMD width + + while (nth < ne00/4 && nth < ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)) { + nth *= 2; + } + + nth = std::min(nth, ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)); + nth = std::min(nth, ne00/4); + + const size_t smem = ggml_metal_pipeline_get_smem(pipeline); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, bid_src0, 1); + ggml_metal_encoder_set_buffer (enc, bid_fuse[0], 2); + ggml_metal_encoder_set_buffer (enc, bid_fuse[1], 3); + ggml_metal_encoder_set_buffer (enc, bid_dst, 4); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne01, ne02, ne03, nth, 1, 1); + + return n_fuse; +} + +int ggml_metal_op_l2_norm(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + float eps; + memcpy(&eps, op->op_params, sizeof(float)); + + int nth = 32; // SIMD width + + ggml_metal_kargs_l2_norm args = { + /*.ne00 =*/ ne00, + /*.ne00_4 =*/ ne00/4, + /*.nb01 =*/ nb01, + /*.eps =*/ eps, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_l2_norm(lib, op); + + while (nth < ne00/4 && nth < ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)) { + nth *= 2; + } + + nth = std::min(nth, ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)); + nth = std::min(nth, ne00/4); + + const size_t smem = ggml_metal_pipeline_get_smem(pipeline); + + const int64_t nrows = ggml_nrows(op->src[0]); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + ggml_metal_encoder_dispatch_threadgroups(enc, nrows, 1, 1, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_group_norm(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + const int32_t ngrp = ((const int32_t *) op->op_params)[0]; + + float eps; + memcpy(&eps, op->op_params + 1, sizeof(float)); + + ggml_metal_kargs_group_norm args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.ngrp =*/ ngrp, + /*.eps =*/ eps, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_group_norm(lib, op); + + int nth = 32; // SIMD width + //while (nth < ne00/4 && nth < ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)) { + // nth *= 2; + //} + + //nth = std::min(nth, ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)); + //nth = std::min(nth, ne00/4); + + const size_t smem = ggml_metal_pipeline_get_smem(pipeline); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + ggml_metal_encoder_dispatch_threadgroups(enc, ngrp, 1, 1, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_norm(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + float eps; + memcpy(&eps, op->op_params, sizeof(float)); + + ggml_metal_kargs_norm args = { + /*.ne00 =*/ ne00, + /*.ne00_4 =*/ ne00/4, + /*.nb01 =*/ nb01, + /*.eps =*/ eps, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_norm(lib, op); + + int nth = 32; // SIMD width + while (nth < ne00/4 && nth < ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)) { + nth *= 2; + } + + nth = std::min(nth, ggml_metal_pipeline_max_theads_per_threadgroup(pipeline)); + nth = std::min(nth, ne00/4); + + const size_t smem = ggml_metal_pipeline_get_smem(pipeline); + + const int64_t nrows = ggml_nrows(op->src[0]); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + ggml_metal_encoder_dispatch_threadgroups(enc, nrows, 1, 1, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_rope(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + // make sure we have one or more position id(ne10) per token(ne02) + GGML_ASSERT(ne10 % ne02 == 0); + GGML_ASSERT(ne10 >= ne02); + + const int nth = std::min(1024, ne00); + + const int n_past = ((const int32_t *) op->op_params)[0]; + const int n_dims = ((const int32_t *) op->op_params)[1]; + //const int mode = ((const int32_t *) op->op_params)[2]; + // skip 3, n_ctx, used in GLM RoPE, unimplemented in metal + const int n_ctx_orig = ((const int32_t *) op->op_params)[4]; + + float freq_base; + float freq_scale; + float ext_factor; + float attn_factor; + float beta_fast; + float beta_slow; + + memcpy(&freq_base, (const int32_t *) op->op_params + 5, sizeof(float)); + memcpy(&freq_scale, (const int32_t *) op->op_params + 6, sizeof(float)); + memcpy(&ext_factor, (const int32_t *) op->op_params + 7, sizeof(float)); + memcpy(&attn_factor, (const int32_t *) op->op_params + 8, sizeof(float)); + memcpy(&beta_fast, (const int32_t *) op->op_params + 9, sizeof(float)); + memcpy(&beta_slow, (const int32_t *) op->op_params + 10, sizeof(float)); + + // mrope + const int sect_0 = ((const int32_t *) op->op_params)[11]; + const int sect_1 = ((const int32_t *) op->op_params)[12]; + const int sect_2 = ((const int32_t *) op->op_params)[13]; + const int sect_3 = ((const int32_t *) op->op_params)[14]; + + ggml_metal_kargs_rope args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + /*.n_past =*/ n_past, + /*.n_dims =*/ n_dims, + /*.n_ctx_orig =*/ n_ctx_orig, + /*.freq_base =*/ freq_base, + /*.freq_scale =*/ freq_scale, + /*.ext_factor =*/ ext_factor, + /*.attn_factor =*/ attn_factor, + /*.beta_fast =*/ beta_fast, + /*.beta_slow =*/ beta_slow, + /* sect_0 =*/ sect_0, + /* sect_1 =*/ sect_1, + /* sect_2 =*/ sect_2, + /* sect_3 =*/ sect_3, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_rope(lib, op); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + if (op->src[2]) { + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[2]), 3); + } else { + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 3); + } + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 4); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne01, ne02, ne03, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_im2col(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + const int32_t s0 = ((const int32_t *)(op->op_params))[0]; + const int32_t s1 = ((const int32_t *)(op->op_params))[1]; + const int32_t p0 = ((const int32_t *)(op->op_params))[2]; + const int32_t p1 = ((const int32_t *)(op->op_params))[3]; + const int32_t d0 = ((const int32_t *)(op->op_params))[4]; + const int32_t d1 = ((const int32_t *)(op->op_params))[5]; + + const bool is_2D = ((const int32_t *)(op->op_params))[6] == 1; + + const int32_t N = op->src[1]->ne[is_2D ? 3 : 2]; + const int32_t IC = op->src[1]->ne[is_2D ? 2 : 1]; + const int32_t IH = is_2D ? op->src[1]->ne[1] : 1; + const int32_t IW = op->src[1]->ne[0]; + + const int32_t KH = is_2D ? op->src[0]->ne[1] : 1; + const int32_t KW = op->src[0]->ne[0]; + + const int32_t OH = is_2D ? op->ne[2] : 1; + const int32_t OW = op->ne[1]; + + const int32_t CHW = IC * KH * KW; + + const uint64_t ofs0 = op->src[1]->nb[is_2D ? 3 : 2] / 4; + const uint64_t ofs1 = op->src[1]->nb[is_2D ? 2 : 1] / 4; + + + ggml_metal_kargs_im2col args = { + /*.ofs0 =*/ ofs0, + /*.ofs1 =*/ ofs1, + /*.IW =*/ IW, + /*.IH =*/ IH, + /*.CHW =*/ CHW, + /*.s0 =*/ s0, + /*.s1 =*/ s1, + /*.p0 =*/ p0, + /*.p1 =*/ p1, + /*.d0 =*/ d0, + /*.d1 =*/ d1, + /*.N =*/ N, + /*.KH =*/ KH, + /*.KW =*/ KW, + /*.KHW =*/ KH * KW, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_im2col(lib, op); + + const uint64_t n_threads = std::min(ggml_metal_pipeline_max_theads_per_threadgroup(pipeline), N); + const int64_t quotient = N / n_threads + (N % n_threads > 0 ? 1 : 0); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_dispatch_threadgroups(enc, quotient * CHW, OH, OW, n_threads, 1, 1); + + return 1; +} + +int ggml_metal_op_conv_transpose_1d(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne1, op->src[1], ne); + GGML_TENSOR_LOCALS(uint64_t, nb1, op->src[1], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + const int32_t s0 = ((const int32_t *)(op->op_params))[0]; + + const int32_t IC = op->src[1]->ne[1]; + const int32_t IL = op->src[1]->ne[0]; + + const int32_t K = op->src[0]->ne[0]; + + const int32_t OL = op->ne[0]; + const int32_t OC = op->ne[1]; + + ggml_metal_kargs_conv_transpose_1d args = { + /*.IC =*/ IC, + /*.IL =*/ IL, + /*.K =*/ K, + /*.s0 =*/ s0, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_conv_transpose_1d(lib, op); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[1]), 2); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 3); + + ggml_metal_encoder_dispatch_threadgroups(enc, OL, OC, 1, 1, 1, 1); + + return 1; +} + +int ggml_metal_op_upscale(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + const float sf0 = (float)ne0/op->src[0]->ne[0]; + const float sf1 = (float)ne1/op->src[0]->ne[1]; + const float sf2 = (float)ne2/op->src[0]->ne[2]; + const float sf3 = (float)ne3/op->src[0]->ne[3]; + + ggml_metal_kargs_upscale args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + /*.sf0 =*/ sf0, + /*.sf1 =*/ sf1, + /*.sf2 =*/ sf2, + /*.sf3 =*/ sf3 + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_upscale(lib, op); + + const int nth = std::min(ggml_metal_pipeline_max_theads_per_threadgroup(pipeline), ne0); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne1, ne2, ne3, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_pad(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + ggml_metal_kargs_pad args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3 + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_pad(lib, op); + + const int nth = std::min(1024, ne0); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne1, ne2, ne3, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_pad_reflect_1d(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + ggml_metal_kargs_pad_reflect_1d args = { + /*.ne00 =*/ ne00, + /*.ne01 =*/ ne01, + /*.ne02 =*/ ne02, + /*.ne03 =*/ ne03, + /*.nb00 =*/ nb00, + /*.nb01 =*/ nb01, + /*.nb02 =*/ nb02, + /*.nb03 =*/ nb03, + /*.ne0 =*/ ne0, + /*.ne1 =*/ ne1, + /*.ne2 =*/ ne2, + /*.ne3 =*/ ne3, + /*.nb0 =*/ nb0, + /*.nb1 =*/ nb1, + /*.nb2 =*/ nb2, + /*.nb3 =*/ nb3, + /*.p0 =*/ ((const int32_t *)(op->op_params))[0], + /*.p1 =*/ ((const int32_t *)(op->op_params))[1] + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_pad_reflect_1d(lib, op); + + const int nth = std::min(1024, ne0); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne1, ne2, ne3, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_arange(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + float start; + float step; + + memcpy(&start, ((const int32_t *) op->op_params) + 0, sizeof(float)); + memcpy(&step, ((const int32_t *) op->op_params) + 2, sizeof(float)); + + ggml_metal_kargs_arange args = { + /*.ne0 =*/ ne0, + /*.start =*/ start, + /*.step =*/ step + }; + + const int nth = std::min(1024, ne0); + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_arange(lib, op); + + //[encoder setComputePipelineState:pipeline]; + //[encoder setBuffer:id_dst offset:offs_dst atIndex:0]; + //[encoder setBytes:&args length:sizeof(args) atIndex:1]; + + //[encoder dispatchThreadgroups:MTLSizeMake(1, 1, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 1); + + ggml_metal_encoder_dispatch_threadgroups(enc, 1, 1, 1, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_timestep_embedding(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + const int dim = op->op_params[0]; + const int max_period = op->op_params[1]; + + ggml_metal_kargs_timestep_embedding args = { + /*.nb1 =*/ nb1, + /*.dim =*/ dim, + /*.max_period =*/ max_period, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_timestep_embedding(lib, op); + + const int nth = std::max(1, std::min(1024, dim/2)); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_dispatch_threadgroups(enc, ne00, 1, 1, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_argmax(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + ggml_metal_kargs_argmax args = { + /*.ne00 = */ ne00, + /*.nb01 = */ nb01, + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_argmax(lib, op); + + const int64_t nrows = ggml_nrows(op->src[0]); + + int nth = 32; // SIMD width + while (nth < ne00 && nth*ne01*ne02*ne03 < 256) { + nth *= 2; + } + + const size_t smem = ggml_metal_pipeline_get_smem(pipeline); + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + ggml_metal_encoder_dispatch_threadgroups(enc, nrows, 1, 1, nth, 1, 1); + + return 1; +} + +int ggml_metal_op_argsort(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + // bitonic sort requires the number of elements to be power of 2 + int64_t ne00_padded = 1; + while (ne00_padded < ne00) { + ne00_padded *= 2; + } + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_argsort(lib, op); + + const int64_t nrows = ggml_nrows(op->src[0]); + + // Metal kernels require the buffer size to be multiple of 16 bytes + // https://developer.apple.com/documentation/metal/mtlcomputecommandencoder/1443142-setthreadgroupmemorylength + const size_t smem = GGML_PAD(ne00_padded*sizeof(int32_t), 16); + + ggml_metal_kargs_argsort args = { + /*.ncols =*/ ne00, + /*.ncols_pad =*/ ne00_padded + }; + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_set_threadgroup_memory_size(enc, smem, 0); + + ggml_metal_encoder_dispatch_threadgroups(enc, 1, nrows, 1, ne00_padded, 1, 1); + + return 1; +} + +int ggml_metal_op_leaky_relu(ggml_metal_op_t ctx, int idx) { + ggml_cgraph * gf = ctx->gf; + ggml_tensor * op = ggml_graph_node(gf, idx); + + ggml_metal_library_t lib = ctx->lib; + ggml_metal_encoder_t enc = ctx->enc; + + GGML_TENSOR_LOCALS( int32_t, ne0, op->src[0], ne); + GGML_TENSOR_LOCALS(uint64_t, nb0, op->src[0], nb); + GGML_TENSOR_LOCALS( int32_t, ne, op, ne); + GGML_TENSOR_LOCALS(uint32_t, nb, op, nb); + + float slope; + memcpy(&slope, op->op_params, sizeof(float)); + + ggml_metal_kargs_leaky_relu args = { + /*.slope =*/ slope + }; + + ggml_metal_pipeline_t pipeline = ggml_metal_library_get_pipeline_unary(lib, op); + + int64_t n = ggml_nelements(op); + + if (n % 4 == 0) { + n /= 4; + } + + ggml_metal_encoder_set_pipeline(enc, pipeline); + ggml_metal_encoder_set_bytes (enc, &args, sizeof(args), 0); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op->src[0]), 1); + ggml_metal_encoder_set_buffer (enc, ggml_metal_get_buffer_id(op), 2); + + ggml_metal_encoder_dispatch_threadgroups(enc, n, 1, 1, 1, 1, 1); + + return 1; +} diff --git a/ggml/src/ggml-metal/ggml-metal-ops.h b/ggml/src/ggml-metal/ggml-metal-ops.h new file mode 100644 index 0000000000..b620de164d --- /dev/null +++ b/ggml/src/ggml-metal/ggml-metal-ops.h @@ -0,0 +1,81 @@ +#pragma once + +#include "ggml-metal-device.h" + +#ifdef __cplusplus +extern "C" { +#endif + +typedef struct ggml_metal_op * ggml_metal_op_t; + +ggml_metal_op_t ggml_metal_op_init( + ggml_metal_device_t dev, + ggml_metal_cmd_buf_t cmd_buf, + struct ggml_cgraph * gf, + int idx_start, + int idx_end, + bool use_fusion, + bool use_concurrency, + bool use_capture, + int debug_graph, + int debug_fusion); + +void ggml_metal_op_free(ggml_metal_op_t ctx); + +int ggml_metal_op_encode(ggml_metal_op_t ctx, int idx); + +// +// available ops: +// + +// tokens per expert +size_t ggml_metal_op_mul_mat_id_extra_tpe(const struct ggml_tensor * op); + +// id map [n_tokens, n_expert] +size_t ggml_metal_op_mul_mat_id_extra_ids(const struct ggml_tensor * op); + +// return true if we should use the FA vector kernel for this op +bool ggml_metal_op_flash_attn_ext_use_vec(const struct ggml_tensor * op); + +size_t ggml_metal_op_flash_attn_ext_extra_tmp(const struct ggml_tensor * op); + +int ggml_metal_op_concat (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_repeat (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_acc (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_scale (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_clamp (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_unary (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_glu (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_sum_rows (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_get_rows (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_set_rows (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_soft_max (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_ssm_conv (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_ssm_scan (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_rwkv (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_cpy (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_pool_2d (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_mul_mat (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_mul_mat_id (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_add_id (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_flash_attn_ext (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_bin (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_rms_norm (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_l2_norm (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_group_norm (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_norm (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_rope (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_im2col (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_conv_transpose_1d (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_upscale (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_pad (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_pad_reflect_1d (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_arange (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_timestep_embedding(ggml_metal_op_t ctx, int idx); +int ggml_metal_op_argmax (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_argsort (ggml_metal_op_t ctx, int idx); +int ggml_metal_op_leaky_relu (ggml_metal_op_t ctx, int idx); + +#ifdef __cplusplus +} +#endif diff --git a/ggml/src/ggml-metal/ggml-metal.cpp b/ggml/src/ggml-metal/ggml-metal.cpp new file mode 100644 index 0000000000..fd0e6ed6e4 --- /dev/null +++ b/ggml/src/ggml-metal/ggml-metal.cpp @@ -0,0 +1,718 @@ +#include "ggml-metal.h" + +#include "ggml-impl.h" +#include "ggml-backend-impl.h" + +#include "ggml-metal-device.h" +#include "ggml-metal-context.h" +#include "ggml-metal-ops.h" + +// globals + +// initialized in ggml_backend_metal_reg +static ggml_backend_reg g_ggml_metal_reg; +static ggml_backend_device g_ggml_metal_device; + +//////////////////////////////////////////////////////////////////////////////// +// backend interface +//////////////////////////////////////////////////////////////////////////////// + +// shared buffer + +static void ggml_backend_metal_buffer_shared_free_buffer(ggml_backend_buffer_t buffer) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(ggml_metal_buffer_is_shared(ctx)); + + ggml_metal_buffer_free(ctx); +} + +static void * ggml_backend_metal_buffer_shared_get_base(ggml_backend_buffer_t buffer) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(ggml_metal_buffer_is_shared(ctx)); + + return ggml_metal_buffer_get_base(ctx); +} + +static void ggml_backend_metal_buffer_shared_memset_tensor(ggml_backend_buffer_t buffer, ggml_tensor * tensor, uint8_t value, size_t offset, size_t size) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(ggml_metal_buffer_is_shared(ctx)); + + ggml_metal_buffer_memset_tensor(ctx, tensor, value, offset, size); +} + +static void ggml_backend_metal_buffer_shared_set_tensor(ggml_backend_buffer_t buffer, ggml_tensor * tensor, const void * data, size_t offset, size_t size) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(ggml_metal_buffer_is_shared(ctx)); + + ggml_metal_buffer_set_tensor(ctx, tensor, data, offset, size); +} + +static void ggml_backend_metal_buffer_shared_get_tensor(ggml_backend_buffer_t buffer, const ggml_tensor * tensor, void * data, size_t offset, size_t size) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(ggml_metal_buffer_is_shared(ctx)); + + ggml_metal_buffer_get_tensor(ctx, tensor, data, offset, size); +} + +static bool ggml_backend_metal_buffer_shared_cpy_tensor(ggml_backend_buffer_t buffer, const ggml_tensor * src, ggml_tensor * dst) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(ggml_metal_buffer_is_shared(ctx)); + + GGML_UNUSED(buffer); + GGML_UNUSED(src); + GGML_UNUSED(dst); + + return false; +} + +static void ggml_backend_metal_buffer_shared_clear(ggml_backend_buffer_t buffer, uint8_t value) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(ggml_metal_buffer_is_shared(ctx)); + + ggml_metal_buffer_clear(ctx, value); +} + +static ggml_backend_buffer_i ggml_backend_metal_buffer_shared_i = { + /* .free_buffer = */ ggml_backend_metal_buffer_shared_free_buffer, + /* .get_base = */ ggml_backend_metal_buffer_shared_get_base, + /* .init_tensor = */ NULL, + /* .memset_tensor = */ ggml_backend_metal_buffer_shared_memset_tensor, + /* .set_tensor = */ ggml_backend_metal_buffer_shared_set_tensor, + /* .get_tensor = */ ggml_backend_metal_buffer_shared_get_tensor, + /* .cpy_tensor = */ ggml_backend_metal_buffer_shared_cpy_tensor, + /* .clear = */ ggml_backend_metal_buffer_shared_clear, + /* .reset = */ NULL, +}; + +// private buffer + +static void ggml_backend_metal_buffer_private_free_buffer(ggml_backend_buffer_t buffer) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(!ggml_metal_buffer_is_shared(ctx)); + + ggml_metal_buffer_free(ctx); +} + +static void * ggml_backend_metal_buffer_private_get_base(ggml_backend_buffer_t buffer) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(!ggml_metal_buffer_is_shared(ctx)); + + return ggml_metal_buffer_get_base(ctx); +} + +static void ggml_backend_metal_buffer_private_memset_tensor(ggml_backend_buffer_t buffer, ggml_tensor * tensor, uint8_t value, size_t offset, size_t size) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(!ggml_metal_buffer_is_shared(ctx)); + + ggml_metal_buffer_memset_tensor(ctx, tensor, value, offset, size); +} + +static void ggml_backend_metal_buffer_private_set_tensor(ggml_backend_buffer_t buffer, ggml_tensor * tensor, const void * data, size_t offset, size_t size) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(!ggml_metal_buffer_is_shared(ctx)); + + ggml_metal_buffer_set_tensor(ctx, tensor, data, offset, size); +} + +static void ggml_backend_metal_buffer_private_get_tensor(ggml_backend_buffer_t buffer, const ggml_tensor * tensor, void * data, size_t offset, size_t size) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(!ggml_metal_buffer_is_shared(ctx)); + + ggml_metal_buffer_get_tensor(ctx, tensor, data, offset, size); +} + +static bool ggml_backend_metal_buffer_private_cpy_tensor(ggml_backend_buffer_t buffer, const ggml_tensor * src, ggml_tensor * dst) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(!ggml_metal_buffer_is_shared(ctx)); + + GGML_UNUSED(buffer); + GGML_UNUSED(src); + GGML_UNUSED(dst); + + return false; +} + +static void ggml_backend_metal_buffer_private_clear(ggml_backend_buffer_t buffer, uint8_t value) { + ggml_metal_buffer_t ctx = (ggml_metal_buffer_t)buffer->context; + + GGML_ASSERT(!ggml_metal_buffer_is_shared(ctx)); + + ggml_metal_buffer_clear(ctx, value); +} + +static ggml_backend_buffer_i ggml_backend_metal_buffer_private_i = { + /* .free_buffer = */ ggml_backend_metal_buffer_private_free_buffer, + /* .get_base = */ ggml_backend_metal_buffer_private_get_base, + /* .init_tensor = */ NULL, + /* .memset_tensor = */ ggml_backend_metal_buffer_private_memset_tensor, + /* .set_tensor = */ ggml_backend_metal_buffer_private_set_tensor, + /* .get_tensor = */ ggml_backend_metal_buffer_private_get_tensor, + /* .cpy_tensor = */ ggml_backend_metal_buffer_private_cpy_tensor, + /* .clear = */ ggml_backend_metal_buffer_private_clear, + /* .reset = */ NULL, +}; + +// +// buffer types +// + +// common method for allocating shread or private Metal buffers +static ggml_backend_buffer_t ggml_backend_metal_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size, bool shared) { + ggml_metal_device_t ctx_dev = (ggml_metal_device_t)buft->device->context; + ggml_metal_buffer_t res = ggml_metal_buffer_init(ctx_dev, size, shared); + + ggml_backend_buffer_i buf_i = ggml_metal_buffer_is_shared(res) + ? ggml_backend_metal_buffer_shared_i + : ggml_backend_metal_buffer_private_i; + + return ggml_backend_buffer_init(buft, buf_i, res, size); +} + +static size_t ggml_backend_metal_buffer_type_get_alloc_size(ggml_backend_buffer_type_t buft, const ggml_tensor * tensor) { + size_t res = ggml_nbytes(tensor); + + // some operations require additional memory for fleeting data: + switch (tensor->op) { + case GGML_OP_MUL_MAT_ID: + { + res += ggml_metal_op_mul_mat_id_extra_tpe(tensor); + res += ggml_metal_op_mul_mat_id_extra_ids(tensor); + } break; + case GGML_OP_FLASH_ATTN_EXT: + { + if (ggml_metal_op_flash_attn_ext_use_vec(tensor)) { + res += ggml_metal_op_flash_attn_ext_extra_tmp(tensor); + } + } break; + default: + break; + } + + return res; + + GGML_UNUSED(buft); +} + +// default (shared) buffer type + +static const char * ggml_backend_metal_buffer_type_shared_get_name(ggml_backend_buffer_type_t buft) { + return "Metal"; + + GGML_UNUSED(buft); +} + +static ggml_backend_buffer_t ggml_backend_metal_buffer_type_shared_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { + return ggml_backend_metal_buffer_type_alloc_buffer(buft, size, true); +} + +static size_t ggml_backend_metal_buffer_type_shared_get_alignment(ggml_backend_buffer_type_t buft) { + return 32; + + GGML_UNUSED(buft); +} + +static size_t ggml_backend_metal_buffer_type_shared_get_max_size(ggml_backend_buffer_type_t buft) { + ggml_metal_device_t ctx_dev = (ggml_metal_device_t)buft->device->context; + + return ggml_metal_device_get_props(ctx_dev)->max_buffer_size; +} + +static size_t ggml_backend_metal_buffer_type_shared_get_alloc_size(ggml_backend_buffer_type_t buft, const ggml_tensor * tensor) { + return ggml_backend_metal_buffer_type_get_alloc_size(buft, tensor); +} + +static bool ggml_backend_metal_buffer_type_shared_is_host(ggml_backend_buffer_type_t buft) { + return false; + + GGML_UNUSED(buft); +} + +static ggml_backend_buffer_type_t ggml_backend_metal_buffer_type_shared(void) { + static ggml_backend_buffer_type ggml_backend_buffer_type_metal = { + /* .iface = */ { + /* .get_name = */ ggml_backend_metal_buffer_type_shared_get_name, + /* .alloc_buffer = */ ggml_backend_metal_buffer_type_shared_alloc_buffer, + /* .get_alignment = */ ggml_backend_metal_buffer_type_shared_get_alignment, + /* .get_max_size = */ ggml_backend_metal_buffer_type_shared_get_max_size, + /* .get_alloc_size = */ ggml_backend_metal_buffer_type_shared_get_alloc_size, + /* .is_host = */ ggml_backend_metal_buffer_type_shared_is_host, + }, + /* .device = */ &g_ggml_metal_device, + /* .context = */ NULL, + }; + + return &ggml_backend_buffer_type_metal; +} + +// default (private) buffer type + +static const char * ggml_backend_metal_buffer_type_private_get_name(ggml_backend_buffer_type_t buft) { + return "Metal_Private"; + + GGML_UNUSED(buft); +} + +static ggml_backend_buffer_t ggml_backend_metal_buffer_type_private_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { + return ggml_backend_metal_buffer_type_alloc_buffer(buft, size, false); +} + +static size_t ggml_backend_metal_buffer_type_private_get_alignment(ggml_backend_buffer_type_t buft) { + return 32; + + GGML_UNUSED(buft); +} + +static size_t ggml_backend_metal_buffer_type_private_get_max_size(ggml_backend_buffer_type_t buft) { + ggml_metal_device_t ctx_dev = (ggml_metal_device_t)buft->device->context; + + return ggml_metal_device_get_props(ctx_dev)->max_buffer_size; +} + +static size_t ggml_backend_metal_buffer_type_private_get_alloc_size(ggml_backend_buffer_type_t buft, const ggml_tensor * tensor) { + return ggml_backend_metal_buffer_type_get_alloc_size(buft, tensor); +} + +static bool ggml_backend_metal_buffer_type_private_is_host(ggml_backend_buffer_type_t buft) { + return false; + + GGML_UNUSED(buft); +} + +static ggml_backend_buffer_type_t ggml_backend_metal_buffer_type_private(void) { + static ggml_backend_buffer_type ggml_backend_buffer_type_metal = { + /* .iface = */ { + /* .get_name = */ ggml_backend_metal_buffer_type_private_get_name, + /* .alloc_buffer = */ ggml_backend_metal_buffer_type_private_alloc_buffer, + /* .get_alignment = */ ggml_backend_metal_buffer_type_private_get_alignment, + /* .get_max_size = */ ggml_backend_metal_buffer_type_private_get_max_size, + /* .get_alloc_size = */ ggml_backend_metal_buffer_type_private_get_alloc_size, + /* .is_host = */ ggml_backend_metal_buffer_type_private_is_host, + }, + /* .device = */ &g_ggml_metal_device, + /* .context = */ NULL, + }; + + return &ggml_backend_buffer_type_metal; +} + +// mapped buffer type + +static const char * ggml_backend_metal_buffer_type_mapped_get_name(ggml_backend_buffer_type_t buft) { + return "Metal_Mapped"; + + GGML_UNUSED(buft); +} + +static ggml_backend_buffer_t ggml_backend_metal_buffer_type_mapped_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { + // for mapped buffers, prefer shared memory + return ggml_backend_metal_buffer_type_alloc_buffer(buft, size, true); +} + +static size_t ggml_backend_metal_buffer_type_mapped_get_alignment(ggml_backend_buffer_type_t buft) { + return 32; + + GGML_UNUSED(buft); +} + +static size_t ggml_backend_metal_buffer_type_mapped_get_max_size(ggml_backend_buffer_type_t buft) { + ggml_metal_device_t ctx_dev = (ggml_metal_device_t)buft->device->context; + + return ggml_metal_device_get_props(ctx_dev)->max_buffer_size; +} + +static size_t ggml_backend_metal_buffer_type_mapped_get_alloc_size(ggml_backend_buffer_type_t buft, const ggml_tensor * tensor) { + return ggml_backend_metal_buffer_type_get_alloc_size(buft, tensor); +} + +static bool ggml_backend_metal_buffer_type_mapped_is_host(ggml_backend_buffer_type_t buft) { + return false; + + GGML_UNUSED(buft); +} + +static ggml_backend_buffer_type_t ggml_backend_metal_buffer_type_mapped(void) { + // note: not obvious, but this buffer type still needs to implement .alloc_buffer: + // https://github.com/ggml-org/llama.cpp/pull/15832#discussion_r2333177099 + static ggml_backend_buffer_type ggml_backend_buffer_type_mapped_metal = { + /* .iface = */ { + /* .get_name = */ ggml_backend_metal_buffer_type_mapped_get_name, + /* .alloc_buffer = */ ggml_backend_metal_buffer_type_mapped_alloc_buffer, + /* .get_alignment = */ ggml_backend_metal_buffer_type_mapped_get_alignment, + /* .get_max_size = */ ggml_backend_metal_buffer_type_mapped_get_max_size, + /* .get_alloc_size = */ ggml_backend_metal_buffer_type_mapped_get_alloc_size, + /* .is_host = */ ggml_backend_metal_buffer_type_mapped_is_host, + }, + /* .device = */ &g_ggml_metal_device, + /* .context = */ NULL, + }; + + return &ggml_backend_buffer_type_mapped_metal; +} + +// backend + +static const char * ggml_backend_metal_name(ggml_backend_t backend) { + return "Metal"; + + GGML_UNUSED(backend); +} + +static void ggml_backend_metal_free(ggml_backend_t backend) { + ggml_metal_t ctx = (ggml_metal_t)backend->context; + + // wait for any ongoing async operations to finish + ggml_metal_synchronize(ctx); + + ggml_metal_free(ctx); + + free(backend); +} + +static void ggml_backend_metal_synchronize(ggml_backend_t backend) { + ggml_metal_t ctx = (ggml_metal_t)backend->context; + + ggml_metal_synchronize(ctx); +} + +static void ggml_backend_metal_set_tensor_async(ggml_backend_t backend, ggml_tensor * tensor, const void * data, size_t offset, size_t size) { + ggml_metal_t ctx = (ggml_metal_t)backend->context; + + ggml_metal_set_tensor_async(ctx, tensor, data, offset, size); +} + +static void ggml_backend_metal_get_tensor_async(ggml_backend_t backend, const ggml_tensor * tensor, void * data, size_t offset, size_t size) { + ggml_metal_t ctx = (ggml_metal_t)backend->context; + + ggml_metal_get_tensor_async(ctx, tensor, data, offset, size); +} + +static bool ggml_backend_metal_cpy_tensor_async(ggml_backend_t backend_src, ggml_backend_t backend_dst, const ggml_tensor * src, ggml_tensor * dst) { + return false; + + GGML_UNUSED(backend_src); + GGML_UNUSED(backend_dst); + GGML_UNUSED(src); + GGML_UNUSED(dst); +} + +static enum ggml_status ggml_backend_metal_graph_compute(ggml_backend_t backend, ggml_cgraph * cgraph) { + ggml_metal_t ctx = (ggml_metal_t)backend->context; + + return ggml_metal_graph_compute(ctx, cgraph); +} + +static void ggml_backend_metal_graph_optimize(ggml_backend_t backend, ggml_cgraph * cgraph) { + ggml_metal_t ctx = (ggml_metal_t)backend->context; + + ggml_metal_graph_optimize(ctx, cgraph); +} + +static void ggml_backend_metal_set_n_cb(ggml_backend_t backend, int n_cb) { + GGML_ASSERT(ggml_backend_is_metal(backend)); + + ggml_metal_t ctx = (ggml_metal_t)backend->context; + + ggml_metal_set_n_cb(ctx, n_cb); + +} + +static ggml_backend_i ggml_backend_metal_i = { + /* .get_name = */ ggml_backend_metal_name, + /* .free = */ ggml_backend_metal_free, + /* .set_tensor_async = */ ggml_backend_metal_set_tensor_async, + /* .get_tensor_async = */ ggml_backend_metal_get_tensor_async, + /* .cpy_tensor_async = */ ggml_backend_metal_cpy_tensor_async, // only needed for multi-GPU setups + /* .synchronize = */ ggml_backend_metal_synchronize, + /* .graph_plan_create = */ NULL, + /* .graph_plan_free = */ NULL, + /* .graph_plan_update = */ NULL, + /* .graph_plan_compute = */ NULL, + /* .graph_compute = */ ggml_backend_metal_graph_compute, + + // the events API is needed only for multi-GPU setups, so likely no need to implement it for Metal + // in any case, these docs seem relevant if we ever decide to implement it: + // https://developer.apple.com/documentation/metal/mtlcommandbuffer#Synchronizing-Passes-with-Events + /* .event_record = */ NULL, + /* .event_wait = */ NULL, + /* .optimize_graph = */ ggml_backend_metal_graph_optimize, +}; + +static ggml_guid_t ggml_backend_metal_guid(void) { + static ggml_guid guid = { 0x81, 0xa1, 0x8b, 0x1e, 0x71, 0xec, 0x79, 0xed, 0x2b, 0x85, 0xdc, 0x8a, 0x61, 0x98, 0x30, 0xe6 }; + return &guid; +} + +ggml_backend_t ggml_backend_metal_init(void) { + ggml_backend_dev_t dev = ggml_backend_reg_dev_get(ggml_backend_metal_reg(), 0); + ggml_metal_device_t ctx_dev = (ggml_metal_device_t)dev->context; + + ggml_metal_t ctx = ggml_metal_init(ctx_dev); + if (ctx == NULL) { + GGML_LOG_ERROR("%s: error: failed to allocate context\n", __func__); + return NULL; + } + + ggml_backend_t backend = (ggml_backend_t) malloc(sizeof(ggml_backend)); + + *backend = { + /* .guid = */ ggml_backend_metal_guid(), + /* .interface = */ ggml_backend_metal_i, + /* .device = */ dev, + /* .context = */ ctx, + }; + + ggml_backend_metal_set_n_cb(backend, 1); + + return backend; +} + +bool ggml_backend_is_metal(ggml_backend_t backend) { + return backend != NULL && ggml_guid_matches(backend->guid, ggml_backend_metal_guid()); +} + +void ggml_backend_metal_set_abort_callback(ggml_backend_t backend, ggml_abort_callback abort_callback, void * user_data) { + GGML_ASSERT(ggml_backend_is_metal(backend)); + + ggml_metal_t ctx = (ggml_metal_t)backend->context; + + ggml_metal_set_abort_callback(ctx, abort_callback, user_data); +} + +bool ggml_backend_metal_supports_family(ggml_backend_t backend, int family) { + GGML_ASSERT(ggml_backend_is_metal(backend)); + + ggml_metal_t ctx = (ggml_metal_t)backend->context; + + return ggml_metal_supports_family(ctx, family); +} + +void ggml_backend_metal_capture_next_compute(ggml_backend_t backend) { + GGML_ASSERT(ggml_backend_is_metal(backend)); + + ggml_metal_t ctx = (ggml_metal_t)backend->context; + + ggml_metal_capture_next_compute(ctx); +} + +// backend device + +static const char * ggml_backend_metal_device_get_name(ggml_backend_dev_t dev) { + return "Metal"; + + GGML_UNUSED(dev); +} + +static const char * ggml_backend_metal_device_get_description(ggml_backend_dev_t dev) { + ggml_metal_device_t ctx_dev = (ggml_metal_device_t)dev->context; + + return ggml_metal_device_get_props(ctx_dev)->name; +} + +static void ggml_backend_metal_device_get_memory(ggml_backend_dev_t dev, size_t * free, size_t * total) { + ggml_metal_device_t ctx_dev = (ggml_metal_device_t)dev->context; + + ggml_metal_device_get_memory(ctx_dev, free, total); +} + +static enum ggml_backend_dev_type ggml_backend_metal_device_get_type(ggml_backend_dev_t dev) { + return GGML_BACKEND_DEVICE_TYPE_GPU; + + GGML_UNUSED(dev); +} + +static void ggml_backend_metal_device_get_props(ggml_backend_dev_t dev, ggml_backend_dev_props * props) { + props->name = ggml_backend_metal_device_get_name(dev); + props->description = ggml_backend_metal_device_get_description(dev); + props->type = ggml_backend_metal_device_get_type(dev); + + ggml_backend_metal_device_get_memory(dev, &props->memory_free, &props->memory_total); + + props->caps = { + /* .async = */ true, + /* .host_buffer = */ false, + /* .buffer_from_host_ptr = */ true, + /* .events = */ false, + }; +} + +static ggml_backend_t ggml_backend_metal_device_init(ggml_backend_dev_t dev, const char * params) { + ggml_metal_device_t ctx_dev = (ggml_metal_device_t)dev->context; + + ggml_metal_t ctx = ggml_metal_init(ctx_dev); + if (ctx == NULL) { + GGML_LOG_ERROR("%s: error: failed to allocate context\n", __func__); + return NULL; + } + + ggml_backend_t backend = (ggml_backend_t) malloc(sizeof(ggml_backend)); + + *backend = { + /* .guid = */ ggml_backend_metal_guid(), + /* .interface = */ ggml_backend_metal_i, + /* .device = */ dev, + /* .context = */ ctx, + }; + + ggml_backend_metal_set_n_cb(backend, 1); + + return backend; + + GGML_UNUSED(params); +} + +static ggml_backend_buffer_type_t ggml_backend_metal_device_get_buffer_type(ggml_backend_dev_t dev) { + ggml_metal_device_t ctx_dev = (ggml_metal_device_t)dev->context; + + const ggml_metal_device_props * props_dev = ggml_metal_device_get_props(ctx_dev); + + return props_dev->use_shared_buffers ? ggml_backend_metal_buffer_type_shared() : ggml_backend_metal_buffer_type_private(); +} + +static ggml_backend_buffer_t ggml_backend_metal_device_buffer_mapped(ggml_backend_dev_t dev, void * ptr, size_t size, size_t max_tensor_size) { + ggml_metal_device_t ctx_dev = (ggml_metal_device_t)dev->context; + + ggml_metal_buffer_t res = ggml_metal_buffer_map(ctx_dev, ptr, size, max_tensor_size); + + return ggml_backend_buffer_init(ggml_backend_metal_buffer_type_mapped(), ggml_backend_metal_buffer_shared_i, res, size); +} + +static bool ggml_backend_metal_device_supports_op(ggml_backend_dev_t dev, const ggml_tensor * op) { + ggml_metal_device_t ctx_dev = (ggml_metal_device_t)dev->context; + + return ggml_metal_device_supports_op(ctx_dev, op); +} + +static bool ggml_backend_metal_device_supports_buft(ggml_backend_dev_t dev, ggml_backend_buffer_type_t buft) { + return + buft->iface.get_name == ggml_backend_metal_buffer_type_shared_get_name || + buft->iface.get_name == ggml_backend_metal_buffer_type_private_get_name || + buft->iface.get_name == ggml_backend_metal_buffer_type_mapped_get_name; + + GGML_UNUSED(dev); +} + +static int64_t get_op_batch_size(const ggml_tensor * op) { + switch (op->op) { + case GGML_OP_MUL_MAT: + return op->ne[1]; + case GGML_OP_MUL_MAT_ID: + return op->ne[2]; + default: + return ggml_nrows(op); + } +} + +static bool ggml_backend_metal_device_offload_op(ggml_backend_dev_t dev, const ggml_tensor * op) { + const int min_batch_size = 32; + + return (op->op == GGML_OP_MUL_MAT || + op->op == GGML_OP_MUL_MAT_ID) && + get_op_batch_size(op) >= min_batch_size; + + GGML_UNUSED(dev); + GGML_UNUSED(op); +} + +static ggml_backend_device_i ggml_backend_metal_device_i = { + /* .get_name = */ ggml_backend_metal_device_get_name, + /* .get_description = */ ggml_backend_metal_device_get_description, + /* .get_memory = */ ggml_backend_metal_device_get_memory, + /* .get_type = */ ggml_backend_metal_device_get_type, + /* .get_props = */ ggml_backend_metal_device_get_props, + /* .init_backend = */ ggml_backend_metal_device_init, + /* .get_buffer_type = */ ggml_backend_metal_device_get_buffer_type, + /* .get_host_buffer_type = */ NULL, + /* .buffer_from_host_ptr = */ ggml_backend_metal_device_buffer_mapped, + /* .supports_op = */ ggml_backend_metal_device_supports_op, + /* .supports_buft = */ ggml_backend_metal_device_supports_buft, + /* .offload_op = */ ggml_backend_metal_device_offload_op, + /* .event_new = */ NULL, + /* .event_free = */ NULL, + /* .event_synchronize = */ NULL, +}; + +// backend registry + +static const char * ggml_backend_metal_reg_get_name(ggml_backend_reg_t reg) { + return "Metal"; + + GGML_UNUSED(reg); +} + +static size_t ggml_backend_metal_reg_device_count(ggml_backend_reg_t reg) { + return 1; + + GGML_UNUSED(reg); +} + +static ggml_backend_dev_t ggml_backend_metal_reg_device_get(ggml_backend_reg_t reg, size_t index) { + GGML_ASSERT(index == 0); + + return &g_ggml_metal_device; + + GGML_UNUSED(reg); + GGML_UNUSED(index); +} + +static ggml_backend_feature g_ggml_backend_metal_features[] = { +#if defined(GGML_METAL_EMBED_LIBRARY) + { "EMBED_LIBRARY", "1" }, +#endif + { NULL, NULL }, +}; + +static ggml_backend_feature * ggml_backend_metal_get_features(ggml_backend_reg_t reg) { + return g_ggml_backend_metal_features; + + GGML_UNUSED(reg); +} + +static void * ggml_backend_metal_get_proc_address(ggml_backend_reg_t reg, const char * name) { + if (strcmp(name, "ggml_backend_get_features") == 0) { + return (void *)ggml_backend_metal_get_features; + } + + return NULL; + + GGML_UNUSED(reg); +} + +static ggml_backend_reg_i ggml_backend_metal_reg_i = { + /* .get_name = */ ggml_backend_metal_reg_get_name, + /* .device_count = */ ggml_backend_metal_reg_device_count, + /* .device_get = */ ggml_backend_metal_reg_device_get, + /* .get_proc_address = */ ggml_backend_metal_get_proc_address, +}; + +ggml_backend_reg_t ggml_backend_metal_reg(void) { + { + g_ggml_metal_reg = { + /* .api_version = */ GGML_BACKEND_API_VERSION, + /* .iface = */ ggml_backend_metal_reg_i, + /* .context = */ NULL, + }; + + g_ggml_metal_device = { + /* .iface = */ ggml_backend_metal_device_i, + /* .reg = */ &g_ggml_metal_reg, + /* .context = */ ggml_metal_device_get(), + }; + } + + return &g_ggml_metal_reg; +} + +GGML_BACKEND_DL_IMPL(ggml_backend_metal_reg) diff --git a/ggml/src/ggml-metal/ggml-metal.m b/ggml/src/ggml-metal/ggml-metal.m deleted file mode 100644 index 2243c174fb..0000000000 --- a/ggml/src/ggml-metal/ggml-metal.m +++ /dev/null @@ -1,6897 +0,0 @@ -#import "ggml-metal.h" - -#import "ggml-impl.h" -#import "ggml-backend-impl.h" -#import "ggml-metal-impl.h" -#import "ggml-metal-common.h" - -#import - -#import - -#undef MIN -#undef MAX -#define MIN(a, b) ((a) < (b) ? (a) : (b)) -#define MAX(a, b) ((a) > (b) ? (a) : (b)) - -// max memory buffers that can be mapped to the device -#define GGML_METAL_MAX_BUFFERS 64 - -// max number of MTLCommandBuffer used to submit a graph for processing -#define GGML_METAL_MAX_COMMAND_BUFFERS 8 - -#ifndef TARGET_OS_VISION -#define TARGET_OS_VISION 0 -#endif - -// create residency sets only on macOS >= 15.0 -#if !TARGET_CPU_X86_64 && TARGET_OS_OSX && __MAC_OS_X_VERSION_MAX_ALLOWED >= 150000 || \ - TARGET_OS_IOS && __IPHONE_OS_VERSION_MAX_ALLOWED >= 180000 || \ - TARGET_OS_TV && __TV_OS_VERSION_MAX_ALLOWED >= 180000 || \ - TARGET_OS_VISION && __VISION_OS_VERSION_MAX_ALLOWED >= 200000 -#define GGML_METAL_HAS_RESIDENCY_SETS 1 -#endif - -// globals - -// overload of MTLGPUFamilyMetal3 (not available in some environments) -static const NSInteger MTLGPUFamilyMetal3_GGML = 5001; - -// initialized in ggml_backend_metal_reg -static struct ggml_backend_reg g_ggml_backend_metal_reg; -static struct ggml_backend_device g_ggml_backend_metal_device; - -// information about a Metal device -// note: assumes single GPU device - the default one -// TODO: support multiple GPU devices -static struct ggml_backend_metal_device_context { - id mtl_device; - int mtl_device_ref_count; - id mtl_library; - - // a single global queue shared by all Metal backends - // technically not needed for devices with unified memory, but enables discrete GPUs support - // ref: https://github.com/ggml-org/llama.cpp/pull/15906 - id mtl_queue; - - NSLock * mtl_lock; - - bool has_simdgroup_reduction; - bool has_simdgroup_mm; - bool has_residency_sets; - bool has_bfloat; - bool use_bfloat; - bool use_fusion; - bool use_concurrency; - bool use_shared_buffers; - bool use_graph_optimize; - - int debug_graph; - int debug_fusion; - - // how many times a given op was fused - uint64_t fuse_cnt[GGML_OP_COUNT]; - - size_t max_size; - - char name[128]; -} g_ggml_ctx_dev_main = { - /*.mtl_device =*/ nil, - /*.mtl_device_ref_count =*/ 0, - /*.mtl_library =*/ nil, - /*.mtl_queue =*/ nil, - /*.mtl_lock =*/ nil, - /*.has_simdgroup_reduction =*/ false, - /*.has_simdgroup_mm =*/ false, - /*.has_residency_sets =*/ false, - /*.has_bfloat =*/ false, - /*.use_bfloat =*/ false, - /*.use_fusion =*/ true, - /*.use_concurrency =*/ true, - /*.use_shared_buffers =*/ true, - /*.use_graph_optimize =*/ true, - /*.debug_graph =*/ 0, - /*.debug_fusion =*/ 0, - /*.fuse_cnt =*/ { 0 }, - /*.max_size =*/ 0, - /*.name =*/ "", -}; - -// acquire -static id ggml_backend_metal_device_acq(struct ggml_backend_metal_device_context * ctx) { - assert(ctx != NULL); - - if (ctx->mtl_lock == nil) { - ctx->mtl_lock = [[NSLock alloc] init]; - } - - if (ctx->mtl_device == nil) { - ctx->mtl_device = MTLCreateSystemDefaultDevice(); - - if (ctx->mtl_device) { - ctx->mtl_queue = [ctx->mtl_device newCommandQueue]; - if (ctx->mtl_queue == nil) { - GGML_LOG_ERROR("%s: error: failed to create command queue\n", __func__); - } - - ctx->has_simdgroup_reduction = [ctx->mtl_device supportsFamily:MTLGPUFamilyApple7]; - ctx->has_simdgroup_reduction |= [ctx->mtl_device supportsFamily:MTLGPUFamilyMetal3_GGML]; - - ctx->has_simdgroup_mm = [ctx->mtl_device supportsFamily:MTLGPUFamilyApple7]; - -#if defined(GGML_METAL_HAS_RESIDENCY_SETS) - ctx->has_residency_sets = getenv("GGML_METAL_NO_RESIDENCY") == nil; -#endif - - ctx->has_bfloat = [ctx->mtl_device supportsFamily:MTLGPUFamilyMetal3_GGML]; - ctx->has_bfloat |= [ctx->mtl_device supportsFamily:MTLGPUFamilyApple6]; - -#if defined(GGML_METAL_USE_BF16) - ctx->use_bfloat = ctx->has_bfloat; -#else - ctx->use_bfloat = false; -#endif - - ctx->use_fusion = getenv("GGML_METAL_FUSION_DISABLE") == nil; - ctx->use_concurrency = getenv("GGML_METAL_CONCURRENCY_DISABLE") == nil; - - { - const char * val = getenv("GGML_METAL_GRAPH_DEBUG"); - ctx->debug_graph = val ? atoi(val) : 0; - } - - { - const char * val = getenv("GGML_METAL_FUSION_DEBUG"); - ctx->debug_fusion = val ? atoi(val) : 0; - } - - ctx->use_shared_buffers = ctx->mtl_device.hasUnifiedMemory; - - if (getenv("GGML_METAL_SHARED_BUFFERS_DISABLE") != NULL) { - ctx->use_shared_buffers = false; - } - - ctx->use_graph_optimize = true; - - if (getenv("GGML_METAL_GRAPH_OPTIMIZE_DISABLE") != NULL) { - ctx->use_graph_optimize = false; - } - - memset(ctx->fuse_cnt, 0, sizeof(ctx->fuse_cnt)); - - ctx->max_size = ctx->mtl_device.maxBufferLength; - - strncpy(ctx->name, [[ctx->mtl_device name] UTF8String], sizeof(ctx->name) - 1); - } - } - - ctx->mtl_device_ref_count++; - - return ctx->mtl_device; -} - -// release -static void ggml_backend_metal_device_rel(struct ggml_backend_metal_device_context * ctx) { - assert(ctx != NULL); - assert(ctx->mtl_device_ref_count > 0); - - ctx->mtl_device_ref_count--; - - if (ctx->mtl_device_ref_count == 0) { - if (ctx->debug_fusion > 0) { - fprintf(stderr, "%s: fusion stats:\n", __func__); - for (int i = 0; i < GGML_OP_COUNT; i++) { - if (ctx->fuse_cnt[i] == 0) { - continue; - } - - // note: cannot use ggml_log here - fprintf(stderr, "%s: - %s: %" PRIu64 "\n", __func__, ggml_op_name((enum ggml_op) i), ctx->fuse_cnt[i]); - } - } - - if (ctx->mtl_lock) { - [ctx->mtl_lock release]; - ctx->mtl_lock = nil; - } - - if (ctx->mtl_library) { - [ctx->mtl_library release]; - ctx->mtl_library = nil; - } - - if (ctx->mtl_queue) { - [ctx->mtl_queue release]; - ctx->mtl_queue = nil; - } - - if (ctx->mtl_device) { - [ctx->mtl_device release]; - ctx->mtl_device = nil; - } - } -} - -// kernels - -struct ggml_metal_kernel { - id pipeline; -}; - -@interface ggml_metal_kernel_wrapper : NSObject - -@property (nonatomic, assign) struct ggml_metal_kernel kernel; - -@end - -@implementation ggml_metal_kernel_wrapper -- (void) dealloc { - [_kernel.pipeline release]; - [super dealloc]; -} -@end - -enum ggml_metal_kernel_type { - GGML_METAL_KERNEL_TYPE_ADD_ID, - GGML_METAL_KERNEL_TYPE_REPEAT_F32, - GGML_METAL_KERNEL_TYPE_REPEAT_F16, - GGML_METAL_KERNEL_TYPE_REPEAT_I32, - GGML_METAL_KERNEL_TYPE_REPEAT_I16, - GGML_METAL_KERNEL_TYPE_SCALE, - GGML_METAL_KERNEL_TYPE_SCALE_4, - GGML_METAL_KERNEL_TYPE_CLAMP, - GGML_METAL_KERNEL_TYPE_TANH, - GGML_METAL_KERNEL_TYPE_RELU, - GGML_METAL_KERNEL_TYPE_SIGMOID, - GGML_METAL_KERNEL_TYPE_GELU, - GGML_METAL_KERNEL_TYPE_GELU_4, - GGML_METAL_KERNEL_TYPE_GELU_ERF, - GGML_METAL_KERNEL_TYPE_GELU_ERF_4, - GGML_METAL_KERNEL_TYPE_GELU_QUICK, - GGML_METAL_KERNEL_TYPE_GELU_QUICK_4, - GGML_METAL_KERNEL_TYPE_SILU, - GGML_METAL_KERNEL_TYPE_SILU_4, - GGML_METAL_KERNEL_TYPE_ELU, - GGML_METAL_KERNEL_TYPE_ABS, - GGML_METAL_KERNEL_TYPE_SGN, - GGML_METAL_KERNEL_TYPE_STEP, - GGML_METAL_KERNEL_TYPE_HARDSWISH, - GGML_METAL_KERNEL_TYPE_HARDSIGMOID, - GGML_METAL_KERNEL_TYPE_EXP, - GGML_METAL_KERNEL_TYPE_SOFT_MAX_F16, - GGML_METAL_KERNEL_TYPE_SOFT_MAX_F16_4, - GGML_METAL_KERNEL_TYPE_SOFT_MAX_F32, - GGML_METAL_KERNEL_TYPE_SOFT_MAX_F32_4, - GGML_METAL_KERNEL_TYPE_DIAG_MASK_INF, - GGML_METAL_KERNEL_TYPE_DIAG_MASK_INF_8, - GGML_METAL_KERNEL_TYPE_GET_ROWS_F32, - GGML_METAL_KERNEL_TYPE_GET_ROWS_F16, - GGML_METAL_KERNEL_TYPE_GET_ROWS_BF16, - GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_0, - GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_1, - GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_0, - GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_1, - GGML_METAL_KERNEL_TYPE_GET_ROWS_Q8_0, - GGML_METAL_KERNEL_TYPE_GET_ROWS_MXFP4, - GGML_METAL_KERNEL_TYPE_GET_ROWS_Q2_K, - GGML_METAL_KERNEL_TYPE_GET_ROWS_Q3_K, - GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_K, - GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_K, - GGML_METAL_KERNEL_TYPE_GET_ROWS_Q6_K, - GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_XXS, - GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_XS, - GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ3_XXS, - GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ3_S, - GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_S, - GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ1_S, - GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ1_M, - GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ4_NL, - GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ4_XS, - GGML_METAL_KERNEL_TYPE_GET_ROWS_I32, - GGML_METAL_KERNEL_TYPE_SET_ROWS_F32, - GGML_METAL_KERNEL_TYPE_SET_ROWS_F16, - GGML_METAL_KERNEL_TYPE_SET_ROWS_BF16, - GGML_METAL_KERNEL_TYPE_SET_ROWS_Q8_0, - GGML_METAL_KERNEL_TYPE_SET_ROWS_Q4_0, - GGML_METAL_KERNEL_TYPE_SET_ROWS_Q4_1, - GGML_METAL_KERNEL_TYPE_SET_ROWS_Q5_0, - GGML_METAL_KERNEL_TYPE_SET_ROWS_Q5_1, - GGML_METAL_KERNEL_TYPE_SET_ROWS_IQ4_NL, - GGML_METAL_KERNEL_TYPE_L2_NORM, - GGML_METAL_KERNEL_TYPE_GROUP_NORM, - GGML_METAL_KERNEL_TYPE_NORM, - GGML_METAL_KERNEL_TYPE_SSM_CONV_F32, - GGML_METAL_KERNEL_TYPE_SSM_SCAN_F32, - GGML_METAL_KERNEL_TYPE_SSM_SCAN_F32_GROUP, - GGML_METAL_KERNEL_TYPE_RWKV_WKV6_F32, - GGML_METAL_KERNEL_TYPE_RWKV_WKV7_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_F32_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_F32_F32_C4, - GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_C4, - GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_1ROW, - GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_L4, - GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F16, - GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_C4, - GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_1ROW, - GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_L4, - GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_BF16, - GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_0_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_1_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_0_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_1_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_Q8_0_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_MXFP4_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F32_F32_R1_2, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F32_F32_R1_3, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F32_F32_R1_4, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F32_F32_R1_5, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_2, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_3, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_4, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_5, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_2, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_3, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_4, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_5, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_2, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_3, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_4, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_5, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_2, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_3, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_4, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_5, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_2, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_3, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_4, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_5, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_2, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_3, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_4, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_5, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_MXFP4_F32_R1_2, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_MXFP4_F32_R1_3, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_MXFP4_F32_R1_4, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_MXFP4_F32_R1_5, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_2, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_3, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_4, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_5, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_2, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_3, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_4, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_5, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_2, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_3, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_4, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_5, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_2, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_3, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_4, - GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_5, - GGML_METAL_KERNEL_TYPE_MUL_MV_Q2_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_Q3_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_Q6_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XXS_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XS_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_IQ3_XXS_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_IQ3_S_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_S_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_IQ1_S_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_IQ1_M_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_IQ4_NL_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_IQ4_XS_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F32_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F16_F32, - //GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F16_F32_1ROW, - //GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F16_F32_L4, - //GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F16_F16, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_BF16_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_0_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_1_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q5_0_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q5_1_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q8_0_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_MXFP4_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q2_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q3_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q5_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q6_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_XXS_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_XS_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ3_XXS_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ3_S_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_S_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ1_S_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ1_M_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ4_NL_F32, - GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ4_XS_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_F32_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_F16_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_BF16_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_0_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_1_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_0_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_1_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_Q8_0_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_MXFP4_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_Q2_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_Q3_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_Q6_K_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XXS_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XS_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_XXS_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_S_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_S_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_IQ1_S_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_IQ1_M_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_IQ4_NL_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_IQ4_XS_F32, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_1, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_2, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_4, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_6, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_8, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_10, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_F32_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_F16_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_BF16_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_0_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_1_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_0_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_1_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q8_0_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MXFP4_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q2_K_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q3_K_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_K_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_K_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q6_K_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_XXS_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_XS_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ3_XXS_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ3_S_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_S_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ1_S_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ1_M_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ4_NL_F16, - GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ4_XS_F16, - GGML_METAL_KERNEL_TYPE_ROPE_NORM_F32, - GGML_METAL_KERNEL_TYPE_ROPE_NORM_F16, - GGML_METAL_KERNEL_TYPE_ROPE_MULTI_F32, - GGML_METAL_KERNEL_TYPE_ROPE_MULTI_F16, - GGML_METAL_KERNEL_TYPE_ROPE_VISION_F32, - GGML_METAL_KERNEL_TYPE_ROPE_VISION_F16, - GGML_METAL_KERNEL_TYPE_ROPE_NEOX_F32, - GGML_METAL_KERNEL_TYPE_ROPE_NEOX_F16, - GGML_METAL_KERNEL_TYPE_IM2COL_F16, - GGML_METAL_KERNEL_TYPE_IM2COL_F32, - GGML_METAL_KERNEL_TYPE_IM2COL_EXT_F16, - GGML_METAL_KERNEL_TYPE_IM2COL_EXT_F32, - GGML_METAL_KERNEL_TYPE_CONV_TRANSPOSE_1D_F32_F32, - GGML_METAL_KERNEL_TYPE_CONV_TRANSPOSE_1D_F16_F32, - GGML_METAL_KERNEL_TYPE_UPSCALE_F32, - GGML_METAL_KERNEL_TYPE_PAD_F32, - GGML_METAL_KERNEL_TYPE_PAD_REFLECT_1D_F32, - GGML_METAL_KERNEL_TYPE_ARANGE_F32, - GGML_METAL_KERNEL_TYPE_TIMESTEP_EMBEDDING_F32, - GGML_METAL_KERNEL_TYPE_ARGSORT_F32_I32_ASC, - GGML_METAL_KERNEL_TYPE_ARGSORT_F32_I32_DESC, - GGML_METAL_KERNEL_TYPE_LEAKY_RELU_F32, - GGML_METAL_KERNEL_TYPE_CPY_F32_F32, - GGML_METAL_KERNEL_TYPE_CPY_F32_F16, - GGML_METAL_KERNEL_TYPE_CPY_F32_BF16, - GGML_METAL_KERNEL_TYPE_CPY_F16_F16, - GGML_METAL_KERNEL_TYPE_CPY_F16_F32, - GGML_METAL_KERNEL_TYPE_CPY_BF16_F32, - GGML_METAL_KERNEL_TYPE_CPY_BF16_BF16, - GGML_METAL_KERNEL_TYPE_CPY_F32_I32, - GGML_METAL_KERNEL_TYPE_CPY_I32_F32, - GGML_METAL_KERNEL_TYPE_CPY_F32_Q8_0, - GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_0, - GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_1, - GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_0, - GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_1, - GGML_METAL_KERNEL_TYPE_CPY_F32_IQ4_NL, - GGML_METAL_KERNEL_TYPE_CPY_Q4_0_F32, - GGML_METAL_KERNEL_TYPE_CPY_Q4_0_F16, - GGML_METAL_KERNEL_TYPE_CPY_Q4_1_F32, - GGML_METAL_KERNEL_TYPE_CPY_Q4_1_F16, - GGML_METAL_KERNEL_TYPE_CPY_Q5_0_F32, - GGML_METAL_KERNEL_TYPE_CPY_Q5_0_F16, - GGML_METAL_KERNEL_TYPE_CPY_Q5_1_F32, - GGML_METAL_KERNEL_TYPE_CPY_Q5_1_F16, - GGML_METAL_KERNEL_TYPE_CPY_Q8_0_F32, - GGML_METAL_KERNEL_TYPE_CPY_Q8_0_F16, - GGML_METAL_KERNEL_TYPE_CONCAT, - GGML_METAL_KERNEL_TYPE_SQR, - GGML_METAL_KERNEL_TYPE_SQRT, - GGML_METAL_KERNEL_TYPE_SIN, - GGML_METAL_KERNEL_TYPE_COS, - GGML_METAL_KERNEL_TYPE_NEG, - GGML_METAL_KERNEL_TYPE_REGLU, - GGML_METAL_KERNEL_TYPE_GEGLU, - GGML_METAL_KERNEL_TYPE_SWIGLU, - GGML_METAL_KERNEL_TYPE_SWIGLU_OAI, - GGML_METAL_KERNEL_TYPE_GEGLU_ERF, - GGML_METAL_KERNEL_TYPE_GEGLU_QUICK, - GGML_METAL_KERNEL_TYPE_SUM_ROWS, - GGML_METAL_KERNEL_TYPE_MEAN, - GGML_METAL_KERNEL_TYPE_POOL_2D_AVG_F32, - GGML_METAL_KERNEL_TYPE_POOL_2D_MAX_F32, - GGML_METAL_KERNEL_TYPE_ARGMAX, - - GGML_METAL_KERNEL_TYPE_COUNT -}; - -struct ggml_metal_command_buffer { - id obj; - - // used to enable concurrent execution of ops in the command buffers - struct ggml_mem_ranges * mem_ranges; -}; - -struct ggml_backend_metal_context { - id device; - id queue; // currently a pointer to the device queue, but might become separate queue [TAG_QUEUE_PER_BACKEND] - - dispatch_queue_t d_queue; - - // the set of pre-compiled kernels for this context - struct ggml_metal_kernel kernels[GGML_METAL_KERNEL_TYPE_COUNT]; - - // additional, inference-time compiled kernels - NSMutableDictionary * kernels_ext; - - // capture state - bool capture_next_compute; - bool capture_started; - - id capture_scope; - - // command buffer state - int n_cb; // number of extra threads used to submit the command buffers - int n_nodes_0; // number of nodes submitted by the main thread - int n_nodes_1; // remaining number of nodes submitted by the n_cb threads - int n_nodes_per_cb; - - struct ggml_cgraph * gf; - - // the callback given to the thread pool - void (^encode_async)(size_t ith); - - // n_cb command buffers + 1 used by the main thread - struct ggml_metal_command_buffer cmd_bufs[GGML_METAL_MAX_COMMAND_BUFFERS + 1]; - - // extra command buffers for things like getting, setting and copying tensors - NSMutableArray * cmd_bufs_ext; - - // the last command buffer queued into the Metal queue with operations relevant to the current Metal backend - id cmd_buf_last; - - // abort ggml_metal_graph_compute if callback returns true - ggml_abort_callback abort_callback; - void * abort_callback_data; -}; - -// MSL code -// TODO: move the contents here when ready -// for now it is easier to work in a separate file -// static NSString * const msl_library_source = @"see metal.metal"; - -#if !GGML_METAL_EMBED_LIBRARY -// Here to assist with NSBundle Path Hack -@interface GGMLMetalClass : NSObject -@end -@implementation GGMLMetalClass -@end -#endif - -static void * ggml_metal_host_malloc(size_t n) { - void * data = NULL; - -#if TARGET_OS_OSX - kern_return_t err = vm_allocate((vm_map_t) mach_task_self(), (void *) &data, n, VM_FLAGS_ANYWHERE); - if (err != KERN_SUCCESS) { - GGML_LOG_ERROR("%s: error: vm_allocate failed\n", __func__); - return NULL; - } -#else - const int result = posix_memalign((void **) &data, sysconf(_SC_PAGESIZE), n); - if (result != 0) { - GGML_LOG_ERROR("%s: error: posix_memalign failed\n", __func__); - return NULL; - } -#endif - - return data; -} - -// load library -// -// - first check if the library is embedded -// - then check if the library is in the bundle -// - if not found, load the source and compile it -// - if that fails, return NULL -static id ggml_metal_load_library(id device, bool use_bfloat) { - const int64_t t_start = ggml_time_us(); - - id metal_library = nil; - NSError * error = nil; - NSString * src = nil; - -#if GGML_METAL_EMBED_LIBRARY - GGML_LOG_INFO("%s: using embedded metal library\n", __func__); - - extern const char ggml_metallib_start[]; - extern const char ggml_metallib_end[]; - - src = [[NSString alloc] initWithBytes:ggml_metallib_start length:(ggml_metallib_end-ggml_metallib_start) encoding:NSUTF8StringEncoding]; - -#else - -#ifdef SWIFT_PACKAGE - NSBundle * bundle = SWIFTPM_MODULE_BUNDLE; -#else - NSBundle * bundle = [NSBundle bundleForClass:[GGMLMetalClass class]]; -#endif - - NSString * path_lib = [bundle pathForResource:@"default" ofType:@"metallib"]; - if (path_lib == nil) { - // Try to find the resource in the directory where the current binary located. - NSString * current_binary = [[NSProcessInfo processInfo] arguments][0]; - NSString * bin_dir = [current_binary stringByDeletingLastPathComponent]; - NSString * default_metallib_path = [NSString pathWithComponents:@[bin_dir, @"default.metallib"]]; - if ([[NSFileManager defaultManager] isReadableFileAtPath:default_metallib_path]) { - GGML_LOG_INFO("%s: found '%s'\n", __func__, [default_metallib_path UTF8String]); - NSDictionary * atts = [[NSFileManager defaultManager] attributesOfItemAtPath:default_metallib_path error:&error]; - if (atts && atts[NSFileType] == NSFileTypeSymbolicLink) { - // Optionally, if this is a symlink, try to resolve it. - default_metallib_path = [[NSFileManager defaultManager] destinationOfSymbolicLinkAtPath:default_metallib_path error:&error]; - if (default_metallib_path && [default_metallib_path length] > 0 && ![[default_metallib_path substringToIndex:1] isEqualToString:@"/"]) { - // It is a relative path, adding the binary directory as directory prefix. - default_metallib_path = [NSString pathWithComponents:@[bin_dir, default_metallib_path]]; - } - if (!default_metallib_path || ![[NSFileManager defaultManager] isReadableFileAtPath:default_metallib_path]) { - // Link to the resource could not be resolved. - default_metallib_path = nil; - } else { - GGML_LOG_INFO("%s: symlink resolved '%s'\n", __func__, [default_metallib_path UTF8String]); - } - } - } else { - // The resource couldn't be found in the binary's directory. - default_metallib_path = nil; - } - path_lib = default_metallib_path; - } - - if (path_lib != nil) { - // pre-compiled library found - NSURL * libURL = [NSURL fileURLWithPath:path_lib]; - GGML_LOG_INFO("%s: loading '%s'\n", __func__, [path_lib UTF8String]); - - metal_library = [device newLibraryWithURL:libURL error:&error]; - if (error) { - GGML_LOG_ERROR("%s: error: %s\n", __func__, [[error description] UTF8String]); - return NULL; - } - } else { - GGML_LOG_INFO("%s: default.metallib not found, loading from source\n", __func__); - - NSString * path_source; - NSString * path_resource = [[NSProcessInfo processInfo].environment objectForKey:@"GGML_METAL_PATH_RESOURCES"]; - - GGML_LOG_INFO("%s: GGML_METAL_PATH_RESOURCES = %s\n", __func__, path_resource ? [path_resource UTF8String] : "nil"); - - if (path_resource) { - path_source = [path_resource stringByAppendingPathComponent:@"ggml-metal.metal"]; - } else { - path_source = [bundle pathForResource:@"ggml-metal" ofType:@"metal"]; - } - - if (path_source == nil) { - GGML_LOG_WARN("%s: error: could not use bundle path to find ggml-metal.metal, falling back to trying cwd\n", __func__); - path_source = @"ggml-metal.metal"; - } - - GGML_LOG_INFO("%s: loading '%s'\n", __func__, [path_source UTF8String]); - - src = [NSString stringWithContentsOfFile:path_source encoding:NSUTF8StringEncoding error:&error]; - if (error) { - GGML_LOG_ERROR("%s: error: %s\n", __func__, [[error description] UTF8String]); - return NULL; - } - } -#endif - - if (!metal_library) { - @autoreleasepool { - // dictionary of preprocessor macros - NSMutableDictionary * prep = [NSMutableDictionary dictionary]; - - if (use_bfloat) { - [prep setObject:@"1" forKey:@"GGML_METAL_USE_BF16"]; - } - -#if GGML_METAL_EMBED_LIBRARY - [prep setObject:@"1" forKey:@"GGML_METAL_EMBED_LIBRARY"]; -#endif - - MTLCompileOptions * options = [MTLCompileOptions new]; - options.preprocessorMacros = prep; - - //[options setFastMathEnabled:false]; - - metal_library = [device newLibraryWithSource:src options:options error:&error]; - if (error) { - GGML_LOG_ERROR("%s: error: %s\n", __func__, [[error description] UTF8String]); - return NULL; - } - -#if !__has_feature(objc_arc) - [options release]; -#endif - } - } - -#if GGML_METAL_EMBED_LIBRARY - [src release]; -#endif // GGML_METAL_EMBED_LIBRARY - - GGML_LOG_INFO("%s: loaded in %.3f sec\n", __func__, (ggml_time_us() - t_start) / 1e6); - - return metal_library; -} - -static struct ggml_backend_metal_context * ggml_metal_init(ggml_backend_dev_t dev) { - GGML_LOG_INFO("%s: allocating\n", __func__); - -#if TARGET_OS_OSX && !GGML_METAL_NDEBUG - // Show all the Metal device instances in the system - NSArray * devices = MTLCopyAllDevices(); - for (id device in devices) { - GGML_LOG_INFO("%s: found device: %s\n", __func__, [[device name] UTF8String]); - } - [devices release]; // since it was created by a *Copy* C method -#endif - - // init context - struct ggml_backend_metal_context * ctx = calloc(1, sizeof(struct ggml_backend_metal_context)); - struct ggml_backend_metal_device_context * ctx_dev = dev->context; - - id device = ctx_dev->mtl_device; - - GGML_LOG_INFO("%s: picking default device: %s\n", __func__, [[device name] UTF8String]); - - ctx->device = device; - - // TODO: question - would it be better to have one queue for the backend and one queue for the device? - // the graph encoders and async ops would use the backend queue while the sync ops would use the device queue? - //ctx->queue = [device newCommandQueue]; [TAG_QUEUE_PER_BACKEND] - ctx->queue = ctx_dev->mtl_queue; - if (ctx->queue == nil) { - GGML_LOG_ERROR("%s: error: failed to create command queue\n", __func__); - return NULL; - } - - ctx->d_queue = dispatch_queue_create("ggml-metal", DISPATCH_QUEUE_CONCURRENT); - - // load library - { - [ctx_dev->mtl_lock lock]; - - if (ctx_dev->mtl_library == nil) { - ctx_dev->mtl_library = ggml_metal_load_library(device, ctx_dev->use_bfloat); - } - - [ctx_dev->mtl_lock unlock]; - } - - id metal_library = ctx_dev->mtl_library; - if (metal_library == nil) { - GGML_LOG_ERROR("%s: error: metal library is nil\n", __func__); - return NULL; - } - - // print MTL GPU family: - GGML_LOG_INFO("%s: GPU name: %s\n", __func__, [[device name] UTF8String]); - - // determine max supported GPU family - // https://developer.apple.com/metal/Metal-Shading-Language-Specification.pdf - // https://developer.apple.com/metal/Metal-Feature-Set-Tables.pdf - { - for (int i = MTLGPUFamilyApple1 + 20; i >= MTLGPUFamilyApple1; --i) { - if ([device supportsFamily:i]) { - GGML_LOG_INFO("%s: GPU family: MTLGPUFamilyApple%d (%d)\n", __func__, i - (int) MTLGPUFamilyApple1 + 1, i); - break; - } - } - - for (int i = MTLGPUFamilyCommon1 + 5; i >= MTLGPUFamilyCommon1; --i) { - if ([device supportsFamily:i]) { - GGML_LOG_INFO("%s: GPU family: MTLGPUFamilyCommon%d (%d)\n", __func__, i - (int) MTLGPUFamilyCommon1 + 1, i); - break; - } - } - - for (int i = MTLGPUFamilyMetal3_GGML + 5; i >= MTLGPUFamilyMetal3_GGML; --i) { - if ([device supportsFamily:i]) { - GGML_LOG_INFO("%s: GPU family: MTLGPUFamilyMetal%d (%d)\n", __func__, i - (int) MTLGPUFamilyMetal3_GGML + 3, i); - break; - } - } - } - - GGML_LOG_INFO("%s: simdgroup reduction = %s\n", __func__, ctx_dev->has_simdgroup_reduction ? "true" : "false"); - GGML_LOG_INFO("%s: simdgroup matrix mul. = %s\n", __func__, ctx_dev->has_simdgroup_mm ? "true" : "false"); - GGML_LOG_INFO("%s: has residency sets = %s\n", __func__, ctx_dev->has_residency_sets ? "true" : "false"); - GGML_LOG_INFO("%s: has bfloat = %s\n", __func__, ctx_dev->has_bfloat ? "true" : "false"); - GGML_LOG_INFO("%s: use bfloat = %s\n", __func__, ctx_dev->use_bfloat ? "true" : "false"); - GGML_LOG_INFO("%s: use fusion = %s\n", __func__, ctx_dev->use_fusion ? "true" : "false"); - GGML_LOG_INFO("%s: use concurrency = %s\n", __func__, ctx_dev->use_concurrency ? "true" : "false"); - GGML_LOG_INFO("%s: use shared buffers = %s\n", __func__, ctx_dev->use_shared_buffers ? "true" : "false"); - GGML_LOG_INFO("%s: use graph optimize = %s\n", __func__, ctx_dev->use_graph_optimize ? "true" : "false"); - GGML_LOG_INFO("%s: hasUnifiedMemory = %s\n", __func__, ctx_dev->mtl_device.hasUnifiedMemory ? "true" : "false"); - - ctx->capture_next_compute = false; - ctx->capture_started = false; - ctx->capture_scope = nil; - - ctx->gf = nil; - ctx->encode_async = nil; - for (int i = 0; i < GGML_METAL_MAX_COMMAND_BUFFERS; ++i) { - ctx->cmd_bufs[i].obj = nil; - - if (ctx_dev->use_concurrency) { - ctx->cmd_bufs[i].mem_ranges = ggml_mem_ranges_init(ctx_dev->debug_graph); - } - } - - ctx->cmd_bufs_ext = [[NSMutableArray alloc] init]; - - ctx->cmd_buf_last = nil; - -#if TARGET_OS_OSX || (TARGET_OS_IOS && __clang_major__ >= 15) - if (@available(macOS 10.12, iOS 16.0, *)) { - GGML_LOG_INFO("%s: recommendedMaxWorkingSetSize = %8.2f MB\n", __func__, device.recommendedMaxWorkingSetSize / 1e6); - } -#endif - - // load kernels - { - NSError * error = nil; - - for (int i = 0; i < GGML_METAL_KERNEL_TYPE_COUNT; ++i) { - ctx->kernels[i].pipeline = nil; - } - -#define GGML_METAL_ADD_KERNEL(e, name, supported) \ - if (supported) { \ - struct ggml_metal_kernel * kernel = &ctx->kernels[e]; \ - id metal_function = [metal_library newFunctionWithName:@"kernel_"#name]; \ - kernel->pipeline = [device newComputePipelineStateWithFunction:metal_function error:&error]; \ - GGML_LOG_DEBUG("%s: loaded %-40s %16p | th_max = %4d | th_width = %4d\n", __func__, "kernel_"#name, (void *) kernel->pipeline, \ - (int) kernel->pipeline.maxTotalThreadsPerThreadgroup, \ - (int) kernel->pipeline.threadExecutionWidth); \ - [metal_function release]; \ - if (error) { \ - GGML_LOG_ERROR("%s: error: load pipeline error: %s\n", __func__, [[error description] UTF8String]); \ - return NULL; \ - } \ - } else { \ - GGML_LOG_WARN("%s: skipping %-40s (not supported)\n", __func__, "kernel_"#name); \ - } - - const bool has_simdgroup_mm = ctx_dev->has_simdgroup_mm; - const bool has_simdgroup_reduction = ctx_dev->has_simdgroup_reduction; - const bool use_bfloat = ctx_dev->use_bfloat; - - // simd_sum and simd_max requires MTLGPUFamilyApple7 - - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ADD_ID, add_id, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_REPEAT_F32, repeat_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_REPEAT_F16, repeat_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_REPEAT_I32, repeat_i32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_REPEAT_I16, repeat_i16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SCALE, scale, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SCALE_4, scale_4, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CLAMP, clamp, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_TANH, tanh, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_RELU, relu, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SIGMOID, sigmoid, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GELU, gelu, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GELU_4, gelu_4, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GELU_ERF, gelu_erf, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GELU_ERF_4, gelu_erf_4, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GELU_QUICK, gelu_quick, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GELU_QUICK_4, gelu_quick_4, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SILU, silu, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SILU_4, silu_4, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ELU, elu, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ABS, abs, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SGN, sgn, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_STEP, step, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_HARDSWISH, hardswish, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_HARDSIGMOID, hardsigmoid, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_EXP, exp, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SOFT_MAX_F16, soft_max_f16, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SOFT_MAX_F16_4, soft_max_f16_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SOFT_MAX_F32, soft_max_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SOFT_MAX_F32_4, soft_max_f32_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_DIAG_MASK_INF, diag_mask_inf, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_DIAG_MASK_INF_8, diag_mask_inf_8, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_F32, get_rows_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_F16, get_rows_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_BF16, get_rows_bf16, use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_0, get_rows_q4_0, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_1, get_rows_q4_1, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_0, get_rows_q5_0, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_1, get_rows_q5_1, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q8_0, get_rows_q8_0, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_MXFP4, get_rows_mxfp4, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q2_K, get_rows_q2_K, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q3_K, get_rows_q3_K, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_K, get_rows_q4_K, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_K, get_rows_q5_K, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_Q6_K, get_rows_q6_K, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_XXS, get_rows_iq2_xxs, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_XS, get_rows_iq2_xs, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ3_XXS, get_rows_iq3_xxs, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ3_S, get_rows_iq3_s, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_S, get_rows_iq2_s, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ1_S, get_rows_iq1_s, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ1_M, get_rows_iq1_m, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ4_NL, get_rows_iq4_nl, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ4_XS, get_rows_iq4_xs, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GET_ROWS_I32, get_rows_i32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SET_ROWS_F32, set_rows_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SET_ROWS_F16, set_rows_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SET_ROWS_BF16, set_rows_bf16, use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SET_ROWS_Q8_0, set_rows_q8_0, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SET_ROWS_Q4_0, set_rows_q4_0, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SET_ROWS_Q4_1, set_rows_q4_1, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SET_ROWS_Q5_0, set_rows_q5_0, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SET_ROWS_Q5_1, set_rows_q5_1, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SET_ROWS_IQ4_NL, set_rows_iq4_nl, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_L2_NORM, l2_norm, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GROUP_NORM, group_norm, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_NORM, norm, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SSM_CONV_F32, ssm_conv_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SSM_SCAN_F32, ssm_scan_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SSM_SCAN_F32_GROUP, ssm_scan_f32_group, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_RWKV_WKV6_F32, rwkv_wkv6_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_RWKV_WKV7_F32, rwkv_wkv7_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_F32_F32, mul_mv_f32_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_F32_F32_C4, mul_mv_f32_f32_c4, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32, mul_mv_bf16_f32, has_simdgroup_reduction && use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_C4, mul_mv_bf16_f32_c4, use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_1ROW, mul_mv_bf16_f32_1row, has_simdgroup_reduction && use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_L4, mul_mv_bf16_f32_l4, has_simdgroup_reduction && use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_BF16, mul_mv_bf16_bf16, has_simdgroup_reduction && use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32, mul_mv_f16_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_C4, mul_mv_f16_f32_c4, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_1ROW, mul_mv_f16_f32_1row, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_L4, mul_mv_f16_f32_l4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F16, mul_mv_f16_f16, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_0_F32, mul_mv_q4_0_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_1_F32, mul_mv_q4_1_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_0_F32, mul_mv_q5_0_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_1_F32, mul_mv_q5_1_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q8_0_F32, mul_mv_q8_0_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_MXFP4_F32, mul_mv_mxfp4_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F32_F32_R1_2, mul_mv_ext_f32_f32_r1_2, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F32_F32_R1_3, mul_mv_ext_f32_f32_r1_3, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F32_F32_R1_4, mul_mv_ext_f32_f32_r1_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F32_F32_R1_5, mul_mv_ext_f32_f32_r1_5, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_2, mul_mv_ext_f16_f32_r1_2, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_3, mul_mv_ext_f16_f32_r1_3, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_4, mul_mv_ext_f16_f32_r1_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_5, mul_mv_ext_f16_f32_r1_5, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_2, mul_mv_ext_q4_0_f32_r1_2, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_3, mul_mv_ext_q4_0_f32_r1_3, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_4, mul_mv_ext_q4_0_f32_r1_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_5, mul_mv_ext_q4_0_f32_r1_5, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_2, mul_mv_ext_q4_1_f32_r1_2, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_3, mul_mv_ext_q4_1_f32_r1_3, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_4, mul_mv_ext_q4_1_f32_r1_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_5, mul_mv_ext_q4_1_f32_r1_5, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_2, mul_mv_ext_q5_0_f32_r1_2, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_3, mul_mv_ext_q5_0_f32_r1_3, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_4, mul_mv_ext_q5_0_f32_r1_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_5, mul_mv_ext_q5_0_f32_r1_5, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_2, mul_mv_ext_q5_1_f32_r1_2, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_3, mul_mv_ext_q5_1_f32_r1_3, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_4, mul_mv_ext_q5_1_f32_r1_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_5, mul_mv_ext_q5_1_f32_r1_5, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_2, mul_mv_ext_q8_0_f32_r1_2, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_3, mul_mv_ext_q8_0_f32_r1_3, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_4, mul_mv_ext_q8_0_f32_r1_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_5, mul_mv_ext_q8_0_f32_r1_5, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_MXFP4_F32_R1_2, mul_mv_ext_mxfp4_f32_r1_2, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_MXFP4_F32_R1_3, mul_mv_ext_mxfp4_f32_r1_3, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_MXFP4_F32_R1_4, mul_mv_ext_mxfp4_f32_r1_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_MXFP4_F32_R1_5, mul_mv_ext_mxfp4_f32_r1_5, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_2, mul_mv_ext_q4_K_f32_r1_2, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_3, mul_mv_ext_q4_K_f32_r1_3, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_4, mul_mv_ext_q4_K_f32_r1_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_5, mul_mv_ext_q4_K_f32_r1_5, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_2, mul_mv_ext_q5_K_f32_r1_2, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_3, mul_mv_ext_q5_K_f32_r1_3, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_4, mul_mv_ext_q5_K_f32_r1_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_5, mul_mv_ext_q5_K_f32_r1_5, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_2, mul_mv_ext_q6_K_f32_r1_2, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_3, mul_mv_ext_q6_K_f32_r1_3, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_4, mul_mv_ext_q6_K_f32_r1_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_5, mul_mv_ext_q6_K_f32_r1_5, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_2, mul_mv_ext_iq4_nl_f32_r1_2, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_3, mul_mv_ext_iq4_nl_f32_r1_3, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_4, mul_mv_ext_iq4_nl_f32_r1_4, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_5, mul_mv_ext_iq4_nl_f32_r1_5, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q2_K_F32, mul_mv_q2_K_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q3_K_F32, mul_mv_q3_K_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_K_F32, mul_mv_q4_K_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_K_F32, mul_mv_q5_K_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_Q6_K_F32, mul_mv_q6_K_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XXS_F32, mul_mv_iq2_xxs_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XS_F32, mul_mv_iq2_xs_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_IQ3_XXS_F32, mul_mv_iq3_xxs_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_IQ3_S_F32, mul_mv_iq3_s_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_S_F32, mul_mv_iq2_s_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_IQ1_S_F32, mul_mv_iq1_s_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_IQ1_M_F32, mul_mv_iq1_m_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_IQ4_NL_F32, mul_mv_iq4_nl_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_IQ4_XS_F32, mul_mv_iq4_xs_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F32_F32, mul_mv_id_f32_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F16_F32, mul_mv_id_f16_f32, has_simdgroup_reduction); - //GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F16_F32_1ROW, mul_mv_id_f16_f32_1row, has_simdgroup_reduction); - //GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F16_F32_L4, mul_mv_id_f16_f32_l4, has_simdgroup_reduction); - //GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F16_F16, mul_mv_id_f16_f16, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_BF16_F32, mul_mv_id_bf16_f32, has_simdgroup_reduction && use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_0_F32, mul_mv_id_q4_0_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_1_F32, mul_mv_id_q4_1_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q5_0_F32, mul_mv_id_q5_0_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q5_1_F32, mul_mv_id_q5_1_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q8_0_F32, mul_mv_id_q8_0_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_MXFP4_F32, mul_mv_id_mxfp4_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q2_K_F32, mul_mv_id_q2_K_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q3_K_F32, mul_mv_id_q3_K_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_K_F32, mul_mv_id_q4_K_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q5_K_F32, mul_mv_id_q5_K_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q6_K_F32, mul_mv_id_q6_K_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_XXS_F32, mul_mv_id_iq2_xxs_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_XS_F32, mul_mv_id_iq2_xs_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ3_XXS_F32, mul_mv_id_iq3_xxs_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ3_S_F32, mul_mv_id_iq3_s_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_S_F32, mul_mv_id_iq2_s_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ1_S_F32, mul_mv_id_iq1_s_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ1_M_F32, mul_mv_id_iq1_m_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ4_NL_F32, mul_mv_id_iq4_nl_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ4_XS_F32, mul_mv_id_iq4_xs_f32, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_F32_F32, mul_mm_f32_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_F16_F32, mul_mm_f16_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_BF16_F32, mul_mm_bf16_f32, has_simdgroup_mm && use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_0_F32, mul_mm_q4_0_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_1_F32, mul_mm_q4_1_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_0_F32, mul_mm_q5_0_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_1_F32, mul_mm_q5_1_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q8_0_F32, mul_mm_q8_0_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_MXFP4_F32, mul_mm_mxfp4_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q2_K_F32, mul_mm_q2_K_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q3_K_F32, mul_mm_q3_K_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_K_F32, mul_mm_q4_K_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_K_F32, mul_mm_q5_K_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_Q6_K_F32, mul_mm_q6_K_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XXS_F32, mul_mm_iq2_xxs_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XS_F32, mul_mm_iq2_xs_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_XXS_F32, mul_mm_iq3_xxs_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_S_F32, mul_mm_iq3_s_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_S_F32, mul_mm_iq2_s_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ1_S_F32, mul_mm_iq1_s_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ1_M_F32, mul_mm_iq1_m_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ4_NL_F32, mul_mm_iq4_nl_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_IQ4_XS_F32, mul_mm_iq4_xs_f32, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_1, mul_mm_id_map0_f16_ne20_1, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_2, mul_mm_id_map0_f16_ne20_2, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_4, mul_mm_id_map0_f16_ne20_4, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_6, mul_mm_id_map0_f16_ne20_6, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_8, mul_mm_id_map0_f16_ne20_8, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_10, mul_mm_id_map0_f16_ne20_10, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_16, mul_mm_id_map0_f16_ne20_16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_F32_F16, mul_mm_id_f32_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_F16_F16, mul_mm_id_f16_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_BF16_F16, mul_mm_id_bf16_f16, has_simdgroup_mm && use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_0_F16, mul_mm_id_q4_0_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_1_F16, mul_mm_id_q4_1_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_0_F16, mul_mm_id_q5_0_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_1_F16, mul_mm_id_q5_1_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q8_0_F16, mul_mm_id_q8_0_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MXFP4_F16, mul_mm_id_mxfp4_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q2_K_F16, mul_mm_id_q2_K_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q3_K_F16, mul_mm_id_q3_K_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_K_F16, mul_mm_id_q4_K_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_K_F16, mul_mm_id_q5_K_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q6_K_F16, mul_mm_id_q6_K_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_XXS_F16, mul_mm_id_iq2_xxs_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_XS_F16, mul_mm_id_iq2_xs_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ3_XXS_F16, mul_mm_id_iq3_xxs_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ3_S_F16, mul_mm_id_iq3_s_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_S_F16, mul_mm_id_iq2_s_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ1_S_F16, mul_mm_id_iq1_s_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ1_M_F16, mul_mm_id_iq1_m_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ4_NL_F16, mul_mm_id_iq4_nl_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ4_XS_F16, mul_mm_id_iq4_xs_f16, has_simdgroup_mm); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ROPE_NORM_F32, rope_norm_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ROPE_NORM_F16, rope_norm_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ROPE_MULTI_F32, rope_multi_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ROPE_MULTI_F16, rope_multi_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ROPE_VISION_F32, rope_vision_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ROPE_VISION_F16, rope_vision_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ROPE_NEOX_F32, rope_neox_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ROPE_NEOX_F16, rope_neox_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_IM2COL_F16, im2col_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_IM2COL_F32, im2col_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_IM2COL_EXT_F16, im2col_ext_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_IM2COL_EXT_F32, im2col_ext_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CONV_TRANSPOSE_1D_F32_F32, conv_transpose_1d_f32_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CONV_TRANSPOSE_1D_F16_F32, conv_transpose_1d_f16_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_UPSCALE_F32, upscale_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_PAD_F32, pad_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_PAD_REFLECT_1D_F32, pad_reflect_1d_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_TIMESTEP_EMBEDDING_F32, timestep_embedding_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ARANGE_F32, arange_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ARGSORT_F32_I32_ASC, argsort_f32_i32_asc, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ARGSORT_F32_I32_DESC, argsort_f32_i32_desc, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_LEAKY_RELU_F32, leaky_relu_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_F32, cpy_f32_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_F16, cpy_f32_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_BF16, cpy_f32_bf16, use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F16_F32, cpy_f16_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F16_F16, cpy_f16_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_BF16_F32, cpy_bf16_f32, use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_BF16_BF16, cpy_bf16_bf16, use_bfloat); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_I32, cpy_f32_i32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_I32_F32, cpy_i32_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q8_0, cpy_f32_q8_0, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_0, cpy_f32_q4_0, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_1, cpy_f32_q4_1, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_0, cpy_f32_q5_0, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_1, cpy_f32_q5_1, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_F32_IQ4_NL, cpy_f32_iq4_nl, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_Q4_0_F32, cpy_q4_0_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_Q4_0_F16, cpy_q4_0_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_Q4_1_F32, cpy_q4_1_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_Q4_1_F16, cpy_q4_1_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_Q5_0_F32, cpy_q5_0_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_Q5_0_F16, cpy_q5_0_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_Q5_1_F32, cpy_q5_1_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_Q5_1_F16, cpy_q5_1_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_Q8_0_F32, cpy_q8_0_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CPY_Q8_0_F16, cpy_q8_0_f16, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_CONCAT, concat, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SQR, sqr, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SQRT, sqrt, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SIN, sin, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_COS, cos, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_NEG, neg, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_REGLU, reglu, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GEGLU, geglu, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SWIGLU, swiglu, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SWIGLU_OAI, swiglu_oai, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GEGLU_ERF, geglu_erf, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_GEGLU_QUICK, geglu_quick, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_SUM_ROWS, sum_rows, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_MEAN, mean, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_ARGMAX, argmax, has_simdgroup_reduction); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_POOL_2D_AVG_F32, pool_2d_avg_f32, true); - GGML_METAL_ADD_KERNEL(GGML_METAL_KERNEL_TYPE_POOL_2D_MAX_F32, pool_2d_max_f32, true); - } - - ctx->kernels_ext = [[NSMutableDictionary alloc] init]; - - return ctx; -} - -static id ggml_metal_get_kernel(struct ggml_backend_metal_context * ctx, const char * name) { - NSString * key = [NSString stringWithUTF8String:name]; - - ggml_metal_kernel_wrapper * obj = [ctx->kernels_ext objectForKey:key]; - if (obj) { - return obj.kernel.pipeline; - } - - return nil; -} - -static id ggml_metal_compile_kernel(ggml_backend_t backend, const char * base, const char * name, MTLFunctionConstantValues * cv) { - struct ggml_backend_metal_context * ctx = backend->context; - struct ggml_backend_metal_device_context * ctx_dev = backend->device->context; - - id res = nil; - - @autoreleasepool { - NSError * error = nil; - - NSString * base_func = [NSString stringWithUTF8String:base]; - - GGML_LOG_DEBUG("%s: compiling kernel: base = '%s', name = '%s'\n", __func__, base, name); - - // TODO: make sure it is thread-safe to compile kernels in parallel - id metal_function = [ctx_dev->mtl_library newFunctionWithName:base_func constantValues:cv error:&error]; - if (!metal_function) { - GGML_LOG_ERROR("%s: error: %s\n", __func__, [[error description] UTF8String]); - - return nil; - } - - struct ggml_metal_kernel kernel = { - /*.pipeline =*/ [ctx_dev->mtl_device newComputePipelineStateWithFunction:metal_function error:&error], - }; - - ggml_metal_kernel_wrapper * obj = [[ggml_metal_kernel_wrapper alloc] init]; - obj.kernel = kernel; - - res = obj.kernel.pipeline; - - NSString * key = [NSString stringWithUTF8String:name]; - [ctx->kernels_ext setObject:obj forKey:key]; - - [metal_function release]; - [obj release]; - - GGML_LOG_DEBUG("%s: loaded %-40s %16p | th_max = %4d | th_width = %4d\n", __func__, name, (void *) kernel.pipeline, - (int) kernel.pipeline.maxTotalThreadsPerThreadgroup, - (int) kernel.pipeline.threadExecutionWidth); - } - - return res; -} - -// tokens per expert -static size_t ggml_metal_mul_mat_id_extra_tpe(const struct ggml_tensor * op) { - assert(op->op == GGML_OP_MUL_MAT_ID); - - const int64_t ne02 = op->src[0]->ne[2]; // n_expert - - return ggml_type_size(GGML_TYPE_I32)*ne02; -} - -// id map [n_tokens, n_expert] -static size_t ggml_metal_mul_mat_id_extra_ids(const struct ggml_tensor * op) { - assert(op->op == GGML_OP_MUL_MAT_ID); - - const int64_t ne02 = op->src[0]->ne[2]; // n_expert - const int64_t ne21 = op->src[2]->ne[1]; // n_token - - return ggml_type_size(GGML_TYPE_I32)*ne02*ne21; -} - -// return true if we should use the FA vector kernel for this op -static bool ggml_metal_flash_attn_ext_use_vec(const struct ggml_tensor * op) { - assert(op->op == GGML_OP_FLASH_ATTN_EXT); - - const int64_t ne00 = op->src[0]->ne[0]; // head size - const int64_t ne01 = op->src[0]->ne[1]; // batch size - - // use vec kernel if the batch size is small and if the head size is supported - return (ne01 < 20) && (ne00 % 32 == 0); -} - -static size_t ggml_metal_flash_attn_ext_extra_tmp(const struct ggml_tensor * op) { - assert(op->op == GGML_OP_FLASH_ATTN_EXT); - - const int64_t nwg = 32; - - const int64_t ne01 = op->src[0]->ne[1]; - const int64_t ne02 = op->src[0]->ne[2]; - const int64_t ne03 = op->src[0]->ne[3]; - const int64_t ne20 = op->src[2]->ne[0]; - - // temp buffer for writing the results from each workgroup - // - ne20: the size of the Value head - // - + 2: the S and M values for each intermediate result - return ggml_type_size(GGML_TYPE_F32)*(ne01*ne02*ne03*nwg*(ne20 + 2)); -} - -static id ggml_metal_get_pipeline_flash_attn_ext( - ggml_backend_t backend, struct ggml_tensor * op, - bool has_mask, - bool has_sinks, - bool has_bias, - bool has_scap, - int32_t nsg) { - struct ggml_backend_metal_context * ctx = backend->context; - - char base[256]; - char name[256]; - - @autoreleasepool { - const int32_t dk = (int32_t) op->src[1]->ne[0]; - const int32_t dv = (int32_t) op->src[2]->ne[0]; - - const int32_t ns10 = op->src[1]->nb[1]/op->src[1]->nb[0]; - const int32_t ns20 = op->src[2]->nb[1]/op->src[2]->nb[0]; - - snprintf(base, 256, "kernel_%s_%s_dk%d_dv%d", - "flash_attn_ext", - ggml_type_name(op->src[1]->type), - dk, - dv); - - snprintf(name, 256, "kernel_%s_%s_dk%d_dv%d_mask=%d_sinks=%d_bias=%d_scap=%d_ns10=%d_ns20=%d_nsg=%d", - "flash_attn_ext", - ggml_type_name(op->src[1]->type), - dk, - dv, - has_mask, - has_sinks, - has_bias, - has_scap, - ns10, - ns20, - nsg); - - id res = ggml_metal_get_kernel(ctx, name); - if (res) { - // kernel found - return res; - } - - MTLFunctionConstantValues * cv = [[MTLFunctionConstantValues alloc] init]; - - [cv setConstantValue:&has_mask type:MTLDataTypeBool atIndex:FC_FLASH_ATTN_EXT + 0]; - [cv setConstantValue:&has_sinks type:MTLDataTypeBool atIndex:FC_FLASH_ATTN_EXT + 1]; - [cv setConstantValue:&has_bias type:MTLDataTypeBool atIndex:FC_FLASH_ATTN_EXT + 2]; - [cv setConstantValue:&has_scap type:MTLDataTypeBool atIndex:FC_FLASH_ATTN_EXT + 3]; - - [cv setConstantValue:&ns10 type:MTLDataTypeInt atIndex:FC_FLASH_ATTN_EXT + 20]; - [cv setConstantValue:&ns20 type:MTLDataTypeInt atIndex:FC_FLASH_ATTN_EXT + 21]; - [cv setConstantValue:&nsg type:MTLDataTypeInt atIndex:FC_FLASH_ATTN_EXT + 22]; - - res = ggml_metal_compile_kernel(backend, base, name, cv); - - [cv release]; - - return res; - } -} - -static id ggml_metal_get_pipeline_flash_attn_ext_vec( - ggml_backend_t backend, struct ggml_tensor * op, - bool has_mask, - bool has_sinks, - bool has_bias, - bool has_scap, - int32_t nsg, - int32_t nwg) { - struct ggml_backend_metal_context * ctx = backend->context; - - char base[256]; - char name[256]; - - @autoreleasepool { - const int32_t dk = (int32_t) op->src[1]->ne[0]; - const int32_t dv = (int32_t) op->src[2]->ne[0]; - - const int32_t ns10 = op->src[1]->nb[1]/op->src[1]->nb[0]; - const int32_t ns20 = op->src[2]->nb[1]/op->src[2]->nb[0]; - - snprintf(base, 256, "kernel_%s_%s_dk%d_dv%d", - "flash_attn_ext_vec", - ggml_type_name(op->src[1]->type), - dk, - dv); - - snprintf(name, 256, "kernel_%s_%s_dk%d_dv%d_mask=%d_sink=%d_bias=%d_softcap=%d_ns10=%d_ns20=%d_nsg=%d_nwg=%d", - "flash_attn_ext_vec", - ggml_type_name(op->src[1]->type), - dk, - dv, - has_mask, - has_sinks, - has_bias, - has_scap, - ns10, - ns20, - nsg, nwg); - - id res = ggml_metal_get_kernel(ctx, name); - if (res) { - // kernel found - return res; - } - - MTLFunctionConstantValues * cv = [[MTLFunctionConstantValues alloc] init]; - - [cv setConstantValue:&has_mask type:MTLDataTypeBool atIndex:FC_FLASH_ATTN_EXT_VEC + 0]; - [cv setConstantValue:&has_sinks type:MTLDataTypeBool atIndex:FC_FLASH_ATTN_EXT_VEC + 1]; - [cv setConstantValue:&has_bias type:MTLDataTypeBool atIndex:FC_FLASH_ATTN_EXT_VEC + 2]; - [cv setConstantValue:&has_scap type:MTLDataTypeBool atIndex:FC_FLASH_ATTN_EXT_VEC + 3]; - - [cv setConstantValue:&ns10 type:MTLDataTypeInt atIndex:FC_FLASH_ATTN_EXT_VEC + 20]; - [cv setConstantValue:&ns20 type:MTLDataTypeInt atIndex:FC_FLASH_ATTN_EXT_VEC + 21]; - [cv setConstantValue:&nsg type:MTLDataTypeInt atIndex:FC_FLASH_ATTN_EXT_VEC + 22]; - [cv setConstantValue:&nwg type:MTLDataTypeInt atIndex:FC_FLASH_ATTN_EXT_VEC + 23]; - - res = ggml_metal_compile_kernel(backend, base, name, cv); - - [cv release]; - - return res; - } -} - -static id ggml_metal_get_pipeline_flash_attn_ext_vec_reduce( - ggml_backend_t backend, struct ggml_tensor * op, - int32_t dv, - int32_t nwg) { - struct ggml_backend_metal_context * ctx = backend->context; - - char base[256]; - char name[256]; - - @autoreleasepool { - snprintf(base, 256, "kernel_flash_attn_ext_vec_reduce"); - snprintf(name, 256, "kernel_flash_attn_ext_vec_reduce_dv=%d_nwg=%d", dv, nwg); - - id res = ggml_metal_get_kernel(ctx, name); - if (res) { - // kernel found - return res; - } - - MTLFunctionConstantValues * cv = [[MTLFunctionConstantValues alloc] init]; - - [cv setConstantValue:&dv type:MTLDataTypeInt atIndex:FC_FLASH_ATTN_EXT_VEC_REDUCE + 0]; - [cv setConstantValue:&nwg type:MTLDataTypeInt atIndex:FC_FLASH_ATTN_EXT_VEC_REDUCE + 1]; - - res = ggml_metal_compile_kernel(backend, base, name, cv); - - [cv release]; - - return res; - } - - GGML_UNUSED(op); -} - -static id ggml_metal_get_pipeline_bin( - ggml_backend_t backend, enum ggml_op op, - int32_t n_fuse, - bool row) { - struct ggml_backend_metal_context * ctx = backend->context; - - char base[256]; - char name[256]; - - @autoreleasepool { - const char * op_str = "undefined"; - switch (op) { - case GGML_OP_ADD: op_str = "add"; break; - case GGML_OP_SUB: op_str = "sub"; break; - case GGML_OP_MUL: op_str = "mul"; break; - case GGML_OP_DIV: op_str = "div"; break; - default: GGML_ABORT("fatal error"); - }; - - if (row) { - snprintf(base, 256, "kernel_%s_row_c4_fuse_%d", op_str, n_fuse); - } else { - snprintf(base, 256, "kernel_%s_fuse_%d", op_str, n_fuse); - } - - snprintf(name, 256, "%s", base); - - id res = ggml_metal_get_kernel(ctx, name); - if (res) { - // kernel found - return res; - } - - return ggml_metal_compile_kernel(backend, base, name, nil); - } -} - -static id ggml_metal_get_pipeline_rms_norm( - ggml_backend_t backend, struct ggml_tensor * op, - int32_t n_fuse) { - struct ggml_backend_metal_context * ctx = backend->context; - - char base[256]; - char name[256]; - - @autoreleasepool { - switch (n_fuse) { - case 1: snprintf(base, 256, "kernel_rms_norm"); break; - case 2: snprintf(base, 256, "kernel_rms_norm_mul"); break; - case 3: snprintf(base, 256, "kernel_rms_norm_mul_add"); break; - default: GGML_ABORT("fatal error"); - } - - snprintf(name, 256, "%s", base); - - id res = ggml_metal_get_kernel(ctx, name); - if (res) { - // kernel found - return res; - } - - return ggml_metal_compile_kernel(backend, base, name, nil); - } - - GGML_UNUSED(op); -} - -static void ggml_metal_free(struct ggml_backend_metal_context * ctx) { - GGML_LOG_INFO("%s: deallocating\n", __func__); - - for (int i = 0; i < GGML_METAL_KERNEL_TYPE_COUNT; ++i) { - [ctx->kernels[i].pipeline release]; - } - - if (ctx->kernels_ext) { - [ctx->kernels_ext release]; - ctx->kernels_ext = nil; - } - - Block_release(ctx->encode_async); - - //[ctx->queue release]; // [TAG_QUEUE_PER_BACKEND] - - for (int i = 0; i < GGML_METAL_MAX_COMMAND_BUFFERS; ++i) { - if (ctx->cmd_bufs[i].obj) { - [ctx->cmd_bufs[i].obj release]; - } - - if (ctx->cmd_bufs[i].mem_ranges) { - ggml_mem_ranges_free(ctx->cmd_bufs[i].mem_ranges); - } - } - - [ctx->cmd_bufs_ext removeAllObjects]; - [ctx->cmd_bufs_ext release]; - - dispatch_release(ctx->d_queue); - - free(ctx); -} - -// temporarily defined here for compatibility between ggml-backend and the old API - -struct ggml_backend_metal_buffer { - void * data; - size_t size; - - id metal; -}; - -struct ggml_backend_metal_buffer_context { - void * all_data; - size_t all_size; - - // if false, the Metal buffer data is allocated in private GPU memory and is not shared with the host - bool is_shared; - - // multiple buffers are used only to avoid the maximum buffer size limitation when using mmap - int n_buffers; - struct ggml_backend_metal_buffer buffers[GGML_METAL_MAX_BUFFERS]; - - // optional MTLResidencySet - // note: cannot use explicity "id" here because it is not available on certain OSes - id rset; - - // pointers to global device objects - id device; - id queue; -}; - -// rset init -static bool ggml_backend_metal_buffer_rset_init( - struct ggml_backend_metal_buffer_context * ctx, - struct ggml_backend_metal_device_context * ctx_dev, - id device) { - ctx->rset = nil; - - if (!ctx_dev->has_residency_sets) { - return true; - } - -#if defined(GGML_METAL_HAS_RESIDENCY_SETS) - if (@available(macOS 15.0, iOS 18.0, tvOS 18.0, visionOS 2.0, *)) { - MTLResidencySetDescriptor * desc = [[MTLResidencySetDescriptor alloc] init]; - desc.label = @"ggml_backend_metal"; - desc.initialCapacity = ctx->n_buffers; - - NSError * error; - ctx->rset = [device newResidencySetWithDescriptor:desc error:&error]; - if (error) { - GGML_LOG_ERROR("%s: error: %s\n", __func__, [[error description] UTF8String]); - [desc release]; - return false; - } - - [desc release]; - - for (int i = 0; i < ctx->n_buffers; i++) { - [ctx->rset addAllocation:ctx->buffers[i].metal]; - } - - [ctx->rset commit]; - [ctx->rset requestResidency]; - - return true; - } -#else - GGML_UNUSED(ctx_dev); - GGML_UNUSED(device); -#endif - - return true; -} - -// rset free -static void ggml_backend_metal_buffer_rset_free(struct ggml_backend_metal_buffer_context * ctx) { -#if defined(GGML_METAL_HAS_RESIDENCY_SETS) - if (@available(macOS 15.0, iOS 18.0, tvOS 18.0, visionOS 2.0, *)) { - if (ctx->rset) { - [ctx->rset endResidency]; - [ctx->rset removeAllAllocations]; - [ctx->rset release]; - } - } -#else - GGML_UNUSED(ctx); -#endif -} - -// finds the Metal buffer that contains the tensor data on the GPU device -// the assumption is that there is 1-to-1 mapping between the host and device memory buffers, so we can find the -// Metal buffer based on the host memory pointer -// -static id ggml_metal_get_buffer(const struct ggml_tensor * t, size_t * offs) { - //GGML_LOG_INFO("%s: data tensor '%16s', offs_data = %8ld, offs_eval = %8ld, offs_cach = %8ld\n", __func__, t->name, offs_data, offs_eval, offs_cach); - - const int64_t tsize = ggml_nbytes(t); - - ggml_backend_buffer_t buffer = t->view_src ? t->view_src->buffer : t->buffer; - - struct ggml_backend_metal_buffer_context * buf_ctx = (struct ggml_backend_metal_buffer_context *) buffer->context; - - // find the view that contains the tensor fully - for (int i = 0; i < buf_ctx->n_buffers; ++i) { - const int64_t ioffs = (int64_t) t->data - (int64_t) buf_ctx->buffers[i].data; - - //GGML_LOG_INFO("ioffs = %10ld, tsize = %10ld, sum = %10ld, buf_ctx->buffers[%d].size = %10ld\n", ioffs, tsize, ioffs + tsize, i, buf_ctx->buffers[i].size); - if (ioffs >= 0 && ioffs + tsize <= (int64_t) buf_ctx->buffers[i].size) { - *offs = (size_t) ioffs; - - //GGML_LOG_INFO("%s: tensor '%16s', offs = %8ld\n", __func__, t->name, *offs); - - return buf_ctx->buffers[i].metal; - } - } - - GGML_LOG_ERROR("%s: error: tensor '%s' buffer is nil\n", __func__, t->name); - - return nil; -} - -static bool ggml_metal_supports_op(const struct ggml_backend_metal_device_context * ctx_dev, const struct ggml_tensor * op) { - const bool has_simdgroup_mm = ctx_dev->has_simdgroup_mm; - const bool has_simdgroup_reduction = ctx_dev->has_simdgroup_reduction; - const bool use_bfloat = ctx_dev->use_bfloat; - - if (!use_bfloat) { - if (op->type == GGML_TYPE_BF16) { - return false; - } - - for (size_t i = 0, n = 3; i < n; ++i) { - if (op->src[i] != NULL && op->src[i]->type == GGML_TYPE_BF16) { - return false; - } - } - } - - switch (op->op) { - case GGML_OP_UNARY: - switch (ggml_get_unary_op(op)) { - case GGML_UNARY_OP_TANH: - case GGML_UNARY_OP_RELU: - case GGML_UNARY_OP_SIGMOID: - case GGML_UNARY_OP_GELU: - case GGML_UNARY_OP_GELU_ERF: - case GGML_UNARY_OP_GELU_QUICK: - case GGML_UNARY_OP_SILU: - case GGML_UNARY_OP_ELU: - case GGML_UNARY_OP_NEG: - case GGML_UNARY_OP_ABS: - case GGML_UNARY_OP_SGN: - case GGML_UNARY_OP_STEP: - case GGML_UNARY_OP_HARDSWISH: - case GGML_UNARY_OP_HARDSIGMOID: - case GGML_UNARY_OP_EXP: - return ggml_is_contiguous(op->src[0]) && op->src[0]->type == GGML_TYPE_F32; - default: - return false; - } - case GGML_OP_GLU: - switch (ggml_get_glu_op(op)) { - case GGML_GLU_OP_REGLU: - case GGML_GLU_OP_GEGLU: - case GGML_GLU_OP_SWIGLU: - case GGML_GLU_OP_SWIGLU_OAI: - case GGML_GLU_OP_GEGLU_ERF: - case GGML_GLU_OP_GEGLU_QUICK: - return ggml_is_contiguous_1(op->src[0]) && op->src[0]->type == GGML_TYPE_F32; - default: - return false; - } - case GGML_OP_NONE: - case GGML_OP_RESHAPE: - case GGML_OP_VIEW: - case GGML_OP_TRANSPOSE: - case GGML_OP_PERMUTE: - case GGML_OP_CONCAT: - return true; - case GGML_OP_ADD: - case GGML_OP_SUB: - case GGML_OP_MUL: - case GGML_OP_DIV: - case GGML_OP_ADD_ID: - return op->src[0]->type == GGML_TYPE_F32; - case GGML_OP_ACC: - case GGML_OP_REPEAT: - case GGML_OP_SCALE: - case GGML_OP_CONV_TRANSPOSE_1D: - return true; - case GGML_OP_CLAMP: - return op->src[0]->type == GGML_TYPE_F32; - case GGML_OP_SQR: - case GGML_OP_SQRT: - case GGML_OP_SIN: - case GGML_OP_COS: - return ggml_is_contiguous(op->src[0]) && op->src[0]->type == GGML_TYPE_F32; - case GGML_OP_LOG: - return false; // TODO: implement - case GGML_OP_SUM_ROWS: - case GGML_OP_MEAN: - case GGML_OP_SOFT_MAX: - case GGML_OP_GROUP_NORM: - return has_simdgroup_reduction && ggml_is_contiguous_rows(op->src[0]); - case GGML_OP_RMS_NORM: - case GGML_OP_L2_NORM: - return has_simdgroup_reduction && (op->ne[0] % 4 == 0 && ggml_is_contiguous_1(op->src[0])); - case GGML_OP_ARGMAX: - return has_simdgroup_reduction; - case GGML_OP_NORM: - return has_simdgroup_reduction && (op->ne[0] % 4 == 0 && ggml_is_contiguous_1(op->src[0])); - case GGML_OP_ROPE: - return true; - case GGML_OP_IM2COL: - return ggml_is_contiguous(op->src[1]) && op->src[1]->type == GGML_TYPE_F32 && (op->type == GGML_TYPE_F16 || op->type == GGML_TYPE_F32); - case GGML_OP_POOL_1D: - return false; - case GGML_OP_UPSCALE: - return op->src[0]->type == GGML_TYPE_F32 && op->op_params[0] == GGML_SCALE_MODE_NEAREST; - case GGML_OP_POOL_2D: - return op->src[0]->type == GGML_TYPE_F32; - case GGML_OP_PAD: - return (ggml_get_op_params_i32(op, 0) == 0) && (ggml_get_op_params_i32(op, 2) == 0) && - (ggml_get_op_params_i32(op, 4) == 0) && (ggml_get_op_params_i32(op, 6) == 0); - case GGML_OP_PAD_REFLECT_1D: - case GGML_OP_TIMESTEP_EMBEDDING: - case GGML_OP_ARGSORT: - case GGML_OP_LEAKY_RELU: - return op->src[0]->type == GGML_TYPE_F32; - case GGML_OP_ARANGE: - return true; - case GGML_OP_FLASH_ATTN_EXT: - // for new head sizes, add checks here - if (op->src[0]->ne[0] != 40 && - op->src[0]->ne[0] != 64 && - op->src[0]->ne[0] != 80 && - op->src[0]->ne[0] != 96 && - op->src[0]->ne[0] != 112 && - op->src[0]->ne[0] != 128 && - op->src[0]->ne[0] != 192 && - op->src[0]->ne[0] != 256) { - return false; - } - if (op->src[0]->ne[0] == 576) { - // DeepSeek sizes - // TODO: disabled for now, until optmized - return false; - } - if (op->src[1]->type != op->src[2]->type) { - return false; - } - return has_simdgroup_mm; // TODO: over-restricted for vec-kernels - case GGML_OP_SSM_CONV: - case GGML_OP_SSM_SCAN: - return has_simdgroup_reduction; - case GGML_OP_RWKV_WKV6: - case GGML_OP_RWKV_WKV7: - return true; - case GGML_OP_MUL_MAT: - case GGML_OP_MUL_MAT_ID: - return has_simdgroup_reduction && - (op->src[0]->type != GGML_TYPE_F32 || op->src[1]->type == GGML_TYPE_F32); - case GGML_OP_CPY: - case GGML_OP_DUP: - case GGML_OP_CONT: - { - switch (op->src[0]->type) { - case GGML_TYPE_F32: - switch (op->type) { - case GGML_TYPE_F32: - case GGML_TYPE_F16: - case GGML_TYPE_BF16: - case GGML_TYPE_Q8_0: - case GGML_TYPE_Q4_0: - case GGML_TYPE_Q4_1: - case GGML_TYPE_Q5_0: - case GGML_TYPE_Q5_1: - case GGML_TYPE_IQ4_NL: - case GGML_TYPE_I32: - return true; - default: - return false; - } - case GGML_TYPE_F16: - switch (op->type) { - case GGML_TYPE_F32: - case GGML_TYPE_F16: - return true; - default: - return false; - } - case GGML_TYPE_BF16: - switch (op->type) { - case GGML_TYPE_F32: - case GGML_TYPE_BF16: - return true; - default: - return false; - } - case GGML_TYPE_Q4_0: - case GGML_TYPE_Q4_1: - case GGML_TYPE_Q5_0: - case GGML_TYPE_Q5_1: - case GGML_TYPE_Q8_0: - switch (op->type) { - case GGML_TYPE_F32: - case GGML_TYPE_F16: - return true; - default: - return false; - } - case GGML_TYPE_I32: - return op->type == GGML_TYPE_F32; - default: - return false; - }; - } - case GGML_OP_DIAG_MASK_INF: - case GGML_OP_GET_ROWS: - { - return op->ne[3] == 1; - } - case GGML_OP_SET_ROWS: - { - if (op->src[0]->type != GGML_TYPE_F32) { - return false; - } - - switch (op->type) { - case GGML_TYPE_F32: - case GGML_TYPE_F16: - case GGML_TYPE_BF16: - case GGML_TYPE_Q8_0: - case GGML_TYPE_Q4_0: - case GGML_TYPE_Q4_1: - case GGML_TYPE_Q5_0: - case GGML_TYPE_Q5_1: - case GGML_TYPE_IQ4_NL: - return true; - default: - return false; - }; - } - default: - return false; - } -} - -struct ggml_metal_encode_context { - ggml_backend_t backend; - - id encoder; - - struct ggml_mem_ranges * mem_ranges; -}; - -static bool ggml_metal_encode_concurrency_reset(struct ggml_metal_encode_context * ctx) { - if (!ctx->mem_ranges) { - return true; - } - - [ctx->encoder memoryBarrierWithScope:MTLBarrierScopeBuffers]; - - ggml_mem_ranges_reset(ctx->mem_ranges); - - return true; -} - -static bool ggml_metal_encode_concurrency_check(struct ggml_metal_encode_context * ctx, const struct ggml_tensor * node) { - if (!ctx->mem_ranges) { - return false; - } - - return ggml_mem_ranges_check(ctx->mem_ranges, node); -} - -static bool ggml_metal_encode_concurrency_add(struct ggml_metal_encode_context * ctx, const struct ggml_tensor * node) { - if (!ctx->mem_ranges) { - return true; - } - - return ggml_mem_ranges_add(ctx->mem_ranges, node); -} - -static int ggml_metal_encode_node(struct ggml_metal_encode_context * ctx_enc, int idx, int idx_end) { - ggml_backend_t backend = ctx_enc->backend; - - id encoder = ctx_enc->encoder; - - struct ggml_backend_metal_context * ctx = backend->context; - struct ggml_backend_metal_device_context * ctx_dev = backend->device->context; - - struct ggml_cgraph * gf = ctx->gf; - - enum ggml_op ops[8]; - - struct ggml_tensor ** nodes = ggml_graph_nodes(gf) + idx; - struct ggml_tensor * node = nodes[0]; - - //GGML_LOG_INFO("%s: encoding node %3d, op = %8s\n", __func__, idx, ggml_op_name(node->op)); - - struct ggml_tensor * src0 = node->src[0]; - struct ggml_tensor * src1 = node->src[1]; - struct ggml_tensor * src2 = node->src[2]; - struct ggml_tensor * dst = node; - - if (ggml_is_empty(dst)) { - return 1; - } - - switch (dst->op) { - case GGML_OP_NONE: - case GGML_OP_RESHAPE: - case GGML_OP_VIEW: - case GGML_OP_TRANSPOSE: - case GGML_OP_PERMUTE: - { - // noop -> next node - } return 1; - default: - { - } break; - } - - if (!ggml_metal_supports_op(ctx_dev, dst)) { - GGML_LOG_ERROR("%s: error: unsupported op '%s'\n", __func__, ggml_op_desc(dst)); - GGML_ABORT("unsupported op"); - } - - const int64_t ne00 = src0 ? src0->ne[0] : 0; - const int64_t ne01 = src0 ? src0->ne[1] : 0; - const int64_t ne02 = src0 ? src0->ne[2] : 0; - const int64_t ne03 = src0 ? src0->ne[3] : 0; - - const uint64_t nb00 = src0 ? src0->nb[0] : 0; - const uint64_t nb01 = src0 ? src0->nb[1] : 0; - const uint64_t nb02 = src0 ? src0->nb[2] : 0; - const uint64_t nb03 = src0 ? src0->nb[3] : 0; - - const int64_t ne10 = src1 ? src1->ne[0] : 0; - const int64_t ne11 = src1 ? src1->ne[1] : 0; - const int64_t ne12 = src1 ? src1->ne[2] : 0; - const int64_t ne13 = src1 ? src1->ne[3] : 0; - - const uint64_t nb10 = src1 ? src1->nb[0] : 0; - const uint64_t nb11 = src1 ? src1->nb[1] : 0; - const uint64_t nb12 = src1 ? src1->nb[2] : 0; - const uint64_t nb13 = src1 ? src1->nb[3] : 0; - - const int64_t ne20 = src2 ? src2->ne[0] : 0; - const int64_t ne21 = src2 ? src2->ne[1] : 0; - const int64_t ne22 = src2 ? src2->ne[2] : 0; GGML_UNUSED(ne22); - const int64_t ne23 = src2 ? src2->ne[3] : 0; GGML_UNUSED(ne23); - - const uint64_t nb20 = src2 ? src2->nb[0] : 0; GGML_UNUSED(nb20); - const uint64_t nb21 = src2 ? src2->nb[1] : 0; - const uint64_t nb22 = src2 ? src2->nb[2] : 0; - const uint64_t nb23 = src2 ? src2->nb[3] : 0; GGML_UNUSED(nb23); - - const int64_t ne0 = dst ? dst->ne[0] : 0; - const int64_t ne1 = dst ? dst->ne[1] : 0; - const int64_t ne2 = dst ? dst->ne[2] : 0; - const int64_t ne3 = dst ? dst->ne[3] : 0; - - const uint64_t nb0 = dst ? dst->nb[0] : 0; - const uint64_t nb1 = dst ? dst->nb[1] : 0; - const uint64_t nb2 = dst ? dst->nb[2] : 0; - const uint64_t nb3 = dst ? dst->nb[3] : 0; - - size_t offs_src[GGML_MAX_SRC]; - - id id_src[GGML_MAX_SRC]; - - enum ggml_type srct[GGML_MAX_SRC]; - - for (int i = 0; i < GGML_MAX_SRC; i++) { - offs_src[i] = 0; - id_src[i] = node->src[i] ? ggml_metal_get_buffer(node->src[i], &offs_src[i]) : nil; - srct[i] = node->src[i] ? node->src[i]->type : GGML_TYPE_COUNT; - } - - // TODO: tmp shorthands - remove - size_t offs_src0 = offs_src[0]; - size_t offs_src1 = offs_src[1]; - size_t offs_src2 = offs_src[2]; - - id id_src0 = id_src[0]; - id id_src1 = id_src[1]; - id id_src2 = id_src[2]; - - const enum ggml_type src0t = src0 ? src0->type : GGML_TYPE_COUNT; - const enum ggml_type src1t = src1 ? src1->type : GGML_TYPE_COUNT; - const enum ggml_type src2t = src2 ? src2->type : GGML_TYPE_COUNT; - const enum ggml_type dstt = dst ? dst->type : GGML_TYPE_COUNT; - - size_t offs_dst = 0; - - id id_dst = dst ? ggml_metal_get_buffer(dst, &offs_dst) : nil; - - int n_fuse = 1; - - // check if the current node can run concurrently with other nodes before it - // the condition is that: - // - the current node cannot write to any previous src or dst ranges - // - the current node cannot read from any previous dst ranges - // - // if the condition is not satisfied, we put a memory barrier and clear all ranges - // otherwise, we add the new ranges to the encoding context and process the node concurrently - // - { - const bool is_concurrent = ggml_metal_encode_concurrency_check(ctx_enc, node); - - if (!is_concurrent) { - ggml_metal_encode_concurrency_reset(ctx_enc); - } - - if (ctx_dev->debug_graph > 0) { - GGML_LOG_DEBUG("%s: node[%5d] - %-12s %s\n", __func__, idx, ggml_op_name(dst->op), is_concurrent ? "(concurrent)" : ""); - } - if (ctx_dev->debug_graph > 1) { - if (src0) { - GGML_LOG_DEBUG("%s: src0 - %4s [%5lld, %5lld, %5lld, %5lld] [%5lld, %5lld, %5lld, %5lld], %d, %s\n", __func__, ggml_type_name(src0t), ne00, ne01, ne02, ne03, nb00, nb01, nb02, nb03, - ggml_is_contiguous(src0), src0->name); - } - if (src1) { - GGML_LOG_DEBUG("%s: src1 - %4s [%5lld, %5lld, %5lld, %5lld] [%5lld, %5lld, %5lld, %5lld], %d, %s\n", __func__, ggml_type_name(src1t), ne10, ne11, ne12, ne13, nb10, nb11, nb12, nb13, - ggml_is_contiguous(src1), src1->name); - } - if (dst) { - GGML_LOG_DEBUG("%s: dst - %4s [%5lld, %5lld, %5lld, %5lld] [%5lld, %5lld, %5lld, %5lld], 1, %s\n", __func__, ggml_type_name(dstt), ne0, ne1, ne2, ne3, nb0, nb1, nb2, nb3, - dst->name); - } - } - } - - id device = ctx_dev->mtl_device; - - switch (dst->op) { - case GGML_OP_CONCAT: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CONCAT].pipeline; - - const int32_t dim = ((const int32_t *) dst->op_params)[0]; - - ggml_metal_kargs_concat args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne10 =*/ ne10, - /*.ne11 =*/ ne11, - /*.ne12 =*/ ne12, - /*.ne13 =*/ ne13, - /*.nb10 =*/ nb10, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb13 =*/ nb13, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.nb0 =*/ nb0, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3, - /*.dim =*/ dim, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; - - const int nth = MIN(1024, ne0); - - [encoder dispatchThreadgroups:MTLSizeMake(ne1, ne2, ne3) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_ADD: - case GGML_OP_SUB: - case GGML_OP_MUL: - case GGML_OP_DIV: - { - GGML_ASSERT(src0t == GGML_TYPE_F32); - GGML_ASSERT(src1t == GGML_TYPE_F32); - - GGML_ASSERT(ggml_is_contiguous_rows(src0)); - GGML_ASSERT(ggml_is_contiguous_rows(src1)); - - const size_t offs = 0; - - bool bcast_row = false; - - ggml_metal_kargs_bin args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne10 =*/ ne10, - /*.ne11 =*/ ne11, - /*.ne12 =*/ ne12, - /*.ne13 =*/ ne13, - /*.nb10 =*/ nb10, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb13 =*/ nb13, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.nb0 =*/ nb0, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3, - /*.offs =*/ offs, - /*.o1 =*/ { offs_src1 }, - }; - - // c[0] = add(a, b[0]) - // c[1] = add(c[0], b[1]) - // c[2] = add(c[1], b[2]) - // ... - if (ctx_dev->use_fusion) { - ops[0] = GGML_OP_ADD; - ops[1] = GGML_OP_ADD; - ops[2] = GGML_OP_ADD; - ops[3] = GGML_OP_ADD; - ops[4] = GGML_OP_ADD; - ops[5] = GGML_OP_ADD; - ops[6] = GGML_OP_ADD; - ops[7] = GGML_OP_ADD; - - size_t offs_fuse; - id id_fuse; - - // note: in metal, we sometimes encode the graph in parallel so we have to avoid fusing nodes - // across splits. idx_end indicates the last node in the current split - for (n_fuse = 0; n_fuse <= 6 && idx + n_fuse + 1 < idx_end; ++n_fuse) { - if (!ggml_can_fuse(gf, idx + n_fuse, ops + n_fuse, 2)) { - break; - } - - if (nodes[n_fuse] != nodes[n_fuse + 1]->src[0]) { - break; - } - - // b[0] === b[1] === ... - if (!ggml_are_same_layout(nodes[n_fuse]->src[1], nodes[n_fuse + 1]->src[1])) { - break; - } - - // only fuse nodes if src1 is in the same Metal buffer - id_fuse = ggml_metal_get_buffer(nodes[n_fuse + 1]->src[1], &offs_fuse); - if (id_fuse != id_src1) { - break; - } - - ctx_dev->fuse_cnt[nodes[n_fuse + 1]->op]++; - - args.o1[n_fuse + 1] = offs_fuse; - } - - ++n_fuse; - - if (ctx_dev->debug_fusion > 1 && n_fuse > 1) { - GGML_LOG_DEBUG("%s: fuse: ADD x %d\n", __func__, n_fuse); - } - } - - id pipeline = nil; - - if (ggml_nelements(src1) == ne10 && ggml_is_contiguous(src1) && ne00 % 4 == 0 && ne10 % 4 == 0) { - GGML_ASSERT(ggml_is_contiguous(src0)); - - // src1 is a row - GGML_ASSERT(ne11 == 1); - - pipeline = ggml_metal_get_pipeline_bin(backend, dst->op, n_fuse, true); - - bcast_row = true; - } else { - pipeline = ggml_metal_get_pipeline_bin(backend, dst->op, n_fuse, false); - } - - if (n_fuse > 1) { - id_dst = ggml_metal_get_buffer(nodes[n_fuse - 1], &offs_dst); - - for (int i = 1; i < n_fuse; ++i) { - if (!ggml_metal_encode_concurrency_check(ctx_enc, nodes[i])) { - ggml_metal_encode_concurrency_reset(ctx_enc); - - break; - } - } - } - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:0 atIndex:2]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; - - if (bcast_row) { - const int64_t n = ggml_nelements(dst)/4; - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } else { - int nth = 32; - - while (16*nth < ne0 && nth < (int) pipeline.maxTotalThreadsPerThreadgroup) { - nth *= 2; - } - - [encoder dispatchThreadgroups:MTLSizeMake(ne01, ne02, ne03) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } - } break; - case GGML_OP_ADD_ID: - { - GGML_ASSERT(src0t == GGML_TYPE_F32); - GGML_ASSERT(src1t == GGML_TYPE_F32); - GGML_ASSERT(src2t == GGML_TYPE_I32); - GGML_ASSERT(dstt == GGML_TYPE_F32); - - GGML_ASSERT(ggml_is_contiguous_rows(src0)); - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ADD_ID].pipeline; - - ggml_metal_kargs_add_id args = { - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb11 =*/ nb11, - /*.nb21 =*/ nb21, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:3]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:4]; - - const int nth = MIN((int) pipeline.maxTotalThreadsPerThreadgroup, ne00); - - [encoder dispatchThreadgroups:MTLSizeMake(ne01, ne02, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_REPEAT: - { - id pipeline; - - switch (src0t) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_REPEAT_F32].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_REPEAT_F16].pipeline; break; - case GGML_TYPE_I32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_REPEAT_I32].pipeline; break; - case GGML_TYPE_I16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_REPEAT_I16].pipeline; break; - default: GGML_ABORT("fatal error"); - } - - ggml_metal_kargs_repeat args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.nb0 =*/ nb0, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - - const int nth = MIN((int) pipeline.maxTotalThreadsPerThreadgroup, ne0); - - [encoder dispatchThreadgroups:MTLSizeMake(ne1, ne2, ne3) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_ACC: - { - GGML_ASSERT(src0t == GGML_TYPE_F32); - GGML_ASSERT(src1t == GGML_TYPE_F32); - GGML_ASSERT(dstt == GGML_TYPE_F32); - - GGML_ASSERT(ggml_is_contiguous(src0)); - GGML_ASSERT(ggml_is_contiguous(src1)); - - const size_t pnb1 = ((const int32_t *) dst->op_params)[0]; - const size_t pnb2 = ((const int32_t *) dst->op_params)[1]; - const size_t pnb3 = ((const int32_t *) dst->op_params)[2]; - const size_t offs = ((const int32_t *) dst->op_params)[3]; - - const bool inplace = (bool) ((const int32_t *) dst->op_params)[4]; - - if (!inplace) { - // run a separete kernel to cpy src->dst - // not sure how to avoid this - // TODO: make a simpler cpy_bytes kernel - - const id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_F32].pipeline; - - ggml_metal_kargs_cpy args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.nb0 =*/ nb0, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - - const int nth = MIN((int) pipeline.maxTotalThreadsPerThreadgroup, ne00); - - [encoder dispatchThreadgroups:MTLSizeMake(ne01, ne02, ne03) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - - ggml_metal_encode_concurrency_reset(ctx_enc); - } - - ggml_metal_kargs_bin args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb00 =*/ nb00, - /*.nb01 =*/ pnb1, - /*.nb02 =*/ pnb2, - /*.nb03 =*/ pnb3, - /*.ne10 =*/ ne10, - /*.ne11 =*/ ne11, - /*.ne12 =*/ ne12, - /*.ne13 =*/ ne13, - /*.nb10 =*/ nb10, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb13 =*/ nb13, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.nb0 =*/ nb0, - /*.nb1 =*/ pnb1, - /*.nb2 =*/ pnb2, - /*.nb3 =*/ pnb3, - /*.offs =*/ offs, - /*.o1 =*/ { offs_src1}, - }; - - //const id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ADD].pipeline; - const id pipeline = ggml_metal_get_pipeline_bin(backend, GGML_OP_ADD, 1, false); - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:0 atIndex:2]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; - - const int nth = MIN((int) pipeline.maxTotalThreadsPerThreadgroup, ne00); - - [encoder dispatchThreadgroups:MTLSizeMake(ne11, ne12, ne13) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_SCALE: - { - GGML_ASSERT(ggml_is_contiguous(src0)); - - float scale; - float bias; - memcpy(&scale, ((const int32_t *) dst->op_params) + 0, sizeof(float)); - memcpy(&bias, ((const int32_t *) dst->op_params) + 1, sizeof(float)); - - int64_t n = ggml_nelements(dst); - - id pipeline = nil; - - if (n % 4 == 0) { - n /= 4; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SCALE_4].pipeline; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SCALE].pipeline; - } - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&scale length:sizeof(scale) atIndex:2]; - [encoder setBytes:&bias length:sizeof(bias) atIndex:3]; - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_OP_CLAMP: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CLAMP].pipeline; - - float min; - float max; - memcpy(&min, ((const int32_t *) dst->op_params) + 0, sizeof(float)); - memcpy(&max, ((const int32_t *) dst->op_params) + 1, sizeof(float)); - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&min length:sizeof(min) atIndex:2]; - [encoder setBytes:&max length:sizeof(max) atIndex:3]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_OP_UNARY: - switch (ggml_get_unary_op(node)) { - // we are not taking into account the strides, so for now require contiguous tensors - GGML_ASSERT(ggml_is_contiguous(src0)); - - case GGML_UNARY_OP_TANH: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_TANH].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_RELU: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_RELU].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_SIGMOID: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SIGMOID].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_GELU: - { - int64_t n = ggml_nelements(dst); - - id pipeline = nil; - - if (n % 4 == 0) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GELU_4].pipeline; - n /= 4; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GELU].pipeline; - } - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_GELU_ERF: - { - int64_t n = ggml_nelements(dst); - - id pipeline = nil; - - if (n % 4 == 0) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GELU_ERF_4].pipeline; - n /= 4; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GELU_ERF].pipeline; - } - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_GELU_QUICK: - { - int64_t n = ggml_nelements(dst); - - id pipeline = nil; - - if (n % 4 == 0) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GELU_QUICK_4].pipeline; - n /= 4; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GELU_QUICK].pipeline; - } - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_SILU: - { - int64_t n = ggml_nelements(dst); - - id pipeline = nil; - - if (n % 4 == 0) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SILU_4].pipeline; - n /= 4; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SILU].pipeline; - } - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_ELU: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ELU].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_NEG: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_NEG].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_ABS: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ABS].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_SGN: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SGN].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_STEP: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_STEP].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_HARDSWISH: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_HARDSWISH].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_HARDSIGMOID: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_HARDSIGMOID].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_UNARY_OP_EXP: - { - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_EXP].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - default: - { - GGML_LOG_WARN("%s: node %3d, op = %8s not implemented\n", __func__, idx, ggml_op_name(dst->op)); - GGML_ABORT("fatal error"); - } - } break; - case GGML_OP_GLU: - { - GGML_ASSERT(ggml_is_contiguous_1(src0)); - - if (src1) { - GGML_ASSERT(ggml_are_same_shape(src0, src1)); - } - - id pipeline = nil; - - switch (ggml_get_glu_op(node)) { - case GGML_GLU_OP_REGLU: - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_REGLU].pipeline; - break; - case GGML_GLU_OP_GEGLU: - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GEGLU].pipeline; - break; - case GGML_GLU_OP_SWIGLU: - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SWIGLU].pipeline; - break; - case GGML_GLU_OP_SWIGLU_OAI: - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SWIGLU_OAI].pipeline; - break; - case GGML_GLU_OP_GEGLU_ERF: - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GEGLU_ERF].pipeline; - break; - case GGML_GLU_OP_GEGLU_QUICK: - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GEGLU_QUICK].pipeline; - break; - default: - GGML_ABORT("fatal error"); - } - - const int32_t swp = ggml_get_op_params_i32(dst, 1); - const float alpha = ggml_get_op_params_f32(dst, 2); - const float limit = ggml_get_op_params_f32(dst, 3); - - const int32_t i00 = swp ? ne0 : 0; - const int32_t i10 = swp ? 0 : ne0; - - ggml_metal_kargs_glu args = { - /*.ne00 =*/ ne00, - /*.nb01 =*/ nb01, - /*.ne10 =*/ src1 ? ne10 : ne00, - /*.nb11 =*/ src1 ? nb11 : nb01, - /*.ne0 =*/ ne0, - /*.nb1 =*/ nb1, - /*.i00 =*/ src1 ? 0 : i00, - /*.i10 =*/ src1 ? 0 : i10, - /*.alpha=*/ alpha, - /*.limit=*/ limit - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - if (src1) { - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - } else { - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - } - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - [encoder setBytes:&args length:sizeof(args) atIndex:3]; - - const int64_t nrows = ggml_nrows(src0); - - const int32_t nth = MIN((int) pipeline.maxTotalThreadsPerThreadgroup, ne00/2); - - [encoder dispatchThreadgroups:MTLSizeMake(nrows, 1, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_SQR: - { - GGML_ASSERT(ggml_is_contiguous(src0)); - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SQR].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_OP_SQRT: - { - GGML_ASSERT(ggml_is_contiguous(src0)); - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SQRT].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_OP_SIN: - { - GGML_ASSERT(ggml_is_contiguous(src0)); - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SIN].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_OP_COS: - { - GGML_ASSERT(ggml_is_contiguous(src0)); - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_COS].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_OP_SUM_ROWS: - case GGML_OP_MEAN: - { - GGML_ASSERT(src0->nb[0] == ggml_type_size(src0->type)); - - id pipeline = nil; - - switch (dst->op) { - case GGML_OP_SUM_ROWS: - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SUM_ROWS].pipeline; - break; - case GGML_OP_MEAN: - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MEAN].pipeline; - break; - default: - GGML_ABORT("fatal error"); - } - - int nth = 32; // SIMD width - - while (nth < ne00 && nth < (int) pipeline.maxTotalThreadsPerThreadgroup) { - nth *= 2; - } - - nth = MIN(nth, (int) pipeline.maxTotalThreadsPerThreadgroup); - nth = MIN(nth, ne00); - - ggml_metal_kargs_sum_rows args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne10 =*/ ne10, - /*.ne11 =*/ ne11, - /*.ne12 =*/ ne12, - /*.ne13 =*/ ne13, - /*.nb10 =*/ nb10, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb13 =*/ nb13, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.nb0 =*/ nb0, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - [encoder setThreadgroupMemoryLength:32*sizeof(float) atIndex:0]; - - [encoder dispatchThreadgroups:MTLSizeMake(ne01, ne02, ne03) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_SOFT_MAX: - { - GGML_ASSERT(!src1 || src1->type == GGML_TYPE_F16 || src1->type == GGML_TYPE_F32); - - int nth = 32; // SIMD width - - id pipeline = nil; - - const bool use_f16 = (src1 && src1->type == GGML_TYPE_F16); - - if (ne00%4 == 0) { - while (nth < ne00/4 && nth*ne01*ne02*ne03 < 256) { - nth *= 2; - } - if (use_f16) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SOFT_MAX_F16_4].pipeline; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SOFT_MAX_F32_4].pipeline; - } - } else { - while (nth < ne00 && nth*ne01*ne02*ne03 < 256) { - nth *= 2; - } - if (use_f16) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SOFT_MAX_F16].pipeline; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SOFT_MAX_F32].pipeline; - } - } - - float scale; - float max_bias; - - memcpy(&scale, ((const int32_t *) dst->op_params) + 0, sizeof(scale)); - memcpy(&max_bias, ((const int32_t *) dst->op_params) + 1, sizeof(max_bias)); - - const uint32_t n_head = src0->ne[2]; - const uint32_t n_head_log2 = 1u << (uint32_t) floorf(log2f((float) n_head)); - - const float m0 = powf(2.0f, -(max_bias ) / n_head_log2); - const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2); - - id h_src0 = id_src0; - - // softmax - - ggml_metal_kargs_soft_max args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne11 =*/ ne11, - /*.ne12 =*/ ne12, - /*.ne13 =*/ ne13, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb13 =*/ nb13, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3, - /*.scale =*/ scale, - /*.max_bias =*/ max_bias, - /*.m0 =*/ m0, - /*.m1 =*/ m1, - /*.n_head_log2 =*/ n_head_log2, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:h_src0 offset:offs_src0 atIndex:0]; - if (id_src1) { - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - } else { - [encoder setBuffer:h_src0 offset:offs_src0 atIndex:1]; - } - if (id_src2) { - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:2]; - } else { - [encoder setBuffer:h_src0 offset:offs_src0 atIndex:2]; - } - [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; - [encoder setBytes:&args length:sizeof(args) atIndex:4]; - - [encoder setThreadgroupMemoryLength:32*sizeof(float) atIndex:0]; - - [encoder dispatchThreadgroups:MTLSizeMake(ne01, ne02, ne03) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_DIAG_MASK_INF: - { - const int n_past = ((const int32_t *)(dst->op_params))[0]; - - id pipeline = nil; - - if (ne00%8 == 0) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_DIAG_MASK_INF_8].pipeline; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_DIAG_MASK_INF].pipeline; - } - - ggml_metal_kargs_diag_mask_inf args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.n_past =*/ n_past, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&args length:sizeof(args) atIndex:2]; - - if (ne00%8 == 0) { - [encoder dispatchThreadgroups:MTLSizeMake(ne00*ne01*ne02/8, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } - else { - [encoder dispatchThreadgroups:MTLSizeMake(ne00, ne01, ne02) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } - } break; - case GGML_OP_SSM_CONV: - { - GGML_ASSERT(src0t == GGML_TYPE_F32); - GGML_ASSERT(src1t == GGML_TYPE_F32); - - GGML_ASSERT(ggml_is_contiguous(src0)); - GGML_ASSERT(ggml_is_contiguous(src1)); - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SSM_CONV_F32].pipeline; - - ggml_metal_kargs_ssm_conv args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.ne10 =*/ ne10, - /*.ne11 =*/ ne11, - /*.nb10 =*/ nb10, - /*.nb11 =*/ nb11, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.nb0 =*/ nb0, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - [encoder setBytes:&args length:sizeof(args) atIndex:3]; - - [encoder dispatchThreadgroups:MTLSizeMake(ne01, ne1, ne02) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_OP_SSM_SCAN: - { - struct ggml_tensor * src3 = node->src[3]; - struct ggml_tensor * src4 = node->src[4]; - struct ggml_tensor * src5 = node->src[5]; - struct ggml_tensor * src6 = node->src[6]; - - GGML_ASSERT(src3); - GGML_ASSERT(src4); - GGML_ASSERT(src5); - GGML_ASSERT(src6); - - size_t offs_src3 = 0; - size_t offs_src4 = 0; - size_t offs_src5 = 0; - size_t offs_src6 = 0; - - id id_src3 = src3 ? ggml_metal_get_buffer(src3, &offs_src3) : nil; - id id_src4 = src4 ? ggml_metal_get_buffer(src4, &offs_src4) : nil; - id id_src5 = src5 ? ggml_metal_get_buffer(src5, &offs_src5) : nil; - id id_src6 = src6 ? ggml_metal_get_buffer(src6, &offs_src6) : nil; - - const int64_t ne30 = src3->ne[0]; - const int64_t ne31 = src3->ne[1]; GGML_UNUSED(ne31); - - const uint64_t nb30 = src3->nb[0]; GGML_UNUSED(nb30); - const uint64_t nb31 = src3->nb[1]; - - const int64_t ne40 = src4->ne[0]; GGML_UNUSED(ne40); - const int64_t ne41 = src4->ne[1]; - const int64_t ne42 = src4->ne[2]; GGML_UNUSED(ne42); - const int64_t ne43 = src4->ne[3]; GGML_UNUSED(ne43); - - const uint64_t nb40 = src4->nb[0]; GGML_UNUSED(nb40); - const uint64_t nb41 = src4->nb[1]; - const uint64_t nb42 = src4->nb[2]; - const uint64_t nb43 = src4->nb[3]; - - const int64_t ne50 = src5->ne[0]; GGML_UNUSED(ne50); - const int64_t ne51 = src5->ne[1]; GGML_UNUSED(ne51); - const int64_t ne52 = src5->ne[2]; GGML_UNUSED(ne52); - const int64_t ne53 = src5->ne[3]; GGML_UNUSED(ne53); - - const uint64_t nb50 = src5->nb[0]; GGML_UNUSED(nb50); - const uint64_t nb51 = src5->nb[1]; - const uint64_t nb52 = src5->nb[2]; - const uint64_t nb53 = src5->nb[3]; - - const int64_t ne60 = src6->ne[0]; GGML_UNUSED(ne60); - - const uint64_t nb60 = src6->nb[0]; GGML_UNUSED(nb60); - - const int64_t d_state = ne00; - const int64_t d_inner = ne01; - const int64_t n_head = ne02; - const int64_t n_group = ne41; - const int64_t n_seq_tokens = ne12; - const int64_t n_seqs = ne13; - - id pipeline = nil; - - if (ne30 == 1) { - // Mamba-2 - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SSM_SCAN_F32_GROUP].pipeline; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SSM_SCAN_F32].pipeline; - } - - ggml_metal_kargs_ssm_scan args = { - /*.d_state =*/ d_state, - /*.d_inner =*/ d_inner, - /*.n_head =*/ n_head, - /*.n_group =*/ n_group, - /*.n_seq_tokens =*/ n_seq_tokens, - /*.n_seqs =*/ n_seqs, - /*.s_off =*/ ggml_nelements(src1) * sizeof(float), - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb13 =*/ nb13, - /*.nb21 =*/ nb21, - /*.nb22 =*/ nb22, - /*.nb31 =*/ nb31, - /*.nb41 =*/ nb41, - /*.nb42 =*/ nb42, - /*.nb43 =*/ nb43, - /*.nb51 =*/ nb51, - /*.nb52 =*/ nb52, - /*.nb53 =*/ nb53, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:2]; - [encoder setBuffer:id_src3 offset:offs_src3 atIndex:3]; - [encoder setBuffer:id_src4 offset:offs_src4 atIndex:4]; - [encoder setBuffer:id_src5 offset:offs_src5 atIndex:5]; - [encoder setBuffer:id_src6 offset:offs_src6 atIndex:6]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:7]; - [encoder setBytes:&args length:sizeof(args) atIndex:8]; - - // One shared memory bucket for each simd group in the threadgroup - // NOTE: Metal kernels require the buffer size to be multiple of 16 bytes - // https://developer.apple.com/documentation/metal/mtlcomputecommandencoder/1443142-setthreadgroupmemorylength - if (d_state >= 32) { - GGML_ASSERT((int64_t)(d_state / 32) <= 32); - const int64_t shmem_size = 32; - GGML_ASSERT(d_state <= (int64_t)pipeline.maxTotalThreadsPerThreadgroup); - [encoder setThreadgroupMemoryLength:(shmem_size)*sizeof(float) atIndex:0]; - } - - if (ne30 == 1) { - // Mamba-2 - [encoder dispatchThreadgroups:MTLSizeMake(d_inner, n_head, n_seqs) threadsPerThreadgroup:MTLSizeMake(d_state, 1, 1)]; - } else { - GGML_ASSERT(d_inner == 1); - [encoder dispatchThreadgroups:MTLSizeMake(n_head, n_seqs, 1) threadsPerThreadgroup:MTLSizeMake(d_state, 1, 1)]; - } - } break; - case GGML_OP_RWKV_WKV6: - { - const int64_t B = dst->src[5]->ne[1]; - const int64_t T = dst->src[0]->ne[2]; - const int64_t C = dst->ne[0]; - const int64_t H = dst->src[0]->ne[1]; - - GGML_ASSERT(dst->src[5]->type == GGML_TYPE_F32); - GGML_ASSERT(C % H == 0); - GGML_ASSERT(C / H == 64); - - size_t offs_src3 = 0; - size_t offs_src4 = 0; - size_t offs_src5 = 0; - - id id_src3 = dst->src[3] ? ggml_metal_get_buffer(dst->src[3], &offs_src3) : nil; - id id_src4 = dst->src[4] ? ggml_metal_get_buffer(dst->src[4], &offs_src4) : nil; - id id_src5 = dst->src[5] ? ggml_metal_get_buffer(dst->src[5], &offs_src5) : nil; - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_RWKV_WKV6_F32].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:2]; - [encoder setBuffer:id_src3 offset:offs_src3 atIndex:3]; - [encoder setBuffer:id_src4 offset:offs_src4 atIndex:4]; - [encoder setBuffer:id_src5 offset:offs_src5 atIndex:5]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:6]; - - [encoder setBytes:&B length:sizeof(B) atIndex:7]; - [encoder setBytes:&T length:sizeof(T) atIndex:8]; - [encoder setBytes:&C length:sizeof(C) atIndex:9]; - [encoder setBytes:&H length:sizeof(H) atIndex:10]; - - [encoder dispatchThreadgroups:MTLSizeMake(B * H, 1, 1) threadsPerThreadgroup:MTLSizeMake(C/ H, 1, 1)]; - } break; - case GGML_OP_RWKV_WKV7: - { - const int64_t B = dst->src[6]->ne[1]; - const int64_t T = dst->src[0]->ne[2]; - const int64_t C = dst->ne[0]; - const int64_t H = dst->src[0]->ne[1]; - - GGML_ASSERT(dst->src[6]->type == GGML_TYPE_F32); - GGML_ASSERT(C % H == 0); - GGML_ASSERT(C / H == 64); - - size_t offs_src3 = 0; - size_t offs_src4 = 0; - size_t offs_src5 = 0; - size_t offs_src6 = 0; - - id id_src3 = dst->src[3] ? ggml_metal_get_buffer(dst->src[3], &offs_src3) : nil; - id id_src4 = dst->src[4] ? ggml_metal_get_buffer(dst->src[4], &offs_src4) : nil; - id id_src5 = dst->src[5] ? ggml_metal_get_buffer(dst->src[5], &offs_src5) : nil; - id id_src6 = dst->src[6] ? ggml_metal_get_buffer(dst->src[6], &offs_src6) : nil; - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_RWKV_WKV7_F32].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:2]; - [encoder setBuffer:id_src3 offset:offs_src3 atIndex:3]; - [encoder setBuffer:id_src4 offset:offs_src4 atIndex:4]; - [encoder setBuffer:id_src5 offset:offs_src5 atIndex:5]; - [encoder setBuffer:id_src6 offset:offs_src6 atIndex:6]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:7]; - - [encoder setBytes:&B length:sizeof(B) atIndex:8]; - [encoder setBytes:&T length:sizeof(T) atIndex:9]; - [encoder setBytes:&C length:sizeof(C) atIndex:10]; - [encoder setBytes:&H length:sizeof(H) atIndex:11]; - - [encoder dispatchThreadgroups:MTLSizeMake(B * H, 1, 1) threadsPerThreadgroup:MTLSizeMake(C/ H, 1, 1)]; - } break; - case GGML_OP_MUL_MAT: - { - GGML_ASSERT(ne00 == ne10); - - GGML_ASSERT(ne12 % ne02 == 0); - GGML_ASSERT(ne13 % ne03 == 0); - - const uint32_t r2 = ne12/ne02; - const uint32_t r3 = ne13/ne03; - - // find the break-even point where the matrix-matrix kernel becomes more efficient compared - // to the matrix-vector kernel - const int ne11_mm_min = 8; - - // first try to use small-batch mat-mv kernels - // these should be efficient for BS [2, ~8] - if (src1t == GGML_TYPE_F32 && (ne00%128 == 0) && - ( - ( - ( - src0t == GGML_TYPE_F32 || // TODO: helper function - src0t == GGML_TYPE_F16 || - src0t == GGML_TYPE_Q4_0 || - src0t == GGML_TYPE_Q4_1 || - src0t == GGML_TYPE_Q5_0 || - src0t == GGML_TYPE_Q5_1 || - src0t == GGML_TYPE_Q8_0 || - src0t == GGML_TYPE_MXFP4 || - src0t == GGML_TYPE_IQ4_NL || - false) && (ne11 >= 2 && ne11 <= 8) - ) || - ( - ( - src0t == GGML_TYPE_Q4_K || - src0t == GGML_TYPE_Q5_K || - src0t == GGML_TYPE_Q6_K || - false) && (ne11 >= 4 && ne11 <= 8) - ) - ) - ) { - // TODO: determine the optimal parameters based on grid utilization - // I still don't know why we should not always use the maximum available threads: - // - // nsg = pipeline.maxTotalThreadsPerThreadgroup / 32 - // - // my current hypothesis is that the work grid is not evenly divisible for different nsg - // values and there can be some tail effects when nsg is high. need to confirm this - // - const int nsg = 2; // num simdgroups per threadgroup - - // num threads along row per simdgroup - int nxpsg = 0; - if (ne00 % 256 == 0 && ne11 < 3) { - nxpsg = 16; - } else if (ne00 % 128 == 0) { - nxpsg = 8; - } else { - nxpsg = 4; - } - - const int nypsg = 32/nxpsg; // num threads along col per simdgroup (i.e. a simdgroup processes that many src0 rows at a time) - const int r0ptg = nypsg*nsg; // num src0 rows per threadgroup - int r1ptg = 4; // num src1 rows per threadgroup - - // note: not sure how optimal are those across all different hardware. there might be someting cleverer - switch (ne11) { - case 2: - r1ptg = 2; break; - case 3: - case 6: - r1ptg = 3; break; - case 4: - case 7: - case 8: - r1ptg = 4; break; - case 5: - r1ptg = 5; break; - }; - - id pipeline = nil; - - switch (src0->type) { - case GGML_TYPE_F32: - switch (r1ptg) { - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F32_F32_R1_2].pipeline; break; - case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F32_F32_R1_3].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F32_F32_R1_4].pipeline; break; - case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F32_F32_R1_5].pipeline; break; - default: GGML_ABORT("not implemented"); - } break; - case GGML_TYPE_F16: - switch (r1ptg) { - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_2].pipeline; break; - case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_3].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_4].pipeline; break; - case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_F16_F32_R1_5].pipeline; break; - default: GGML_ABORT("not implemented"); - } break; - case GGML_TYPE_Q4_0: - switch (r1ptg) { - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_2].pipeline; break; - case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_3].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_4].pipeline; break; - case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_0_F32_R1_5].pipeline; break; - default: GGML_ABORT("not implemented"); - } break; - case GGML_TYPE_Q4_1: - switch (r1ptg) { - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_2].pipeline; break; - case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_3].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_4].pipeline; break; - case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_1_F32_R1_5].pipeline; break; - default: GGML_ABORT("not implemented"); - } break; - case GGML_TYPE_Q5_0: - switch (r1ptg) { - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_2].pipeline; break; - case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_3].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_4].pipeline; break; - case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_0_F32_R1_5].pipeline; break; - default: GGML_ABORT("not implemented"); - } break; - case GGML_TYPE_Q5_1: - switch (r1ptg) { - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_2].pipeline; break; - case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_3].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_4].pipeline; break; - case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_1_F32_R1_5].pipeline; break; - default: GGML_ABORT("not implemented"); - } break; - case GGML_TYPE_Q8_0: - switch (r1ptg) { - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_2].pipeline; break; - case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_3].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_4].pipeline; break; - case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q8_0_F32_R1_5].pipeline; break; - default: GGML_ABORT("not implemented"); - } break; - case GGML_TYPE_MXFP4: - switch (r1ptg) { - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_MXFP4_F32_R1_2].pipeline; break; - case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_MXFP4_F32_R1_3].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_MXFP4_F32_R1_4].pipeline; break; - case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_MXFP4_F32_R1_5].pipeline; break; - default: GGML_ABORT("not implemented"); - } break; - case GGML_TYPE_Q4_K: - switch (r1ptg) { - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_2].pipeline; break; - case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_3].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_4].pipeline; break; - case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q4_K_F32_R1_5].pipeline; break; - default: GGML_ABORT("not implemented"); - } break; - case GGML_TYPE_Q5_K: - switch (r1ptg) { - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_2].pipeline; break; - case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_3].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_4].pipeline; break; - case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q5_K_F32_R1_5].pipeline; break; - default: GGML_ABORT("not implemented"); - } break; - case GGML_TYPE_Q6_K: - switch (r1ptg) { - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_2].pipeline; break; - case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_3].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_4].pipeline; break; - case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_Q6_K_F32_R1_5].pipeline; break; - default: GGML_ABORT("not implemented"); - } break; - case GGML_TYPE_IQ4_NL: - switch (r1ptg) { - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_2].pipeline; break; - case 3: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_3].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_4].pipeline; break; - case 5: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_EXT_IQ4_NL_F32_R1_5].pipeline; break; - default: GGML_ABORT("not implemented"); - } break; - default: GGML_ABORT("not implemented"); - } - - ggml_metal_kargs_mul_mv_ext args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne10 =*/ ne10, - /*.ne11 =*/ ne11, - /*.ne12 =*/ ne12, - /*.nb10 =*/ nb10, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb13 =*/ nb13, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.r2 =*/ r2, - /*.r3 =*/ r3, - /*.nsg =*/ nsg, - /*.nxpsg =*/ nxpsg, - /*.r1ptg =*/ r1ptg, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; - - //printf("ne01 = %lld nr0ptg = %d\n", ne01, nr0ptg); - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + r0ptg - 1)/r0ptg, (ne11 + r1ptg - 1)/r1ptg, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(32, nsg, 1)]; - } else - // for now the matrix-matrix multiplication kernel only works on A14+/M1+ SoCs - // AMD GPU and older A-chips will reuse matrix-vector multiplication kernel - if ([device supportsFamily:MTLGPUFamilyApple7] && - !ggml_is_transposed(src0) && - !ggml_is_transposed(src1) && - src1t == GGML_TYPE_F32 && - ne00 % 32 == 0 && ne00 >= 64 && - (ne11 > ne11_mm_min || (ggml_is_quantized(src0t) && ne12 > 1))) { - //printf("matrix: ne00 = %6d, ne01 = %6d, ne02 = %6d, ne11 = %6d, ne12 = %6d\n", ne00, ne01, ne02, ne11, ne12); - - // some Metal matrix data types require aligned pointers - // ref: https://developer.apple.com/metal/Metal-Shading-Language-Specification.pdf (Table 2.5) - switch (src0->type) { - case GGML_TYPE_F32: GGML_ASSERT(nb01 % 16 == 0); break; - case GGML_TYPE_F16: GGML_ASSERT(nb01 % 8 == 0); break; - case GGML_TYPE_BF16: GGML_ASSERT(nb01 % 8 == 0); break; - default: break; - } - - id pipeline = nil; - - switch (src0->type) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_F32_F32 ].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_F16_F32 ].pipeline; break; - case GGML_TYPE_BF16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_BF16_F32 ].pipeline; break; - case GGML_TYPE_Q4_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_0_F32 ].pipeline; break; - case GGML_TYPE_Q4_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_1_F32 ].pipeline; break; - case GGML_TYPE_Q5_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_0_F32 ].pipeline; break; - case GGML_TYPE_Q5_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_1_F32 ].pipeline; break; - case GGML_TYPE_Q8_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q8_0_F32 ].pipeline; break; - case GGML_TYPE_MXFP4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_MXFP4_F32 ].pipeline; break; - case GGML_TYPE_Q2_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q2_K_F32 ].pipeline; break; - case GGML_TYPE_Q3_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q3_K_F32 ].pipeline; break; - case GGML_TYPE_Q4_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q4_K_F32 ].pipeline; break; - case GGML_TYPE_Q5_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q5_K_F32 ].pipeline; break; - case GGML_TYPE_Q6_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_Q6_K_F32 ].pipeline; break; - case GGML_TYPE_IQ2_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XXS_F32].pipeline; break; - case GGML_TYPE_IQ2_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_XS_F32 ].pipeline; break; - case GGML_TYPE_IQ3_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_XXS_F32].pipeline; break; - case GGML_TYPE_IQ3_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ3_S_F32 ].pipeline; break; - case GGML_TYPE_IQ2_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ2_S_F32 ].pipeline; break; - case GGML_TYPE_IQ1_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ1_S_F32 ].pipeline; break; - case GGML_TYPE_IQ1_M: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ1_M_F32 ].pipeline; break; - case GGML_TYPE_IQ4_NL: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ4_NL_F32 ].pipeline; break; - case GGML_TYPE_IQ4_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_IQ4_XS_F32 ].pipeline; break; - default: GGML_ABORT("MUL MAT-MAT not implemented"); - } - - ggml_metal_kargs_mul_mm args = { - /*.ne00 =*/ ne00, - /*.ne02 =*/ ne02, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne12 =*/ ne12, - /*.nb10 =*/ nb10, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb13 =*/ nb13, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.r2 =*/ r2, - /*.r3 =*/ r3, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; - - [encoder setThreadgroupMemoryLength:8192 atIndex:0]; - [encoder dispatchThreadgroups:MTLSizeMake((ne11 + 31)/32, (ne01 + 63)/64, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(128, 1, 1)]; - } else { - id pipeline = nil; - - int nsg = 0; // number of simdgroups - int nr0 = 0; // number of src0 rows per simdgroup - int nr1 = 1; // number of src1 rows per threadgroup - - size_t smem = 0; // shared memory - - // use custom matrix x vector kernel - switch (src0t) { - case GGML_TYPE_F32: - { - GGML_ASSERT(src1t == GGML_TYPE_F32); - nsg = 1; - nr0 = 1; - nr1 = 4; - if (ne00 == 4) { - nr0 = 32; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F32_F32_C4].pipeline; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F32_F32].pipeline; - } - } break; - case GGML_TYPE_F16: - { - nsg = 1; - nr0 = 1; - if (src1t == GGML_TYPE_F32) { - if (ne00 == 4) { - nr0 = 32; - nr1 = 4; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_C4].pipeline; - } else if (ne11 * ne12 < 4) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_1ROW].pipeline; - } else if (ne00 >= 128 && ne01 >= 8 && ne00%4 == 0) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32_L4].pipeline; - nr1 = ne11; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F32].pipeline; - nr1 = 4; - } - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_F16_F16].pipeline; - nr1 = 4; - } - } break; - case GGML_TYPE_BF16: - { - nsg = 1; - nr0 = 1; - if (src1t == GGML_TYPE_F32) { - if (ne00 == 4) { - nr0 = 32; - nr1 = 4; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_C4].pipeline; - } else if (ne11 * ne12 < 4) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_1ROW].pipeline; - } else if (ne00 >= 128 && ne01 >= 8 && ne00%4 == 0) { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32_L4].pipeline; - nr1 = ne11; - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_F32].pipeline; - nr1 = 4; - } - } else { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_BF16_BF16].pipeline; - nr1 = 4; - } - } break; - case GGML_TYPE_Q4_0: - { - nsg = N_SG_Q4_0; - nr0 = N_R0_Q4_0; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_0_F32].pipeline; - } break; - case GGML_TYPE_Q4_1: - { - nsg = N_SG_Q4_1; - nr0 = N_R0_Q4_1; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_1_F32].pipeline; - } break; - case GGML_TYPE_Q5_0: - { - nsg = N_SG_Q5_0; - nr0 = N_R0_Q5_0; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_0_F32].pipeline; - } break; - case GGML_TYPE_Q5_1: - { - nsg = N_SG_Q5_1; - nr0 = N_R0_Q5_1; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_1_F32].pipeline; - } break; - case GGML_TYPE_Q8_0: - { - nsg = N_SG_Q8_0; - nr0 = N_R0_Q8_0; - smem = 32*sizeof(float)*N_R0_Q8_0; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q8_0_F32].pipeline; - } break; - case GGML_TYPE_MXFP4: - { - nsg = N_SG_MXFP4; - nr0 = N_R0_MXFP4; - smem = 32*sizeof(float); - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_MXFP4_F32].pipeline; - } break; - case GGML_TYPE_Q2_K: - { - nsg = N_SG_Q2_K; - nr0 = N_R0_Q2_K; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q2_K_F32].pipeline; - } break; - case GGML_TYPE_Q3_K: - { - nsg = N_SG_Q3_K; - nr0 = N_R0_Q3_K; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q3_K_F32].pipeline; - } break; - case GGML_TYPE_Q4_K: - { - nsg = N_SG_Q4_K; - nr0 = N_R0_Q4_K; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q4_K_F32].pipeline; - } break; - case GGML_TYPE_Q5_K: - { - nsg = N_SG_Q5_K; - nr0 = N_R0_Q5_K; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q5_K_F32].pipeline; - } break; - case GGML_TYPE_Q6_K: - { - nsg = N_SG_Q6_K; - nr0 = N_R0_Q6_K; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_Q6_K_F32].pipeline; - } break; - case GGML_TYPE_IQ2_XXS: - { - nsg = N_SG_IQ2_XXS; - nr0 = N_R0_IQ2_XXS; - smem = 256*8+128; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XXS_F32].pipeline; - } break; - case GGML_TYPE_IQ2_XS: - { - nsg = N_SG_IQ2_XS; - nr0 = N_R0_IQ2_XS; - smem = 512*8+128; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_XS_F32].pipeline; - } break; - case GGML_TYPE_IQ3_XXS: - { - nsg = N_SG_IQ3_XXS; - nr0 = N_R0_IQ3_XXS; - smem = 256*4+128; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ3_XXS_F32].pipeline; - } break; - case GGML_TYPE_IQ3_S: - { - nsg = N_SG_IQ3_S; - nr0 = N_R0_IQ3_S; - smem = 512*4; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ3_S_F32].pipeline; - } break; - case GGML_TYPE_IQ2_S: - { - nsg = N_SG_IQ2_S; - nr0 = N_R0_IQ2_S; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ2_S_F32].pipeline; - } break; - case GGML_TYPE_IQ1_S: - { - nsg = N_SG_IQ1_S; - nr0 = N_R0_IQ1_S; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ1_S_F32].pipeline; - } break; - case GGML_TYPE_IQ1_M: - { - nsg = N_SG_IQ1_M; - nr0 = N_R0_IQ1_M; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ1_M_F32].pipeline; - } break; - case GGML_TYPE_IQ4_NL: - { - nsg = N_SG_IQ4_NL; - nr0 = N_R0_IQ4_NL; - smem = 32*sizeof(float); - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ4_NL_F32].pipeline; - } break; - case GGML_TYPE_IQ4_XS: - { - nsg = N_SG_IQ4_XS; - nr0 = N_R0_IQ4_XS; - smem = 32*sizeof(float); - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_IQ4_XS_F32].pipeline; - } break; - default: - { - GGML_LOG_ERROR("Asserting on type %d\n", (int)src0t); - GGML_ABORT("not implemented"); - } - }; - - ggml_metal_kargs_mul_mv args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne10 =*/ ne10, - /*.ne11 =*/ ne11, - /*.ne12 =*/ ne12, - /*.nb10 =*/ nb10, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb13 =*/ nb13, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.r2 =*/ r2, - /*.r3 =*/ r3, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; - - if (smem > 0) { - [encoder setThreadgroupMemoryLength:smem atIndex:0]; - } - - if (src0t == GGML_TYPE_Q8_0) { - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + nr0 - 1)/(nr0), (ne11 + nr1 - 1)/nr1, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(32, nsg, 1)]; - } else { - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + nr0*nsg - 1)/(nr0*nsg), (ne11 + nr1 - 1)/nr1, ne12*ne13) threadsPerThreadgroup:MTLSizeMake(32, nsg, 1)]; - } - } - } break; - case GGML_OP_MUL_MAT_ID: - { - // src2 = ids - GGML_ASSERT(src2t == GGML_TYPE_I32); - - GGML_ASSERT(!ggml_is_transposed(src0)); - GGML_ASSERT(!ggml_is_transposed(src1)); - - GGML_ASSERT(src1t == GGML_TYPE_F32); - - GGML_ASSERT(ne03 == 1); - GGML_ASSERT(ne13 == 1); - - const uint32_t r2 = 1; - const uint32_t r3 = 1; - - // find the break-even point where the matrix-matrix kernel becomes more efficient compared - // to the matrix-vector kernel - // ne20 = n_used_experts - // ne21 = n_rows (batch size) - const int ne21_mm_id_min = 32; - - // for now the matrix-matrix multiplication kernel only works on A14+/M1+ SoCs - // AMD GPU and older A-chips will reuse matrix-vector multiplication kernel - if ([device supportsFamily:MTLGPUFamilyApple7] && - ne00 % 32 == 0 && ne00 >= 64 && - (ne21 >= ne21_mm_id_min)) { - GGML_ASSERT(ne00 % 4 == 0); - - // some Metal matrix data types require aligned pointers - // ref: https://developer.apple.com/metal/Metal-Shading-Language-Specification.pdf (Table 2.5) - switch (src0->type) { - case GGML_TYPE_F32: GGML_ASSERT(nb01 % 16 == 0); break; - case GGML_TYPE_F16: GGML_ASSERT(nb01 % 8 == 0); break; - case GGML_TYPE_BF16: GGML_ASSERT(nb01 % 8 == 0); break; - default: break; - } - - // extra buffers for intermediate id mapping - size_t offs_tpe = offs_dst + ggml_nbytes(dst); - size_t offs_ids = offs_tpe + ggml_metal_mul_mat_id_extra_tpe(dst); - - { - ggml_metal_kargs_mul_mm_id_map0 args = { - ne02, - ne10, - ne11, // n_expert_used (bcast) - nb11, - nb12, - ne21, // n_tokens - ne20, // n_expert_used - nb21, - }; - - id pipeline = nil; - - pipeline = nil; - - switch (ne20) { - case 1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_1 ].pipeline; break; - case 2: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_2 ].pipeline; break; - case 4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_4 ].pipeline; break; - case 6: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_6 ].pipeline; break; - case 8: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_8 ].pipeline; break; - case 10: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_10].pipeline; break; - case 16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MAP0_F16_NE20_16].pipeline; break; - default: GGML_ABORT("missing specialization for ne20 = %d", (int) ne20); - } - - GGML_ASSERT(ne02 <= (int) pipeline.maxTotalThreadsPerThreadgroup); - - const size_t smem = ne02*ne20*sizeof(uint16_t); - - GGML_ASSERT(smem <= device.maxThreadgroupMemoryLength); - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_tpe atIndex:2]; - [encoder setBuffer:id_dst offset:offs_ids atIndex:3]; - [encoder setThreadgroupMemoryLength:smem atIndex:0]; - - [encoder dispatchThreadgroups:MTLSizeMake(1, 1, 1) threadsPerThreadgroup:MTLSizeMake(ne02, 1, 1)]; - } - - // this barrier is always needed because the next kernel has to wait for the id maps to be computed - ggml_metal_encode_concurrency_reset(ctx_enc); - - { - id pipeline = nil; - - switch (src0->type) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_F32_F16 ].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_F16_F16 ].pipeline; break; - case GGML_TYPE_BF16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_BF16_F16 ].pipeline; break; - case GGML_TYPE_Q4_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_0_F16 ].pipeline; break; - case GGML_TYPE_Q4_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_1_F16 ].pipeline; break; - case GGML_TYPE_Q5_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_0_F16 ].pipeline; break; - case GGML_TYPE_Q5_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_1_F16 ].pipeline; break; - case GGML_TYPE_Q8_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q8_0_F16 ].pipeline; break; - case GGML_TYPE_MXFP4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_MXFP4_F16 ].pipeline; break; - case GGML_TYPE_Q2_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q2_K_F16 ].pipeline; break; - case GGML_TYPE_Q3_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q3_K_F16 ].pipeline; break; - case GGML_TYPE_Q4_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q4_K_F16 ].pipeline; break; - case GGML_TYPE_Q5_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q5_K_F16 ].pipeline; break; - case GGML_TYPE_Q6_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_Q6_K_F16 ].pipeline; break; - case GGML_TYPE_IQ2_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_XXS_F16].pipeline; break; - case GGML_TYPE_IQ2_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_XS_F16 ].pipeline; break; - case GGML_TYPE_IQ3_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ3_XXS_F16].pipeline; break; - case GGML_TYPE_IQ3_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ3_S_F16 ].pipeline; break; - case GGML_TYPE_IQ2_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ2_S_F16 ].pipeline; break; - case GGML_TYPE_IQ1_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ1_S_F16 ].pipeline; break; - case GGML_TYPE_IQ1_M: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ1_M_F16 ].pipeline; break; - case GGML_TYPE_IQ4_NL: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ4_NL_F16 ].pipeline; break; - case GGML_TYPE_IQ4_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MM_ID_IQ4_XS_F16 ].pipeline; break; - default: GGML_ABORT("MUL_MAT_ID not implemented"); - } - - ggml_metal_kargs_mul_mm_id args = { - /*.ne00 =*/ ne00, - /*.ne02 =*/ ne02, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne11 =*/ ne11, // n_expert_used (bcast) - /*.nb10 =*/ nb10, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb13 =*/ nb13, - /*.ne20 =*/ ne20, // n_expert_used - /*.ne21 =*/ ne21, // n_tokens - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.r2 =*/ r2, - /*.r3 =*/ r3, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; - [encoder setBuffer:id_dst offset:offs_tpe atIndex:3]; - [encoder setBuffer:id_dst offset:offs_ids atIndex:4]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:5]; - - [encoder setThreadgroupMemoryLength:8192 atIndex:0]; - [encoder dispatchThreadgroups:MTLSizeMake((ne21 + 31)/32, (ne01 + 63)/64, ne02) threadsPerThreadgroup:MTLSizeMake(128, 1, 1)]; - } - } else { - id pipeline = nil; - - int nsg = 0; // number of simdgroups - int nr0 = 0; // number of src0 rows per simdgroup - int nr1 = 1; // number of src1 rows per threadgroup - - size_t smem = 0; // shared memory - - // use custom matrix x vector kernel - switch (src0t) { - case GGML_TYPE_F32: - { - GGML_ASSERT(src1t == GGML_TYPE_F32); - nsg = 1; - nr0 = 1; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F32_F32].pipeline; - } break; - case GGML_TYPE_F16: - { - GGML_ASSERT(src1t == GGML_TYPE_F32); - nsg = 1; - nr0 = 1; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_F16_F32].pipeline; - } break; - case GGML_TYPE_BF16: - { - GGML_ASSERT(src1t == GGML_TYPE_F32); - nsg = 1; - nr0 = 1; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_BF16_F32].pipeline; - } break; - case GGML_TYPE_Q4_0: - { - nsg = N_SG_Q4_0; - nr0 = N_R0_Q4_0; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_0_F32].pipeline; - } break; - case GGML_TYPE_Q4_1: - { - nsg = N_SG_Q4_1; - nr0 = N_R0_Q4_1; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_1_F32].pipeline; - } break; - case GGML_TYPE_Q5_0: - { - nsg = N_SG_Q5_0; - nr0 = N_R0_Q5_0; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q5_0_F32].pipeline; - } break; - case GGML_TYPE_Q5_1: - { - nsg = N_SG_Q5_1; - nr0 = N_R0_Q5_1; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q5_1_F32].pipeline; - } break; - case GGML_TYPE_Q8_0: - { - nsg = N_SG_Q8_0; - nr0 = N_R0_Q8_0; - smem = 32*sizeof(float)*N_R0_Q8_0; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q8_0_F32].pipeline; - } break; - case GGML_TYPE_MXFP4: - { - nsg = N_SG_MXFP4; - nr0 = N_R0_MXFP4; - smem = 32*sizeof(float); - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_MXFP4_F32].pipeline; - } break; - case GGML_TYPE_Q2_K: - { - nsg = N_SG_Q2_K; - nr0 = N_R0_Q2_K; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q2_K_F32].pipeline; - } break; - case GGML_TYPE_Q3_K: - { - nsg = N_SG_Q3_K; - nr0 = N_R0_Q3_K; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q3_K_F32].pipeline; - } break; - case GGML_TYPE_Q4_K: - { - nsg = N_SG_Q4_K; - nr0 = N_R0_Q4_K; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q4_K_F32].pipeline; - } break; - case GGML_TYPE_Q5_K: - { - nsg = N_SG_Q5_K; - nr0 = N_R0_Q5_K; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q5_K_F32].pipeline; - } break; - case GGML_TYPE_Q6_K: - { - nsg = N_SG_Q6_K; - nr0 = N_R0_Q6_K; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_Q6_K_F32].pipeline; - } break; - case GGML_TYPE_IQ2_XXS: - { - nsg = N_SG_IQ2_XXS; - nr0 = N_R0_IQ2_XXS; - smem = 256*8+128; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_XXS_F32].pipeline; - } break; - case GGML_TYPE_IQ2_XS: - { - nsg = N_SG_IQ2_XS; - nr0 = N_R0_IQ2_XS; - smem = 512*8+128; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_XS_F32].pipeline; - } break; - case GGML_TYPE_IQ3_XXS: - { - nsg = N_SG_IQ3_XXS; - nr0 = N_R0_IQ3_XXS; - smem = 256*4+128; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ3_XXS_F32].pipeline; - } break; - case GGML_TYPE_IQ3_S: - { - nsg = N_SG_IQ3_S; - nr0 = N_R0_IQ3_S; - smem = 512*4; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ3_S_F32].pipeline; - } break; - case GGML_TYPE_IQ2_S: - { - nsg = N_SG_IQ2_S; - nr0 = N_R0_IQ2_S; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ2_S_F32].pipeline; - } break; - case GGML_TYPE_IQ1_S: - { - nsg = N_SG_IQ1_S; - nr0 = N_R0_IQ1_S; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ1_S_F32].pipeline; - } break; - case GGML_TYPE_IQ1_M: - { - nsg = N_SG_IQ1_M; - nr0 = N_R0_IQ1_M; - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ1_M_F32].pipeline; - } break; - case GGML_TYPE_IQ4_NL: - { - nsg = N_SG_IQ4_NL; - nr0 = N_R0_IQ4_NL; - smem = 32*sizeof(float); - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ4_NL_F32].pipeline; - } break; - case GGML_TYPE_IQ4_XS: - { - nsg = N_SG_IQ4_XS; - nr0 = N_R0_IQ4_XS; - smem = 32*sizeof(float); - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_MUL_MV_ID_IQ4_XS_F32].pipeline; - } break; - default: - { - GGML_LOG_ERROR("Asserting on type %d\n", (int)src2t); - GGML_ABORT("not implemented"); - } - }; - - if (ggml_is_quantized(src0t)) { - GGML_ASSERT(ne00 >= nsg*nr0); - } - - ggml_metal_kargs_mul_mv_id args = { - /*.nei0 =*/ ne20, - /*.nei1 =*/ ne21, - /*.nbi1 =*/ nb21, - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.ne10 =*/ ne10, - /*.ne11 =*/ ne11, - /*.ne12 =*/ ne12, - /*.ne13 =*/ ne13, - /*.nb10 =*/ nb10, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.nb1 =*/ nb1, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:4]; - - const int64_t _ne1 = 1; - const int64_t ne123 = ne20*ne21; - - if (smem > 0) { - [encoder setThreadgroupMemoryLength:smem atIndex:0]; - } - - if (src0t == GGML_TYPE_Q8_0) { - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + nr0 - 1)/(nr0), (_ne1 + nr1 - 1)/nr1, ne123) threadsPerThreadgroup:MTLSizeMake(32, nsg, 1)]; - } else { - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + nr0*nsg - 1)/(nr0*nsg), (_ne1 + nr1 - 1)/nr1, ne123) threadsPerThreadgroup:MTLSizeMake(32, nsg, 1)]; - } - } - } break; - case GGML_OP_GET_ROWS: - { - id pipeline = nil; - - switch (src0->type) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_F32 ].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_F16 ].pipeline; break; - case GGML_TYPE_BF16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_BF16 ].pipeline; break; - case GGML_TYPE_Q4_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_0 ].pipeline; break; - case GGML_TYPE_Q4_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_1 ].pipeline; break; - case GGML_TYPE_Q5_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_0 ].pipeline; break; - case GGML_TYPE_Q5_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_1 ].pipeline; break; - case GGML_TYPE_Q8_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q8_0 ].pipeline; break; - case GGML_TYPE_MXFP4: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_MXFP4 ].pipeline; break; - case GGML_TYPE_Q2_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q2_K ].pipeline; break; - case GGML_TYPE_Q3_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q3_K ].pipeline; break; - case GGML_TYPE_Q4_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q4_K ].pipeline; break; - case GGML_TYPE_Q5_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q5_K ].pipeline; break; - case GGML_TYPE_Q6_K: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_Q6_K ].pipeline; break; - case GGML_TYPE_IQ2_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_XXS].pipeline; break; - case GGML_TYPE_IQ2_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_XS ].pipeline; break; - case GGML_TYPE_IQ3_XXS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ3_XXS].pipeline; break; - case GGML_TYPE_IQ3_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ3_S ].pipeline; break; - case GGML_TYPE_IQ2_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ2_S ].pipeline; break; - case GGML_TYPE_IQ1_S: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ1_S ].pipeline; break; - case GGML_TYPE_IQ1_M: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ1_M ].pipeline; break; - case GGML_TYPE_IQ4_NL: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ4_NL ].pipeline; break; - case GGML_TYPE_IQ4_XS: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_IQ4_XS ].pipeline; break; - case GGML_TYPE_I32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GET_ROWS_I32 ].pipeline; break; - default: GGML_ABORT("not implemented"); - } - - ggml_metal_kargs_get_rows args = { - /*.ne00 =*/ ne00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.ne10 =*/ ne10, - /*.nb10 =*/ nb10, - /*.nb11 =*/ nb11, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; - - [encoder dispatchThreadgroups:MTLSizeMake(ne10, ne11, 1) threadsPerThreadgroup:MTLSizeMake(32, 1, 1)]; - } break; - case GGML_OP_SET_ROWS: - { - id pipeline = nil; - - switch (dst->type) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SET_ROWS_F32 ].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SET_ROWS_F16 ].pipeline; break; - case GGML_TYPE_BF16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SET_ROWS_BF16 ].pipeline; break; - case GGML_TYPE_Q8_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SET_ROWS_Q8_0 ].pipeline; break; - case GGML_TYPE_Q4_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SET_ROWS_Q4_0 ].pipeline; break; - case GGML_TYPE_Q4_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SET_ROWS_Q4_1 ].pipeline; break; - case GGML_TYPE_Q5_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SET_ROWS_Q5_0 ].pipeline; break; - case GGML_TYPE_Q5_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SET_ROWS_Q5_1 ].pipeline; break; - case GGML_TYPE_IQ4_NL: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_SET_ROWS_IQ4_NL].pipeline; break; - default: GGML_ABORT("not implemented"); - } - - const int32_t nk0 = ne0/ggml_blck_size(dst->type); - - int nth = 32; // SIMD width - - while (nth < nk0 && nth < (int) pipeline.maxTotalThreadsPerThreadgroup) { - nth *= 2; - } - - int nrptg = 1; - if (nth > nk0) { - nrptg = (nth + nk0 - 1)/nk0; - nth = nk0; - - if (nrptg*nth > (int) pipeline.maxTotalThreadsPerThreadgroup) { - nrptg--; - } - } - - nth = MIN(nth, nk0); - - ggml_metal_kargs_set_rows args = { - /*.nk0 =*/ nk0, - /*.ne01 =*/ ne01, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne11 =*/ ne11, - /*.ne12 =*/ ne12, - /*.nb10 =*/ nb10, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:3]; - - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + nrptg - 1)/nrptg, ne02, ne03) threadsPerThreadgroup:MTLSizeMake(nth, nrptg, 1)]; - } break; - case GGML_OP_RMS_NORM: - { - GGML_ASSERT(ne00 % 4 == 0); - GGML_ASSERT(ggml_is_contiguous_rows(src0)); - - float eps; - memcpy(&eps, dst->op_params, sizeof(float)); - - ggml_metal_kargs_rms_norm args = { - /*.ne00 =*/ ne00, - /*.ne00_4 =*/ ne00/4, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3, - /*.eps =*/ eps, - /*.nef1 =*/ { ne01 }, - /*.nef2 =*/ { ne02 }, - /*.nef3 =*/ { ne03 }, - /*.nbf1 =*/ { nb01 }, - /*.nbf2 =*/ { nb02 }, - /*.nbf3 =*/ { nb03 }, - }; - - size_t offs_fuse[2] = { 0, 0 }; - id id_fuse[2] = { id_src0, id_src0 }; - - // d[0] = rms_norm(a) - // d[1] = mul(d[0], b) - // d[2] = add(d[1], c) - if (ctx_dev->use_fusion) { - ops[0] = GGML_OP_RMS_NORM; - ops[1] = GGML_OP_MUL; - ops[2] = GGML_OP_ADD; - - for (n_fuse = 0; n_fuse <= 1 && idx + n_fuse + 1 < idx_end; ++n_fuse) { - if (!ggml_can_fuse(gf, idx + n_fuse, ops + n_fuse, 2)) { - break; - } - - if (nodes[n_fuse] != nodes[n_fuse + 1]->src[0]) { - break; - } - - if (nodes[n_fuse + 1]->src[1]->ne[0] != node->ne[0]) { - break; - } - - if (!ggml_is_contiguous_rows(nodes[n_fuse + 1]->src[1])) { - break; - } - - if (nodes[n_fuse + 1]->type != GGML_TYPE_F32) { - break; - } - - ctx_dev->fuse_cnt[nodes[n_fuse + 1]->op]++; - - id_fuse[n_fuse] = ggml_metal_get_buffer(nodes[n_fuse + 1]->src[1], &offs_fuse[n_fuse]); - - args.nef1[n_fuse + 1] = nodes[n_fuse + 1]->src[1]->ne[1]; - args.nef2[n_fuse + 1] = nodes[n_fuse + 1]->src[1]->ne[2]; - args.nef3[n_fuse + 1] = nodes[n_fuse + 1]->src[1]->ne[3]; - - args.nbf1[n_fuse + 1] = nodes[n_fuse + 1]->src[1]->nb[1]; - args.nbf2[n_fuse + 1] = nodes[n_fuse + 1]->src[1]->nb[2]; - args.nbf3[n_fuse + 1] = nodes[n_fuse + 1]->src[1]->nb[3]; - } - - ++n_fuse; - - if (ctx_dev->debug_fusion > 1 && n_fuse > 1) { - if (n_fuse == 2) { - GGML_LOG_DEBUG("%s: fuse: RMS_NORM + MUL\n", __func__); - } - if (n_fuse == 3) { - GGML_LOG_DEBUG("%s: fuse: RMS_NORM + MUL + ADD\n", __func__); - } - } - } - - if (n_fuse > 1) { - id_dst = ggml_metal_get_buffer(nodes[n_fuse - 1], &offs_dst); - - for (int i = 1; i < n_fuse; ++i) { - if (!ggml_metal_encode_concurrency_check(ctx_enc, nodes[i])) { - ggml_metal_encode_concurrency_reset(ctx_enc); - - break; - } - } - } - - const id pipeline = ggml_metal_get_pipeline_rms_norm(backend, node, n_fuse); - - int nth = 32; // SIMD width - - while (nth < ne00/4 && nth < (int) pipeline.maxTotalThreadsPerThreadgroup) { - nth *= 2; - } - - nth = MIN(nth, (int) pipeline.maxTotalThreadsPerThreadgroup); - nth = MIN(nth, ne00/4); - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_fuse[0] offset:offs_fuse[0] atIndex:2]; - [encoder setBuffer:id_fuse[1] offset:offs_fuse[1] atIndex:3]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:4]; - - [encoder setThreadgroupMemoryLength:32*sizeof(float) atIndex:0]; - - [encoder dispatchThreadgroups:MTLSizeMake(ne01, ne02, ne03) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_L2_NORM: - { - GGML_ASSERT(ne00 % 4 == 0); - GGML_ASSERT(ggml_is_contiguous_1(src0)); - - float eps; - memcpy(&eps, dst->op_params, sizeof(float)); - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_L2_NORM].pipeline; - - int nth = 32; // SIMD width - - while (nth < ne00/4 && nth < (int) pipeline.maxTotalThreadsPerThreadgroup) { - nth *= 2; - } - - nth = MIN(nth, (int) pipeline.maxTotalThreadsPerThreadgroup); - nth = MIN(nth, ne00/4); - - ggml_metal_kargs_l2_norm args = { - /*.ne00 =*/ ne00, - /*.ne00_4 =*/ ne00/4, - /*.nb01 =*/ nb01, - /*.eps =*/ eps, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - - [encoder setThreadgroupMemoryLength:32*sizeof(float) atIndex:0]; - - const int64_t nrows = ggml_nrows(src0); - - [encoder dispatchThreadgroups:MTLSizeMake(nrows, 1, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_GROUP_NORM: - { - GGML_ASSERT(ggml_is_contiguous(src0)); - - float eps; - memcpy(&eps, dst->op_params + 1, sizeof(float)); - - const int32_t n_groups = ((const int32_t *) dst->op_params)[0]; - - int nth = 32; // SIMD width - - //while (nth < ne00/4 && nth < 1024) { - // nth *= 2; - //} - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_GROUP_NORM].pipeline; - - ggml_metal_kargs_group_norm args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.n_groups =*/ n_groups, - /*.eps =*/ eps, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&args length:sizeof(args) atIndex:2]; - [encoder setThreadgroupMemoryLength:32*sizeof(float) atIndex:0]; - - [encoder dispatchThreadgroups:MTLSizeMake(n_groups, 1, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_NORM: - { - GGML_ASSERT(ne00 % 4 == 0); - GGML_ASSERT(ggml_is_contiguous_1(src0)); - - float eps; - memcpy(&eps, dst->op_params, sizeof(float)); - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_NORM].pipeline; - - int nth = 32; // SIMD width - - while (nth < ne00/4 && nth < (int) pipeline.maxTotalThreadsPerThreadgroup) { - nth *= 2; - } - - nth = MIN(nth, (int) pipeline.maxTotalThreadsPerThreadgroup); - nth = MIN(nth, ne00/4); - - ggml_metal_kargs_norm args = { - /*.ne00 =*/ ne00, - /*.ne00_4 =*/ ne00/4, - /*.nb01 =*/ nb01, - /*.eps =*/ eps, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - - [encoder setThreadgroupMemoryLength:32*sizeof(float) atIndex:0]; - - const int64_t nrows = ggml_nrows(src0); - - [encoder dispatchThreadgroups:MTLSizeMake(nrows, 1, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_ROPE: - { - // make sure we have one or more position id(ne10) per token(ne02) - GGML_ASSERT(ne10 % ne02 == 0); - GGML_ASSERT(ne10 >= ne02); - - const int nth = MIN(1024, ne00); - - const int n_past = ((const int32_t *) dst->op_params)[0]; - const int n_dims = ((const int32_t *) dst->op_params)[1]; - const int mode = ((const int32_t *) dst->op_params)[2]; - // skip 3, n_ctx, used in GLM RoPE, unimplemented in metal - const int n_ctx_orig = ((const int32_t *) dst->op_params)[4]; - - float freq_base; - float freq_scale; - float ext_factor; - float attn_factor; - float beta_fast; - float beta_slow; - - memcpy(&freq_base, (const int32_t *) dst->op_params + 5, sizeof(float)); - memcpy(&freq_scale, (const int32_t *) dst->op_params + 6, sizeof(float)); - memcpy(&ext_factor, (const int32_t *) dst->op_params + 7, sizeof(float)); - memcpy(&attn_factor, (const int32_t *) dst->op_params + 8, sizeof(float)); - memcpy(&beta_fast, (const int32_t *) dst->op_params + 9, sizeof(float)); - memcpy(&beta_slow, (const int32_t *) dst->op_params + 10, sizeof(float)); - - const bool is_neox = mode & GGML_ROPE_TYPE_NEOX; - const bool is_mrope = mode & GGML_ROPE_TYPE_MROPE; - const bool is_vision = mode == GGML_ROPE_TYPE_VISION; - - // mrope - const int sect_0 = ((const int32_t *) dst->op_params)[11]; - const int sect_1 = ((const int32_t *) dst->op_params)[12]; - const int sect_2 = ((const int32_t *) dst->op_params)[13]; - const int sect_3 = ((const int32_t *) dst->op_params)[14]; - - id pipeline = nil; - - if (is_neox) { - switch (src0->type) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ROPE_NEOX_F32].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ROPE_NEOX_F16].pipeline; break; - default: GGML_ABORT("fatal error"); - }; - } else if (is_mrope && !is_vision) { - GGML_ASSERT(ne10*4 >= ne02); // need at least 4 pos per token - switch (src0->type) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ROPE_MULTI_F32].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ROPE_MULTI_F16].pipeline; break; - default: GGML_ABORT("fatal error"); - }; - } else if (is_vision) { - GGML_ASSERT(ne10*4 >= ne02); // need at least 4 pos per token - switch (src0->type) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ROPE_VISION_F32].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ROPE_VISION_F16].pipeline; break; - default: GGML_ABORT("fatal error"); - }; - } else { - switch (src0->type) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ROPE_NORM_F32].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ROPE_NORM_F16].pipeline; break; - default: GGML_ABORT("fatal error"); - }; - } - - ggml_metal_kargs_rope args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.nb0 =*/ nb0, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3, - /*.n_past =*/ n_past, - /*.n_dims =*/ n_dims, - /*.n_ctx_orig =*/ n_ctx_orig, - /*.freq_base =*/ freq_base, - /*.freq_scale =*/ freq_scale, - /*.ext_factor =*/ ext_factor, - /*.attn_factor =*/ attn_factor, - /*.beta_fast =*/ beta_fast, - /*.beta_slow =*/ beta_slow, - /* sect_0 =*/ sect_0, - /* sect_1 =*/ sect_1, - /* sect_2 =*/ sect_2, - /* sect_3 =*/ sect_3, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; - if (id_src2 != nil) { - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:3]; - } else { - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:3]; - } - [encoder setBuffer:id_dst offset:offs_dst atIndex:4]; - - [encoder dispatchThreadgroups:MTLSizeMake(ne01, ne02, ne03) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_IM2COL: - { - GGML_ASSERT(ggml_is_contiguous(src1)); - GGML_ASSERT(src1->type == GGML_TYPE_F32); - GGML_ASSERT( dst->type == GGML_TYPE_F16 || dst->type == GGML_TYPE_F32); - - const int32_t s0 = ((const int32_t *)(dst->op_params))[0]; - const int32_t s1 = ((const int32_t *)(dst->op_params))[1]; - const int32_t p0 = ((const int32_t *)(dst->op_params))[2]; - const int32_t p1 = ((const int32_t *)(dst->op_params))[3]; - const int32_t d0 = ((const int32_t *)(dst->op_params))[4]; - const int32_t d1 = ((const int32_t *)(dst->op_params))[5]; - - const bool is_2D = ((const int32_t *)(dst->op_params))[6] == 1; - - const int32_t N = src1->ne[is_2D ? 3 : 2]; - const int32_t IC = src1->ne[is_2D ? 2 : 1]; - const int32_t IH = is_2D ? src1->ne[1] : 1; - const int32_t IW = src1->ne[0]; - - const int32_t KH = is_2D ? src0->ne[1] : 1; - const int32_t KW = src0->ne[0]; - - const int32_t OH = is_2D ? dst->ne[2] : 1; - const int32_t OW = dst->ne[1]; - - const int32_t CHW = IC * KH * KW; - - const uint64_t ofs0 = src1->nb[is_2D ? 3 : 2] / 4; - const uint64_t ofs1 = src1->nb[is_2D ? 2 : 1] / 4; - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_IM2COL_F32].pipeline; - - const bool is_gt_mttpt = ((size_t)(N * KH * KW)) > pipeline.maxTotalThreadsPerThreadgroup; - - switch (dst->type) { - case GGML_TYPE_F32: { - pipeline = (is_gt_mttpt ? - ctx->kernels[GGML_METAL_KERNEL_TYPE_IM2COL_EXT_F32].pipeline - : - ctx->kernels[GGML_METAL_KERNEL_TYPE_IM2COL_F32].pipeline); - } break; - case GGML_TYPE_F16: { - pipeline = (is_gt_mttpt ? - ctx->kernels[GGML_METAL_KERNEL_TYPE_IM2COL_EXT_F16].pipeline - : - ctx->kernels[GGML_METAL_KERNEL_TYPE_IM2COL_F16].pipeline); - } break; - default: GGML_ABORT("fatal error"); - }; - - ggml_metal_kargs_im2col args = { - /*.ofs0 =*/ ofs0, - /*.ofs1 =*/ ofs1, - /*.IW =*/ IW, - /*.IH =*/ IH, - /*.CHW =*/ CHW, - /*.s0 =*/ s0, - /*.s1 =*/ s1, - /*.p0 =*/ p0, - /*.p1 =*/ p1, - /*.d0 =*/ d0, - /*.d1 =*/ d1, - /*.N =*/ N, - /*.KH =*/ KH, - /*.KW =*/ KW, - /*.KHW =*/ KH * KW, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&args length:sizeof(args) atIndex:2]; - - if (is_gt_mttpt) { - const uint64_t n_threads = MIN(pipeline.maxTotalThreadsPerThreadgroup, (uint64_t)N); - - const int64_t quotient = N / n_threads + (N % n_threads > 0 ? 1 : 0); - - [encoder dispatchThreadgroups:MTLSizeMake(quotient * CHW, OH, OW) threadsPerThreadgroup:MTLSizeMake(n_threads, 1, 1)]; - } else { - [encoder dispatchThreadgroups:MTLSizeMake(IC, OH, OW) threadsPerThreadgroup:MTLSizeMake(N, KH, KW)]; - } - } break; - case GGML_OP_CONV_TRANSPOSE_1D: - { - GGML_ASSERT(ggml_is_contiguous(src0)); - GGML_ASSERT(ggml_is_contiguous(src1)); - GGML_ASSERT(src0->type == GGML_TYPE_F16 || src0->type == GGML_TYPE_F32); - GGML_ASSERT(src1->type == GGML_TYPE_F32); - GGML_ASSERT( dst->type == GGML_TYPE_F32); - - const int32_t s0 = ((const int32_t *)(dst->op_params))[0]; - - const int32_t IC = src1->ne[1]; - const int32_t IL = src1->ne[0]; - - const int32_t K = src0->ne[0]; - - const int32_t OL = dst->ne[0]; - const int32_t OC = dst->ne[1]; - - id pipeline; - - switch (src0->type) { - case GGML_TYPE_F32: { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CONV_TRANSPOSE_1D_F32_F32].pipeline; - } break; - case GGML_TYPE_F16: { - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CONV_TRANSPOSE_1D_F16_F32].pipeline; - } break; - default: GGML_ABORT("fatal error"); - }; - - ggml_metal_kargs_conv_transpose_1d args = { - /*.IC =*/ IC, - /*.IL =*/ IL, - /*.K =*/ K, - /*.s0 =*/ s0, - /*.nb0 =*/ nb0, - /*.nb1 =*/ nb1, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - [encoder setBytes:&args length:sizeof(args) atIndex:3]; - - [encoder dispatchThreadgroups:MTLSizeMake(OL, OC, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_OP_UPSCALE: - { - GGML_ASSERT(src0->type == GGML_TYPE_F32); - - const float sf0 = (float)ne0/src0->ne[0]; - const float sf1 = (float)ne1/src0->ne[1]; - const float sf2 = (float)ne2/src0->ne[2]; - const float sf3 = (float)ne3/src0->ne[3]; - - const id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_UPSCALE_F32].pipeline; - - ggml_metal_kargs_upscale args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.nb0 =*/ nb0, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3, - /*.sf0 =*/ sf0, - /*.sf1 =*/ sf1, - /*.sf2 =*/ sf2, - /*.sf3 =*/ sf3 - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&args length:sizeof(args) atIndex:2]; - - const int nth = MIN((int) pipeline.maxTotalThreadsPerThreadgroup, ne0); - - [encoder dispatchThreadgroups:MTLSizeMake(ne1, ne2, ne3) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_PAD: - { - GGML_ASSERT(src0->type == GGML_TYPE_F32); - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_PAD_F32].pipeline; - - ggml_metal_kargs_pad args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.nb0 =*/ nb0, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3 - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&args length:sizeof(args) atIndex:2]; - - const int nth = MIN(1024, ne0); - - [encoder dispatchThreadgroups:MTLSizeMake(ne1, ne2, ne3) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_PAD_REFLECT_1D: - { - GGML_ASSERT(src0->type == GGML_TYPE_F32); - - const int32_t p0 = ((const int32_t *)(dst->op_params))[0]; - const int32_t p1 = ((const int32_t *)(dst->op_params))[1]; - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_PAD_REFLECT_1D_F32].pipeline; - - ggml_metal_kargs_pad_reflect_1d args = { - /*.ne00 =*/ ne00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.nb0 =*/ nb0, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3, - /*.p0 =*/ p0, - /*.p1 =*/ p1 - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&args length:sizeof(args) atIndex:2]; - - const int nth = MIN(1024, ne0); - - [encoder dispatchThreadgroups:MTLSizeMake(ne1, ne2, ne3) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_ARANGE: - { - GGML_ASSERT(dst->type == GGML_TYPE_F32); - - float start; - float step; - - memcpy(&start, ((const int32_t *) dst->op_params) + 0, sizeof(float)); - memcpy(&step, ((const int32_t *) dst->op_params) + 2, sizeof(float)); - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ARANGE_F32].pipeline; - - ggml_metal_kargs_arange args = { - /*.ne0 =*/ ne0, - /*.start =*/ start, - /*.step =*/ step - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:0]; - [encoder setBytes:&args length:sizeof(args) atIndex:1]; - - const int nth = MIN(1024, ne0); - - [encoder dispatchThreadgroups:MTLSizeMake(1, 1, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_TIMESTEP_EMBEDDING: - { - GGML_ASSERT(src0->type == GGML_TYPE_F32); - - const int dim = dst->op_params[0]; - const int max_period = dst->op_params[1]; - - const int half = dim / 2; - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_TIMESTEP_EMBEDDING_F32].pipeline; - - ggml_metal_kargs_timestep_embedding args = { - /*.nb1 =*/ nb1, - /*.dim =*/ dim, - /*.max_period =*/ max_period - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&args length:sizeof(args) atIndex:2]; - - const int nth = MIN(1024, half); - - [encoder dispatchThreadgroups:MTLSizeMake(ne00, 1, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - case GGML_OP_ARGSORT: - { - GGML_ASSERT(src0->type == GGML_TYPE_F32); - GGML_ASSERT( dst->type == GGML_TYPE_I32); - - const int nrows = ggml_nrows(src0); - - enum ggml_sort_order order = (enum ggml_sort_order) dst->op_params[0]; - - // bitonic sort requires the number of elements to be power of 2 - int64_t ne00_padded = 1; - while (ne00_padded < ne00) { - ne00_padded *= 2; - } - - // Metal kernels require the buffer size to be multiple of 16 bytes - // https://developer.apple.com/documentation/metal/mtlcomputecommandencoder/1443142-setthreadgroupmemorylength - const int mem_size = GGML_PAD(ne00_padded*sizeof(int32_t), 16); - - id pipeline = nil; - - switch (order) { - case GGML_SORT_ORDER_ASC: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ARGSORT_F32_I32_ASC].pipeline; break; - case GGML_SORT_ORDER_DESC: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ARGSORT_F32_I32_DESC].pipeline; break; - default: GGML_ABORT("fatal error"); - }; - - ggml_metal_kargs_argsort args = { - /*.ncols =*/ ne00, - /*.ncols_pad =*/ ne00_padded - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&args length:sizeof(args) atIndex:2]; - [encoder setThreadgroupMemoryLength:mem_size atIndex:0]; - - [encoder dispatchThreadgroups:MTLSizeMake(1, nrows, 1) threadsPerThreadgroup:MTLSizeMake(ne00_padded, 1, 1)]; - } break; - case GGML_OP_LEAKY_RELU: - { - GGML_ASSERT(src0->type == GGML_TYPE_F32); - - float slope; - memcpy(&slope, dst->op_params, sizeof(float)); - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_LEAKY_RELU_F32].pipeline; - - ggml_metal_kargs_leaky_relu args = { - /*.slope =*/ slope - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&args length:sizeof(args) atIndex:2]; - - const int64_t n = ggml_nelements(dst); - - [encoder dispatchThreadgroups:MTLSizeMake(n, 1, 1) threadsPerThreadgroup:MTLSizeMake(1, 1, 1)]; - } break; - case GGML_OP_FLASH_ATTN_EXT: - { - GGML_ASSERT(ne00 % 4 == 0); - GGML_ASSERT(ne11 % 32 == 0); - - GGML_ASSERT(src0->type == GGML_TYPE_F32); - GGML_ASSERT(src1->type == src2->type); - - //GGML_ASSERT(ggml_are_same_shape (src1, src2)); - GGML_ASSERT(ne11 == ne21); - GGML_ASSERT(ne12 == ne22); - - struct ggml_tensor * src3 = node->src[3]; // mask - struct ggml_tensor * src4 = node->src[4]; // sinks - - size_t offs_src3 = 0; - size_t offs_src4 = 0; - - id id_src3 = src3 ? ggml_metal_get_buffer(src3, &offs_src3) : nil; - id id_src4 = src4 ? ggml_metal_get_buffer(src4, &offs_src4) : nil; - - GGML_ASSERT(!src3 || src3->type == GGML_TYPE_F16); - GGML_ASSERT(!src3 || src3->ne[1] >= GGML_PAD(src0->ne[1], 8) && - "the Flash-Attention Metal kernel requires the mask to be padded to 8 and at least n_queries big"); - - const int64_t ne30 = src3 ? src3->ne[0] : 0; GGML_UNUSED(ne30); - //const int64_t ne31 = src3 ? src3->ne[1] : 0; - const int64_t ne32 = src3 ? src3->ne[2] : 0; GGML_UNUSED(ne32); - const int64_t ne33 = src3 ? src3->ne[3] : 0; GGML_UNUSED(ne33); - - const uint64_t nb30 = src3 ? src3->nb[0] : 0; GGML_UNUSED(nb30); - const uint64_t nb31 = src3 ? src3->nb[1] : 0; - const uint64_t nb32 = src3 ? src3->nb[2] : 0; GGML_UNUSED(nb32); - const uint64_t nb33 = src3 ? src3->nb[3] : 0; GGML_UNUSED(nb33); - - float scale; - float max_bias; - float logit_softcap; - - memcpy(&scale, ((const int32_t *) dst->op_params) + 0, sizeof(scale)); - memcpy(&max_bias, ((const int32_t *) dst->op_params) + 1, sizeof(max_bias)); - memcpy(&logit_softcap, ((const int32_t *) dst->op_params) + 2, sizeof(logit_softcap)); - - if (logit_softcap != 0.0f) { - scale /= logit_softcap; - } - - const bool has_mask = src3 != NULL; - const bool has_sinks = src4 != NULL; - const bool has_bias = max_bias != 0.0f; - const bool has_scap = logit_softcap != 0.0f; - - const uint32_t n_head = src0->ne[2]; - const uint32_t n_head_log2 = 1u << (uint32_t) floorf(log2f((float) n_head)); - - const float m0 = powf(2.0f, -(max_bias ) / n_head_log2); - const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2); - - GGML_ASSERT(ne01 < 65536); - - if (!ggml_metal_flash_attn_ext_use_vec(dst)) { - // half8x8 kernel - const int64_t nqptg = 8; // queries per threadgroup !! sync with kernel template arguments !! - const int64_t ncpsg = 64; // cache values per simdgroup !! sync with kernel template arguments !! - - GGML_ASSERT(nqptg <= 32); - GGML_ASSERT(nqptg % 8 == 0); - GGML_ASSERT(ncpsg % 32 == 0); - - const int is_q = ggml_is_quantized(src1->type) ? 1 : 0; - - // 2*(2*ncpsg) - // ncpsg soft_max values + ncpsg mask values - // - // 16*32*(nsg) - // the shared memory needed for the simdgroups to load the KV cache - // each thread loads (dequantizes) 16 head elements, there are 32 threads in th SG - // -#define FATTN_SMEM(nsg) (GGML_PAD((nqptg*(ne00 + 2*GGML_PAD(ne20, 64) + 2*(2*ncpsg)) + is_q*(16*32*(nsg)))*(sizeof(float)/2), 16)) - - //int64_t nsgmax = 4; - // - //if (is_q) { - // nsgmax = 2; - // while (true) { - // const size_t smem = FATTN_SMEM(nsgmax); - // if (smem > device.maxThreadgroupMemoryLength/2) { - // break; - // } - // nsgmax *= 2; - // } - // nsgmax /= 2; - //} - - // simdgroups per threadgroup (a.k.a. warps) - //nsg = ne01 <= nqptg ? MAX(4, MIN(nsgmax, MIN(ne11/ncpsg, (int64_t) pipeline.maxTotalThreadsPerThreadgroup/32))) : 4; - int32_t nsg = 4; - - const size_t smem = FATTN_SMEM(nsg); - - ggml_metal_kargs_flash_attn_ext args = { - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne11 =*/ ne11, - /*.ne_12_2 =*/ ne12, - /*.ne_12_3 =*/ ne13, - /*.ns10 =*/ nb11/nb10, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb13 =*/ nb13, - /*.ns20 =*/ nb21/nb20, - /*.nb21 =*/ nb21, - /*.nb22 =*/ nb22, - /*.nb23 =*/ nb23, - /*.ne32 =*/ ne32, - /*.ne33 =*/ ne33, - /*.nb31 =*/ nb31, - /*.nb32 =*/ nb32, - /*.nb33 =*/ nb33, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.scale =*/ scale, - /*.max_bias =*/ max_bias, - /*.m0 =*/ m0, - /*.m1 =*/ m1, - /*.n_head_log2 =*/ n_head_log2, - /*.logit_softcap =*/ logit_softcap, - }; - - id pipeline = ggml_metal_get_pipeline_flash_attn_ext(backend, node, has_mask, has_sinks, has_bias, has_scap, nsg); - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:3]; - if (id_src3) { - [encoder setBuffer:id_src3 offset:offs_src3 atIndex:4]; - } else { - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:4]; - } - if (id_src4) { - [encoder setBuffer:id_src4 offset:offs_src4 atIndex:5]; - } else { - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:5]; - } - - [encoder setBuffer:id_dst offset:offs_dst atIndex:6]; - - //printf("smem: %zu, max: %zu, nsg = %d, ne02 = %d, ne12 = %d\n", smem, device.maxThreadgroupMemoryLength, (int) nsg, ne02, ne12); - GGML_ASSERT(smem <= device.maxThreadgroupMemoryLength); - [encoder setThreadgroupMemoryLength:smem atIndex:0]; - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + nqptg - 1)/nqptg, ne02, ne03) threadsPerThreadgroup:MTLSizeMake(32, nsg, 1)]; -#undef FATTN_SMEM - } else { - // half4x4 kernel - const int64_t nqptg = 1; // queries per threadgroup !! sync with kernel template arguments !! - const int64_t ncpsg = 32; // cache values per simdgroup !! sync with kernel template arguments !! - const int64_t nkpsg = 1*ncpsg; - - GGML_ASSERT(nqptg <= 32); - GGML_ASSERT(nqptg % 1 == 0); - GGML_ASSERT(ncpsg % 32 == 0); - - // ne00 + 2*ncpsg*(nsg) - // for each query, we load it as f16 in shared memory (ne00) - // and store the soft_max values and the mask - // - // ne20*(nsg) - // each simdgroup has a full f32 head vector in shared mem to accumulate results - // -#define FATTN_SMEM(nsg) (GGML_PAD((nqptg*(GGML_PAD(ne00, 128) + 4*ncpsg*(nsg)) + 2*GGML_PAD(ne20, 128)*(nsg))*(sizeof(float)/2), 16)) - - int64_t nsgmax = 2; - while (true) { - const size_t smem = FATTN_SMEM(nsgmax); - // avoid using more than half of the threadgroup memory - can cause slow downs especially for large head sizes - if (smem > device.maxThreadgroupMemoryLength/2) { - break; - } - nsgmax *= 2; - } - nsgmax /= 2; - - // simdgroups per threadgroup (a.k.a. warps) - //const int64_t nsgt = MAX(2, MIN(nsgmax, MIN((ne11 + nkpsg - 1)/(nkpsg), (int64_t) pipeline.maxTotalThreadsPerThreadgroup/32))); - const int64_t nsgt = MAX(2, MIN(nsgmax, MIN((ne11 + nkpsg - 1)/(nkpsg), (int64_t) 1024/32))); - - int64_t nsg = 1; - while (nsg <= nsgt) { - nsg *= 2; - } - nsg /= 2; - - // workgroups - // each workgroup handles nsg*nkpsg cache values - int32_t nwg = 1; - if (false) { - // for small KV caches, we could launch a single workgroup and write the results directly to dst/ - // however, this does not lead to significant improvement, so disabled - nwg = 1; - nsg = 4; - } else { - nwg = 32; - nsg = 1; - while (2*nwg*nsg*nkpsg < ne11 && nsg < 4) { - nsg *= 2; - } - } - - ggml_metal_kargs_flash_attn_ext_vec args = { - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne11 =*/ ne11, - /*.ne_12_2 =*/ ne12, - /*.ne_12_3 =*/ ne13, - /*.ns10 =*/ nb11/nb10, - /*.nb11 =*/ nb11, - /*.nb12 =*/ nb12, - /*.nb13 =*/ nb13, - /*.ns20 =*/ nb21/nb20, - /*.nb21 =*/ nb21, - /*.nb22 =*/ nb22, - /*.nb23 =*/ nb23, - /*.ne32 =*/ ne32, - /*.ne33 =*/ ne33, - /*.nb31 =*/ nb31, - /*.nb32 =*/ nb32, - /*.nb33 =*/ nb33, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.scale =*/ scale, - /*.max_bias =*/ max_bias, - /*.m0 =*/ m0, - /*.m1 =*/ m1, - /*.n_head_log2 =*/ n_head_log2, - /*.logit_softcap =*/ logit_softcap, - }; - - id pipeline = ggml_metal_get_pipeline_flash_attn_ext_vec(backend, node, has_mask, has_sinks, has_bias, has_scap, nsg, nwg); - - GGML_ASSERT(nsg*32 <= (int) pipeline.maxTotalThreadsPerThreadgroup); - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_src1 offset:offs_src1 atIndex:2]; - [encoder setBuffer:id_src2 offset:offs_src2 atIndex:3]; - if (id_src3) { - [encoder setBuffer:id_src3 offset:offs_src3 atIndex:4]; - } else { - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:4]; - } - if (id_src4) { - [encoder setBuffer:id_src4 offset:offs_src4 atIndex:5]; - } else { - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:5]; - } - - const size_t smem = FATTN_SMEM(nsg); - - //printf("smem: %zu, max: %zu, nsg = %d, nsgmax = %d\n", smem, device.maxThreadgroupMemoryLength, (int) nsg, (int) nsgmax); - GGML_ASSERT(smem <= device.maxThreadgroupMemoryLength); - - if (nwg == 1) { - // using 1 workgroup -> write the result directly into dst - [encoder setBuffer:id_dst offset:offs_dst atIndex:6]; - - [encoder setThreadgroupMemoryLength:smem atIndex:0]; - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + nqptg - 1)/nqptg, ne02, ne03*nwg) threadsPerThreadgroup:MTLSizeMake(32, nsg, 1)]; - } else { - // sanity checks - GGML_ASSERT(ne01*ne02*ne03 == ne1*ne2*ne3); - GGML_ASSERT(ne1*ne2*ne3 <= (1u << 31)); - - // write the results from each workgroup into a temp buffer - const size_t offs_tmp = offs_dst + ggml_nbytes(dst); - [encoder setBuffer:id_dst offset:offs_tmp atIndex:6]; - - [encoder setThreadgroupMemoryLength:smem atIndex:0]; - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + nqptg - 1)/nqptg, ne02, ne03*nwg) threadsPerThreadgroup:MTLSizeMake(32, nsg, 1)]; - - // sync the 2 kernels - ggml_metal_encode_concurrency_reset(ctx_enc); - - // reduce the results from the workgroups - { - const int32_t nrows = ne1*ne2*ne3; - - ggml_metal_kargs_flash_attn_ext_vec_reduce args0 = { - nrows, - }; - - id pipeline0 = ggml_metal_get_pipeline_flash_attn_ext_vec_reduce(backend, node, ne20, nwg); - - [encoder setComputePipelineState:pipeline0]; - [encoder setBytes:&args0 length:sizeof(args0) atIndex:0]; - [encoder setBuffer:id_dst offset:offs_tmp atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - - //printf("ne1 = %d, ne2 = %d, ne3 = %d, ne20 = %d\n", ne1, ne2, ne3, ne20); - [encoder dispatchThreadgroups:MTLSizeMake(nrows, 1, 1) threadsPerThreadgroup:MTLSizeMake(32*nwg, 1, 1)]; - } - } -#undef FATTN_SMEM - } - } break; - case GGML_OP_DUP: - case GGML_OP_CPY: - case GGML_OP_CONT: - { - id pipeline = nil; - - switch (src0t) { - case GGML_TYPE_F32: - { - GGML_ASSERT(ne0 % ggml_blck_size(dst->type) == 0); - - switch (dstt) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_F32].pipeline; break; - case GGML_TYPE_I32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_I32].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_F16].pipeline; break; - case GGML_TYPE_BF16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_BF16].pipeline; break; - case GGML_TYPE_Q8_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q8_0].pipeline; break; - case GGML_TYPE_Q4_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_0].pipeline; break; - case GGML_TYPE_Q4_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q4_1].pipeline; break; - case GGML_TYPE_Q5_0: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_0].pipeline; break; - case GGML_TYPE_Q5_1: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_Q5_1].pipeline; break; - case GGML_TYPE_IQ4_NL: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F32_IQ4_NL].pipeline; break; - default: GGML_ABORT("not implemented"); - }; - } break; - case GGML_TYPE_I32: - { - switch (dstt) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_I32_F32].pipeline; break; - default: GGML_ABORT("not implemented"); - }; - } break; - case GGML_TYPE_F16: - { - switch (dstt) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F16_F32].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_F16_F16].pipeline; break; - default: GGML_ABORT("not implemented"); - }; - } break; - case GGML_TYPE_BF16: - { - switch (dstt) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_BF16_F32].pipeline; break; - case GGML_TYPE_BF16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_BF16_BF16].pipeline; break; - default: GGML_ABORT("not implemented"); - }; - } break; - case GGML_TYPE_Q4_0: - { - switch (dstt) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_Q4_0_F32].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_Q4_0_F16].pipeline; break; - default: GGML_ABORT("not implemented"); - }; - } break; - case GGML_TYPE_Q4_1: - { - switch (dstt) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_Q4_1_F32].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_Q4_1_F16].pipeline; break; - default: GGML_ABORT("not implemented"); - }; - } break; - case GGML_TYPE_Q5_0: - { - switch (dstt) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_Q5_0_F32].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_Q5_0_F16].pipeline; break; - default: GGML_ABORT("not implemented"); - }; - } break; - case GGML_TYPE_Q5_1: - { - switch (dstt) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_Q5_1_F32].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_Q5_1_F16].pipeline; break; - default: GGML_ABORT("not implemented"); - }; - } break; - case GGML_TYPE_Q8_0: - { - switch (dstt) { - case GGML_TYPE_F32: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_Q8_0_F32].pipeline; break; - case GGML_TYPE_F16: pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_CPY_Q8_0_F16].pipeline; break; - default: GGML_ABORT("not implemented"); - }; - } break; - default: GGML_ABORT("not implemented"); - } - - GGML_ASSERT(ne00 % ggml_blck_size(src0->type) == 0); - - // TODO: support - //const int32_t nk00 = ne00/ggml_blck_size(dst->type); - const int32_t nk00 = ne00; - - int nth = 32; // SIMD width - - while (nth < nk00 && nth < (int) pipeline.maxTotalThreadsPerThreadgroup) { - nth *= 2; - } - - nth = MIN(nth, (int) pipeline.maxTotalThreadsPerThreadgroup); - - // when rows are small, we can batch them together in a single threadgroup - int nrptg = 1; - - // TODO: relax this constraint in the future - if (ggml_blck_size(src0->type) == 1 && ggml_blck_size(dst->type) == 1) { - if (nth > nk00) { - nrptg = (nth + nk00 - 1)/nk00; - nth = nk00; - - if (nrptg*nth > (int) pipeline.maxTotalThreadsPerThreadgroup) { - nrptg--; - } - } - } - - nth = MIN(nth, nk00); - - ggml_metal_kargs_cpy args = { - /*.ne00 =*/ nk00, - /*.ne01 =*/ ne01, - /*.ne02 =*/ ne02, - /*.ne03 =*/ ne03, - /*.nb00 =*/ nb00, - /*.nb01 =*/ nb01, - /*.nb02 =*/ nb02, - /*.nb03 =*/ nb03, - /*.ne0 =*/ ne0, - /*.ne1 =*/ ne1, - /*.ne2 =*/ ne2, - /*.ne3 =*/ ne3, - /*.nb0 =*/ nb0, - /*.nb1 =*/ nb1, - /*.nb2 =*/ nb2, - /*.nb3 =*/ nb3, - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBytes:&args length:sizeof(args) atIndex:0]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:1]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:2]; - - [encoder dispatchThreadgroups:MTLSizeMake((ne01 + nrptg - 1)/nrptg, ne02, ne03) threadsPerThreadgroup:MTLSizeMake(nth, nrptg, 1)]; - } break; - case GGML_OP_POOL_2D: - { - GGML_ASSERT(ggml_is_contiguous(src0)); - GGML_ASSERT(src0t == GGML_TYPE_F32 && src0t == dstt); - - const int32_t * opts = dst->op_params; - enum ggml_op_pool op = opts[0]; - - id pipeline = nil; - switch (src0t) { - case GGML_TYPE_F32: { - switch(op) { - case GGML_OP_POOL_AVG: - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_POOL_2D_AVG_F32].pipeline; break; - case GGML_OP_POOL_MAX: - pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_POOL_2D_MAX_F32].pipeline; break; - default: GGML_ASSERT(false && "not implemented"); - } - } break; - default: GGML_ASSERT(false && "not implemented"); - } - - const int32_t k0 = opts[1]; - const int32_t k1 = opts[2]; - const int32_t s0 = opts[3]; - const int32_t s1 = opts[4]; - const int32_t p0 = opts[5]; - const int32_t p1 = opts[6]; - - const int64_t IH = src0->ne[1]; - const int64_t IW = src0->ne[0]; - - const int64_t N = dst->ne[3]; - const int64_t OC = dst->ne[2]; - const int64_t OH = dst->ne[1]; - const int64_t OW = dst->ne[0]; - - const int64_t parallel_elements = N * OC * OH * OW; - const int64_t n_threads = MIN((int64_t)[pipeline maxTotalThreadsPerThreadgroup], parallel_elements); - const int64_t n_tg = (parallel_elements + n_threads - 1) / n_threads; - - ggml_metal_kargs_pool_2d args_pool_2d = { - /* .k0 = */ k0, - /* .k1 = */ k1, - /* .s0 = */ s0, - /* .s1 = */ s1, - /* .p0 = */ p0, - /* .p1 = */ p1, - /* .IH = */ IH, - /* .IW = */ IW, - /* .OH = */ OH, - /* .OW = */ OW, - /* .parallel_elements = */ parallel_elements - }; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&args_pool_2d length:sizeof(args_pool_2d) atIndex:2]; - - [encoder dispatchThreadgroups:MTLSizeMake(n_tg, 1, 1) threadsPerThreadgroup:MTLSizeMake(n_threads, 1, 1)]; - } break; - case GGML_OP_ARGMAX: - { - GGML_ASSERT(src0->type == GGML_TYPE_F32); - GGML_ASSERT(ggml_is_contiguous_1(src0)); - GGML_ASSERT(nb00 == ggml_type_size(src0->type)); - - const int64_t nrows = ggml_nrows(src0); - - int nth = 32; // SIMD width - while (nth < ne00 && nth*ne01*ne02*ne03 < 256) { - nth *= 2; - } - - id pipeline = ctx->kernels[GGML_METAL_KERNEL_TYPE_ARGMAX].pipeline; - - [encoder setComputePipelineState:pipeline]; - [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; - [encoder setBuffer:id_dst offset:offs_dst atIndex:1]; - [encoder setBytes:&ne00 length:sizeof( int64_t) atIndex:2]; - [encoder setBytes:&nb01 length:sizeof(uint64_t) atIndex:3]; - [encoder setThreadgroupMemoryLength:32*sizeof(float) atIndex:0]; - [encoder setThreadgroupMemoryLength:32*sizeof(int32_t) atIndex:1]; - - [encoder dispatchThreadgroups:MTLSizeMake(nrows, 1, 1) threadsPerThreadgroup:MTLSizeMake(nth, 1, 1)]; - } break; - default: - { - GGML_LOG_ERROR("%s: error: node %3d, op = %8s not implemented\n", __func__, idx, ggml_op_name(dst->op)); - GGML_ABORT("fatal error"); - } - } - - if (ctx_dev->debug_graph > 0) { - if (n_fuse > 1) { - GGML_LOG_DEBUG("%s: fuse %d ops\n", __func__, n_fuse); - } - } - - // update the mem ranges in the encoding context - for (int i = 0; i < n_fuse; ++i) { - if (!ggml_metal_encode_concurrency_add(ctx_enc, nodes[i])) { - ggml_metal_encode_concurrency_reset(ctx_enc); - } - } - - return n_fuse; -} - -static enum ggml_status ggml_metal_graph_compute( - ggml_backend_t backend, - struct ggml_cgraph * gf) { - struct ggml_backend_metal_context * ctx = backend->context; - struct ggml_backend_metal_device_context * ctx_dev = backend->device->context; - - // number of nodes encoded by the main thread (empirically determined) - const int n_main = 64; - - // number of threads in addition to the main thread - const int n_cb = ctx->n_cb; - - // submit the ggml compute graph to the GPU by creating command buffers and encoding the ops in them - // the first n_nodes_0 are encoded and submitted for processing directly by the calling thread - // while these nodes are processing, we start n_cb threads to enqueue the rest of the nodes - // each thread creates it's own command buffer and enqueues the ops in parallel - // - // tests on M1 Pro and M2 Ultra using LLaMA models, show that optimal values for n_cb are 1 or 2 - - @autoreleasepool { - ctx->gf = gf; - - ctx->n_nodes_0 = MIN(n_main, gf->n_nodes); - ctx->n_nodes_1 = gf->n_nodes - ctx->n_nodes_0; - - ctx->n_nodes_per_cb = (ctx->n_nodes_1 + ctx->n_cb - 1) / ctx->n_cb; - - const bool should_capture = ctx->capture_next_compute; - if (should_capture) { - ctx->capture_next_compute = false; - - // make sure all previous computations have finished before starting the capture - if (ctx->cmd_buf_last) { - [ctx->cmd_buf_last waitUntilCompleted]; - ctx->cmd_buf_last = nil; - } - - if (!ctx->capture_started) { - // create capture scope - ctx->capture_scope = [[MTLCaptureManager sharedCaptureManager] newCaptureScopeWithDevice:ctx_dev->mtl_device]; - - MTLCaptureDescriptor * descriptor = [MTLCaptureDescriptor new]; - descriptor.captureObject = ctx->capture_scope; - descriptor.destination = MTLCaptureDestinationGPUTraceDocument; - descriptor.outputURL = [NSURL fileURLWithPath:[NSString stringWithFormat:@"/tmp/perf-metal.gputrace"]]; - - NSError * error = nil; - if (![[MTLCaptureManager sharedCaptureManager] startCaptureWithDescriptor:descriptor error:&error]) { - GGML_LOG_ERROR("%s: error: unable to start capture '%s'\n", __func__, [[error localizedDescription] UTF8String]); - } else { - [ctx->capture_scope beginScope]; - ctx->capture_started = true; - } - } - } - - // the main thread commits the first few commands immediately - // cmd_buf[n_cb] - { - id cmd_buf = [ctx->queue commandBufferWithUnretainedReferences]; - [cmd_buf retain]; - - if (ctx->cmd_bufs[n_cb].obj) { - [ctx->cmd_bufs[n_cb].obj release]; - } - ctx->cmd_bufs[n_cb].obj = cmd_buf; - - [cmd_buf enqueue]; - - ctx->encode_async(n_cb); - } - - // remember the command buffer for the next iteration - ctx->cmd_buf_last = ctx->cmd_bufs[n_cb].obj; - - // prepare the rest of the command buffers asynchronously (optional) - // cmd_buf[0.. n_cb) - for (int cb_idx = 0; cb_idx < n_cb; ++cb_idx) { - id cmd_buf = [ctx->queue commandBufferWithUnretainedReferences]; - [cmd_buf retain]; - - if (ctx->cmd_bufs[cb_idx].obj) { - [ctx->cmd_bufs[cb_idx].obj release]; - } - ctx->cmd_bufs[cb_idx].obj = cmd_buf; - - // always enqueue the first two command buffers - // enqueue all of the command buffers if we don't need to abort - if (cb_idx < 2 || ctx->abort_callback == NULL) { - [cmd_buf enqueue]; - - // update the pointer to the last queued command buffer - // this is needed to implement synchronize() - ctx->cmd_buf_last = cmd_buf; - } - } - - dispatch_apply(n_cb, ctx->d_queue, ctx->encode_async); - - // for debugging: block until graph is computed - //[ctx->cmd_buf_last waitUntilCompleted]; - - // enter here only when capturing in order to wait for all computation to finish - // otherwise, we leave the graph to compute asynchronously - if (!should_capture && ctx->capture_started) { - // wait for completion and check status of each command buffer - // needed to detect if the device ran out-of-memory for example (#1881) - { - id cmd_buf = ctx->cmd_bufs[n_cb].obj; - [cmd_buf waitUntilCompleted]; - - MTLCommandBufferStatus status = [cmd_buf status]; - if (status != MTLCommandBufferStatusCompleted) { - GGML_LOG_INFO("%s: command buffer %d failed with status %lu\n", __func__, n_cb, status); - if (status == MTLCommandBufferStatusError) { - GGML_LOG_INFO("error: %s\n", [[cmd_buf error].localizedDescription UTF8String]); - } - - return GGML_STATUS_FAILED; - } - } - - for (int i = 0; i < n_cb; ++i) { - id cmd_buf = ctx->cmd_bufs[i].obj; - [cmd_buf waitUntilCompleted]; - - MTLCommandBufferStatus status = [cmd_buf status]; - if (status != MTLCommandBufferStatusCompleted) { - GGML_LOG_INFO("%s: command buffer %d failed with status %lu\n", __func__, i, status); - if (status == MTLCommandBufferStatusError) { - GGML_LOG_INFO("error: %s\n", [[cmd_buf error].localizedDescription UTF8String]); - } - - return GGML_STATUS_FAILED; - } - - id next_buffer = (i + 1 < n_cb ? ctx->cmd_bufs[i + 1].obj : nil); - if (!next_buffer) { - continue; - } - - const bool next_queued = ([next_buffer status] != MTLCommandBufferStatusNotEnqueued); - if (next_queued) { - continue; - } - - if (ctx->abort_callback && ctx->abort_callback(ctx->abort_callback_data)) { - GGML_LOG_INFO("%s: command buffer %d aborted", __func__, i); - return GGML_STATUS_ABORTED; - } - - [next_buffer commit]; - } - - [ctx->capture_scope endScope]; - [[MTLCaptureManager sharedCaptureManager] stopCapture]; - } - } - - return GGML_STATUS_SUCCESS; -} - -//////////////////////////////////////////////////////////////////////////////// -// backend interface -//////////////////////////////////////////////////////////////////////////////// - -// shared buffer - -static void ggml_backend_metal_buffer_shared_free_buffer(ggml_backend_buffer_t buffer) { - struct ggml_backend_metal_buffer_context * ctx = (struct ggml_backend_metal_buffer_context *)buffer->context; - - for (int i = 0; i < ctx->n_buffers; i++) { - [ctx->buffers[i].metal release]; - } - - ggml_backend_metal_buffer_rset_free(ctx); - - GGML_ASSERT(ctx->is_shared); - - { -#if TARGET_OS_OSX - vm_deallocate((vm_map_t)mach_task_self(), (vm_address_t)ctx->all_data, ctx->all_size); -#else - free(ctx->all_data); -#endif - } - - free(ctx); -} - -static void * ggml_backend_metal_buffer_shared_get_base(ggml_backend_buffer_t buffer) { - struct ggml_backend_metal_buffer_context * ctx = (struct ggml_backend_metal_buffer_context *)buffer->context; - - return ctx->all_data; -} - -static void ggml_backend_metal_buffer_shared_memset_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, uint8_t value, size_t offset, size_t size) { - struct ggml_backend_metal_buffer_context * ctx = (struct ggml_backend_metal_buffer_context *)buffer->context; - - GGML_ASSERT(ctx->is_shared); - - memset((char *)tensor->data + offset, value, size); -} - -static void ggml_backend_metal_buffer_shared_set_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { - struct ggml_backend_metal_buffer_context * ctx = (struct ggml_backend_metal_buffer_context *)buffer->context; - - GGML_ASSERT(ctx->is_shared); - - memcpy((char *)tensor->data + offset, data, size); -} - -static void ggml_backend_metal_buffer_shared_get_tensor(ggml_backend_buffer_t buffer, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { - struct ggml_backend_metal_buffer_context * ctx = (struct ggml_backend_metal_buffer_context *)buffer->context; - - GGML_ASSERT(ctx->is_shared); - - memcpy(data, (const char *)tensor->data + offset, size); -} - -static bool ggml_backend_metal_buffer_shared_cpy_tensor(ggml_backend_buffer_t buffer, const struct ggml_tensor * src, struct ggml_tensor * dst) { - GGML_UNUSED(buffer); - GGML_UNUSED(src); - GGML_UNUSED(dst); - - return false; -} - -static void ggml_backend_metal_buffer_shared_clear(ggml_backend_buffer_t buffer, uint8_t value) { - struct ggml_backend_metal_buffer_context * ctx = (struct ggml_backend_metal_buffer_context *)buffer->context; - - GGML_ASSERT(ctx->is_shared); - - memset(ctx->all_data, value, ctx->all_size); -} - -static struct ggml_backend_buffer_i ggml_backend_metal_buffer_shared_i = { - /* .free_buffer = */ ggml_backend_metal_buffer_shared_free_buffer, - /* .get_base = */ ggml_backend_metal_buffer_shared_get_base, - /* .init_tensor = */ NULL, - /* .memset_tensor = */ ggml_backend_metal_buffer_shared_memset_tensor, - /* .set_tensor = */ ggml_backend_metal_buffer_shared_set_tensor, - /* .get_tensor = */ ggml_backend_metal_buffer_shared_get_tensor, - /* .cpy_tensor = */ ggml_backend_metal_buffer_shared_cpy_tensor, - /* .clear = */ ggml_backend_metal_buffer_shared_clear, - /* .reset = */ NULL, -}; - -// private buffer - -static void ggml_backend_metal_buffer_private_free_buffer(ggml_backend_buffer_t buffer) { - struct ggml_backend_metal_buffer_context * ctx = (struct ggml_backend_metal_buffer_context *)buffer->context; - - for (int i = 0; i < ctx->n_buffers; i++) { - [ctx->buffers[i].metal release]; - } - - ggml_backend_metal_buffer_rset_free(ctx); - - GGML_ASSERT(!ctx->is_shared); - - free(ctx); -} - -static void * ggml_backend_metal_buffer_private_get_base(ggml_backend_buffer_t buffer) { - struct ggml_backend_metal_buffer_context * ctx = (struct ggml_backend_metal_buffer_context *)buffer->context; - - return ctx->all_data; -} - -static void ggml_backend_metal_buffer_private_memset_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, uint8_t value, size_t offset, size_t size) { - struct ggml_backend_metal_buffer_context * ctx = (struct ggml_backend_metal_buffer_context *)buffer->context; - - GGML_ASSERT(!ctx->is_shared); - - @autoreleasepool { - // dst - size_t buf_dst_offset = 0; - id buf_dst = ggml_metal_get_buffer(tensor, &buf_dst_offset); - - buf_dst_offset += offset; - - id queue = ctx->queue; - id cmd_buf = [queue commandBufferWithUnretainedReferences]; - - { - id encoder = [cmd_buf blitCommandEncoder]; - - [encoder fillBuffer:buf_dst - range:NSMakeRange(buf_dst_offset, buf_dst_offset + size) - value:value]; - - [encoder endEncoding]; - } - - [cmd_buf commit]; - [cmd_buf waitUntilCompleted]; - } -} - -static void ggml_backend_metal_buffer_private_set_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { - struct ggml_backend_metal_buffer_context * ctx = (struct ggml_backend_metal_buffer_context *)buffer->context; - - GGML_ASSERT(!ctx->is_shared); - - @autoreleasepool { - // src - void * data_ptr = (void *)(uintptr_t) data; // "const cast" the src data - id buf_src = [ctx->device newBufferWithBytesNoCopy:data_ptr - length:size - options:MTLResourceStorageModeShared - deallocator:nil]; - - // dst - size_t buf_dst_offset = 0; - id buf_dst = ggml_metal_get_buffer(tensor, &buf_dst_offset); - - buf_dst_offset += offset; - - // note: for experimentation purposes, here we use a semaphore to wait for the copy to complete - // this is alternative to waitUntilCompleted, which should be faster, but don't seem to make much difference - dispatch_semaphore_t completion_semaphore = dispatch_semaphore_create(0); - - id queue = ctx->queue; - id cmd_buf = [queue commandBufferWithUnretainedReferences]; - - { - id encoder = [cmd_buf blitCommandEncoder]; - - [encoder copyFromBuffer:buf_src - sourceOffset:0 - toBuffer:buf_dst - destinationOffset:buf_dst_offset - size:size]; - - [encoder endEncoding]; - } - - [cmd_buf addCompletedHandler:^(id cb) { - // TODO: can check for errors here - GGML_UNUSED(cb); - - dispatch_semaphore_signal(completion_semaphore); - }]; - - [cmd_buf commit]; - - dispatch_semaphore_wait(completion_semaphore, DISPATCH_TIME_FOREVER); - //[cmd_buf waitUntilCompleted]; - } -} - -static void ggml_backend_metal_buffer_private_get_tensor(ggml_backend_buffer_t buffer, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { - struct ggml_backend_metal_buffer_context * ctx = (struct ggml_backend_metal_buffer_context *)buffer->context; - - GGML_ASSERT(!ctx->is_shared); - - @autoreleasepool { - // src - size_t buf_src_offset = 0; - id buf_src = ggml_metal_get_buffer(tensor, &buf_src_offset); - - buf_src_offset += offset; - - // dst - id buf_dst = [ctx->device newBufferWithBytesNoCopy:data - length:size - options:MTLResourceStorageModeShared - deallocator:nil]; - - id queue = ctx->queue; - id cmd_buf = [queue commandBufferWithUnretainedReferences]; - - { - id encoder = [cmd_buf blitCommandEncoder]; - - [encoder copyFromBuffer:buf_src - sourceOffset:buf_src_offset - toBuffer:buf_dst - destinationOffset:0 - size:size]; - - [encoder endEncoding]; - } - - [cmd_buf commit]; - [cmd_buf waitUntilCompleted]; - } -} - -static bool ggml_backend_metal_buffer_private_cpy_tensor(ggml_backend_buffer_t buffer, const struct ggml_tensor * src, struct ggml_tensor * dst) { - GGML_UNUSED(buffer); - GGML_UNUSED(src); - GGML_UNUSED(dst); - - return false; -} - -static void ggml_backend_metal_buffer_private_clear(ggml_backend_buffer_t buffer, uint8_t value) { - struct ggml_backend_metal_buffer_context * ctx = (struct ggml_backend_metal_buffer_context *)buffer->context; - - GGML_ASSERT(!ctx->is_shared); - - @autoreleasepool { - id queue = ctx->queue; - id cmd_buf = [queue commandBufferWithUnretainedReferences]; - - { - id encoder = [cmd_buf blitCommandEncoder]; - - [encoder fillBuffer:ctx->buffers[0].metal - range:NSMakeRange(0, ctx->buffers[0].size) - value:value]; - - [encoder endEncoding]; - } - - [cmd_buf commit]; - [cmd_buf waitUntilCompleted]; - } -} - -static struct ggml_backend_buffer_i ggml_backend_metal_buffer_private_i = { - /* .free_buffer = */ ggml_backend_metal_buffer_private_free_buffer, - /* .get_base = */ ggml_backend_metal_buffer_private_get_base, - /* .init_tensor = */ NULL, - /* .memset_tensor = */ ggml_backend_metal_buffer_private_memset_tensor, - /* .set_tensor = */ ggml_backend_metal_buffer_private_set_tensor, - /* .get_tensor = */ ggml_backend_metal_buffer_private_get_tensor, - /* .cpy_tensor = */ ggml_backend_metal_buffer_private_cpy_tensor, - /* .clear = */ ggml_backend_metal_buffer_private_clear, - /* .reset = */ NULL, -}; - -// -// buffer types -// - -static void ggml_backend_metal_log_allocated_size(id device, size_t size_aligned) { -#ifndef GGML_METAL_NDEBUG -#if TARGET_OS_OSX || (TARGET_OS_IOS && __clang_major__ >= 15) - if (@available(macOS 10.12, iOS 16.0, *)) { - GGML_LOG_DEBUG("%s: allocated buffer, size = %8.2f MiB, (%8.2f / %8.2f)\n", - __func__, - size_aligned / 1024.0 / 1024.0, - device.currentAllocatedSize / 1024.0 / 1024.0, - device.recommendedMaxWorkingSetSize / 1024.0 / 1024.0); - - if (device.currentAllocatedSize > device.recommendedMaxWorkingSetSize) { - GGML_LOG_WARN("%s: warning: current allocated size is greater than the recommended max working set size\n", __func__); - } - } else { - GGML_LOG_INFO("%s: allocated buffer, size = %8.2f MiB, (%8.2f)\n", - __func__, - size_aligned / 1024.0 / 1024.0, - device.currentAllocatedSize / 1024.0 / 1024.0); - } -#endif -#endif - GGML_UNUSED(device); - GGML_UNUSED(size_aligned); -} - -// common method for allocating shread or private Metal buffers -static ggml_backend_buffer_t ggml_backend_metal_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size, bool shared) { - struct ggml_backend_metal_buffer_context * ctx = calloc(1, sizeof(struct ggml_backend_metal_buffer_context)); - - const size_t size_page = sysconf(_SC_PAGESIZE); - - size_t size_aligned = size; - if ((size_aligned % size_page) != 0) { - size_aligned += (size_page - (size_aligned % size_page)); - } - - struct ggml_backend_metal_device_context * ctx_dev = (struct ggml_backend_metal_device_context *)buft->device->context; - - GGML_ASSERT(ctx_dev->mtl_device != nil); - - id device = ctx_dev->mtl_device; - - // allocate shared buffer if the device supports it and it is required by the buffer type - if (ctx_dev->use_shared_buffers && shared) { - ctx->all_data = ggml_metal_host_malloc(size_aligned); - ctx->is_shared = true; - } else { - // dummy, non-NULL value - we'll populate this after creating the Metal buffer below - ctx->all_data = (void *) 0x000000400ULL; - ctx->is_shared = false; - } - ctx->all_size = size_aligned; - - ctx->device = device; - ctx->queue = ctx_dev->mtl_queue; - - ctx->n_buffers = 1; - - if (ctx->all_data != NULL) { - ctx->buffers[0].size = size; - ctx->buffers[0].metal = nil; - - if (size_aligned > 0) { - if (ctx_dev->use_shared_buffers) { - ctx->buffers[0].metal = [device newBufferWithBytesNoCopy:ctx->all_data - length:size_aligned - options:MTLResourceStorageModeShared - deallocator:nil]; - } else { - ctx->buffers[0].metal = [device newBufferWithLength:size_aligned options:MTLResourceStorageModePrivate]; - - ctx->all_data = (void *) (ctx->buffers[0].metal.gpuAddress); - } - } - - ctx->buffers[0].data = ctx->all_data; - } - - if (size_aligned > 0 && (ctx->all_data == NULL || ctx->buffers[0].metal == nil)) { - GGML_LOG_ERROR("%s: error: failed to allocate buffer, size = %8.2f MiB\n", __func__, size_aligned / 1024.0 / 1024.0); - free(ctx); - return NULL; - } - - if (!ggml_backend_metal_buffer_rset_init(ctx, ctx_dev, device)) { - GGML_LOG_ERROR("%s: error: failed to initialize residency set\n", __func__); - free(ctx); - return NULL; - } - - //ggml_backend_metal_log_allocated_size(device, size_aligned); - - struct ggml_backend_buffer_i buf_i = ctx->is_shared ? ggml_backend_metal_buffer_shared_i : ggml_backend_metal_buffer_private_i; - - return ggml_backend_buffer_init(buft, buf_i, ctx, size); -} - -static size_t ggml_backend_metal_buffer_type_get_alloc_size(ggml_backend_buffer_type_t buft, const struct ggml_tensor * tensor) { - size_t res = ggml_nbytes(tensor); - - // some operations require additional memory for fleeting data: - switch (tensor->op) { - case GGML_OP_MUL_MAT_ID: - { - res += ggml_metal_mul_mat_id_extra_tpe(tensor); - res += ggml_metal_mul_mat_id_extra_ids(tensor); - } break; - case GGML_OP_FLASH_ATTN_EXT: - { - if (ggml_metal_flash_attn_ext_use_vec(tensor)) { - res += ggml_metal_flash_attn_ext_extra_tmp(tensor); - } - } break; - default: - break; - } - - return res; - - GGML_UNUSED(buft); -} - -// default (shared) buffer type - -static const char * ggml_backend_metal_buffer_type_shared_get_name(ggml_backend_buffer_type_t buft) { - return "Metal"; - - GGML_UNUSED(buft); -} - -static ggml_backend_buffer_t ggml_backend_metal_buffer_type_shared_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { - return ggml_backend_metal_buffer_type_alloc_buffer(buft, size, true); -} - -static size_t ggml_backend_metal_buffer_type_shared_get_alignment(ggml_backend_buffer_type_t buft) { - return 32; - - GGML_UNUSED(buft); -} - -static size_t ggml_backend_metal_buffer_type_shared_get_max_size(ggml_backend_buffer_type_t buft) { - const size_t max_size = ((struct ggml_backend_metal_device_context *)buft->device->context)->max_size; - - return max_size; -} - -static size_t ggml_backend_metal_buffer_type_shared_get_alloc_size(ggml_backend_buffer_type_t buft, const struct ggml_tensor * tensor) { - return ggml_backend_metal_buffer_type_get_alloc_size(buft, tensor); -} - -static bool ggml_backend_metal_buffer_type_shared_is_host(ggml_backend_buffer_type_t buft) { - return false; - - GGML_UNUSED(buft); -} - -static ggml_backend_buffer_type_t ggml_backend_metal_buffer_type_shared(void) { - static struct ggml_backend_buffer_type ggml_backend_buffer_type_metal = { - /* .iface = */ { - /* .get_name = */ ggml_backend_metal_buffer_type_shared_get_name, - /* .alloc_buffer = */ ggml_backend_metal_buffer_type_shared_alloc_buffer, - /* .get_alignment = */ ggml_backend_metal_buffer_type_shared_get_alignment, - /* .get_max_size = */ ggml_backend_metal_buffer_type_shared_get_max_size, - /* .get_alloc_size = */ ggml_backend_metal_buffer_type_shared_get_alloc_size, - /* .is_host = */ ggml_backend_metal_buffer_type_shared_is_host, - }, - /* .device = */ &g_ggml_backend_metal_device, - /* .context = */ NULL, - }; - - return &ggml_backend_buffer_type_metal; -} - -// default (private) buffer type - -static const char * ggml_backend_metal_buffer_type_private_get_name(ggml_backend_buffer_type_t buft) { - return "Metal_Private"; - - GGML_UNUSED(buft); -} - -static ggml_backend_buffer_t ggml_backend_metal_buffer_type_private_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { - return ggml_backend_metal_buffer_type_alloc_buffer(buft, size, false); -} - -static size_t ggml_backend_metal_buffer_type_private_get_alignment(ggml_backend_buffer_type_t buft) { - return 32; - - GGML_UNUSED(buft); -} - -static size_t ggml_backend_metal_buffer_type_private_get_max_size(ggml_backend_buffer_type_t buft) { - const size_t max_size = ((struct ggml_backend_metal_device_context *)buft->device->context)->max_size; - - return max_size; -} - -static size_t ggml_backend_metal_buffer_type_private_get_alloc_size(ggml_backend_buffer_type_t buft, const struct ggml_tensor * tensor) { - return ggml_backend_metal_buffer_type_get_alloc_size(buft, tensor); -} - -static bool ggml_backend_metal_buffer_type_private_is_host(ggml_backend_buffer_type_t buft) { - return false; - - GGML_UNUSED(buft); -} - -static ggml_backend_buffer_type_t ggml_backend_metal_buffer_type_private(void) { - static struct ggml_backend_buffer_type ggml_backend_buffer_type_metal = { - /* .iface = */ { - /* .get_name = */ ggml_backend_metal_buffer_type_private_get_name, - /* .alloc_buffer = */ ggml_backend_metal_buffer_type_private_alloc_buffer, - /* .get_alignment = */ ggml_backend_metal_buffer_type_private_get_alignment, - /* .get_max_size = */ ggml_backend_metal_buffer_type_private_get_max_size, - /* .get_alloc_size = */ ggml_backend_metal_buffer_type_private_get_alloc_size, - /* .is_host = */ ggml_backend_metal_buffer_type_private_is_host, - }, - /* .device = */ &g_ggml_backend_metal_device, - /* .context = */ NULL, - }; - - return &ggml_backend_buffer_type_metal; -} - -// mapped buffer type - -static const char * ggml_backend_metal_buffer_type_mapped_get_name(ggml_backend_buffer_type_t buft) { - return "Metal_Mapped"; - - GGML_UNUSED(buft); -} - -static ggml_backend_buffer_t ggml_backend_metal_buffer_type_mapped_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { - // for mapped buffers, prefer shared memory - return ggml_backend_metal_buffer_type_alloc_buffer(buft, size, true); -} - -static size_t ggml_backend_metal_buffer_type_mapped_get_alignment(ggml_backend_buffer_type_t buft) { - return 32; - - GGML_UNUSED(buft); -} - -static size_t ggml_backend_metal_buffer_type_mapped_get_max_size(ggml_backend_buffer_type_t buft) { - const size_t max_size = ((struct ggml_backend_metal_device_context *)buft->device->context)->max_size; - - return max_size; -} - -static size_t ggml_backend_metal_buffer_type_mapped_get_alloc_size(ggml_backend_buffer_type_t buft, const struct ggml_tensor * tensor) { - return ggml_backend_metal_buffer_type_get_alloc_size(buft, tensor); -} - -static bool ggml_backend_metal_buffer_type_mapped_is_host(ggml_backend_buffer_type_t buft) { - return false; - - GGML_UNUSED(buft); -} - -static ggml_backend_buffer_type_t ggml_backend_metal_buffer_type_mapped(void) { - // note: not obvious, but this buffer type still needs to implement .alloc_buffer: - // https://github.com/ggml-org/llama.cpp/pull/15832#discussion_r2333177099 - static struct ggml_backend_buffer_type ggml_backend_buffer_type_mapped_metal = { - /* .iface = */ { - /* .get_name = */ ggml_backend_metal_buffer_type_mapped_get_name, - /* .alloc_buffer = */ ggml_backend_metal_buffer_type_mapped_alloc_buffer, - /* .get_alignment = */ ggml_backend_metal_buffer_type_mapped_get_alignment, - /* .get_max_size = */ ggml_backend_metal_buffer_type_mapped_get_max_size, - /* .get_alloc_size = */ ggml_backend_metal_buffer_type_mapped_get_alloc_size, - /* .is_host = */ ggml_backend_metal_buffer_type_mapped_is_host, - }, - /* .device = */ &g_ggml_backend_metal_device, - /* .context = */ NULL, - }; - - return &ggml_backend_buffer_type_mapped_metal; -} - -// backend - -static const char * ggml_backend_metal_name(ggml_backend_t backend) { - return "Metal"; - - GGML_UNUSED(backend); -} - -static void ggml_backend_metal_free(ggml_backend_t backend) { - struct ggml_backend_metal_context * ctx = backend->context; - - ggml_metal_free(ctx); - - free(backend); -} - -static void ggml_backend_metal_synchronize(ggml_backend_t backend) { - struct ggml_backend_metal_context * ctx = backend->context; - - // wait for any backend operations to finish - if (ctx->cmd_buf_last) { - [ctx->cmd_buf_last waitUntilCompleted]; - ctx->cmd_buf_last = nil; - } - - // release any completed command buffers - if (ctx->cmd_bufs_ext.count > 0) { - for (size_t i = 0; i < ctx->cmd_bufs_ext.count; ++i) { - id cmd_buf = ctx->cmd_bufs_ext[i]; - - MTLCommandBufferStatus status = [cmd_buf status]; - if (status != MTLCommandBufferStatusCompleted) { - GGML_LOG_ERROR("%s: error: command buffer %d failed with status %d\n", __func__, (int) i, (int) status); - if (status == MTLCommandBufferStatusError) { - GGML_LOG_ERROR("error: %s\n", [[cmd_buf error].localizedDescription UTF8String]); - } - GGML_ABORT("fatal error"); - } - - [cmd_buf release]; - } - - [ctx->cmd_bufs_ext removeAllObjects]; - } -} - -static void ggml_backend_metal_set_tensor_async(ggml_backend_t backend, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { - struct ggml_backend_metal_context * ctx = backend->context; - struct ggml_backend_metal_device_context * ctx_dev = backend->device->context; - - @autoreleasepool { - id device = ctx_dev->mtl_device; - - // wrap the source data into a Metal buffer - id buf_src = [device newBufferWithBytes:data - length:size - options:MTLResourceStorageModeShared]; - - size_t buf_dst_offset = 0; - id buf_dst = ggml_metal_get_buffer(tensor, &buf_dst_offset); - - if (buf_dst == nil) { - GGML_ABORT("%s: failed to find buffer for tensor '%s'\n", __func__, tensor->name); - } - - buf_dst_offset += offset; - - // queue the copy operation into the queue of the Metal context - // this will be queued at the end, after any currently ongoing GPU operations - id cmd_buf = [ctx->queue commandBufferWithUnretainedReferences]; - id encoder = [cmd_buf blitCommandEncoder]; - - [encoder copyFromBuffer:buf_src - sourceOffset:0 - toBuffer:buf_dst - destinationOffset:buf_dst_offset - size:size]; - - [encoder endEncoding]; - [cmd_buf commit]; - - // do not wait here for completion - //[cmd_buf waitUntilCompleted]; - - // instead, remember a reference to the command buffer and wait for it later if needed - [ctx->cmd_bufs_ext addObject:cmd_buf]; - ctx->cmd_buf_last = cmd_buf; - - [cmd_buf retain]; - } -} - -static void ggml_backend_metal_get_tensor_async(ggml_backend_t backend, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { - struct ggml_backend_metal_context * ctx = backend->context; - struct ggml_backend_metal_device_context * ctx_dev = backend->device->context; - - @autoreleasepool { - id device = ctx_dev->mtl_device; - - id buf_dst = [device newBufferWithBytesNoCopy:data - length:size - options:MTLResourceStorageModeShared - deallocator:nil]; - - size_t buf_src_offset = 0; - id buf_src = ggml_metal_get_buffer(tensor, &buf_src_offset); - - if (buf_src == nil) { - GGML_ABORT("%s: failed to find buffer for tensor '%s'\n", __func__, tensor->name); - } - - buf_src_offset += offset; - - // queue the copy operation into the queue of the Metal context - // this will be queued at the end, after any currently ongoing GPU operations - id cmd_buf = [ctx->queue commandBufferWithUnretainedReferences]; - id encoder = [cmd_buf blitCommandEncoder]; - - [encoder copyFromBuffer:buf_src - sourceOffset:buf_src_offset - toBuffer:buf_dst - destinationOffset:0 - size:size]; - - [encoder endEncoding]; - [cmd_buf commit]; - - // do not wait here for completion - //[cmd_buf waitUntilCompleted]; - - // instead, remember a reference to the command buffer and wait for it later if needed - [ctx->cmd_bufs_ext addObject:cmd_buf]; - ctx->cmd_buf_last = cmd_buf; - - [cmd_buf retain]; - } -} - -static bool ggml_backend_metal_cpy_tensor_async(ggml_backend_t backend_src, ggml_backend_t backend_dst, const struct ggml_tensor * src, struct ggml_tensor * dst) { - return false; - - GGML_UNUSED(backend_src); - GGML_UNUSED(backend_dst); - GGML_UNUSED(src); - GGML_UNUSED(dst); -} - -static enum ggml_status ggml_backend_metal_graph_compute(ggml_backend_t backend, struct ggml_cgraph * cgraph) { - return ggml_metal_graph_compute(backend, cgraph); -} - -static void ggml_backend_metal_graph_optimize(ggml_backend_t backend, struct ggml_cgraph * cgraph) { - struct ggml_backend_metal_device_context * ctx_dev = backend->device->context; - - //const int64_t t_start = ggml_time_us(); - - if (ctx_dev->use_graph_optimize) { - ggml_metal_graph_optimize(cgraph); - } - - //printf("%s: graph optimize took %.3f ms\n", __func__, (ggml_time_us() - t_start) / 1000.0); -} - -static void ggml_backend_metal_set_n_cb(ggml_backend_t backend, int n_cb) { - GGML_ASSERT(ggml_backend_is_metal(backend)); - - struct ggml_backend_metal_context * ctx = (struct ggml_backend_metal_context *)backend->context; - - if (ctx->n_cb != n_cb) { - ctx->n_cb = MIN(n_cb, GGML_METAL_MAX_COMMAND_BUFFERS); - - if (ctx->n_cb > 2) { - GGML_LOG_WARN("%s: n_cb = %d, using n_cb > 2 is not recommended and can degrade the performance in some cases\n", __func__, n_cb); - } - } - - if (ctx->encode_async) { - Block_release(ctx->encode_async); - } - - ctx->encode_async = Block_copy(^(size_t iter) { - const int cb_idx = iter; - const int n_cb_l = ctx->n_cb; - - const int n_nodes_0 = ctx->n_nodes_0; - const int n_nodes_1 = ctx->n_nodes_1; - - const int n_nodes_per_cb = ctx->n_nodes_per_cb; - - id cmd_buf = ctx->cmd_bufs[cb_idx].obj; - struct ggml_mem_ranges * mem_ranges = ctx->cmd_bufs[cb_idx].mem_ranges; - - if (mem_ranges) { - ggml_mem_ranges_reset(mem_ranges); - } - - id encoder; - - struct ggml_backend_metal_device_context * ctx_dev = backend->device->context; - - if (ctx_dev->use_concurrency) { - encoder = [cmd_buf computeCommandEncoderWithDispatchType: MTLDispatchTypeConcurrent]; - } else { - encoder = [cmd_buf computeCommandEncoder]; - } - - int node_start = 0; - int node_end = n_nodes_0; - - if (cb_idx < n_cb_l) { - node_start = n_nodes_0 + ( (cb_idx + 0) * n_nodes_per_cb); - node_end = n_nodes_0 + (MIN((cb_idx == n_cb_l - 1) ? n_nodes_1 : (cb_idx + 1) * n_nodes_per_cb, n_nodes_1)); - } - - const bool should_capture = ctx->capture_next_compute; - - struct ggml_metal_encode_context ctx_enc = { - /*.backend =*/ backend, - /*.encoder =*/ encoder, - /*.mem_ranges =*/ mem_ranges, - }; - - for (int idx = node_start; idx < node_end;) { - if (should_capture) { - [encoder pushDebugGroup:[NSString stringWithCString:ggml_op_desc(ggml_graph_node(ctx->gf, idx)) encoding:NSUTF8StringEncoding]]; - } - - const int res = ggml_metal_encode_node(&ctx_enc, idx, node_end); - if (idx + res > node_end) { - GGML_ABORT("fusion error: nodes spanning multiple encoders have been fused. this indicates a bug in the fusion logic %s", - "https://github.com/ggml-org/llama.cpp/pull/14849"); - } - - if (should_capture) { - [encoder popDebugGroup]; - } - - if (res == 0) { - break; - } - - idx += res; - } - - [encoder endEncoding]; - - if (cb_idx < 2 || ctx->abort_callback == NULL) { - [cmd_buf commit]; - } - }); -} - -static struct ggml_backend_i ggml_backend_metal_i = { - /* .get_name = */ ggml_backend_metal_name, - /* .free = */ ggml_backend_metal_free, - /* .set_tensor_async = */ ggml_backend_metal_set_tensor_async, - /* .get_tensor_async = */ ggml_backend_metal_get_tensor_async, - /* .cpy_tensor_async = */ ggml_backend_metal_cpy_tensor_async, // only needed for multi-GPU setups - /* .synchronize = */ ggml_backend_metal_synchronize, - /* .graph_plan_create = */ NULL, - /* .graph_plan_free = */ NULL, - /* .graph_plan_update = */ NULL, - /* .graph_plan_compute = */ NULL, - /* .graph_compute = */ ggml_backend_metal_graph_compute, - - // the events API is needed only for multi-GPU setups, so likely no need to implement it for Metal - // in any case, these docs seem relevant if we ever decide to implement it: - // https://developer.apple.com/documentation/metal/mtlcommandbuffer#Synchronizing-Passes-with-Events - /* .event_record = */ NULL, - /* .event_wait = */ NULL, - /* .optimize_graph = */ ggml_backend_metal_graph_optimize, -}; - -static ggml_guid_t ggml_backend_metal_guid(void) { - static ggml_guid guid = { 0x81, 0xa1, 0x8b, 0x1e, 0x71, 0xec, 0x79, 0xed, 0x2b, 0x85, 0xdc, 0x8a, 0x61, 0x98, 0x30, 0xe6 }; - return &guid; -} - -// TODO: remove in the future -ggml_backend_t ggml_backend_metal_init(void) { - ggml_backend_dev_t dev = ggml_backend_reg_dev_get(ggml_backend_metal_reg(), 0); - - struct ggml_backend_metal_context * ctx = ggml_metal_init(dev); - if (ctx == NULL) { - GGML_LOG_ERROR("%s: error: failed to allocate context\n", __func__); - return NULL; - } - - ggml_backend_t backend = malloc(sizeof(struct ggml_backend)); - - *backend = (struct ggml_backend) { - /* .guid = */ ggml_backend_metal_guid(), - /* .interface = */ ggml_backend_metal_i, - /* .device = */ dev, - /* .context = */ ctx, - }; - - ggml_backend_metal_set_n_cb(backend, 1); - - return backend; -} - -bool ggml_backend_is_metal(ggml_backend_t backend) { - return backend != NULL && ggml_guid_matches(backend->guid, ggml_backend_metal_guid()); -} - -void ggml_backend_metal_set_abort_callback(ggml_backend_t backend, ggml_abort_callback abort_callback, void * user_data) { - GGML_ASSERT(ggml_backend_is_metal(backend)); - - struct ggml_backend_metal_context * ctx = (struct ggml_backend_metal_context *)backend->context; - - ctx->abort_callback = abort_callback; - ctx->abort_callback_data = user_data; -} - -bool ggml_backend_metal_supports_family(ggml_backend_t backend, int family) { - GGML_ASSERT(ggml_backend_is_metal(backend)); - - struct ggml_backend_metal_device_context * ctx_dev = backend->device->context; - - GGML_ASSERT(ctx_dev->mtl_device != nil); - - return [ctx_dev->mtl_device supportsFamily:(MTLGPUFamilyApple1 + family - 1)]; -} - -void ggml_backend_metal_capture_next_compute(ggml_backend_t backend) { - GGML_ASSERT(ggml_backend_is_metal(backend)); - - struct ggml_backend_metal_context * ctx = (struct ggml_backend_metal_context *)backend->context; - ctx->capture_next_compute = true; -} - -// backend device - -static const char * ggml_backend_metal_device_get_name(ggml_backend_dev_t dev) { - return "Metal"; - - GGML_UNUSED(dev); -} - -static const char * ggml_backend_metal_device_get_description(ggml_backend_dev_t dev) { - struct ggml_backend_metal_device_context * ctx_dev = (struct ggml_backend_metal_device_context *)dev->context; - - return ctx_dev->name; -} - -static void ggml_backend_metal_device_get_memory(ggml_backend_dev_t dev, size_t * free, size_t * total) { - if (@available(macOS 10.12, iOS 16.0, *)) { - struct ggml_backend_metal_device_context * ctx_dev = (struct ggml_backend_metal_device_context *)dev->context; - id device = ctx_dev->mtl_device; - - *total = device.recommendedMaxWorkingSetSize; - *free = *total - device.currentAllocatedSize; - } else { - *free = 1; - *total = 1; - } -} - -static enum ggml_backend_dev_type ggml_backend_metal_device_get_type(ggml_backend_dev_t dev) { - return GGML_BACKEND_DEVICE_TYPE_GPU; - - GGML_UNUSED(dev); -} - -static void ggml_backend_metal_device_get_props(ggml_backend_dev_t dev, struct ggml_backend_dev_props * props) { - props->name = ggml_backend_metal_device_get_name(dev); - props->description = ggml_backend_metal_device_get_description(dev); - props->type = ggml_backend_metal_device_get_type(dev); - ggml_backend_metal_device_get_memory(dev, &props->memory_free, &props->memory_total); - props->caps = (struct ggml_backend_dev_caps) { - /* .async = */ true, - /* .host_buffer = */ false, - /* .buffer_from_host_ptr = */ true, - /* .events = */ false, - }; -} - -static ggml_backend_t ggml_backend_metal_device_init(ggml_backend_dev_t dev, const char * params) { - struct ggml_backend_metal_context * ctx = ggml_metal_init(dev); - if (ctx == NULL) { - GGML_LOG_ERROR("%s: error: failed to allocate context\n", __func__); - return NULL; - } - - ggml_backend_t backend = malloc(sizeof(struct ggml_backend)); - - *backend = (struct ggml_backend) { - /* .guid = */ ggml_backend_metal_guid(), - /* .interface = */ ggml_backend_metal_i, - /* .device = */ dev, - /* .context = */ ctx, - }; - - ggml_backend_metal_set_n_cb(backend, 1); - - return backend; - - GGML_UNUSED(params); -} - -static ggml_backend_buffer_type_t ggml_backend_metal_device_get_buffer_type(ggml_backend_dev_t dev) { - struct ggml_backend_metal_device_context * ctx_dev = dev->context; - - return ctx_dev->use_shared_buffers ? ggml_backend_metal_buffer_type_shared() : ggml_backend_metal_buffer_type_private(); -} - -static ggml_backend_buffer_t ggml_backend_metal_device_buffer_mapped(ggml_backend_dev_t dev, void * ptr, size_t size, size_t max_tensor_size) { - struct ggml_backend_metal_buffer_context * ctx = calloc(1, sizeof(struct ggml_backend_metal_buffer_context)); - - ctx->all_data = ptr; - ctx->all_size = size; - - ctx->is_shared = true; - - ctx->n_buffers = 0; - - const size_t size_page = sysconf(_SC_PAGESIZE); - - // page-align the data ptr - { - const uintptr_t offs = (uintptr_t) ptr % size_page; - ptr = (void *) ((char *) ptr - offs); - size += offs; - } - - size_t size_aligned = size; - if ((size_aligned % size_page) != 0) { - size_aligned += (size_page - (size_aligned % size_page)); - } - - struct ggml_backend_metal_device_context * ctx_dev = (struct ggml_backend_metal_device_context *)dev->context; - - GGML_ASSERT(ctx_dev->mtl_device != nil); - - id device = ctx_dev->mtl_device; - - ctx->device = device; - ctx->queue = ctx_dev->mtl_queue; - - // the buffer fits into the max buffer size allowed by the device - if (size_aligned <= device.maxBufferLength) { - ctx->buffers[ctx->n_buffers].data = ptr; - ctx->buffers[ctx->n_buffers].size = size; - ctx->buffers[ctx->n_buffers].metal = nil; - - if (size_aligned > 0) { - ctx->buffers[ctx->n_buffers].metal = [device newBufferWithBytesNoCopy:ptr length:size_aligned options:MTLResourceStorageModeShared deallocator:nil]; - - if (ctx->buffers[ctx->n_buffers].metal == nil) { - GGML_LOG_ERROR("%s: error: failed to allocate buffer, size = %8.2f MiB\n", __func__, size_aligned / 1024.0 / 1024.0); - return false; - } - } - - ggml_backend_metal_log_allocated_size(device, size_aligned); - - ++ctx->n_buffers; - } else { - // this overlap between the views will guarantee that the tensor with the maximum size will fully fit into - // one of the views - const size_t size_ovlp = ((max_tensor_size + size_page - 1) / size_page + 1) * size_page; // round-up 2 pages just in case - const size_t size_step = device.maxBufferLength - size_ovlp; - const size_t size_view = device.maxBufferLength; - - for (size_t i = 0; i < size; i += size_step) { - const size_t size_step_aligned = (i + size_view <= size) ? size_view : (size_aligned - i); - - ctx->buffers[ctx->n_buffers].data = (void *) ((uint8_t *) ptr + i); - ctx->buffers[ctx->n_buffers].size = size_step_aligned; - ctx->buffers[ctx->n_buffers].metal = nil; - - if (size_step_aligned > 0) { - ctx->buffers[ctx->n_buffers].metal = [device newBufferWithBytesNoCopy:(void *) ((uint8_t *) ptr + i) length:size_step_aligned options:MTLResourceStorageModeShared deallocator:nil]; - - if (ctx->buffers[ctx->n_buffers].metal == nil) { - GGML_LOG_ERROR("%s: error: failed to allocate buffer, size = %8.2f MiB\n", __func__, size_step_aligned / 1024.0 / 1024.0); - return false; - } - } - - ggml_backend_metal_log_allocated_size(device, size_step_aligned); - - if (i + size_step < size) { - GGML_LOG_INFO("\n"); - } - - ++ctx->n_buffers; - } - } - - if (!ggml_backend_metal_buffer_rset_init(ctx, ctx_dev, device)) { - GGML_LOG_ERROR("%s: error: failed to initialize residency set\n", __func__); - free(ctx); - return NULL; - } - - return ggml_backend_buffer_init(ggml_backend_metal_buffer_type_mapped(), ggml_backend_metal_buffer_shared_i, ctx, size); -} - -static bool ggml_backend_metal_device_supports_op(ggml_backend_dev_t dev, const struct ggml_tensor * op) { - struct ggml_backend_metal_device_context * ctx_dev = dev->context; - - return ggml_metal_supports_op(ctx_dev, op); -} - -static bool ggml_backend_metal_device_supports_buft(ggml_backend_dev_t dev, ggml_backend_buffer_type_t buft) { - return - buft->iface.get_name == ggml_backend_metal_buffer_type_shared_get_name || - buft->iface.get_name == ggml_backend_metal_buffer_type_private_get_name || - buft->iface.get_name == ggml_backend_metal_buffer_type_mapped_get_name; - - GGML_UNUSED(dev); -} - -static int64_t get_op_batch_size(const struct ggml_tensor * op) { - switch (op->op) { - case GGML_OP_MUL_MAT: - return op->ne[1]; - case GGML_OP_MUL_MAT_ID: - return op->ne[2]; - default: - return ggml_nrows(op); - } -} - -static bool ggml_backend_metal_device_offload_op(ggml_backend_dev_t dev, const struct ggml_tensor * op) { - const int min_batch_size = 32; - - return (op->op == GGML_OP_MUL_MAT || - op->op == GGML_OP_MUL_MAT_ID) && - get_op_batch_size(op) >= min_batch_size; - - GGML_UNUSED(dev); - GGML_UNUSED(op); -} - -static struct ggml_backend_device_i ggml_backend_metal_device_i = { - /* .get_name = */ ggml_backend_metal_device_get_name, - /* .get_description = */ ggml_backend_metal_device_get_description, - /* .get_memory = */ ggml_backend_metal_device_get_memory, - /* .get_type = */ ggml_backend_metal_device_get_type, - /* .get_props = */ ggml_backend_metal_device_get_props, - /* .init_backend = */ ggml_backend_metal_device_init, - /* .get_buffer_type = */ ggml_backend_metal_device_get_buffer_type, - /* .get_host_buffer_type = */ NULL, - /* .buffer_from_host_ptr = */ ggml_backend_metal_device_buffer_mapped, - /* .supports_op = */ ggml_backend_metal_device_supports_op, - /* .supports_buft = */ ggml_backend_metal_device_supports_buft, - /* .offload_op = */ ggml_backend_metal_device_offload_op, - /* .event_new = */ NULL, - /* .event_free = */ NULL, - /* .event_synchronize = */ NULL, -}; - -// backend registry - -static const char * ggml_backend_metal_reg_get_name(ggml_backend_reg_t reg) { - return "Metal"; - - GGML_UNUSED(reg); -} - -static size_t ggml_backend_metal_reg_device_count(ggml_backend_reg_t reg) { - return 1; - - GGML_UNUSED(reg); -} - -static ggml_backend_dev_t ggml_backend_metal_reg_device_get(ggml_backend_reg_t reg, size_t index) { - GGML_ASSERT(index == 0); - - return &g_ggml_backend_metal_device; - - GGML_UNUSED(reg); - GGML_UNUSED(index); -} - -static struct ggml_backend_feature g_ggml_backend_metal_features[] = { -#if defined(GGML_METAL_EMBED_LIBRARY) - { "EMBED_LIBRARY", "1" }, -#endif -#if defined(GGML_METAL_USE_BF16) - { "BF16", "1" }, -#endif - { nil, nil }, -}; - -static struct ggml_backend_feature * ggml_backend_metal_get_features(ggml_backend_reg_t reg) { - return g_ggml_backend_metal_features; - - GGML_UNUSED(reg); -} - -static void * ggml_backend_metal_get_proc_address(ggml_backend_reg_t reg, const char * name) { - if (strcmp(name, "ggml_backend_get_features") == 0) { - return (void *)ggml_backend_metal_get_features; - } - - return NULL; - - GGML_UNUSED(reg); -} -static struct ggml_backend_reg_i ggml_backend_metal_reg_i = { - /* .get_name = */ ggml_backend_metal_reg_get_name, - /* .device_count = */ ggml_backend_metal_reg_device_count, - /* .device_get = */ ggml_backend_metal_reg_device_get, - /* .get_proc_address = */ ggml_backend_metal_get_proc_address, -}; - -// called upon program exit -static void ggml_metal_cleanup(void) { - ggml_backend_metal_device_rel(&g_ggml_ctx_dev_main); -} - -// TODO: make thread-safe -ggml_backend_reg_t ggml_backend_metal_reg(void) { - ggml_backend_metal_device_acq(&g_ggml_ctx_dev_main); - - // register cleanup callback - // TODO: not ideal, but not sure if there is a better way to do this in Objective-C - atexit(ggml_metal_cleanup); - - { - g_ggml_backend_metal_reg = (struct ggml_backend_reg) { - /* .api_version = */ GGML_BACKEND_API_VERSION, - /* .iface = */ ggml_backend_metal_reg_i, - /* .context = */ NULL, - }; - - g_ggml_backend_metal_device = (struct ggml_backend_device) { - /* .iface = */ ggml_backend_metal_device_i, - /* .reg = */ &g_ggml_backend_metal_reg, - /* .context = */ &g_ggml_ctx_dev_main, - }; - } - - return &g_ggml_backend_metal_reg; -} - -GGML_BACKEND_DL_IMPL(ggml_backend_metal_reg) diff --git a/ggml/src/ggml-metal/ggml-metal.metal b/ggml/src/ggml-metal/ggml-metal.metal index 5057e264f6..20ceb1fe04 100644 --- a/ggml/src/ggml-metal/ggml-metal.metal +++ b/ggml/src/ggml-metal/ggml-metal.metal @@ -27,11 +27,11 @@ using namespace metal; // .../usr/bin/metal -dM -E -c ggml/src/ggml-metal/ggml-metal.metal // .../usr/bin/metal -dM -E -c -target air64-apple-ios14.0 ggml/src/ggml-metal/ggml-metal.metal // -#if __METAL_VERSION__ < 310 && defined(GGML_METAL_USE_BF16) -#undef GGML_METAL_USE_BF16 +#if __METAL_VERSION__ < 310 && defined(GGML_METAL_HAS_BF16) +#undef GGML_METAL_HAS_BF16 #endif -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) typedef matrix bfloat4x4; #endif @@ -87,7 +87,7 @@ void dequantize_f16_t4(device const half4 * src, short il, thread type4 & reg) { reg = (type4)(*(src)); } -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template void dequantize_bf16(device const bfloat4x4 * src, short il, thread type4x4 & reg) { reg = (type4x4)(*src); @@ -1222,53 +1222,78 @@ typedef decltype(kernel_div_row_c4_fuse_impl<1>) kernel_div_row_c4_fuse_t; template [[host_name("kernel_div_row_c4_fuse_1")]] kernel kernel_div_row_c4_fuse_t kernel_div_row_c4_fuse_impl<1>; -kernel void kernel_scale( +kernel void kernel_scale_f32( + constant ggml_metal_kargs_scale & args, device const float * src0, device float * dst, - constant float & scale, - constant float & bias, uint tpig[[thread_position_in_grid]]) { - dst[tpig] = src0[tpig] * scale + bias; + dst[tpig] = src0[tpig] * args.scale + args.bias; } -kernel void kernel_scale_4( +kernel void kernel_scale_f32_4( + constant ggml_metal_kargs_scale & args, device const float4 * src0, device float4 * dst, - constant float & scale, - constant float & bias, uint tpig[[thread_position_in_grid]]) { - dst[tpig] = src0[tpig] * scale + bias; + dst[tpig] = src0[tpig] * args.scale + args.bias; } -kernel void kernel_clamp( +kernel void kernel_clamp_f32( + constant ggml_metal_kargs_clamp & args, device const float * src0, device float * dst, - constant float & min, - constant float & max, uint tpig[[thread_position_in_grid]]) { - dst[tpig] = src0[tpig] < min ? min : (src0[tpig] > max ? max : src0[tpig]); + dst[tpig] = clamp(src0[tpig], args.min, args.max); } -kernel void kernel_relu( +kernel void kernel_clamp_f32_4( + constant ggml_metal_kargs_clamp & args, + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = clamp(src0[tpig], args.min, args.max); +} + +kernel void kernel_relu_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { dst[tpig] = max(0.0f, src0[tpig]); } -kernel void kernel_sigmoid( +kernel void kernel_relu_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = max(0.0f, src0[tpig]); +} + +kernel void kernel_sigmoid_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { dst[tpig] = 1.0f / (1.0f + exp(-src0[tpig])); } -kernel void kernel_tanh( +kernel void kernel_sigmoid_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = 1.0f / (1.0f + exp(-src0[tpig])); +} + +kernel void kernel_tanh_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { - device const float & x = src0[tpig]; - dst[tpig] = precise::tanh(x); + dst[tpig] = precise::tanh(src0[tpig]); +} + +kernel void kernel_tanh_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = precise::tanh(src0[tpig]); } constant float GELU_COEF_A = 0.044715f; @@ -1276,7 +1301,7 @@ constant float GELU_QUICK_COEF = -1.702f; constant float SQRT_2_OVER_PI = 0.79788456080286535587989211986876f; constant float SQRT_2_INV = 0.70710678118654752440084436210484f; -kernel void kernel_gelu( +kernel void kernel_gelu_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { @@ -1285,7 +1310,7 @@ kernel void kernel_gelu( dst[tpig] = 0.5f*x*(1.0f + precise::tanh(SQRT_2_OVER_PI*x*(1.0f + GELU_COEF_A*x*x))); } -kernel void kernel_gelu_4( +kernel void kernel_gelu_f32_4( device const float4 * src0, device float4 * dst, uint tpig[[thread_position_in_grid]]) { @@ -1298,7 +1323,7 @@ kernel void kernel_gelu_4( dst[tpig] = 0.5f*x*(1.0f + precise::tanh(SQRT_2_OVER_PI*x*(1.0f + GELU_COEF_A*x*x))); } -kernel void kernel_gelu_quick( +kernel void kernel_gelu_quick_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { @@ -1307,7 +1332,7 @@ kernel void kernel_gelu_quick( dst[tpig] = x*(1.0f/(1.0f+exp(GELU_QUICK_COEF*x))); } -kernel void kernel_gelu_quick_4( +kernel void kernel_gelu_quick_f32_4( device const float4 * src0, device float4 * dst, uint tpig[[thread_position_in_grid]]) { @@ -1334,7 +1359,7 @@ T erf_approx(T x) { return sign_x * y; } -kernel void kernel_gelu_erf( +kernel void kernel_gelu_erf_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { @@ -1343,7 +1368,7 @@ kernel void kernel_gelu_erf( dst[tpig] = 0.5f*x*(1.0f+erf_approx(x*SQRT_2_INV)); } -kernel void kernel_gelu_erf_4( +kernel void kernel_gelu_erf_f32_4( device const float4 * src0, device float4 * dst, uint tpig[[thread_position_in_grid]]) { @@ -1352,7 +1377,7 @@ kernel void kernel_gelu_erf_4( dst[tpig] = 0.5f*x*(1.0f+erf_approx(x*SQRT_2_INV)); } -kernel void kernel_silu( +kernel void kernel_silu_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { @@ -1360,7 +1385,7 @@ kernel void kernel_silu( dst[tpig] = x / (1.0f + exp(-x)); } -kernel void kernel_silu_4( +kernel void kernel_silu_f32_4( device const float4 * src0, device float4 * dst, uint tpig[[thread_position_in_grid]]) { @@ -1368,99 +1393,202 @@ kernel void kernel_silu_4( dst[tpig] = x / (1.0f + exp(-x)); } -kernel void kernel_elu( +kernel void kernel_elu_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { - device const float & x = src0[tpig]; + const float x = src0[tpig]; dst[tpig] = (x > 0.0f) ? x : (exp(x) - 1.0f); } -kernel void kernel_sqr( +kernel void kernel_elu_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + const float4 x = src0[tpig]; + dst[tpig][0] = (x[0] > 0.0f) ? x[0] : (exp(x[0]) - 1.0f); + dst[tpig][1] = (x[1] > 0.0f) ? x[1] : (exp(x[1]) - 1.0f); + dst[tpig][2] = (x[2] > 0.0f) ? x[2] : (exp(x[2]) - 1.0f); + dst[tpig][3] = (x[3] > 0.0f) ? x[3] : (exp(x[3]) - 1.0f); +} + +kernel void kernel_sqr_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { dst[tpig] = src0[tpig] * src0[tpig]; } -kernel void kernel_sqrt( +kernel void kernel_sqr_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = src0[tpig] * src0[tpig]; +} + +kernel void kernel_sqrt_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { dst[tpig] = sqrt(src0[tpig]); } -kernel void kernel_sin( +kernel void kernel_sqrt_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = sqrt(src0[tpig]); +} + +kernel void kernel_sin_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { dst[tpig] = sin(src0[tpig]); } -kernel void kernel_cos( +kernel void kernel_sin_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = sin(src0[tpig]); +} + +kernel void kernel_cos_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { dst[tpig] = cos(src0[tpig]); } -kernel void kernel_neg( +kernel void kernel_cos_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = cos(src0[tpig]); +} + +kernel void kernel_log_f32( + device const float * src0, + device float * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = log(src0[tpig]); +} + +kernel void kernel_log_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = log(src0[tpig]); +} + +kernel void kernel_neg_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { dst[tpig] = -src0[tpig]; } -kernel void kernel_abs( +kernel void kernel_neg_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = -src0[tpig]; +} + +kernel void kernel_abs_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { dst[tpig] = fabs(src0[tpig]); } -kernel void kernel_sgn( - device const float * src0, - device float * dst, +kernel void kernel_abs_f32_4( + device const float4 * src0, + device float4 * dst, uint tpig[[thread_position_in_grid]]) { - device const float & x = src0[tpig]; - dst[tpig] = (x > 0.0f) ? 1.0f : ((x < 0.0f) ? -1.0f : 0.0f); + dst[tpig] = fabs(src0[tpig]); } -kernel void kernel_step( +kernel void kernel_sgn_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { - dst[tpig] = src0[tpig] > 0.0f ? 1.0f : 0.0f; + dst[tpig] = sign(src0[tpig]); } -kernel void kernel_hardswish( +kernel void kernel_sgn_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = sign(src0[tpig]); +} + +kernel void kernel_step_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { - device const float & x = src0[tpig]; + dst[tpig] = step(0.0f, src0[tpig]); +} + +kernel void kernel_step_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = step(0.0f, src0[tpig]); +} + +kernel void kernel_hardswish_f32( + device const float * src0, + device float * dst, + uint tpig[[thread_position_in_grid]]) { + const float x = src0[tpig]; dst[tpig] = x * fmin(1.0f, fmax(0.0f, (x + 3.0f) / 6.0f)); } -kernel void kernel_hardsigmoid( +kernel void kernel_hardswish_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + const float4 x = src0[tpig]; + dst[tpig] = x * fmin(1.0f, fmax(0.0f, (x + 3.0f) / 6.0f)); +} + +kernel void kernel_hardsigmoid_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { - device const float & x = src0[tpig]; + const float x = src0[tpig]; dst[tpig] = fmin(1.0f, fmax(0.0f, (x + 3.0f) / 6.0f)); } -kernel void kernel_exp( +kernel void kernel_hardsigmoid_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + const float4 x = src0[tpig]; + dst[tpig] = fmin(1.0f, fmax(0.0f, (x + 3.0f) / 6.0f)); +} + +kernel void kernel_exp_f32( device const float * src0, device float * dst, uint tpig[[thread_position_in_grid]]) { dst[tpig] = exp(src0[tpig]); } -kernel void kernel_reglu( +kernel void kernel_exp_f32_4( + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + dst[tpig] = exp(src0[tpig]); +} + +kernel void kernel_reglu_f32( + constant ggml_metal_kargs_glu & args, device const char * src0, device const char * src1, device char * dst, - constant ggml_metal_kargs_glu & args, uint tgpig[[threadgroup_position_in_grid]], uint tpitg[[thread_position_in_threadgroup]], uint ntg[[threads_per_threadgroup]]) { @@ -1476,11 +1604,11 @@ kernel void kernel_reglu( } } -kernel void kernel_geglu( +kernel void kernel_geglu_f32( + constant ggml_metal_kargs_glu & args, device const char * src0, device const char * src1, device char * dst, - constant ggml_metal_kargs_glu & args, uint tgpig[[threadgroup_position_in_grid]], uint tpitg[[thread_position_in_threadgroup]], uint ntg[[threads_per_threadgroup]]) { @@ -1498,11 +1626,11 @@ kernel void kernel_geglu( } } -kernel void kernel_swiglu( +kernel void kernel_swiglu_f32( + constant ggml_metal_kargs_glu & args, device const char * src0, device const char * src1, device char * dst, - constant ggml_metal_kargs_glu & args, uint tgpig[[threadgroup_position_in_grid]], uint tpitg[[thread_position_in_threadgroup]], uint ntg[[threads_per_threadgroup]]) { @@ -1520,11 +1648,11 @@ kernel void kernel_swiglu( } } -kernel void kernel_swiglu_oai( +kernel void kernel_swiglu_oai_f32( + constant ggml_metal_kargs_glu & args, device const char * src0, device const char * src1, device char * dst, - constant ggml_metal_kargs_glu & args, uint tgpig[[threadgroup_position_in_grid]], uint tpitg[[thread_position_in_threadgroup]], uint ntg[[threads_per_threadgroup]]) { @@ -1546,11 +1674,11 @@ kernel void kernel_swiglu_oai( } } -kernel void kernel_geglu_erf( +kernel void kernel_geglu_erf_f32( + constant ggml_metal_kargs_glu & args, device const char * src0, device const char * src1, device char * dst, - constant ggml_metal_kargs_glu & args, uint tgpig[[threadgroup_position_in_grid]], uint tpitg[[thread_position_in_threadgroup]], uint ntg[[threads_per_threadgroup]]) { @@ -1568,11 +1696,11 @@ kernel void kernel_geglu_erf( } } -kernel void kernel_geglu_quick( +kernel void kernel_geglu_quick_f32( + constant ggml_metal_kargs_glu & args, device const char * src0, device const char * src1, device char * dst, - constant ggml_metal_kargs_glu & args, uint tgpig[[threadgroup_position_in_grid]], uint tpitg[[thread_position_in_threadgroup]], uint ntg[[threads_per_threadgroup]]) { @@ -1642,16 +1770,16 @@ kernel void kernel_sum_rows( typedef decltype(kernel_sum_rows) kernel_sum_rows_t; -template [[host_name("kernel_sum_rows")]] kernel kernel_sum_rows_t kernel_sum_rows; -template [[host_name("kernel_mean")]] kernel kernel_sum_rows_t kernel_sum_rows; +template [[host_name("kernel_sum_rows_f32")]] kernel kernel_sum_rows_t kernel_sum_rows; +template [[host_name("kernel_mean_f32")]] kernel kernel_sum_rows_t kernel_sum_rows; template kernel void kernel_soft_max( + constant ggml_metal_kargs_soft_max & args, device const char * src0, device const char * src1, device const char * src2, device char * dst, - constant ggml_metal_kargs_soft_max & args, threadgroup float * buf [[threadgroup(0)]], uint3 tgpig[[threadgroup_position_in_grid]], uint3 tpitg[[thread_position_in_threadgroup]], @@ -1753,11 +1881,11 @@ kernel void kernel_soft_max( template kernel void kernel_soft_max_4( + constant ggml_metal_kargs_soft_max & args, device const char * src0, device const char * src1, device const char * src2, device char * dst, - constant ggml_metal_kargs_soft_max & args, threadgroup float * buf [[threadgroup(0)]], uint3 tgpig[[threadgroup_position_in_grid]], uint3 tpitg[[thread_position_in_threadgroup]], @@ -1867,53 +1995,12 @@ template [[host_name("kernel_soft_max_f32")]] kernel kernel_soft_max_t kerne template [[host_name("kernel_soft_max_f16_4")]] kernel kernel_soft_max_4_t kernel_soft_max_4; template [[host_name("kernel_soft_max_f32_4")]] kernel kernel_soft_max_4_t kernel_soft_max_4; -kernel void kernel_diag_mask_inf( - device const float * src0, - device float * dst, - constant ggml_metal_kargs_diag_mask_inf & args, - uint3 tpig[[thread_position_in_grid]]) { - const int64_t i02 = tpig[2]; - const int64_t i01 = tpig[1]; - const int64_t i00 = tpig[0]; - - if (i00 > args.n_past + i01) { - dst[i02*args.ne01*args.ne00 + i01*args.ne00 + i00] = -INFINITY; - } else { - dst[i02*args.ne01*args.ne00 + i01*args.ne00 + i00] = src0[i02*args.ne01*args.ne00 + i01*args.ne00 + i00]; - } -} - -kernel void kernel_diag_mask_inf_8( - device const float4 * src0, - device float4 * dst, - constant ggml_metal_kargs_diag_mask_inf & args, - uint3 tpig[[thread_position_in_grid]]) { - - const int64_t i = 2*tpig[0]; - - dst[i+0] = src0[i+0]; - dst[i+1] = src0[i+1]; - int64_t i4 = 4*i; - const int64_t i02 = i4/(args.ne00*args.ne01); i4 -= i02*args.ne00*args.ne01; - const int64_t i01 = i4/(args.ne00); i4 -= i01*args.ne00; - const int64_t i00 = i4; - for (int k = 3; k >= 0; --k) { - if (i00 + 4 + k <= args.n_past + i01) { - break; - } - dst[i+1][k] = -INFINITY; - if (i00 + k > args.n_past + i01) { - dst[i][k] = -INFINITY; - } - } -} - // ref: ggml.c:ggml_compute_forward_ssm_conv_f32 -kernel void kernel_ssm_conv_f32( +kernel void kernel_ssm_conv_f32_f32( + constant ggml_metal_kargs_ssm_conv & args, device const void * src0, device const void * src1, device float * dst, - constant ggml_metal_kargs_ssm_conv & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tpitg[[thread_position_in_threadgroup]], uint3 ntg[[threads_per_threadgroup]]) { @@ -1942,6 +2029,7 @@ kernel void kernel_ssm_conv_f32( // ref: ggml.c:ggml_compute_forward_ssm_scan_f32, Mamba-1 part kernel void kernel_ssm_scan_f32( + constant ggml_metal_kargs_ssm_scan & args, device const void * src0, device const void * src1, device const void * src2, @@ -1951,7 +2039,6 @@ kernel void kernel_ssm_scan_f32( device const void * src6, device float * dst, threadgroup float * shared [[threadgroup(0)]], - constant ggml_metal_kargs_ssm_scan & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tpitg[[thread_position_in_threadgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]], @@ -2057,7 +2144,8 @@ kernel void kernel_ssm_scan_f32( } // ref: ggml.c:ggml_compute_forward_ssm_scan_f32, Mamba-2 part -kernel void kernel_ssm_scan_f32_group( +kernel void kernel_ssm_scan_group_f32( + constant ggml_metal_kargs_ssm_scan & args, device const void * src0, device const void * src1, device const void * src2, @@ -2067,7 +2155,6 @@ kernel void kernel_ssm_scan_f32_group( device const void * src6, device float * dst, threadgroup float * shared [[threadgroup(0)]], - constant ggml_metal_kargs_ssm_scan & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tpitg[[thread_position_in_threadgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]], @@ -2346,24 +2433,22 @@ kernel void kernel_rwkv_wkv7_f32( } } -kernel void kernel_argmax( - device const void * x, - device int32_t * dst, - constant int64_t & ncols, - constant uint64_t & nb01, - threadgroup float * shared_maxval [[threadgroup(0)]], - threadgroup int32_t * shared_argmax [[threadgroup(1)]], +kernel void kernel_argmax_f32( + constant ggml_metal_kargs_argmax & args, + device const char * src0, + device char * dst, + threadgroup char * shmem [[threadgroup(0)]], uint tgpig[[threadgroup_position_in_grid]], uint tpitg[[thread_position_in_threadgroup]], uint sgitg[[simdgroup_index_in_threadgroup]], uint tiisg[[thread_index_in_simdgroup]], uint ntg[[threads_per_threadgroup]]) { - device const float * x_row = (device const float *) ((device const char *) x + tgpig * nb01); + device const float * x_row = (device const float *) ((device const char *) src0 + tgpig * args.nb01); float lmax = -INFINITY; int32_t larg = -1; - for (int i00 = tpitg; i00 < ncols; i00 += ntg) { + for (int i00 = tpitg; i00 < args.ne00; i00 += ntg) { if (x_row[i00] > lmax) { lmax = x_row[i00]; larg = i00; @@ -2374,6 +2459,11 @@ kernel void kernel_argmax( float max_val = simd_max(lmax); int32_t arg_val = simd_max(select(-1, larg, lmax == max_val)); + device int32_t * dst_i32 = (device int32_t *) dst; + + threadgroup float * shared_maxval = (threadgroup float *) shmem; + threadgroup int32_t * shared_argmax = (threadgroup int32_t *) shmem + N_SIMDWIDTH; + if (ntg > N_SIMDWIDTH) { if (sgitg == 0) { shared_maxval[tiisg] = -INFINITY; @@ -2395,15 +2485,15 @@ kernel void kernel_argmax( float max_val_reduced = simd_max(max_val); int32_t arg_val_reduced = simd_max(select(-1, arg_val, max_val == max_val_reduced)); - dst[tgpig] = arg_val_reduced; + dst_i32[tgpig] = arg_val_reduced; return; } - dst[tgpig] = arg_val; + dst_i32[tgpig] = arg_val; } -kernel void kernel_norm( +kernel void kernel_norm_f32( constant ggml_metal_kargs_norm & args, device const char * src0, device char * dst, @@ -2537,11 +2627,11 @@ kernel void kernel_rms_norm_fuse_impl( typedef decltype(kernel_rms_norm_fuse_impl<1>) kernel_rms_norm_fuse_t; -template [[host_name("kernel_rms_norm")]] kernel kernel_rms_norm_fuse_t kernel_rms_norm_fuse_impl<1>; -template [[host_name("kernel_rms_norm_mul")]] kernel kernel_rms_norm_fuse_t kernel_rms_norm_fuse_impl<2>; -template [[host_name("kernel_rms_norm_mul_add")]] kernel kernel_rms_norm_fuse_t kernel_rms_norm_fuse_impl<3>; +template [[host_name("kernel_rms_norm_f32")]] kernel kernel_rms_norm_fuse_t kernel_rms_norm_fuse_impl<1>; +template [[host_name("kernel_rms_norm_mul_f32")]] kernel kernel_rms_norm_fuse_t kernel_rms_norm_fuse_impl<2>; +template [[host_name("kernel_rms_norm_mul_add_f32")]] kernel kernel_rms_norm_fuse_t kernel_rms_norm_fuse_impl<3>; -kernel void kernel_l2_norm( +kernel void kernel_l2_norm_f32( constant ggml_metal_kargs_l2_norm & args, device const char * src0, device char * dst, @@ -2584,10 +2674,10 @@ kernel void kernel_l2_norm( } } -kernel void kernel_group_norm( +kernel void kernel_group_norm_f32( + constant ggml_metal_kargs_group_norm & args, device const float * src0, device float * dst, - constant ggml_metal_kargs_group_norm & args, threadgroup float * buf [[threadgroup(0)]], uint tgpig[[threadgroup_position_in_grid]], uint tpitg[[thread_position_in_threadgroup]], @@ -2595,7 +2685,7 @@ kernel void kernel_group_norm( uint tiisg[[thread_index_in_simdgroup]], uint ntg[[threads_per_threadgroup]]) { const int64_t ne = args.ne00*args.ne01*args.ne02; - const int64_t gs = args.ne00*args.ne01*((args.ne02 + args.n_groups - 1) / args.n_groups); + const int64_t gs = args.ne00*args.ne01*((args.ne02 + args.ngrp - 1) / args.ngrp); int start = tgpig * gs; int end = start + gs; @@ -2793,7 +2883,9 @@ static inline void helper_mv_reduce_and_write( } } -template +constant short FC_mul_mv_nsg [[function_constant(FC_MUL_MV + 0)]]; + +template void mul_vec_q_n_f32_impl( args_t args, device const char * src0, @@ -2803,6 +2895,8 @@ void mul_vec_q_n_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; + constexpr short NQ = 16; const int nb = args.ne00/QK4_0; @@ -2887,7 +2981,7 @@ kernel void kernel_mul_mv_q4_0_f32( uint3 tgpig[[threadgroup_position_in_grid]], ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - mul_vec_q_n_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + mul_vec_q_n_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } kernel void kernel_mul_mv_q4_1_f32( @@ -2899,7 +2993,7 @@ kernel void kernel_mul_mv_q4_1_f32( uint3 tgpig[[threadgroup_position_in_grid]], ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - mul_vec_q_n_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + mul_vec_q_n_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } kernel void kernel_mul_mv_q5_0_f32( @@ -2911,7 +3005,7 @@ kernel void kernel_mul_mv_q5_0_f32( uint3 tgpig[[threadgroup_position_in_grid]], ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - mul_vec_q_n_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + mul_vec_q_n_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } kernel void kernel_mul_mv_q5_1_f32( @@ -2923,10 +3017,10 @@ kernel void kernel_mul_mv_q5_1_f32( uint3 tgpig[[threadgroup_position_in_grid]], ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - mul_vec_q_n_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + mul_vec_q_n_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_q8_0_f32_impl( args_t args, device const char * src0, @@ -2936,6 +3030,8 @@ void kernel_mul_mv_q8_0_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; + constexpr short NQ = 8; const int nb = args.ne00/QK8_0; @@ -3007,7 +3103,7 @@ kernel void kernel_mul_mv_q8_0_f32( uint3 tgpig[[threadgroup_position_in_grid]], ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_q8_0_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + kernel_mul_mv_q8_0_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } // mat-vec kernel processing in chunks of float4 @@ -3314,104 +3410,215 @@ template [[host_name("kernel_mul_mv_ext_q6_K_f32_r1_3")]] kernel mul_mv_ext_q4x4 template [[host_name("kernel_mul_mv_ext_q6_K_f32_r1_4")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<4, block_q6_K, 256, dequantize_q6_K>; template [[host_name("kernel_mul_mv_ext_q6_K_f32_r1_5")]] kernel mul_mv_ext_q4x4_f32_t kernel_mul_mv_ext_q4x4_f32_disp<5, block_q6_K, 256, dequantize_q6_K>; -#define N_MV_T_T 4 - -template -void kernel_mul_mv_impl( +template +void kernel_mul_mv_t_t_impl( args_t args, device const char * src0, device const char * src1, device char * dst, + threadgroup char * shmem, uint3 tgpig, - ushort tiisg) { - const int r0 = tgpig.x; - const int rb = tgpig.y*N_MV_T_T; + ushort tiisg, + ushort sgitg) { + const short NSG = FC_mul_mv_nsg; + + constexpr short NB = 32; + constexpr short NF = 8; + + const int nb = args.ne00/NB; + + const int r0 = tgpig.x*NR0; + const int r1 = tgpig.y; const int im = tgpig.z; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; - const uint64_t offset0 = r0*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + //const uint64_t offset0 = r0*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - device const T0 * x = (device const T0 *) (src0 + offset0); + //device const T0 * x = (device const T0 *) (src0 + offset0); + device const T1 * y = (device const T1 *) (src1 + offset1); - device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1; + // pointers to src0 rows + device const T0 * ax [NR0]; + FOR_UNROLL (short row = 0; row < NR0; ++row) { + const uint64_t offset0 = (r0 + row)*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; - if (args.ne00 < 128) { - for (int row = 0; row < N_MV_T_T; ++row) { - int r1 = rb + row; - if (r1 >= args.ne11) { - break; - } + ax[row] = (device const T0 *) ((device char *) src0 + offset0); + } - const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; + float sumf[NR0] = { 0.f }; - device const T1 * y = (device const T1 *) (src1 + offset1); + const short ix = tiisg/(NW/NF); + const short il = tiisg%(NW/NF); - float sumf = 0; - for (int i = tiisg; i < args.ne00; i += 32) { - sumf += (T0) x[i] * (T1) y[i]; - } + const int ib0 = sgitg*NF + ix; - float sum_all = simd_sum(sumf); - if (tiisg == 0) { - dst_f32[(uint64_t)r1*args.ne0 + r0] = sum_all; - } + T1 yl[NF]; + + device const T1 * yb = y + (ib0*NB + il*NF); + + for (int ib = ib0; ib < nb; ib += NSG*NF) { + for (short i = 0; i < NF; ++i) { + yl[i] = yb[i]; } - } else { - device const T04 * x4 = (device const T04 *) x; - for (int row = 0; row < N_MV_T_T; ++row) { - int r1 = rb + row; - if (r1 >= args.ne11) { - break; + + for (short row = 0; row < NR0; row++) { + device const T0 * xb = ax[row] + (ib*NB + il*NF); + + float sumq = 0.f; + FOR_UNROLL (short i = 0; i < NF; ++i) { + sumq += xb[i] * yl[i]; } - const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; + sumf[row] += sumq; + } - device const T1 * y = (device const T1 *) (src1 + offset1); - device const T14 * y4 = (device const T14 *) y; + yb += NSG*NF*NW; + } - float sumf = 0; - for (int i = tiisg; i < args.ne00/4; i += 32) { - sumf += dot((float4) x4[i], (float4) y4[i]); - } - - float sum_all = simd_sum(sumf); - if (tiisg == 0) { - for (int i = 4*(args.ne00/4); i < args.ne00; ++i) sum_all += (float) (x[i] * y[i]); - dst_f32[(uint64_t)r1*args.ne0 + r0] = sum_all; - } + for (int i = nb*NB + sgitg*NW + tiisg; i < args.ne00; i += NW*NSG) { + for (short row = 0; row < NR0; row++) { + sumf[row] += ax[row][i] * y[i]; } } + + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + + helper_mv_reduce_and_write(dst_f32, sumf, r0, args.ne01, tiisg, sgitg, shmem); } -template -kernel void kernel_mul_mv( +template +kernel void kernel_mul_mv_t_t( constant ggml_metal_kargs_mul_mv & args, device const char * src0, device const char * src1, device char * dst, + threadgroup char * shmem [[threadgroup(0)]], uint3 tgpig[[threadgroup_position_in_grid]], - ushort tiisg[[thread_index_in_simdgroup]]) { - kernel_mul_mv_impl( - args, - src0, - src1, - dst, - tgpig, - tiisg); + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + kernel_mul_mv_t_t_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } -typedef decltype(kernel_mul_mv) mul_mv_t; +typedef decltype(kernel_mul_mv_t_t) mul_mv_t_t; -template [[host_name("kernel_mul_mv_f32_f32")]] kernel mul_mv_t kernel_mul_mv; -template [[host_name("kernel_mul_mv_f16_f32")]] kernel mul_mv_t kernel_mul_mv; -template [[host_name("kernel_mul_mv_f16_f16")]] kernel mul_mv_t kernel_mul_mv; -#if defined(GGML_METAL_USE_BF16) -template [[host_name("kernel_mul_mv_bf16_f32")]] kernel mul_mv_t kernel_mul_mv; -template [[host_name("kernel_mul_mv_bf16_bf16")]] kernel mul_mv_t kernel_mul_mv; +template [[host_name("kernel_mul_mv_f32_f32")]] kernel mul_mv_t_t kernel_mul_mv_t_t; +template [[host_name("kernel_mul_mv_f16_f32")]] kernel mul_mv_t_t kernel_mul_mv_t_t; +template [[host_name("kernel_mul_mv_f16_f16")]] kernel mul_mv_t_t kernel_mul_mv_t_t; +#if defined(GGML_METAL_HAS_BF16) +template [[host_name("kernel_mul_mv_bf16_f32")]] kernel mul_mv_t_t kernel_mul_mv_t_t; +template [[host_name("kernel_mul_mv_bf16_bf16")]] kernel mul_mv_t_t kernel_mul_mv_t_t; #endif +template +void kernel_mul_mv_t_t_4_impl( + args_t args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem, + uint3 tgpig, + ushort tiisg, + ushort sgitg) { + const short NSG = FC_mul_mv_nsg; + + constexpr short NB = 32; + constexpr short NF = 16; + constexpr short NF4 = NF/4; + + const int nb = args.ne00/NB; + + const int r0 = tgpig.x*NR0; + const int r1 = tgpig.y; + const int im = tgpig.z; + + const uint i12 = im%args.ne12; + const uint i13 = im/args.ne12; + + //const uint64_t offset0 = r0*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; + + device const T1 * y = (device const T1 *) (src1 + offset1); + device const T14 * y4 = (device const T14 *) (src1 + offset1); + + // pointers to src0 rows + device const T0 * ax [NR0]; + device const T04 * ax4[NR0]; + FOR_UNROLL (short row = 0; row < NR0; ++row) { + const uint64_t offset0 = (r0 + row)*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; + + ax [row] = (device const T0 *) ((device char *) src0 + offset0); + ax4[row] = (device const T04 *) ((device char *) src0 + offset0); + } + + float sumf[NR0] = { 0.f }; + + const short ix = tiisg/(NW/NF); + const short il = tiisg%(NW/NF); + + const int ib0 = sgitg*NF + ix; + + T14 yl4[NF4]; + + device const T14 * yb4 = y4 + (ib0*NB + il*NF)/4; + + for (int ib = ib0; ib < nb; ib += NSG*NF) { + for (short i = 0; i < NF4; ++i) { + yl4[i] = yb4[i]; + } + + for (short row = 0; row < NR0; row++) { + device const T04 * xb4 = ax4[row] + (ib*NB + il*NF)/4; + + float sumq = 0.f; + FOR_UNROLL (short i = 0; i < NF4; ++i) { + sumq += dot(float4(xb4[i]), float4(yl4[i])); + } + + sumf[row] += sumq; + } + + yb4 += NSG*NF*NW/4; + } + + for (int i = nb*NB + sgitg*NW + tiisg; i < args.ne00; i += NW*NSG) { + for (short row = 0; row < NR0; row++) { + sumf[row] += ax[row][i] * y[i]; + } + } + + device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; + + helper_mv_reduce_and_write(dst_f32, sumf, r0, args.ne01, tiisg, sgitg, shmem); +} + +template +kernel void kernel_mul_mv_t_t_4( + constant ggml_metal_kargs_mul_mv & args, + device const char * src0, + device const char * src1, + device char * dst, + threadgroup char * shmem [[threadgroup(0)]], + uint3 tgpig[[threadgroup_position_in_grid]], + ushort tiisg[[thread_index_in_simdgroup]], + ushort sgitg[[simdgroup_index_in_threadgroup]]) { + kernel_mul_mv_t_t_4_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); +} + +typedef decltype(kernel_mul_mv_t_t_4) mul_mv_t_t_4; + +template [[host_name("kernel_mul_mv_f32_f32_4")]] kernel mul_mv_t_t_4 kernel_mul_mv_t_t_4; +template [[host_name("kernel_mul_mv_f16_f32_4")]] kernel mul_mv_t_t_4 kernel_mul_mv_t_t_4; +template [[host_name("kernel_mul_mv_f16_f16_4")]] kernel mul_mv_t_t_4 kernel_mul_mv_t_t_4; +#if defined(GGML_METAL_HAS_BF16) +template [[host_name("kernel_mul_mv_bf16_f32_4")]] kernel mul_mv_t_t_4 kernel_mul_mv_t_t_4; +template [[host_name("kernel_mul_mv_bf16_bf16_4")]] kernel mul_mv_t_t_4 kernel_mul_mv_t_t_4; +#endif + +#define N_MV_T_T 4 + template void kernel_mul_mv_c4_impl( args_t args, @@ -3472,112 +3679,10 @@ typedef decltype(kernel_mul_mv_c4) mul_mv_c4_t; template [[host_name("kernel_mul_mv_f32_f32_c4")]] kernel mul_mv_c4_t kernel_mul_mv_c4; template [[host_name("kernel_mul_mv_f16_f32_c4")]] kernel mul_mv_c4_t kernel_mul_mv_c4; -#if defined(GGML_METAL_USE_BF16) -template [[host_name("kernel_mul_mv_bf16_f32_c4")]] kernel mul_mv_c4_t kernel_mul_mv_c4; -#endif - -template -kernel void kernel_mul_mv_1row( - constant ggml_metal_kargs_mul_mv & args, - device const char * src0, - device const char * src1, - device char * dst, - uint3 tgpig[[threadgroup_position_in_grid]], - ushort tiisg[[thread_index_in_simdgroup]]) { - - const int r0 = tgpig.x; - const int r1 = tgpig.y; - const int im = tgpig.z; - - const uint i12 = im%args.ne12; - const uint i13 = im/args.ne12; - - const uint64_t offset0 = r0*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; - const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - - device const T * x = (device const T *) (src0 + offset0); - device const float * y = (device const float *) (src1 + offset1); - - device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1 + (uint64_t)r1*args.ne0; - - float sumf = 0; - if (args.ne00 < 128) { - for (int i = tiisg; i < args.ne00; i += 32) { - sumf += (float) x[i] * (float) y[i]; - } - float sum_all = simd_sum(sumf); - if (tiisg == 0) { - dst_f32[r0] = sum_all; - } - } else { - device const T4 * x4 = (device const T4 *) x; - device const float4 * y4 = (device const float4 *) y; - - for (int i = tiisg; i < args.ne00/4; i += 32) { - sumf += dot((float4) x4[i], y4[i]); - } - - float sum_all = simd_sum(sumf); - - if (tiisg == 0) { - for (int i = 4*(args.ne00/4); i < args.ne00; ++i) sum_all += (float) (x[i] * y[i]); - dst_f32[r0] = sum_all; - } - } -} - -typedef decltype(kernel_mul_mv_1row) mul_mv_1row_t; - -template [[host_name("kernel_mul_mv_f16_f32_1row")]] kernel mul_mv_1row_t kernel_mul_mv_1row; -#if defined(GGML_METAL_USE_BF16) -template [[host_name("kernel_mul_mv_bf16_f32_1row")]] kernel mul_mv_1row_t kernel_mul_mv_1row; -#endif - -// Assumes row size (ne00) is a multiple of 4 -template -kernel void kernel_mul_mv_l4( - constant ggml_metal_kargs_mul_mv & args, - device const char * src0, - device const char * src1, - device char * dst, - uint3 tgpig[[threadgroup_position_in_grid]], - ushort tiisg[[thread_index_in_simdgroup]]) { - - const int nrows = args.ne11; - const int r0 = tgpig.x; - const int im = tgpig.z; - - const uint i12 = im%args.ne12; - const uint i13 = im/args.ne12; - - const uint64_t offset0 = r0*args.nb01 + (i12/args.r2)*args.nb02 + (i13/args.r3)*args.nb03; - - device const T4 * x4 = (device const T4 *) (src0 + offset0); - - device float * dst_f32 = (device float *) dst + (uint64_t)im*args.ne0*args.ne1; - - for (int r1 = 0; r1 < nrows; ++r1) { - const uint64_t offset1 = r1*args.nb11 + (i12 )*args.nb12 + (i13 )*args.nb13; - - device const float4 * y4 = (device const float4 *) (src1 + offset1); - - float sumf = 0; - for (int i = tiisg; i < args.ne00/4; i += 32) { - sumf += dot((float4) x4[i], y4[i]); - } - - float sum_all = simd_sum(sumf); - if (tiisg == 0) { - dst_f32[(uint64_t)r1*args.ne0 + r0] = sum_all; - } - } -} - -typedef decltype(kernel_mul_mv_l4) mul_mv_l4_t; - -template [[host_name("kernel_mul_mv_f16_f32_l4")]] kernel mul_mv_l4_t kernel_mul_mv_l4; -#if defined(GGML_METAL_USE_BF16) -template [[host_name("kernel_mul_mv_bf16_f32_l4")]] kernel mul_mv_l4_t kernel_mul_mv_l4; +template [[host_name("kernel_mul_mv_f16_f16_c4")]] kernel mul_mv_c4_t kernel_mul_mv_c4; +#if defined(GGML_METAL_HAS_BF16) +template [[host_name("kernel_mul_mv_bf16_f32_c4")]] kernel mul_mv_c4_t kernel_mul_mv_c4; +template [[host_name("kernel_mul_mv_bf16_bf16_c4")]] kernel mul_mv_c4_t kernel_mul_mv_c4; #endif static float rope_yarn_ramp(const float low, const float high, const int i0) { @@ -3879,62 +3984,63 @@ template [[host_name("kernel_rope_multi_f16")]] kernel kernel_rope_multi_t kerne template [[host_name("kernel_rope_vision_f32")]] kernel kernel_rope_vision_t kernel_rope_vision; template [[host_name("kernel_rope_vision_f16")]] kernel kernel_rope_vision_t kernel_rope_vision; -typedef void (im2col_t)( - device const float * x, - device char * dst, - constant ggml_metal_kargs_im2col & args, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tgpg[[threadgroups_per_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]); - -template -kernel void kernel_im2col( - device const float * x, - device char * dst, - constant ggml_metal_kargs_im2col & args, - uint3 tgpig[[threadgroup_position_in_grid]], - uint3 tgpg[[threadgroups_per_grid]], - uint3 tpitg[[thread_position_in_threadgroup]], - uint3 ntg[[threads_per_threadgroup]]) { -// const int64_t IC = tgpg[0]; - const int64_t OH = tgpg[1]; - const int64_t OW = tgpg[2]; - -// const int64_t N = ntg[0]; - const int64_t KH = ntg[1]; - const int64_t KW = ntg[2]; - - const int64_t in = tpitg[0]; - const int64_t ikh = tpitg[1]; - const int64_t ikw = tpitg[2]; - - const int64_t iic = tgpig[0]; - const int64_t ioh = tgpig[1]; - const int64_t iow = tgpig[2]; - - const int64_t iiw = iow*args.s0 + ikw*args.d0 - args.p0; - const int64_t iih = ioh*args.s1 + ikh*args.d1 - args.p1; - - const int64_t offset_dst = (in*OH*OW + ioh*OW + iow)*args.CHW + (iic*(KH*KW) + ikh*KW + ikw); - - device T * pdst = (device T *) (dst); - - if (iih < 0 || iih >= args.IH || iiw < 0 || iiw >= args.IW) { - pdst[offset_dst] = 0.0f; - } else { - const int64_t offset_src = in*args.ofs0 + iic*args.ofs1 + iih*args.IW + iiw; - pdst[offset_dst] = x[offset_src]; - } -} - -template [[host_name("kernel_im2col_f32")]] kernel im2col_t kernel_im2col; -template [[host_name("kernel_im2col_f16")]] kernel im2col_t kernel_im2col; +// TODO: obolete -- remove +//typedef void (im2col_t)( +// constant ggml_metal_kargs_im2col & args, +// device const float * x, +// device char * dst, +// uint3 tgpig[[threadgroup_position_in_grid]], +// uint3 tgpg[[threadgroups_per_grid]], +// uint3 tpitg[[thread_position_in_threadgroup]], +// uint3 ntg[[threads_per_threadgroup]]); +// +//template +//kernel void kernel_im2col( +// constant ggml_metal_kargs_im2col & args, +// device const float * x, +// device char * dst, +// uint3 tgpig[[threadgroup_position_in_grid]], +// uint3 tgpg[[threadgroups_per_grid]], +// uint3 tpitg[[thread_position_in_threadgroup]], +// uint3 ntg[[threads_per_threadgroup]]) { +//// const int64_t IC = tgpg[0]; +// const int64_t OH = tgpg[1]; +// const int64_t OW = tgpg[2]; +// +//// const int64_t N = ntg[0]; +// const int64_t KH = ntg[1]; +// const int64_t KW = ntg[2]; +// +// const int64_t in = tpitg[0]; +// const int64_t ikh = tpitg[1]; +// const int64_t ikw = tpitg[2]; +// +// const int64_t iic = tgpig[0]; +// const int64_t ioh = tgpig[1]; +// const int64_t iow = tgpig[2]; +// +// const int64_t iiw = iow*args.s0 + ikw*args.d0 - args.p0; +// const int64_t iih = ioh*args.s1 + ikh*args.d1 - args.p1; +// +// const int64_t offset_dst = (in*OH*OW + ioh*OW + iow)*args.CHW + (iic*(KH*KW) + ikh*KW + ikw); +// +// device T * pdst = (device T *) (dst); +// +// if (iih < 0 || iih >= args.IH || iiw < 0 || iiw >= args.IW) { +// pdst[offset_dst] = 0.0f; +// } else { +// const int64_t offset_src = in*args.ofs0 + iic*args.ofs1 + iih*args.IW + iiw; +// pdst[offset_dst] = x[offset_src]; +// } +//} +// +//template [[host_name("kernel_im2col_f32")]] kernel im2col_t kernel_im2col; +//template [[host_name("kernel_im2col_f16")]] kernel im2col_t kernel_im2col; typedef void (im2col_ext_t)( + constant ggml_metal_kargs_im2col & args, device const float * x, device char * dst, - constant ggml_metal_kargs_im2col & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tgpg[[threadgroups_per_grid]], uint3 tpitg[[thread_position_in_threadgroup]], @@ -3942,16 +4048,16 @@ typedef void (im2col_ext_t)( template kernel void kernel_im2col_ext( + constant ggml_metal_kargs_im2col & args, device const float * x, device char * dst, - constant ggml_metal_kargs_im2col & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tgpg[[threadgroups_per_grid]], // tgpg[0] = D x IC x KH x KW, CHW = IC x KH x KW uint3 tpitg[[thread_position_in_threadgroup]], uint3 ntg[[threads_per_threadgroup]]) { // [M, 1, 1] const int64_t KHW = (int64_t)args.KHW; - const int64_t d = tgpig[0] / args.CHW; + const int64_t d = tgpig[0] / args.CHW; const int64_t chw = tgpig[0] % args.CHW; const int64_t tgpig_0 = chw / KHW; // 0 ~ (IC - 1) const int64_t HW = tgpig[0] % KHW; @@ -3985,19 +4091,19 @@ template [[host_name("kernel_im2col_ext_f32")]] kernel im2col_ext_t kernel_im2co template [[host_name("kernel_im2col_ext_f16")]] kernel im2col_ext_t kernel_im2col_ext; typedef void (conv_transpose_1d_t)( + constant ggml_metal_kargs_conv_transpose_1d & args, device const float * src0, device const float * src1, device char * dst, - constant ggml_metal_kargs_conv_transpose_1d & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tgpg[[threadgroups_per_grid]]); template kernel void kernel_conv_transpose_1d( + constant ggml_metal_kargs_conv_transpose_1d & args, device const T * src0, device const float * src1, device char * dst, - constant ggml_metal_kargs_conv_transpose_1d & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tgpg[[threadgroups_per_grid]]) { @@ -4021,26 +4127,26 @@ kernel void kernel_conv_transpose_1d( template [[host_name("kernel_conv_transpose_1d_f32_f32")]] kernel void kernel_conv_transpose_1d( + constant ggml_metal_kargs_conv_transpose_1d & args, device const float * src0, device const float * src1, device char * dst, - constant ggml_metal_kargs_conv_transpose_1d & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tgpg[[threadgroups_per_grid]]); template [[host_name("kernel_conv_transpose_1d_f16_f32")]] kernel void kernel_conv_transpose_1d( + constant ggml_metal_kargs_conv_transpose_1d & args, device const half * src0, device const float * src1, device char * dst, - constant ggml_metal_kargs_conv_transpose_1d & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tgpg[[threadgroups_per_grid]]); kernel void kernel_upscale_f32( + constant ggml_metal_kargs_upscale & args, device const char * src0, device char * dst, - constant ggml_metal_kargs_upscale & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tpitg[[thread_position_in_threadgroup]], uint3 ntg[[threads_per_threadgroup]]) { @@ -4064,9 +4170,9 @@ kernel void kernel_upscale_f32( } kernel void kernel_pad_f32( + constant ggml_metal_kargs_pad & args, device const char * src0, device char * dst, - constant ggml_metal_kargs_pad & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tpitg[[thread_position_in_threadgroup]], uint3 ntg[[threads_per_threadgroup]]) { @@ -4100,9 +4206,9 @@ kernel void kernel_pad_f32( } kernel void kernel_pad_reflect_1d_f32( + constant ggml_metal_kargs_pad_reflect_1d & args, device const char * src0, device char * dst, - constant ggml_metal_kargs_pad_reflect_1d & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tgpg[[threadgroups_per_grid]], uint3 tpitg[[thread_position_in_threadgroup]], @@ -4133,8 +4239,8 @@ kernel void kernel_pad_reflect_1d_f32( } kernel void kernel_arange_f32( - device char * dst, constant ggml_metal_kargs_arange & args, + device char * dst, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tpitg[[thread_position_in_threadgroup]], uint3 ntg[[threads_per_threadgroup]]) { @@ -4147,9 +4253,9 @@ kernel void kernel_arange_f32( } kernel void kernel_timestep_embedding_f32( + constant ggml_metal_kargs_timestep_embedding & args, device const char * src0, device char * dst, - constant ggml_metal_kargs_timestep_embedding & args, uint3 tgpig[[threadgroup_position_in_grid]], uint3 tpitg[[thread_position_in_threadgroup]], uint3 ntg[[threads_per_threadgroup]]) { @@ -4173,19 +4279,19 @@ kernel void kernel_timestep_embedding_f32( // bitonic sort implementation following the CUDA kernels as reference typedef void (argsort_t)( - device const float * x, - device int32_t * dst, constant ggml_metal_kargs_argsort & args, + device const float * x, + device int32_t * dst, threadgroup int32_t * shared_values [[threadgroup(0)]], uint3 tgpig[[threadgroup_position_in_grid]], uint3 tpitg[[thread_position_in_threadgroup]]); template kernel void kernel_argsort_f32_i32( - device const float * x, - device int32_t * dst, constant ggml_metal_kargs_argsort & args, - threadgroup int32_t * shared_values [[threadgroup(0)]], + device const float * x, + device int32_t * dst, + threadgroup int32_t * shared_values [[threadgroup(0)]], uint3 tgpig[[threadgroup_position_in_grid]], uint3 tpitg[[thread_position_in_threadgroup]]) { // bitonic sort @@ -4238,11 +4344,21 @@ template [[host_name("kernel_argsort_f32_i32_asc")]] kernel argsort_t kernel_ar template [[host_name("kernel_argsort_f32_i32_desc")]] kernel argsort_t kernel_argsort_f32_i32; kernel void kernel_leaky_relu_f32( + constant ggml_metal_kargs_leaky_relu & args, device const float * src0, device float * dst, - constant ggml_metal_kargs_leaky_relu & args, uint tpig[[thread_position_in_grid]]) { - dst[tpig] = src0[tpig] > 0.0f ? src0[tpig] : src0[tpig] * args.slope; + const float x = src0[tpig]; + dst[tpig] = x > 0.0f ? x : x * args.slope; +} + +kernel void kernel_leaky_relu_f32_4( + constant ggml_metal_kargs_leaky_relu & args, + device const float4 * src0, + device float4 * dst, + uint tpig[[thread_position_in_grid]]) { + const float4 x = src0[tpig]; + dst[tpig] = float4(x > 0.0f)*x + float4(x <= 0.0f)*(x * args.slope); } constant bool FC_flash_attn_ext_has_mask [[function_constant(FC_FLASH_ATTN_EXT + 0)]]; @@ -4884,7 +5000,7 @@ template [[host_name("kernel_flash_attn_ext_f16_dk192_dv128")]] kernel flash_at template [[host_name("kernel_flash_attn_ext_f16_dk256_dv256")]] kernel flash_attn_ext_t kernel_flash_attn_ext; template [[host_name("kernel_flash_attn_ext_f16_dk576_dv512")]] kernel flash_attn_ext_t kernel_flash_attn_ext; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_flash_attn_ext_bf16_dk40_dv40" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; template [[host_name("kernel_flash_attn_ext_bf16_dk64_dv64" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; template [[host_name("kernel_flash_attn_ext_bf16_dk80_dv80" )]] kernel flash_attn_ext_t kernel_flash_attn_ext; @@ -5450,7 +5566,7 @@ kernel void kernel_flash_attn_ext_vec( typedef decltype(kernel_flash_attn_ext_vec) flash_attn_ext_vec_t; template [[host_name("kernel_flash_attn_ext_vec_f16_dk64_dv64")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_flash_attn_ext_vec_bf16_dk64_dv64")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; #endif template [[host_name("kernel_flash_attn_ext_vec_q4_0_dk64_dv64")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; @@ -5460,7 +5576,7 @@ template [[host_name("kernel_flash_attn_ext_vec_q5_1_dk64_dv64")]] kernel flas template [[host_name("kernel_flash_attn_ext_vec_q8_0_dk64_dv64")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; template [[host_name("kernel_flash_attn_ext_vec_f16_dk96_dv96")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_flash_attn_ext_vec_bf16_dk96_dv96")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; #endif template [[host_name("kernel_flash_attn_ext_vec_q4_0_dk96_dv96")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; @@ -5470,7 +5586,7 @@ template [[host_name("kernel_flash_attn_ext_vec_q5_1_dk96_dv96")]] kernel flas template [[host_name("kernel_flash_attn_ext_vec_q8_0_dk96_dv96")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; template [[host_name("kernel_flash_attn_ext_vec_f16_dk128_dv128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_flash_attn_ext_vec_bf16_dk128_dv128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; #endif template [[host_name("kernel_flash_attn_ext_vec_q4_0_dk128_dv128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; @@ -5480,7 +5596,7 @@ template [[host_name("kernel_flash_attn_ext_vec_q5_1_dk128_dv128")]] kernel flas template [[host_name("kernel_flash_attn_ext_vec_q8_0_dk128_dv128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; template [[host_name("kernel_flash_attn_ext_vec_f16_dk192_dv192")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_flash_attn_ext_vec_bf16_dk192_dv192")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; #endif template [[host_name("kernel_flash_attn_ext_vec_q4_0_dk192_dv192")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; @@ -5490,7 +5606,7 @@ template [[host_name("kernel_flash_attn_ext_vec_q5_1_dk192_dv192")]] kernel flas template [[host_name("kernel_flash_attn_ext_vec_q8_0_dk192_dv192")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; template [[host_name("kernel_flash_attn_ext_vec_f16_dk192_dv128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_flash_attn_ext_vec_bf16_dk192_dv128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; #endif template [[host_name("kernel_flash_attn_ext_vec_q4_0_dk192_dv128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; @@ -5500,7 +5616,7 @@ template [[host_name("kernel_flash_attn_ext_vec_q5_1_dk192_dv128")]] kernel flas template [[host_name("kernel_flash_attn_ext_vec_q8_0_dk192_dv128")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; template [[host_name("kernel_flash_attn_ext_vec_f16_dk256_dv256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_flash_attn_ext_vec_bf16_dk256_dv256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; #endif template [[host_name("kernel_flash_attn_ext_vec_q4_0_dk256_dv256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; @@ -5510,7 +5626,7 @@ template [[host_name("kernel_flash_attn_ext_vec_q5_1_dk256_dv256")]] kernel flas template [[host_name("kernel_flash_attn_ext_vec_q8_0_dk256_dv256")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; template [[host_name("kernel_flash_attn_ext_vec_f16_dk576_dv512")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_flash_attn_ext_vec_bf16_dk576_dv512")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; #endif template [[host_name("kernel_flash_attn_ext_vec_q4_0_dk576_dv512")]] kernel flash_attn_ext_vec_t kernel_flash_attn_ext_vec; @@ -5603,12 +5719,12 @@ template [[host_name("kernel_cpy_f32_f32")]] kernel kernel_cpy_t kernel_cpy; template [[host_name("kernel_cpy_f32_i32")]] kernel kernel_cpy_t kernel_cpy; template [[host_name("kernel_cpy_i32_f32")]] kernel kernel_cpy_t kernel_cpy; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_cpy_f32_bf16")]] kernel kernel_cpy_t kernel_cpy; #endif template [[host_name("kernel_cpy_f16_f32")]] kernel kernel_cpy_t kernel_cpy; template [[host_name("kernel_cpy_f16_f16")]] kernel kernel_cpy_t kernel_cpy; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_cpy_bf16_f32")]] kernel kernel_cpy_t kernel_cpy; template [[host_name("kernel_cpy_bf16_bf16")]] kernel kernel_cpy_t kernel_cpy; #endif @@ -5850,7 +5966,7 @@ kernel void kernel_concat( } } -template +template void kernel_mul_mv_q2_K_f32_impl( args_t args, device const char * src0, @@ -5860,13 +5976,15 @@ void kernel_mul_mv_q2_K_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; const int nb = args.ne00/QK_K; + const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -5950,10 +6068,10 @@ kernel void kernel_mul_mv_q2_K_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_q2_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_q2_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_q3_K_f32_impl( args_t args, device const char * src0, @@ -5963,6 +6081,7 @@ void kernel_mul_mv_q3_K_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; const int nb = args.ne00/QK_K; @@ -5970,7 +6089,7 @@ void kernel_mul_mv_q3_K_f32_impl( const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -6114,10 +6233,10 @@ kernel void kernel_mul_mv_q3_K_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_q3_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_q3_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_q4_K_f32_impl( args_t args, device const char * src0, @@ -6127,6 +6246,8 @@ void kernel_mul_mv_q4_K_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; + const uint16_t kmask1 = 0x3f3f; const uint16_t kmask2 = 0x0f0f; const uint16_t kmask3 = 0xc0c0; @@ -6142,7 +6263,7 @@ void kernel_mul_mv_q4_K_f32_impl( const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -6236,10 +6357,10 @@ kernel void kernel_mul_mv_q4_K_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_q4_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_q4_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_q5_K_f32_impl( args_t args, device const char * src0, @@ -6249,6 +6370,7 @@ void kernel_mul_mv_q5_K_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; const int nb = args.ne00/QK_K; @@ -6256,7 +6378,7 @@ void kernel_mul_mv_q5_K_f32_impl( const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -6367,10 +6489,10 @@ kernel void kernel_mul_mv_q5_K_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_q5_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_q5_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_q6_K_f32_impl( args_t args, device const char * src0, @@ -6380,6 +6502,7 @@ void kernel_mul_mv_q6_K_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; const uint8_t kmask1 = 0x03; const uint8_t kmask2 = 0x0C; @@ -6392,7 +6515,7 @@ void kernel_mul_mv_q6_K_f32_impl( const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -6476,12 +6599,12 @@ kernel void kernel_mul_mv_q6_K_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_q6_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_q6_K_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } // ======================= "True" 2-bit -template +template void kernel_mul_mv_iq2_xxs_f32_impl( args_t args, device const char * src0, @@ -6491,13 +6614,15 @@ void kernel_mul_mv_iq2_xxs_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; const int nb = args.ne00/QK_K; + const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -6584,10 +6709,10 @@ kernel void kernel_mul_mv_iq2_xxs_f32( uint3 tgpig[[threadgroup_position_in_grid]], ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq2_xxs_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + kernel_mul_mv_iq2_xxs_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_iq2_xs_f32_impl( args_t args, device const char * src0, @@ -6597,13 +6722,15 @@ void kernel_mul_mv_iq2_xs_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; const int nb = args.ne00/QK_K; + const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -6701,10 +6828,10 @@ kernel void kernel_mul_mv_iq2_xs_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq2_xs_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + kernel_mul_mv_iq2_xs_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_iq3_xxs_f32_impl( args_t args, device const char * src0, @@ -6714,13 +6841,15 @@ void kernel_mul_mv_iq3_xxs_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; const int nb = args.ne00/QK_K; + const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -6811,10 +6940,10 @@ kernel void kernel_mul_mv_iq3_xxs_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq3_xxs_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + kernel_mul_mv_iq3_xxs_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_iq3_s_f32_impl( args_t args, device const char * src0, @@ -6824,13 +6953,15 @@ void kernel_mul_mv_iq3_s_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; const int nb = args.ne00/QK_K; + const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -6921,10 +7052,10 @@ kernel void kernel_mul_mv_iq3_s_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq3_s_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + kernel_mul_mv_iq3_s_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_iq2_s_f32_impl( args_t args, device const char * src0, @@ -6934,13 +7065,15 @@ void kernel_mul_mv_iq2_s_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; const int nb = args.ne00/QK_K; + const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -7032,10 +7165,10 @@ kernel void kernel_mul_mv_iq2_s_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq2_s_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + kernel_mul_mv_iq2_s_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_iq1_s_f32_impl( args_t args, device const char * src0, @@ -7045,13 +7178,15 @@ void kernel_mul_mv_iq1_s_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; const int nb = args.ne00/QK_K; + const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -7129,10 +7264,10 @@ kernel void kernel_mul_mv_iq1_s_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq1_s_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_iq1_s_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_iq1_m_f32_impl( args_t args, device const char * src0, @@ -7142,6 +7277,7 @@ void kernel_mul_mv_iq1_m_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; const int nb = args.ne00/QK_K; @@ -7149,7 +7285,7 @@ void kernel_mul_mv_iq1_m_f32_impl( const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -7237,10 +7373,10 @@ kernel void kernel_mul_mv_iq1_m_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq1_m_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); + kernel_mul_mv_iq1_m_f32_impl(args, src0, src1, dst, nullptr, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_iq4_nl_f32_impl( args_t args, device const char * src0, @@ -7250,6 +7386,7 @@ void kernel_mul_mv_iq4_nl_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; threadgroup float * shmem_f32 = (threadgroup float *) shmem; const int nb = args.ne00/QK4_NL; @@ -7258,7 +7395,7 @@ void kernel_mul_mv_iq4_nl_f32_impl( const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -7343,10 +7480,10 @@ kernel void kernel_mul_mv_iq4_nl_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq4_nl_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + kernel_mul_mv_iq4_nl_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_iq4_xs_f32_impl( args_t args, device const char * src0, @@ -7356,13 +7493,15 @@ void kernel_mul_mv_iq4_xs_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; threadgroup float * shmem_f32 = (threadgroup float *) shmem; const int nb = args.ne00/QK_K; + const int r0 = tgpig.x; const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -7448,10 +7587,10 @@ kernel void kernel_mul_mv_iq4_xs_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_iq4_xs_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + kernel_mul_mv_iq4_xs_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } -template +template void kernel_mul_mv_mxfp4_f32_impl( args_t args, device const char * src0, @@ -7461,6 +7600,7 @@ void kernel_mul_mv_mxfp4_f32_impl( uint3 tgpig, ushort tiisg, ushort sgitg) { + const short NSG = FC_mul_mv_nsg; threadgroup float * shmem_f32 = (threadgroup float *) shmem; const int nb = args.ne00/QK_MXFP4; @@ -7469,7 +7609,7 @@ void kernel_mul_mv_mxfp4_f32_impl( const int r1 = tgpig.y; const int im = tgpig.z; - const int first_row = (r0 * nsg + sgitg) * nr0; + const int first_row = (r0 * NSG + sgitg) * nr0; const uint i12 = im%args.ne12; const uint i13 = im/args.ne12; @@ -7537,7 +7677,7 @@ kernel void kernel_mul_mv_mxfp4_f32( ushort tiisg[[thread_index_in_simdgroup]], ushort sgitg[[simdgroup_index_in_threadgroup]]) { - kernel_mul_mv_mxfp4_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); + kernel_mul_mv_mxfp4_f32_impl(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } template @@ -7880,13 +8020,13 @@ kernel void kernel_mul_mm_id_map0( typedef decltype(kernel_mul_mm_id_map0<1>) kernel_mul_mm_id_map0_t; -template [[host_name("kernel_mul_mm_id_map0_f16_ne20_1" )]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<1>; -template [[host_name("kernel_mul_mm_id_map0_f16_ne20_2" )]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<2>; -template [[host_name("kernel_mul_mm_id_map0_f16_ne20_4" )]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<4>; -template [[host_name("kernel_mul_mm_id_map0_f16_ne20_6" )]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<6>; -template [[host_name("kernel_mul_mm_id_map0_f16_ne20_8" )]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<8>; -template [[host_name("kernel_mul_mm_id_map0_f16_ne20_10")]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<10>; -template [[host_name("kernel_mul_mm_id_map0_f16_ne20_16")]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<16>; +template [[host_name("kernel_mul_mm_id_map0_ne20_1" )]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<1>; +template [[host_name("kernel_mul_mm_id_map0_ne20_2" )]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<2>; +template [[host_name("kernel_mul_mm_id_map0_ne20_4" )]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<4>; +template [[host_name("kernel_mul_mm_id_map0_ne20_6" )]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<6>; +template [[host_name("kernel_mul_mm_id_map0_ne20_8" )]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<8>; +template [[host_name("kernel_mul_mm_id_map0_ne20_10")]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<10>; +template [[host_name("kernel_mul_mm_id_map0_ne20_16")]] kernel kernel_mul_mm_id_map0_t kernel_mul_mm_id_map0<16>; template kernel void kernel_mul_mm_id( @@ -8050,7 +8190,7 @@ typedef decltype(kernel_get_rows_f) get_rows_f_t; template [[host_name("kernel_get_rows_f32")]] kernel get_rows_f_t kernel_get_rows_f; template [[host_name("kernel_get_rows_f16")]] kernel get_rows_f_t kernel_get_rows_f; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_get_rows_bf16")]] kernel get_rows_f_t kernel_get_rows_f; #endif @@ -8085,7 +8225,7 @@ typedef decltype(kernel_set_rows_f) set_rows_f_t; template [[host_name("kernel_set_rows_f32")]] kernel set_rows_f_t kernel_set_rows_f; template [[host_name("kernel_set_rows_f16")]] kernel set_rows_f_t kernel_set_rows_f; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_set_rows_bf16")]] kernel set_rows_f_t kernel_set_rows_f; #endif @@ -8106,7 +8246,7 @@ typedef decltype(kernel_mul_mm; template [[host_name("kernel_mul_mm_f16_f32")]] kernel mul_mm_t kernel_mul_mm; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_mul_mm_bf16_f32")]] kernel mul_mm_t kernel_mul_mm; #endif template [[host_name("kernel_mul_mm_q4_0_f32")]] kernel mul_mm_t kernel_mul_mm; @@ -8138,7 +8278,7 @@ typedef decltype(kernel_mul_mm_id; template [[host_name("kernel_mul_mm_id_f16_f16")]] kernel mul_mm_id kernel_mul_mm_id; -#if defined(GGML_METAL_USE_BF16) +#if defined(GGML_METAL_HAS_BF16) template [[host_name("kernel_mul_mm_id_bf16_f16")]] kernel mul_mm_id kernel_mul_mm_id; #endif template [[host_name("kernel_mul_mm_id_q4_0_f16")]] kernel mul_mm_id kernel_mul_mm_id; @@ -8213,7 +8353,7 @@ void mmv_fn( impl_fn(args, src0, src1, dst, shmem, tgpig, tiisg, sgitg); } -typedef decltype(mmv_fn>) mul_mv_impl_fn_t; +typedef decltype(mmv_fn>) mul_mv_impl_fn_t; template kernel void kernel_mul_mv_id( @@ -8278,44 +8418,52 @@ kernel void kernel_mul_mv_id( sgitg); } -typedef decltype(kernel_mul_mv_id>>) kernel_mul_mv_id_t; +typedef decltype(kernel_mul_mv_id>>) kernel_mul_mv_id_t; -template [[host_name("kernel_mul_mv_id_f32_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_f16_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -#if defined(GGML_METAL_USE_BF16) -template [[host_name("kernel_mul_mv_id_bf16_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +typedef decltype(kernel_mul_mv_id>>) kernel_mul_mv_id_4_t; + +template [[host_name("kernel_mul_mv_id_f32_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_f16_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +#if defined(GGML_METAL_HAS_BF16) +template [[host_name("kernel_mul_mv_id_bf16_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +#endif +template [[host_name("kernel_mul_mv_id_f32_f32_4")]] kernel kernel_mul_mv_id_4_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_f16_f32_4")]] kernel kernel_mul_mv_id_4_t kernel_mul_mv_id>>; +#if defined(GGML_METAL_HAS_BF16) +template [[host_name("kernel_mul_mv_id_bf16_f32_4")]] kernel kernel_mul_mv_id_4_t kernel_mul_mv_id>>; #endif -template [[host_name("kernel_mul_mv_id_q8_0_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_q4_0_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_q4_1_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_q5_0_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_q5_1_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_q8_0_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_mxfp4_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_q4_0_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_q4_1_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_q5_0_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_q5_1_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_q2_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_q3_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_q4_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_q5_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_q6_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_iq1_s_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_iq1_m_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_iq2_xxs_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_iq2_xs_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_iq3_xxs_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_iq3_s_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_iq2_s_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_iq4_nl_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; -template [[host_name("kernel_mul_mv_id_iq4_xs_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_mxfp4_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; + +template [[host_name("kernel_mul_mv_id_q2_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_q3_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_q4_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_q5_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_q6_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_iq1_s_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_iq1_m_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_iq2_xxs_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_iq2_xs_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_iq3_xxs_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_iq3_s_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_iq2_s_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_iq4_nl_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; +template [[host_name("kernel_mul_mv_id_iq4_xs_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id>>; kernel void kernel_pool_2d_max_f32( + constant ggml_metal_kargs_pool_2d & args, device const float * src0, device float * dst, - constant ggml_metal_kargs_pool_2d & args, uint gid[[thread_position_in_grid]]) { - if (gid >= args.parallel_elements) { + if (gid >= args.np) { return; } @@ -8348,12 +8496,12 @@ kernel void kernel_pool_2d_max_f32( } kernel void kernel_pool_2d_avg_f32( + constant ggml_metal_kargs_pool_2d & args, device const float * src0, device float * dst, - constant ggml_metal_kargs_pool_2d & args, uint gid[[thread_position_in_grid]]) { - if (gid >= args.parallel_elements) { + if (gid >= args.np) { return; } diff --git a/ggml/src/ggml-webgpu/ggml-webgpu.cpp b/ggml/src/ggml-webgpu/ggml-webgpu.cpp index df6a3ed95a..8e7b986df8 100644 --- a/ggml/src/ggml-webgpu/ggml-webgpu.cpp +++ b/ggml/src/ggml-webgpu/ggml-webgpu.cpp @@ -116,6 +116,10 @@ struct webgpu_context_struct { wgpu::Queue queue; wgpu::Limits limits; + // Separate this out from limits since on some Metal systems, the limit returned by + // querying the limits is higher than the actual allowed maximum. + uint32_t max_wg_size_x; + std::recursive_mutex mutex; webgpu_buf_pool param_buf_pool; @@ -124,7 +128,15 @@ struct webgpu_context_struct { wgpu::ComputePipeline memset_pipeline; wgpu::ComputePipeline mul_mat_pipeline[30][2]; wgpu::ComputePipeline set_rows_pipeline; + wgpu::ComputePipeline get_rows_pipeline[30]; + wgpu::ComputePipeline get_rows_f32_no_vec_pipeline; wgpu::ComputePipeline cpy_pipeline; + wgpu::ComputePipeline add_pipeline[2]; + wgpu::ComputePipeline add_ip_pipeline[2]; + wgpu::ComputePipeline mul_pipeline[2]; + wgpu::ComputePipeline mul_ip_pipeline[2]; + wgpu::ComputePipeline rms_norm_pipeline; + wgpu::ComputePipeline rms_norm_ip_pipeline; size_t memset_bytes_per_thread; @@ -232,14 +244,15 @@ static void ggml_backend_webgpu_wait_on_submission(webgpu_context & ctx) { std::lock_guard lock(ctx->mutex); if (ctx->callback_futures.empty()) { // no existing callbacks, wait on queue submission - ctx->instance.WaitAny(ctx->queue.OnSubmittedWorkDone( - wgpu::CallbackMode::AllowSpontaneous, - [](wgpu::QueueWorkDoneStatus status, wgpu::StringView message) { - if (status != wgpu::QueueWorkDoneStatus::Success) { - GGML_LOG_ERROR("ggml_webgpu: Failed to submit commands: %s\n", std::string(message).c_str()); - } - }), - UINT64_MAX); + ctx->instance.WaitAny( + ctx->queue.OnSubmittedWorkDone(wgpu::CallbackMode::AllowSpontaneous, + [](wgpu::QueueWorkDoneStatus status, wgpu::StringView message) { + if (status != wgpu::QueueWorkDoneStatus::Success) { + GGML_LOG_ERROR("ggml_webgpu: Failed to submit commands: %s\n", + std::string(message).c_str()); + } + }), + UINT64_MAX); } else { // existing callbacks, wait on them ctx->instance.WaitAny(ctx->callback_futures.size(), ctx->callback_futures.data(), UINT64_MAX); @@ -286,10 +299,7 @@ static void ggml_backend_webgpu_submit_queue(webgpu_context & ctx) { // Check for errrors in SET_ROWS operations for (auto & error_bufs : staged_set_row_error_bufs) { wgpu::Future f = error_bufs.host_buf.MapAsync( - wgpu::MapMode::Read, - 0, - error_bufs.host_buf.GetSize(), - wgpu::CallbackMode::AllowSpontaneous, + wgpu::MapMode::Read, 0, error_bufs.host_buf.GetSize(), wgpu::CallbackMode::AllowSpontaneous, [ctx, error_bufs](wgpu::MapAsyncStatus status, wgpu::StringView message) { if (status != wgpu::MapAsyncStatus::Success) { GGML_LOG_ERROR("ggml_webgpu: Failed to map error buffer: %s\n", std::string(message).c_str()); @@ -311,10 +321,7 @@ static void ggml_backend_webgpu_map_buffer(webgpu_context & ctx, wgpu::MapMode mode, size_t offset, size_t size) { - ctx->instance.WaitAny(buffer.MapAsync(mode, - offset, - size, - wgpu::CallbackMode::AllowSpontaneous, + ctx->instance.WaitAny(buffer.MapAsync(mode, offset, size, wgpu::CallbackMode::AllowSpontaneous, [](wgpu::MapAsyncStatus status, wgpu::StringView message) { if (status != wgpu::MapAsyncStatus::Success) { GGML_LOG_ERROR("ggml_webgpu: Failed to map buffer: %s\n", @@ -351,7 +358,8 @@ static void ggml_backend_webgpu_build_and_enqueue(webgpu_context & std::vector params, std::vector bind_group_entries, uint32_t wg_x, - bool submit_and_wait = false) { + const char * bind_group_label = nullptr, + bool submit_and_wait = false) { webgpu_pool_bufs params_bufs = ctx->param_buf_pool.alloc_bufs(); ggml_backend_webgpu_map_buffer(ctx, params_bufs.host_buf, wgpu::MapMode::Write, 0, params_bufs.host_buf.GetSize()); @@ -372,6 +380,9 @@ static void ggml_backend_webgpu_build_and_enqueue(webgpu_context & bind_group_desc.layout = pipeline.GetBindGroupLayout(0); bind_group_desc.entryCount = bind_group_entries.size(); bind_group_desc.entries = bind_group_entries.data(); + if (bind_group_label) { + bind_group_desc.label = bind_group_label; + } wgpu::BindGroup bind_group = ctx->device.CreateBindGroup(&bind_group_desc); wgpu::CommandEncoder encoder = ctx->device.CreateCommandEncoder(); @@ -415,9 +426,9 @@ static void ggml_backend_webgpu_buffer_memset(webgpu_context & ctx, std::vector entries = { { .binding = 0, .buffer = buf, .offset = 0, .size = buf.GetSize() } }; - size_t bytes_per_wg = ctx->limits.maxComputeWorkgroupSizeX * ctx->memset_bytes_per_thread; + size_t bytes_per_wg = ctx->max_wg_size_x * ctx->memset_bytes_per_thread; uint32_t wg_x = ((size + 3) + bytes_per_wg - 1) / bytes_per_wg; - ggml_backend_webgpu_build_and_enqueue(ctx, ctx->memset_pipeline, params, entries, wg_x, true); + ggml_backend_webgpu_build_and_enqueue(ctx, ctx->memset_pipeline, params, entries, wg_x, "MEMSET", true); } /** End WebGPU Actions */ @@ -461,26 +472,26 @@ static size_t ggml_webgpu_tensor_binding_size(webgpu_context & ctx, ggml_tensor ~(WEBGPU_STORAGE_BUF_BINDING_MULT - 1); } +// Used to determine if two tensors are the same for in-place operations +static bool ggml_webgpu_tensor_equal(ggml_tensor * a, ggml_tensor * b) { + return (ggml_webgpu_tensor_buf(a).Get() == ggml_webgpu_tensor_buf(b).Get()) && + (ggml_webgpu_tensor_offset(a) == ggml_webgpu_tensor_offset(b)); +} + static void ggml_webgpu_cpy(webgpu_context & ctx, ggml_tensor * src, ggml_tensor * dst) { uint32_t ne = (uint32_t) ggml_nelements(dst); - std::vector params = { ne, - (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, src) / ggml_type_size(src->type)), - (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, dst) / ggml_type_size(dst->type)), - // Convert byte-strides to element-strides - (uint32_t) (src->nb[0] / ggml_type_size(src->type)), - (uint32_t) (src->nb[1] / ggml_type_size(src->type)), - (uint32_t) (src->nb[2] / ggml_type_size(src->type)), - (uint32_t) (src->nb[3] / ggml_type_size(src->type)), - (uint32_t) (dst->nb[0] / ggml_type_size(dst->type)), - (uint32_t) (dst->nb[1] / ggml_type_size(dst->type)), - (uint32_t) (dst->nb[2] / ggml_type_size(dst->type)), - (uint32_t) (dst->nb[3] / ggml_type_size(dst->type)), - // Logical shape β€” same for both tensors even if permuted - (uint32_t) src->ne[0], - (uint32_t) src->ne[1], - (uint32_t) src->ne[2], - (uint32_t) src->ne[3] }; + std::vector params = { + ne, (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, src) / ggml_type_size(src->type)), + (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, dst) / ggml_type_size(dst->type)), + // Convert byte-strides to element-strides + (uint32_t) (src->nb[0] / ggml_type_size(src->type)), (uint32_t) (src->nb[1] / ggml_type_size(src->type)), + (uint32_t) (src->nb[2] / ggml_type_size(src->type)), (uint32_t) (src->nb[3] / ggml_type_size(src->type)), + (uint32_t) (dst->nb[0] / ggml_type_size(dst->type)), (uint32_t) (dst->nb[1] / ggml_type_size(dst->type)), + (uint32_t) (dst->nb[2] / ggml_type_size(dst->type)), (uint32_t) (dst->nb[3] / ggml_type_size(dst->type)), + // Logical shape β€” same for both tensors even if permuted + (uint32_t) src->ne[0], (uint32_t) src->ne[1], (uint32_t) src->ne[2], (uint32_t) src->ne[3] + }; std::vector entries = { { .binding = 0, @@ -493,9 +504,9 @@ static void ggml_webgpu_cpy(webgpu_context & ctx, ggml_tensor * src, ggml_tensor .size = ggml_webgpu_tensor_binding_size(ctx, dst) } }; - size_t max_wg_size = ctx->limits.maxComputeWorkgroupSizeX; + size_t max_wg_size = ctx->max_wg_size_x; uint32_t wg_x = (ne + max_wg_size - 1) / max_wg_size; - ggml_backend_webgpu_build_and_enqueue(ctx, ctx->cpy_pipeline, params, entries, wg_x); + ggml_backend_webgpu_build_and_enqueue(ctx, ctx->cpy_pipeline, params, entries, wg_x, ggml_op_name(dst->op)); } static void ggml_webgpu_set_rows(webgpu_context & ctx, ggml_tensor * src, ggml_tensor * idx, ggml_tensor * dst) { @@ -509,27 +520,21 @@ static void ggml_webgpu_set_rows(webgpu_context & ctx, ggml_tensor * src, ggml_t error_bufs.host_buf.Unmap(); } - std::vector params = { (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, src) / ggml_type_size(src->type)), - (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, idx) / ggml_type_size(idx->type)), - (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, dst) / ggml_type_size(dst->type)), - // Convert byte-strides to element-strides - (uint32_t) (src->nb[1] / ggml_type_size(src->type)), - (uint32_t) (src->nb[2] / ggml_type_size(src->type)), - (uint32_t) (src->nb[3] / ggml_type_size(src->type)), - (uint32_t) (idx->nb[0] / ggml_type_size(idx->type)), - (uint32_t) (idx->nb[1] / ggml_type_size(idx->type)), - (uint32_t) (idx->nb[2] / ggml_type_size(idx->type)), - (uint32_t) (dst->nb[1] / ggml_type_size(dst->type)), - (uint32_t) (dst->nb[2] / ggml_type_size(dst->type)), - (uint32_t) (dst->nb[3] / ggml_type_size(dst->type)), - // Shape of src - (uint32_t) src->ne[0], - (uint32_t) src->ne[1], - (uint32_t) src->ne[2], - (uint32_t) src->ne[3], - // Shape of idx - (uint32_t) (idx->ne[1]), - (uint32_t) (idx->ne[2]) }; + std::vector params = { + (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, src) / ggml_type_size(src->type)), + (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, idx) / ggml_type_size(idx->type)), + (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, dst) / ggml_type_size(dst->type)), + // Convert byte-strides to element-strides + (uint32_t) (src->nb[1] / ggml_type_size(src->type)), (uint32_t) (src->nb[2] / ggml_type_size(src->type)), + (uint32_t) (src->nb[3] / ggml_type_size(src->type)), (uint32_t) (idx->nb[0] / ggml_type_size(idx->type)), + (uint32_t) (idx->nb[1] / ggml_type_size(idx->type)), (uint32_t) (idx->nb[2] / ggml_type_size(idx->type)), + (uint32_t) (dst->nb[1] / ggml_type_size(dst->type)), (uint32_t) (dst->nb[2] / ggml_type_size(dst->type)), + (uint32_t) (dst->nb[3] / ggml_type_size(dst->type)), + // Shape of src + (uint32_t) src->ne[0], (uint32_t) src->ne[1], (uint32_t) src->ne[2], (uint32_t) src->ne[3], + // Shape of idx + (uint32_t) (idx->ne[1]), (uint32_t) (idx->ne[2]) + }; std::vector entries = { { .binding = 0, @@ -547,13 +552,55 @@ static void ggml_webgpu_set_rows(webgpu_context & ctx, ggml_tensor * src, ggml_t { .binding = 3, .buffer = error_bufs.dev_buf, .offset = 0, .size = error_bufs.dev_buf.GetSize() } }; - size_t max_wg_size = ctx->limits.maxComputeWorkgroupSizeX; + size_t max_wg_size = ctx->max_wg_size_x; uint32_t wg_x = (src->ne[1] * src->ne[2] * src->ne[3] + max_wg_size - 1) / max_wg_size; std::lock_guard lock(ctx->mutex); ctx->staged_set_row_error_bufs.push_back(error_bufs); - ggml_backend_webgpu_build_and_enqueue(ctx, ctx->set_rows_pipeline, params, entries, wg_x); + ggml_backend_webgpu_build_and_enqueue(ctx, ctx->set_rows_pipeline, params, entries, wg_x, ggml_op_name(dst->op)); +} + +static void ggml_webgpu_get_rows(webgpu_context & ctx, ggml_tensor * src, ggml_tensor * idx, ggml_tensor * dst) { + std::vector params = { + (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, src) / ggml_type_size(src->type)), + (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, idx) / ggml_type_size(idx->type)), + (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, dst) / ggml_type_size(dst->type)), + // Convert byte-strides to element-strides + (uint32_t) (src->nb[1] / ggml_type_size(src->type)), (uint32_t) (src->nb[2] / ggml_type_size(src->type)), + (uint32_t) (src->nb[3] / ggml_type_size(src->type)), (uint32_t) (idx->nb[0] / ggml_type_size(idx->type)), + (uint32_t) (idx->nb[1] / ggml_type_size(idx->type)), (uint32_t) (idx->nb[2] / ggml_type_size(idx->type)), + (uint32_t) (dst->nb[1] / ggml_type_size(dst->type)), (uint32_t) (dst->nb[2] / ggml_type_size(dst->type)), + (uint32_t) (dst->nb[3] / ggml_type_size(dst->type)), + // Shape of dst + (uint32_t) dst->ne[0], (uint32_t) dst->ne[1], (uint32_t) dst->ne[2], (uint32_t) dst->ne[3], + // Shape of idx + (uint32_t) (idx->ne[1]), (uint32_t) (idx->ne[2]) + }; + + std::vector entries = { + { .binding = 0, + .buffer = ggml_webgpu_tensor_buf(src), + .offset = ggml_webgpu_tensor_align_offset(ctx, src), + .size = ggml_webgpu_tensor_binding_size(ctx, src) }, + { .binding = 1, + .buffer = ggml_webgpu_tensor_buf(idx), + .offset = ggml_webgpu_tensor_align_offset(ctx, idx), + .size = ggml_webgpu_tensor_binding_size(ctx, idx) }, + { .binding = 2, + .buffer = ggml_webgpu_tensor_buf(dst), + .offset = ggml_webgpu_tensor_align_offset(ctx, dst), + .size = ggml_webgpu_tensor_binding_size(ctx, dst) } + }; + + size_t max_wg_size = ctx->max_wg_size_x; + uint32_t wg_x = (dst->ne[1] * dst->ne[2] * dst->ne[3] + max_wg_size - 1) / max_wg_size; + + wgpu::ComputePipeline pipeline = ctx->get_rows_pipeline[src->type]; + if (src->type == GGML_TYPE_F32 && dst->ne[0] % 4 != 0) { + pipeline = ctx->get_rows_f32_no_vec_pipeline; + } + ggml_backend_webgpu_build_and_enqueue(ctx, pipeline, params, entries, wg_x, ggml_op_name(dst->op)); } static void ggml_webgpu_mul_mat(webgpu_context & ctx, ggml_tensor * src0, ggml_tensor * src1, ggml_tensor * dst) { @@ -593,7 +640,104 @@ static void ggml_webgpu_mul_mat(webgpu_context & ctx, ggml_tensor * src0, ggml_t uint32_t wg_x = (dst->ne[0] * dst->ne[1] * dst->ne[2] * dst->ne[3] + WEBGPU_MUL_MAT_WG_SIZE - 1) / WEBGPU_MUL_MAT_WG_SIZE; - ggml_backend_webgpu_build_and_enqueue(ctx, ctx->mul_mat_pipeline[src0->type][src1->type], params, entries, wg_x); + ggml_backend_webgpu_build_and_enqueue(ctx, ctx->mul_mat_pipeline[src0->type][src1->type], params, entries, wg_x, + ggml_op_name(dst->op)); +} + +static void ggml_webgpu_binary_op(webgpu_context & ctx, + ggml_tensor * src0, + ggml_tensor * src1, + ggml_tensor * dst, + wgpu::ComputePipeline & pipeline, + bool in_place) { + std::vector params = { + (uint32_t) ggml_nelements(dst), + (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, src0) / ggml_type_size(src0->type)), + (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, src1) / ggml_type_size(src1->type)), + (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, dst) / ggml_type_size(dst->type)), + (uint32_t) (src1->nb[0] / ggml_type_size(src1->type)), + (uint32_t) (src1->nb[1] / ggml_type_size(src1->type)), + (uint32_t) (src1->nb[2] / ggml_type_size(src1->type)), + (uint32_t) (src1->nb[3] / ggml_type_size(src1->type)), + (uint32_t) src0->ne[0], + (uint32_t) src0->ne[1], + (uint32_t) src0->ne[2], + (uint32_t) src1->ne[0], + (uint32_t) src1->ne[1], + (uint32_t) src1->ne[2], + (uint32_t) src1->ne[3], + }; + + std::vector entries = { + { .binding = 0, + .buffer = ggml_webgpu_tensor_buf(src0), + .offset = ggml_webgpu_tensor_align_offset(ctx, src0), + .size = ggml_webgpu_tensor_binding_size(ctx, src0) }, + { .binding = 1, + .buffer = ggml_webgpu_tensor_buf(src1), + .offset = ggml_webgpu_tensor_align_offset(ctx, src1), + .size = ggml_webgpu_tensor_binding_size(ctx, src1) } + }; + if (!in_place) { + entries.push_back({ .binding = 2, + .buffer = ggml_webgpu_tensor_buf(dst), + .offset = ggml_webgpu_tensor_align_offset(ctx, dst), + .size = ggml_webgpu_tensor_binding_size(ctx, dst) }); + } + + size_t max_wg_size = ctx->max_wg_size_x; + uint32_t wg_x = (ggml_nelements(dst) + max_wg_size - 1) / max_wg_size; + ggml_backend_webgpu_build_and_enqueue(ctx, pipeline, params, entries, wg_x, ggml_op_name(dst->op)); +} + +static void ggml_webgpu_rms_norm(webgpu_context & ctx, ggml_tensor * src, ggml_tensor * dst) { + bool in_place = ggml_webgpu_tensor_equal(src, dst); + + uint32_t eps; + memcpy(&eps, dst->op_params, sizeof(float)); + + std::vector params = { + (uint32_t) (ggml_webgpu_tensor_misalignment(ctx, src) / ggml_type_size(src->type)), + }; + if (!in_place) { + params.push_back((uint32_t) (ggml_webgpu_tensor_misalignment(ctx, dst) / ggml_type_size(dst->type))); + } + params.push_back((uint32_t) (src->nb[1] / ggml_type_size(src->type))); + params.push_back((uint32_t) (src->nb[2] / ggml_type_size(src->type))); + params.push_back((uint32_t) (src->nb[3] / ggml_type_size(src->type))); + if (!in_place) { + params.push_back((uint32_t) (dst->nb[1] / ggml_type_size(dst->type))); + params.push_back((uint32_t) (dst->nb[2] / ggml_type_size(dst->type))); + params.push_back((uint32_t) (dst->nb[3] / ggml_type_size(dst->type))); + } + params.push_back((uint32_t) src->ne[0]); + params.push_back((uint32_t) src->ne[1]); + params.push_back((uint32_t) src->ne[2]); + params.push_back((uint32_t) src->ne[3]); + params.push_back(eps); // epsilon, will be bitcast to float in shader + + std::vector entries = { + { .binding = 0, + .buffer = ggml_webgpu_tensor_buf(src), + .offset = ggml_webgpu_tensor_align_offset(ctx, src), + .size = ggml_webgpu_tensor_binding_size(ctx, src) } + }; + if (!in_place) { + entries.push_back({ .binding = 1, + .buffer = ggml_webgpu_tensor_buf(dst), + .offset = ggml_webgpu_tensor_align_offset(ctx, dst), + .size = ggml_webgpu_tensor_binding_size(ctx, dst) }); + } + + wgpu::ComputePipeline pipeline; + if (in_place) { + pipeline = ctx->rms_norm_ip_pipeline; + } else { + pipeline = ctx->rms_norm_pipeline; + } + size_t max_wg_size = ctx->max_wg_size_x; + uint32_t wg_x = (src->ne[1] * src->ne[2] * src->ne[3] + max_wg_size - 1) / max_wg_size; + ggml_backend_webgpu_build_and_enqueue(ctx, pipeline, params, entries, wg_x, ggml_op_name(dst->op)); } // Returns true if node has enqueued work into the queue, false otherwise @@ -615,20 +759,34 @@ static bool ggml_webgpu_encode_node(webgpu_context ctx, ggml_tensor * node) { case GGML_OP_RESHAPE: return false; case GGML_OP_CPY: - { - ggml_webgpu_cpy(ctx, src0, node); - break; - } + ggml_webgpu_cpy(ctx, src0, node); + break; case GGML_OP_SET_ROWS: - { - ggml_webgpu_set_rows(ctx, src0, src1, node); - break; - } + ggml_webgpu_set_rows(ctx, src0, src1, node); + break; + case GGML_OP_GET_ROWS: + ggml_webgpu_get_rows(ctx, src0, src1, node); + break; case GGML_OP_MUL_MAT: - { - ggml_webgpu_mul_mat(ctx, src0, src1, node); - break; + ggml_webgpu_mul_mat(ctx, src0, src1, node); + break; + case GGML_OP_ADD: + if (ggml_webgpu_tensor_equal(src0, node)) { + ggml_webgpu_binary_op(ctx, src0, src1, node, ctx->add_ip_pipeline[node->type], true); + } else { + ggml_webgpu_binary_op(ctx, src0, src1, node, ctx->add_pipeline[node->type], false); } + break; + case GGML_OP_MUL: + if (ggml_webgpu_tensor_equal(src0, node)) { + ggml_webgpu_binary_op(ctx, src0, src1, node, ctx->mul_ip_pipeline[node->type], true); + } else { + ggml_webgpu_binary_op(ctx, src0, src1, node, ctx->mul_pipeline[node->type], false); + } + break; + case GGML_OP_RMS_NORM: + ggml_webgpu_rms_norm(ctx, src0, node); + break; default: return false; } @@ -731,8 +889,8 @@ static void ggml_backend_webgpu_buffer_set_tensor(ggml_backend_buffer_t buffer, ((uint8_t *) &val32)[i] = ((const uint8_t *) data)[size - remaining_size + i]; } // memset the remaining bytes - ggml_backend_webgpu_buffer_memset( - webgpu_ctx, buf_ctx->buffer, val32, total_offset + (size - remaining_size), remaining_size); + ggml_backend_webgpu_buffer_memset(webgpu_ctx, buf_ctx->buffer, val32, total_offset + (size - remaining_size), + remaining_size); } else { // wait for WriteBuffer to complete ggml_backend_webgpu_wait_on_submission(webgpu_ctx); @@ -766,11 +924,8 @@ static void ggml_backend_webgpu_buffer_get_tensor(ggml_backend_buffer_t buffer, if (webgpu_ctx->get_tensor_staging_buf) { webgpu_ctx->get_tensor_staging_buf.Destroy(); } - ggml_webgpu_create_buffer(device, - webgpu_ctx->get_tensor_staging_buf, - final_size, - wgpu::BufferUsage::CopyDst | wgpu::BufferUsage::MapRead, - "get_tensor_staging_buf"); + ggml_webgpu_create_buffer(device, webgpu_ctx->get_tensor_staging_buf, final_size, + wgpu::BufferUsage::CopyDst | wgpu::BufferUsage::MapRead, "get_tensor_staging_buf"); } // Copy the data from the buffer to the staging buffer @@ -824,8 +979,7 @@ static ggml_backend_buffer_t ggml_backend_webgpu_buffer_type_alloc_buffer(ggml_b ggml_backend_webgpu_device_context * ctx = static_cast(buft->device->context); wgpu::Buffer buf; - ggml_webgpu_create_buffer(ctx->webgpu_ctx->device, - buf, + ggml_webgpu_create_buffer(ctx->webgpu_ctx->device, buf, (size + WEBGPU_STORAGE_BUF_BINDING_MULT - 1) & ~(WEBGPU_STORAGE_BUF_BINDING_MULT - 1), wgpu::BufferUsage::Storage | wgpu::BufferUsage::CopySrc | wgpu::BufferUsage::CopyDst, "allocated_buffer"); @@ -890,9 +1044,17 @@ static ggml_guid_t ggml_backend_webgpu_guid(void) { return reinterpret_cast((void *) guid_str); } +// The max workgroup size is a common constant +static std::vector ggml_webgpu_max_wg_size_entry(webgpu_context & webgpu_ctx) { + std::vector constants(1); + constants[0].key = "wg_size"; + constants[0].value = webgpu_ctx->max_wg_size_x; + return constants; +} + static void ggml_webgpu_init_memset_pipeline(webgpu_context & webgpu_ctx) { // we use the maximum workgroup size for the memset pipeline - size_t max_wg_size = webgpu_ctx->limits.maxComputeWorkgroupSizeX; + size_t max_wg_size = webgpu_ctx->max_wg_size_x; size_t max_threads = max_wg_size * webgpu_ctx->limits.maxComputeWorkgroupsPerDimension; // Size the bytes_per_thread so that the largest buffer size can be handled webgpu_ctx->memset_bytes_per_thread = @@ -906,109 +1068,142 @@ static void ggml_webgpu_init_memset_pipeline(webgpu_context & webgpu_ctx) { } static void ggml_webgpu_init_mul_mat_pipeline(webgpu_context & webgpu_ctx) { - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_F32][GGML_TYPE_F32], - wgsl_mul_mat_f32_f32, - "mul_mat_f32_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_F16][GGML_TYPE_F16], - wgsl_mul_mat_f16_f16, - "mul_mat_f16_f16"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_F16][GGML_TYPE_F32], - wgsl_mul_mat_f16_f32, - "mul_mat_f16_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q4_0][GGML_TYPE_F32], - wgsl_mul_mat_q4_0_f32, - "mul_mat_q4_0_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q4_1][GGML_TYPE_F32], - wgsl_mul_mat_q4_1_f32, - "mul_mat_q4_1_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q5_0][GGML_TYPE_F32], - wgsl_mul_mat_q5_0_f32, - "mul_mat_q5_0_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q5_1][GGML_TYPE_F32], - wgsl_mul_mat_q5_1_f32, - "mul_mat_q5_1_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q8_0][GGML_TYPE_F32], - wgsl_mul_mat_q8_0_f32, - "mul_mat_q8_0_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q2_K][GGML_TYPE_F32], - wgsl_mul_mat_q2_k_f32, - "mul_mat_q2_k_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q3_K][GGML_TYPE_F32], - wgsl_mul_mat_q3_k_f32, - "mul_mat_q3_k_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q4_K][GGML_TYPE_F32], - wgsl_mul_mat_q4_k_f32, - "mul_mat_q4_k_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q5_K][GGML_TYPE_F32], - wgsl_mul_mat_q5_k_f32, - "mul_mat_q5_k_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q6_K][GGML_TYPE_F32], - wgsl_mul_mat_q6_k_f32, - "mul_mat_q6_k_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ2_XXS][GGML_TYPE_F32], - wgsl_mul_mat_iq2_xxs_f32, - "mul_mat_iq2_xxs_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ2_XS][GGML_TYPE_F32], - wgsl_mul_mat_iq2_xs_f32, - "mul_mat_iq2_xs_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ2_S][GGML_TYPE_F32], - wgsl_mul_mat_iq2_s_f32, - "mul_mat_iq2_s_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ3_XXS][GGML_TYPE_F32], - wgsl_mul_mat_iq3_xxs_f32, - "mul_mat_iq3_xxs_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ3_S][GGML_TYPE_F32], - wgsl_mul_mat_iq3_s_f32, - "mul_mat_iq3_s_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ1_S][GGML_TYPE_F32], - wgsl_mul_mat_iq1_s_f32, - "mul_mat_iq1_s_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ1_M][GGML_TYPE_F32], - wgsl_mul_mat_iq1_m_f32, - "mul_mat_iq1_m_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ4_NL][GGML_TYPE_F32], - wgsl_mul_mat_iq4_nl_f32, - "mul_mat_iq4_nl_f32"); - ggml_webgpu_create_pipeline(webgpu_ctx->device, - webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ4_XS][GGML_TYPE_F32], - wgsl_mul_mat_iq4_xs_f32, - "mul_mat_iq4_xs_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_F32][GGML_TYPE_F32], + wgsl_mul_mat_f32_f32, "mul_mat_f32_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_F16][GGML_TYPE_F16], + wgsl_mul_mat_f16_f16, "mul_mat_f16_f16"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_F16][GGML_TYPE_F32], + wgsl_mul_mat_f16_f32, "mul_mat_f16_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q4_0][GGML_TYPE_F32], + wgsl_mul_mat_q4_0_f32, "mul_mat_q4_0_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q4_1][GGML_TYPE_F32], + wgsl_mul_mat_q4_1_f32, "mul_mat_q4_1_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q5_0][GGML_TYPE_F32], + wgsl_mul_mat_q5_0_f32, "mul_mat_q5_0_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q5_1][GGML_TYPE_F32], + wgsl_mul_mat_q5_1_f32, "mul_mat_q5_1_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q8_0][GGML_TYPE_F32], + wgsl_mul_mat_q8_0_f32, "mul_mat_q8_0_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q2_K][GGML_TYPE_F32], + wgsl_mul_mat_q2_k_f32, "mul_mat_q2_k_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q3_K][GGML_TYPE_F32], + wgsl_mul_mat_q3_k_f32, "mul_mat_q3_k_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q4_K][GGML_TYPE_F32], + wgsl_mul_mat_q4_k_f32, "mul_mat_q4_k_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q5_K][GGML_TYPE_F32], + wgsl_mul_mat_q5_k_f32, "mul_mat_q5_k_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_Q6_K][GGML_TYPE_F32], + wgsl_mul_mat_q6_k_f32, "mul_mat_q6_k_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ2_XXS][GGML_TYPE_F32], + wgsl_mul_mat_iq2_xxs_f32, "mul_mat_iq2_xxs_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ2_XS][GGML_TYPE_F32], + wgsl_mul_mat_iq2_xs_f32, "mul_mat_iq2_xs_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ2_S][GGML_TYPE_F32], + wgsl_mul_mat_iq2_s_f32, "mul_mat_iq2_s_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ3_XXS][GGML_TYPE_F32], + wgsl_mul_mat_iq3_xxs_f32, "mul_mat_iq3_xxs_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ3_S][GGML_TYPE_F32], + wgsl_mul_mat_iq3_s_f32, "mul_mat_iq3_s_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ1_S][GGML_TYPE_F32], + wgsl_mul_mat_iq1_s_f32, "mul_mat_iq1_s_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ1_M][GGML_TYPE_F32], + wgsl_mul_mat_iq1_m_f32, "mul_mat_iq1_m_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ4_NL][GGML_TYPE_F32], + wgsl_mul_mat_iq4_nl_f32, "mul_mat_iq4_nl_f32"); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_mat_pipeline[GGML_TYPE_IQ4_XS][GGML_TYPE_F32], + wgsl_mul_mat_iq4_xs_f32, "mul_mat_iq4_xs_f32"); } static void ggml_webgpu_init_set_rows_pipeline(webgpu_context & webgpu_ctx) { - std::vector constants(1); - constants[0].key = "wg_size"; - constants[0].value = webgpu_ctx->limits.maxComputeWorkgroupSizeX; - ggml_webgpu_create_pipeline( - webgpu_ctx->device, webgpu_ctx->set_rows_pipeline, wgsl_set_rows, "set_rows", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->set_rows_pipeline, wgsl_set_rows, "set_rows", + ggml_webgpu_max_wg_size_entry(webgpu_ctx)); +} + +static void ggml_webgpu_init_get_rows_pipeline(webgpu_context & webgpu_ctx) { + std::vector constants = ggml_webgpu_max_wg_size_entry(webgpu_ctx); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_F32], wgsl_get_rows_f32_vec, + "get_rows_f32_vec", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_f32_no_vec_pipeline, wgsl_get_rows_f32, + "get_rows_f32", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_F16], wgsl_get_rows_f16, + "get_rows_f16", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_I32], wgsl_get_rows_i32, + "get_rows_i32", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_Q4_0], wgsl_get_rows_q4_0, + "get_rows_q4_0", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_Q4_1], wgsl_get_rows_q4_1, + "get_rows_q4_1", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_Q5_0], wgsl_get_rows_q5_0, + "get_rows_q5_0", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_Q5_1], wgsl_get_rows_q5_1, + "get_rows_q5_1", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_Q8_0], wgsl_get_rows_q8_0, + "get_rows_q8_0", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_Q2_K], wgsl_get_rows_q2_k, + "get_rows_q2_k", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_Q3_K], wgsl_get_rows_q3_k, + "get_rows_q3_k", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_Q4_K], wgsl_get_rows_q4_k, + "get_rows_q4_k", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_Q5_K], wgsl_get_rows_q5_k, + "get_rows_q5_k", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_Q6_K], wgsl_get_rows_q6_k, + "get_rows_q6_k", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_IQ2_XXS], + wgsl_get_rows_iq2_xxs, "get_rows_iq2_xxs", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_IQ2_XS], + wgsl_get_rows_iq2_xs, "get_rows_iq2_xs", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_IQ2_S], wgsl_get_rows_iq2_s, + "get_rows_iq2_s", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_IQ3_XXS], + wgsl_get_rows_iq3_xxs, "get_rows_iq3_xxs", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_IQ3_S], wgsl_get_rows_iq3_s, + "get_rows_iq3_s", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_IQ1_S], wgsl_get_rows_iq1_s, + "get_rows_iq1_s", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_IQ1_M], wgsl_get_rows_iq1_m, + "get_rows_iq1_m", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_IQ4_NL], + wgsl_get_rows_iq4_nl, "get_rows_iq4_nl", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->get_rows_pipeline[GGML_TYPE_IQ4_XS], + wgsl_get_rows_iq4_xs, "get_rows_iq4_xs", constants); } static void ggml_webgpu_init_cpy_pipeline(webgpu_context & webgpu_ctx) { - std::vector constants(1); - constants[0].key = "wg_size"; - constants[0].value = webgpu_ctx->limits.maxComputeWorkgroupSizeX; - ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->cpy_pipeline, wgsl_cpy, "cpy", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->cpy_pipeline, wgsl_cpy, "cpy", + ggml_webgpu_max_wg_size_entry(webgpu_ctx)); +} + +static void ggml_webgpu_init_add_pipeline(webgpu_context & webgpu_ctx) { + std::vector constants = ggml_webgpu_max_wg_size_entry(webgpu_ctx); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->add_pipeline[GGML_TYPE_F32], wgsl_add_f32, "add_f32", + constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->add_pipeline[GGML_TYPE_F16], wgsl_add_f16, "add_f16", + constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->add_ip_pipeline[GGML_TYPE_F32], wgsl_add_in_place_f32, + "add_in_place_f32", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->add_ip_pipeline[GGML_TYPE_F16], wgsl_add_in_place_f16, + "add_in_place_f16", constants); +} + +static void ggml_webgpu_init_mul_pipeline(webgpu_context & webgpu_ctx) { + std::vector constants = ggml_webgpu_max_wg_size_entry(webgpu_ctx); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_pipeline[GGML_TYPE_F32], wgsl_mul_f32, "mul_f32", + constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_pipeline[GGML_TYPE_F16], wgsl_mul_f16, "mul_f16", + constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_ip_pipeline[GGML_TYPE_F32], wgsl_mul_in_place_f32, + "mul_in_place_f32", constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->mul_ip_pipeline[GGML_TYPE_F16], wgsl_mul_in_place_f16, + "mul_in_place_f16", constants); +} + +static void ggml_webgpu_init_rms_norm_pipeline(webgpu_context & webgpu_ctx) { + std::vector constants = ggml_webgpu_max_wg_size_entry(webgpu_ctx); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->rms_norm_pipeline, wgsl_rms_norm, "rms_norm", + constants); + ggml_webgpu_create_pipeline(webgpu_ctx->device, webgpu_ctx->rms_norm_ip_pipeline, wgsl_rms_norm_in_place, + "rms_norm_in_place", constants); } static ggml_backend_t ggml_backend_webgpu_device_init(ggml_backend_dev_t dev, const char * params) { @@ -1058,24 +1253,77 @@ static bool ggml_backend_webgpu_device_supports_buft(ggml_backend_dev_t dev, ggm return buft->iface.get_name == ggml_backend_webgpu_buffer_type_get_name; } -static bool ggml_backend_webgpu_device_supports_op(ggml_backend_dev_t dev, const ggml_tensor * op) { - GGML_UNUSED(dev); +static bool ggml_webgpu_supported_qtype(ggml_type type) { + switch (type) { + case GGML_TYPE_Q4_0: + case GGML_TYPE_Q4_1: + case GGML_TYPE_Q5_0: + case GGML_TYPE_Q5_1: + case GGML_TYPE_Q8_0: + case GGML_TYPE_Q2_K: + case GGML_TYPE_Q3_K: + case GGML_TYPE_Q4_K: + case GGML_TYPE_Q5_K: + case GGML_TYPE_Q6_K: + case GGML_TYPE_IQ2_XXS: + case GGML_TYPE_IQ2_XS: + case GGML_TYPE_IQ2_S: + case GGML_TYPE_IQ3_XXS: + case GGML_TYPE_IQ3_S: + case GGML_TYPE_IQ1_S: + case GGML_TYPE_IQ1_M: + case GGML_TYPE_IQ4_NL: + case GGML_TYPE_IQ4_XS: + return true; + default: + return false; + } +} +static bool ggml_backend_webgpu_device_supports_op(ggml_backend_dev_t dev, const ggml_tensor * op) { + ggml_backend_webgpu_device_context * ctx = static_cast(dev->context); + + webgpu_context webgpu_ctx = ctx->webgpu_ctx; + + ggml_tensor * src0 = op->src[0]; + ggml_tensor * src1 = op->src[1]; + // on smaller devices (or CI), tensors may be larger than the max storage buffer size + if (ggml_nbytes(op) > webgpu_ctx->limits.maxStorageBufferBindingSize || + (src0 != nullptr && ggml_nbytes(src0) > webgpu_ctx->limits.maxStorageBufferBindingSize) || + (src1 != nullptr && ggml_nbytes(src1) > webgpu_ctx->limits.maxStorageBufferBindingSize)) { + return false; + } + + bool supports_op = false; switch (op->op) { case GGML_OP_NONE: case GGML_OP_VIEW: case GGML_OP_PERMUTE: case GGML_OP_TRANSPOSE: case GGML_OP_RESHAPE: - return true; + supports_op = true; + break; + case GGML_OP_ADD: + case GGML_OP_MUL: + supports_op = (op->type == GGML_TYPE_F32 || op->type == GGML_TYPE_F16) && (op->src[0]->type == op->type) && + (op->src[1]->type == op->type); + break; case GGML_OP_CPY: case GGML_OP_SET_ROWS: - return op->type == GGML_TYPE_F16 && op->src[0]->type == GGML_TYPE_F32; + supports_op = (op->type == GGML_TYPE_F16 && op->src[0]->type == GGML_TYPE_F32); + break; + case GGML_OP_GET_ROWS: + if (op->src[0]->type == GGML_TYPE_F32 || op->src[0]->type == GGML_TYPE_F16 || + op->src[0]->type == GGML_TYPE_I32 || ggml_webgpu_supported_qtype(op->src[0]->type)) { + supports_op = (op->type == GGML_TYPE_F32); + } + break; case GGML_OP_MUL_MAT: { switch (op->src[1]->type) { case GGML_TYPE_F16: - return op->src[0]->type == GGML_TYPE_F16; + supports_op = (op->src[0]->type == GGML_TYPE_F16); + break; case GGML_TYPE_F32: switch (op->src[0]->type) { case GGML_TYPE_F32: @@ -1099,17 +1347,30 @@ static bool ggml_backend_webgpu_device_supports_op(ggml_backend_dev_t dev, const case GGML_TYPE_IQ1_M: case GGML_TYPE_IQ4_NL: case GGML_TYPE_IQ4_XS: - return true; + supports_op = true; + break; default: - return false; + break; } default: - return false; + break; } + break; } + case GGML_OP_RMS_NORM: + supports_op = op->type == GGML_TYPE_F32 && op->src[0]->type == GGML_TYPE_F32; + break; default: - return false; + break; } +#ifdef GGML_WEBGPU_DEBUG + if (!supports_op) { + WEBGPU_LOG_DEBUG("not supported: " << ggml_op_name(op->op) << " with types dst: " << ggml_type_name(op->type) + << ", src0: " << (op->src[0] ? ggml_type_name(op->src[0]->type) : "null") + << ", src1: " << (op->src[1] ? ggml_type_name(op->src[1]->type) : "null")); + } +#endif + return supports_op; } static struct ggml_backend_device_i ggml_backend_webgpu_device_i = { @@ -1155,18 +1416,20 @@ static ggml_backend_dev_t ggml_backend_webgpu_reg_get_device(ggml_backend_reg_t webgpu_context ctx = reg_ctx->webgpu_ctx; wgpu::RequestAdapterOptions options = {}; - ctx->instance.WaitAny( - ctx->instance.RequestAdapter(&options, wgpu::CallbackMode::AllowSpontaneous, - [&ctx](wgpu::RequestAdapterStatus status, wgpu::Adapter adapter, const char * message) { - if (status != wgpu::RequestAdapterStatus::Success) { - GGML_LOG_ERROR("ggml_webgpu: Failed to get an adapter: %s\n", message); - return; - } - ctx->adapter = std::move(adapter); - }), UINT64_MAX); + ctx->instance.WaitAny(ctx->instance.RequestAdapter( + &options, wgpu::CallbackMode::AllowSpontaneous, + [&ctx](wgpu::RequestAdapterStatus status, wgpu::Adapter adapter, const char * message) { + if (status != wgpu::RequestAdapterStatus::Success) { + GGML_LOG_ERROR("ggml_webgpu: Failed to get an adapter: %s\n", message); + return; + } + ctx->adapter = std::move(adapter); + }), + UINT64_MAX); GGML_ASSERT(ctx->adapter != nullptr); ctx->adapter.GetLimits(&ctx->limits); + ctx->max_wg_size_x = 288; // default value wgpu::AdapterInfo info{}; ctx->adapter.GetInfo(&info); @@ -1182,21 +1445,21 @@ static ggml_backend_dev_t ggml_backend_webgpu_reg_get_device(ggml_backend_reg_t wgpu::CallbackMode::AllowSpontaneous, [](const wgpu::Device & device, wgpu::DeviceLostReason reason, wgpu::StringView message) { GGML_UNUSED(device); - GGML_LOG_ERROR( - "ggml_webgpu: Device lost! Reason: %d, Message: %s\n", static_cast(reason), std::string(message).c_str()); + GGML_LOG_ERROR("ggml_webgpu: Device lost! Reason: %d, Message: %s\n", static_cast(reason), + std::string(message).c_str()); }); dev_desc.SetUncapturedErrorCallback( [](const wgpu::Device & device, wgpu::ErrorType reason, wgpu::StringView message) { GGML_UNUSED(device); - GGML_LOG_ERROR( - "ggml_webgpu: Device error! Reason: %d, Message: %s\n", static_cast(reason), std::string(message).c_str()); + GGML_LOG_ERROR("ggml_webgpu: Device error! Reason: %d, Message: %s\n", static_cast(reason), + std::string(message).c_str()); }); ctx->instance.WaitAny(ctx->adapter.RequestDevice( - &dev_desc, - wgpu::CallbackMode::AllowSpontaneous, + &dev_desc, wgpu::CallbackMode::AllowSpontaneous, [ctx](wgpu::RequestDeviceStatus status, wgpu::Device device, wgpu::StringView message) { if (status != wgpu::RequestDeviceStatus::Success) { - GGML_LOG_ERROR("ggml_webgpu: Failed to get a device: %s\n", std::string(message).c_str()); + GGML_LOG_ERROR("ggml_webgpu: Failed to get a device: %s\n", + std::string(message).c_str()); return; } ctx->device = std::move(device); @@ -1208,34 +1471,28 @@ static ggml_backend_dev_t ggml_backend_webgpu_reg_get_device(ggml_backend_reg_t ctx->queue = ctx->device.GetQueue(); // Create buffer pool for shader parameters - ctx->param_buf_pool.init(ctx->device, - WEBGPU_NUM_PARAM_BUFS, - WEBGPU_PARAMS_BUF_SIZE_BYTES, + ctx->param_buf_pool.init(ctx->device, WEBGPU_NUM_PARAM_BUFS, WEBGPU_PARAMS_BUF_SIZE_BYTES, wgpu::BufferUsage::CopyDst | wgpu::BufferUsage::Uniform, wgpu::BufferUsage::CopySrc | wgpu::BufferUsage::MapWrite); - ctx->set_rows_error_buf_pool.init(ctx->device, - WEBGPU_NUM_SET_ROWS_ERROR_BUFS, - WEBGPU_SET_ROWS_ERROR_BUF_SIZE_BYTES, + ctx->set_rows_error_buf_pool.init(ctx->device, WEBGPU_NUM_SET_ROWS_ERROR_BUFS, WEBGPU_SET_ROWS_ERROR_BUF_SIZE_BYTES, wgpu::BufferUsage::CopySrc | wgpu::BufferUsage::Storage, wgpu::BufferUsage::CopyDst | wgpu::BufferUsage::MapRead); ggml_webgpu_init_memset_pipeline(ctx); ggml_webgpu_init_mul_mat_pipeline(ctx); ggml_webgpu_init_set_rows_pipeline(ctx); + ggml_webgpu_init_get_rows_pipeline(ctx); ggml_webgpu_init_cpy_pipeline(ctx); + ggml_webgpu_init_add_pipeline(ctx); + ggml_webgpu_init_mul_pipeline(ctx); + ggml_webgpu_init_rms_norm_pipeline(ctx); #ifdef GGML_WEBGPU_DEBUG // Initialize debug buffers - ggml_webgpu_create_buffer(ctx->device, - ctx->debug_host_buf, - WEBGPU_DEBUG_BUF_ELEMS * sizeof(uint32_t), - wgpu::BufferUsage::CopyDst | wgpu::BufferUsage::MapRead, - "debug_host_buf"); - ggml_webgpu_create_buffer(ctx->device, - ctx->debug_dev_buf, - WEBGPU_DEBUG_BUF_ELEMS * sizeof(uint32_t), - wgpu::BufferUsage::Storage | wgpu::BufferUsage::CopySrc, - "debug_dev_buf"); + ggml_webgpu_create_buffer(ctx->device, ctx->debug_host_buf, WEBGPU_DEBUG_BUF_ELEMS * sizeof(uint32_t), + wgpu::BufferUsage::CopyDst | wgpu::BufferUsage::MapRead, "debug_host_buf"); + ggml_webgpu_create_buffer(ctx->device, ctx->debug_dev_buf, WEBGPU_DEBUG_BUF_ELEMS * sizeof(uint32_t), + wgpu::BufferUsage::Storage | wgpu::BufferUsage::CopySrc, "debug_dev_buf"); #endif static ggml_backend_webgpu_device_context device_ctx; @@ -1246,12 +1503,8 @@ static ggml_backend_dev_t ggml_backend_webgpu_reg_get_device(ggml_backend_reg_t GGML_LOG_INFO( "ggml_webgpu: adapter_info: vendor_id: %u | vendor: %s | architecture: %s | device_id: %u | name: %s | " "device_desc: %s\n", - info.vendorID, - std::string(info.vendor).c_str(), - std::string(info.architecture).c_str(), - info.deviceID, - std::string(info.device).c_str(), - std::string(info.description).c_str()); + info.vendorID, std::string(info.vendor).c_str(), std::string(info.architecture).c_str(), info.deviceID, + std::string(info.device).c_str(), std::string(info.description).c_str()); // See GGML Backend Device Interface section static ggml_backend_device device = { diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/add.tmpl.wgsl b/ggml/src/ggml-webgpu/wgsl-shaders/add.tmpl.wgsl new file mode 100644 index 0000000000..f261cbb553 --- /dev/null +++ b/ggml/src/ggml-webgpu/wgsl-shaders/add.tmpl.wgsl @@ -0,0 +1,44 @@ +#define(VARIANTS) + +[ + { + "REPLS": { + "TYPE" : "f32", + } + }, + { + "REPLS": { + "TYPE" : "f16", + } + } +] + +#end(VARIANTS) + +#define(SHADER) + +enable f16; + +#include "binary_head.tmpl" + +@group(0) @binding(0) +var src0: array<{{TYPE}}>; + +@group(0) @binding(1) +var src1: array<{{TYPE}}>; + +@group(0) @binding(2) +var dst: array<{{TYPE}}>; + +@group(0) @binding(3) +var params: Params; + +override wg_size: u32; +@compute @workgroup_size(wg_size) +fn main(@builtin(global_invocation_id) gid: vec3) { + if (gid.x < params.ne) { + dst[params.offset_dst + gid.x] = src0[params.offset_src0 + gid.x] + src1[params.offset_src1 + src1_index(gid.x)]; + } +} + +#end(SHADER) diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/add_in_place.tmpl.wgsl b/ggml/src/ggml-webgpu/wgsl-shaders/add_in_place.tmpl.wgsl new file mode 100644 index 0000000000..903f7bdbcc --- /dev/null +++ b/ggml/src/ggml-webgpu/wgsl-shaders/add_in_place.tmpl.wgsl @@ -0,0 +1,41 @@ +#define(VARIANTS) + +[ + { + "REPLS": { + "TYPE" : "f32", + } + }, + { + "REPLS": { + "TYPE" : "f16", + } + } +] + +#end(VARIANTS) + +#define(SHADER) + +enable f16; + +#include "binary_head.tmpl" + +@group(0) @binding(0) +var src0: array<{{TYPE}}>; + +@group(0) @binding(1) +var src1: array<{{TYPE}}>; + +@group(0) @binding(2) +var params: Params; + +override wg_size: u32; +@compute @workgroup_size(wg_size) +fn main(@builtin(global_invocation_id) gid: vec3) { + if (gid.x < params.ne) { + src0[params.offset_dst + gid.x] = src0[params.offset_src0 + gid.x] + src1[params.offset_src1 + src1_index(gid.x)]; + } +} + +#end(SHADER) diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/binary_head.tmpl b/ggml/src/ggml-webgpu/wgsl-shaders/binary_head.tmpl new file mode 100644 index 0000000000..4b254f468d --- /dev/null +++ b/ggml/src/ggml-webgpu/wgsl-shaders/binary_head.tmpl @@ -0,0 +1,45 @@ +struct Params { + ne: u32, + + // offsets in elements + offset_src0: u32, + offset_src1: u32, + offset_dst: u32, + + stride_src1_0: u32, + stride_src1_1: u32, + stride_src1_2: u32, + stride_src1_3: u32, + + a_ne0: u32, + a_ne1: u32, + a_ne2: u32, + + b_ne0: u32, + b_ne1: u32, + b_ne2: u32, + b_ne3: u32, +}; + +fn src1_index(_i: u32) -> u32 { + var i = _i; + let a_i3 = i / (params.a_ne2 * params.a_ne1 * params.a_ne0); + i = i % (params.a_ne2 * params.a_ne1 * params.a_ne0); + let a_i2 = i / (params.a_ne1 * params.a_ne0); + i = i % (params.a_ne1 * params.a_ne0); + let a_i1 = i / params.a_ne0; + let a_i0 = i % params.a_ne0; + + // handle repetition of b + // index loops back to the beginning and repeats after elements are exhausted = modulo + let b_i0 = a_i0 % params.b_ne0; + let b_i1 = a_i1 % params.b_ne1; + let b_i2 = a_i2 % params.b_ne2; + let b_i3 = a_i3 % params.b_ne3; + + // compute index for position in b's flat array + return b_i0 * params.stride_src1_0 + + b_i1 * params.stride_src1_1 + + b_i2 * params.stride_src1_2 + + b_i3 * params.stride_src1_3; +} diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/common_decls.tmpl b/ggml/src/ggml-webgpu/wgsl-shaders/common_decls.tmpl new file mode 100644 index 0000000000..389c97bb51 --- /dev/null +++ b/ggml/src/ggml-webgpu/wgsl-shaders/common_decls.tmpl @@ -0,0 +1,930 @@ +#decl(BYTE_HELPERS) + +fn get_byte(value: u32, index: u32) -> u32 { + return (value >> (index * 8)) & 0xFF; +} + +fn get_byte_i32(value: u32, index: u32) -> i32 { + return bitcast(((value >> (index * 8)) & 0xFF) << 24) >> 24; +} + +#enddecl(BYTE_HELPERS) + +#decl(Q4_0_T) +struct q4_0 { + d: f16, + qs: array +}; +#enddecl(Q4_0_T) + +#decl(Q4_1_T) +struct q4_1 { + d: f16, + m: f16, + qs: array +}; +#enddecl(Q4_1_T) + +#decl(Q5_0_T) +struct q5_0 { + d: f16, + qh: array, + qs: array +}; +#enddecl(Q5_0_T) + +#decl(Q5_1_T) +struct q5_1 { + d: f16, + m: f16, + qh: u32, + qs: array +}; +#enddecl(Q5_1_T) + +#decl(Q8_0_T) +struct q8_0 { + d: f16, + qs: array +}; +#enddecl(Q8_0_T) + +#decl(Q8_1_T) +struct q8_1 { + d: f16, + m: f16, + qs: array +}; +#enddecl(Q8_1_T) + +#decl(Q2_K_T) +struct q2_k { + scales: array, + qs: array, + d: f16, + dmin: f16 +}; +#enddecl(Q2_K_T) + +#decl(Q3_K_T) +struct q3_k { + hmask: array, + qs: array, + scales: array, + d: f16 +}; +#enddecl(Q3_K_T) + +#decl(Q45_K_SCALE_MIN) + +fn get_scale_min(is: u32, scales: array) -> vec2 { + if (is < 4) { + let sc_byte = get_byte(scales[is / 4], is % 4); + let min_byte = get_byte(scales[(is + 4) / 4], is % 4); + return vec2(f32(sc_byte & 63), f32(min_byte & 63)); + } else { + let sc_min_lo = get_byte(scales[(is + 4) / 4], (is + 4) % 4); + let sc_hi = get_byte(scales[(is - 4) / 4], (is - 4) % 4); + let min_hi = get_byte(scales[is / 4], is % 4); + let sc = (sc_min_lo & 0xF) | ((sc_hi >> 6) << 4); + let m = (sc_min_lo >> 4) | ((min_hi >> 6) << 4); + return vec2(f32(sc), f32(m)); + } +} + +#enddecl(Q45_K_SCALE_MIN) + +#decl(Q4_K_T) +struct q4_k { + d: f16, + dmin: f16, + scales: array, + qs: array +}; +#enddecl(Q4_K_T) + +#decl(Q5_K_T) +struct q5_k { + d: f16, + dmin: f16, + scales: array, + qh: array, + qs: array +}; +#enddecl(Q5_K_T) + +#decl(Q6_K_T) +struct q6_k { + ql: array, + qh: array, + scales: array, + d: f16 +}; +#enddecl(Q6_K_T) + +#decl(IQ2_XXS_T) +struct iq2_xxs { + d: f16, + qs: array +}; +#enddecl(IQ2_XXS_T) + +#decl(IQ2_XS_T) +struct iq2_xs { + d: f16, + qs: array, + scales: array +}; +#enddecl(IQ2_XS_T) + +#decl(IQ2_S_T) +struct iq2_s { + d: f16, + qs: array, + qh: array, + scales: array +}; +#enddecl(IQ2_S_T) + +#decl(IQ3_XSS_T) +struct iq3_xxs { + d: f16, + qs: array +}; +#enddecl(IQ3_XSS_T) + +#decl(IQ3_S_T) +struct iq3_s { + d: f16, + qs: array, + qh: array, + signs: array, + scales: array +}; +#enddecl(IQ3_S_T) + +#decl(IQ1_S_T) +struct iq1_s { + d: f16, + qs: array, + qh: array +}; +#enddecl(IQ1_S_T) + +#decl(IQ1_M_T) +struct iq1_m { + qs: array, + qh: array, + scales: array +}; +#enddecl(IQ1_M_T) + +#decl(IQ4_NL_T) +struct iq4_nl { + d: f16, + qs: array, +}; +#enddecl(IQ4_NL_T) + +#decl(IQ4_XS_T) +struct iq4_xs { + d: f16, + scales_h: f16, + scales_l: u32, + qs: array +}; +#enddecl(IQ4_XS_T) + +#decl(IQ23_TABLES) +const kmask_iq2xs : array = array( + 0x08040201u, // 1, 2, 4, 8 + 0x80402010u // 16, 32, 64, 128 +); + +const ksigns_iq2xs: array = array( + 0x03828100,0x87060584,0x8b0a0988,0x0f8e8d0c, + 0x93121190,0x17969514,0x1b9a9918,0x9f1e1d9c, + 0xa32221a0,0x27a6a524,0x2baaa928,0xaf2e2dac, + 0x33b2b130,0xb73635b4,0xbb3a39b8,0x3fbebd3c, + 0xc34241c0,0x47c6c544,0x4bcac948,0xcf4e4dcc, + 0x53d2d150,0xd75655d4,0xdb5a59d8,0x5fdedd5c, + 0x63e2e160,0xe76665e4,0xeb6a69e8,0x6feeed6c, + 0xf37271f0,0x77f6f574,0x7bfaf978,0xff7e7dfc +); +#enddecl(IQ23_TABLES) + +#decl(IQ2_XXS_GRID) +const iq2xxs_grid = array( + 0x08080808, 0x08080808, 0x0808082b, 0x08080808, 0x08081919, 0x08080808, 0x08082b08, 0x08080808, + 0x08082b2b, 0x08080808, 0x08190819, 0x08080808, 0x08191908, 0x08080808, 0x082b0808, 0x08080808, + 0x082b082b, 0x08080808, 0x082b2b08, 0x08080808, 0x082b2b2b, 0x08080808, 0x19080819, 0x08080808, + 0x19081908, 0x08080808, 0x19190808, 0x08080808, 0x19192b08, 0x08080808, 0x192b0819, 0x08080808, + 0x192b1908, 0x08080808, 0x2b080808, 0x08080808, 0x2b08082b, 0x08080808, 0x2b082b2b, 0x08080808, + 0x2b2b082b, 0x08080808, 0x08080819, 0x08080819, 0x08081908, 0x08080819, 0x08190808, 0x08080819, + 0x08191919, 0x08080819, 0x19080808, 0x08080819, 0x2b081908, 0x08080819, 0x2b192b08, 0x08080819, + 0x08080808, 0x0808082b, 0x0808082b, 0x0808082b, 0x082b082b, 0x0808082b, 0x2b08082b, 0x0808082b, + 0x08080819, 0x08081908, 0x08081908, 0x08081908, 0x08190808, 0x08081908, 0x082b0819, 0x08081908, + 0x082b1908, 0x08081908, 0x19080808, 0x08081908, 0x1908082b, 0x08081908, 0x19082b08, 0x08081908, + 0x192b0808, 0x08081908, 0x2b080819, 0x08081908, 0x2b081908, 0x08081908, 0x2b190808, 0x08081908, + 0x2b2b1908, 0x08081908, 0x08080808, 0x08081919, 0x0808082b, 0x08081919, 0x08082b08, 0x08081919, + 0x082b0808, 0x08081919, 0x1908192b, 0x08081919, 0x192b2b19, 0x08081919, 0x2b080808, 0x08081919, + 0x2b190819, 0x08081919, 0x08082b19, 0x0808192b, 0x08190808, 0x0808192b, 0x19080808, 0x0808192b, + 0x2b081908, 0x0808192b, 0x2b2b1908, 0x0808192b, 0x08080808, 0x08082b08, 0x08081919, 0x08082b08, + 0x08082b08, 0x08082b08, 0x08191908, 0x08082b08, 0x082b2b08, 0x08082b08, 0x19080819, 0x08082b08, + 0x19081908, 0x08082b08, 0x19190808, 0x08082b08, 0x1919082b, 0x08082b08, 0x2b082b08, 0x08082b08, + 0x08081908, 0x08082b19, 0x19080808, 0x08082b19, 0x0808082b, 0x08082b2b, 0x08191908, 0x08082b2b, + 0x08080819, 0x08190808, 0x08081908, 0x08190808, 0x08190808, 0x08190808, 0x082b0819, 0x08190808, + 0x19080808, 0x08190808, 0x192b0808, 0x08190808, 0x2b081908, 0x08190808, 0x2b190808, 0x08190808, + 0x2b191919, 0x08190808, 0x08080808, 0x08190819, 0x08082b08, 0x08190819, 0x082b0808, 0x08190819, + 0x19190808, 0x08190819, 0x19192b2b, 0x08190819, 0x2b080808, 0x08190819, 0x082b1908, 0x0819082b, + 0x19081919, 0x0819082b, 0x08080808, 0x08191908, 0x08082b08, 0x08191908, 0x082b0808, 0x08191908, + 0x082b1919, 0x08191908, 0x19082b19, 0x08191908, 0x2b080808, 0x08191908, 0x08192b08, 0x08191919, + 0x192b082b, 0x08191919, 0x08080808, 0x0819192b, 0x0819192b, 0x0819192b, 0x08080819, 0x08192b08, + 0x08081908, 0x08192b08, 0x08190808, 0x08192b08, 0x19080808, 0x08192b08, 0x2b080819, 0x08192b08, + 0x08080808, 0x08192b19, 0x08081919, 0x08192b19, 0x2b2b0808, 0x08192b19, 0x19190819, 0x08192b2b, + 0x08080808, 0x082b0808, 0x0808082b, 0x082b0808, 0x08082b2b, 0x082b0808, 0x19081908, 0x082b0808, + 0x192b0819, 0x082b0808, 0x2b080808, 0x082b0808, 0x2b08082b, 0x082b0808, 0x082b2b19, 0x082b0819, + 0x19082b08, 0x082b0819, 0x08080808, 0x082b082b, 0x0808082b, 0x082b082b, 0x08080819, 0x082b1908, + 0x08081908, 0x082b1908, 0x08190808, 0x082b1908, 0x19080808, 0x082b1908, 0x1919192b, 0x082b1908, + 0x08080808, 0x082b1919, 0x19080819, 0x082b1919, 0x192b1908, 0x082b1919, 0x2b190808, 0x082b192b, + 0x08082b08, 0x082b2b08, 0x082b0808, 0x082b2b08, 0x2b191908, 0x082b2b08, 0x19081908, 0x082b2b2b, + 0x08080819, 0x19080808, 0x08081908, 0x19080808, 0x08190808, 0x19080808, 0x08192b08, 0x19080808, + 0x082b0819, 0x19080808, 0x082b1908, 0x19080808, 0x19080808, 0x19080808, 0x19082b08, 0x19080808, + 0x1919192b, 0x19080808, 0x192b0808, 0x19080808, 0x2b080819, 0x19080808, 0x2b081908, 0x19080808, + 0x2b190808, 0x19080808, 0x08080808, 0x19080819, 0x082b0808, 0x19080819, 0x192b0819, 0x19080819, + 0x2b080808, 0x19080819, 0x2b081919, 0x19080819, 0x08080819, 0x1908082b, 0x08190808, 0x1908082b, + 0x19082b08, 0x1908082b, 0x1919192b, 0x1908082b, 0x192b2b08, 0x1908082b, 0x08080808, 0x19081908, + 0x08082b08, 0x19081908, 0x082b0808, 0x19081908, 0x2b080808, 0x19081908, 0x2b192b19, 0x19081908, + 0x0819082b, 0x19081919, 0x082b1908, 0x19081919, 0x08080808, 0x1908192b, 0x08080819, 0x19082b08, + 0x08081908, 0x19082b08, 0x08190808, 0x19082b08, 0x19080808, 0x19082b08, 0x19081919, 0x19082b08, + 0x08080808, 0x19082b19, 0x19192b08, 0x19082b19, 0x192b0819, 0x19082b19, 0x2b08082b, 0x19082b19, + 0x19081919, 0x19082b2b, 0x2b190808, 0x19082b2b, 0x08080808, 0x19190808, 0x08082b08, 0x19190808, + 0x08190819, 0x19190808, 0x08192b19, 0x19190808, 0x082b0808, 0x19190808, 0x2b080808, 0x19190808, + 0x2b082b08, 0x19190808, 0x08081908, 0x19190819, 0x1908082b, 0x19190819, 0x2b2b1908, 0x19190819, + 0x2b190819, 0x1919082b, 0x2b190808, 0x19191908, 0x2b19082b, 0x19191908, 0x08082b2b, 0x19191919, + 0x08080819, 0x1919192b, 0x19191908, 0x1919192b, 0x08080808, 0x19192b08, 0x08190819, 0x19192b08, + 0x08192b19, 0x19192b08, 0x192b1908, 0x19192b08, 0x19080808, 0x19192b19, 0x08082b08, 0x19192b2b, + 0x08081908, 0x192b0808, 0x08190808, 0x192b0808, 0x19080808, 0x192b0808, 0x192b2b08, 0x192b0808, + 0x08080808, 0x192b0819, 0x19191919, 0x192b0819, 0x08192b08, 0x192b082b, 0x192b0808, 0x192b082b, + 0x08080808, 0x192b1908, 0x08081919, 0x192b1908, 0x08190808, 0x192b1919, 0x0819082b, 0x192b1919, + 0x2b081908, 0x192b1919, 0x1908082b, 0x192b2b08, 0x08080808, 0x2b080808, 0x0808082b, 0x2b080808, + 0x08082b2b, 0x2b080808, 0x19080819, 0x2b080808, 0x2b08082b, 0x2b080808, 0x08081908, 0x2b080819, + 0x08192b08, 0x2b080819, 0x19080808, 0x2b080819, 0x08190819, 0x2b08082b, 0x08080819, 0x2b081908, + 0x08081908, 0x2b081908, 0x08190808, 0x2b081908, 0x08191919, 0x2b081908, 0x19080808, 0x2b081908, + 0x192b0808, 0x2b081908, 0x08080808, 0x2b081919, 0x1908192b, 0x2b081919, 0x2b191908, 0x2b081919, + 0x08082b19, 0x2b08192b, 0x19080808, 0x2b08192b, 0x192b0808, 0x2b08192b, 0x0808082b, 0x2b082b08, + 0x08081908, 0x2b082b19, 0x08190819, 0x2b082b2b, 0x08081908, 0x2b190808, 0x08190808, 0x2b190808, + 0x082b1908, 0x2b190808, 0x19080808, 0x2b190808, 0x2b2b0819, 0x2b190808, 0x0819192b, 0x2b190819, + 0x2b080808, 0x2b190819, 0x19081919, 0x2b19082b, 0x08080808, 0x2b191908, 0x082b082b, 0x2b191908, + 0x19081908, 0x2b191908, 0x19190819, 0x2b191919, 0x2b080819, 0x2b192b08, 0x082b0808, 0x2b192b19, + 0x0808082b, 0x2b2b0808, 0x19190808, 0x2b2b0808, 0x2b081919, 0x2b2b0808, 0x08082b19, 0x2b2b0819, + 0x08080808, 0x2b2b082b, 0x08192b08, 0x2b2b1908, 0x19190808, 0x2b2b2b08, 0x08081908, 0x2b2b2b19 +); +#enddecl(IQ2_XXS_GRID) + +#decl(IQ2_XS_GRID) +const iq2xs_grid = array( + 0x08080808, 0x08080808, 0x0808082b, 0x08080808, 0x08081919, 0x08080808, 0x08082b08, 0x08080808, + 0x08082b2b, 0x08080808, 0x08190819, 0x08080808, 0x08191908, 0x08080808, 0x0819192b, 0x08080808, + 0x08192b19, 0x08080808, 0x082b0808, 0x08080808, 0x082b082b, 0x08080808, 0x082b1919, 0x08080808, + 0x082b2b08, 0x08080808, 0x19080819, 0x08080808, 0x19081908, 0x08080808, 0x1908192b, 0x08080808, + 0x19082b19, 0x08080808, 0x19190808, 0x08080808, 0x1919082b, 0x08080808, 0x19191919, 0x08080808, + 0x19192b08, 0x08080808, 0x192b0819, 0x08080808, 0x192b1908, 0x08080808, 0x2b080808, 0x08080808, + 0x2b08082b, 0x08080808, 0x2b081919, 0x08080808, 0x2b082b08, 0x08080808, 0x2b190819, 0x08080808, + 0x2b191908, 0x08080808, 0x2b192b19, 0x08080808, 0x2b2b0808, 0x08080808, 0x08080819, 0x08080819, + 0x08081908, 0x08080819, 0x0808192b, 0x08080819, 0x08082b19, 0x08080819, 0x08190808, 0x08080819, + 0x0819082b, 0x08080819, 0x08191919, 0x08080819, 0x08192b08, 0x08080819, 0x08192b2b, 0x08080819, + 0x082b0819, 0x08080819, 0x082b1908, 0x08080819, 0x19080808, 0x08080819, 0x1908082b, 0x08080819, + 0x19081919, 0x08080819, 0x19082b08, 0x08080819, 0x19190819, 0x08080819, 0x19191908, 0x08080819, + 0x192b0808, 0x08080819, 0x192b2b08, 0x08080819, 0x2b080819, 0x08080819, 0x2b081908, 0x08080819, + 0x2b190808, 0x08080819, 0x08080808, 0x0808082b, 0x0808082b, 0x0808082b, 0x08081919, 0x0808082b, + 0x08082b08, 0x0808082b, 0x08190819, 0x0808082b, 0x08191908, 0x0808082b, 0x082b0808, 0x0808082b, + 0x19080819, 0x0808082b, 0x19081908, 0x0808082b, 0x19190808, 0x0808082b, 0x19191919, 0x0808082b, + 0x2b080808, 0x0808082b, 0x2b082b2b, 0x0808082b, 0x08080819, 0x08081908, 0x08081908, 0x08081908, + 0x0808192b, 0x08081908, 0x08082b19, 0x08081908, 0x08190808, 0x08081908, 0x0819082b, 0x08081908, + 0x08191919, 0x08081908, 0x08192b08, 0x08081908, 0x082b0819, 0x08081908, 0x082b1908, 0x08081908, + 0x19080808, 0x08081908, 0x1908082b, 0x08081908, 0x19081919, 0x08081908, 0x19082b08, 0x08081908, + 0x19190819, 0x08081908, 0x19191908, 0x08081908, 0x1919192b, 0x08081908, 0x192b0808, 0x08081908, + 0x2b080819, 0x08081908, 0x2b081908, 0x08081908, 0x2b190808, 0x08081908, 0x08080808, 0x08081919, + 0x0808082b, 0x08081919, 0x08081919, 0x08081919, 0x08082b08, 0x08081919, 0x08190819, 0x08081919, + 0x08191908, 0x08081919, 0x082b0808, 0x08081919, 0x19080819, 0x08081919, 0x19081908, 0x08081919, + 0x19190808, 0x08081919, 0x192b0819, 0x08081919, 0x2b080808, 0x08081919, 0x08080819, 0x0808192b, + 0x08081908, 0x0808192b, 0x08190808, 0x0808192b, 0x082b192b, 0x0808192b, 0x19080808, 0x0808192b, + 0x1908082b, 0x0808192b, 0x2b081908, 0x0808192b, 0x08080808, 0x08082b08, 0x0808082b, 0x08082b08, + 0x08081919, 0x08082b08, 0x08082b08, 0x08082b08, 0x08082b2b, 0x08082b08, 0x08190819, 0x08082b08, + 0x08191908, 0x08082b08, 0x082b0808, 0x08082b08, 0x082b1919, 0x08082b08, 0x19080819, 0x08082b08, + 0x19081908, 0x08082b08, 0x19190808, 0x08082b08, 0x19192b08, 0x08082b08, 0x2b080808, 0x08082b08, + 0x2b2b0808, 0x08082b08, 0x2b2b2b2b, 0x08082b08, 0x08080819, 0x08082b19, 0x08081908, 0x08082b19, + 0x08190808, 0x08082b19, 0x19080808, 0x08082b19, 0x2b080819, 0x08082b19, 0x2b082b19, 0x08082b19, + 0x08080808, 0x08082b2b, 0x082b0808, 0x08082b2b, 0x082b2b08, 0x08082b2b, 0x2b19192b, 0x08082b2b, + 0x2b2b0808, 0x08082b2b, 0x08080819, 0x08190808, 0x08081908, 0x08190808, 0x0808192b, 0x08190808, + 0x08082b19, 0x08190808, 0x08190808, 0x08190808, 0x0819082b, 0x08190808, 0x08191919, 0x08190808, + 0x08192b08, 0x08190808, 0x082b0819, 0x08190808, 0x082b1908, 0x08190808, 0x19080808, 0x08190808, + 0x1908082b, 0x08190808, 0x19081919, 0x08190808, 0x19082b08, 0x08190808, 0x19190819, 0x08190808, + 0x19191908, 0x08190808, 0x192b0808, 0x08190808, 0x192b2b2b, 0x08190808, 0x2b080819, 0x08190808, + 0x2b081908, 0x08190808, 0x2b190808, 0x08190808, 0x08080808, 0x08190819, 0x0808082b, 0x08190819, + 0x08081919, 0x08190819, 0x08082b08, 0x08190819, 0x08190819, 0x08190819, 0x08191908, 0x08190819, + 0x082b0808, 0x08190819, 0x19080819, 0x08190819, 0x19081908, 0x08190819, 0x19190808, 0x08190819, + 0x2b080808, 0x08190819, 0x2b191908, 0x08190819, 0x2b19192b, 0x08190819, 0x08080819, 0x0819082b, + 0x08081908, 0x0819082b, 0x0808192b, 0x0819082b, 0x08190808, 0x0819082b, 0x19080808, 0x0819082b, + 0x192b0808, 0x0819082b, 0x08080808, 0x08191908, 0x0808082b, 0x08191908, 0x08081919, 0x08191908, + 0x08082b08, 0x08191908, 0x08190819, 0x08191908, 0x08191908, 0x08191908, 0x082b0808, 0x08191908, + 0x19080819, 0x08191908, 0x19081908, 0x08191908, 0x19082b19, 0x08191908, 0x19190808, 0x08191908, + 0x192b1908, 0x08191908, 0x2b080808, 0x08191908, 0x08080819, 0x08191919, 0x08081908, 0x08191919, + 0x08190808, 0x08191919, 0x19080808, 0x08191919, 0x08080808, 0x0819192b, 0x08191908, 0x0819192b, + 0x19082b19, 0x0819192b, 0x08080819, 0x08192b08, 0x08081908, 0x08192b08, 0x08190808, 0x08192b08, + 0x0819082b, 0x08192b08, 0x19080808, 0x08192b08, 0x19191908, 0x08192b08, 0x2b08192b, 0x08192b08, + 0x08080808, 0x08192b19, 0x08081919, 0x08192b19, 0x192b192b, 0x08192b19, 0x19190819, 0x08192b2b, + 0x2b2b2b19, 0x08192b2b, 0x08080808, 0x082b0808, 0x0808082b, 0x082b0808, 0x08081919, 0x082b0808, + 0x08082b08, 0x082b0808, 0x08082b2b, 0x082b0808, 0x08190819, 0x082b0808, 0x08191908, 0x082b0808, + 0x082b0808, 0x082b0808, 0x19080819, 0x082b0808, 0x19081908, 0x082b0808, 0x19190808, 0x082b0808, + 0x2b080808, 0x082b0808, 0x2b2b0808, 0x082b0808, 0x08080819, 0x082b0819, 0x08081908, 0x082b0819, + 0x08190808, 0x082b0819, 0x19080808, 0x082b0819, 0x19082b08, 0x082b0819, 0x192b1919, 0x082b0819, + 0x08080808, 0x082b082b, 0x082b082b, 0x082b082b, 0x2b080808, 0x082b082b, 0x2b2b2b08, 0x082b082b, + 0x08080819, 0x082b1908, 0x08081908, 0x082b1908, 0x08190808, 0x082b1908, 0x082b2b19, 0x082b1908, + 0x19080808, 0x082b1908, 0x08080808, 0x082b1919, 0x19080819, 0x082b1919, 0x1919082b, 0x082b1919, + 0x2b192b19, 0x082b1919, 0x08080819, 0x082b192b, 0x08192b2b, 0x082b192b, 0x2b2b192b, 0x082b192b, + 0x08080808, 0x082b2b08, 0x08082b08, 0x082b2b08, 0x08082b2b, 0x082b2b08, 0x082b0808, 0x082b2b08, + 0x19191919, 0x082b2b08, 0x2b082b08, 0x082b2b08, 0x2b2b082b, 0x082b2b08, 0x192b2b08, 0x082b2b19, + 0x2b190808, 0x082b2b19, 0x08082b08, 0x082b2b2b, 0x082b0808, 0x082b2b2b, 0x2b08082b, 0x082b2b2b, + 0x2b082b08, 0x082b2b2b, 0x2b082b2b, 0x082b2b2b, 0x08080819, 0x19080808, 0x08081908, 0x19080808, + 0x0808192b, 0x19080808, 0x08082b19, 0x19080808, 0x08190808, 0x19080808, 0x0819082b, 0x19080808, + 0x08191919, 0x19080808, 0x08192b08, 0x19080808, 0x082b0819, 0x19080808, 0x082b1908, 0x19080808, + 0x19080808, 0x19080808, 0x1908082b, 0x19080808, 0x19081919, 0x19080808, 0x19082b08, 0x19080808, + 0x19082b2b, 0x19080808, 0x19190819, 0x19080808, 0x19191908, 0x19080808, 0x192b0808, 0x19080808, + 0x192b1919, 0x19080808, 0x2b080819, 0x19080808, 0x2b081908, 0x19080808, 0x2b190808, 0x19080808, + 0x08080808, 0x19080819, 0x0808082b, 0x19080819, 0x08081919, 0x19080819, 0x08082b08, 0x19080819, + 0x08190819, 0x19080819, 0x08191908, 0x19080819, 0x082b0808, 0x19080819, 0x19080819, 0x19080819, + 0x19081908, 0x19080819, 0x19190808, 0x19080819, 0x2b080808, 0x19080819, 0x2b081919, 0x19080819, + 0x2b2b082b, 0x19080819, 0x08080819, 0x1908082b, 0x08081908, 0x1908082b, 0x08190808, 0x1908082b, + 0x0819082b, 0x1908082b, 0x082b2b19, 0x1908082b, 0x19080808, 0x1908082b, 0x08080808, 0x19081908, + 0x0808082b, 0x19081908, 0x08081919, 0x19081908, 0x08082b08, 0x19081908, 0x08190819, 0x19081908, + 0x08191908, 0x19081908, 0x08192b19, 0x19081908, 0x082b0808, 0x19081908, 0x19080819, 0x19081908, + 0x19081908, 0x19081908, 0x19190808, 0x19081908, 0x2b080808, 0x19081908, 0x2b191908, 0x19081908, + 0x08080819, 0x19081919, 0x08081908, 0x19081919, 0x08190808, 0x19081919, 0x082b1908, 0x19081919, + 0x19080808, 0x19081919, 0x2b192b2b, 0x19081919, 0x08080808, 0x1908192b, 0x08082b2b, 0x1908192b, + 0x19081908, 0x1908192b, 0x19190808, 0x1908192b, 0x08080819, 0x19082b08, 0x08081908, 0x19082b08, + 0x08190808, 0x19082b08, 0x19080808, 0x19082b08, 0x19081919, 0x19082b08, 0x19191908, 0x19082b08, + 0x192b082b, 0x19082b08, 0x08080808, 0x19082b19, 0x08190819, 0x19082b19, 0x19081908, 0x19082b19, + 0x19190808, 0x19082b19, 0x192b2b19, 0x19082b19, 0x08081908, 0x19082b2b, 0x08080808, 0x19190808, + 0x0808082b, 0x19190808, 0x08081919, 0x19190808, 0x08082b08, 0x19190808, 0x08190819, 0x19190808, + 0x08191908, 0x19190808, 0x082b0808, 0x19190808, 0x082b2b08, 0x19190808, 0x19080819, 0x19190808, + 0x19081908, 0x19190808, 0x19190808, 0x19190808, 0x2b080808, 0x19190808, 0x08080819, 0x19190819, + 0x08081908, 0x19190819, 0x08190808, 0x19190819, 0x08191919, 0x19190819, 0x19080808, 0x19190819, + 0x1908082b, 0x19190819, 0x08080808, 0x1919082b, 0x19081908, 0x1919082b, 0x2b2b2b2b, 0x1919082b, + 0x08080819, 0x19191908, 0x08081908, 0x19191908, 0x08190808, 0x19191908, 0x082b0819, 0x19191908, + 0x19080808, 0x19191908, 0x192b0808, 0x19191908, 0x2b080819, 0x19191908, 0x2b2b0819, 0x19191908, + 0x08080808, 0x19191919, 0x08082b08, 0x19191919, 0x2b080808, 0x19191919, 0x2b082b08, 0x19191919, + 0x082b0819, 0x1919192b, 0x192b2b08, 0x1919192b, 0x2b2b0819, 0x1919192b, 0x08080808, 0x19192b08, + 0x08191908, 0x19192b08, 0x19080819, 0x19192b08, 0x19190808, 0x19192b08, 0x2b192b19, 0x19192b08, + 0x08192b2b, 0x19192b19, 0x19080808, 0x19192b19, 0x1908082b, 0x19192b19, 0x2b081919, 0x19192b2b, + 0x08080819, 0x192b0808, 0x08081908, 0x192b0808, 0x08190808, 0x192b0808, 0x19080808, 0x192b0808, + 0x19191908, 0x192b0808, 0x192b082b, 0x192b0808, 0x2b08192b, 0x192b0808, 0x2b2b2b19, 0x192b0808, + 0x08080808, 0x192b0819, 0x082b1908, 0x192b082b, 0x19082b2b, 0x192b082b, 0x2b19082b, 0x192b082b, + 0x08080808, 0x192b1908, 0x0819192b, 0x192b1908, 0x08190808, 0x192b1919, 0x19080808, 0x192b1919, + 0x19081919, 0x192b1919, 0x2b2b1908, 0x192b1919, 0x08080819, 0x192b2b08, 0x192b2b2b, 0x192b2b08, + 0x082b1919, 0x192b2b19, 0x0808192b, 0x192b2b2b, 0x19191908, 0x192b2b2b, 0x192b082b, 0x192b2b2b, + 0x08080808, 0x2b080808, 0x0808082b, 0x2b080808, 0x08081919, 0x2b080808, 0x08082b08, 0x2b080808, + 0x08190819, 0x2b080808, 0x08191908, 0x2b080808, 0x082b0808, 0x2b080808, 0x082b2b2b, 0x2b080808, + 0x19080819, 0x2b080808, 0x19081908, 0x2b080808, 0x19190808, 0x2b080808, 0x2b080808, 0x2b080808, + 0x2b08082b, 0x2b080808, 0x2b2b2b08, 0x2b080808, 0x2b2b2b2b, 0x2b080808, 0x08080819, 0x2b080819, + 0x08081908, 0x2b080819, 0x0808192b, 0x2b080819, 0x08190808, 0x2b080819, 0x19080808, 0x2b080819, + 0x19190819, 0x2b080819, 0x19192b19, 0x2b080819, 0x08080808, 0x2b08082b, 0x082b0808, 0x2b08082b, + 0x2b080808, 0x2b08082b, 0x2b08082b, 0x2b08082b, 0x2b2b0808, 0x2b08082b, 0x2b2b2b08, 0x2b08082b, + 0x08080819, 0x2b081908, 0x08081908, 0x2b081908, 0x08190808, 0x2b081908, 0x0819082b, 0x2b081908, + 0x08191919, 0x2b081908, 0x19080808, 0x2b081908, 0x192b0808, 0x2b081908, 0x2b082b19, 0x2b081908, + 0x08080808, 0x2b081919, 0x19081908, 0x2b081919, 0x2b2b1919, 0x2b081919, 0x08192b08, 0x2b08192b, + 0x192b2b2b, 0x2b08192b, 0x08080808, 0x2b082b08, 0x08082b08, 0x2b082b08, 0x082b1919, 0x2b082b08, + 0x19192b2b, 0x2b082b08, 0x2b080808, 0x2b082b08, 0x2b08082b, 0x2b082b08, 0x2b2b2b08, 0x2b082b08, + 0x0808192b, 0x2b082b19, 0x082b082b, 0x2b082b2b, 0x2b080808, 0x2b082b2b, 0x2b082b08, 0x2b082b2b, + 0x2b19192b, 0x2b082b2b, 0x2b2b2b08, 0x2b082b2b, 0x08080819, 0x2b190808, 0x08081908, 0x2b190808, + 0x08190808, 0x2b190808, 0x19080808, 0x2b190808, 0x1919192b, 0x2b190808, 0x2b081908, 0x2b190808, + 0x08080808, 0x2b190819, 0x082b082b, 0x2b190819, 0x192b1908, 0x2b190819, 0x1919192b, 0x2b19082b, + 0x2b082b19, 0x2b19082b, 0x08080808, 0x2b191908, 0x08081919, 0x2b191908, 0x19081908, 0x2b191908, + 0x19190808, 0x2b191908, 0x19192b08, 0x2b191908, 0x082b2b19, 0x2b191919, 0x2b190808, 0x2b191919, + 0x2b19082b, 0x2b191919, 0x19080819, 0x2b19192b, 0x19190819, 0x2b192b08, 0x2b2b192b, 0x2b192b08, + 0x19082b19, 0x2b192b19, 0x08191919, 0x2b192b2b, 0x192b0808, 0x2b192b2b, 0x08080808, 0x2b2b0808, + 0x0808082b, 0x2b2b0808, 0x08082b08, 0x2b2b0808, 0x08082b2b, 0x2b2b0808, 0x082b0808, 0x2b2b0808, + 0x082b2b2b, 0x2b2b0808, 0x2b2b0808, 0x2b2b0808, 0x19190819, 0x2b2b0819, 0x19192b19, 0x2b2b0819, + 0x2b2b192b, 0x2b2b0819, 0x08080808, 0x2b2b082b, 0x0808082b, 0x2b2b082b, 0x08082b08, 0x2b2b082b, + 0x082b2b2b, 0x2b2b082b, 0x2b080808, 0x2b2b082b, 0x2b2b0808, 0x2b2b082b, 0x19080808, 0x2b2b1908, + 0x2b191919, 0x2b2b1908, 0x192b1919, 0x2b2b192b, 0x2b192b08, 0x2b2b192b, 0x08082b2b, 0x2b2b2b08, + 0x082b0808, 0x2b2b2b08, 0x082b082b, 0x2b2b2b08, 0x082b2b08, 0x2b2b2b08, 0x2b2b0808, 0x2b2b2b08, + 0x2b2b2b08, 0x2b2b2b08, 0x08081908, 0x2b2b2b19, 0x2b081908, 0x2b2b2b19, 0x2b08192b, 0x2b2b2b19, + 0x082b2b08, 0x2b2b2b2b, 0x082b2b2b, 0x2b2b2b2b, 0x2b190819, 0x2b2b2b2b, 0x2b2b2b2b, 0x2b2b2b2b +); +#enddecl(IQ2_XS_GRID) + +#decl(IQ2_S_GRID) +const iq2s_grid = array( + 0x08080808, 0x08080808, 0x0808082b, 0x08080808, 0x08081919, 0x08080808, 0x08082b08, 0x08080808, + 0x08082b2b, 0x08080808, 0x08190819, 0x08080808, 0x08191908, 0x08080808, 0x0819192b, 0x08080808, + 0x08192b19, 0x08080808, 0x082b0808, 0x08080808, 0x082b082b, 0x08080808, 0x082b1919, 0x08080808, + 0x082b2b08, 0x08080808, 0x19080819, 0x08080808, 0x19081908, 0x08080808, 0x1908192b, 0x08080808, + 0x19082b19, 0x08080808, 0x19190808, 0x08080808, 0x1919082b, 0x08080808, 0x19191919, 0x08080808, + 0x19192b08, 0x08080808, 0x192b0819, 0x08080808, 0x192b1908, 0x08080808, 0x192b192b, 0x08080808, + 0x192b2b19, 0x08080808, 0x2b080808, 0x08080808, 0x2b08082b, 0x08080808, 0x2b081919, 0x08080808, + 0x2b082b08, 0x08080808, 0x2b190819, 0x08080808, 0x2b191908, 0x08080808, 0x2b2b0808, 0x08080808, + 0x2b2b1919, 0x08080808, 0x2b2b2b2b, 0x08080808, 0x08080819, 0x08080819, 0x08081908, 0x08080819, + 0x0808192b, 0x08080819, 0x08082b19, 0x08080819, 0x08190808, 0x08080819, 0x0819082b, 0x08080819, + 0x08191919, 0x08080819, 0x08192b08, 0x08080819, 0x082b0819, 0x08080819, 0x082b1908, 0x08080819, + 0x19080808, 0x08080819, 0x1908082b, 0x08080819, 0x19081919, 0x08080819, 0x19082b08, 0x08080819, + 0x19190819, 0x08080819, 0x19191908, 0x08080819, 0x1919192b, 0x08080819, 0x19192b19, 0x08080819, + 0x192b0808, 0x08080819, 0x192b1919, 0x08080819, 0x192b2b08, 0x08080819, 0x2b080819, 0x08080819, + 0x2b081908, 0x08080819, 0x2b190808, 0x08080819, 0x2b19082b, 0x08080819, 0x2b191919, 0x08080819, + 0x2b2b0819, 0x08080819, 0x2b2b1908, 0x08080819, 0x08080808, 0x0808082b, 0x0808082b, 0x0808082b, + 0x08081919, 0x0808082b, 0x08082b08, 0x0808082b, 0x08190819, 0x0808082b, 0x08191908, 0x0808082b, + 0x082b0808, 0x0808082b, 0x082b2b2b, 0x0808082b, 0x19080819, 0x0808082b, 0x19081908, 0x0808082b, + 0x1908192b, 0x0808082b, 0x19082b19, 0x0808082b, 0x19190808, 0x0808082b, 0x19191919, 0x0808082b, + 0x2b080808, 0x0808082b, 0x2b081919, 0x0808082b, 0x2b082b2b, 0x0808082b, 0x2b191908, 0x0808082b, + 0x2b2b082b, 0x0808082b, 0x08080819, 0x08081908, 0x08081908, 0x08081908, 0x0808192b, 0x08081908, + 0x08082b19, 0x08081908, 0x08190808, 0x08081908, 0x0819082b, 0x08081908, 0x08191919, 0x08081908, + 0x08192b08, 0x08081908, 0x082b0819, 0x08081908, 0x082b1908, 0x08081908, 0x082b192b, 0x08081908, + 0x082b2b19, 0x08081908, 0x19080808, 0x08081908, 0x1908082b, 0x08081908, 0x19081919, 0x08081908, + 0x19082b08, 0x08081908, 0x19082b2b, 0x08081908, 0x19190819, 0x08081908, 0x19191908, 0x08081908, + 0x1919192b, 0x08081908, 0x19192b19, 0x08081908, 0x192b0808, 0x08081908, 0x192b082b, 0x08081908, + 0x192b1919, 0x08081908, 0x2b080819, 0x08081908, 0x2b081908, 0x08081908, 0x2b08192b, 0x08081908, + 0x2b082b19, 0x08081908, 0x2b190808, 0x08081908, 0x2b191919, 0x08081908, 0x2b192b08, 0x08081908, + 0x2b2b0819, 0x08081908, 0x2b2b1908, 0x08081908, 0x08080808, 0x08081919, 0x0808082b, 0x08081919, + 0x08081919, 0x08081919, 0x08082b08, 0x08081919, 0x08082b2b, 0x08081919, 0x08190819, 0x08081919, + 0x08191908, 0x08081919, 0x0819192b, 0x08081919, 0x08192b19, 0x08081919, 0x082b0808, 0x08081919, + 0x082b1919, 0x08081919, 0x082b2b08, 0x08081919, 0x19080819, 0x08081919, 0x19081908, 0x08081919, + 0x1908192b, 0x08081919, 0x19082b19, 0x08081919, 0x19190808, 0x08081919, 0x1919082b, 0x08081919, + 0x19191919, 0x08081919, 0x19192b08, 0x08081919, 0x192b0819, 0x08081919, 0x192b1908, 0x08081919, + 0x2b080808, 0x08081919, 0x2b08082b, 0x08081919, 0x2b081919, 0x08081919, 0x2b082b08, 0x08081919, + 0x2b190819, 0x08081919, 0x2b191908, 0x08081919, 0x2b2b0808, 0x08081919, 0x08080819, 0x0808192b, + 0x08081908, 0x0808192b, 0x0808192b, 0x0808192b, 0x08082b19, 0x0808192b, 0x08190808, 0x0808192b, + 0x08191919, 0x0808192b, 0x19080808, 0x0808192b, 0x19081919, 0x0808192b, 0x19082b08, 0x0808192b, + 0x19190819, 0x0808192b, 0x19191908, 0x0808192b, 0x192b0808, 0x0808192b, 0x2b080819, 0x0808192b, + 0x2b081908, 0x0808192b, 0x2b190808, 0x0808192b, 0x08080808, 0x08082b08, 0x0808082b, 0x08082b08, + 0x08081919, 0x08082b08, 0x08082b08, 0x08082b08, 0x08190819, 0x08082b08, 0x08191908, 0x08082b08, + 0x0819192b, 0x08082b08, 0x08192b19, 0x08082b08, 0x082b0808, 0x08082b08, 0x082b1919, 0x08082b08, + 0x082b2b2b, 0x08082b08, 0x19080819, 0x08082b08, 0x19081908, 0x08082b08, 0x1908192b, 0x08082b08, + 0x19082b19, 0x08082b08, 0x19190808, 0x08082b08, 0x1919082b, 0x08082b08, 0x19191919, 0x08082b08, + 0x19192b08, 0x08082b08, 0x192b0819, 0x08082b08, 0x192b1908, 0x08082b08, 0x2b080808, 0x08082b08, + 0x2b081919, 0x08082b08, 0x2b191908, 0x08082b08, 0x2b2b2b2b, 0x08082b08, 0x08080819, 0x08082b19, + 0x08081908, 0x08082b19, 0x08190808, 0x08082b19, 0x0819082b, 0x08082b19, 0x08191919, 0x08082b19, + 0x08192b08, 0x08082b19, 0x082b0819, 0x08082b19, 0x19080808, 0x08082b19, 0x19081919, 0x08082b19, + 0x19082b08, 0x08082b19, 0x19190819, 0x08082b19, 0x19191908, 0x08082b19, 0x192b0808, 0x08082b19, + 0x2b080819, 0x08082b19, 0x2b190808, 0x08082b19, 0x08080808, 0x08082b2b, 0x08190819, 0x08082b2b, + 0x08191908, 0x08082b2b, 0x082b082b, 0x08082b2b, 0x082b2b08, 0x08082b2b, 0x082b2b2b, 0x08082b2b, + 0x19190808, 0x08082b2b, 0x2b192b19, 0x08082b2b, 0x08080819, 0x08190808, 0x08081908, 0x08190808, + 0x0808192b, 0x08190808, 0x08082b19, 0x08190808, 0x08190808, 0x08190808, 0x0819082b, 0x08190808, + 0x08191919, 0x08190808, 0x08192b08, 0x08190808, 0x082b0819, 0x08190808, 0x082b1908, 0x08190808, + 0x082b192b, 0x08190808, 0x19080808, 0x08190808, 0x1908082b, 0x08190808, 0x19081919, 0x08190808, + 0x19082b08, 0x08190808, 0x19190819, 0x08190808, 0x19191908, 0x08190808, 0x1919192b, 0x08190808, + 0x19192b19, 0x08190808, 0x192b0808, 0x08190808, 0x192b082b, 0x08190808, 0x192b1919, 0x08190808, + 0x192b2b08, 0x08190808, 0x2b080819, 0x08190808, 0x2b081908, 0x08190808, 0x2b08192b, 0x08190808, + 0x2b190808, 0x08190808, 0x2b191919, 0x08190808, 0x2b192b08, 0x08190808, 0x2b2b0819, 0x08190808, + 0x2b2b1908, 0x08190808, 0x08080808, 0x08190819, 0x0808082b, 0x08190819, 0x08081919, 0x08190819, + 0x08082b08, 0x08190819, 0x08082b2b, 0x08190819, 0x08190819, 0x08190819, 0x08191908, 0x08190819, + 0x0819192b, 0x08190819, 0x08192b19, 0x08190819, 0x082b0808, 0x08190819, 0x082b082b, 0x08190819, + 0x082b1919, 0x08190819, 0x082b2b08, 0x08190819, 0x19080819, 0x08190819, 0x19081908, 0x08190819, + 0x1908192b, 0x08190819, 0x19082b19, 0x08190819, 0x19190808, 0x08190819, 0x1919082b, 0x08190819, + 0x19191919, 0x08190819, 0x19192b08, 0x08190819, 0x192b0819, 0x08190819, 0x192b1908, 0x08190819, + 0x2b080808, 0x08190819, 0x2b08082b, 0x08190819, 0x2b081919, 0x08190819, 0x2b082b08, 0x08190819, + 0x2b190819, 0x08190819, 0x2b191908, 0x08190819, 0x08080819, 0x0819082b, 0x08081908, 0x0819082b, + 0x08082b19, 0x0819082b, 0x08190808, 0x0819082b, 0x08191919, 0x0819082b, 0x082b0819, 0x0819082b, + 0x082b1908, 0x0819082b, 0x19080808, 0x0819082b, 0x19081919, 0x0819082b, 0x19190819, 0x0819082b, + 0x19191908, 0x0819082b, 0x2b080819, 0x0819082b, 0x2b081908, 0x0819082b, 0x2b190808, 0x0819082b, + 0x08080808, 0x08191908, 0x0808082b, 0x08191908, 0x08081919, 0x08191908, 0x08082b08, 0x08191908, + 0x08190819, 0x08191908, 0x08191908, 0x08191908, 0x0819192b, 0x08191908, 0x08192b19, 0x08191908, + 0x082b0808, 0x08191908, 0x082b1919, 0x08191908, 0x082b2b08, 0x08191908, 0x19080819, 0x08191908, + 0x19081908, 0x08191908, 0x1908192b, 0x08191908, 0x19082b19, 0x08191908, 0x19190808, 0x08191908, + 0x1919082b, 0x08191908, 0x19191919, 0x08191908, 0x19192b08, 0x08191908, 0x192b0819, 0x08191908, + 0x192b1908, 0x08191908, 0x2b080808, 0x08191908, 0x2b08082b, 0x08191908, 0x2b081919, 0x08191908, + 0x2b082b08, 0x08191908, 0x2b190819, 0x08191908, 0x2b191908, 0x08191908, 0x2b2b0808, 0x08191908, + 0x08080819, 0x08191919, 0x08081908, 0x08191919, 0x0808192b, 0x08191919, 0x08082b19, 0x08191919, + 0x08190808, 0x08191919, 0x0819082b, 0x08191919, 0x08191919, 0x08191919, 0x08192b08, 0x08191919, + 0x082b0819, 0x08191919, 0x082b1908, 0x08191919, 0x19080808, 0x08191919, 0x1908082b, 0x08191919, + 0x19081919, 0x08191919, 0x19082b08, 0x08191919, 0x19190819, 0x08191919, 0x19191908, 0x08191919, + 0x192b0808, 0x08191919, 0x2b080819, 0x08191919, 0x2b081908, 0x08191919, 0x2b190808, 0x08191919, + 0x08080808, 0x0819192b, 0x08081919, 0x0819192b, 0x08082b08, 0x0819192b, 0x08190819, 0x0819192b, + 0x08191908, 0x0819192b, 0x082b0808, 0x0819192b, 0x19080819, 0x0819192b, 0x19081908, 0x0819192b, + 0x19190808, 0x0819192b, 0x2b080808, 0x0819192b, 0x2b2b2b2b, 0x0819192b, 0x08080819, 0x08192b08, + 0x08081908, 0x08192b08, 0x0808192b, 0x08192b08, 0x08082b19, 0x08192b08, 0x08190808, 0x08192b08, + 0x08191919, 0x08192b08, 0x08192b08, 0x08192b08, 0x082b0819, 0x08192b08, 0x19080808, 0x08192b08, + 0x1908082b, 0x08192b08, 0x19081919, 0x08192b08, 0x19082b08, 0x08192b08, 0x19190819, 0x08192b08, + 0x19191908, 0x08192b08, 0x192b0808, 0x08192b08, 0x2b080819, 0x08192b08, 0x2b081908, 0x08192b08, + 0x08080808, 0x08192b19, 0x0808082b, 0x08192b19, 0x08081919, 0x08192b19, 0x08082b08, 0x08192b19, + 0x08190819, 0x08192b19, 0x08191908, 0x08192b19, 0x082b0808, 0x08192b19, 0x19080819, 0x08192b19, + 0x19081908, 0x08192b19, 0x19190808, 0x08192b19, 0x192b2b19, 0x08192b19, 0x2b2b082b, 0x08192b19, + 0x08081908, 0x08192b2b, 0x08190808, 0x08192b2b, 0x19080808, 0x08192b2b, 0x1919192b, 0x08192b2b, + 0x08080808, 0x082b0808, 0x0808082b, 0x082b0808, 0x08081919, 0x082b0808, 0x08082b08, 0x082b0808, + 0x08190819, 0x082b0808, 0x08191908, 0x082b0808, 0x0819192b, 0x082b0808, 0x08192b19, 0x082b0808, + 0x082b0808, 0x082b0808, 0x082b1919, 0x082b0808, 0x082b2b2b, 0x082b0808, 0x19080819, 0x082b0808, + 0x19081908, 0x082b0808, 0x19190808, 0x082b0808, 0x1919082b, 0x082b0808, 0x19191919, 0x082b0808, + 0x192b1908, 0x082b0808, 0x2b080808, 0x082b0808, 0x2b082b2b, 0x082b0808, 0x2b191908, 0x082b0808, + 0x2b2b2b2b, 0x082b0808, 0x08080819, 0x082b0819, 0x08081908, 0x082b0819, 0x08190808, 0x082b0819, + 0x0819082b, 0x082b0819, 0x08191919, 0x082b0819, 0x082b0819, 0x082b0819, 0x19080808, 0x082b0819, + 0x1908082b, 0x082b0819, 0x19081919, 0x082b0819, 0x19190819, 0x082b0819, 0x19191908, 0x082b0819, + 0x192b0808, 0x082b0819, 0x2b080819, 0x082b0819, 0x2b081908, 0x082b0819, 0x2b190808, 0x082b0819, + 0x08080808, 0x082b082b, 0x08082b2b, 0x082b082b, 0x082b082b, 0x082b082b, 0x082b2b08, 0x082b082b, + 0x082b2b2b, 0x082b082b, 0x19081908, 0x082b082b, 0x19190808, 0x082b082b, 0x2b082b08, 0x082b082b, + 0x2b082b2b, 0x082b082b, 0x2b2b2b08, 0x082b082b, 0x08080819, 0x082b1908, 0x08081908, 0x082b1908, + 0x0808192b, 0x082b1908, 0x08082b19, 0x082b1908, 0x08190808, 0x082b1908, 0x08191919, 0x082b1908, + 0x08192b08, 0x082b1908, 0x082b0819, 0x082b1908, 0x082b1908, 0x082b1908, 0x19080808, 0x082b1908, + 0x1908082b, 0x082b1908, 0x19081919, 0x082b1908, 0x19082b08, 0x082b1908, 0x19190819, 0x082b1908, + 0x19191908, 0x082b1908, 0x192b0808, 0x082b1908, 0x2b080819, 0x082b1908, 0x2b081908, 0x082b1908, + 0x2b190808, 0x082b1908, 0x08080808, 0x082b1919, 0x08081919, 0x082b1919, 0x08082b08, 0x082b1919, + 0x08190819, 0x082b1919, 0x08191908, 0x082b1919, 0x082b0808, 0x082b1919, 0x19080819, 0x082b1919, + 0x19081908, 0x082b1919, 0x19190808, 0x082b1919, 0x192b192b, 0x082b1919, 0x2b080808, 0x082b1919, + 0x08080819, 0x082b192b, 0x08081908, 0x082b192b, 0x08190808, 0x082b192b, 0x19080808, 0x082b192b, + 0x19192b19, 0x082b192b, 0x08080808, 0x082b2b08, 0x08081919, 0x082b2b08, 0x08190819, 0x082b2b08, + 0x08191908, 0x082b2b08, 0x19080819, 0x082b2b08, 0x19081908, 0x082b2b08, 0x19190808, 0x082b2b08, + 0x2b082b2b, 0x082b2b08, 0x2b2b2b2b, 0x082b2b08, 0x08080819, 0x082b2b19, 0x08081908, 0x082b2b19, + 0x08190808, 0x082b2b19, 0x2b191919, 0x082b2b19, 0x08082b2b, 0x082b2b2b, 0x082b082b, 0x082b2b2b, + 0x192b1908, 0x082b2b2b, 0x2b082b08, 0x082b2b2b, 0x2b082b2b, 0x082b2b2b, 0x08080819, 0x19080808, + 0x08081908, 0x19080808, 0x0808192b, 0x19080808, 0x08082b19, 0x19080808, 0x08190808, 0x19080808, + 0x0819082b, 0x19080808, 0x08191919, 0x19080808, 0x08192b08, 0x19080808, 0x08192b2b, 0x19080808, + 0x082b0819, 0x19080808, 0x082b1908, 0x19080808, 0x082b192b, 0x19080808, 0x19080808, 0x19080808, + 0x1908082b, 0x19080808, 0x19081919, 0x19080808, 0x19082b08, 0x19080808, 0x19082b2b, 0x19080808, + 0x19190819, 0x19080808, 0x19191908, 0x19080808, 0x1919192b, 0x19080808, 0x19192b19, 0x19080808, + 0x192b0808, 0x19080808, 0x192b082b, 0x19080808, 0x192b1919, 0x19080808, 0x2b080819, 0x19080808, + 0x2b081908, 0x19080808, 0x2b190808, 0x19080808, 0x2b191919, 0x19080808, 0x2b192b08, 0x19080808, + 0x2b2b0819, 0x19080808, 0x2b2b1908, 0x19080808, 0x08080808, 0x19080819, 0x0808082b, 0x19080819, + 0x08081919, 0x19080819, 0x08082b08, 0x19080819, 0x08190819, 0x19080819, 0x08191908, 0x19080819, + 0x0819192b, 0x19080819, 0x08192b19, 0x19080819, 0x082b0808, 0x19080819, 0x082b082b, 0x19080819, + 0x082b1919, 0x19080819, 0x19080819, 0x19080819, 0x19081908, 0x19080819, 0x1908192b, 0x19080819, + 0x19082b19, 0x19080819, 0x19190808, 0x19080819, 0x1919082b, 0x19080819, 0x19191919, 0x19080819, + 0x19192b08, 0x19080819, 0x192b0819, 0x19080819, 0x192b1908, 0x19080819, 0x2b080808, 0x19080819, + 0x2b08082b, 0x19080819, 0x2b081919, 0x19080819, 0x2b082b08, 0x19080819, 0x2b190819, 0x19080819, + 0x2b191908, 0x19080819, 0x2b2b0808, 0x19080819, 0x08080819, 0x1908082b, 0x08081908, 0x1908082b, + 0x08190808, 0x1908082b, 0x0819082b, 0x1908082b, 0x08191919, 0x1908082b, 0x08192b08, 0x1908082b, + 0x082b1908, 0x1908082b, 0x19080808, 0x1908082b, 0x19081919, 0x1908082b, 0x19082b08, 0x1908082b, + 0x19190819, 0x1908082b, 0x19191908, 0x1908082b, 0x192b0808, 0x1908082b, 0x2b080819, 0x1908082b, + 0x2b081908, 0x1908082b, 0x08080808, 0x19081908, 0x0808082b, 0x19081908, 0x08081919, 0x19081908, + 0x08082b08, 0x19081908, 0x08082b2b, 0x19081908, 0x08190819, 0x19081908, 0x08191908, 0x19081908, + 0x0819192b, 0x19081908, 0x08192b19, 0x19081908, 0x082b0808, 0x19081908, 0x082b082b, 0x19081908, + 0x082b1919, 0x19081908, 0x082b2b08, 0x19081908, 0x19080819, 0x19081908, 0x19081908, 0x19081908, + 0x1908192b, 0x19081908, 0x19082b19, 0x19081908, 0x19190808, 0x19081908, 0x1919082b, 0x19081908, + 0x19191919, 0x19081908, 0x19192b08, 0x19081908, 0x192b0819, 0x19081908, 0x192b1908, 0x19081908, + 0x2b080808, 0x19081908, 0x2b08082b, 0x19081908, 0x2b081919, 0x19081908, 0x2b082b08, 0x19081908, + 0x2b190819, 0x19081908, 0x2b191908, 0x19081908, 0x2b2b0808, 0x19081908, 0x08080819, 0x19081919, + 0x08081908, 0x19081919, 0x0808192b, 0x19081919, 0x08082b19, 0x19081919, 0x08190808, 0x19081919, + 0x0819082b, 0x19081919, 0x08191919, 0x19081919, 0x08192b08, 0x19081919, 0x082b0819, 0x19081919, + 0x082b1908, 0x19081919, 0x19080808, 0x19081919, 0x1908082b, 0x19081919, 0x19081919, 0x19081919, + 0x19082b08, 0x19081919, 0x19190819, 0x19081919, 0x19191908, 0x19081919, 0x192b0808, 0x19081919, + 0x192b2b2b, 0x19081919, 0x2b080819, 0x19081919, 0x2b081908, 0x19081919, 0x2b190808, 0x19081919, + 0x08080808, 0x1908192b, 0x0808082b, 0x1908192b, 0x08081919, 0x1908192b, 0x08082b08, 0x1908192b, + 0x08190819, 0x1908192b, 0x08191908, 0x1908192b, 0x082b0808, 0x1908192b, 0x19080819, 0x1908192b, + 0x19081908, 0x1908192b, 0x19190808, 0x1908192b, 0x2b080808, 0x1908192b, 0x2b2b1919, 0x1908192b, + 0x08080819, 0x19082b08, 0x08081908, 0x19082b08, 0x08082b19, 0x19082b08, 0x08190808, 0x19082b08, + 0x0819082b, 0x19082b08, 0x08191919, 0x19082b08, 0x08192b08, 0x19082b08, 0x082b0819, 0x19082b08, + 0x082b1908, 0x19082b08, 0x19080808, 0x19082b08, 0x1908082b, 0x19082b08, 0x19081919, 0x19082b08, + 0x19082b08, 0x19082b08, 0x19190819, 0x19082b08, 0x19191908, 0x19082b08, 0x192b0808, 0x19082b08, + 0x2b081908, 0x19082b08, 0x2b190808, 0x19082b08, 0x08080808, 0x19082b19, 0x0808082b, 0x19082b19, + 0x08081919, 0x19082b19, 0x08082b08, 0x19082b19, 0x08190819, 0x19082b19, 0x08191908, 0x19082b19, + 0x082b0808, 0x19082b19, 0x19080819, 0x19082b19, 0x19081908, 0x19082b19, 0x19190808, 0x19082b19, + 0x2b080808, 0x19082b19, 0x2b19192b, 0x19082b19, 0x08080819, 0x19082b2b, 0x08081908, 0x19082b2b, + 0x08190808, 0x19082b2b, 0x19080808, 0x19082b2b, 0x08080808, 0x19190808, 0x0808082b, 0x19190808, + 0x08081919, 0x19190808, 0x08082b08, 0x19190808, 0x08190819, 0x19190808, 0x08191908, 0x19190808, + 0x0819192b, 0x19190808, 0x08192b19, 0x19190808, 0x082b0808, 0x19190808, 0x082b082b, 0x19190808, + 0x082b1919, 0x19190808, 0x082b2b08, 0x19190808, 0x19080819, 0x19190808, 0x19081908, 0x19190808, + 0x1908192b, 0x19190808, 0x19082b19, 0x19190808, 0x19190808, 0x19190808, 0x1919082b, 0x19190808, + 0x19191919, 0x19190808, 0x19192b08, 0x19190808, 0x192b0819, 0x19190808, 0x192b1908, 0x19190808, + 0x2b080808, 0x19190808, 0x2b08082b, 0x19190808, 0x2b081919, 0x19190808, 0x2b082b08, 0x19190808, + 0x2b190819, 0x19190808, 0x2b191908, 0x19190808, 0x08080819, 0x19190819, 0x08081908, 0x19190819, + 0x0808192b, 0x19190819, 0x08082b19, 0x19190819, 0x08190808, 0x19190819, 0x0819082b, 0x19190819, + 0x08191919, 0x19190819, 0x08192b08, 0x19190819, 0x082b0819, 0x19190819, 0x082b1908, 0x19190819, + 0x19080808, 0x19190819, 0x1908082b, 0x19190819, 0x19081919, 0x19190819, 0x19082b08, 0x19190819, + 0x19190819, 0x19190819, 0x19191908, 0x19190819, 0x192b0808, 0x19190819, 0x2b080819, 0x19190819, + 0x2b081908, 0x19190819, 0x2b190808, 0x19190819, 0x08080808, 0x1919082b, 0x08081919, 0x1919082b, + 0x08082b08, 0x1919082b, 0x08190819, 0x1919082b, 0x08191908, 0x1919082b, 0x082b0808, 0x1919082b, + 0x19080819, 0x1919082b, 0x19081908, 0x1919082b, 0x19190808, 0x1919082b, 0x192b2b19, 0x1919082b, + 0x2b080808, 0x1919082b, 0x08080819, 0x19191908, 0x08081908, 0x19191908, 0x0808192b, 0x19191908, + 0x08082b19, 0x19191908, 0x08190808, 0x19191908, 0x0819082b, 0x19191908, 0x08191919, 0x19191908, + 0x08192b08, 0x19191908, 0x082b0819, 0x19191908, 0x082b1908, 0x19191908, 0x19080808, 0x19191908, + 0x1908082b, 0x19191908, 0x19081919, 0x19191908, 0x19082b08, 0x19191908, 0x19190819, 0x19191908, + 0x19191908, 0x19191908, 0x192b0808, 0x19191908, 0x2b080819, 0x19191908, 0x2b081908, 0x19191908, + 0x2b190808, 0x19191908, 0x08080808, 0x19191919, 0x0808082b, 0x19191919, 0x08081919, 0x19191919, + 0x08082b08, 0x19191919, 0x08190819, 0x19191919, 0x08191908, 0x19191919, 0x082b0808, 0x19191919, + 0x19080819, 0x19191919, 0x19081908, 0x19191919, 0x19190808, 0x19191919, 0x2b080808, 0x19191919, + 0x08080819, 0x1919192b, 0x08081908, 0x1919192b, 0x08190808, 0x1919192b, 0x082b192b, 0x1919192b, + 0x19080808, 0x1919192b, 0x08080808, 0x19192b08, 0x0808082b, 0x19192b08, 0x08081919, 0x19192b08, + 0x08082b08, 0x19192b08, 0x08190819, 0x19192b08, 0x08191908, 0x19192b08, 0x082b0808, 0x19192b08, + 0x19080819, 0x19192b08, 0x19081908, 0x19192b08, 0x19190808, 0x19192b08, 0x19192b2b, 0x19192b08, + 0x2b080808, 0x19192b08, 0x08080819, 0x19192b19, 0x08081908, 0x19192b19, 0x08190808, 0x19192b19, + 0x19080808, 0x19192b19, 0x08080808, 0x19192b2b, 0x08192b19, 0x19192b2b, 0x2b081919, 0x19192b2b, + 0x2b2b2b08, 0x19192b2b, 0x08080819, 0x192b0808, 0x08081908, 0x192b0808, 0x0808192b, 0x192b0808, + 0x08190808, 0x192b0808, 0x0819082b, 0x192b0808, 0x08191919, 0x192b0808, 0x08192b08, 0x192b0808, + 0x082b0819, 0x192b0808, 0x082b1908, 0x192b0808, 0x19080808, 0x192b0808, 0x19081919, 0x192b0808, + 0x19082b08, 0x192b0808, 0x19190819, 0x192b0808, 0x19191908, 0x192b0808, 0x192b0808, 0x192b0808, + 0x2b081908, 0x192b0808, 0x2b190808, 0x192b0808, 0x08080808, 0x192b0819, 0x0808082b, 0x192b0819, + 0x08081919, 0x192b0819, 0x08082b08, 0x192b0819, 0x08190819, 0x192b0819, 0x08191908, 0x192b0819, + 0x082b0808, 0x192b0819, 0x19080819, 0x192b0819, 0x19081908, 0x192b0819, 0x19190808, 0x192b0819, + 0x2b080808, 0x192b0819, 0x2b192b19, 0x192b0819, 0x08081908, 0x192b082b, 0x08190808, 0x192b082b, + 0x19080808, 0x192b082b, 0x1919192b, 0x192b082b, 0x2b2b0819, 0x192b082b, 0x08080808, 0x192b1908, + 0x08081919, 0x192b1908, 0x08082b08, 0x192b1908, 0x08190819, 0x192b1908, 0x08191908, 0x192b1908, + 0x082b0808, 0x192b1908, 0x19080819, 0x192b1908, 0x19081908, 0x192b1908, 0x19190808, 0x192b1908, + 0x2b080808, 0x192b1908, 0x08080819, 0x192b1919, 0x08081908, 0x192b1919, 0x08190808, 0x192b1919, + 0x19080808, 0x192b1919, 0x19082b2b, 0x192b1919, 0x192b2b08, 0x192b1919, 0x2b19082b, 0x192b1919, + 0x08080808, 0x192b192b, 0x2b191908, 0x192b192b, 0x08080819, 0x192b2b08, 0x08081908, 0x192b2b08, + 0x08190808, 0x192b2b08, 0x192b1919, 0x192b2b08, 0x2b192b08, 0x192b2b08, 0x08080808, 0x192b2b19, + 0x082b2b2b, 0x192b2b19, 0x1908082b, 0x192b2b2b, 0x2b2b0819, 0x192b2b2b, 0x08080808, 0x2b080808, + 0x0808082b, 0x2b080808, 0x08081919, 0x2b080808, 0x08082b08, 0x2b080808, 0x08190819, 0x2b080808, + 0x08191908, 0x2b080808, 0x08192b19, 0x2b080808, 0x082b0808, 0x2b080808, 0x082b1919, 0x2b080808, + 0x19080819, 0x2b080808, 0x19081908, 0x2b080808, 0x19190808, 0x2b080808, 0x1919082b, 0x2b080808, + 0x19191919, 0x2b080808, 0x19192b08, 0x2b080808, 0x192b0819, 0x2b080808, 0x2b080808, 0x2b080808, + 0x2b081919, 0x2b080808, 0x2b190819, 0x2b080808, 0x2b191908, 0x2b080808, 0x08080819, 0x2b080819, + 0x08081908, 0x2b080819, 0x08082b19, 0x2b080819, 0x08190808, 0x2b080819, 0x0819082b, 0x2b080819, + 0x08191919, 0x2b080819, 0x08192b08, 0x2b080819, 0x082b0819, 0x2b080819, 0x082b1908, 0x2b080819, + 0x19080808, 0x2b080819, 0x1908082b, 0x2b080819, 0x19081919, 0x2b080819, 0x19082b08, 0x2b080819, + 0x19190819, 0x2b080819, 0x19191908, 0x2b080819, 0x2b080819, 0x2b080819, 0x2b081908, 0x2b080819, + 0x2b190808, 0x2b080819, 0x2b2b2b19, 0x2b080819, 0x08080808, 0x2b08082b, 0x08081919, 0x2b08082b, + 0x08082b2b, 0x2b08082b, 0x08190819, 0x2b08082b, 0x08191908, 0x2b08082b, 0x19080819, 0x2b08082b, + 0x19081908, 0x2b08082b, 0x19190808, 0x2b08082b, 0x08080819, 0x2b081908, 0x08081908, 0x2b081908, + 0x0808192b, 0x2b081908, 0x08082b19, 0x2b081908, 0x08190808, 0x2b081908, 0x0819082b, 0x2b081908, + 0x08191919, 0x2b081908, 0x08192b08, 0x2b081908, 0x082b0819, 0x2b081908, 0x19080808, 0x2b081908, + 0x1908082b, 0x2b081908, 0x19081919, 0x2b081908, 0x19082b08, 0x2b081908, 0x19190819, 0x2b081908, + 0x19191908, 0x2b081908, 0x192b0808, 0x2b081908, 0x2b080819, 0x2b081908, 0x2b081908, 0x2b081908, + 0x2b190808, 0x2b081908, 0x08080808, 0x2b081919, 0x0808082b, 0x2b081919, 0x08081919, 0x2b081919, + 0x08082b08, 0x2b081919, 0x08190819, 0x2b081919, 0x08191908, 0x2b081919, 0x082b0808, 0x2b081919, + 0x19080819, 0x2b081919, 0x19081908, 0x2b081919, 0x19190808, 0x2b081919, 0x2b080808, 0x2b081919, + 0x2b082b2b, 0x2b081919, 0x08080819, 0x2b08192b, 0x08081908, 0x2b08192b, 0x08190808, 0x2b08192b, + 0x082b2b19, 0x2b08192b, 0x19080808, 0x2b08192b, 0x08080808, 0x2b082b08, 0x08081919, 0x2b082b08, + 0x08190819, 0x2b082b08, 0x08191908, 0x2b082b08, 0x19080819, 0x2b082b08, 0x19081908, 0x2b082b08, + 0x19190808, 0x2b082b08, 0x2b2b082b, 0x2b082b08, 0x08080819, 0x2b082b19, 0x08081908, 0x2b082b19, + 0x19080808, 0x2b082b19, 0x192b1919, 0x2b082b19, 0x082b082b, 0x2b082b2b, 0x19192b08, 0x2b082b2b, + 0x19192b2b, 0x2b082b2b, 0x2b08082b, 0x2b082b2b, 0x2b2b082b, 0x2b082b2b, 0x08080819, 0x2b190808, + 0x08081908, 0x2b190808, 0x08082b19, 0x2b190808, 0x08190808, 0x2b190808, 0x0819082b, 0x2b190808, + 0x08191919, 0x2b190808, 0x08192b08, 0x2b190808, 0x082b1908, 0x2b190808, 0x19080808, 0x2b190808, + 0x1908082b, 0x2b190808, 0x19081919, 0x2b190808, 0x19082b08, 0x2b190808, 0x19190819, 0x2b190808, + 0x19191908, 0x2b190808, 0x192b0808, 0x2b190808, 0x2b080819, 0x2b190808, 0x2b081908, 0x2b190808, + 0x2b190808, 0x2b190808, 0x08080808, 0x2b190819, 0x08081919, 0x2b190819, 0x08190819, 0x2b190819, + 0x08191908, 0x2b190819, 0x19080819, 0x2b190819, 0x19081908, 0x2b190819, 0x19190808, 0x2b190819, + 0x19192b2b, 0x2b190819, 0x08080819, 0x2b19082b, 0x08081908, 0x2b19082b, 0x08190808, 0x2b19082b, + 0x19080808, 0x2b19082b, 0x2b2b192b, 0x2b19082b, 0x08080808, 0x2b191908, 0x0808082b, 0x2b191908, + 0x08081919, 0x2b191908, 0x08082b08, 0x2b191908, 0x08190819, 0x2b191908, 0x08191908, 0x2b191908, + 0x082b0808, 0x2b191908, 0x19080819, 0x2b191908, 0x19081908, 0x2b191908, 0x19190808, 0x2b191908, + 0x2b080808, 0x2b191908, 0x2b19192b, 0x2b191908, 0x08080819, 0x2b191919, 0x08081908, 0x2b191919, + 0x08190808, 0x2b191919, 0x19080808, 0x2b191919, 0x2b192b08, 0x2b191919, 0x2b2b0819, 0x2b191919, + 0x08080808, 0x2b19192b, 0x1908192b, 0x2b19192b, 0x192b1908, 0x2b19192b, 0x08080819, 0x2b192b08, + 0x08081908, 0x2b192b08, 0x08190808, 0x2b192b08, 0x082b192b, 0x2b192b08, 0x19080808, 0x2b192b08, + 0x2b2b2b19, 0x2b192b08, 0x08080808, 0x2b192b19, 0x19082b19, 0x2b192b19, 0x1919082b, 0x2b192b19, + 0x2b190808, 0x2b192b2b, 0x08080808, 0x2b2b0808, 0x08081919, 0x2b2b0808, 0x08082b2b, 0x2b2b0808, + 0x08191908, 0x2b2b0808, 0x082b082b, 0x2b2b0808, 0x082b2b2b, 0x2b2b0808, 0x19080819, 0x2b2b0808, + 0x19081908, 0x2b2b0808, 0x19190808, 0x2b2b0808, 0x2b2b082b, 0x2b2b0808, 0x2b2b2b2b, 0x2b2b0808, + 0x19080808, 0x2b2b0819, 0x192b1919, 0x2b2b0819, 0x0808082b, 0x2b2b082b, 0x08082b2b, 0x2b2b082b, + 0x082b082b, 0x2b2b082b, 0x082b2b08, 0x2b2b082b, 0x082b2b2b, 0x2b2b082b, 0x2b08082b, 0x2b2b082b, + 0x2b082b08, 0x2b2b082b, 0x2b082b2b, 0x2b2b082b, 0x2b2b2b08, 0x2b2b082b, 0x08080819, 0x2b2b1908, + 0x08081908, 0x2b2b1908, 0x08190808, 0x2b2b1908, 0x19080808, 0x2b2b1908, 0x2b082b19, 0x2b2b1908, + 0x2b2b1908, 0x2b2b1908, 0x08080808, 0x2b2b1919, 0x08192b19, 0x2b2b1919, 0x19190819, 0x2b2b192b, + 0x08082b2b, 0x2b2b2b08, 0x082b2b08, 0x2b2b2b08, 0x2b2b082b, 0x2b2b2b08, 0x19191908, 0x2b2b2b19, + 0x2b08192b, 0x2b2b2b19, 0x08082b08, 0x2b2b2b2b, 0x08082b2b, 0x2b2b2b2b, 0x082b0808, 0x2b2b2b2b, + 0x082b082b, 0x2b2b2b2b, 0x082b2b08, 0x2b2b2b2b, 0x2b082b08, 0x2b2b2b2b, 0x2b2b2b2b, 0x2b2b2b2b +); +#enddecl(IQ2_S_GRID) + +#decl(IQ3_XSS_GRID) + +const iq3xxs_grid = array( + 0x04040404, 0x04040414, 0x04040424, 0x04040c0c, 0x04040c1c, 0x04040c3e, 0x04041404, 0x04041414, + 0x04041c0c, 0x04042414, 0x04043e1c, 0x04043e2c, 0x040c040c, 0x040c041c, 0x040c0c04, 0x040c0c14, + 0x040c140c, 0x040c142c, 0x040c1c04, 0x040c1c14, 0x040c240c, 0x040c2c24, 0x040c3e04, 0x04140404, + 0x04140414, 0x04140424, 0x04140c0c, 0x04141404, 0x04141414, 0x04141c0c, 0x04141c1c, 0x04141c3e, + 0x04142c0c, 0x04142c3e, 0x04143e2c, 0x041c040c, 0x041c043e, 0x041c0c04, 0x041c0c14, 0x041c142c, + 0x041c3e04, 0x04240c1c, 0x04241c3e, 0x04242424, 0x04242c3e, 0x04243e1c, 0x04243e2c, 0x042c040c, + 0x042c043e, 0x042c1c14, 0x042c2c14, 0x04341c2c, 0x04343424, 0x043e0c04, 0x043e0c24, 0x043e0c34, + 0x043e241c, 0x043e340c, 0x0c04040c, 0x0c04041c, 0x0c040c04, 0x0c040c14, 0x0c04140c, 0x0c04141c, + 0x0c041c04, 0x0c041c14, 0x0c041c24, 0x0c04243e, 0x0c042c04, 0x0c0c0404, 0x0c0c0414, 0x0c0c0c0c, + 0x0c0c1404, 0x0c0c1414, 0x0c14040c, 0x0c14041c, 0x0c140c04, 0x0c140c14, 0x0c14140c, 0x0c141c04, + 0x0c143e14, 0x0c1c0404, 0x0c1c0414, 0x0c1c1404, 0x0c1c1c0c, 0x0c1c2434, 0x0c1c3434, 0x0c24040c, + 0x0c24042c, 0x0c242c04, 0x0c2c1404, 0x0c2c1424, 0x0c2c2434, 0x0c2c3e0c, 0x0c34042c, 0x0c3e1414, + 0x0c3e2404, 0x14040404, 0x14040414, 0x14040c0c, 0x14040c1c, 0x14041404, 0x14041414, 0x14041434, + 0x14041c0c, 0x14042414, 0x140c040c, 0x140c041c, 0x140c042c, 0x140c0c04, 0x140c0c14, 0x140c140c, + 0x140c1c04, 0x140c341c, 0x140c343e, 0x140c3e04, 0x14140404, 0x14140414, 0x14140c0c, 0x14140c3e, + 0x14141404, 0x14141414, 0x14141c3e, 0x14142404, 0x14142c2c, 0x141c040c, 0x141c0c04, 0x141c0c24, + 0x141c3e04, 0x141c3e24, 0x14241c2c, 0x14242c1c, 0x142c041c, 0x142c143e, 0x142c240c, 0x142c3e24, + 0x143e040c, 0x143e041c, 0x143e0c34, 0x143e242c, 0x1c04040c, 0x1c040c04, 0x1c040c14, 0x1c04140c, + 0x1c04141c, 0x1c042c04, 0x1c04342c, 0x1c043e14, 0x1c0c0404, 0x1c0c0414, 0x1c0c1404, 0x1c0c1c0c, + 0x1c0c2424, 0x1c0c2434, 0x1c14040c, 0x1c14041c, 0x1c140c04, 0x1c14142c, 0x1c142c14, 0x1c143e14, + 0x1c1c0c0c, 0x1c1c1c1c, 0x1c241c04, 0x1c24243e, 0x1c243e14, 0x1c2c0404, 0x1c2c0434, 0x1c2c1414, + 0x1c2c2c2c, 0x1c340c24, 0x1c341c34, 0x1c34341c, 0x1c3e1c1c, 0x1c3e3404, 0x24040424, 0x24040c3e, + 0x24041c2c, 0x24041c3e, 0x24042c1c, 0x24042c3e, 0x240c3e24, 0x24141404, 0x24141c3e, 0x24142404, + 0x24143404, 0x24143434, 0x241c043e, 0x241c242c, 0x24240424, 0x24242c0c, 0x24243424, 0x242c142c, + 0x242c241c, 0x242c3e04, 0x243e042c, 0x243e0c04, 0x243e0c14, 0x243e1c04, 0x2c040c14, 0x2c04240c, + 0x2c043e04, 0x2c0c0404, 0x2c0c0434, 0x2c0c1434, 0x2c0c2c2c, 0x2c140c24, 0x2c141c14, 0x2c143e14, + 0x2c1c0414, 0x2c1c2c1c, 0x2c240c04, 0x2c24141c, 0x2c24143e, 0x2c243e14, 0x2c2c0414, 0x2c2c1c0c, + 0x2c342c04, 0x2c3e1424, 0x2c3e2414, 0x34041424, 0x34042424, 0x34042434, 0x34043424, 0x340c140c, + 0x340c340c, 0x34140c3e, 0x34143424, 0x341c1c04, 0x341c1c34, 0x34242424, 0x342c042c, 0x342c2c14, + 0x34341c1c, 0x343e041c, 0x343e140c, 0x3e04041c, 0x3e04042c, 0x3e04043e, 0x3e040c04, 0x3e041c14, + 0x3e042c14, 0x3e0c1434, 0x3e0c2404, 0x3e140c14, 0x3e14242c, 0x3e142c14, 0x3e1c0404, 0x3e1c0c2c, + 0x3e1c1c1c, 0x3e1c3404, 0x3e24140c, 0x3e24240c, 0x3e2c0404, 0x3e2c0414, 0x3e2c1424, 0x3e341c04 +); +#enddecl(IQ3_XSS_GRID) + +#decl(IQ3_S_GRID) + +const iq3s_grid = array( + 0x01010101, 0x01010103, 0x01010105, 0x0101010b, 0x0101010f, 0x01010301, 0x01010303, 0x01010305, + 0x01010309, 0x0101030d, 0x01010501, 0x01010503, 0x0101050b, 0x01010707, 0x01010901, 0x01010905, + 0x0101090b, 0x0101090f, 0x01010b03, 0x01010b07, 0x01010d01, 0x01010d05, 0x01010f03, 0x01010f09, + 0x01010f0f, 0x01030101, 0x01030103, 0x01030105, 0x01030109, 0x01030301, 0x01030303, 0x0103030b, + 0x01030501, 0x01030507, 0x0103050f, 0x01030703, 0x0103070b, 0x01030909, 0x01030d03, 0x01030d0b, + 0x01030f05, 0x01050101, 0x01050103, 0x0105010b, 0x0105010f, 0x01050301, 0x01050307, 0x0105030d, + 0x01050503, 0x0105050b, 0x01050701, 0x01050709, 0x01050905, 0x0105090b, 0x0105090f, 0x01050b03, + 0x01050b07, 0x01050f01, 0x01050f07, 0x01070107, 0x01070303, 0x0107030b, 0x01070501, 0x01070505, + 0x01070703, 0x01070707, 0x0107070d, 0x01070909, 0x01070b01, 0x01070b05, 0x01070d0f, 0x01070f03, + 0x01070f0b, 0x01090101, 0x01090307, 0x0109030f, 0x01090503, 0x01090509, 0x01090705, 0x01090901, + 0x01090907, 0x01090b03, 0x01090f01, 0x010b0105, 0x010b0109, 0x010b0501, 0x010b0505, 0x010b050d, + 0x010b0707, 0x010b0903, 0x010b090b, 0x010b090f, 0x010b0d0d, 0x010b0f07, 0x010d010d, 0x010d0303, + 0x010d0307, 0x010d0703, 0x010d0b05, 0x010d0f03, 0x010f0101, 0x010f0105, 0x010f0109, 0x010f0501, + 0x010f0505, 0x010f050d, 0x010f0707, 0x010f0b01, 0x010f0b09, 0x03010101, 0x03010103, 0x03010105, + 0x03010109, 0x03010301, 0x03010303, 0x03010307, 0x0301030b, 0x0301030f, 0x03010501, 0x03010505, + 0x03010703, 0x03010709, 0x0301070d, 0x03010b09, 0x03010b0d, 0x03010d03, 0x03010f05, 0x03030101, + 0x03030103, 0x03030107, 0x0303010d, 0x03030301, 0x03030309, 0x03030503, 0x03030701, 0x03030707, + 0x03030903, 0x03030b01, 0x03030b05, 0x03030f01, 0x03030f0d, 0x03050101, 0x03050305, 0x0305030b, + 0x0305030f, 0x03050501, 0x03050509, 0x03050705, 0x03050901, 0x03050907, 0x03050b0b, 0x03050d01, + 0x03050f05, 0x03070103, 0x03070109, 0x0307010f, 0x03070301, 0x03070307, 0x03070503, 0x0307050f, + 0x03070701, 0x03070709, 0x03070903, 0x03070d05, 0x03070f01, 0x03090107, 0x0309010b, 0x03090305, + 0x03090309, 0x03090703, 0x03090707, 0x03090905, 0x0309090d, 0x03090b01, 0x03090b09, 0x030b0103, + 0x030b0301, 0x030b0307, 0x030b0503, 0x030b0701, 0x030b0705, 0x030b0b03, 0x030d0501, 0x030d0509, + 0x030d050f, 0x030d0909, 0x030d090d, 0x030f0103, 0x030f0107, 0x030f0301, 0x030f0305, 0x030f0503, + 0x030f070b, 0x030f0903, 0x030f0d05, 0x030f0f01, 0x05010101, 0x05010103, 0x05010107, 0x0501010b, + 0x0501010f, 0x05010301, 0x05010305, 0x05010309, 0x0501030d, 0x05010503, 0x05010507, 0x0501050f, + 0x05010701, 0x05010705, 0x05010903, 0x05010907, 0x0501090b, 0x05010b01, 0x05010b05, 0x05010d0f, + 0x05010f01, 0x05010f07, 0x05010f0b, 0x05030101, 0x05030105, 0x05030301, 0x05030307, 0x0503030f, + 0x05030505, 0x0503050b, 0x05030703, 0x05030709, 0x05030905, 0x05030b03, 0x05050103, 0x05050109, + 0x0505010f, 0x05050503, 0x05050507, 0x05050701, 0x0505070f, 0x05050903, 0x05050b07, 0x05050b0f, + 0x05050f03, 0x05050f09, 0x05070101, 0x05070105, 0x0507010b, 0x05070303, 0x05070505, 0x05070509, + 0x05070703, 0x05070707, 0x05070905, 0x05070b01, 0x05070d0d, 0x05090103, 0x0509010f, 0x05090501, + 0x05090507, 0x05090705, 0x0509070b, 0x05090903, 0x05090f05, 0x05090f0b, 0x050b0109, 0x050b0303, + 0x050b0505, 0x050b070f, 0x050b0901, 0x050b0b07, 0x050b0f01, 0x050d0101, 0x050d0105, 0x050d010f, + 0x050d0503, 0x050d0b0b, 0x050d0d03, 0x050f010b, 0x050f0303, 0x050f050d, 0x050f0701, 0x050f0907, + 0x050f0b01, 0x07010105, 0x07010303, 0x07010307, 0x0701030b, 0x0701030f, 0x07010505, 0x07010703, + 0x07010707, 0x0701070b, 0x07010905, 0x07010909, 0x0701090f, 0x07010b03, 0x07010d07, 0x07010f03, + 0x07030103, 0x07030107, 0x0703010b, 0x07030309, 0x07030503, 0x07030507, 0x07030901, 0x07030d01, + 0x07030f05, 0x07030f0d, 0x07050101, 0x07050305, 0x07050501, 0x07050705, 0x07050709, 0x07050b01, + 0x07070103, 0x07070301, 0x07070309, 0x07070503, 0x07070507, 0x0707050f, 0x07070701, 0x07070903, + 0x07070907, 0x0707090f, 0x07070b0b, 0x07070f07, 0x07090107, 0x07090303, 0x0709030d, 0x07090505, + 0x07090703, 0x07090b05, 0x07090d01, 0x07090d09, 0x070b0103, 0x070b0301, 0x070b0305, 0x070b050b, + 0x070b0705, 0x070b0909, 0x070b0b0d, 0x070b0f07, 0x070d030d, 0x070d0903, 0x070f0103, 0x070f0107, + 0x070f0501, 0x070f0505, 0x070f070b, 0x09010101, 0x09010109, 0x09010305, 0x09010501, 0x09010509, + 0x0901050f, 0x09010705, 0x09010903, 0x09010b01, 0x09010f01, 0x09030105, 0x0903010f, 0x09030303, + 0x09030307, 0x09030505, 0x09030701, 0x0903070b, 0x09030907, 0x09030b03, 0x09030b0b, 0x09050103, + 0x09050107, 0x09050301, 0x0905030b, 0x09050503, 0x09050707, 0x09050901, 0x09050b0f, 0x09050d05, + 0x09050f01, 0x09070109, 0x09070303, 0x09070307, 0x09070501, 0x09070505, 0x09070703, 0x0907070b, + 0x09090101, 0x09090105, 0x09090509, 0x0909070f, 0x09090901, 0x09090f03, 0x090b010b, 0x090b010f, + 0x090b0503, 0x090b0d05, 0x090d0307, 0x090d0709, 0x090d0d01, 0x090f0301, 0x090f030b, 0x090f0701, + 0x090f0907, 0x090f0b03, 0x0b010105, 0x0b010301, 0x0b010309, 0x0b010505, 0x0b010901, 0x0b010909, + 0x0b01090f, 0x0b010b05, 0x0b010d0d, 0x0b010f09, 0x0b030103, 0x0b030107, 0x0b03010b, 0x0b030305, + 0x0b030503, 0x0b030705, 0x0b030f05, 0x0b050101, 0x0b050303, 0x0b050507, 0x0b050701, 0x0b05070d, + 0x0b050b07, 0x0b070105, 0x0b07010f, 0x0b070301, 0x0b07050f, 0x0b070909, 0x0b070b03, 0x0b070d0b, + 0x0b070f07, 0x0b090103, 0x0b090109, 0x0b090501, 0x0b090705, 0x0b09090d, 0x0b0b0305, 0x0b0b050d, + 0x0b0b0b03, 0x0b0b0b07, 0x0b0d0905, 0x0b0f0105, 0x0b0f0109, 0x0b0f0505, 0x0d010303, 0x0d010307, + 0x0d01030b, 0x0d010703, 0x0d010707, 0x0d010d01, 0x0d030101, 0x0d030501, 0x0d03050f, 0x0d030d09, + 0x0d050305, 0x0d050709, 0x0d050905, 0x0d050b0b, 0x0d050d05, 0x0d050f01, 0x0d070101, 0x0d070309, + 0x0d070503, 0x0d070901, 0x0d09050b, 0x0d090907, 0x0d090d05, 0x0d0b0101, 0x0d0b0107, 0x0d0b0709, + 0x0d0b0d01, 0x0d0d010b, 0x0d0d0901, 0x0d0f0303, 0x0d0f0307, 0x0f010101, 0x0f010109, 0x0f01010f, + 0x0f010501, 0x0f010505, 0x0f01070d, 0x0f010901, 0x0f010b09, 0x0f010d05, 0x0f030105, 0x0f030303, + 0x0f030509, 0x0f030907, 0x0f03090b, 0x0f050103, 0x0f050109, 0x0f050301, 0x0f05030d, 0x0f050503, + 0x0f050701, 0x0f050b03, 0x0f070105, 0x0f070705, 0x0f07070b, 0x0f070b07, 0x0f090103, 0x0f09010b, + 0x0f090307, 0x0f090501, 0x0f090b01, 0x0f0b0505, 0x0f0b0905, 0x0f0d0105, 0x0f0d0703, 0x0f0f0101 +); +#enddecl(IQ3_S_GRID) + +#decl(IQ1_GRID) + +const IQ1_DELTA: f32 = 0.125; + +const iq1_grid = array( + 0xfffdffff, 0xfff7fff0, 0xffccfff5, 0xffdfffc0, 0xffd7ffdd, 0xff30ffd5, 0xff03ff0c, 0xff10ff01, + 0xff7dff7f, 0xff75ff77, 0xff5fff40, 0xff57ff5d, 0xfcf3ff55, 0xfcccfcf0, 0xfcc1fcc3, 0xfcc5fcc4, + 0xfc3cfcd0, 0xfc34fc31, 0xfc00fc0d, 0xfc1cfc05, 0xfc11fc13, 0xfc70fc17, 0xfc43fc4c, 0xfc50fc41, + 0xfdfdfdff, 0xfdf5fdf7, 0xfddffdc0, 0xfdd7fddd, 0xfd30fdd5, 0xfd04fd0c, 0xfd14fd13, 0xfd7dfd7f, + 0xfd75fd77, 0xfd40fd4c, 0xfd5ffd44, 0xfd57fd5d, 0xf3ccfd55, 0xf3c1f3c3, 0xf33cf3d0, 0xf300f334, + 0xf313f305, 0xf34cf310, 0xf350f344, 0xf0f3f0fc, 0xf0f1f0f0, 0xf0c7f0c0, 0xf0d4f0c5, 0xf030f03f, + 0xf00ff035, 0xf003f00c, 0xf001f000, 0xf01ff004, 0xf010f01d, 0xf015f017, 0xf04cf07c, 0xf047f040, + 0xf05cf045, 0xf050f053, 0xf054f051, 0xf1c4f1c3, 0xf133f13c, 0xf10df10f, 0xf107f100, 0xf11cf11f, + 0xf114f111, 0xf14cf170, 0xf144f143, 0xf7fdf7ff, 0xf7f5f7f7, 0xf7dff7c0, 0xf7d7f7dd, 0xf730f7d5, + 0xf701f70c, 0xf77ff710, 0xf777f77d, 0xf740f775, 0xf75df75f, 0xf755f757, 0xf4ccf4f0, 0xf4c4f4c3, + 0xf4d0f4d3, 0xf40ff43c, 0xf400f40c, 0xf413f41c, 0xf44cf414, 0xf441f443, 0xf450f444, 0xf5fdf5ff, + 0xf5f5f5f7, 0xf5dff5c0, 0xf5d7f5dd, 0xf530f5d5, 0xf504f50c, 0xf510f51c, 0xf57df57f, 0xf577f570, + 0xf540f575, 0xf55df55f, 0xf555f557, 0xcfcccfcf, 0xcfc4cfc3, 0xcfd0cfd3, 0xcf33cf3c, 0xcf00cf0f, + 0xcf1ccf07, 0xcf10cf13, 0xcf4ccf14, 0xcf41cf43, 0xcf50cf5c, 0xccf3ccfc, 0xccf4ccf1, 0xcccdcccf, + 0xccc7ccc0, 0xccd3ccdc, 0xcc30ccd4, 0xcc0fcc35, 0xcc0dcc0c, 0xcc00cc03, 0xcc04cc01, 0xcc10cc1f, + 0xcc4dcc73, 0xcc5ccc40, 0xcdcccc53, 0xcdc1cdc3, 0xcd3fcdd0, 0xcd34cd31, 0xcd00cd0d, 0xcd05cd07, + 0xcd11cd13, 0xcd4ccd70, 0xcd41cd43, 0xc3fccd50, 0xc3f4c3f1, 0xc3c0c3c3, 0xc3c4c3c7, 0xc3d1c3dc, + 0xc330c33c, 0xc337c331, 0xc30cc335, 0xc300c303, 0xc304c301, 0xc310c31d, 0xc373c317, 0xc34fc374, + 0xc340c343, 0xc344c347, 0xc35cc345, 0xc350c353, 0xc0fdc354, 0xc0f5c0f0, 0xc0c3c0cc, 0xc0c1c0c0, + 0xc0dfc0c4, 0xc0d0c0dd, 0xc0d5c0d7, 0xc033c03c, 0xc031c030, 0xc00dc00c, 0xc000c003, 0xc004c001, + 0xc01cc005, 0xc010c013, 0xc014c011, 0xc07dc07f, 0xc070c073, 0xc075c077, 0xc04cc04f, 0xc040c043, + 0xc044c041, 0xc05fc045, 0xc050c05d, 0xc1f3c1fc, 0xc1f1c1f0, 0xc1c1c1c0, 0xc1c5c1c7, 0xc1d1c1dc, + 0xc13dc13f, 0xc130c133, 0xc135c137, 0xc100c10c, 0xc107c101, 0xc11cc104, 0xc110c113, 0xc114c117, + 0xc171c115, 0xc14dc175, 0xc153c140, 0xc7ccc154, 0xc7d0c7c1, 0xc733c73c, 0xc734c731, 0xc700c70f, + 0xc705c707, 0xc71cc71f, 0xc711c713, 0xc770c714, 0xc743c74c, 0xc4cfc750, 0xc4c0c4cd, 0xc4dcc4c5, + 0xc43dc4d0, 0xc430c433, 0xc40cc437, 0xc400c403, 0xc404c401, 0xc41fc405, 0xc415c410, 0xc44cc474, + 0xc440c44d, 0xc45cc447, 0xc454c451, 0xc5c1c5f4, 0xc5d1c5d3, 0xc531c533, 0xc50fc534, 0xc500c50d, + 0xc51cc507, 0xc514c511, 0xc54cc570, 0xc545c541, 0xdffddfff, 0xdff5dff7, 0xdfdfdfc0, 0xdfd0dfdd, + 0xdfd5dfd7, 0xdf0cdf30, 0xdf1cdf04, 0xdf7fdf10, 0xdf77df7d, 0xdf40df75, 0xdf5ddf5f, 0xdf57df50, + 0xdcf0df55, 0xdcc3dccc, 0xdcd0dcc4, 0xdc33dc3d, 0xdc00dc34, 0xdc05dc07, 0xdc13dc1c, 0xdc11dc10, + 0xdc4fdc70, 0xdc44dc41, 0xddfcdc50, 0xddf5ddf7, 0xddc0ddcc, 0xdddddddf, 0xddd5ddd7, 0xdd0cdd30, + 0xdd04dd01, 0xdd7cdd10, 0xdd75dd77, 0xdd40dd4c, 0xdd5ddd5f, 0xdd55dd57, 0xd3c3d3f0, 0xd3c4d3c1, + 0xd333d3d0, 0xd331d330, 0xd30dd334, 0xd307d300, 0xd311d305, 0xd34cd370, 0xd344d343, 0xd350d35c, + 0xd0c0d0f4, 0xd0d4d0dc, 0xd030d03f, 0xd00cd037, 0xd000d003, 0xd01dd004, 0xd017d010, 0xd04fd074, + 0xd040d043, 0xd045d047, 0xd053d05c, 0xd054d051, 0xd1cfd1f0, 0xd1c4d1cd, 0xd13cd1d0, 0xd100d134, + 0xd11cd11f, 0xd173d114, 0xd14fd171, 0xd7ffd145, 0xd7f7d7fd, 0xd7c0d7f5, 0xd7ddd7df, 0xd7d5d7d7, + 0xd70cd730, 0xd710d703, 0xd77dd77f, 0xd775d777, 0xd75dd75f, 0xd755d757, 0xd4ccd4f4, 0xd4c4d4c3, + 0xd431d4d0, 0xd40dd434, 0xd41cd400, 0xd411d413, 0xd470d414, 0xd441d44f, 0xd453d444, 0xd5ffd450, + 0xd5f7d5fd, 0xd5dfd5f5, 0xd5d7d5dd, 0xd530d5d5, 0xd501d50c, 0xd510d504, 0xd57dd57f, 0xd575d577, + 0xd55fd540, 0xd557d55d, 0x3ff0d555, 0x3fc13fcc, 0x3f343fd0, 0x3f003f0d, 0x3f053f07, 0x3f133f1c, + 0x3f433f11, 0x3f5c3f44, 0x3cff3f51, 0x3cf33cfc, 0x3cf43cf1, 0x3cc03ccd, 0x3cc73cc1, 0x3cdc3cc5, + 0x3cd43cd1, 0x3c373c30, 0x3c0c3c35, 0x3c003c03, 0x3c043c01, 0x3c103c05, 0x3c153c17, 0x3c733c7c, + 0x3c4f3c71, 0x3c403c4d, 0x3c5c3c5f, 0x3df03c5d, 0x3dc33dcc, 0x3dd03dc1, 0x3d0d3d3c, 0x3d053d00, + 0x3d143d13, 0x3d433d74, 0x33fc3d50, 0x33c433c0, 0x333033d4, 0x33353337, 0x3303330c, 0x33013300, + 0x331d331c, 0x33173310, 0x337c3315, 0x33743371, 0x334d334f, 0x335f3340, 0x3354335c, 0x30fd30fc, + 0x30f530f0, 0x30c330cc, 0x30c130c0, 0x30df30c4, 0x30d530d0, 0x3033303c, 0x30313030, 0x300f3034, + 0x3003300c, 0x30013000, 0x30043007, 0x3013301c, 0x30113010, 0x307d3014, 0x30703073, 0x304c3077, + 0x30403043, 0x30443041, 0x30503045, 0x30553057, 0x31f031fc, 0x31c331f4, 0x31c731c0, 0x31dc31c5, + 0x31d431d3, 0x313d313f, 0x31373130, 0x310c310f, 0x3100310d, 0x31043101, 0x3110311d, 0x317c3117, + 0x31753170, 0x31403143, 0x3153315c, 0x37f03151, 0x37c037cc, 0x37d037c5, 0x3734373d, 0x3700370f, + 0x371c3707, 0x37113713, 0x37703714, 0x3743374c, 0x37443741, 0x34fc3750, 0x34f134f0, 0x34cf34f5, + 0x34c034c3, 0x34dc34c7, 0x34d134d3, 0x3430343f, 0x340c3435, 0x3403340d, 0x34013400, 0x341f3404, + 0x3410341d, 0x34153411, 0x34743471, 0x3440344d, 0x34473441, 0x3453345c, 0x34543451, 0x353335c1, + 0x35343531, 0x35073500, 0x35133505, 0x35433514, 0x0ffc3550, 0x0ff00ff3, 0x0ff40ff1, 0x0fc00fcd, + 0x0fdc0fc5, 0x0fd40fd3, 0x0f300f3f, 0x0f0c0f37, 0x0f000f03, 0x0f040f01, 0x0f170f10, 0x0f740f71, + 0x0f470f40, 0x0f5c0f5f, 0x0f540f51, 0x0cf70cf0, 0x0cf50cf4, 0x0cc30ccc, 0x0cc10cc0, 0x0cc40cc7, + 0x0cd00cdf, 0x0cd70cd1, 0x0c3c0cd5, 0x0c300c33, 0x0c340c31, 0x0c0c0c0f, 0x0c030c0d, 0x0c010c00, + 0x0c040c07, 0x0c1c0c05, 0x0c100c13, 0x0c140c11, 0x0c700c7d, 0x0c430c4c, 0x0c410c40, 0x0c5f0c44, + 0x0c550c50, 0x0df10dfc, 0x0dc00dcd, 0x0ddc0dc5, 0x0d3d0dd3, 0x0d350d30, 0x0d030d0c, 0x0d010d00, + 0x0d1d0d04, 0x0d700d10, 0x0d4d0d4f, 0x0d440d40, 0x0d530d45, 0x03f003f3, 0x03c303cc, 0x03c103c0, + 0x03c403c7, 0x03d003dc, 0x03d503d7, 0x0333033c, 0x03310330, 0x03350334, 0x030c030f, 0x03000303, + 0x03070301, 0x03050304, 0x031d031c, 0x03100313, 0x03140311, 0x0377037f, 0x034c0375, 0x03400343, + 0x03440341, 0x0353035c, 0x03550350, 0x00fd00fc, 0x00f000f3, 0x00f400f1, 0x00cc00cf, 0x00c300cd, + 0x00c100c0, 0x00c500c4, 0x00d300dc, 0x00d100d0, 0x003f00d4, 0x003d003c, 0x00300033, 0x00370031, + 0x000f0034, 0x000d000c, 0x00000003, 0x00070001, 0x00050004, 0x001c001f, 0x00100013, 0x00170011, + 0x00150014, 0x0073007c, 0x00740070, 0x004f0075, 0x0043004c, 0x00410040, 0x00440047, 0x0053005c, + 0x00510050, 0x01ff0054, 0x01fd01fc, 0x01f101f3, 0x01f401f7, 0x01c301cc, 0x01c701c0, 0x01df01c4, + 0x01dd01dc, 0x01d001d3, 0x01d701d1, 0x013c01d4, 0x01310130, 0x01340137, 0x010f0135, 0x010d010c, + 0x01000103, 0x01070101, 0x01050104, 0x0113011c, 0x01140110, 0x0170017d, 0x01770171, 0x01750174, + 0x0140014c, 0x015d0145, 0x01510150, 0x01540157, 0x07f007f3, 0x07f407f1, 0x07c007cf, 0x07dc07c7, + 0x073007d5, 0x07350737, 0x0703070c, 0x07010700, 0x07040707, 0x071d071f, 0x07100713, 0x0774077d, + 0x074d074f, 0x07470740, 0x0754075c, 0x04fd04fc, 0x04f504f0, 0x04c304cc, 0x04c104c0, 0x04d004c4, + 0x0433043c, 0x04310430, 0x040f0434, 0x040d040c, 0x04000403, 0x04070401, 0x04050404, 0x0413041c, + 0x04110410, 0x047c0414, 0x04740470, 0x0443044c, 0x04410440, 0x04440447, 0x05f30450, 0x05c005f7, + 0x05df05c5, 0x05d105d0, 0x053005d4, 0x05340537, 0x0500050c, 0x05070501, 0x051d0504, 0x05170510, + 0x057c0515, 0x054d0575, 0x05410540, 0x05450547, 0x1ff0055c, 0x1fc11fc3, 0x1fd01fc4, 0x1f0f1f33, + 0x1f011f00, 0x1f051f07, 0x1f131f1c, 0x1f141f11, 0x1f411f7c, 0x1cfc1f50, 0x1cf11cf3, 0x1ccd1cf4, + 0x1cdc1cc0, 0x1cd11cdd, 0x1c301cd4, 0x1c0c1c34, 0x1c011c00, 0x1c101c04, 0x1c151c11, 0x1c751c73, + 0x1c401c4d, 0x1c511c5c, 0x1dcc1c54, 0x1dc41dc1, 0x1d3c1d3f, 0x1d001d31, 0x1d071d01, 0x1d701d1f, + 0x1d411d4c, 0x13cc1d50, 0x13c013cd, 0x13c513c1, 0x13d113dc, 0x133f13d4, 0x1330133d, 0x13351337, + 0x1303130c, 0x13011300, 0x13051304, 0x131d131f, 0x13731310, 0x13741370, 0x134d134f, 0x13401343, + 0x13471341, 0x135c1345, 0x13541353, 0x10f710f0, 0x10cc10f5, 0x10c110c0, 0x103310c4, 0x10311030, + 0x100f1034, 0x1003100c, 0x10011000, 0x101c1004, 0x10101013, 0x10141011, 0x10741071, 0x104c1075, + 0x10411040, 0x10451044, 0x1050105d, 0x10571051, 0x11f411fd, 0x11df11c0, 0x11d711d1, 0x113f11d4, + 0x11371130, 0x110c1135, 0x11001103, 0x11071101, 0x111f1105, 0x11171110, 0x117d117f, 0x11751170, + 0x11411143, 0x11441147, 0x1153115f, 0x11551151, 0x17c417c1, 0x173c17d0, 0x1700170d, 0x171c1705, + 0x17701714, 0x1747174c, 0x14fc1751, 0x14cf14f3, 0x14dc14c0, 0x14d114d3, 0x143f14d4, 0x1430143c, + 0x14371431, 0x1403140c, 0x14011400, 0x141f1404, 0x14151410, 0x1473147d, 0x14401475, 0x1453145c, + 0x14541450, 0x15c115cc, 0x153c15c7, 0x15341533, 0x1500150f, 0x15051507, 0x15101513, 0x15711514, + 0x15471543, 0x15511545, 0x7ffd7fff, 0x7ff57ff7, 0x7fdd7fdf, 0x7fd57fd7, 0x7f0f7f30, 0x7f037f0c, + 0x7f047f01, 0x7f7f7f10, 0x7f777f7d, 0x7f407f75, 0x7f5d7f5f, 0x7f557f57, 0x7ccc7cf0, 0x7cc17cc3, + 0x7cd07cc4, 0x7c337c3c, 0x7c0f7c34, 0x7c007c0d, 0x7c077c01, 0x7c137c04, 0x7c147c11, 0x7c747c70, + 0x7c417c43, 0x7c507c44, 0x7dfd7dff, 0x7df57df7, 0x7ddf7dc0, 0x7dd77ddd, 0x7d0c7dd5, 0x7d047d03, + 0x7d7f7d10, 0x7d777d7d, 0x7d407d75, 0x7d5d7d5f, 0x7d557d57, 0x73c473c3, 0x7333733c, 0x7300730c, + 0x731c7305, 0x73147313, 0x73447343, 0x70f470fc, 0x70c070cd, 0x70d170c5, 0x703f70d4, 0x7030703c, + 0x700c7037, 0x70007003, 0x70047001, 0x70107005, 0x70177011, 0x707c7015, 0x70717073, 0x704f7074, + 0x7040704d, 0x70517047, 0x71c171cc, 0x71d071c4, 0x7133713c, 0x71357134, 0x7100710f, 0x71057104, + 0x7111711c, 0x71707115, 0x7145714c, 0x77ff7153, 0x77f777fd, 0x77c077f5, 0x77dd77df, 0x77d577d7, + 0x7730773c, 0x7703770c, 0x77107704, 0x777f7714, 0x7777777d, 0x77407775, 0x775d775f, 0x77557757, + 0x74f174f0, 0x74c374cc, 0x74d074c1, 0x7433743c, 0x74347431, 0x740d740f, 0x74057400, 0x7413741c, + 0x74417470, 0x74507444, 0x75fd75ff, 0x75f575f7, 0x75df75c0, 0x75d775dd, 0x753075d5, 0x7503750c, + 0x757f7501, 0x7577757d, 0x75407575, 0x755d755f, 0x75557557, 0x4fcc4ff0, 0x4fc74fc1, 0x4fd04fc4, + 0x4f314f3c, 0x4f004f34, 0x4f054f07, 0x4f154f14, 0x4f4c4f70, 0x4f414f43, 0x4f504f44, 0x4cf34cfc, + 0x4cf44cf1, 0x4cc04ccf, 0x4cc54cc7, 0x4cd34cdc, 0x4cd44cd1, 0x4c304c3f, 0x4c0c4c0f, 0x4c004c03, + 0x4c044c01, 0x4c104c1d, 0x4c714c73, 0x4c404c4d, 0x4c5c4c47, 0x4c514c53, 0x4df04c54, 0x4dc34dcc, + 0x4dd04dc4, 0x4d314d33, 0x4d0f4d34, 0x4d004d0d, 0x4d114d07, 0x4d704d14, 0x4d414d43, 0x43fc4d54, + 0x43f143f3, 0x43c043cf, 0x43d143c7, 0x4335433f, 0x4303430c, 0x43014300, 0x43044307, 0x431c431f, + 0x4310431d, 0x43714373, 0x4343434d, 0x43474340, 0x4354435c, 0x40f040ff, 0x40f540f7, 0x40cc40cf, + 0x40c040c3, 0x40c440c1, 0x40d040dc, 0x40d540d4, 0x4033403c, 0x40314030, 0x400f4034, 0x400d400c, + 0x40004003, 0x40074001, 0x40054004, 0x4013401c, 0x40114010, 0x407c4014, 0x40774070, 0x404d404c, + 0x40404043, 0x40444041, 0x405f4045, 0x4050405d, 0x40554057, 0x41f341fc, 0x41c041cf, 0x41df41c4, + 0x41d441d1, 0x41374130, 0x410c4134, 0x4100410d, 0x41044101, 0x41174110, 0x4173417d, 0x41754174, + 0x4143414d, 0x41534140, 0x41544151, 0x47c147f0, 0x47d047c4, 0x4731473c, 0x470d470f, 0x47014700, + 0x47134705, 0x47704710, 0x4741474c, 0x47504744, 0x44f144f3, 0x44cf44f4, 0x44c044cd, 0x44c544c7, + 0x44dc44df, 0x44d144d3, 0x443d443f, 0x44374430, 0x440c4435, 0x44004403, 0x44044401, 0x4410441d, + 0x44154411, 0x4473447c, 0x444d444f, 0x44454440, 0x4451445c, 0x45c045f0, 0x453345d0, 0x45344531, + 0x4500450f, 0x451c4507, 0x454c4570, 0x45404543, 0x5fff4541, 0x5ff75ffd, 0x5fc05ff5, 0x5fdd5fdf, + 0x5fd55fd7, 0x5f0c5f30, 0x5f015f03, 0x5f7f5f04, 0x5f775f7d, 0x5f405f75, 0x5f5d5f5f, 0x5f555f57, + 0x5cf45cf0, 0x5cc35ccc, 0x5cc45cc1, 0x5c315cc5, 0x5c0c5c34, 0x5c075c00, 0x5c1c5c05, 0x5c705c13, + 0x5c4d5c4f, 0x5c445c41, 0x5df75dfd, 0x5dcf5df5, 0x5ddd5dc4, 0x5dd55dd7, 0x5d0c5d30, 0x5d045d01, + 0x5d7f5d10, 0x5d775d7d, 0x5d405d75, 0x5d5d5d5f, 0x5d555d57, 0x53d053c4, 0x5333533c, 0x5303530f, + 0x53075300, 0x531c5305, 0x53115310, 0x53145317, 0x50f15370, 0x50cf50f4, 0x50c050cd, 0x50d150c7, + 0x503d50d4, 0x500c5030, 0x50005003, 0x50045001, 0x50155010, 0x5073507c, 0x50715070, 0x504d5074, + 0x50475040, 0x51cc51f0, 0x51c551c1, 0x51d051dc, 0x51315133, 0x510d5135, 0x51015100, 0x511f5107, + 0x5171511d, 0x5140514f, 0x51445141, 0x5153515c, 0x57ff5151, 0x57f757fd, 0x57df57f5, 0x57d757dd, + 0x570c57d5, 0x57015703, 0x577f5704, 0x5777577d, 0x57405775, 0x575d575f, 0x57555757, 0x54c354f0, + 0x54dc54c4, 0x543c54d0, 0x5400540f, 0x541c5405, 0x54145411, 0x5441544f, 0x55fd55ff, 0x55f555f7, + 0x55dd55df, 0x55d555d7, 0x5503550c, 0x557f5501, 0x5577557d, 0x55405575, 0x555d555f, 0x55555557 +); + +#enddecl(IQ1_GRID) + +#decl(IQ4_GRID) + +const kvalues_iq4nl = array( + -127, -104, -83, -65, -49, -35, -22, -10, 1, 13, 25, 38, 53, 69, 89, 113 +); + +#enddecl(IQ4_GRID) diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/embed_wgsl.py b/ggml/src/ggml-webgpu/wgsl-shaders/embed_wgsl.py index cc8def7f13..d9dfd7d6f4 100755 --- a/ggml/src/ggml-webgpu/wgsl-shaders/embed_wgsl.py +++ b/ggml/src/ggml-webgpu/wgsl-shaders/embed_wgsl.py @@ -27,6 +27,26 @@ def replace_placeholders(shader_text, replacements): return shader_text +def expand_includes(shader, input_dir): + """ + Replace #include "file" lines in the text with the contents of that file. + Searches for files relative to input_dir. + """ + include_pattern = re.compile(r'^\s*#include\s+"([^"]+)"\s*$', re.MULTILINE) + + def replacer(match): + fname = match.group(1) + file_path = os.path.join(input_dir, fname) + if not os.path.exists(file_path): + raise FileNotFoundError(f"Included file not found: {file_path}") + with open(file_path, "r", encoding="utf-8") as f: + included_code = f.read() + # Recursively expand includes inside the included file + return expand_includes(included_code, input_dir) + + return include_pattern.sub(replacer, shader) + + def write_shader(shader_name, shader_code, output_dir, outfile): if output_dir: wgsl_filename = os.path.join(output_dir, f"{shader_name}.wgsl") @@ -35,8 +55,9 @@ def write_shader(shader_name, shader_code, output_dir, outfile): outfile.write(f'const char* wgsl_{shader_name} = R"({shader_code})";\n\n') -def generate_variants(shader_path, output_dir, outfile): - shader_base_name = shader_path.split("/")[-1].split(".")[0] +def generate_variants(fname, input_dir, output_dir, outfile): + shader_path = os.path.join(input_dir, fname) + shader_base_name = fname.split(".")[0] with open(shader_path, "r", encoding="utf-8") as f: text = f.read() @@ -46,11 +67,21 @@ def generate_variants(shader_path, output_dir, outfile): except ValueError: write_shader(shader_base_name, text, output_dir, outfile) else: - decls_map = parse_decls(extract_block(text, "DECLS")) - shader_template = extract_block(text, "SHADER") + try: + decls_map = parse_decls(extract_block(text, "DECLS")) + except ValueError: + decls_map = {} + with open(os.path.join(input_dir, "common_decls.tmpl"), "r", encoding="utf-8") as f: + common_decls = f.read() + decls_map.update(parse_decls(common_decls)) + + shader_template = extract_block(text, "SHADER") for variant in variants: - decls = variant["DECLS"] + if "DECLS" in variant: + decls = variant["DECLS"] + else: + decls = [] decls_code = "" for key in decls: if key not in decls_map: @@ -59,8 +90,16 @@ def generate_variants(shader_path, output_dir, outfile): shader_variant = replace_placeholders(shader_template, variant["REPLS"]) final_shader = re.sub(r'\bDECLS\b', decls_code, shader_variant) + final_shader = expand_includes(final_shader, input_dir) - output_name = f"{shader_base_name}_" + "_".join([variant["REPLS"]["SRC0_TYPE"], variant["REPLS"]["SRC1_TYPE"]]) + if "SRC0_TYPE" in variant["REPLS"] and "SRC1_TYPE" in variant["REPLS"]: + output_name = f"{shader_base_name}_" + "_".join([variant["REPLS"]["SRC0_TYPE"], variant["REPLS"]["SRC1_TYPE"]]) + elif "TYPE_SUFFIX" in variant["REPLS"]: + output_name = f"{shader_base_name}_" + variant["REPLS"]["TYPE_SUFFIX"] + elif "TYPE" in variant["REPLS"]: + output_name = f"{shader_base_name}_" + variant["REPLS"]["TYPE"] + else: + output_name = shader_base_name write_shader(output_name, final_shader, output_dir, outfile) @@ -78,7 +117,7 @@ def main(): out.write("// Auto-generated shader embedding\n\n") for fname in sorted(os.listdir(args.input_dir)): if fname.endswith(".wgsl"): - generate_variants(os.path.join(args.input_dir, fname), args.output_dir, out) + generate_variants(fname, args.input_dir, args.output_dir, out) if __name__ == "__main__": diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/get_rows.tmpl.wgsl b/ggml/src/ggml-webgpu/wgsl-shaders/get_rows.tmpl.wgsl new file mode 100644 index 0000000000..e3fe311b26 --- /dev/null +++ b/ggml/src/ggml-webgpu/wgsl-shaders/get_rows.tmpl.wgsl @@ -0,0 +1,874 @@ +#define(VARIANTS) + +[ + { + "REPLS": { + "TYPE" : "vec4", + "TYPE_SUFFIX": "f32_vec", + "DST_TYPE": "vec4", + "BLOCK_SIZE": 4 + }, + "DECLS": ["F32_VEC"] + }, + { + "REPLS": { + "TYPE" : "f32", + "DST_TYPE": "f32", + "BLOCK_SIZE": 1 + }, + "DECLS": ["F32"] + }, + { + "REPLS": { + "TYPE" : "f16", + "DST_TYPE": "f32", + "BLOCK_SIZE": 1 + }, + "DECLS": ["F16"] + }, + { + "REPLS": { + "TYPE" : "i32", + "DST_TYPE": "i32", + "BLOCK_SIZE": 1 + }, + "DECLS": ["I32"] + }, + { + "REPLS": { + "TYPE" : "q4_0", + "DST_TYPE": "f32", + "BLOCK_SIZE": 32 + }, + "DECLS": ["BYTE_HELPERS", "Q4_0_T", "Q4_0"] + }, + { + "REPLS": { + "TYPE" : "q4_1", + "DST_TYPE": "f32", + "BLOCK_SIZE": 32 + }, + "DECLS": ["BYTE_HELPERS", "Q4_1_T", "Q4_1"] + }, + { + "REPLS": { + "TYPE" : "q5_0", + "DST_TYPE": "f32", + "BLOCK_SIZE": 32 + }, + "DECLS": ["BYTE_HELPERS", "Q5_0_T", "Q5_0"] + }, + { + "REPLS": { + "TYPE" : "q5_1", + "DST_TYPE": "f32", + "BLOCK_SIZE": 32 + }, + "DECLS": ["BYTE_HELPERS", "Q5_1_T", "Q5_1"] + }, + { + "REPLS": { + "TYPE" : "q8_0", + "DST_TYPE": "f32", + "BLOCK_SIZE": 32 + }, + "DECLS": ["BYTE_HELPERS", "Q8_0_T", "Q8_0"] + }, + { + "REPLS": { + "TYPE" : "q2_k", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256 + }, + "DECLS": ["BYTE_HELPERS", "Q2_K_T", "Q2_K"] + }, + { + "REPLS": { + "TYPE" : "q3_k", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256 + }, + "DECLS": ["BYTE_HELPERS", "Q3_K_T", "Q3_K"] + }, + { + "REPLS": { + "TYPE" : "q4_k", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256 + }, + "DECLS": ["Q45_K_SCALE_MIN", "BYTE_HELPERS", "Q4_K_T", "Q4_K"] + }, + { + "REPLS": { + "TYPE" : "q5_k", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256 + }, + "DECLS": ["Q45_K_SCALE_MIN", "BYTE_HELPERS", "Q5_K_T", "Q5_K"] + }, + { + "REPLS": { + "TYPE" : "q6_k", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256 + }, + "DECLS": ["BYTE_HELPERS", "Q6_K_T", "Q6_K"] + }, + { + "REPLS": { + "TYPE" : "iq2_xxs", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256 + }, + "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ2_XXS_GRID", "IQ2_XXS_T", "IQ2_XXS"] + }, + { + "REPLS": { + "TYPE" : "iq2_xs", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256 + }, + "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ2_XS_GRID", "IQ2_XS_T", "IQ2_XS"] + }, + { + "REPLS": { + "TYPE": "iq2_s", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256 + }, + "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ2_S_GRID", "IQ2_S_T", "IQ2_S"] + }, + { + "REPLS": { + "TYPE": "iq3_xxs", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256 + }, + "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ3_XSS_GRID", "IQ3_XSS_T", "IQ3_XSS"] + }, + { + "REPLS": { + "TYPE": "iq3_s", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256 + }, + "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ3_S_GRID", "IQ3_S_T", "IQ3_S"] + }, + { + "REPLS": { + "TYPE": "iq1_s", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256 + }, + "DECLS": ["BYTE_HELPERS", "IQ1_GRID", "IQ1_S_T", "IQ1_S"] + }, + { + "REPLS": { + "TYPE": "iq1_m", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256 + }, + "DECLS": ["BYTE_HELPERS", "IQ1_GRID", "IQ1_M_T", "IQ1_M"] + }, + { + "REPLS": { + "TYPE": "iq4_nl", + "DST_TYPE": "f32", + "BLOCK_SIZE": 32, + }, + "DECLS": ["BYTE_HELPERS", "IQ4_GRID", "IQ4_NL_T", "IQ4_NL"] + }, + { + "REPLS": { + "TYPE": "iq4_xs", + "DST_TYPE": "f32", + "BLOCK_SIZE": 256, + }, + "DECLS": ["BYTE_HELPERS", "IQ4_GRID", "IQ4_XS_T", "IQ4_XS"] + } +] + +#end(VARIANTS) + +#define(DECLS) + +#decl(F32_VEC) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + dst[(dst_base / 4) + offset] = src[(src_base / 4) + offset]; +} +#enddecl(F32_VEC) + +#decl(F32) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + dst[dst_base + offset] = src[src_base + offset]; +} +#enddecl(F32) + +#decl(F16) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + dst[dst_base + offset] = f32(src[src_base + offset]); +} +#enddecl(F16) + +#decl(I32) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + dst[dst_base + offset] = src[src_base + offset]; +} +#enddecl(I32) + +#decl(Q4_0) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block_q4_0 = src[src_base + offset]; + let d = f32(block_q4_0.d); + for (var j: u32 = 0; j < 4; j++) { + let q_packed = bitcast(vec2(block_q4_0.qs[2 * j], block_q4_0.qs[2 * j + 1])); + for (var k: u32 = 0; k < 4; k++) { + let q_byte = get_byte(q_packed, k); + let q_hi = (f32((q_byte >> 4) & 0xF) - 8.0f) * d; + let q_lo = (f32(q_byte & 0xF) - 8.0f) * d; + let dst_offset = dst_base + offset * 32 + j * 4 + k; + dst[dst_offset] = q_lo; + dst[dst_offset + 16] = q_hi; + } + } +} +#enddecl(Q4_0) + +#decl(Q4_1) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block_q4_1 = src[src_base + offset]; + let d = f32(block_q4_1.d); + let m = f32(block_q4_1.m); + for (var j: u32 = 0; j < 4; j++) { + let q_packed = block_q4_1.qs[j]; + for (var k: u32 = 0; k < 4; k++) { + let q_byte = get_byte(q_packed, k); + let q_hi = f32((q_byte >> 4) & 0xF) * d + m; + let q_lo = f32(q_byte & 0xF) * d + m; + let dst_offset = dst_base + offset * 32 + j * 4 + k; + dst[dst_offset] = q_lo; + dst[dst_offset + 16] = q_hi; + } + } +} +#enddecl(Q4_1) + +#decl(Q5_0) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block_q5_0 = src[src_base + offset]; + let d = f32(block_q5_0.d); + let qh_packed = bitcast(vec2(block_q5_0.qh[0], block_q5_0.qh[1])); + for (var j: u32 = 0; j < 4; j++) { + let q_packed = bitcast(vec2(block_q5_0.qs[2 * j], block_q5_0.qs[2 * j + 1])); + for (var k: u32 = 0; k < 4; k++) { + let q_byte = get_byte(q_packed, k); + let qh_hi = (qh_packed >> (j * 4 + k + 12)) & 0x10; + let q_hi = (f32(((q_byte >> 4) & 0xF) | qh_hi) - 16.0) * d; + let qh_lo = ((qh_packed >> (j * 4 + k)) << 4) & 0x10; + let q_lo = (f32((q_byte & 0xF) | qh_lo) - 16.0) * d; + let dst_offset = dst_base + offset * 32 + j * 4 + k; + dst[dst_offset] = q_lo; + dst[dst_offset + 16] = q_hi; + } + } +} + +#enddecl(Q5_0) + +#decl(Q5_1) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block_q5_1 = src[src_base + offset]; + let d = f32(block_q5_1.d); + let m = f32(block_q5_1.m); + for (var j: u32 = 0; j < 4; j++) { + let q_packed = block_q5_1.qs[j]; + for (var k: u32 = 0; k < 4; k++) { + let q_byte = get_byte(q_packed, k); + let qh_hi = (block_q5_1.qh >> (j * 4 + k + 12)) & 0x10; + let q_hi = f32(((q_byte >> 4) & 0xF) | qh_hi) * d + m; + let qh_lo = ((block_q5_1.qh >> (j * 4 + k)) << 4) & 0x10; + let q_lo = f32((q_byte & 0xF) | qh_lo) * d + m; + let dst_offset = dst_base + offset * 32 + j * 4 + k; + dst[dst_offset] = q_lo; + dst[dst_offset + 16] = q_hi; + } + } +} +#enddecl(Q5_1) + +#decl(Q8_0) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block_q8_0 = src[src_base + offset]; + let d = f32(block_q8_0.d); + for (var j: u32 = 0; j < 8; j++) { + let q_packed = bitcast(vec2(block_q8_0.qs[2 * j], block_q8_0.qs[2 * j + 1])); + for (var k: u32 = 0; k < 4; k++) { + let q_byte = get_byte_i32(q_packed, k); + let q_val = f32(q_byte) * d; + let dst_offset = dst_base + offset * 32 + j * 4 + k; + dst[dst_offset] = q_val; + } + } +} +#enddecl(Q8_0) + +#decl(Q2_K) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + let m = f32(block.dmin); + var dst_i = dst_base + offset * 256; + var is: u32 = 0; + // 2 halves of the block (128 elements each) + for (var q_b_idx: u32 = 0; q_b_idx < 64; q_b_idx += 32) { + // 4 groups (each group has 2 blocks of 16 elements) + for (var shift: u32 = 0; shift < 8; shift += 2) { + // 2 blocks + for (var k: u32 = 0; k < 32; k += 16) { + let sc = get_byte(block.scales[is / 4], is % 4); + is++; + let dl = d * f32(sc & 0xF); + let ml = m * f32(sc >> 4); + for (var l: u32 = 0u; l < 16; l++) { + let q_idx = q_b_idx + k + l; + let q_byte = get_byte(block.qs[q_idx / 4], q_idx % 4); + let qs_val = (q_byte >> shift) & 3; + dst[dst_i] = (f32(qs_val) * dl - ml); + dst_i++; + } + } + } + } +} +#enddecl(Q2_K) + +#decl(Q3_K) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + + // extract 6-bit scales, which consist of 4-bits from first 8 bytes of scale, + // and 2-bits from the last 4 bytes + let kmask1: u32 = 0x03030303; + let kmask2: u32 = 0x0f0f0f0f; + var scale_vals: array; + for (var i: u32 = 0; i < 4; i++) { + scale_vals[i] = bitcast(vec2(block.scales[2 * i], block.scales[2 * i + 1])); + } + var tmp: u32 = scale_vals[2]; + scale_vals[2] = ((scale_vals[0] >> 4) & kmask2) | (((tmp >> 4) & kmask1) << 4); + scale_vals[3] = ((scale_vals[1] >> 4) & kmask2) | (((tmp >> 6) & kmask1) << 4); + scale_vals[0] = (scale_vals[0] & kmask2) | ((tmp & kmask1) << 4); + scale_vals[1] = (scale_vals[1] & kmask2) | (((tmp >> 2) & kmask1) << 4); + + // convert arrays of f16 -> u32 + var hmask_vals: array; + for (var i: u32 = 0; i < 8; i++) { + hmask_vals[i] = bitcast(vec2(block.hmask[2 * i], block.hmask[2 * i + 1])); + } + var qs_vals: array; + for (var i: u32 = 0; i < 16; i++) { + qs_vals[i] = bitcast(vec2(block.qs[2 * i], block.qs[2 * i + 1])); + } + + var dst_i = dst_base + offset * 256; + var is: u32 = 0; + var m: u32 = 1; + // 2 halves of the block (128 elements each) + for (var q_b_idx: u32 = 0; q_b_idx < 64; q_b_idx += 32) { + // 4 groups (each group has 2 blocks of 16 elements) + for (var shift: u32 = 0; shift < 8; shift += 2) { + // 2 blocks + for (var k: u32 = 0; k < 32; k += 16) { + let sc = get_byte(scale_vals[is / 4], is % 4); + is++; + let dl = d * (f32(sc) - 32.0); + for (var l: u32 = 0u; l < 16u; l++) { + let q_idx = q_b_idx + k + l; + let hm_idx = k + l; + let q_byte = get_byte(qs_vals[q_idx / 4], q_idx % 4); + let hmask_byte = get_byte(hmask_vals[hm_idx / 4], hm_idx % 4); + let hm = select(4.0, 0.0, (hmask_byte & m) != 0); + let qs_val = (q_byte >> shift) & 3; + dst[dst_i] = (f32(qs_val) - hm) * dl; + dst_i++; + } + } + m <<= 1; + } + } +} +#enddecl(Q3_K) + +#decl(Q4_K) +// 8 blocks of 32 elements each +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + let m = f32(block.dmin); + var dst_i = dst_base + offset * 256; + var is: u32 = 0; + // 2 blocks each iteration + for (var q_b_idx: u32 = 0; q_b_idx < 128; q_b_idx += 32) { + for (var shift: u32 = 0; shift < 8; shift += 4) { + let scale_min = get_scale_min(is, block.scales); + is++; + let dl = d * scale_min.x; + let ml = m * scale_min.y; + for (var l: u32 = 0; l < 32; l++) { + let q_idx = q_b_idx + l; + let q_byte = get_byte(block.qs[q_idx / 4], q_idx % 4); + let qs_val = (q_byte >> shift) & 0xF; + dst[dst_i] = (f32(qs_val) * dl - ml); + dst_i++; + } + } + } +} +#enddecl(Q4_K) + +#decl(Q5_K) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + let m = f32(block.dmin); + var dst_i = dst_base + offset * 256; + var is: u32 = 0; + var u: u32 = 1; + // 2 blocks each iteration + for (var q_b_idx: u32 = 0; q_b_idx < 128; q_b_idx += 32) { + for (var shift: u32 = 0; shift < 8; shift += 4) { + let scale_min = get_scale_min(is, block.scales); + is++; + let dl = d * scale_min.x; + let ml = m * scale_min.y; + for (var l: u32 = 0; l < 32; l++) { + let q_idx = q_b_idx + l; + let q_byte = get_byte(block.qs[q_idx / 4], q_idx % 4); + let qh_byte = get_byte(block.qh[l / 4], l % 4); + let qs_val = (q_byte >> shift) & 0xF; + let qh_val = select(0.0, 16.0, (qh_byte & u) != 0); + dst[dst_i] = (f32(qs_val) + qh_val) * dl - ml; + dst_i++; + } + u <<= 1; + } + } +} +#enddecl(Q5_K) + +#decl(Q6_K) +// 16 blocks of 16 elements each +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + + // convert arrays of f16 -> u32 + var ql_vals: array; + for (var i: u32 = 0; i < 32; i++) { + ql_vals[i] = bitcast(vec2(block.ql[2 * i], block.ql[2 * i + 1])); + } + var qh_vals: array; + for (var i: u32 = 0; i < 16; i++) { + qh_vals[i] = bitcast(vec2(block.qh[2 * i], block.qh[2 * i + 1])); + } + var scale_vals: array; + for (var i: u32 = 0; i < 4; i++) { + scale_vals[i] = bitcast(vec2(block.scales[2 * i], block.scales[2 * i + 1])); + } + + var dst_i = dst_base + offset * 256; + var qh_b_idx: u32 = 0; + var sc_b_idx: u32 = 0; + for (var ql_b_idx: u32 = 0; ql_b_idx < 128; ql_b_idx += 64) { + for (var l: u32 = 0; l < 32; l++) { + let ql13_b = get_byte(ql_vals[(ql_b_idx + l) / 4], (ql_b_idx + l) % 4); + let ql24_b = get_byte(ql_vals[(ql_b_idx + l + 32) / 4], (ql_b_idx + l + 32) % 4); + let qh_b = get_byte(qh_vals[(qh_b_idx + l) / 4], (qh_b_idx + l) % 4); + + let q1 = f32((ql13_b & 0xF) | ((qh_b & 3) << 4)) - 32.0; + let q2 = f32((ql24_b & 0xF) | (((qh_b >> 2) & 3) << 4)) - 32.0; + let q3 = f32((ql13_b >> 4) | (((qh_b >> 4) & 3) << 4)) - 32.0; + let q4 = f32((ql24_b >> 4) | (((qh_b >> 6) & 3) << 4)) - 32.0; + + let is = l/16; + let is1 = sc_b_idx + is; + let sc1 = get_byte_i32(scale_vals[is1 / 4], is1 % 4); + let is2 = sc_b_idx + is + 2; + let sc2 = get_byte_i32(scale_vals[is2 / 4], is2 % 4); + let is3 = sc_b_idx + is + 4; + let sc3 = get_byte_i32(scale_vals[is3 / 4], is3 % 4); + let is4 = sc_b_idx + is + 6; + let sc4 = get_byte_i32(scale_vals[is4 / 4], is4 % 4); + + dst[dst_i + l] = (q1 * f32(sc1)) * d; + dst[dst_i + l + 32] = (q2 * f32(sc2)) * d; + dst[dst_i + l + 64] = (q3 * f32(sc3)) * d; + dst[dst_i + l + 96] = (q4 * f32(sc4)) * d; + } + dst_i += 128; + qh_b_idx += 32; + sc_b_idx += 8; + } +} + +#enddecl(Q6_K) + +#decl(IQ2_XXS) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + var dst_i = dst_base + offset * 256; + for (var ib: u32 = 0; ib < 32; ib += 4) { + let aux0 = bitcast(vec2(block.qs[ib], block.qs[ib + 1])); + let aux1 = bitcast(vec2(block.qs[ib + 2], block.qs[ib + 3])); + let db = d * (0.5 + f32(aux1 >> 28)) * 0.25; + for (var l: u32 = 0; l < 4; l++) { + let ig = get_byte(aux0, l) * 8; + let is = (aux1 >> (7 * l)) & 127; + let signs = get_byte(ksigns_iq2xs[is / 4], is % 4); + for (var j: u32 = 0; j < 8; j++) { + let g = get_byte(iq2xxs_grid[(ig + j) / 4], (ig + j) % 4); + let m = select(1.0, -1.0, (get_byte(kmask_iq2xs[j / 4], j % 4) & signs) != 0); + dst[dst_i] = db * f32(g) * m; + dst_i++; + } + } + } +} +#enddecl(IQ2_XXS) + +#decl(IQ2_XS) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + var dst_i = dst_base + offset * 256; + var scale_vals = array( + bitcast(vec2(block.scales[0], block.scales[1])), + bitcast(vec2(block.scales[2], block.scales[3])) + ); + for (var ib: u32 = 0; ib < 32; ib += 4) { + let s = get_byte(scale_vals[ib / 16], (ib % 16) / 4); + let db = array( + d * (0.5 + f32(s & 0xF)) * 0.25, + d * (0.5 + f32(s >> 4)) * 0.25 + ); + for (var l: u32 = 0; l < 4; l++) { + let qs_val = bitcast(vec2(block.qs[ib + l], 0.0)); + let ig = (qs_val & 511) * 8; + let is = qs_val >> 9; + let signs = get_byte(ksigns_iq2xs[is / 4], is % 4); + let dl = db[l/2]; + for (var j: u32 = 0; j < 8; j++) { + let g = get_byte(iq2xs_grid[(ig + j) / 4], (ig + j) % 4); + let m = select(1.0, -1.0, (get_byte(kmask_iq2xs[j / 4], j % 4) & signs) != 0); + dst[dst_i] = dl * f32(g) * m; + dst_i++; + } + } + } +} +#enddecl(IQ2_XS) + +#decl(IQ2_S) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + var dst_i = dst_base + offset * 256; + var qs_vals : array; + for (var i: u32 = 0; i < 16; i++) { + qs_vals[i] = bitcast(vec2(block.qs[i * 2], block.qs[i * 2 + 1])); + } + var qh_vals = array( + bitcast(vec2(block.qh[0], block.qh[1])), + bitcast(vec2(block.qh[2], block.qh[3])) + ); + var scale_vals = array( + bitcast(vec2(block.scales[0], block.scales[1])), + bitcast(vec2(block.scales[2], block.scales[3])) + ); + for (var ib: u32 = 0; ib < 8; ib ++) { + let s = get_byte(scale_vals[ib / 4], ib % 4); + let db = array( + d * (0.5 + f32(s & 0xF)) * 0.25, + d * (0.5 + f32(s >> 4)) * 0.25 + ); + let qs_w = qs_vals[ib]; + for (var l: u32 = 0; l < 4; l++) { + let qh_b = (get_byte(qh_vals[ib / 4], ib % 4) << (8 - 2 * l)) & 0x300; + let ig = (get_byte(qs_w, l) | qh_b) * 8; + let signs = get_byte(qs_vals[ib + 8], l); + let dl = db[l/2]; + for (var j: u32 = 0; j < 8; j++) { + let g = get_byte(iq2s_grid[(ig + j) / 4], (ig + j) % 4); + let m = select(1.0, -1.0, (get_byte(kmask_iq2xs[j / 4], j % 4) & signs) != 0); + dst[dst_i] = dl * f32(g) * m; + dst_i++; + } + } + } +} + +#enddecl(IQ2_S) + +#decl(IQ3_XSS) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + var dst_i = dst_base + offset * 256; + for (var ib: u32 = 0; ib < 16; ib += 2) { + let sc_sign = bitcast(vec2(block.qs[ib + 32], block.qs[ib + 33])); + let db = d * (0.5 + f32(sc_sign >> 28)) * 0.5; + for (var l: u32 = 0; l < 4; l++) { + let is = (sc_sign >> (7 * l)) & 127; + let signs = get_byte(ksigns_iq2xs[is / 4], is % 4); + let ig_val = bitcast(vec2(block.qs[ib * 2 + l], 0.0)); + let ig1 = get_byte(ig_val, 0); + let ig2 = get_byte(ig_val, 1); + for (var j: u32 = 0; j < 4; j++) { + let g1 = get_byte(iq3xxs_grid[ig1], j); + let g2 = get_byte(iq3xxs_grid[ig2], j); + let m1 = select(1.0, -1.0, (get_byte(kmask_iq2xs[0], j) & signs) != 0); + let m2 = select(1.0, -1.0, (get_byte(kmask_iq2xs[1], j) & signs) != 0); + dst[dst_i] = db * f32(g1) * m1; + dst[dst_i + 4] = db * f32(g2) * m2; + dst_i++; + } + dst_i += 4; + } + } +} +#enddecl(IQ3_XSS) + +#decl(IQ3_S) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + var dst_i = dst_base + offset * 256; + var qh_vals = array( + bitcast(vec2(block.qh[0], block.qh[1])), + bitcast(vec2(block.qh[2], block.qh[3])) + ); + var sign_vals: array; + for (var i: u32 = 0; i < 8; i++) { + sign_vals[i] = bitcast(vec2(block.signs[i * 2], block.signs[i * 2 + 1])); + } + var scale_vals = bitcast(vec2(block.scales[0], block.scales[1])); + for (var ib: u32 = 0; ib < 4; ib++) { + let s = get_byte(scale_vals, ib); + let db = array( + d * (1.0 + 2.0 * f32(s & 0xF)), + d * (1.0 + 2.0 * f32(s >> 4)) + ); + for (var k: u32 = 0; k < 2; k++) { + let dl = db[k]; + let qh_byte = get_byte(qh_vals[ib / 2], (ib % 2) * 2 + k); + let sign_w = sign_vals[ib * 2 + k]; + for (var l: u32 = 0; l < 4; l++) { + let signs = get_byte(sign_w, l); + let ig_val = bitcast(vec2(block.qs[ib * 8 + k * 4 + l], 0.0)); + let ig1 = get_byte(ig_val, 0) | ((qh_byte << ((8 - (2 * l)))) & 256); + let ig2 = get_byte(ig_val, 1) | ((qh_byte << ((7 - (2 * l)))) & 256); + for (var j: u32 = 0; j < 4; j++) { + let g1 = get_byte(iq3s_grid[ig1], j); + let g2 = get_byte(iq3s_grid[ig2], j); + let m1 = select(1.0, -1.0, (get_byte(kmask_iq2xs[0], j) & signs) != 0); + let m2 = select(1.0, -1.0, (get_byte(kmask_iq2xs[1], j) & signs) != 0); + dst[dst_i] = dl * f32(g1) * m1; + dst[dst_i + 4] = dl * f32(g2) * m2; + dst_i++; + } + dst_i += 4; + } + } + } +} +#enddecl(IQ3_S) + +#decl(IQ1_S) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + var dst_i = dst_base + offset * 256; + for (var ib: u32 = 0; ib < 8; ib++) { + let qh = bitcast(vec2(block.qh[ib], 0.0)); + let dl = d * (2 * f32((qh >> 12) & 7) + 1); + let delta = select(IQ1_DELTA, -IQ1_DELTA, (qh & 0x8000) != 0); + let qs_w = bitcast(vec2(block.qs[ib * 2], block.qs[ib * 2 + 1])); + for (var l: u32 = 0; l < 4; l++) { + let ig = (get_byte(qs_w, l) | (((qh >> (3 * l)) & 7) << 8)) * 8; + for (var j: u32 = 0; j < 8; j++) { + let gw = iq1_grid[(ig + j) / 16]; + let g = (gw >> (((ig + j) % 16) * 2)) & 3; + let gs = bitcast(g << 30) >> 30; + dst[dst_i] = dl * (f32(gs) + delta); + dst_i++; + } + } + } +} + +#enddecl(IQ1_S) + +#decl(IQ1_M) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + + let scale = ((block.scales[0] >> 12) & 0xF) | ((block.scales[0] >> 24) & 0x00F0) | ((block.scales[1] >> 4) & 0x0F00) | ((block.scales[1] >> 16) & 0xF000); + let d = f32(bitcast>(scale).x); + var dst_i = dst_base + offset * 256; + for (var ib: u32 = 0; ib < 8; ib++) { + let sw = (block.scales[ib / 4] >> (16 * ((ib / 2) % 2))) & 0xFFFF; + let s1 : u32 = (sw >> (6 * (ib % 2))) & 0x7; + let s2 : u32 = (sw >> (6 * (ib % 2) + 3)) & 0x7; + var dl = array( + d * f32(2 * s1 + 1), + d * f32(2 * s2 + 1) + ); + + let qh = block.qh[ib / 2] >> (16 * (ib % 2)); + var idx = array( + get_byte(block.qs[ib], 0) | ((qh << 8) & 0x700), + get_byte(block.qs[ib], 1) | ((qh << 4) & 0x700), + get_byte(block.qs[ib], 2) | ((qh) & 0x700), + get_byte(block.qs[ib], 3) | ((qh >> 4) & 0x700) + ); + var delta = array( + select(IQ1_DELTA, -IQ1_DELTA, (qh & 0x08) != 0), + select(IQ1_DELTA, -IQ1_DELTA, (qh & 0x80) != 0), + select(IQ1_DELTA, -IQ1_DELTA, ((qh >> 8) & 0x08) != 0), + select(IQ1_DELTA, -IQ1_DELTA, ((qh >> 8) & 0x80) != 0) + ); + for (var l: u32 = 0; l < 4; l++) { + let ig = idx[l] * 8; + for (var j: u32 = 0; j < 8; j++) { + let gw = iq1_grid[(ig + j) / 16]; + let g = (gw >> (((ig + j) % 16) * 2)) & 3; + let gs = bitcast(g << 30) >> 30; + dst[dst_i] = dl[l/2] * (f32(gs) + delta[l]); + dst_i++; + } + } + } +} + +#enddecl(IQ1_M) + +#decl(IQ4_NL) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + var dst_i = dst_base + offset * 32; + var qs: array; + for (var i: u32 = 0; i < 4; i++) { + qs[i] = bitcast(vec2(block.qs[i * 2], block.qs[i * 2 + 1])); + } + for (var j: u32 = 0; j < 16; j++) { + let qsb = get_byte(qs[j / 4], j % 4); + dst[dst_i] = d * f32(kvalues_iq4nl[qsb & 0xF]); + dst[dst_i + 16] = d * f32(kvalues_iq4nl[qsb >> 4]); + dst_i++; + } +} +#enddecl(IQ4_NL) + +#decl(IQ4_XS) +fn copy_elements(src_base: u32, dst_base: u32, offset: u32) { + let block = src[src_base + offset]; + let d = f32(block.d); + let scales_h = bitcast(vec2(block.scales_h, 0.0)); + var dst_i = dst_base + offset * 256; + for (var ib: u32 = 0; ib < 8; ib++) { + let ls = ((get_byte(block.scales_l, ib / 2) >> (4 * (ib % 2))) & 0xF) | (((scales_h >> (2 * ib)) & 3) << 4); + let dl = d * (f32(ls) - 32.0); + for (var j: u32 = 0; j < 16; j++) { + let iqs = ib * 16 + j; + let qsb = get_byte(block.qs[iqs / 4], iqs % 4); + dst[dst_i] = dl * f32(kvalues_iq4nl[qsb & 0xF]); + dst[dst_i + 16] = dl * f32(kvalues_iq4nl[qsb >> 4]); + dst_i++; + } + dst_i += 16; + } +} +#enddecl(IQ4_XS) + +#end(DECLS) + +#define(SHADER) + +enable f16; + +DECLS + +@group(0) @binding(0) +var src: array<{{TYPE}}>; + +@group(0) @binding(1) +var idx: array; + +@group(0) @binding(2) +var dst: array<{{DST_TYPE}}>; + +struct Params { + offset_src: u32, // in elements + offset_idx: u32, // in elements + offset_dst: u32, // in elements + + // Strides (in elements) + stride_src1: u32, + stride_src2: u32, + stride_src3: u32, + + stride_idx0: u32, + stride_idx1: u32, + stride_idx2: u32, + + stride_dst1: u32, + stride_dst2: u32, + stride_dst3: u32, + + // Shape of dst + ne0: u32, + n_rows: u32, + ne2: u32, + ne3: u32, + + // Shape of idx + idx1: u32, + idx2: u32, +}; + +@group(0) @binding(3) +var params: Params; + +override wg_size: u32; +@compute @workgroup_size(wg_size) +fn main(@builtin(global_invocation_id) gid: vec3) { + if (gid.x >= params.n_rows * params.ne2 * params.ne3) { + return; + } + var i = gid.x; + let i_dst3 = i / (params.ne2 * params.n_rows); + + i = i % (params.ne2 * params.n_rows); + let i_dst2 = i / params.n_rows; + let i_dst1 = i % params.n_rows; + + let i_idx2 = i_dst3 % params.idx2; + let i_idx1 = i_dst2 % params.idx1; + let i_idx0 = i_dst1; + + let i_idx = params.offset_idx + i_idx0 * params.stride_idx0 + i_idx1 * params.stride_idx1 + i_idx2 * params.stride_idx2; + + let idx_val = u32(idx[i_idx]); + + let i_src_row = params.offset_src + idx_val * params.stride_src1 + i_dst2 * params.stride_src2 + i_dst3 * params.stride_src3; + let i_dst_row = params.offset_dst + i_dst1 * params.stride_dst1 + i_dst2 * params.stride_dst2 + i_dst3 * params.stride_dst3; + + for (var i: u32 = 0; i < params.ne0/{{BLOCK_SIZE}}; i++) { + copy_elements(i_src_row, i_dst_row, i); + } +} + +#end(SHADER) diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/mul.tmpl.wgsl b/ggml/src/ggml-webgpu/wgsl-shaders/mul.tmpl.wgsl new file mode 100644 index 0000000000..12506e1420 --- /dev/null +++ b/ggml/src/ggml-webgpu/wgsl-shaders/mul.tmpl.wgsl @@ -0,0 +1,44 @@ +#define(VARIANTS) + +[ + { + "REPLS": { + "TYPE" : "f32", + } + }, + { + "REPLS": { + "TYPE" : "f16", + } + } +] + +#end(VARIANTS) + +#define(SHADER) + +enable f16; + +#include "binary_head.tmpl" + +@group(0) @binding(0) +var src0: array<{{TYPE}}>; + +@group(0) @binding(1) +var src1: array<{{TYPE}}>; + +@group(0) @binding(2) +var dst: array<{{TYPE}}>; + +@group(0) @binding(3) +var params: Params; + +override wg_size: u32; +@compute @workgroup_size(wg_size) +fn main(@builtin(global_invocation_id) gid: vec3) { + if (gid.x < params.ne) { + dst[params.offset_dst + gid.x] = src0[params.offset_src0 + gid.x] * src1[params.offset_src1 + src1_index(gid.x)]; + } +} + +#end(SHADER) diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/mul_in_place.tmpl.wgsl b/ggml/src/ggml-webgpu/wgsl-shaders/mul_in_place.tmpl.wgsl new file mode 100644 index 0000000000..e467e59edb --- /dev/null +++ b/ggml/src/ggml-webgpu/wgsl-shaders/mul_in_place.tmpl.wgsl @@ -0,0 +1,41 @@ +#define(VARIANTS) + +[ + { + "REPLS": { + "TYPE" : "f32", + } + }, + { + "REPLS": { + "TYPE" : "f16", + } + } +] + +#end(VARIANTS) + +#define(SHADER) + +enable f16; + +#include "binary_head.tmpl" + +@group(0) @binding(0) +var src0: array<{{TYPE}}>; + +@group(0) @binding(1) +var src1: array<{{TYPE}}>; + +@group(0) @binding(2) +var params: Params; + +override wg_size: u32; +@compute @workgroup_size(wg_size) +fn main(@builtin(global_invocation_id) gid: vec3) { + if (gid.x < params.ne) { + src0[params.offset_dst + gid.x] = src0[params.offset_src0 + gid.x] * src1[params.offset_src1 + src1_index(gid.x)]; + } +} + +#end(SHADER) diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat.tmpl.wgsl b/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat.tmpl.wgsl index 79465c298d..25e2185de8 100644 --- a/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat.tmpl.wgsl +++ b/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat.tmpl.wgsl @@ -31,7 +31,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 32 }, - "DECLS": ["BYTE_HELPERS", "Q4_0"] + "DECLS": ["BYTE_HELPERS", "Q4_0_T", "Q4_0"] }, { "REPLS": { @@ -39,7 +39,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 32 }, - "DECLS": ["BYTE_HELPERS", "Q4_1"] + "DECLS": ["BYTE_HELPERS", "Q4_1_T", "Q4_1"] }, { "REPLS": { @@ -47,7 +47,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 32 }, - "DECLS": ["BYTE_HELPERS", "Q5_0"] + "DECLS": ["BYTE_HELPERS", "Q5_0_T", "Q5_0"] }, { "REPLS": { @@ -55,7 +55,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 32 }, - "DECLS": ["BYTE_HELPERS", "Q5_1"] + "DECLS": ["BYTE_HELPERS", "Q5_1_T", "Q5_1"] }, { "REPLS": { @@ -63,7 +63,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 32 }, - "DECLS": ["BYTE_HELPERS", "Q8_0"] + "DECLS": ["BYTE_HELPERS", "Q8_0_T", "Q8_0"] }, { "REPLS": { @@ -71,7 +71,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256 }, - "DECLS": ["BYTE_HELPERS", "Q2_K"] + "DECLS": ["BYTE_HELPERS", "Q2_K_T", "Q2_K"] }, { "REPLS": { @@ -79,7 +79,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256 }, - "DECLS": ["BYTE_HELPERS", "Q3_K"] + "DECLS": ["BYTE_HELPERS", "Q3_K_T", "Q3_K"] }, { "REPLS": { @@ -87,7 +87,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256 }, - "DECLS": ["Q45_K_SCALE_MIN", "BYTE_HELPERS", "Q4_K"] + "DECLS": ["Q45_K_SCALE_MIN", "BYTE_HELPERS", "Q4_K_T", "Q4_K"] }, { "REPLS": { @@ -95,7 +95,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256 }, - "DECLS": ["Q45_K_SCALE_MIN", "BYTE_HELPERS", "Q5_K"] + "DECLS": ["Q45_K_SCALE_MIN", "BYTE_HELPERS", "Q5_K_T", "Q5_K"] }, { "REPLS": { @@ -103,7 +103,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256 }, - "DECLS": ["BYTE_HELPERS", "Q6_K"] + "DECLS": ["BYTE_HELPERS", "Q6_K_T", "Q6_K"] }, { "REPLS": { @@ -111,7 +111,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256 }, - "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ2_XXS"] + "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ2_XXS_GRID", "IQ2_XXS_T", "IQ2_XXS"] }, { "REPLS": { @@ -119,7 +119,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256 }, - "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ2_XS"] + "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ2_XS_GRID", "IQ2_XS_T", "IQ2_XS"] }, { "REPLS": { @@ -127,7 +127,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256 }, - "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ2_S"] + "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ2_S_GRID", "IQ2_S_T", "IQ2_S"] }, { "REPLS": { @@ -135,7 +135,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256 }, - "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ3_XSS"] + "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ3_XSS_GRID", "IQ3_XSS_T", "IQ3_XSS"] }, { "REPLS": { @@ -143,7 +143,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256 }, - "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ3_S"] + "DECLS": ["BYTE_HELPERS", "IQ23_TABLES", "IQ3_S_GRID", "IQ3_S_T", "IQ3_S"] }, { "REPLS": { @@ -151,7 +151,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256 }, - "DECLS": ["BYTE_HELPERS", "IQ1_TABLE","IQ1_S"] + "DECLS": ["BYTE_HELPERS", "IQ1_GRID", "IQ1_S_T", "IQ1_S"] }, { "REPLS": { @@ -159,7 +159,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256 }, - "DECLS": ["BYTE_HELPERS", "IQ1_TABLE","IQ1_M"] + "DECLS": ["BYTE_HELPERS", "IQ1_GRID", "IQ1_M_T", "IQ1_M"] }, { "REPLS": { @@ -167,7 +167,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 32, }, - "DECLS": ["BYTE_HELPERS", "IQ4_TABLE", "IQ4_NL"] + "DECLS": ["BYTE_HELPERS", "IQ4_GRID", "IQ4_NL_T", "IQ4_NL"] }, { "REPLS": { @@ -175,7 +175,7 @@ "SRC1_TYPE": "f32", "BLOCK_SIZE": 256, }, - "DECLS": ["BYTE_HELPERS", "IQ4_TABLE", "IQ4_XS"] + "DECLS": ["BYTE_HELPERS", "IQ4_GRID", "IQ4_XS_T", "IQ4_XS"] } ] @@ -183,18 +183,6 @@ #define(DECLS) -#decl(BYTE_HELPERS) - -fn get_byte(value: u32, index: u32) -> u32 { - return (value >> (index * 8)) & 0xFF; -} - -fn get_byte_i32(value: u32, index: u32) -> i32 { - return bitcast(((value >> (index * 8)) & 0xFF) << 24) >> 24; -} - -#enddecl(BYTE_HELPERS) - #decl(FLOAT) fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { return f32(src0[src0_idx_base + offset]) * f32(src1[src1_idx_base + offset]); @@ -202,11 +190,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(FLOAT) #decl(Q4_0) -struct q4_0 { - d: f16, - qs: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block_q4_0 = src0[src0_idx_base + offset]; let d = f32(block_q4_0.d); @@ -227,12 +210,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(Q4_0) #decl(Q4_1) -struct q4_1 { - d: f16, - m: f16, - qs: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block_q4_1 = src0[src0_idx_base + offset]; let d = f32(block_q4_1.d); @@ -254,12 +231,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(Q4_1) #decl(Q5_0) -struct q5_0 { - d: f16, - qh: array, - qs: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block_q5_0 = src0[src0_idx_base + offset]; let d = f32(block_q5_0.d); @@ -283,13 +254,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(Q5_0) #decl(Q5_1) -struct q5_1 { - d: f16, - m: f16, - qh: u32, - qs: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block_q5_1 = src0[src0_idx_base + offset]; let d = f32(block_q5_1.d); @@ -313,11 +277,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(Q5_1) #decl(Q8_0) -struct q8_0 { - d: f16, - qs: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block_q8_0 = src0[src0_idx_base + offset]; let d = f32(block_q8_0.d); @@ -336,12 +295,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(Q8_0) #decl(Q8_1) -struct q8_1 { - d: f16, - m: f16, - qs: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block_q8_1 = src0[src0_idx_base + offset]; let d = f32(block_q8_1.d); @@ -362,13 +315,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #decl(Q2_K) // 16 blocks of 16 elements each -struct q2_k { - scales: array, - qs: array, - d: f16, - dmin: f16 -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); @@ -403,13 +349,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #decl(Q3_K) // 16 blocks of 16 elements each -struct q3_k { - hmask: array, - qs: array, - scales: array, // 6-bit quantized values - d: f16 -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); @@ -470,34 +409,8 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(Q3_K) -#decl(Q45_K_SCALE_MIN) - -fn get_scale_min(is: u32, scales: array) -> vec2 { - if (is < 4) { - let sc_byte = get_byte(scales[is / 4], is % 4); - let min_byte = get_byte(scales[(is + 4) / 4], is % 4); - return vec2(f32(sc_byte & 63), f32(min_byte & 63)); - } else { - let sc_min_lo = get_byte(scales[(is + 4) / 4], (is + 4) % 4); - let sc_hi = get_byte(scales[(is - 4) / 4], (is - 4) % 4); - let min_hi = get_byte(scales[is / 4], is % 4); - let sc = (sc_min_lo & 0xF) | ((sc_hi >> 6) << 4); - let m = (sc_min_lo >> 4) | ((min_hi >> 6) << 4); - return vec2(f32(sc), f32(m)); - } -} - -#enddecl(Q45_K_SCALE_MIN) - #decl(Q4_K) // 8 blocks of 32 elements each -struct q4_k { - d: f16, - dmin: f16, - scales: array, - qs: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); @@ -528,14 +441,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #decl(Q5_K) // 8 blocks of 32 elements each -struct q5_k { - d: f16, - dmin: f16, - scales: array, - qh: array, - qs: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); @@ -570,13 +475,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #decl(Q6_K) // 16 blocks of 16 elements each -struct q6_k { - ql: array, - qh: array, - scales: array, - d: f16 -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); @@ -634,98 +532,7 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(Q6_K) -#decl(IQ23_TABLES) -const kmask_iq2xs : array = array( - 0x08040201u, // 1, 2, 4, 8 - 0x80402010u // 16, 32, 64, 128 -); - -const ksigns_iq2xs: array = array( - 0x03828100,0x87060584,0x8b0a0988,0x0f8e8d0c, - 0x93121190,0x17969514,0x1b9a9918,0x9f1e1d9c, - 0xa32221a0,0x27a6a524,0x2baaa928,0xaf2e2dac, - 0x33b2b130,0xb73635b4,0xbb3a39b8,0x3fbebd3c, - 0xc34241c0,0x47c6c544,0x4bcac948,0xcf4e4dcc, - 0x53d2d150,0xd75655d4,0xdb5a59d8,0x5fdedd5c, - 0x63e2e160,0xe76665e4,0xeb6a69e8,0x6feeed6c, - 0xf37271f0,0x77f6f574,0x7bfaf978,0xff7e7dfc -); -#enddecl(IQ23_TABLES) - #decl(IQ2_XXS) - -const iq2xxs_grid = array( - 0x08080808, 0x08080808, 0x0808082b, 0x08080808, 0x08081919, 0x08080808, 0x08082b08, 0x08080808, - 0x08082b2b, 0x08080808, 0x08190819, 0x08080808, 0x08191908, 0x08080808, 0x082b0808, 0x08080808, - 0x082b082b, 0x08080808, 0x082b2b08, 0x08080808, 0x082b2b2b, 0x08080808, 0x19080819, 0x08080808, - 0x19081908, 0x08080808, 0x19190808, 0x08080808, 0x19192b08, 0x08080808, 0x192b0819, 0x08080808, - 0x192b1908, 0x08080808, 0x2b080808, 0x08080808, 0x2b08082b, 0x08080808, 0x2b082b2b, 0x08080808, - 0x2b2b082b, 0x08080808, 0x08080819, 0x08080819, 0x08081908, 0x08080819, 0x08190808, 0x08080819, - 0x08191919, 0x08080819, 0x19080808, 0x08080819, 0x2b081908, 0x08080819, 0x2b192b08, 0x08080819, - 0x08080808, 0x0808082b, 0x0808082b, 0x0808082b, 0x082b082b, 0x0808082b, 0x2b08082b, 0x0808082b, - 0x08080819, 0x08081908, 0x08081908, 0x08081908, 0x08190808, 0x08081908, 0x082b0819, 0x08081908, - 0x082b1908, 0x08081908, 0x19080808, 0x08081908, 0x1908082b, 0x08081908, 0x19082b08, 0x08081908, - 0x192b0808, 0x08081908, 0x2b080819, 0x08081908, 0x2b081908, 0x08081908, 0x2b190808, 0x08081908, - 0x2b2b1908, 0x08081908, 0x08080808, 0x08081919, 0x0808082b, 0x08081919, 0x08082b08, 0x08081919, - 0x082b0808, 0x08081919, 0x1908192b, 0x08081919, 0x192b2b19, 0x08081919, 0x2b080808, 0x08081919, - 0x2b190819, 0x08081919, 0x08082b19, 0x0808192b, 0x08190808, 0x0808192b, 0x19080808, 0x0808192b, - 0x2b081908, 0x0808192b, 0x2b2b1908, 0x0808192b, 0x08080808, 0x08082b08, 0x08081919, 0x08082b08, - 0x08082b08, 0x08082b08, 0x08191908, 0x08082b08, 0x082b2b08, 0x08082b08, 0x19080819, 0x08082b08, - 0x19081908, 0x08082b08, 0x19190808, 0x08082b08, 0x1919082b, 0x08082b08, 0x2b082b08, 0x08082b08, - 0x08081908, 0x08082b19, 0x19080808, 0x08082b19, 0x0808082b, 0x08082b2b, 0x08191908, 0x08082b2b, - 0x08080819, 0x08190808, 0x08081908, 0x08190808, 0x08190808, 0x08190808, 0x082b0819, 0x08190808, - 0x19080808, 0x08190808, 0x192b0808, 0x08190808, 0x2b081908, 0x08190808, 0x2b190808, 0x08190808, - 0x2b191919, 0x08190808, 0x08080808, 0x08190819, 0x08082b08, 0x08190819, 0x082b0808, 0x08190819, - 0x19190808, 0x08190819, 0x19192b2b, 0x08190819, 0x2b080808, 0x08190819, 0x082b1908, 0x0819082b, - 0x19081919, 0x0819082b, 0x08080808, 0x08191908, 0x08082b08, 0x08191908, 0x082b0808, 0x08191908, - 0x082b1919, 0x08191908, 0x19082b19, 0x08191908, 0x2b080808, 0x08191908, 0x08192b08, 0x08191919, - 0x192b082b, 0x08191919, 0x08080808, 0x0819192b, 0x0819192b, 0x0819192b, 0x08080819, 0x08192b08, - 0x08081908, 0x08192b08, 0x08190808, 0x08192b08, 0x19080808, 0x08192b08, 0x2b080819, 0x08192b08, - 0x08080808, 0x08192b19, 0x08081919, 0x08192b19, 0x2b2b0808, 0x08192b19, 0x19190819, 0x08192b2b, - 0x08080808, 0x082b0808, 0x0808082b, 0x082b0808, 0x08082b2b, 0x082b0808, 0x19081908, 0x082b0808, - 0x192b0819, 0x082b0808, 0x2b080808, 0x082b0808, 0x2b08082b, 0x082b0808, 0x082b2b19, 0x082b0819, - 0x19082b08, 0x082b0819, 0x08080808, 0x082b082b, 0x0808082b, 0x082b082b, 0x08080819, 0x082b1908, - 0x08081908, 0x082b1908, 0x08190808, 0x082b1908, 0x19080808, 0x082b1908, 0x1919192b, 0x082b1908, - 0x08080808, 0x082b1919, 0x19080819, 0x082b1919, 0x192b1908, 0x082b1919, 0x2b190808, 0x082b192b, - 0x08082b08, 0x082b2b08, 0x082b0808, 0x082b2b08, 0x2b191908, 0x082b2b08, 0x19081908, 0x082b2b2b, - 0x08080819, 0x19080808, 0x08081908, 0x19080808, 0x08190808, 0x19080808, 0x08192b08, 0x19080808, - 0x082b0819, 0x19080808, 0x082b1908, 0x19080808, 0x19080808, 0x19080808, 0x19082b08, 0x19080808, - 0x1919192b, 0x19080808, 0x192b0808, 0x19080808, 0x2b080819, 0x19080808, 0x2b081908, 0x19080808, - 0x2b190808, 0x19080808, 0x08080808, 0x19080819, 0x082b0808, 0x19080819, 0x192b0819, 0x19080819, - 0x2b080808, 0x19080819, 0x2b081919, 0x19080819, 0x08080819, 0x1908082b, 0x08190808, 0x1908082b, - 0x19082b08, 0x1908082b, 0x1919192b, 0x1908082b, 0x192b2b08, 0x1908082b, 0x08080808, 0x19081908, - 0x08082b08, 0x19081908, 0x082b0808, 0x19081908, 0x2b080808, 0x19081908, 0x2b192b19, 0x19081908, - 0x0819082b, 0x19081919, 0x082b1908, 0x19081919, 0x08080808, 0x1908192b, 0x08080819, 0x19082b08, - 0x08081908, 0x19082b08, 0x08190808, 0x19082b08, 0x19080808, 0x19082b08, 0x19081919, 0x19082b08, - 0x08080808, 0x19082b19, 0x19192b08, 0x19082b19, 0x192b0819, 0x19082b19, 0x2b08082b, 0x19082b19, - 0x19081919, 0x19082b2b, 0x2b190808, 0x19082b2b, 0x08080808, 0x19190808, 0x08082b08, 0x19190808, - 0x08190819, 0x19190808, 0x08192b19, 0x19190808, 0x082b0808, 0x19190808, 0x2b080808, 0x19190808, - 0x2b082b08, 0x19190808, 0x08081908, 0x19190819, 0x1908082b, 0x19190819, 0x2b2b1908, 0x19190819, - 0x2b190819, 0x1919082b, 0x2b190808, 0x19191908, 0x2b19082b, 0x19191908, 0x08082b2b, 0x19191919, - 0x08080819, 0x1919192b, 0x19191908, 0x1919192b, 0x08080808, 0x19192b08, 0x08190819, 0x19192b08, - 0x08192b19, 0x19192b08, 0x192b1908, 0x19192b08, 0x19080808, 0x19192b19, 0x08082b08, 0x19192b2b, - 0x08081908, 0x192b0808, 0x08190808, 0x192b0808, 0x19080808, 0x192b0808, 0x192b2b08, 0x192b0808, - 0x08080808, 0x192b0819, 0x19191919, 0x192b0819, 0x08192b08, 0x192b082b, 0x192b0808, 0x192b082b, - 0x08080808, 0x192b1908, 0x08081919, 0x192b1908, 0x08190808, 0x192b1919, 0x0819082b, 0x192b1919, - 0x2b081908, 0x192b1919, 0x1908082b, 0x192b2b08, 0x08080808, 0x2b080808, 0x0808082b, 0x2b080808, - 0x08082b2b, 0x2b080808, 0x19080819, 0x2b080808, 0x2b08082b, 0x2b080808, 0x08081908, 0x2b080819, - 0x08192b08, 0x2b080819, 0x19080808, 0x2b080819, 0x08190819, 0x2b08082b, 0x08080819, 0x2b081908, - 0x08081908, 0x2b081908, 0x08190808, 0x2b081908, 0x08191919, 0x2b081908, 0x19080808, 0x2b081908, - 0x192b0808, 0x2b081908, 0x08080808, 0x2b081919, 0x1908192b, 0x2b081919, 0x2b191908, 0x2b081919, - 0x08082b19, 0x2b08192b, 0x19080808, 0x2b08192b, 0x192b0808, 0x2b08192b, 0x0808082b, 0x2b082b08, - 0x08081908, 0x2b082b19, 0x08190819, 0x2b082b2b, 0x08081908, 0x2b190808, 0x08190808, 0x2b190808, - 0x082b1908, 0x2b190808, 0x19080808, 0x2b190808, 0x2b2b0819, 0x2b190808, 0x0819192b, 0x2b190819, - 0x2b080808, 0x2b190819, 0x19081919, 0x2b19082b, 0x08080808, 0x2b191908, 0x082b082b, 0x2b191908, - 0x19081908, 0x2b191908, 0x19190819, 0x2b191919, 0x2b080819, 0x2b192b08, 0x082b0808, 0x2b192b19, - 0x0808082b, 0x2b2b0808, 0x19190808, 0x2b2b0808, 0x2b081919, 0x2b2b0808, 0x08082b19, 0x2b2b0819, - 0x08080808, 0x2b2b082b, 0x08192b08, 0x2b2b1908, 0x19190808, 0x2b2b2b08, 0x08081908, 0x2b2b2b19 -); - -struct iq2_xxs { - d: f16, - qs: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); @@ -753,143 +560,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(IQ2_XXS) #decl(IQ2_XS) -const iq2xs_grid = array( - 0x08080808, 0x08080808, 0x0808082b, 0x08080808, 0x08081919, 0x08080808, 0x08082b08, 0x08080808, - 0x08082b2b, 0x08080808, 0x08190819, 0x08080808, 0x08191908, 0x08080808, 0x0819192b, 0x08080808, - 0x08192b19, 0x08080808, 0x082b0808, 0x08080808, 0x082b082b, 0x08080808, 0x082b1919, 0x08080808, - 0x082b2b08, 0x08080808, 0x19080819, 0x08080808, 0x19081908, 0x08080808, 0x1908192b, 0x08080808, - 0x19082b19, 0x08080808, 0x19190808, 0x08080808, 0x1919082b, 0x08080808, 0x19191919, 0x08080808, - 0x19192b08, 0x08080808, 0x192b0819, 0x08080808, 0x192b1908, 0x08080808, 0x2b080808, 0x08080808, - 0x2b08082b, 0x08080808, 0x2b081919, 0x08080808, 0x2b082b08, 0x08080808, 0x2b190819, 0x08080808, - 0x2b191908, 0x08080808, 0x2b192b19, 0x08080808, 0x2b2b0808, 0x08080808, 0x08080819, 0x08080819, - 0x08081908, 0x08080819, 0x0808192b, 0x08080819, 0x08082b19, 0x08080819, 0x08190808, 0x08080819, - 0x0819082b, 0x08080819, 0x08191919, 0x08080819, 0x08192b08, 0x08080819, 0x08192b2b, 0x08080819, - 0x082b0819, 0x08080819, 0x082b1908, 0x08080819, 0x19080808, 0x08080819, 0x1908082b, 0x08080819, - 0x19081919, 0x08080819, 0x19082b08, 0x08080819, 0x19190819, 0x08080819, 0x19191908, 0x08080819, - 0x192b0808, 0x08080819, 0x192b2b08, 0x08080819, 0x2b080819, 0x08080819, 0x2b081908, 0x08080819, - 0x2b190808, 0x08080819, 0x08080808, 0x0808082b, 0x0808082b, 0x0808082b, 0x08081919, 0x0808082b, - 0x08082b08, 0x0808082b, 0x08190819, 0x0808082b, 0x08191908, 0x0808082b, 0x082b0808, 0x0808082b, - 0x19080819, 0x0808082b, 0x19081908, 0x0808082b, 0x19190808, 0x0808082b, 0x19191919, 0x0808082b, - 0x2b080808, 0x0808082b, 0x2b082b2b, 0x0808082b, 0x08080819, 0x08081908, 0x08081908, 0x08081908, - 0x0808192b, 0x08081908, 0x08082b19, 0x08081908, 0x08190808, 0x08081908, 0x0819082b, 0x08081908, - 0x08191919, 0x08081908, 0x08192b08, 0x08081908, 0x082b0819, 0x08081908, 0x082b1908, 0x08081908, - 0x19080808, 0x08081908, 0x1908082b, 0x08081908, 0x19081919, 0x08081908, 0x19082b08, 0x08081908, - 0x19190819, 0x08081908, 0x19191908, 0x08081908, 0x1919192b, 0x08081908, 0x192b0808, 0x08081908, - 0x2b080819, 0x08081908, 0x2b081908, 0x08081908, 0x2b190808, 0x08081908, 0x08080808, 0x08081919, - 0x0808082b, 0x08081919, 0x08081919, 0x08081919, 0x08082b08, 0x08081919, 0x08190819, 0x08081919, - 0x08191908, 0x08081919, 0x082b0808, 0x08081919, 0x19080819, 0x08081919, 0x19081908, 0x08081919, - 0x19190808, 0x08081919, 0x192b0819, 0x08081919, 0x2b080808, 0x08081919, 0x08080819, 0x0808192b, - 0x08081908, 0x0808192b, 0x08190808, 0x0808192b, 0x082b192b, 0x0808192b, 0x19080808, 0x0808192b, - 0x1908082b, 0x0808192b, 0x2b081908, 0x0808192b, 0x08080808, 0x08082b08, 0x0808082b, 0x08082b08, - 0x08081919, 0x08082b08, 0x08082b08, 0x08082b08, 0x08082b2b, 0x08082b08, 0x08190819, 0x08082b08, - 0x08191908, 0x08082b08, 0x082b0808, 0x08082b08, 0x082b1919, 0x08082b08, 0x19080819, 0x08082b08, - 0x19081908, 0x08082b08, 0x19190808, 0x08082b08, 0x19192b08, 0x08082b08, 0x2b080808, 0x08082b08, - 0x2b2b0808, 0x08082b08, 0x2b2b2b2b, 0x08082b08, 0x08080819, 0x08082b19, 0x08081908, 0x08082b19, - 0x08190808, 0x08082b19, 0x19080808, 0x08082b19, 0x2b080819, 0x08082b19, 0x2b082b19, 0x08082b19, - 0x08080808, 0x08082b2b, 0x082b0808, 0x08082b2b, 0x082b2b08, 0x08082b2b, 0x2b19192b, 0x08082b2b, - 0x2b2b0808, 0x08082b2b, 0x08080819, 0x08190808, 0x08081908, 0x08190808, 0x0808192b, 0x08190808, - 0x08082b19, 0x08190808, 0x08190808, 0x08190808, 0x0819082b, 0x08190808, 0x08191919, 0x08190808, - 0x08192b08, 0x08190808, 0x082b0819, 0x08190808, 0x082b1908, 0x08190808, 0x19080808, 0x08190808, - 0x1908082b, 0x08190808, 0x19081919, 0x08190808, 0x19082b08, 0x08190808, 0x19190819, 0x08190808, - 0x19191908, 0x08190808, 0x192b0808, 0x08190808, 0x192b2b2b, 0x08190808, 0x2b080819, 0x08190808, - 0x2b081908, 0x08190808, 0x2b190808, 0x08190808, 0x08080808, 0x08190819, 0x0808082b, 0x08190819, - 0x08081919, 0x08190819, 0x08082b08, 0x08190819, 0x08190819, 0x08190819, 0x08191908, 0x08190819, - 0x082b0808, 0x08190819, 0x19080819, 0x08190819, 0x19081908, 0x08190819, 0x19190808, 0x08190819, - 0x2b080808, 0x08190819, 0x2b191908, 0x08190819, 0x2b19192b, 0x08190819, 0x08080819, 0x0819082b, - 0x08081908, 0x0819082b, 0x0808192b, 0x0819082b, 0x08190808, 0x0819082b, 0x19080808, 0x0819082b, - 0x192b0808, 0x0819082b, 0x08080808, 0x08191908, 0x0808082b, 0x08191908, 0x08081919, 0x08191908, - 0x08082b08, 0x08191908, 0x08190819, 0x08191908, 0x08191908, 0x08191908, 0x082b0808, 0x08191908, - 0x19080819, 0x08191908, 0x19081908, 0x08191908, 0x19082b19, 0x08191908, 0x19190808, 0x08191908, - 0x192b1908, 0x08191908, 0x2b080808, 0x08191908, 0x08080819, 0x08191919, 0x08081908, 0x08191919, - 0x08190808, 0x08191919, 0x19080808, 0x08191919, 0x08080808, 0x0819192b, 0x08191908, 0x0819192b, - 0x19082b19, 0x0819192b, 0x08080819, 0x08192b08, 0x08081908, 0x08192b08, 0x08190808, 0x08192b08, - 0x0819082b, 0x08192b08, 0x19080808, 0x08192b08, 0x19191908, 0x08192b08, 0x2b08192b, 0x08192b08, - 0x08080808, 0x08192b19, 0x08081919, 0x08192b19, 0x192b192b, 0x08192b19, 0x19190819, 0x08192b2b, - 0x2b2b2b19, 0x08192b2b, 0x08080808, 0x082b0808, 0x0808082b, 0x082b0808, 0x08081919, 0x082b0808, - 0x08082b08, 0x082b0808, 0x08082b2b, 0x082b0808, 0x08190819, 0x082b0808, 0x08191908, 0x082b0808, - 0x082b0808, 0x082b0808, 0x19080819, 0x082b0808, 0x19081908, 0x082b0808, 0x19190808, 0x082b0808, - 0x2b080808, 0x082b0808, 0x2b2b0808, 0x082b0808, 0x08080819, 0x082b0819, 0x08081908, 0x082b0819, - 0x08190808, 0x082b0819, 0x19080808, 0x082b0819, 0x19082b08, 0x082b0819, 0x192b1919, 0x082b0819, - 0x08080808, 0x082b082b, 0x082b082b, 0x082b082b, 0x2b080808, 0x082b082b, 0x2b2b2b08, 0x082b082b, - 0x08080819, 0x082b1908, 0x08081908, 0x082b1908, 0x08190808, 0x082b1908, 0x082b2b19, 0x082b1908, - 0x19080808, 0x082b1908, 0x08080808, 0x082b1919, 0x19080819, 0x082b1919, 0x1919082b, 0x082b1919, - 0x2b192b19, 0x082b1919, 0x08080819, 0x082b192b, 0x08192b2b, 0x082b192b, 0x2b2b192b, 0x082b192b, - 0x08080808, 0x082b2b08, 0x08082b08, 0x082b2b08, 0x08082b2b, 0x082b2b08, 0x082b0808, 0x082b2b08, - 0x19191919, 0x082b2b08, 0x2b082b08, 0x082b2b08, 0x2b2b082b, 0x082b2b08, 0x192b2b08, 0x082b2b19, - 0x2b190808, 0x082b2b19, 0x08082b08, 0x082b2b2b, 0x082b0808, 0x082b2b2b, 0x2b08082b, 0x082b2b2b, - 0x2b082b08, 0x082b2b2b, 0x2b082b2b, 0x082b2b2b, 0x08080819, 0x19080808, 0x08081908, 0x19080808, - 0x0808192b, 0x19080808, 0x08082b19, 0x19080808, 0x08190808, 0x19080808, 0x0819082b, 0x19080808, - 0x08191919, 0x19080808, 0x08192b08, 0x19080808, 0x082b0819, 0x19080808, 0x082b1908, 0x19080808, - 0x19080808, 0x19080808, 0x1908082b, 0x19080808, 0x19081919, 0x19080808, 0x19082b08, 0x19080808, - 0x19082b2b, 0x19080808, 0x19190819, 0x19080808, 0x19191908, 0x19080808, 0x192b0808, 0x19080808, - 0x192b1919, 0x19080808, 0x2b080819, 0x19080808, 0x2b081908, 0x19080808, 0x2b190808, 0x19080808, - 0x08080808, 0x19080819, 0x0808082b, 0x19080819, 0x08081919, 0x19080819, 0x08082b08, 0x19080819, - 0x08190819, 0x19080819, 0x08191908, 0x19080819, 0x082b0808, 0x19080819, 0x19080819, 0x19080819, - 0x19081908, 0x19080819, 0x19190808, 0x19080819, 0x2b080808, 0x19080819, 0x2b081919, 0x19080819, - 0x2b2b082b, 0x19080819, 0x08080819, 0x1908082b, 0x08081908, 0x1908082b, 0x08190808, 0x1908082b, - 0x0819082b, 0x1908082b, 0x082b2b19, 0x1908082b, 0x19080808, 0x1908082b, 0x08080808, 0x19081908, - 0x0808082b, 0x19081908, 0x08081919, 0x19081908, 0x08082b08, 0x19081908, 0x08190819, 0x19081908, - 0x08191908, 0x19081908, 0x08192b19, 0x19081908, 0x082b0808, 0x19081908, 0x19080819, 0x19081908, - 0x19081908, 0x19081908, 0x19190808, 0x19081908, 0x2b080808, 0x19081908, 0x2b191908, 0x19081908, - 0x08080819, 0x19081919, 0x08081908, 0x19081919, 0x08190808, 0x19081919, 0x082b1908, 0x19081919, - 0x19080808, 0x19081919, 0x2b192b2b, 0x19081919, 0x08080808, 0x1908192b, 0x08082b2b, 0x1908192b, - 0x19081908, 0x1908192b, 0x19190808, 0x1908192b, 0x08080819, 0x19082b08, 0x08081908, 0x19082b08, - 0x08190808, 0x19082b08, 0x19080808, 0x19082b08, 0x19081919, 0x19082b08, 0x19191908, 0x19082b08, - 0x192b082b, 0x19082b08, 0x08080808, 0x19082b19, 0x08190819, 0x19082b19, 0x19081908, 0x19082b19, - 0x19190808, 0x19082b19, 0x192b2b19, 0x19082b19, 0x08081908, 0x19082b2b, 0x08080808, 0x19190808, - 0x0808082b, 0x19190808, 0x08081919, 0x19190808, 0x08082b08, 0x19190808, 0x08190819, 0x19190808, - 0x08191908, 0x19190808, 0x082b0808, 0x19190808, 0x082b2b08, 0x19190808, 0x19080819, 0x19190808, - 0x19081908, 0x19190808, 0x19190808, 0x19190808, 0x2b080808, 0x19190808, 0x08080819, 0x19190819, - 0x08081908, 0x19190819, 0x08190808, 0x19190819, 0x08191919, 0x19190819, 0x19080808, 0x19190819, - 0x1908082b, 0x19190819, 0x08080808, 0x1919082b, 0x19081908, 0x1919082b, 0x2b2b2b2b, 0x1919082b, - 0x08080819, 0x19191908, 0x08081908, 0x19191908, 0x08190808, 0x19191908, 0x082b0819, 0x19191908, - 0x19080808, 0x19191908, 0x192b0808, 0x19191908, 0x2b080819, 0x19191908, 0x2b2b0819, 0x19191908, - 0x08080808, 0x19191919, 0x08082b08, 0x19191919, 0x2b080808, 0x19191919, 0x2b082b08, 0x19191919, - 0x082b0819, 0x1919192b, 0x192b2b08, 0x1919192b, 0x2b2b0819, 0x1919192b, 0x08080808, 0x19192b08, - 0x08191908, 0x19192b08, 0x19080819, 0x19192b08, 0x19190808, 0x19192b08, 0x2b192b19, 0x19192b08, - 0x08192b2b, 0x19192b19, 0x19080808, 0x19192b19, 0x1908082b, 0x19192b19, 0x2b081919, 0x19192b2b, - 0x08080819, 0x192b0808, 0x08081908, 0x192b0808, 0x08190808, 0x192b0808, 0x19080808, 0x192b0808, - 0x19191908, 0x192b0808, 0x192b082b, 0x192b0808, 0x2b08192b, 0x192b0808, 0x2b2b2b19, 0x192b0808, - 0x08080808, 0x192b0819, 0x082b1908, 0x192b082b, 0x19082b2b, 0x192b082b, 0x2b19082b, 0x192b082b, - 0x08080808, 0x192b1908, 0x0819192b, 0x192b1908, 0x08190808, 0x192b1919, 0x19080808, 0x192b1919, - 0x19081919, 0x192b1919, 0x2b2b1908, 0x192b1919, 0x08080819, 0x192b2b08, 0x192b2b2b, 0x192b2b08, - 0x082b1919, 0x192b2b19, 0x0808192b, 0x192b2b2b, 0x19191908, 0x192b2b2b, 0x192b082b, 0x192b2b2b, - 0x08080808, 0x2b080808, 0x0808082b, 0x2b080808, 0x08081919, 0x2b080808, 0x08082b08, 0x2b080808, - 0x08190819, 0x2b080808, 0x08191908, 0x2b080808, 0x082b0808, 0x2b080808, 0x082b2b2b, 0x2b080808, - 0x19080819, 0x2b080808, 0x19081908, 0x2b080808, 0x19190808, 0x2b080808, 0x2b080808, 0x2b080808, - 0x2b08082b, 0x2b080808, 0x2b2b2b08, 0x2b080808, 0x2b2b2b2b, 0x2b080808, 0x08080819, 0x2b080819, - 0x08081908, 0x2b080819, 0x0808192b, 0x2b080819, 0x08190808, 0x2b080819, 0x19080808, 0x2b080819, - 0x19190819, 0x2b080819, 0x19192b19, 0x2b080819, 0x08080808, 0x2b08082b, 0x082b0808, 0x2b08082b, - 0x2b080808, 0x2b08082b, 0x2b08082b, 0x2b08082b, 0x2b2b0808, 0x2b08082b, 0x2b2b2b08, 0x2b08082b, - 0x08080819, 0x2b081908, 0x08081908, 0x2b081908, 0x08190808, 0x2b081908, 0x0819082b, 0x2b081908, - 0x08191919, 0x2b081908, 0x19080808, 0x2b081908, 0x192b0808, 0x2b081908, 0x2b082b19, 0x2b081908, - 0x08080808, 0x2b081919, 0x19081908, 0x2b081919, 0x2b2b1919, 0x2b081919, 0x08192b08, 0x2b08192b, - 0x192b2b2b, 0x2b08192b, 0x08080808, 0x2b082b08, 0x08082b08, 0x2b082b08, 0x082b1919, 0x2b082b08, - 0x19192b2b, 0x2b082b08, 0x2b080808, 0x2b082b08, 0x2b08082b, 0x2b082b08, 0x2b2b2b08, 0x2b082b08, - 0x0808192b, 0x2b082b19, 0x082b082b, 0x2b082b2b, 0x2b080808, 0x2b082b2b, 0x2b082b08, 0x2b082b2b, - 0x2b19192b, 0x2b082b2b, 0x2b2b2b08, 0x2b082b2b, 0x08080819, 0x2b190808, 0x08081908, 0x2b190808, - 0x08190808, 0x2b190808, 0x19080808, 0x2b190808, 0x1919192b, 0x2b190808, 0x2b081908, 0x2b190808, - 0x08080808, 0x2b190819, 0x082b082b, 0x2b190819, 0x192b1908, 0x2b190819, 0x1919192b, 0x2b19082b, - 0x2b082b19, 0x2b19082b, 0x08080808, 0x2b191908, 0x08081919, 0x2b191908, 0x19081908, 0x2b191908, - 0x19190808, 0x2b191908, 0x19192b08, 0x2b191908, 0x082b2b19, 0x2b191919, 0x2b190808, 0x2b191919, - 0x2b19082b, 0x2b191919, 0x19080819, 0x2b19192b, 0x19190819, 0x2b192b08, 0x2b2b192b, 0x2b192b08, - 0x19082b19, 0x2b192b19, 0x08191919, 0x2b192b2b, 0x192b0808, 0x2b192b2b, 0x08080808, 0x2b2b0808, - 0x0808082b, 0x2b2b0808, 0x08082b08, 0x2b2b0808, 0x08082b2b, 0x2b2b0808, 0x082b0808, 0x2b2b0808, - 0x082b2b2b, 0x2b2b0808, 0x2b2b0808, 0x2b2b0808, 0x19190819, 0x2b2b0819, 0x19192b19, 0x2b2b0819, - 0x2b2b192b, 0x2b2b0819, 0x08080808, 0x2b2b082b, 0x0808082b, 0x2b2b082b, 0x08082b08, 0x2b2b082b, - 0x082b2b2b, 0x2b2b082b, 0x2b080808, 0x2b2b082b, 0x2b2b0808, 0x2b2b082b, 0x19080808, 0x2b2b1908, - 0x2b191919, 0x2b2b1908, 0x192b1919, 0x2b2b192b, 0x2b192b08, 0x2b2b192b, 0x08082b2b, 0x2b2b2b08, - 0x082b0808, 0x2b2b2b08, 0x082b082b, 0x2b2b2b08, 0x082b2b08, 0x2b2b2b08, 0x2b2b0808, 0x2b2b2b08, - 0x2b2b2b08, 0x2b2b2b08, 0x08081908, 0x2b2b2b19, 0x2b081908, 0x2b2b2b19, 0x2b08192b, 0x2b2b2b19, - 0x082b2b08, 0x2b2b2b2b, 0x082b2b2b, 0x2b2b2b2b, 0x2b190819, 0x2b2b2b2b, 0x2b2b2b2b, 0x2b2b2b2b -); - -struct iq2_xs { - d: f16, - qs: array, - scales: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); @@ -925,273 +595,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(IQ2_XS) #decl(IQ2_S) - -const iq2s_grid = array( - 0x08080808, 0x08080808, 0x0808082b, 0x08080808, 0x08081919, 0x08080808, 0x08082b08, 0x08080808, - 0x08082b2b, 0x08080808, 0x08190819, 0x08080808, 0x08191908, 0x08080808, 0x0819192b, 0x08080808, - 0x08192b19, 0x08080808, 0x082b0808, 0x08080808, 0x082b082b, 0x08080808, 0x082b1919, 0x08080808, - 0x082b2b08, 0x08080808, 0x19080819, 0x08080808, 0x19081908, 0x08080808, 0x1908192b, 0x08080808, - 0x19082b19, 0x08080808, 0x19190808, 0x08080808, 0x1919082b, 0x08080808, 0x19191919, 0x08080808, - 0x19192b08, 0x08080808, 0x192b0819, 0x08080808, 0x192b1908, 0x08080808, 0x192b192b, 0x08080808, - 0x192b2b19, 0x08080808, 0x2b080808, 0x08080808, 0x2b08082b, 0x08080808, 0x2b081919, 0x08080808, - 0x2b082b08, 0x08080808, 0x2b190819, 0x08080808, 0x2b191908, 0x08080808, 0x2b2b0808, 0x08080808, - 0x2b2b1919, 0x08080808, 0x2b2b2b2b, 0x08080808, 0x08080819, 0x08080819, 0x08081908, 0x08080819, - 0x0808192b, 0x08080819, 0x08082b19, 0x08080819, 0x08190808, 0x08080819, 0x0819082b, 0x08080819, - 0x08191919, 0x08080819, 0x08192b08, 0x08080819, 0x082b0819, 0x08080819, 0x082b1908, 0x08080819, - 0x19080808, 0x08080819, 0x1908082b, 0x08080819, 0x19081919, 0x08080819, 0x19082b08, 0x08080819, - 0x19190819, 0x08080819, 0x19191908, 0x08080819, 0x1919192b, 0x08080819, 0x19192b19, 0x08080819, - 0x192b0808, 0x08080819, 0x192b1919, 0x08080819, 0x192b2b08, 0x08080819, 0x2b080819, 0x08080819, - 0x2b081908, 0x08080819, 0x2b190808, 0x08080819, 0x2b19082b, 0x08080819, 0x2b191919, 0x08080819, - 0x2b2b0819, 0x08080819, 0x2b2b1908, 0x08080819, 0x08080808, 0x0808082b, 0x0808082b, 0x0808082b, - 0x08081919, 0x0808082b, 0x08082b08, 0x0808082b, 0x08190819, 0x0808082b, 0x08191908, 0x0808082b, - 0x082b0808, 0x0808082b, 0x082b2b2b, 0x0808082b, 0x19080819, 0x0808082b, 0x19081908, 0x0808082b, - 0x1908192b, 0x0808082b, 0x19082b19, 0x0808082b, 0x19190808, 0x0808082b, 0x19191919, 0x0808082b, - 0x2b080808, 0x0808082b, 0x2b081919, 0x0808082b, 0x2b082b2b, 0x0808082b, 0x2b191908, 0x0808082b, - 0x2b2b082b, 0x0808082b, 0x08080819, 0x08081908, 0x08081908, 0x08081908, 0x0808192b, 0x08081908, - 0x08082b19, 0x08081908, 0x08190808, 0x08081908, 0x0819082b, 0x08081908, 0x08191919, 0x08081908, - 0x08192b08, 0x08081908, 0x082b0819, 0x08081908, 0x082b1908, 0x08081908, 0x082b192b, 0x08081908, - 0x082b2b19, 0x08081908, 0x19080808, 0x08081908, 0x1908082b, 0x08081908, 0x19081919, 0x08081908, - 0x19082b08, 0x08081908, 0x19082b2b, 0x08081908, 0x19190819, 0x08081908, 0x19191908, 0x08081908, - 0x1919192b, 0x08081908, 0x19192b19, 0x08081908, 0x192b0808, 0x08081908, 0x192b082b, 0x08081908, - 0x192b1919, 0x08081908, 0x2b080819, 0x08081908, 0x2b081908, 0x08081908, 0x2b08192b, 0x08081908, - 0x2b082b19, 0x08081908, 0x2b190808, 0x08081908, 0x2b191919, 0x08081908, 0x2b192b08, 0x08081908, - 0x2b2b0819, 0x08081908, 0x2b2b1908, 0x08081908, 0x08080808, 0x08081919, 0x0808082b, 0x08081919, - 0x08081919, 0x08081919, 0x08082b08, 0x08081919, 0x08082b2b, 0x08081919, 0x08190819, 0x08081919, - 0x08191908, 0x08081919, 0x0819192b, 0x08081919, 0x08192b19, 0x08081919, 0x082b0808, 0x08081919, - 0x082b1919, 0x08081919, 0x082b2b08, 0x08081919, 0x19080819, 0x08081919, 0x19081908, 0x08081919, - 0x1908192b, 0x08081919, 0x19082b19, 0x08081919, 0x19190808, 0x08081919, 0x1919082b, 0x08081919, - 0x19191919, 0x08081919, 0x19192b08, 0x08081919, 0x192b0819, 0x08081919, 0x192b1908, 0x08081919, - 0x2b080808, 0x08081919, 0x2b08082b, 0x08081919, 0x2b081919, 0x08081919, 0x2b082b08, 0x08081919, - 0x2b190819, 0x08081919, 0x2b191908, 0x08081919, 0x2b2b0808, 0x08081919, 0x08080819, 0x0808192b, - 0x08081908, 0x0808192b, 0x0808192b, 0x0808192b, 0x08082b19, 0x0808192b, 0x08190808, 0x0808192b, - 0x08191919, 0x0808192b, 0x19080808, 0x0808192b, 0x19081919, 0x0808192b, 0x19082b08, 0x0808192b, - 0x19190819, 0x0808192b, 0x19191908, 0x0808192b, 0x192b0808, 0x0808192b, 0x2b080819, 0x0808192b, - 0x2b081908, 0x0808192b, 0x2b190808, 0x0808192b, 0x08080808, 0x08082b08, 0x0808082b, 0x08082b08, - 0x08081919, 0x08082b08, 0x08082b08, 0x08082b08, 0x08190819, 0x08082b08, 0x08191908, 0x08082b08, - 0x0819192b, 0x08082b08, 0x08192b19, 0x08082b08, 0x082b0808, 0x08082b08, 0x082b1919, 0x08082b08, - 0x082b2b2b, 0x08082b08, 0x19080819, 0x08082b08, 0x19081908, 0x08082b08, 0x1908192b, 0x08082b08, - 0x19082b19, 0x08082b08, 0x19190808, 0x08082b08, 0x1919082b, 0x08082b08, 0x19191919, 0x08082b08, - 0x19192b08, 0x08082b08, 0x192b0819, 0x08082b08, 0x192b1908, 0x08082b08, 0x2b080808, 0x08082b08, - 0x2b081919, 0x08082b08, 0x2b191908, 0x08082b08, 0x2b2b2b2b, 0x08082b08, 0x08080819, 0x08082b19, - 0x08081908, 0x08082b19, 0x08190808, 0x08082b19, 0x0819082b, 0x08082b19, 0x08191919, 0x08082b19, - 0x08192b08, 0x08082b19, 0x082b0819, 0x08082b19, 0x19080808, 0x08082b19, 0x19081919, 0x08082b19, - 0x19082b08, 0x08082b19, 0x19190819, 0x08082b19, 0x19191908, 0x08082b19, 0x192b0808, 0x08082b19, - 0x2b080819, 0x08082b19, 0x2b190808, 0x08082b19, 0x08080808, 0x08082b2b, 0x08190819, 0x08082b2b, - 0x08191908, 0x08082b2b, 0x082b082b, 0x08082b2b, 0x082b2b08, 0x08082b2b, 0x082b2b2b, 0x08082b2b, - 0x19190808, 0x08082b2b, 0x2b192b19, 0x08082b2b, 0x08080819, 0x08190808, 0x08081908, 0x08190808, - 0x0808192b, 0x08190808, 0x08082b19, 0x08190808, 0x08190808, 0x08190808, 0x0819082b, 0x08190808, - 0x08191919, 0x08190808, 0x08192b08, 0x08190808, 0x082b0819, 0x08190808, 0x082b1908, 0x08190808, - 0x082b192b, 0x08190808, 0x19080808, 0x08190808, 0x1908082b, 0x08190808, 0x19081919, 0x08190808, - 0x19082b08, 0x08190808, 0x19190819, 0x08190808, 0x19191908, 0x08190808, 0x1919192b, 0x08190808, - 0x19192b19, 0x08190808, 0x192b0808, 0x08190808, 0x192b082b, 0x08190808, 0x192b1919, 0x08190808, - 0x192b2b08, 0x08190808, 0x2b080819, 0x08190808, 0x2b081908, 0x08190808, 0x2b08192b, 0x08190808, - 0x2b190808, 0x08190808, 0x2b191919, 0x08190808, 0x2b192b08, 0x08190808, 0x2b2b0819, 0x08190808, - 0x2b2b1908, 0x08190808, 0x08080808, 0x08190819, 0x0808082b, 0x08190819, 0x08081919, 0x08190819, - 0x08082b08, 0x08190819, 0x08082b2b, 0x08190819, 0x08190819, 0x08190819, 0x08191908, 0x08190819, - 0x0819192b, 0x08190819, 0x08192b19, 0x08190819, 0x082b0808, 0x08190819, 0x082b082b, 0x08190819, - 0x082b1919, 0x08190819, 0x082b2b08, 0x08190819, 0x19080819, 0x08190819, 0x19081908, 0x08190819, - 0x1908192b, 0x08190819, 0x19082b19, 0x08190819, 0x19190808, 0x08190819, 0x1919082b, 0x08190819, - 0x19191919, 0x08190819, 0x19192b08, 0x08190819, 0x192b0819, 0x08190819, 0x192b1908, 0x08190819, - 0x2b080808, 0x08190819, 0x2b08082b, 0x08190819, 0x2b081919, 0x08190819, 0x2b082b08, 0x08190819, - 0x2b190819, 0x08190819, 0x2b191908, 0x08190819, 0x08080819, 0x0819082b, 0x08081908, 0x0819082b, - 0x08082b19, 0x0819082b, 0x08190808, 0x0819082b, 0x08191919, 0x0819082b, 0x082b0819, 0x0819082b, - 0x082b1908, 0x0819082b, 0x19080808, 0x0819082b, 0x19081919, 0x0819082b, 0x19190819, 0x0819082b, - 0x19191908, 0x0819082b, 0x2b080819, 0x0819082b, 0x2b081908, 0x0819082b, 0x2b190808, 0x0819082b, - 0x08080808, 0x08191908, 0x0808082b, 0x08191908, 0x08081919, 0x08191908, 0x08082b08, 0x08191908, - 0x08190819, 0x08191908, 0x08191908, 0x08191908, 0x0819192b, 0x08191908, 0x08192b19, 0x08191908, - 0x082b0808, 0x08191908, 0x082b1919, 0x08191908, 0x082b2b08, 0x08191908, 0x19080819, 0x08191908, - 0x19081908, 0x08191908, 0x1908192b, 0x08191908, 0x19082b19, 0x08191908, 0x19190808, 0x08191908, - 0x1919082b, 0x08191908, 0x19191919, 0x08191908, 0x19192b08, 0x08191908, 0x192b0819, 0x08191908, - 0x192b1908, 0x08191908, 0x2b080808, 0x08191908, 0x2b08082b, 0x08191908, 0x2b081919, 0x08191908, - 0x2b082b08, 0x08191908, 0x2b190819, 0x08191908, 0x2b191908, 0x08191908, 0x2b2b0808, 0x08191908, - 0x08080819, 0x08191919, 0x08081908, 0x08191919, 0x0808192b, 0x08191919, 0x08082b19, 0x08191919, - 0x08190808, 0x08191919, 0x0819082b, 0x08191919, 0x08191919, 0x08191919, 0x08192b08, 0x08191919, - 0x082b0819, 0x08191919, 0x082b1908, 0x08191919, 0x19080808, 0x08191919, 0x1908082b, 0x08191919, - 0x19081919, 0x08191919, 0x19082b08, 0x08191919, 0x19190819, 0x08191919, 0x19191908, 0x08191919, - 0x192b0808, 0x08191919, 0x2b080819, 0x08191919, 0x2b081908, 0x08191919, 0x2b190808, 0x08191919, - 0x08080808, 0x0819192b, 0x08081919, 0x0819192b, 0x08082b08, 0x0819192b, 0x08190819, 0x0819192b, - 0x08191908, 0x0819192b, 0x082b0808, 0x0819192b, 0x19080819, 0x0819192b, 0x19081908, 0x0819192b, - 0x19190808, 0x0819192b, 0x2b080808, 0x0819192b, 0x2b2b2b2b, 0x0819192b, 0x08080819, 0x08192b08, - 0x08081908, 0x08192b08, 0x0808192b, 0x08192b08, 0x08082b19, 0x08192b08, 0x08190808, 0x08192b08, - 0x08191919, 0x08192b08, 0x08192b08, 0x08192b08, 0x082b0819, 0x08192b08, 0x19080808, 0x08192b08, - 0x1908082b, 0x08192b08, 0x19081919, 0x08192b08, 0x19082b08, 0x08192b08, 0x19190819, 0x08192b08, - 0x19191908, 0x08192b08, 0x192b0808, 0x08192b08, 0x2b080819, 0x08192b08, 0x2b081908, 0x08192b08, - 0x08080808, 0x08192b19, 0x0808082b, 0x08192b19, 0x08081919, 0x08192b19, 0x08082b08, 0x08192b19, - 0x08190819, 0x08192b19, 0x08191908, 0x08192b19, 0x082b0808, 0x08192b19, 0x19080819, 0x08192b19, - 0x19081908, 0x08192b19, 0x19190808, 0x08192b19, 0x192b2b19, 0x08192b19, 0x2b2b082b, 0x08192b19, - 0x08081908, 0x08192b2b, 0x08190808, 0x08192b2b, 0x19080808, 0x08192b2b, 0x1919192b, 0x08192b2b, - 0x08080808, 0x082b0808, 0x0808082b, 0x082b0808, 0x08081919, 0x082b0808, 0x08082b08, 0x082b0808, - 0x08190819, 0x082b0808, 0x08191908, 0x082b0808, 0x0819192b, 0x082b0808, 0x08192b19, 0x082b0808, - 0x082b0808, 0x082b0808, 0x082b1919, 0x082b0808, 0x082b2b2b, 0x082b0808, 0x19080819, 0x082b0808, - 0x19081908, 0x082b0808, 0x19190808, 0x082b0808, 0x1919082b, 0x082b0808, 0x19191919, 0x082b0808, - 0x192b1908, 0x082b0808, 0x2b080808, 0x082b0808, 0x2b082b2b, 0x082b0808, 0x2b191908, 0x082b0808, - 0x2b2b2b2b, 0x082b0808, 0x08080819, 0x082b0819, 0x08081908, 0x082b0819, 0x08190808, 0x082b0819, - 0x0819082b, 0x082b0819, 0x08191919, 0x082b0819, 0x082b0819, 0x082b0819, 0x19080808, 0x082b0819, - 0x1908082b, 0x082b0819, 0x19081919, 0x082b0819, 0x19190819, 0x082b0819, 0x19191908, 0x082b0819, - 0x192b0808, 0x082b0819, 0x2b080819, 0x082b0819, 0x2b081908, 0x082b0819, 0x2b190808, 0x082b0819, - 0x08080808, 0x082b082b, 0x08082b2b, 0x082b082b, 0x082b082b, 0x082b082b, 0x082b2b08, 0x082b082b, - 0x082b2b2b, 0x082b082b, 0x19081908, 0x082b082b, 0x19190808, 0x082b082b, 0x2b082b08, 0x082b082b, - 0x2b082b2b, 0x082b082b, 0x2b2b2b08, 0x082b082b, 0x08080819, 0x082b1908, 0x08081908, 0x082b1908, - 0x0808192b, 0x082b1908, 0x08082b19, 0x082b1908, 0x08190808, 0x082b1908, 0x08191919, 0x082b1908, - 0x08192b08, 0x082b1908, 0x082b0819, 0x082b1908, 0x082b1908, 0x082b1908, 0x19080808, 0x082b1908, - 0x1908082b, 0x082b1908, 0x19081919, 0x082b1908, 0x19082b08, 0x082b1908, 0x19190819, 0x082b1908, - 0x19191908, 0x082b1908, 0x192b0808, 0x082b1908, 0x2b080819, 0x082b1908, 0x2b081908, 0x082b1908, - 0x2b190808, 0x082b1908, 0x08080808, 0x082b1919, 0x08081919, 0x082b1919, 0x08082b08, 0x082b1919, - 0x08190819, 0x082b1919, 0x08191908, 0x082b1919, 0x082b0808, 0x082b1919, 0x19080819, 0x082b1919, - 0x19081908, 0x082b1919, 0x19190808, 0x082b1919, 0x192b192b, 0x082b1919, 0x2b080808, 0x082b1919, - 0x08080819, 0x082b192b, 0x08081908, 0x082b192b, 0x08190808, 0x082b192b, 0x19080808, 0x082b192b, - 0x19192b19, 0x082b192b, 0x08080808, 0x082b2b08, 0x08081919, 0x082b2b08, 0x08190819, 0x082b2b08, - 0x08191908, 0x082b2b08, 0x19080819, 0x082b2b08, 0x19081908, 0x082b2b08, 0x19190808, 0x082b2b08, - 0x2b082b2b, 0x082b2b08, 0x2b2b2b2b, 0x082b2b08, 0x08080819, 0x082b2b19, 0x08081908, 0x082b2b19, - 0x08190808, 0x082b2b19, 0x2b191919, 0x082b2b19, 0x08082b2b, 0x082b2b2b, 0x082b082b, 0x082b2b2b, - 0x192b1908, 0x082b2b2b, 0x2b082b08, 0x082b2b2b, 0x2b082b2b, 0x082b2b2b, 0x08080819, 0x19080808, - 0x08081908, 0x19080808, 0x0808192b, 0x19080808, 0x08082b19, 0x19080808, 0x08190808, 0x19080808, - 0x0819082b, 0x19080808, 0x08191919, 0x19080808, 0x08192b08, 0x19080808, 0x08192b2b, 0x19080808, - 0x082b0819, 0x19080808, 0x082b1908, 0x19080808, 0x082b192b, 0x19080808, 0x19080808, 0x19080808, - 0x1908082b, 0x19080808, 0x19081919, 0x19080808, 0x19082b08, 0x19080808, 0x19082b2b, 0x19080808, - 0x19190819, 0x19080808, 0x19191908, 0x19080808, 0x1919192b, 0x19080808, 0x19192b19, 0x19080808, - 0x192b0808, 0x19080808, 0x192b082b, 0x19080808, 0x192b1919, 0x19080808, 0x2b080819, 0x19080808, - 0x2b081908, 0x19080808, 0x2b190808, 0x19080808, 0x2b191919, 0x19080808, 0x2b192b08, 0x19080808, - 0x2b2b0819, 0x19080808, 0x2b2b1908, 0x19080808, 0x08080808, 0x19080819, 0x0808082b, 0x19080819, - 0x08081919, 0x19080819, 0x08082b08, 0x19080819, 0x08190819, 0x19080819, 0x08191908, 0x19080819, - 0x0819192b, 0x19080819, 0x08192b19, 0x19080819, 0x082b0808, 0x19080819, 0x082b082b, 0x19080819, - 0x082b1919, 0x19080819, 0x19080819, 0x19080819, 0x19081908, 0x19080819, 0x1908192b, 0x19080819, - 0x19082b19, 0x19080819, 0x19190808, 0x19080819, 0x1919082b, 0x19080819, 0x19191919, 0x19080819, - 0x19192b08, 0x19080819, 0x192b0819, 0x19080819, 0x192b1908, 0x19080819, 0x2b080808, 0x19080819, - 0x2b08082b, 0x19080819, 0x2b081919, 0x19080819, 0x2b082b08, 0x19080819, 0x2b190819, 0x19080819, - 0x2b191908, 0x19080819, 0x2b2b0808, 0x19080819, 0x08080819, 0x1908082b, 0x08081908, 0x1908082b, - 0x08190808, 0x1908082b, 0x0819082b, 0x1908082b, 0x08191919, 0x1908082b, 0x08192b08, 0x1908082b, - 0x082b1908, 0x1908082b, 0x19080808, 0x1908082b, 0x19081919, 0x1908082b, 0x19082b08, 0x1908082b, - 0x19190819, 0x1908082b, 0x19191908, 0x1908082b, 0x192b0808, 0x1908082b, 0x2b080819, 0x1908082b, - 0x2b081908, 0x1908082b, 0x08080808, 0x19081908, 0x0808082b, 0x19081908, 0x08081919, 0x19081908, - 0x08082b08, 0x19081908, 0x08082b2b, 0x19081908, 0x08190819, 0x19081908, 0x08191908, 0x19081908, - 0x0819192b, 0x19081908, 0x08192b19, 0x19081908, 0x082b0808, 0x19081908, 0x082b082b, 0x19081908, - 0x082b1919, 0x19081908, 0x082b2b08, 0x19081908, 0x19080819, 0x19081908, 0x19081908, 0x19081908, - 0x1908192b, 0x19081908, 0x19082b19, 0x19081908, 0x19190808, 0x19081908, 0x1919082b, 0x19081908, - 0x19191919, 0x19081908, 0x19192b08, 0x19081908, 0x192b0819, 0x19081908, 0x192b1908, 0x19081908, - 0x2b080808, 0x19081908, 0x2b08082b, 0x19081908, 0x2b081919, 0x19081908, 0x2b082b08, 0x19081908, - 0x2b190819, 0x19081908, 0x2b191908, 0x19081908, 0x2b2b0808, 0x19081908, 0x08080819, 0x19081919, - 0x08081908, 0x19081919, 0x0808192b, 0x19081919, 0x08082b19, 0x19081919, 0x08190808, 0x19081919, - 0x0819082b, 0x19081919, 0x08191919, 0x19081919, 0x08192b08, 0x19081919, 0x082b0819, 0x19081919, - 0x082b1908, 0x19081919, 0x19080808, 0x19081919, 0x1908082b, 0x19081919, 0x19081919, 0x19081919, - 0x19082b08, 0x19081919, 0x19190819, 0x19081919, 0x19191908, 0x19081919, 0x192b0808, 0x19081919, - 0x192b2b2b, 0x19081919, 0x2b080819, 0x19081919, 0x2b081908, 0x19081919, 0x2b190808, 0x19081919, - 0x08080808, 0x1908192b, 0x0808082b, 0x1908192b, 0x08081919, 0x1908192b, 0x08082b08, 0x1908192b, - 0x08190819, 0x1908192b, 0x08191908, 0x1908192b, 0x082b0808, 0x1908192b, 0x19080819, 0x1908192b, - 0x19081908, 0x1908192b, 0x19190808, 0x1908192b, 0x2b080808, 0x1908192b, 0x2b2b1919, 0x1908192b, - 0x08080819, 0x19082b08, 0x08081908, 0x19082b08, 0x08082b19, 0x19082b08, 0x08190808, 0x19082b08, - 0x0819082b, 0x19082b08, 0x08191919, 0x19082b08, 0x08192b08, 0x19082b08, 0x082b0819, 0x19082b08, - 0x082b1908, 0x19082b08, 0x19080808, 0x19082b08, 0x1908082b, 0x19082b08, 0x19081919, 0x19082b08, - 0x19082b08, 0x19082b08, 0x19190819, 0x19082b08, 0x19191908, 0x19082b08, 0x192b0808, 0x19082b08, - 0x2b081908, 0x19082b08, 0x2b190808, 0x19082b08, 0x08080808, 0x19082b19, 0x0808082b, 0x19082b19, - 0x08081919, 0x19082b19, 0x08082b08, 0x19082b19, 0x08190819, 0x19082b19, 0x08191908, 0x19082b19, - 0x082b0808, 0x19082b19, 0x19080819, 0x19082b19, 0x19081908, 0x19082b19, 0x19190808, 0x19082b19, - 0x2b080808, 0x19082b19, 0x2b19192b, 0x19082b19, 0x08080819, 0x19082b2b, 0x08081908, 0x19082b2b, - 0x08190808, 0x19082b2b, 0x19080808, 0x19082b2b, 0x08080808, 0x19190808, 0x0808082b, 0x19190808, - 0x08081919, 0x19190808, 0x08082b08, 0x19190808, 0x08190819, 0x19190808, 0x08191908, 0x19190808, - 0x0819192b, 0x19190808, 0x08192b19, 0x19190808, 0x082b0808, 0x19190808, 0x082b082b, 0x19190808, - 0x082b1919, 0x19190808, 0x082b2b08, 0x19190808, 0x19080819, 0x19190808, 0x19081908, 0x19190808, - 0x1908192b, 0x19190808, 0x19082b19, 0x19190808, 0x19190808, 0x19190808, 0x1919082b, 0x19190808, - 0x19191919, 0x19190808, 0x19192b08, 0x19190808, 0x192b0819, 0x19190808, 0x192b1908, 0x19190808, - 0x2b080808, 0x19190808, 0x2b08082b, 0x19190808, 0x2b081919, 0x19190808, 0x2b082b08, 0x19190808, - 0x2b190819, 0x19190808, 0x2b191908, 0x19190808, 0x08080819, 0x19190819, 0x08081908, 0x19190819, - 0x0808192b, 0x19190819, 0x08082b19, 0x19190819, 0x08190808, 0x19190819, 0x0819082b, 0x19190819, - 0x08191919, 0x19190819, 0x08192b08, 0x19190819, 0x082b0819, 0x19190819, 0x082b1908, 0x19190819, - 0x19080808, 0x19190819, 0x1908082b, 0x19190819, 0x19081919, 0x19190819, 0x19082b08, 0x19190819, - 0x19190819, 0x19190819, 0x19191908, 0x19190819, 0x192b0808, 0x19190819, 0x2b080819, 0x19190819, - 0x2b081908, 0x19190819, 0x2b190808, 0x19190819, 0x08080808, 0x1919082b, 0x08081919, 0x1919082b, - 0x08082b08, 0x1919082b, 0x08190819, 0x1919082b, 0x08191908, 0x1919082b, 0x082b0808, 0x1919082b, - 0x19080819, 0x1919082b, 0x19081908, 0x1919082b, 0x19190808, 0x1919082b, 0x192b2b19, 0x1919082b, - 0x2b080808, 0x1919082b, 0x08080819, 0x19191908, 0x08081908, 0x19191908, 0x0808192b, 0x19191908, - 0x08082b19, 0x19191908, 0x08190808, 0x19191908, 0x0819082b, 0x19191908, 0x08191919, 0x19191908, - 0x08192b08, 0x19191908, 0x082b0819, 0x19191908, 0x082b1908, 0x19191908, 0x19080808, 0x19191908, - 0x1908082b, 0x19191908, 0x19081919, 0x19191908, 0x19082b08, 0x19191908, 0x19190819, 0x19191908, - 0x19191908, 0x19191908, 0x192b0808, 0x19191908, 0x2b080819, 0x19191908, 0x2b081908, 0x19191908, - 0x2b190808, 0x19191908, 0x08080808, 0x19191919, 0x0808082b, 0x19191919, 0x08081919, 0x19191919, - 0x08082b08, 0x19191919, 0x08190819, 0x19191919, 0x08191908, 0x19191919, 0x082b0808, 0x19191919, - 0x19080819, 0x19191919, 0x19081908, 0x19191919, 0x19190808, 0x19191919, 0x2b080808, 0x19191919, - 0x08080819, 0x1919192b, 0x08081908, 0x1919192b, 0x08190808, 0x1919192b, 0x082b192b, 0x1919192b, - 0x19080808, 0x1919192b, 0x08080808, 0x19192b08, 0x0808082b, 0x19192b08, 0x08081919, 0x19192b08, - 0x08082b08, 0x19192b08, 0x08190819, 0x19192b08, 0x08191908, 0x19192b08, 0x082b0808, 0x19192b08, - 0x19080819, 0x19192b08, 0x19081908, 0x19192b08, 0x19190808, 0x19192b08, 0x19192b2b, 0x19192b08, - 0x2b080808, 0x19192b08, 0x08080819, 0x19192b19, 0x08081908, 0x19192b19, 0x08190808, 0x19192b19, - 0x19080808, 0x19192b19, 0x08080808, 0x19192b2b, 0x08192b19, 0x19192b2b, 0x2b081919, 0x19192b2b, - 0x2b2b2b08, 0x19192b2b, 0x08080819, 0x192b0808, 0x08081908, 0x192b0808, 0x0808192b, 0x192b0808, - 0x08190808, 0x192b0808, 0x0819082b, 0x192b0808, 0x08191919, 0x192b0808, 0x08192b08, 0x192b0808, - 0x082b0819, 0x192b0808, 0x082b1908, 0x192b0808, 0x19080808, 0x192b0808, 0x19081919, 0x192b0808, - 0x19082b08, 0x192b0808, 0x19190819, 0x192b0808, 0x19191908, 0x192b0808, 0x192b0808, 0x192b0808, - 0x2b081908, 0x192b0808, 0x2b190808, 0x192b0808, 0x08080808, 0x192b0819, 0x0808082b, 0x192b0819, - 0x08081919, 0x192b0819, 0x08082b08, 0x192b0819, 0x08190819, 0x192b0819, 0x08191908, 0x192b0819, - 0x082b0808, 0x192b0819, 0x19080819, 0x192b0819, 0x19081908, 0x192b0819, 0x19190808, 0x192b0819, - 0x2b080808, 0x192b0819, 0x2b192b19, 0x192b0819, 0x08081908, 0x192b082b, 0x08190808, 0x192b082b, - 0x19080808, 0x192b082b, 0x1919192b, 0x192b082b, 0x2b2b0819, 0x192b082b, 0x08080808, 0x192b1908, - 0x08081919, 0x192b1908, 0x08082b08, 0x192b1908, 0x08190819, 0x192b1908, 0x08191908, 0x192b1908, - 0x082b0808, 0x192b1908, 0x19080819, 0x192b1908, 0x19081908, 0x192b1908, 0x19190808, 0x192b1908, - 0x2b080808, 0x192b1908, 0x08080819, 0x192b1919, 0x08081908, 0x192b1919, 0x08190808, 0x192b1919, - 0x19080808, 0x192b1919, 0x19082b2b, 0x192b1919, 0x192b2b08, 0x192b1919, 0x2b19082b, 0x192b1919, - 0x08080808, 0x192b192b, 0x2b191908, 0x192b192b, 0x08080819, 0x192b2b08, 0x08081908, 0x192b2b08, - 0x08190808, 0x192b2b08, 0x192b1919, 0x192b2b08, 0x2b192b08, 0x192b2b08, 0x08080808, 0x192b2b19, - 0x082b2b2b, 0x192b2b19, 0x1908082b, 0x192b2b2b, 0x2b2b0819, 0x192b2b2b, 0x08080808, 0x2b080808, - 0x0808082b, 0x2b080808, 0x08081919, 0x2b080808, 0x08082b08, 0x2b080808, 0x08190819, 0x2b080808, - 0x08191908, 0x2b080808, 0x08192b19, 0x2b080808, 0x082b0808, 0x2b080808, 0x082b1919, 0x2b080808, - 0x19080819, 0x2b080808, 0x19081908, 0x2b080808, 0x19190808, 0x2b080808, 0x1919082b, 0x2b080808, - 0x19191919, 0x2b080808, 0x19192b08, 0x2b080808, 0x192b0819, 0x2b080808, 0x2b080808, 0x2b080808, - 0x2b081919, 0x2b080808, 0x2b190819, 0x2b080808, 0x2b191908, 0x2b080808, 0x08080819, 0x2b080819, - 0x08081908, 0x2b080819, 0x08082b19, 0x2b080819, 0x08190808, 0x2b080819, 0x0819082b, 0x2b080819, - 0x08191919, 0x2b080819, 0x08192b08, 0x2b080819, 0x082b0819, 0x2b080819, 0x082b1908, 0x2b080819, - 0x19080808, 0x2b080819, 0x1908082b, 0x2b080819, 0x19081919, 0x2b080819, 0x19082b08, 0x2b080819, - 0x19190819, 0x2b080819, 0x19191908, 0x2b080819, 0x2b080819, 0x2b080819, 0x2b081908, 0x2b080819, - 0x2b190808, 0x2b080819, 0x2b2b2b19, 0x2b080819, 0x08080808, 0x2b08082b, 0x08081919, 0x2b08082b, - 0x08082b2b, 0x2b08082b, 0x08190819, 0x2b08082b, 0x08191908, 0x2b08082b, 0x19080819, 0x2b08082b, - 0x19081908, 0x2b08082b, 0x19190808, 0x2b08082b, 0x08080819, 0x2b081908, 0x08081908, 0x2b081908, - 0x0808192b, 0x2b081908, 0x08082b19, 0x2b081908, 0x08190808, 0x2b081908, 0x0819082b, 0x2b081908, - 0x08191919, 0x2b081908, 0x08192b08, 0x2b081908, 0x082b0819, 0x2b081908, 0x19080808, 0x2b081908, - 0x1908082b, 0x2b081908, 0x19081919, 0x2b081908, 0x19082b08, 0x2b081908, 0x19190819, 0x2b081908, - 0x19191908, 0x2b081908, 0x192b0808, 0x2b081908, 0x2b080819, 0x2b081908, 0x2b081908, 0x2b081908, - 0x2b190808, 0x2b081908, 0x08080808, 0x2b081919, 0x0808082b, 0x2b081919, 0x08081919, 0x2b081919, - 0x08082b08, 0x2b081919, 0x08190819, 0x2b081919, 0x08191908, 0x2b081919, 0x082b0808, 0x2b081919, - 0x19080819, 0x2b081919, 0x19081908, 0x2b081919, 0x19190808, 0x2b081919, 0x2b080808, 0x2b081919, - 0x2b082b2b, 0x2b081919, 0x08080819, 0x2b08192b, 0x08081908, 0x2b08192b, 0x08190808, 0x2b08192b, - 0x082b2b19, 0x2b08192b, 0x19080808, 0x2b08192b, 0x08080808, 0x2b082b08, 0x08081919, 0x2b082b08, - 0x08190819, 0x2b082b08, 0x08191908, 0x2b082b08, 0x19080819, 0x2b082b08, 0x19081908, 0x2b082b08, - 0x19190808, 0x2b082b08, 0x2b2b082b, 0x2b082b08, 0x08080819, 0x2b082b19, 0x08081908, 0x2b082b19, - 0x19080808, 0x2b082b19, 0x192b1919, 0x2b082b19, 0x082b082b, 0x2b082b2b, 0x19192b08, 0x2b082b2b, - 0x19192b2b, 0x2b082b2b, 0x2b08082b, 0x2b082b2b, 0x2b2b082b, 0x2b082b2b, 0x08080819, 0x2b190808, - 0x08081908, 0x2b190808, 0x08082b19, 0x2b190808, 0x08190808, 0x2b190808, 0x0819082b, 0x2b190808, - 0x08191919, 0x2b190808, 0x08192b08, 0x2b190808, 0x082b1908, 0x2b190808, 0x19080808, 0x2b190808, - 0x1908082b, 0x2b190808, 0x19081919, 0x2b190808, 0x19082b08, 0x2b190808, 0x19190819, 0x2b190808, - 0x19191908, 0x2b190808, 0x192b0808, 0x2b190808, 0x2b080819, 0x2b190808, 0x2b081908, 0x2b190808, - 0x2b190808, 0x2b190808, 0x08080808, 0x2b190819, 0x08081919, 0x2b190819, 0x08190819, 0x2b190819, - 0x08191908, 0x2b190819, 0x19080819, 0x2b190819, 0x19081908, 0x2b190819, 0x19190808, 0x2b190819, - 0x19192b2b, 0x2b190819, 0x08080819, 0x2b19082b, 0x08081908, 0x2b19082b, 0x08190808, 0x2b19082b, - 0x19080808, 0x2b19082b, 0x2b2b192b, 0x2b19082b, 0x08080808, 0x2b191908, 0x0808082b, 0x2b191908, - 0x08081919, 0x2b191908, 0x08082b08, 0x2b191908, 0x08190819, 0x2b191908, 0x08191908, 0x2b191908, - 0x082b0808, 0x2b191908, 0x19080819, 0x2b191908, 0x19081908, 0x2b191908, 0x19190808, 0x2b191908, - 0x2b080808, 0x2b191908, 0x2b19192b, 0x2b191908, 0x08080819, 0x2b191919, 0x08081908, 0x2b191919, - 0x08190808, 0x2b191919, 0x19080808, 0x2b191919, 0x2b192b08, 0x2b191919, 0x2b2b0819, 0x2b191919, - 0x08080808, 0x2b19192b, 0x1908192b, 0x2b19192b, 0x192b1908, 0x2b19192b, 0x08080819, 0x2b192b08, - 0x08081908, 0x2b192b08, 0x08190808, 0x2b192b08, 0x082b192b, 0x2b192b08, 0x19080808, 0x2b192b08, - 0x2b2b2b19, 0x2b192b08, 0x08080808, 0x2b192b19, 0x19082b19, 0x2b192b19, 0x1919082b, 0x2b192b19, - 0x2b190808, 0x2b192b2b, 0x08080808, 0x2b2b0808, 0x08081919, 0x2b2b0808, 0x08082b2b, 0x2b2b0808, - 0x08191908, 0x2b2b0808, 0x082b082b, 0x2b2b0808, 0x082b2b2b, 0x2b2b0808, 0x19080819, 0x2b2b0808, - 0x19081908, 0x2b2b0808, 0x19190808, 0x2b2b0808, 0x2b2b082b, 0x2b2b0808, 0x2b2b2b2b, 0x2b2b0808, - 0x19080808, 0x2b2b0819, 0x192b1919, 0x2b2b0819, 0x0808082b, 0x2b2b082b, 0x08082b2b, 0x2b2b082b, - 0x082b082b, 0x2b2b082b, 0x082b2b08, 0x2b2b082b, 0x082b2b2b, 0x2b2b082b, 0x2b08082b, 0x2b2b082b, - 0x2b082b08, 0x2b2b082b, 0x2b082b2b, 0x2b2b082b, 0x2b2b2b08, 0x2b2b082b, 0x08080819, 0x2b2b1908, - 0x08081908, 0x2b2b1908, 0x08190808, 0x2b2b1908, 0x19080808, 0x2b2b1908, 0x2b082b19, 0x2b2b1908, - 0x2b2b1908, 0x2b2b1908, 0x08080808, 0x2b2b1919, 0x08192b19, 0x2b2b1919, 0x19190819, 0x2b2b192b, - 0x08082b2b, 0x2b2b2b08, 0x082b2b08, 0x2b2b2b08, 0x2b2b082b, 0x2b2b2b08, 0x19191908, 0x2b2b2b19, - 0x2b08192b, 0x2b2b2b19, 0x08082b08, 0x2b2b2b2b, 0x08082b2b, 0x2b2b2b2b, 0x082b0808, 0x2b2b2b2b, - 0x082b082b, 0x2b2b2b2b, 0x082b2b08, 0x2b2b2b2b, 0x2b082b08, 0x2b2b2b2b, 0x2b2b2b2b, 0x2b2b2b2b -); - -struct iq2_s { - d: f16, - qs: array, - qh: array, - scales: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); @@ -1236,47 +639,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(IQ2_S) #decl(IQ3_XSS) - -const iq3xxs_grid = array( - 0x04040404, 0x04040414, 0x04040424, 0x04040c0c, 0x04040c1c, 0x04040c3e, 0x04041404, 0x04041414, - 0x04041c0c, 0x04042414, 0x04043e1c, 0x04043e2c, 0x040c040c, 0x040c041c, 0x040c0c04, 0x040c0c14, - 0x040c140c, 0x040c142c, 0x040c1c04, 0x040c1c14, 0x040c240c, 0x040c2c24, 0x040c3e04, 0x04140404, - 0x04140414, 0x04140424, 0x04140c0c, 0x04141404, 0x04141414, 0x04141c0c, 0x04141c1c, 0x04141c3e, - 0x04142c0c, 0x04142c3e, 0x04143e2c, 0x041c040c, 0x041c043e, 0x041c0c04, 0x041c0c14, 0x041c142c, - 0x041c3e04, 0x04240c1c, 0x04241c3e, 0x04242424, 0x04242c3e, 0x04243e1c, 0x04243e2c, 0x042c040c, - 0x042c043e, 0x042c1c14, 0x042c2c14, 0x04341c2c, 0x04343424, 0x043e0c04, 0x043e0c24, 0x043e0c34, - 0x043e241c, 0x043e340c, 0x0c04040c, 0x0c04041c, 0x0c040c04, 0x0c040c14, 0x0c04140c, 0x0c04141c, - 0x0c041c04, 0x0c041c14, 0x0c041c24, 0x0c04243e, 0x0c042c04, 0x0c0c0404, 0x0c0c0414, 0x0c0c0c0c, - 0x0c0c1404, 0x0c0c1414, 0x0c14040c, 0x0c14041c, 0x0c140c04, 0x0c140c14, 0x0c14140c, 0x0c141c04, - 0x0c143e14, 0x0c1c0404, 0x0c1c0414, 0x0c1c1404, 0x0c1c1c0c, 0x0c1c2434, 0x0c1c3434, 0x0c24040c, - 0x0c24042c, 0x0c242c04, 0x0c2c1404, 0x0c2c1424, 0x0c2c2434, 0x0c2c3e0c, 0x0c34042c, 0x0c3e1414, - 0x0c3e2404, 0x14040404, 0x14040414, 0x14040c0c, 0x14040c1c, 0x14041404, 0x14041414, 0x14041434, - 0x14041c0c, 0x14042414, 0x140c040c, 0x140c041c, 0x140c042c, 0x140c0c04, 0x140c0c14, 0x140c140c, - 0x140c1c04, 0x140c341c, 0x140c343e, 0x140c3e04, 0x14140404, 0x14140414, 0x14140c0c, 0x14140c3e, - 0x14141404, 0x14141414, 0x14141c3e, 0x14142404, 0x14142c2c, 0x141c040c, 0x141c0c04, 0x141c0c24, - 0x141c3e04, 0x141c3e24, 0x14241c2c, 0x14242c1c, 0x142c041c, 0x142c143e, 0x142c240c, 0x142c3e24, - 0x143e040c, 0x143e041c, 0x143e0c34, 0x143e242c, 0x1c04040c, 0x1c040c04, 0x1c040c14, 0x1c04140c, - 0x1c04141c, 0x1c042c04, 0x1c04342c, 0x1c043e14, 0x1c0c0404, 0x1c0c0414, 0x1c0c1404, 0x1c0c1c0c, - 0x1c0c2424, 0x1c0c2434, 0x1c14040c, 0x1c14041c, 0x1c140c04, 0x1c14142c, 0x1c142c14, 0x1c143e14, - 0x1c1c0c0c, 0x1c1c1c1c, 0x1c241c04, 0x1c24243e, 0x1c243e14, 0x1c2c0404, 0x1c2c0434, 0x1c2c1414, - 0x1c2c2c2c, 0x1c340c24, 0x1c341c34, 0x1c34341c, 0x1c3e1c1c, 0x1c3e3404, 0x24040424, 0x24040c3e, - 0x24041c2c, 0x24041c3e, 0x24042c1c, 0x24042c3e, 0x240c3e24, 0x24141404, 0x24141c3e, 0x24142404, - 0x24143404, 0x24143434, 0x241c043e, 0x241c242c, 0x24240424, 0x24242c0c, 0x24243424, 0x242c142c, - 0x242c241c, 0x242c3e04, 0x243e042c, 0x243e0c04, 0x243e0c14, 0x243e1c04, 0x2c040c14, 0x2c04240c, - 0x2c043e04, 0x2c0c0404, 0x2c0c0434, 0x2c0c1434, 0x2c0c2c2c, 0x2c140c24, 0x2c141c14, 0x2c143e14, - 0x2c1c0414, 0x2c1c2c1c, 0x2c240c04, 0x2c24141c, 0x2c24143e, 0x2c243e14, 0x2c2c0414, 0x2c2c1c0c, - 0x2c342c04, 0x2c3e1424, 0x2c3e2414, 0x34041424, 0x34042424, 0x34042434, 0x34043424, 0x340c140c, - 0x340c340c, 0x34140c3e, 0x34143424, 0x341c1c04, 0x341c1c34, 0x34242424, 0x342c042c, 0x342c2c14, - 0x34341c1c, 0x343e041c, 0x343e140c, 0x3e04041c, 0x3e04042c, 0x3e04043e, 0x3e040c04, 0x3e041c14, - 0x3e042c14, 0x3e0c1434, 0x3e0c2404, 0x3e140c14, 0x3e14242c, 0x3e142c14, 0x3e1c0404, 0x3e1c0c2c, - 0x3e1c1c1c, 0x3e1c3404, 0x3e24140c, 0x3e24240c, 0x3e2c0404, 0x3e2c0414, 0x3e2c1424, 0x3e341c04 -); - -struct iq3_xxs { - d: f16, - qs: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); @@ -1309,82 +671,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(IQ3_XSS) #decl(IQ3_S) - -const iq3s_grid = array( - 0x01010101, 0x01010103, 0x01010105, 0x0101010b, 0x0101010f, 0x01010301, 0x01010303, 0x01010305, - 0x01010309, 0x0101030d, 0x01010501, 0x01010503, 0x0101050b, 0x01010707, 0x01010901, 0x01010905, - 0x0101090b, 0x0101090f, 0x01010b03, 0x01010b07, 0x01010d01, 0x01010d05, 0x01010f03, 0x01010f09, - 0x01010f0f, 0x01030101, 0x01030103, 0x01030105, 0x01030109, 0x01030301, 0x01030303, 0x0103030b, - 0x01030501, 0x01030507, 0x0103050f, 0x01030703, 0x0103070b, 0x01030909, 0x01030d03, 0x01030d0b, - 0x01030f05, 0x01050101, 0x01050103, 0x0105010b, 0x0105010f, 0x01050301, 0x01050307, 0x0105030d, - 0x01050503, 0x0105050b, 0x01050701, 0x01050709, 0x01050905, 0x0105090b, 0x0105090f, 0x01050b03, - 0x01050b07, 0x01050f01, 0x01050f07, 0x01070107, 0x01070303, 0x0107030b, 0x01070501, 0x01070505, - 0x01070703, 0x01070707, 0x0107070d, 0x01070909, 0x01070b01, 0x01070b05, 0x01070d0f, 0x01070f03, - 0x01070f0b, 0x01090101, 0x01090307, 0x0109030f, 0x01090503, 0x01090509, 0x01090705, 0x01090901, - 0x01090907, 0x01090b03, 0x01090f01, 0x010b0105, 0x010b0109, 0x010b0501, 0x010b0505, 0x010b050d, - 0x010b0707, 0x010b0903, 0x010b090b, 0x010b090f, 0x010b0d0d, 0x010b0f07, 0x010d010d, 0x010d0303, - 0x010d0307, 0x010d0703, 0x010d0b05, 0x010d0f03, 0x010f0101, 0x010f0105, 0x010f0109, 0x010f0501, - 0x010f0505, 0x010f050d, 0x010f0707, 0x010f0b01, 0x010f0b09, 0x03010101, 0x03010103, 0x03010105, - 0x03010109, 0x03010301, 0x03010303, 0x03010307, 0x0301030b, 0x0301030f, 0x03010501, 0x03010505, - 0x03010703, 0x03010709, 0x0301070d, 0x03010b09, 0x03010b0d, 0x03010d03, 0x03010f05, 0x03030101, - 0x03030103, 0x03030107, 0x0303010d, 0x03030301, 0x03030309, 0x03030503, 0x03030701, 0x03030707, - 0x03030903, 0x03030b01, 0x03030b05, 0x03030f01, 0x03030f0d, 0x03050101, 0x03050305, 0x0305030b, - 0x0305030f, 0x03050501, 0x03050509, 0x03050705, 0x03050901, 0x03050907, 0x03050b0b, 0x03050d01, - 0x03050f05, 0x03070103, 0x03070109, 0x0307010f, 0x03070301, 0x03070307, 0x03070503, 0x0307050f, - 0x03070701, 0x03070709, 0x03070903, 0x03070d05, 0x03070f01, 0x03090107, 0x0309010b, 0x03090305, - 0x03090309, 0x03090703, 0x03090707, 0x03090905, 0x0309090d, 0x03090b01, 0x03090b09, 0x030b0103, - 0x030b0301, 0x030b0307, 0x030b0503, 0x030b0701, 0x030b0705, 0x030b0b03, 0x030d0501, 0x030d0509, - 0x030d050f, 0x030d0909, 0x030d090d, 0x030f0103, 0x030f0107, 0x030f0301, 0x030f0305, 0x030f0503, - 0x030f070b, 0x030f0903, 0x030f0d05, 0x030f0f01, 0x05010101, 0x05010103, 0x05010107, 0x0501010b, - 0x0501010f, 0x05010301, 0x05010305, 0x05010309, 0x0501030d, 0x05010503, 0x05010507, 0x0501050f, - 0x05010701, 0x05010705, 0x05010903, 0x05010907, 0x0501090b, 0x05010b01, 0x05010b05, 0x05010d0f, - 0x05010f01, 0x05010f07, 0x05010f0b, 0x05030101, 0x05030105, 0x05030301, 0x05030307, 0x0503030f, - 0x05030505, 0x0503050b, 0x05030703, 0x05030709, 0x05030905, 0x05030b03, 0x05050103, 0x05050109, - 0x0505010f, 0x05050503, 0x05050507, 0x05050701, 0x0505070f, 0x05050903, 0x05050b07, 0x05050b0f, - 0x05050f03, 0x05050f09, 0x05070101, 0x05070105, 0x0507010b, 0x05070303, 0x05070505, 0x05070509, - 0x05070703, 0x05070707, 0x05070905, 0x05070b01, 0x05070d0d, 0x05090103, 0x0509010f, 0x05090501, - 0x05090507, 0x05090705, 0x0509070b, 0x05090903, 0x05090f05, 0x05090f0b, 0x050b0109, 0x050b0303, - 0x050b0505, 0x050b070f, 0x050b0901, 0x050b0b07, 0x050b0f01, 0x050d0101, 0x050d0105, 0x050d010f, - 0x050d0503, 0x050d0b0b, 0x050d0d03, 0x050f010b, 0x050f0303, 0x050f050d, 0x050f0701, 0x050f0907, - 0x050f0b01, 0x07010105, 0x07010303, 0x07010307, 0x0701030b, 0x0701030f, 0x07010505, 0x07010703, - 0x07010707, 0x0701070b, 0x07010905, 0x07010909, 0x0701090f, 0x07010b03, 0x07010d07, 0x07010f03, - 0x07030103, 0x07030107, 0x0703010b, 0x07030309, 0x07030503, 0x07030507, 0x07030901, 0x07030d01, - 0x07030f05, 0x07030f0d, 0x07050101, 0x07050305, 0x07050501, 0x07050705, 0x07050709, 0x07050b01, - 0x07070103, 0x07070301, 0x07070309, 0x07070503, 0x07070507, 0x0707050f, 0x07070701, 0x07070903, - 0x07070907, 0x0707090f, 0x07070b0b, 0x07070f07, 0x07090107, 0x07090303, 0x0709030d, 0x07090505, - 0x07090703, 0x07090b05, 0x07090d01, 0x07090d09, 0x070b0103, 0x070b0301, 0x070b0305, 0x070b050b, - 0x070b0705, 0x070b0909, 0x070b0b0d, 0x070b0f07, 0x070d030d, 0x070d0903, 0x070f0103, 0x070f0107, - 0x070f0501, 0x070f0505, 0x070f070b, 0x09010101, 0x09010109, 0x09010305, 0x09010501, 0x09010509, - 0x0901050f, 0x09010705, 0x09010903, 0x09010b01, 0x09010f01, 0x09030105, 0x0903010f, 0x09030303, - 0x09030307, 0x09030505, 0x09030701, 0x0903070b, 0x09030907, 0x09030b03, 0x09030b0b, 0x09050103, - 0x09050107, 0x09050301, 0x0905030b, 0x09050503, 0x09050707, 0x09050901, 0x09050b0f, 0x09050d05, - 0x09050f01, 0x09070109, 0x09070303, 0x09070307, 0x09070501, 0x09070505, 0x09070703, 0x0907070b, - 0x09090101, 0x09090105, 0x09090509, 0x0909070f, 0x09090901, 0x09090f03, 0x090b010b, 0x090b010f, - 0x090b0503, 0x090b0d05, 0x090d0307, 0x090d0709, 0x090d0d01, 0x090f0301, 0x090f030b, 0x090f0701, - 0x090f0907, 0x090f0b03, 0x0b010105, 0x0b010301, 0x0b010309, 0x0b010505, 0x0b010901, 0x0b010909, - 0x0b01090f, 0x0b010b05, 0x0b010d0d, 0x0b010f09, 0x0b030103, 0x0b030107, 0x0b03010b, 0x0b030305, - 0x0b030503, 0x0b030705, 0x0b030f05, 0x0b050101, 0x0b050303, 0x0b050507, 0x0b050701, 0x0b05070d, - 0x0b050b07, 0x0b070105, 0x0b07010f, 0x0b070301, 0x0b07050f, 0x0b070909, 0x0b070b03, 0x0b070d0b, - 0x0b070f07, 0x0b090103, 0x0b090109, 0x0b090501, 0x0b090705, 0x0b09090d, 0x0b0b0305, 0x0b0b050d, - 0x0b0b0b03, 0x0b0b0b07, 0x0b0d0905, 0x0b0f0105, 0x0b0f0109, 0x0b0f0505, 0x0d010303, 0x0d010307, - 0x0d01030b, 0x0d010703, 0x0d010707, 0x0d010d01, 0x0d030101, 0x0d030501, 0x0d03050f, 0x0d030d09, - 0x0d050305, 0x0d050709, 0x0d050905, 0x0d050b0b, 0x0d050d05, 0x0d050f01, 0x0d070101, 0x0d070309, - 0x0d070503, 0x0d070901, 0x0d09050b, 0x0d090907, 0x0d090d05, 0x0d0b0101, 0x0d0b0107, 0x0d0b0709, - 0x0d0b0d01, 0x0d0d010b, 0x0d0d0901, 0x0d0f0303, 0x0d0f0307, 0x0f010101, 0x0f010109, 0x0f01010f, - 0x0f010501, 0x0f010505, 0x0f01070d, 0x0f010901, 0x0f010b09, 0x0f010d05, 0x0f030105, 0x0f030303, - 0x0f030509, 0x0f030907, 0x0f03090b, 0x0f050103, 0x0f050109, 0x0f050301, 0x0f05030d, 0x0f050503, - 0x0f050701, 0x0f050b03, 0x0f070105, 0x0f070705, 0x0f07070b, 0x0f070b07, 0x0f090103, 0x0f09010b, - 0x0f090307, 0x0f090501, 0x0f090b01, 0x0f0b0505, 0x0f0b0905, 0x0f0d0105, 0x0f0d0703, 0x0f0f0101 -); - -struct iq3_s { - d: f16, - qs: array, - qh: array, - signs: array, - scales: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); @@ -1431,151 +717,7 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { } #enddecl(IQ3_S) -#decl(IQ1_TABLE) - -const IQ1_DELTA: f32 = 0.125; - -const iq1_grid = array( - 0xfffdffff, 0xfff7fff0, 0xffccfff5, 0xffdfffc0, 0xffd7ffdd, 0xff30ffd5, 0xff03ff0c, 0xff10ff01, - 0xff7dff7f, 0xff75ff77, 0xff5fff40, 0xff57ff5d, 0xfcf3ff55, 0xfcccfcf0, 0xfcc1fcc3, 0xfcc5fcc4, - 0xfc3cfcd0, 0xfc34fc31, 0xfc00fc0d, 0xfc1cfc05, 0xfc11fc13, 0xfc70fc17, 0xfc43fc4c, 0xfc50fc41, - 0xfdfdfdff, 0xfdf5fdf7, 0xfddffdc0, 0xfdd7fddd, 0xfd30fdd5, 0xfd04fd0c, 0xfd14fd13, 0xfd7dfd7f, - 0xfd75fd77, 0xfd40fd4c, 0xfd5ffd44, 0xfd57fd5d, 0xf3ccfd55, 0xf3c1f3c3, 0xf33cf3d0, 0xf300f334, - 0xf313f305, 0xf34cf310, 0xf350f344, 0xf0f3f0fc, 0xf0f1f0f0, 0xf0c7f0c0, 0xf0d4f0c5, 0xf030f03f, - 0xf00ff035, 0xf003f00c, 0xf001f000, 0xf01ff004, 0xf010f01d, 0xf015f017, 0xf04cf07c, 0xf047f040, - 0xf05cf045, 0xf050f053, 0xf054f051, 0xf1c4f1c3, 0xf133f13c, 0xf10df10f, 0xf107f100, 0xf11cf11f, - 0xf114f111, 0xf14cf170, 0xf144f143, 0xf7fdf7ff, 0xf7f5f7f7, 0xf7dff7c0, 0xf7d7f7dd, 0xf730f7d5, - 0xf701f70c, 0xf77ff710, 0xf777f77d, 0xf740f775, 0xf75df75f, 0xf755f757, 0xf4ccf4f0, 0xf4c4f4c3, - 0xf4d0f4d3, 0xf40ff43c, 0xf400f40c, 0xf413f41c, 0xf44cf414, 0xf441f443, 0xf450f444, 0xf5fdf5ff, - 0xf5f5f5f7, 0xf5dff5c0, 0xf5d7f5dd, 0xf530f5d5, 0xf504f50c, 0xf510f51c, 0xf57df57f, 0xf577f570, - 0xf540f575, 0xf55df55f, 0xf555f557, 0xcfcccfcf, 0xcfc4cfc3, 0xcfd0cfd3, 0xcf33cf3c, 0xcf00cf0f, - 0xcf1ccf07, 0xcf10cf13, 0xcf4ccf14, 0xcf41cf43, 0xcf50cf5c, 0xccf3ccfc, 0xccf4ccf1, 0xcccdcccf, - 0xccc7ccc0, 0xccd3ccdc, 0xcc30ccd4, 0xcc0fcc35, 0xcc0dcc0c, 0xcc00cc03, 0xcc04cc01, 0xcc10cc1f, - 0xcc4dcc73, 0xcc5ccc40, 0xcdcccc53, 0xcdc1cdc3, 0xcd3fcdd0, 0xcd34cd31, 0xcd00cd0d, 0xcd05cd07, - 0xcd11cd13, 0xcd4ccd70, 0xcd41cd43, 0xc3fccd50, 0xc3f4c3f1, 0xc3c0c3c3, 0xc3c4c3c7, 0xc3d1c3dc, - 0xc330c33c, 0xc337c331, 0xc30cc335, 0xc300c303, 0xc304c301, 0xc310c31d, 0xc373c317, 0xc34fc374, - 0xc340c343, 0xc344c347, 0xc35cc345, 0xc350c353, 0xc0fdc354, 0xc0f5c0f0, 0xc0c3c0cc, 0xc0c1c0c0, - 0xc0dfc0c4, 0xc0d0c0dd, 0xc0d5c0d7, 0xc033c03c, 0xc031c030, 0xc00dc00c, 0xc000c003, 0xc004c001, - 0xc01cc005, 0xc010c013, 0xc014c011, 0xc07dc07f, 0xc070c073, 0xc075c077, 0xc04cc04f, 0xc040c043, - 0xc044c041, 0xc05fc045, 0xc050c05d, 0xc1f3c1fc, 0xc1f1c1f0, 0xc1c1c1c0, 0xc1c5c1c7, 0xc1d1c1dc, - 0xc13dc13f, 0xc130c133, 0xc135c137, 0xc100c10c, 0xc107c101, 0xc11cc104, 0xc110c113, 0xc114c117, - 0xc171c115, 0xc14dc175, 0xc153c140, 0xc7ccc154, 0xc7d0c7c1, 0xc733c73c, 0xc734c731, 0xc700c70f, - 0xc705c707, 0xc71cc71f, 0xc711c713, 0xc770c714, 0xc743c74c, 0xc4cfc750, 0xc4c0c4cd, 0xc4dcc4c5, - 0xc43dc4d0, 0xc430c433, 0xc40cc437, 0xc400c403, 0xc404c401, 0xc41fc405, 0xc415c410, 0xc44cc474, - 0xc440c44d, 0xc45cc447, 0xc454c451, 0xc5c1c5f4, 0xc5d1c5d3, 0xc531c533, 0xc50fc534, 0xc500c50d, - 0xc51cc507, 0xc514c511, 0xc54cc570, 0xc545c541, 0xdffddfff, 0xdff5dff7, 0xdfdfdfc0, 0xdfd0dfdd, - 0xdfd5dfd7, 0xdf0cdf30, 0xdf1cdf04, 0xdf7fdf10, 0xdf77df7d, 0xdf40df75, 0xdf5ddf5f, 0xdf57df50, - 0xdcf0df55, 0xdcc3dccc, 0xdcd0dcc4, 0xdc33dc3d, 0xdc00dc34, 0xdc05dc07, 0xdc13dc1c, 0xdc11dc10, - 0xdc4fdc70, 0xdc44dc41, 0xddfcdc50, 0xddf5ddf7, 0xddc0ddcc, 0xdddddddf, 0xddd5ddd7, 0xdd0cdd30, - 0xdd04dd01, 0xdd7cdd10, 0xdd75dd77, 0xdd40dd4c, 0xdd5ddd5f, 0xdd55dd57, 0xd3c3d3f0, 0xd3c4d3c1, - 0xd333d3d0, 0xd331d330, 0xd30dd334, 0xd307d300, 0xd311d305, 0xd34cd370, 0xd344d343, 0xd350d35c, - 0xd0c0d0f4, 0xd0d4d0dc, 0xd030d03f, 0xd00cd037, 0xd000d003, 0xd01dd004, 0xd017d010, 0xd04fd074, - 0xd040d043, 0xd045d047, 0xd053d05c, 0xd054d051, 0xd1cfd1f0, 0xd1c4d1cd, 0xd13cd1d0, 0xd100d134, - 0xd11cd11f, 0xd173d114, 0xd14fd171, 0xd7ffd145, 0xd7f7d7fd, 0xd7c0d7f5, 0xd7ddd7df, 0xd7d5d7d7, - 0xd70cd730, 0xd710d703, 0xd77dd77f, 0xd775d777, 0xd75dd75f, 0xd755d757, 0xd4ccd4f4, 0xd4c4d4c3, - 0xd431d4d0, 0xd40dd434, 0xd41cd400, 0xd411d413, 0xd470d414, 0xd441d44f, 0xd453d444, 0xd5ffd450, - 0xd5f7d5fd, 0xd5dfd5f5, 0xd5d7d5dd, 0xd530d5d5, 0xd501d50c, 0xd510d504, 0xd57dd57f, 0xd575d577, - 0xd55fd540, 0xd557d55d, 0x3ff0d555, 0x3fc13fcc, 0x3f343fd0, 0x3f003f0d, 0x3f053f07, 0x3f133f1c, - 0x3f433f11, 0x3f5c3f44, 0x3cff3f51, 0x3cf33cfc, 0x3cf43cf1, 0x3cc03ccd, 0x3cc73cc1, 0x3cdc3cc5, - 0x3cd43cd1, 0x3c373c30, 0x3c0c3c35, 0x3c003c03, 0x3c043c01, 0x3c103c05, 0x3c153c17, 0x3c733c7c, - 0x3c4f3c71, 0x3c403c4d, 0x3c5c3c5f, 0x3df03c5d, 0x3dc33dcc, 0x3dd03dc1, 0x3d0d3d3c, 0x3d053d00, - 0x3d143d13, 0x3d433d74, 0x33fc3d50, 0x33c433c0, 0x333033d4, 0x33353337, 0x3303330c, 0x33013300, - 0x331d331c, 0x33173310, 0x337c3315, 0x33743371, 0x334d334f, 0x335f3340, 0x3354335c, 0x30fd30fc, - 0x30f530f0, 0x30c330cc, 0x30c130c0, 0x30df30c4, 0x30d530d0, 0x3033303c, 0x30313030, 0x300f3034, - 0x3003300c, 0x30013000, 0x30043007, 0x3013301c, 0x30113010, 0x307d3014, 0x30703073, 0x304c3077, - 0x30403043, 0x30443041, 0x30503045, 0x30553057, 0x31f031fc, 0x31c331f4, 0x31c731c0, 0x31dc31c5, - 0x31d431d3, 0x313d313f, 0x31373130, 0x310c310f, 0x3100310d, 0x31043101, 0x3110311d, 0x317c3117, - 0x31753170, 0x31403143, 0x3153315c, 0x37f03151, 0x37c037cc, 0x37d037c5, 0x3734373d, 0x3700370f, - 0x371c3707, 0x37113713, 0x37703714, 0x3743374c, 0x37443741, 0x34fc3750, 0x34f134f0, 0x34cf34f5, - 0x34c034c3, 0x34dc34c7, 0x34d134d3, 0x3430343f, 0x340c3435, 0x3403340d, 0x34013400, 0x341f3404, - 0x3410341d, 0x34153411, 0x34743471, 0x3440344d, 0x34473441, 0x3453345c, 0x34543451, 0x353335c1, - 0x35343531, 0x35073500, 0x35133505, 0x35433514, 0x0ffc3550, 0x0ff00ff3, 0x0ff40ff1, 0x0fc00fcd, - 0x0fdc0fc5, 0x0fd40fd3, 0x0f300f3f, 0x0f0c0f37, 0x0f000f03, 0x0f040f01, 0x0f170f10, 0x0f740f71, - 0x0f470f40, 0x0f5c0f5f, 0x0f540f51, 0x0cf70cf0, 0x0cf50cf4, 0x0cc30ccc, 0x0cc10cc0, 0x0cc40cc7, - 0x0cd00cdf, 0x0cd70cd1, 0x0c3c0cd5, 0x0c300c33, 0x0c340c31, 0x0c0c0c0f, 0x0c030c0d, 0x0c010c00, - 0x0c040c07, 0x0c1c0c05, 0x0c100c13, 0x0c140c11, 0x0c700c7d, 0x0c430c4c, 0x0c410c40, 0x0c5f0c44, - 0x0c550c50, 0x0df10dfc, 0x0dc00dcd, 0x0ddc0dc5, 0x0d3d0dd3, 0x0d350d30, 0x0d030d0c, 0x0d010d00, - 0x0d1d0d04, 0x0d700d10, 0x0d4d0d4f, 0x0d440d40, 0x0d530d45, 0x03f003f3, 0x03c303cc, 0x03c103c0, - 0x03c403c7, 0x03d003dc, 0x03d503d7, 0x0333033c, 0x03310330, 0x03350334, 0x030c030f, 0x03000303, - 0x03070301, 0x03050304, 0x031d031c, 0x03100313, 0x03140311, 0x0377037f, 0x034c0375, 0x03400343, - 0x03440341, 0x0353035c, 0x03550350, 0x00fd00fc, 0x00f000f3, 0x00f400f1, 0x00cc00cf, 0x00c300cd, - 0x00c100c0, 0x00c500c4, 0x00d300dc, 0x00d100d0, 0x003f00d4, 0x003d003c, 0x00300033, 0x00370031, - 0x000f0034, 0x000d000c, 0x00000003, 0x00070001, 0x00050004, 0x001c001f, 0x00100013, 0x00170011, - 0x00150014, 0x0073007c, 0x00740070, 0x004f0075, 0x0043004c, 0x00410040, 0x00440047, 0x0053005c, - 0x00510050, 0x01ff0054, 0x01fd01fc, 0x01f101f3, 0x01f401f7, 0x01c301cc, 0x01c701c0, 0x01df01c4, - 0x01dd01dc, 0x01d001d3, 0x01d701d1, 0x013c01d4, 0x01310130, 0x01340137, 0x010f0135, 0x010d010c, - 0x01000103, 0x01070101, 0x01050104, 0x0113011c, 0x01140110, 0x0170017d, 0x01770171, 0x01750174, - 0x0140014c, 0x015d0145, 0x01510150, 0x01540157, 0x07f007f3, 0x07f407f1, 0x07c007cf, 0x07dc07c7, - 0x073007d5, 0x07350737, 0x0703070c, 0x07010700, 0x07040707, 0x071d071f, 0x07100713, 0x0774077d, - 0x074d074f, 0x07470740, 0x0754075c, 0x04fd04fc, 0x04f504f0, 0x04c304cc, 0x04c104c0, 0x04d004c4, - 0x0433043c, 0x04310430, 0x040f0434, 0x040d040c, 0x04000403, 0x04070401, 0x04050404, 0x0413041c, - 0x04110410, 0x047c0414, 0x04740470, 0x0443044c, 0x04410440, 0x04440447, 0x05f30450, 0x05c005f7, - 0x05df05c5, 0x05d105d0, 0x053005d4, 0x05340537, 0x0500050c, 0x05070501, 0x051d0504, 0x05170510, - 0x057c0515, 0x054d0575, 0x05410540, 0x05450547, 0x1ff0055c, 0x1fc11fc3, 0x1fd01fc4, 0x1f0f1f33, - 0x1f011f00, 0x1f051f07, 0x1f131f1c, 0x1f141f11, 0x1f411f7c, 0x1cfc1f50, 0x1cf11cf3, 0x1ccd1cf4, - 0x1cdc1cc0, 0x1cd11cdd, 0x1c301cd4, 0x1c0c1c34, 0x1c011c00, 0x1c101c04, 0x1c151c11, 0x1c751c73, - 0x1c401c4d, 0x1c511c5c, 0x1dcc1c54, 0x1dc41dc1, 0x1d3c1d3f, 0x1d001d31, 0x1d071d01, 0x1d701d1f, - 0x1d411d4c, 0x13cc1d50, 0x13c013cd, 0x13c513c1, 0x13d113dc, 0x133f13d4, 0x1330133d, 0x13351337, - 0x1303130c, 0x13011300, 0x13051304, 0x131d131f, 0x13731310, 0x13741370, 0x134d134f, 0x13401343, - 0x13471341, 0x135c1345, 0x13541353, 0x10f710f0, 0x10cc10f5, 0x10c110c0, 0x103310c4, 0x10311030, - 0x100f1034, 0x1003100c, 0x10011000, 0x101c1004, 0x10101013, 0x10141011, 0x10741071, 0x104c1075, - 0x10411040, 0x10451044, 0x1050105d, 0x10571051, 0x11f411fd, 0x11df11c0, 0x11d711d1, 0x113f11d4, - 0x11371130, 0x110c1135, 0x11001103, 0x11071101, 0x111f1105, 0x11171110, 0x117d117f, 0x11751170, - 0x11411143, 0x11441147, 0x1153115f, 0x11551151, 0x17c417c1, 0x173c17d0, 0x1700170d, 0x171c1705, - 0x17701714, 0x1747174c, 0x14fc1751, 0x14cf14f3, 0x14dc14c0, 0x14d114d3, 0x143f14d4, 0x1430143c, - 0x14371431, 0x1403140c, 0x14011400, 0x141f1404, 0x14151410, 0x1473147d, 0x14401475, 0x1453145c, - 0x14541450, 0x15c115cc, 0x153c15c7, 0x15341533, 0x1500150f, 0x15051507, 0x15101513, 0x15711514, - 0x15471543, 0x15511545, 0x7ffd7fff, 0x7ff57ff7, 0x7fdd7fdf, 0x7fd57fd7, 0x7f0f7f30, 0x7f037f0c, - 0x7f047f01, 0x7f7f7f10, 0x7f777f7d, 0x7f407f75, 0x7f5d7f5f, 0x7f557f57, 0x7ccc7cf0, 0x7cc17cc3, - 0x7cd07cc4, 0x7c337c3c, 0x7c0f7c34, 0x7c007c0d, 0x7c077c01, 0x7c137c04, 0x7c147c11, 0x7c747c70, - 0x7c417c43, 0x7c507c44, 0x7dfd7dff, 0x7df57df7, 0x7ddf7dc0, 0x7dd77ddd, 0x7d0c7dd5, 0x7d047d03, - 0x7d7f7d10, 0x7d777d7d, 0x7d407d75, 0x7d5d7d5f, 0x7d557d57, 0x73c473c3, 0x7333733c, 0x7300730c, - 0x731c7305, 0x73147313, 0x73447343, 0x70f470fc, 0x70c070cd, 0x70d170c5, 0x703f70d4, 0x7030703c, - 0x700c7037, 0x70007003, 0x70047001, 0x70107005, 0x70177011, 0x707c7015, 0x70717073, 0x704f7074, - 0x7040704d, 0x70517047, 0x71c171cc, 0x71d071c4, 0x7133713c, 0x71357134, 0x7100710f, 0x71057104, - 0x7111711c, 0x71707115, 0x7145714c, 0x77ff7153, 0x77f777fd, 0x77c077f5, 0x77dd77df, 0x77d577d7, - 0x7730773c, 0x7703770c, 0x77107704, 0x777f7714, 0x7777777d, 0x77407775, 0x775d775f, 0x77557757, - 0x74f174f0, 0x74c374cc, 0x74d074c1, 0x7433743c, 0x74347431, 0x740d740f, 0x74057400, 0x7413741c, - 0x74417470, 0x74507444, 0x75fd75ff, 0x75f575f7, 0x75df75c0, 0x75d775dd, 0x753075d5, 0x7503750c, - 0x757f7501, 0x7577757d, 0x75407575, 0x755d755f, 0x75557557, 0x4fcc4ff0, 0x4fc74fc1, 0x4fd04fc4, - 0x4f314f3c, 0x4f004f34, 0x4f054f07, 0x4f154f14, 0x4f4c4f70, 0x4f414f43, 0x4f504f44, 0x4cf34cfc, - 0x4cf44cf1, 0x4cc04ccf, 0x4cc54cc7, 0x4cd34cdc, 0x4cd44cd1, 0x4c304c3f, 0x4c0c4c0f, 0x4c004c03, - 0x4c044c01, 0x4c104c1d, 0x4c714c73, 0x4c404c4d, 0x4c5c4c47, 0x4c514c53, 0x4df04c54, 0x4dc34dcc, - 0x4dd04dc4, 0x4d314d33, 0x4d0f4d34, 0x4d004d0d, 0x4d114d07, 0x4d704d14, 0x4d414d43, 0x43fc4d54, - 0x43f143f3, 0x43c043cf, 0x43d143c7, 0x4335433f, 0x4303430c, 0x43014300, 0x43044307, 0x431c431f, - 0x4310431d, 0x43714373, 0x4343434d, 0x43474340, 0x4354435c, 0x40f040ff, 0x40f540f7, 0x40cc40cf, - 0x40c040c3, 0x40c440c1, 0x40d040dc, 0x40d540d4, 0x4033403c, 0x40314030, 0x400f4034, 0x400d400c, - 0x40004003, 0x40074001, 0x40054004, 0x4013401c, 0x40114010, 0x407c4014, 0x40774070, 0x404d404c, - 0x40404043, 0x40444041, 0x405f4045, 0x4050405d, 0x40554057, 0x41f341fc, 0x41c041cf, 0x41df41c4, - 0x41d441d1, 0x41374130, 0x410c4134, 0x4100410d, 0x41044101, 0x41174110, 0x4173417d, 0x41754174, - 0x4143414d, 0x41534140, 0x41544151, 0x47c147f0, 0x47d047c4, 0x4731473c, 0x470d470f, 0x47014700, - 0x47134705, 0x47704710, 0x4741474c, 0x47504744, 0x44f144f3, 0x44cf44f4, 0x44c044cd, 0x44c544c7, - 0x44dc44df, 0x44d144d3, 0x443d443f, 0x44374430, 0x440c4435, 0x44004403, 0x44044401, 0x4410441d, - 0x44154411, 0x4473447c, 0x444d444f, 0x44454440, 0x4451445c, 0x45c045f0, 0x453345d0, 0x45344531, - 0x4500450f, 0x451c4507, 0x454c4570, 0x45404543, 0x5fff4541, 0x5ff75ffd, 0x5fc05ff5, 0x5fdd5fdf, - 0x5fd55fd7, 0x5f0c5f30, 0x5f015f03, 0x5f7f5f04, 0x5f775f7d, 0x5f405f75, 0x5f5d5f5f, 0x5f555f57, - 0x5cf45cf0, 0x5cc35ccc, 0x5cc45cc1, 0x5c315cc5, 0x5c0c5c34, 0x5c075c00, 0x5c1c5c05, 0x5c705c13, - 0x5c4d5c4f, 0x5c445c41, 0x5df75dfd, 0x5dcf5df5, 0x5ddd5dc4, 0x5dd55dd7, 0x5d0c5d30, 0x5d045d01, - 0x5d7f5d10, 0x5d775d7d, 0x5d405d75, 0x5d5d5d5f, 0x5d555d57, 0x53d053c4, 0x5333533c, 0x5303530f, - 0x53075300, 0x531c5305, 0x53115310, 0x53145317, 0x50f15370, 0x50cf50f4, 0x50c050cd, 0x50d150c7, - 0x503d50d4, 0x500c5030, 0x50005003, 0x50045001, 0x50155010, 0x5073507c, 0x50715070, 0x504d5074, - 0x50475040, 0x51cc51f0, 0x51c551c1, 0x51d051dc, 0x51315133, 0x510d5135, 0x51015100, 0x511f5107, - 0x5171511d, 0x5140514f, 0x51445141, 0x5153515c, 0x57ff5151, 0x57f757fd, 0x57df57f5, 0x57d757dd, - 0x570c57d5, 0x57015703, 0x577f5704, 0x5777577d, 0x57405775, 0x575d575f, 0x57555757, 0x54c354f0, - 0x54dc54c4, 0x543c54d0, 0x5400540f, 0x541c5405, 0x54145411, 0x5441544f, 0x55fd55ff, 0x55f555f7, - 0x55dd55df, 0x55d555d7, 0x5503550c, 0x557f5501, 0x5577557d, 0x55405575, 0x555d555f, 0x55555557 -); - -#enddecl(IQ1_TABLE) - #decl(IQ1_S) - -struct iq1_s { - d: f16, - qs: array, - qh: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); @@ -1603,13 +745,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(IQ1_S) #decl(IQ1_M) - -struct iq1_m { - qs: array, - qh: array, - scales: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; @@ -1655,21 +790,7 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(IQ1_M) -#decl(IQ4_TABLE) - -const kvalues_iq4nl = array( - -127, -104, -83, -65, -49, -35, -22, -10, 1, 13, 25, 38, 53, 69, 89, 113 -); - -#enddecl(IQ4_TABLE) - #decl(IQ4_NL) - -struct iq4_nl { - d: f16, - qs: array, -} - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); @@ -1691,14 +812,6 @@ fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { #enddecl(IQ4_NL) #decl(IQ4_XS) - -struct iq4_xs { - d: f16, - scales_h: f16, - scales_l: u32, - qs: array -}; - fn multiply_add(src0_idx_base: u32, src1_idx_base: u32, offset: u32) -> f32 { let block = src0[src0_idx_base + offset]; let d = f32(block.d); diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/rms_norm.wgsl b/ggml/src/ggml-webgpu/wgsl-shaders/rms_norm.wgsl new file mode 100644 index 0000000000..f919a51336 --- /dev/null +++ b/ggml/src/ggml-webgpu/wgsl-shaders/rms_norm.wgsl @@ -0,0 +1,57 @@ +@group(0) @binding(0) +var src: array; + +@group(0) @binding(1) +var dst: array; + +struct Params { + offset_src: u32, // in elements + offset_dst: u32, // in elements + + // Strides (in elements) + stride_src1: u32, + stride_src2: u32, + stride_src3: u32, + + stride_dst1: u32, + stride_dst2: u32, + stride_dst3: u32, + + // Shape of src/dst + ne0: u32, + ne1: u32, + ne2: u32, + ne3: u32, + + eps: u32 +}; + +@group(0) @binding(2) +var params: Params; + +override wg_size: u32; +@compute @workgroup_size(wg_size) +fn main(@builtin(global_invocation_id) gid: vec3) { + if (gid.x >= params.ne1 * params.ne2 * params.ne3) { + return; + } + + // one thread per row + var i = gid.x; + let i3 = i / (params.ne2 * params.ne1); + i = i % (params.ne2 * params.ne1); + let i2 = i / params.ne1; + let i1 = i % params.ne1; + let i_src_row = params.offset_src + i3 * params.stride_src3 + i2 * params.stride_src2 + i1 * params.stride_src1; + let i_dst_row = params.offset_src + i3 * params.stride_dst3 + i2 * params.stride_dst2 + i1 * params.stride_dst1; + + var sum = 0.0f; + for (var j: u32 = 0; j < params.ne0; j++) { + sum += src[i_src_row + j] * src[i_src_row + j]; + } + let eps = bitcast(params.eps); + let scale = 1.0/sqrt(sum/f32(params.ne0) + eps); + for (var j: u32 = 0; j < params.ne0; j++) { + dst[i_dst_row + j] = scale * src[i_src_row + j]; + } +} diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/rms_norm_in_place.wgsl b/ggml/src/ggml-webgpu/wgsl-shaders/rms_norm_in_place.wgsl new file mode 100644 index 0000000000..ae84f556d6 --- /dev/null +++ b/ggml/src/ggml-webgpu/wgsl-shaders/rms_norm_in_place.wgsl @@ -0,0 +1,48 @@ +@group(0) @binding(0) +var a: array; + +struct Params { + offset: u32, // in elements + + // Strides (in elements) + stride1: u32, + stride2: u32, + stride3: u32, + + // Shape + ne0: u32, + ne1: u32, + ne2: u32, + ne3: u32, + + eps: u32 +}; + +@group(0) @binding(1) +var params: Params; + +override wg_size: u32; +@compute @workgroup_size(wg_size) +fn main(@builtin(global_invocation_id) gid: vec3) { + if (gid.x >= params.ne1 * params.ne2 * params.ne3) { + return; + } + + // one thread per row + var i = gid.x; + let i3 = i / (params.ne2 * params.ne1); + i = i % (params.ne2 * params.ne1); + let i2 = i / params.ne1; + let i1 = i % params.ne1; + let i_row = params.offset + i3 * params.stride3 + i2 * params.stride2 + i1 * params.stride1; + + var sum = 0.0f; + for (var j: u32 = 0; j < params.ne0; j++) { + sum += a[i_row + j] * a[i_row + j]; + } + let eps = bitcast(params.eps); + let scale = 1.0/sqrt(sum/f32(params.ne0) + eps); + for (var j: u32 = 0; j < params.ne0; j++) { + a[i_row + j] = scale * a[i_row + j]; + } +} diff --git a/ggml/src/ggml-webgpu/wgsl-shaders/set_rows.wgsl b/ggml/src/ggml-webgpu/wgsl-shaders/set_rows.wgsl index 4bd6f94a23..3567713dc2 100644 --- a/ggml/src/ggml-webgpu/wgsl-shaders/set_rows.wgsl +++ b/ggml/src/ggml-webgpu/wgsl-shaders/set_rows.wgsl @@ -52,7 +52,6 @@ fn main(@builtin(global_invocation_id) gid: vec3) { } var i = gid.x; let i_src3 = i / (params.ne2 * params.n_rows); - let i_dst3 = i / (params.ne2 * 3); i = i % (params.ne2 * params.n_rows); let i_src2 = i / params.n_rows; diff --git a/src/llama-cparams.h b/src/llama-cparams.h index dbbaba9f62..eae7b839f4 100644 --- a/src/llama-cparams.h +++ b/src/llama-cparams.h @@ -4,7 +4,7 @@ #include -#define LLAMA_MAX_SEQ 64 +#define LLAMA_MAX_SEQ 256 struct llama_cparams { uint32_t n_ctx; // context size used during inference diff --git a/src/llama-hparams.h b/src/llama-hparams.h index 116d728e8c..202cbbd1b2 100644 --- a/src/llama-hparams.h +++ b/src/llama-hparams.h @@ -149,7 +149,7 @@ struct llama_hparams { bool causal_attn = true; bool use_alibi = false; bool attn_soft_cap = false; - bool use_kq_norm = true; + bool use_kq_norm = false; // for Classifiers uint32_t n_cls_out = 1; diff --git a/src/llama-model.cpp b/src/llama-model.cpp index 009e03625e..b4fb644918 100644 --- a/src/llama-model.cpp +++ b/src/llama-model.cpp @@ -36,6 +36,7 @@ const char * llm_type_name(llm_type type) { case LLM_TYPE_80M: return "80M"; case LLM_TYPE_109M: return "109M"; case LLM_TYPE_137M: return "137M"; + case LLM_TYPE_140M: return "140M"; case LLM_TYPE_160M: return "160M"; case LLM_TYPE_190M: return "190M"; case LLM_TYPE_220M: return "220M"; @@ -44,6 +45,7 @@ const char * llm_type_name(llm_type type) { case LLM_TYPE_270M: return "270M"; case LLM_TYPE_335M: return "335M"; case LLM_TYPE_350M: return "350M"; + case LLM_TYPE_360M: return "360M"; case LLM_TYPE_410M: return "410M"; case LLM_TYPE_450M: return "450M"; case LLM_TYPE_475M: return "475M"; @@ -51,6 +53,7 @@ const char * llm_type_name(llm_type type) { case LLM_TYPE_700M: return "700M"; case LLM_TYPE_770M: return "770M"; case LLM_TYPE_780M: return "780M"; + case LLM_TYPE_950M: return "950M"; case LLM_TYPE_0_3B: return "0.3B"; case LLM_TYPE_0_5B: return "0.5B"; case LLM_TYPE_0_6B: return "0.6B"; @@ -623,19 +626,32 @@ void llama_model::load_hparams(llama_model_loader & ml) { ml.get_key(LLM_KV_EXPERT_FEED_FORWARD_LENGTH, hparams.n_ff_exp); ml.get_key(LLM_KV_INTERLEAVE_MOE_LAYER_STEP, hparams.n_moe_layer_step); - hparams.swa_type = LLAMA_SWA_TYPE_CHUNKED; - hparams.n_swa = 8192; // should this be a gguf kv? currently it's the same for Scout and Maverick - hparams.set_swa_pattern(4); // pattern: 3 chunked - 1 full + const bool found_swa = ml.get_key(LLM_KV_ATTENTION_SLIDING_WINDOW, hparams.n_swa, false); + if (found_swa && hparams.n_swa == 0) { + hparams.swa_type = LLAMA_SWA_TYPE_NONE; + hparams.n_no_rope_layer_step = hparams.n_layer; // always use rope + } else { + hparams.swa_type = LLAMA_SWA_TYPE_CHUNKED; + hparams.n_swa = 8192; + hparams.set_swa_pattern(4); // pattern: 3 chunked - 1 full + } switch (hparams.n_expert) { + case 0: { + // MobileLLM (no MoE) + switch (hparams.n_embd) { + case 2048: type = LLM_TYPE_140M; break; + case 4096: type = LLM_TYPE_360M; break; + case 6144: type = LLM_TYPE_950M; break; + default: type = LLM_TYPE_UNKNOWN; + } + } break; case 16: type = LLM_TYPE_17B_16E; break; case 128: type = LLM_TYPE_17B_128E; break; default: type = LLM_TYPE_UNKNOWN; } - if (type == LLM_TYPE_17B_128E) { - hparams.use_kq_norm = false; - } + hparams.use_kq_norm = type != LLM_TYPE_17B_128E; } break; case LLM_ARCH_ARCEE: { @@ -2548,9 +2564,8 @@ bool llama_model::load_tensors(llama_model_loader & ml) { output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); } - GGML_ASSERT(hparams.n_moe_layer_step > 0 && "Llama 4 requires n_moe_layer_step > 0"); for (int i = 0; i < n_layer; ++i) { - bool is_moe_layer = (i + 1) % hparams.n_moe_layer_step == 0; + bool is_moe_layer = hparams.n_moe_layer_step > 0 && (i + 1) % hparams.n_moe_layer_step == 0; auto & layer = layers[i]; @@ -6423,6 +6438,14 @@ struct llm_build_llama : public llm_graph_context { cb(Kcur, "Kcur", il); cb(Vcur, "Vcur", il); + if (hparams.use_kq_norm) { + // Llama4TextL2Norm + Qcur = ggml_rms_norm(ctx0, Qcur, hparams.f_norm_rms_eps); + Kcur = ggml_rms_norm(ctx0, Kcur, hparams.f_norm_rms_eps); + cb(Qcur, "Qcur_normed", il); + cb(Kcur, "Kcur_normed", il); + } + cur = build_attn(inp_attn, model.layers[il].wo, model.layers[il].bo, Qcur, Kcur, Vcur, nullptr, nullptr, nullptr, kq_scale, il); @@ -6530,7 +6553,8 @@ struct llm_build_llama_iswa : public llm_graph_context { for (int il = 0; il < n_layer; ++il) { ggml_tensor * inpSA = inpL; - const bool use_rope = (il + 1) % hparams.n_no_rope_layer_step != 0; + const bool use_rope = hparams.n_no_rope_layer_step > 0 && + (il + 1) % hparams.n_no_rope_layer_step != 0; // norm cur = build_norm(inpL, @@ -19399,7 +19423,11 @@ ggml_cgraph * llama_model::build_graph(const llm_graph_params & params) const { } break; case LLM_ARCH_LLAMA4: { - llm = std::make_unique(*this, params); + if (hparams.swa_type == LLAMA_SWA_TYPE_NONE) { + llm = std::make_unique(*this, params); + } else { + llm = std::make_unique(*this, params); + } } break; case LLM_ARCH_DECI: { diff --git a/src/llama-model.h b/src/llama-model.h index 62ff8348f3..9b4eb27953 100644 --- a/src/llama-model.h +++ b/src/llama-model.h @@ -28,6 +28,7 @@ enum llm_type { LLM_TYPE_80M, LLM_TYPE_109M, LLM_TYPE_137M, + LLM_TYPE_140M, LLM_TYPE_160M, LLM_TYPE_190M, LLM_TYPE_220M, @@ -36,6 +37,7 @@ enum llm_type { LLM_TYPE_270M, LLM_TYPE_335M, LLM_TYPE_350M, + LLM_TYPE_360M, LLM_TYPE_410M, LLM_TYPE_450M, LLM_TYPE_475M, @@ -43,6 +45,7 @@ enum llm_type { LLM_TYPE_700M, LLM_TYPE_770M, LLM_TYPE_780M, + LLM_TYPE_950M, LLM_TYPE_0_3B, LLM_TYPE_0_5B, LLM_TYPE_0_6B, diff --git a/tests/test-backend-ops.cpp b/tests/test-backend-ops.cpp index b54a1a4e82..893c2af313 100644 --- a/tests/test-backend-ops.cpp +++ b/tests/test-backend-ops.cpp @@ -6071,6 +6071,10 @@ static std::vector> make_test_cases_eval() { //add_test_bin_bcast(type, {3, 3, 2560, 1280}, {2, 1, 1, 1}); } + // single in-place tests, especially important for WebGPU backend since kernels for in-place vs. not are different + test_cases.emplace_back(new test_bin_bcast(ggml_add_inplace, GGML_TYPE_F32, {16, 5, 4, 3}, {1, 1, 1, 1}, 16)); + test_cases.emplace_back(new test_bin_bcast(ggml_mul_inplace, GGML_TYPE_F32, {16, 5, 4, 3}, {1, 1, 1, 1}, 16)); + // fusion test_cases.emplace_back(new test_bin_bcast(ggml_add, GGML_TYPE_F32, {10, 5, 4, 3}, {2, 1, 1, 1}, 2)); test_cases.emplace_back(new test_bin_bcast(ggml_add, GGML_TYPE_F32, {16, 5, 4, 3}, {1, 2, 1, 1}, 3)); @@ -6325,12 +6329,20 @@ static std::vector> make_test_cases_eval() { } for (ggml_type type : {GGML_TYPE_F16, GGML_TYPE_F32}) { - test_cases.emplace_back(new test_sqr(type)); - test_cases.emplace_back(new test_sqrt(type)); - test_cases.emplace_back(new test_log(type)); - test_cases.emplace_back(new test_sin(type)); - test_cases.emplace_back(new test_cos(type)); - test_cases.emplace_back(new test_clamp(type)); + test_cases.emplace_back(new test_sqr (type)); + test_cases.emplace_back(new test_sqrt (type)); + test_cases.emplace_back(new test_log (type)); + test_cases.emplace_back(new test_sin (type)); + test_cases.emplace_back(new test_cos (type)); + test_cases.emplace_back(new test_clamp (type)); + test_cases.emplace_back(new test_leaky_relu(type)); + test_cases.emplace_back(new test_sqr (type, {7, 1, 5, 3})); + test_cases.emplace_back(new test_sqrt (type, {7, 1, 5, 3})); + test_cases.emplace_back(new test_log (type, {7, 1, 5, 3})); + test_cases.emplace_back(new test_sin (type, {7, 1, 5, 3})); + test_cases.emplace_back(new test_cos (type, {7, 1, 5, 3})); + test_cases.emplace_back(new test_clamp (type, {7, 1, 5, 3})); + test_cases.emplace_back(new test_leaky_relu(type, {7, 1, 5, 3})); } test_cases.emplace_back(new test_diag_mask_inf(GGML_TYPE_F32, {10, 10, 1, 1}, 5)); diff --git a/tools/server/public/index.html.gz b/tools/server/public/index.html.gz index 0e3e1f9834..909fa1cceb 100644 Binary files a/tools/server/public/index.html.gz and b/tools/server/public/index.html.gz differ diff --git a/tools/server/server.cpp b/tools/server/server.cpp index 519704fad7..c0ffe17692 100644 --- a/tools/server/server.cpp +++ b/tools/server/server.cpp @@ -111,6 +111,7 @@ static bool server_task_type_need_logits(server_task_type task_type) { struct slot_params { bool stream = true; + bool include_usage = false; bool cache_prompt = true; // remember the prompt to avoid reprocessing all prompt bool return_tokens = false; bool return_progress = false; @@ -310,17 +311,19 @@ struct server_task { params.verbose = params_base.verbosity > 9; params.timings_per_token = json_value(data, "timings_per_token", false); - params.stream = json_value(data, "stream", false); - params.cache_prompt = json_value(data, "cache_prompt", true); - params.return_tokens = json_value(data, "return_tokens", false); - params.return_progress = json_value(data, "return_progress", false); - params.n_predict = json_value(data, "n_predict", json_value(data, "max_tokens", defaults.n_predict)); - params.n_indent = json_value(data, "n_indent", defaults.n_indent); - params.n_keep = json_value(data, "n_keep", defaults.n_keep); - params.n_discard = json_value(data, "n_discard", defaults.n_discard); - //params.t_max_prompt_ms = json_value(data, "t_max_prompt_ms", defaults.t_max_prompt_ms); // TODO: implement - params.t_max_predict_ms = json_value(data, "t_max_predict_ms", defaults.t_max_predict_ms); - params.response_fields = json_value(data, "response_fields", std::vector()); + params.stream = json_value(data, "stream", false); + auto stream_opt = json_value(data, "stream_options", json::object()); + params.include_usage = json_value(stream_opt, "include_usage", false); + params.cache_prompt = json_value(data, "cache_prompt", true); + params.return_tokens = json_value(data, "return_tokens", false); + params.return_progress = json_value(data, "return_progress", false); + params.n_predict = json_value(data, "n_predict", json_value(data, "max_tokens", defaults.n_predict)); + params.n_indent = json_value(data, "n_indent", defaults.n_indent); + params.n_keep = json_value(data, "n_keep", defaults.n_keep); + params.n_discard = json_value(data, "n_discard", defaults.n_discard); + //params.t_max_prompt_ms = json_value(data, "t_max_prompt_ms", defaults.t_max_prompt_ms); // TODO: implement + params.t_max_predict_ms = json_value(data, "t_max_predict_ms", defaults.t_max_predict_ms); + params.response_fields = json_value(data, "response_fields", std::vector()); params.sampling.top_k = json_value(data, "top_k", defaults.sampling.top_k); params.sampling.top_p = json_value(data, "top_p", defaults.sampling.top_p); @@ -775,6 +778,7 @@ struct server_task_result_cmpl_final : server_task_result { llama_tokens tokens; bool stream; + bool include_usage; result_timings timings; std::string prompt; @@ -982,21 +986,23 @@ struct server_task_result_cmpl_final : server_task_result { {"object", "chat.completion.chunk"}, }); - // OpenAI API spec for chat.completion.chunks specifies an empty `choices` array for the last chunk when including usage - // https://platform.openai.com/docs/api-reference/chat_streaming/streaming#chat_streaming/streaming-choices - deltas.push_back({ - {"choices", json::array()}, - {"created", t}, - {"id", oaicompat_cmpl_id}, - {"model", oaicompat_model}, - {"system_fingerprint", build_info}, - {"object", "chat.completion.chunk"}, - {"usage", json { - {"completion_tokens", n_decoded}, - {"prompt_tokens", n_prompt_tokens}, - {"total_tokens", n_decoded + n_prompt_tokens}, - }}, - }); + if (include_usage) { + // OpenAI API spec for chat.completion.chunks specifies an empty `choices` array for the last chunk when including usage + // https://platform.openai.com/docs/api-reference/chat_streaming/streaming#chat_streaming/streaming-choices + deltas.push_back({ + {"choices", json::array()}, + {"created", t}, + {"id", oaicompat_cmpl_id}, + {"model", oaicompat_model}, + {"system_fingerprint", build_info}, + {"object", "chat.completion.chunk"}, + {"usage", json { + {"completion_tokens", n_decoded}, + {"prompt_tokens", n_prompt_tokens}, + {"total_tokens", n_decoded + n_prompt_tokens}, + }}, + }); + } if (timings.prompt_n >= 0) { deltas.back().push_back({"timings", timings.to_json()}); @@ -2815,6 +2821,7 @@ struct server_context { res->verbose = slot.params.verbose; res->stream = slot.params.stream; + res->include_usage = slot.params.include_usage; res->oaicompat = slot.params.oaicompat; res->oaicompat_model = slot.params.oaicompat_model; res->oaicompat_cmpl_id = slot.params.oaicompat_cmpl_id; @@ -5261,6 +5268,42 @@ int main(int argc, char ** argv) { svr->Get (params.api_prefix + "/slots", handle_slots); svr->Post(params.api_prefix + "/slots/:id_slot", handle_slots_action); + // SPA fallback route - serve index.html for any route that doesn't match API endpoints + // This enables client-side routing for dynamic routes like /chat/[id] + if (params.webui && params.public_path.empty()) { + // Only add fallback when using embedded static files + svr->Get(".*", [](const httplib::Request & req, httplib::Response & res) { + // Skip API routes - they should have been handled above + if (req.path.find("/v1/") != std::string::npos || + req.path.find("/health") != std::string::npos || + req.path.find("/metrics") != std::string::npos || + req.path.find("/props") != std::string::npos || + req.path.find("/models") != std::string::npos || + req.path.find("/api/tags") != std::string::npos || + req.path.find("/completions") != std::string::npos || + req.path.find("/chat/completions") != std::string::npos || + req.path.find("/embeddings") != std::string::npos || + req.path.find("/tokenize") != std::string::npos || + req.path.find("/detokenize") != std::string::npos || + req.path.find("/lora-adapters") != std::string::npos || + req.path.find("/slots") != std::string::npos) { + return false; // Let other handlers process API routes + } + + // Serve index.html for all other routes (SPA fallback) + if (req.get_header_value("Accept-Encoding").find("gzip") == std::string::npos) { + res.set_content("Error: gzip is not supported by this browser", "text/plain"); + } else { + res.set_header("Content-Encoding", "gzip"); + // COEP and COOP headers, required by pyodide (python interpreter) + res.set_header("Cross-Origin-Embedder-Policy", "require-corp"); + res.set_header("Cross-Origin-Opener-Policy", "same-origin"); + res.set_content(reinterpret_cast(index_html_gz), index_html_gz_len, "text/html; charset=utf-8"); + } + return false; + }); + } + // // Start the server // diff --git a/tools/server/tests/unit/test_basic.py b/tools/server/tests/unit/test_basic.py index 58ade52be6..829af2ebe7 100644 --- a/tools/server/tests/unit/test_basic.py +++ b/tools/server/tests/unit/test_basic.py @@ -92,7 +92,7 @@ def test_no_webui(): url = f"http://{server.server_host}:{server.server_port}" res = requests.get(url) assert res.status_code == 200 - assert "" in res.text + assert "" in res.text server.stop() # with --no-webui diff --git a/tools/server/tests/unit/test_chat_completion.py b/tools/server/tests/unit/test_chat_completion.py index 53421d1b57..2979ed4bb7 100644 --- a/tools/server/tests/unit/test_chat_completion.py +++ b/tools/server/tests/unit/test_chat_completion.py @@ -271,8 +271,10 @@ def test_chat_completion_with_timings_per_token(): "max_tokens": 10, "messages": [{"role": "user", "content": "test"}], "stream": True, + "stream_options": {"include_usage": True}, "timings_per_token": True, }) + stats_received = False for i, data in enumerate(res): if i == 0: # Check first role message for stream=True @@ -288,6 +290,8 @@ def test_chat_completion_with_timings_per_token(): assert "predicted_per_second" in data["timings"] assert "predicted_n" in data["timings"] assert data["timings"]["predicted_n"] <= 10 + stats_received = True + assert stats_received def test_logprobs(): diff --git a/tools/server/webui/.gitignore b/tools/server/webui/.gitignore index a547bf36d8..cc54bb717f 100644 --- a/tools/server/webui/.gitignore +++ b/tools/server/webui/.gitignore @@ -1,24 +1,27 @@ -# Logs -logs -*.log -npm-debug.log* -yarn-debug.log* -yarn-error.log* -pnpm-debug.log* -lerna-debug.log* - +test-results node_modules -dist -dist-ssr -*.local -# Editor directories and files -.vscode/* -!.vscode/extensions.json -.idea +# Output +.output +.vercel +.netlify +.wrangler +/.svelte-kit +/build + +# OS .DS_Store -*.suo -*.ntvs* -*.njsproj -*.sln -*.sw? +Thumbs.db + +# Env +.env +.env.* +!.env.example +!.env.test + +# Vite +vite.config.js.timestamp-* +vite.config.ts.timestamp-* + +*storybook.log +storybook-static diff --git a/tools/server/webui/.npmrc b/tools/server/webui/.npmrc new file mode 100644 index 0000000000..b6f27f1359 --- /dev/null +++ b/tools/server/webui/.npmrc @@ -0,0 +1 @@ +engine-strict=true diff --git a/tools/server/webui/.prettierignore b/tools/server/webui/.prettierignore index c0cb165b37..7d74fe2468 100644 --- a/tools/server/webui/.prettierignore +++ b/tools/server/webui/.prettierignore @@ -1,10 +1,9 @@ -**/.vscode -**/.github -**/.git -**/.svn -**/.hg -**/node_modules -**/dist -**/build +# Package Managers +package-lock.json +pnpm-lock.yaml +yarn.lock +bun.lock +bun.lockb -*.config.js +# Miscellaneous +/static/ diff --git a/tools/server/webui/.prettierrc b/tools/server/webui/.prettierrc new file mode 100644 index 0000000000..8103a0b5d2 --- /dev/null +++ b/tools/server/webui/.prettierrc @@ -0,0 +1,16 @@ +{ + "useTabs": true, + "singleQuote": true, + "trailingComma": "none", + "printWidth": 100, + "plugins": ["prettier-plugin-svelte", "prettier-plugin-tailwindcss"], + "overrides": [ + { + "files": "*.svelte", + "options": { + "parser": "svelte" + } + } + ], + "tailwindStylesheet": "./src/app.css" +} diff --git a/tools/server/webui/.storybook/ModeWatcherDecorator.svelte b/tools/server/webui/.storybook/ModeWatcherDecorator.svelte new file mode 100644 index 0000000000..8bded8b3f1 --- /dev/null +++ b/tools/server/webui/.storybook/ModeWatcherDecorator.svelte @@ -0,0 +1,36 @@ + + + + +{#if children} + {@const Component = children} + + +{/if} diff --git a/tools/server/webui/.storybook/TooltipProviderDecorator.svelte b/tools/server/webui/.storybook/TooltipProviderDecorator.svelte new file mode 100644 index 0000000000..9aad1eaa4a --- /dev/null +++ b/tools/server/webui/.storybook/TooltipProviderDecorator.svelte @@ -0,0 +1,13 @@ + + + + {@render children()} + diff --git a/tools/server/webui/.storybook/main.ts b/tools/server/webui/.storybook/main.ts new file mode 100644 index 0000000000..7145bcb7eb --- /dev/null +++ b/tools/server/webui/.storybook/main.ts @@ -0,0 +1,17 @@ +import type { StorybookConfig } from '@storybook/sveltekit'; + +const config: StorybookConfig = { + stories: ['../src/**/*.mdx', '../src/**/*.stories.@(js|ts|svelte)'], + addons: [ + '@storybook/addon-svelte-csf', + '@chromatic-com/storybook', + '@storybook/addon-docs', + '@storybook/addon-a11y', + '@storybook/addon-vitest' + ], + framework: { + name: '@storybook/sveltekit', + options: {} + } +}; +export default config; diff --git a/tools/server/webui/.storybook/preview.ts b/tools/server/webui/.storybook/preview.ts new file mode 100644 index 0000000000..fb91386af4 --- /dev/null +++ b/tools/server/webui/.storybook/preview.ts @@ -0,0 +1,34 @@ +import type { Preview } from '@storybook/sveltekit'; +import '../src/app.css'; +import ModeWatcherDecorator from './ModeWatcherDecorator.svelte'; +import TooltipProviderDecorator from './TooltipProviderDecorator.svelte'; + +const preview: Preview = { + parameters: { + controls: { + matchers: { + color: /(background|color)$/i, + date: /Date$/i + } + }, + backgrounds: { + disable: true + } + }, + decorators: [ + (story) => ({ + Component: ModeWatcherDecorator, + props: { + children: story + } + }), + (story) => ({ + Component: TooltipProviderDecorator, + props: { + children: story + } + }) + ] +}; + +export default preview; diff --git a/tools/server/webui/.storybook/vitest.setup.ts b/tools/server/webui/.storybook/vitest.setup.ts new file mode 100644 index 0000000000..e0c1753c84 --- /dev/null +++ b/tools/server/webui/.storybook/vitest.setup.ts @@ -0,0 +1,11 @@ +import { setProjectAnnotations } from '@storybook/sveltekit'; +import * as previewAnnotations from './preview'; +import { beforeAll } from 'vitest'; + +const project = setProjectAnnotations([previewAnnotations]); + +beforeAll(async () => { + if (project.beforeAll) { + await project.beforeAll(); + } +}); diff --git a/tools/server/webui/README.md b/tools/server/webui/README.md new file mode 100644 index 0000000000..9d16f34d86 --- /dev/null +++ b/tools/server/webui/README.md @@ -0,0 +1,66 @@ +# llama.cpp Web UI + +A modern, feature-rich web interface for llama.cpp built with SvelteKit. This UI provides an intuitive chat interface with advanced file handling, conversation management, and comprehensive model interaction capabilities. + +## Features + +- **Modern Chat Interface** - Clean, responsive design with dark/light mode +- **File Attachments** - Support for images, text files, PDFs, and audio with rich previews and drag-and-drop support +- **Conversation Management** - Create, edit, branch, and search conversations +- **Advanced Markdown** - Code highlighting, math formulas (KaTeX), and content blocks +- **Reasoning Content** - Support for models with thinking blocks +- **Keyboard Shortcuts** - Keyboard navigation (Shift+Ctrl/Cmd+O for new chat, Shift+Ctrl/Cmdt+E for edit conversation, Shift+Ctrl/Cmdt+D for delete conversation, Ctrl/Cmd+K for search, Ctrl/Cmd+V for paste, Ctrl/Cmd+B for opening/collapsing sidebar) +- **Request Tracking** - Monitor processing with slots endpoint integration +- **UI Testing** - Storybook component library with automated tests + +## Development + +Install dependencies: + +```bash +npm install +``` + +Start the development server + Storybook: + +```bash +npm run dev +``` + +This will start both the SvelteKit dev server and Storybook on port 6006. + +## Building + +Create a production build: + +```bash +npm run build +``` + +The build outputs static files to `../public` directory for deployment with llama.cpp server. + +## Testing + +Run the test suite: + +```bash +# E2E tests +npm run test:e2e + +# Unit tests +npm run test:unit + +# UI tests +npm run test:ui + +# All tests +npm run test +``` + +## Architecture + +- **Framework**: SvelteKit with Svelte 5 runes +- **Components**: ShadCN UI + bits-ui design system +- **Database**: IndexedDB with Dexie for local storage +- **Build**: Static adapter for deployment with llama.cpp server +- **Testing**: Playwright (E2E) + Vitest (unit) + Storybook (components) diff --git a/tools/server/webui/components.json b/tools/server/webui/components.json new file mode 100644 index 0000000000..224bd70acf --- /dev/null +++ b/tools/server/webui/components.json @@ -0,0 +1,16 @@ +{ + "$schema": "https://shadcn-svelte.com/schema.json", + "tailwind": { + "css": "src/app.css", + "baseColor": "neutral" + }, + "aliases": { + "components": "$lib/components", + "utils": "$lib/components/ui/utils", + "ui": "$lib/components/ui", + "hooks": "$lib/hooks", + "lib": "$lib" + }, + "typescript": true, + "registry": "https://shadcn-svelte.com/registry" +} diff --git a/tools/server/webui/e2e/demo.test.ts b/tools/server/webui/e2e/demo.test.ts new file mode 100644 index 0000000000..9985ce113e --- /dev/null +++ b/tools/server/webui/e2e/demo.test.ts @@ -0,0 +1,6 @@ +import { expect, test } from '@playwright/test'; + +test('home page has expected h1', async ({ page }) => { + await page.goto('/'); + await expect(page.locator('h1')).toBeVisible(); +}); diff --git a/tools/server/webui/eslint.config.js b/tools/server/webui/eslint.config.js index 7c0d39b89b..5baea57f33 100644 --- a/tools/server/webui/eslint.config.js +++ b/tools/server/webui/eslint.config.js @@ -1,26 +1,49 @@ -import js from '@eslint/js' -import globals from 'globals' -import reactHooks from 'eslint-plugin-react-hooks' -import reactRefresh from 'eslint-plugin-react-refresh' -import tseslint from 'typescript-eslint' +// For more info, see https://github.com/storybookjs/eslint-plugin-storybook#configuration-flat-config-format +import storybook from 'eslint-plugin-storybook'; -export default tseslint.config( - { ignores: ['dist'] }, - { - extends: [js.configs.recommended, ...tseslint.configs.recommended], - files: ['**/*.{ts,tsx}'], - languageOptions: { - ecmaVersion: 2020, - globals: globals.browser, - }, - plugins: { - 'react-hooks': reactHooks, - 'react-refresh': reactRefresh, - }, - rules: { - ...reactHooks.configs.recommended.rules, - 'react-refresh/only-export-components': 'off', - '@typescript-eslint/no-unused-vars': 'off', - }, - }, -) +import prettier from 'eslint-config-prettier'; +import { includeIgnoreFile } from '@eslint/compat'; +import js from '@eslint/js'; +import svelte from 'eslint-plugin-svelte'; +import globals from 'globals'; +import { fileURLToPath } from 'node:url'; +import ts from 'typescript-eslint'; +import svelteConfig from './svelte.config.js'; + +const gitignorePath = fileURLToPath(new URL('./.gitignore', import.meta.url)); + +export default ts.config( + includeIgnoreFile(gitignorePath), + js.configs.recommended, + ...ts.configs.recommended, + ...svelte.configs.recommended, + prettier, + ...svelte.configs.prettier, + { + languageOptions: { + globals: { ...globals.browser, ...globals.node } + }, + rules: { + // typescript-eslint strongly recommend that you do not use the no-undef lint rule on TypeScript projects. + // see: https://typescript-eslint.io/troubleshooting/faqs/eslint/#i-get-errors-from-the-no-undef-rule-about-global-variables-not-being-defined-even-though-there-are-no-typescript-errors + 'no-undef': 'off', + 'svelte/no-at-html-tags': 'off' + } + }, + { + files: ['**/*.svelte', '**/*.svelte.ts', '**/*.svelte.js'], + languageOptions: { + parserOptions: { + projectService: true, + extraFileExtensions: ['.svelte'], + parser: ts.parser, + svelteConfig + } + } + }, + { + // Exclude Storybook files from main ESLint rules + ignores: ['.storybook/**/*'] + }, + storybook.configs['flat/recommended'] +); diff --git a/tools/server/webui/index.html b/tools/server/webui/index.html deleted file mode 100644 index 471f46b3ad..0000000000 --- a/tools/server/webui/index.html +++ /dev/null @@ -1,16 +0,0 @@ - - - - - - - πŸ¦™ llama.cpp - chat - - -
- - - diff --git a/tools/server/webui/package-lock.json b/tools/server/webui/package-lock.json index a05cbcfe5c..c300ecaa77 100644 --- a/tools/server/webui/package-lock.json +++ b/tools/server/webui/package-lock.json @@ -1,6620 +1,8486 @@ { - "name": "webui", - "version": "0.0.0", - "lockfileVersion": 3, - "requires": true, - "packages": { - "": { - "name": "webui", - "version": "0.0.0", - "dependencies": { - "@heroicons/react": "^2.2.0", - "@sec-ant/readable-stream": "^0.6.0", - "@tailwindcss/postcss": "^4.1.1", - "@tailwindcss/vite": "^4.1.1", - "@vscode/markdown-it-katex": "^1.1.1", - "autoprefixer": "^10.4.20", - "daisyui": "^5.0.12", - "dexie": "^4.0.11", - "highlight.js": "^11.10.0", - "katex": "^0.16.15", - "pdfjs-dist": "^5.2.133", - "postcss": "^8.4.49", - "react": "^18.3.1", - "react-dom": "^18.3.1", - "react-dropzone": "^14.3.8", - "react-hot-toast": "^2.5.2", - "react-markdown": "^9.0.3", - "react-router": "^7.1.5", - "rehype-highlight": "^7.0.2", - "rehype-katex": "^7.0.1", - "remark-breaks": "^4.0.0", - "remark-gfm": "^4.0.0", - "remark-math": "^6.0.0", - "tailwindcss": "^4.1.1", - "textlinestream": "^1.1.1", - "vite-plugin-singlefile": "^2.0.3" - }, - "devDependencies": { - "@eslint/js": "^9.17.0", - "@types/markdown-it": "^14.1.2", - "@types/node": "^22.13.1", - "@types/react": "^18.3.18", - "@types/react-dom": "^18.3.5", - "@vitejs/plugin-react": "^4.3.4", - "eslint": "^9.17.0", - "eslint-plugin-react-hooks": "^5.0.0", - "eslint-plugin-react-refresh": "^0.4.16", - "fflate": "^0.8.2", - "globals": "^15.14.0", - "prettier": "^3.4.2", - "sass-embedded": "^1.83.4", - "typescript": "~5.6.2", - "typescript-eslint": "^8.18.2", - "vite": "^6.0.5" - } - }, - "node_modules/@alloc/quick-lru": { - "version": "5.2.0", - "resolved": "https://registry.npmjs.org/@alloc/quick-lru/-/quick-lru-5.2.0.tgz", - "integrity": "sha512-UrcABB+4bUrFABwbluTIBErXwvbsU/V7TZWfmbgJfbkwiBuziS9gxdODUyuiecfdGQ85jglMW6juS3+z5TsKLw==", - "license": "MIT", - "engines": { - "node": ">=10" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/@ampproject/remapping": { - "version": "2.3.0", - "resolved": "https://registry.npmjs.org/@ampproject/remapping/-/remapping-2.3.0.tgz", - "integrity": "sha512-30iZtAPgz+LTIYoeivqYo853f02jBYSd5uGnGpkFV0M3xOt9aN73erkgYAmZU43x4VfqcnLxW9Kpg3R5LC4YYw==", - "dev": true, - "license": "Apache-2.0", - "dependencies": { - "@jridgewell/gen-mapping": "^0.3.5", - "@jridgewell/trace-mapping": "^0.3.24" - }, - "engines": { - "node": ">=6.0.0" - } - }, - "node_modules/@babel/code-frame": { - "version": "7.26.2", - "resolved": "https://registry.npmjs.org/@babel/code-frame/-/code-frame-7.26.2.tgz", - "integrity": "sha512-RJlIHRueQgwWitWgF8OdFYGZX328Ax5BCemNGlqHfplnRT9ESi8JkFlvaVYbS+UubVY6dpv87Fs2u5M29iNFVQ==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/helper-validator-identifier": "^7.25.9", - "js-tokens": "^4.0.0", - "picocolors": "^1.0.0" - }, - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@babel/compat-data": { - "version": "7.26.5", - "resolved": "https://registry.npmjs.org/@babel/compat-data/-/compat-data-7.26.5.tgz", - "integrity": "sha512-XvcZi1KWf88RVbF9wn8MN6tYFloU5qX8KjuF3E1PVBmJ9eypXfs4GRiJwLuTZL0iSnJUKn1BFPa5BPZZJyFzPg==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@babel/core": { - "version": "7.26.7", - "resolved": "https://registry.npmjs.org/@babel/core/-/core-7.26.7.tgz", - "integrity": "sha512-SRijHmF0PSPgLIBYlWnG0hyeJLwXE2CgpsXaMOrtt2yp9/86ALw6oUlj9KYuZ0JN07T4eBMVIW4li/9S1j2BGA==", - "dev": true, - "license": "MIT", - "dependencies": { - "@ampproject/remapping": "^2.2.0", - "@babel/code-frame": "^7.26.2", - "@babel/generator": "^7.26.5", - "@babel/helper-compilation-targets": "^7.26.5", - "@babel/helper-module-transforms": "^7.26.0", - "@babel/helpers": "^7.26.7", - "@babel/parser": "^7.26.7", - "@babel/template": "^7.25.9", - "@babel/traverse": "^7.26.7", - "@babel/types": "^7.26.7", - "convert-source-map": "^2.0.0", - "debug": "^4.1.0", - "gensync": "^1.0.0-beta.2", - "json5": "^2.2.3", - "semver": "^6.3.1" - }, - "engines": { - "node": ">=6.9.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/babel" - } - }, - "node_modules/@babel/generator": { - "version": "7.26.5", - "resolved": "https://registry.npmjs.org/@babel/generator/-/generator-7.26.5.tgz", - "integrity": "sha512-2caSP6fN9I7HOe6nqhtft7V4g7/V/gfDsC3Ag4W7kEzzvRGKqiv0pu0HogPiZ3KaVSoNDhUws6IJjDjpfmYIXw==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/parser": "^7.26.5", - "@babel/types": "^7.26.5", - "@jridgewell/gen-mapping": "^0.3.5", - "@jridgewell/trace-mapping": "^0.3.25", - "jsesc": "^3.0.2" - }, - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@babel/helper-compilation-targets": { - "version": "7.26.5", - "resolved": "https://registry.npmjs.org/@babel/helper-compilation-targets/-/helper-compilation-targets-7.26.5.tgz", - "integrity": "sha512-IXuyn5EkouFJscIDuFF5EsiSolseme1s0CZB+QxVugqJLYmKdxI1VfIBOst0SUu4rnk2Z7kqTwmoO1lp3HIfnA==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/compat-data": "^7.26.5", - "@babel/helper-validator-option": "^7.25.9", - "browserslist": "^4.24.0", - "lru-cache": "^5.1.1", - "semver": "^6.3.1" - }, - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@babel/helper-module-imports": { - "version": "7.25.9", - "resolved": "https://registry.npmjs.org/@babel/helper-module-imports/-/helper-module-imports-7.25.9.tgz", - "integrity": "sha512-tnUA4RsrmflIM6W6RFTLFSXITtl0wKjgpnLgXyowocVPrbYrLUXSBXDgTs8BlbmIzIdlBySRQjINYs2BAkiLtw==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/traverse": "^7.25.9", - "@babel/types": "^7.25.9" - }, - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@babel/helper-module-transforms": { - "version": "7.26.0", - "resolved": "https://registry.npmjs.org/@babel/helper-module-transforms/-/helper-module-transforms-7.26.0.tgz", - "integrity": "sha512-xO+xu6B5K2czEnQye6BHA7DolFFmS3LB7stHZFaOLb1pAwO1HWLS8fXA+eh0A2yIvltPVmx3eNNDBJA2SLHXFw==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/helper-module-imports": "^7.25.9", - "@babel/helper-validator-identifier": "^7.25.9", - "@babel/traverse": "^7.25.9" - }, - "engines": { - "node": ">=6.9.0" - }, - "peerDependencies": { - "@babel/core": "^7.0.0" - } - }, - "node_modules/@babel/helper-plugin-utils": { - "version": "7.26.5", - "resolved": "https://registry.npmjs.org/@babel/helper-plugin-utils/-/helper-plugin-utils-7.26.5.tgz", - "integrity": "sha512-RS+jZcRdZdRFzMyr+wcsaqOmld1/EqTghfaBGQQd/WnRdzdlvSZ//kF7U8VQTxf1ynZ4cjUcYgjVGx13ewNPMg==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@babel/helper-string-parser": { - "version": "7.25.9", - "resolved": "https://registry.npmjs.org/@babel/helper-string-parser/-/helper-string-parser-7.25.9.tgz", - "integrity": "sha512-4A/SCr/2KLd5jrtOMFzaKjVtAei3+2r/NChoBNoZ3EyP/+GlhoaEGoWOZUmFmoITP7zOJyHIMm+DYRd8o3PvHA==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@babel/helper-validator-identifier": { - "version": "7.25.9", - "resolved": "https://registry.npmjs.org/@babel/helper-validator-identifier/-/helper-validator-identifier-7.25.9.tgz", - "integrity": "sha512-Ed61U6XJc3CVRfkERJWDz4dJwKe7iLmmJsbOGu9wSloNSFttHV0I8g6UAgb7qnK5ly5bGLPd4oXZlxCdANBOWQ==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@babel/helper-validator-option": { - "version": "7.25.9", - "resolved": "https://registry.npmjs.org/@babel/helper-validator-option/-/helper-validator-option-7.25.9.tgz", - "integrity": "sha512-e/zv1co8pp55dNdEcCynfj9X7nyUKUXoUEwfXqaZt0omVOmDe9oOTdKStH4GmAw6zxMFs50ZayuMfHDKlO7Tfw==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@babel/helpers": { - "version": "7.26.7", - "resolved": "https://registry.npmjs.org/@babel/helpers/-/helpers-7.26.7.tgz", - "integrity": "sha512-8NHiL98vsi0mbPQmYAGWwfcFaOy4j2HY49fXJCfuDcdE7fMIsH9a7GdaeXpIBsbT7307WU8KCMp5pUVDNL4f9A==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/template": "^7.25.9", - "@babel/types": "^7.26.7" - }, - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@babel/parser": { - "version": "7.26.7", - "resolved": "https://registry.npmjs.org/@babel/parser/-/parser-7.26.7.tgz", - "integrity": "sha512-kEvgGGgEjRUutvdVvZhbn/BxVt+5VSpwXz1j3WYXQbXDo8KzFOPNG2GQbdAiNq8g6wn1yKk7C/qrke03a84V+w==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/types": "^7.26.7" - }, - "bin": { - "parser": "bin/babel-parser.js" - }, - "engines": { - "node": ">=6.0.0" - } - }, - "node_modules/@babel/plugin-transform-react-jsx-self": { - "version": "7.25.9", - "resolved": "https://registry.npmjs.org/@babel/plugin-transform-react-jsx-self/-/plugin-transform-react-jsx-self-7.25.9.tgz", - "integrity": "sha512-y8quW6p0WHkEhmErnfe58r7x0A70uKphQm8Sp8cV7tjNQwK56sNVK0M73LK3WuYmsuyrftut4xAkjjgU0twaMg==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/helper-plugin-utils": "^7.25.9" - }, - "engines": { - "node": ">=6.9.0" - }, - "peerDependencies": { - "@babel/core": "^7.0.0-0" - } - }, - "node_modules/@babel/plugin-transform-react-jsx-source": { - "version": "7.25.9", - "resolved": "https://registry.npmjs.org/@babel/plugin-transform-react-jsx-source/-/plugin-transform-react-jsx-source-7.25.9.tgz", - "integrity": "sha512-+iqjT8xmXhhYv4/uiYd8FNQsraMFZIfxVSqxxVSZP0WbbSAWvBXAul0m/zu+7Vv4O/3WtApy9pmaTMiumEZgfg==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/helper-plugin-utils": "^7.25.9" - }, - "engines": { - "node": ">=6.9.0" - }, - "peerDependencies": { - "@babel/core": "^7.0.0-0" - } - }, - "node_modules/@babel/template": { - "version": "7.25.9", - "resolved": "https://registry.npmjs.org/@babel/template/-/template-7.25.9.tgz", - "integrity": "sha512-9DGttpmPvIxBb/2uwpVo3dqJ+O6RooAFOS+lB+xDqoE2PVCE8nfoHMdZLpfCQRLwvohzXISPZcgxt80xLfsuwg==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/code-frame": "^7.25.9", - "@babel/parser": "^7.25.9", - "@babel/types": "^7.25.9" - }, - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@babel/traverse": { - "version": "7.26.7", - "resolved": "https://registry.npmjs.org/@babel/traverse/-/traverse-7.26.7.tgz", - "integrity": "sha512-1x1sgeyRLC3r5fQOM0/xtQKsYjyxmFjaOrLJNtZ81inNjyJHGIolTULPiSc/2qe1/qfpFLisLQYFnnZl7QoedA==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/code-frame": "^7.26.2", - "@babel/generator": "^7.26.5", - "@babel/parser": "^7.26.7", - "@babel/template": "^7.25.9", - "@babel/types": "^7.26.7", - "debug": "^4.3.1", - "globals": "^11.1.0" - }, - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@babel/traverse/node_modules/globals": { - "version": "11.12.0", - "resolved": "https://registry.npmjs.org/globals/-/globals-11.12.0.tgz", - "integrity": "sha512-WOBp/EEGUiIsJSp7wcv/y6MO+lV9UoncWqxuFfm8eBwzWNgyfBd6Gz+IeKQ9jCmyhoH99g15M3T+QaVHFjizVA==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=4" - } - }, - "node_modules/@babel/types": { - "version": "7.26.7", - "resolved": "https://registry.npmjs.org/@babel/types/-/types-7.26.7.tgz", - "integrity": "sha512-t8kDRGrKXyp6+tjUh7hw2RLyclsW4TRoRvRHtSyAX9Bb5ldlFh+90YAYY6awRXrlB4G5G2izNeGySpATlFzmOg==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/helper-string-parser": "^7.25.9", - "@babel/helper-validator-identifier": "^7.25.9" - }, - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/@bufbuild/protobuf": { - "version": "2.2.3", - "resolved": "https://registry.npmjs.org/@bufbuild/protobuf/-/protobuf-2.2.3.tgz", - "integrity": "sha512-tFQoXHJdkEOSwj5tRIZSPNUuXK3RaR7T1nUrPgbYX1pUbvqqaaZAsfo+NXBPsz5rZMSKVFrgK1WL8Q/MSLvprg==", - "devOptional": true, - "license": "(Apache-2.0 AND BSD-3-Clause)" - }, - "node_modules/@esbuild/aix-ppc64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/aix-ppc64/-/aix-ppc64-0.24.2.tgz", - "integrity": "sha512-thpVCb/rhxE/BnMLQ7GReQLLN8q9qbHmI55F4489/ByVg2aQaQ6kbcLb6FHkocZzQhxc4gx0sCk0tJkKBFzDhA==", - "cpu": [ - "ppc64" - ], - "license": "MIT", - "optional": true, - "os": [ - "aix" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/android-arm": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/android-arm/-/android-arm-0.24.2.tgz", - "integrity": "sha512-tmwl4hJkCfNHwFB3nBa8z1Uy3ypZpxqxfTQOcHX+xRByyYgunVbZ9MzUUfb0RxaHIMnbHagwAxuTL+tnNM+1/Q==", - "cpu": [ - "arm" - ], - "license": "MIT", - "optional": true, - "os": [ - "android" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/android-arm64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/android-arm64/-/android-arm64-0.24.2.tgz", - "integrity": "sha512-cNLgeqCqV8WxfcTIOeL4OAtSmL8JjcN6m09XIgro1Wi7cF4t/THaWEa7eL5CMoMBdjoHOTh/vwTO/o2TRXIyzg==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "android" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/android-x64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/android-x64/-/android-x64-0.24.2.tgz", - "integrity": "sha512-B6Q0YQDqMx9D7rvIcsXfmJfvUYLoP722bgfBlO5cGvNVb5V/+Y7nhBE3mHV9OpxBf4eAS2S68KZztiPaWq4XYw==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "android" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/darwin-arm64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/darwin-arm64/-/darwin-arm64-0.24.2.tgz", - "integrity": "sha512-kj3AnYWc+CekmZnS5IPu9D+HWtUI49hbnyqk0FLEJDbzCIQt7hg7ucF1SQAilhtYpIujfaHr6O0UHlzzSPdOeA==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "darwin" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/darwin-x64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/darwin-x64/-/darwin-x64-0.24.2.tgz", - "integrity": "sha512-WeSrmwwHaPkNR5H3yYfowhZcbriGqooyu3zI/3GGpF8AyUdsrrP0X6KumITGA9WOyiJavnGZUwPGvxvwfWPHIA==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "darwin" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/freebsd-arm64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/freebsd-arm64/-/freebsd-arm64-0.24.2.tgz", - "integrity": "sha512-UN8HXjtJ0k/Mj6a9+5u6+2eZ2ERD7Edt1Q9IZiB5UZAIdPnVKDoG7mdTVGhHJIeEml60JteamR3qhsr1r8gXvg==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "freebsd" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/freebsd-x64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/freebsd-x64/-/freebsd-x64-0.24.2.tgz", - "integrity": "sha512-TvW7wE/89PYW+IevEJXZ5sF6gJRDY/14hyIGFXdIucxCsbRmLUcjseQu1SyTko+2idmCw94TgyaEZi9HUSOe3Q==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "freebsd" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/linux-arm": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/linux-arm/-/linux-arm-0.24.2.tgz", - "integrity": "sha512-n0WRM/gWIdU29J57hJyUdIsk0WarGd6To0s+Y+LwvlC55wt+GT/OgkwoXCXvIue1i1sSNWblHEig00GBWiJgfA==", - "cpu": [ - "arm" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/linux-arm64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/linux-arm64/-/linux-arm64-0.24.2.tgz", - "integrity": "sha512-7HnAD6074BW43YvvUmE/35Id9/NB7BeX5EoNkK9obndmZBUk8xmJJeU7DwmUeN7tkysslb2eSl6CTrYz6oEMQg==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/linux-ia32": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/linux-ia32/-/linux-ia32-0.24.2.tgz", - "integrity": "sha512-sfv0tGPQhcZOgTKO3oBE9xpHuUqguHvSo4jl+wjnKwFpapx+vUDcawbwPNuBIAYdRAvIDBfZVvXprIj3HA+Ugw==", - "cpu": [ - "ia32" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/linux-loong64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/linux-loong64/-/linux-loong64-0.24.2.tgz", - "integrity": "sha512-CN9AZr8kEndGooS35ntToZLTQLHEjtVB5n7dl8ZcTZMonJ7CCfStrYhrzF97eAecqVbVJ7APOEe18RPI4KLhwQ==", - "cpu": [ - "loong64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/linux-mips64el": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/linux-mips64el/-/linux-mips64el-0.24.2.tgz", - "integrity": "sha512-iMkk7qr/wl3exJATwkISxI7kTcmHKE+BlymIAbHO8xanq/TjHaaVThFF6ipWzPHryoFsesNQJPE/3wFJw4+huw==", - "cpu": [ - "mips64el" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/linux-ppc64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/linux-ppc64/-/linux-ppc64-0.24.2.tgz", - "integrity": "sha512-shsVrgCZ57Vr2L8mm39kO5PPIb+843FStGt7sGGoqiiWYconSxwTiuswC1VJZLCjNiMLAMh34jg4VSEQb+iEbw==", - "cpu": [ - "ppc64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/linux-riscv64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/linux-riscv64/-/linux-riscv64-0.24.2.tgz", - "integrity": "sha512-4eSFWnU9Hhd68fW16GD0TINewo1L6dRrB+oLNNbYyMUAeOD2yCK5KXGK1GH4qD/kT+bTEXjsyTCiJGHPZ3eM9Q==", - "cpu": [ - "riscv64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/linux-s390x": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/linux-s390x/-/linux-s390x-0.24.2.tgz", - "integrity": "sha512-S0Bh0A53b0YHL2XEXC20bHLuGMOhFDO6GN4b3YjRLK//Ep3ql3erpNcPlEFed93hsQAjAQDNsvcK+hV90FubSw==", - "cpu": [ - "s390x" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/linux-x64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/linux-x64/-/linux-x64-0.24.2.tgz", - "integrity": "sha512-8Qi4nQcCTbLnK9WoMjdC9NiTG6/E38RNICU6sUNqK0QFxCYgoARqVqxdFmWkdonVsvGqWhmm7MO0jyTqLqwj0Q==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/netbsd-arm64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/netbsd-arm64/-/netbsd-arm64-0.24.2.tgz", - "integrity": "sha512-wuLK/VztRRpMt9zyHSazyCVdCXlpHkKm34WUyinD2lzK07FAHTq0KQvZZlXikNWkDGoT6x3TD51jKQ7gMVpopw==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "netbsd" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/netbsd-x64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/netbsd-x64/-/netbsd-x64-0.24.2.tgz", - "integrity": "sha512-VefFaQUc4FMmJuAxmIHgUmfNiLXY438XrL4GDNV1Y1H/RW3qow68xTwjZKfj/+Plp9NANmzbH5R40Meudu8mmw==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "netbsd" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/openbsd-arm64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/openbsd-arm64/-/openbsd-arm64-0.24.2.tgz", - "integrity": "sha512-YQbi46SBct6iKnszhSvdluqDmxCJA+Pu280Av9WICNwQmMxV7nLRHZfjQzwbPs3jeWnuAhE9Jy0NrnJ12Oz+0A==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "openbsd" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/openbsd-x64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/openbsd-x64/-/openbsd-x64-0.24.2.tgz", - "integrity": "sha512-+iDS6zpNM6EnJyWv0bMGLWSWeXGN/HTaF/LXHXHwejGsVi+ooqDfMCCTerNFxEkM3wYVcExkeGXNqshc9iMaOA==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "openbsd" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/sunos-x64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/sunos-x64/-/sunos-x64-0.24.2.tgz", - "integrity": "sha512-hTdsW27jcktEvpwNHJU4ZwWFGkz2zRJUz8pvddmXPtXDzVKTTINmlmga3ZzwcuMpUvLw7JkLy9QLKyGpD2Yxig==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "sunos" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/win32-arm64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/win32-arm64/-/win32-arm64-0.24.2.tgz", - "integrity": "sha512-LihEQ2BBKVFLOC9ZItT9iFprsE9tqjDjnbulhHoFxYQtQfai7qfluVODIYxt1PgdoyQkz23+01rzwNwYfutxUQ==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "win32" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/win32-ia32": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/win32-ia32/-/win32-ia32-0.24.2.tgz", - "integrity": "sha512-q+iGUwfs8tncmFC9pcnD5IvRHAzmbwQ3GPS5/ceCyHdjXubwQWI12MKWSNSMYLJMq23/IUCvJMS76PDqXe1fxA==", - "cpu": [ - "ia32" - ], - "license": "MIT", - "optional": true, - "os": [ - "win32" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@esbuild/win32-x64": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/@esbuild/win32-x64/-/win32-x64-0.24.2.tgz", - "integrity": "sha512-7VTgWzgMGvup6aSqDPLiW5zHaxYJGTO4OokMjIlrCtf+VpEL+cXKtCvg723iguPYI5oaUNdS+/V7OU2gvXVWEg==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "win32" - ], - "engines": { - "node": ">=18" - } - }, - "node_modules/@eslint-community/eslint-utils": { - "version": "4.4.1", - "resolved": "https://registry.npmjs.org/@eslint-community/eslint-utils/-/eslint-utils-4.4.1.tgz", - "integrity": "sha512-s3O3waFUrMV8P/XaF/+ZTp1X9XBZW1a4B97ZnjQF2KYWaFD2A8KyFBsrsfSjEmjn3RGWAIuvlneuZm3CUK3jbA==", - "dev": true, - "license": "MIT", - "dependencies": { - "eslint-visitor-keys": "^3.4.3" - }, - "engines": { - "node": "^12.22.0 || ^14.17.0 || >=16.0.0" - }, - "funding": { - "url": "https://opencollective.com/eslint" - }, - "peerDependencies": { - "eslint": "^6.0.0 || ^7.0.0 || >=8.0.0" - } - }, - "node_modules/@eslint-community/eslint-utils/node_modules/eslint-visitor-keys": { - "version": "3.4.3", - "resolved": "https://registry.npmjs.org/eslint-visitor-keys/-/eslint-visitor-keys-3.4.3.tgz", - "integrity": "sha512-wpc+LXeiyiisxPlEkUzU6svyS1frIO3Mgxj1fdy7Pm8Ygzguax2N3Fa/D/ag1WqbOprdI+uY6wMUl8/a2G+iag==", - "dev": true, - "license": "Apache-2.0", - "engines": { - "node": "^12.22.0 || ^14.17.0 || >=16.0.0" - }, - "funding": { - "url": "https://opencollective.com/eslint" - } - }, - "node_modules/@eslint-community/regexpp": { - "version": "4.12.1", - "resolved": "https://registry.npmjs.org/@eslint-community/regexpp/-/regexpp-4.12.1.tgz", - "integrity": "sha512-CCZCDJuduB9OUkFkY2IgppNZMi2lBQgD2qzwXkEia16cge2pijY/aXi96CJMquDMn3nJdlPV1A5KrJEXwfLNzQ==", - "dev": true, - "license": "MIT", - "engines": { - "node": "^12.0.0 || ^14.0.0 || >=16.0.0" - } - }, - "node_modules/@eslint/config-array": { - "version": "0.19.2", - "resolved": "https://registry.npmjs.org/@eslint/config-array/-/config-array-0.19.2.tgz", - "integrity": "sha512-GNKqxfHG2ySmJOBSHg7LxeUx4xpuCoFjacmlCoYWEbaPXLwvfIjixRI12xCQZeULksQb23uiA8F40w5TojpV7w==", - "dev": true, - "license": "Apache-2.0", - "dependencies": { - "@eslint/object-schema": "^2.1.6", - "debug": "^4.3.1", - "minimatch": "^3.1.2" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - } - }, - "node_modules/@eslint/core": { - "version": "0.10.0", - "resolved": "https://registry.npmjs.org/@eslint/core/-/core-0.10.0.tgz", - "integrity": "sha512-gFHJ+xBOo4G3WRlR1e/3G8A6/KZAH6zcE/hkLRCZTi/B9avAG365QhFA8uOGzTMqgTghpn7/fSnscW++dpMSAw==", - "dev": true, - "license": "Apache-2.0", - "dependencies": { - "@types/json-schema": "^7.0.15" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - } - }, - "node_modules/@eslint/eslintrc": { - "version": "3.2.0", - "resolved": "https://registry.npmjs.org/@eslint/eslintrc/-/eslintrc-3.2.0.tgz", - "integrity": "sha512-grOjVNN8P3hjJn/eIETF1wwd12DdnwFDoyceUJLYYdkpbwq3nLi+4fqrTAONx7XDALqlL220wC/RHSC/QTI/0w==", - "dev": true, - "license": "MIT", - "dependencies": { - "ajv": "^6.12.4", - "debug": "^4.3.2", - "espree": "^10.0.1", - "globals": "^14.0.0", - "ignore": "^5.2.0", - "import-fresh": "^3.2.1", - "js-yaml": "^4.1.0", - "minimatch": "^3.1.2", - "strip-json-comments": "^3.1.1" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "url": "https://opencollective.com/eslint" - } - }, - "node_modules/@eslint/eslintrc/node_modules/globals": { - "version": "14.0.0", - "resolved": "https://registry.npmjs.org/globals/-/globals-14.0.0.tgz", - "integrity": "sha512-oahGvuMGQlPw/ivIYBjVSrWAfWLBeku5tpPE2fOPLi+WHffIWbuh2tCjhyQhTBPMf5E9jDEH4FOmTYgYwbKwtQ==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=18" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/@eslint/js": { - "version": "9.19.0", - "resolved": "https://registry.npmjs.org/@eslint/js/-/js-9.19.0.tgz", - "integrity": "sha512-rbq9/g38qjfqFLOVPvwjIvFFdNziEC5S65jmjPw5r6A//QH+W91akh9irMwjDN8zKUTak6W9EsAv4m/7Wnw0UQ==", - "dev": true, - "license": "MIT", - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - } - }, - "node_modules/@eslint/object-schema": { - "version": "2.1.6", - "resolved": "https://registry.npmjs.org/@eslint/object-schema/-/object-schema-2.1.6.tgz", - "integrity": "sha512-RBMg5FRL0I0gs51M/guSAj5/e14VQ4tpZnQNWwuDT66P14I43ItmPfIZRhO9fUVIPOAQXU47atlywZ/czoqFPA==", - "dev": true, - "license": "Apache-2.0", - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - } - }, - "node_modules/@eslint/plugin-kit": { - "version": "0.2.5", - "resolved": "https://registry.npmjs.org/@eslint/plugin-kit/-/plugin-kit-0.2.5.tgz", - "integrity": "sha512-lB05FkqEdUg2AA0xEbUz0SnkXT1LcCTa438W4IWTUh4hdOnVbQyOJ81OrDXsJk/LSiJHubgGEFoR5EHq1NsH1A==", - "dev": true, - "license": "Apache-2.0", - "dependencies": { - "@eslint/core": "^0.10.0", - "levn": "^0.4.1" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - } - }, - "node_modules/@heroicons/react": { - "version": "2.2.0", - "resolved": "https://registry.npmjs.org/@heroicons/react/-/react-2.2.0.tgz", - "integrity": "sha512-LMcepvRaS9LYHJGsF0zzmgKCUim/X3N/DQKc4jepAXJ7l8QxJ1PmxJzqplF2Z3FE4PqBAIGyJAQ/w4B5dsqbtQ==", - "license": "MIT", - "peerDependencies": { - "react": ">= 16 || ^19.0.0-rc" - } - }, - "node_modules/@humanfs/core": { - "version": "0.19.1", - "resolved": "https://registry.npmjs.org/@humanfs/core/-/core-0.19.1.tgz", - "integrity": "sha512-5DyQ4+1JEUzejeK1JGICcideyfUbGixgS9jNgex5nqkW+cY7WZhxBigmieN5Qnw9ZosSNVC9KQKyb+GUaGyKUA==", - "dev": true, - "license": "Apache-2.0", - "engines": { - "node": ">=18.18.0" - } - }, - "node_modules/@humanfs/node": { - "version": "0.16.6", - "resolved": "https://registry.npmjs.org/@humanfs/node/-/node-0.16.6.tgz", - "integrity": "sha512-YuI2ZHQL78Q5HbhDiBA1X4LmYdXCKCMQIfw0pw7piHJwyREFebJUvrQN4cMssyES6x+vfUbx1CIpaQUKYdQZOw==", - "dev": true, - "license": "Apache-2.0", - "dependencies": { - "@humanfs/core": "^0.19.1", - "@humanwhocodes/retry": "^0.3.0" - }, - "engines": { - "node": ">=18.18.0" - } - }, - "node_modules/@humanfs/node/node_modules/@humanwhocodes/retry": { - "version": "0.3.1", - "resolved": "https://registry.npmjs.org/@humanwhocodes/retry/-/retry-0.3.1.tgz", - "integrity": "sha512-JBxkERygn7Bv/GbN5Rv8Ul6LVknS+5Bp6RgDC/O8gEBU/yeH5Ui5C/OlWrTb6qct7LjjfT6Re2NxB0ln0yYybA==", - "dev": true, - "license": "Apache-2.0", - "engines": { - "node": ">=18.18" - }, - "funding": { - "type": "github", - "url": "https://github.com/sponsors/nzakas" - } - }, - "node_modules/@humanwhocodes/module-importer": { - "version": "1.0.1", - "resolved": "https://registry.npmjs.org/@humanwhocodes/module-importer/-/module-importer-1.0.1.tgz", - "integrity": "sha512-bxveV4V8v5Yb4ncFTT3rPSgZBOpCkjfK0y4oVVVJwIuDVBRMDXrPyXRL988i5ap9m9bnyEEjWfm5WkBmtffLfA==", - "dev": true, - "license": "Apache-2.0", - "engines": { - "node": ">=12.22" - }, - "funding": { - "type": "github", - "url": "https://github.com/sponsors/nzakas" - } - }, - "node_modules/@humanwhocodes/retry": { - "version": "0.4.1", - "resolved": "https://registry.npmjs.org/@humanwhocodes/retry/-/retry-0.4.1.tgz", - "integrity": "sha512-c7hNEllBlenFTHBky65mhq8WD2kbN9Q6gk0bTk8lSBvc554jpXSkST1iePudpt7+A/AQvuHs9EMqjHDXMY1lrA==", - "dev": true, - "license": "Apache-2.0", - "engines": { - "node": ">=18.18" - }, - "funding": { - "type": "github", - "url": "https://github.com/sponsors/nzakas" - } - }, - "node_modules/@jridgewell/gen-mapping": { - "version": "0.3.8", - "resolved": "https://registry.npmjs.org/@jridgewell/gen-mapping/-/gen-mapping-0.3.8.tgz", - "integrity": "sha512-imAbBGkb+ebQyxKgzv5Hu2nmROxoDOXHh80evxdoXNOrvAnVx7zimzc1Oo5h9RlfV4vPXaE2iM5pOFbvOCClWA==", - "devOptional": true, - "license": "MIT", - "dependencies": { - "@jridgewell/set-array": "^1.2.1", - "@jridgewell/sourcemap-codec": "^1.4.10", - "@jridgewell/trace-mapping": "^0.3.24" - }, - "engines": { - "node": ">=6.0.0" - } - }, - "node_modules/@jridgewell/resolve-uri": { - "version": "3.1.2", - "resolved": "https://registry.npmjs.org/@jridgewell/resolve-uri/-/resolve-uri-3.1.2.tgz", - "integrity": "sha512-bRISgCIjP20/tbWSPWMEi54QVPRZExkuD9lJL+UIxUKtwVJA8wW1Trb1jMs1RFXo1CBTNZ/5hpC9QvmKWdopKw==", - "devOptional": true, - "license": "MIT", - "engines": { - "node": ">=6.0.0" - } - }, - "node_modules/@jridgewell/set-array": { - "version": "1.2.1", - "resolved": "https://registry.npmjs.org/@jridgewell/set-array/-/set-array-1.2.1.tgz", - "integrity": "sha512-R8gLRTZeyp03ymzP/6Lil/28tGeGEzhx1q2k703KGWRAI1VdvPIXdG70VJc2pAMw3NA6JKL5hhFu1sJX0Mnn/A==", - "devOptional": true, - "license": "MIT", - "engines": { - "node": ">=6.0.0" - } - }, - "node_modules/@jridgewell/source-map": { - "version": "0.3.6", - "resolved": "https://registry.npmjs.org/@jridgewell/source-map/-/source-map-0.3.6.tgz", - "integrity": "sha512-1ZJTZebgqllO79ue2bm3rIGud/bOe0pP5BjSRCRxxYkEZS8STV7zN84UBbiYu7jy+eCKSnVIUgoWWE/tt+shMQ==", - "license": "MIT", - "optional": true, - "peer": true, - "dependencies": { - "@jridgewell/gen-mapping": "^0.3.5", - "@jridgewell/trace-mapping": "^0.3.25" - } - }, - "node_modules/@jridgewell/sourcemap-codec": { - "version": "1.5.0", - "resolved": "https://registry.npmjs.org/@jridgewell/sourcemap-codec/-/sourcemap-codec-1.5.0.tgz", - "integrity": "sha512-gv3ZRaISU3fjPAgNsriBRqGWQL6quFx04YMPW/zD8XMLsU32mhCCbfbO6KZFLjvYpCZ8zyDEgqsgf+PwPaM7GQ==", - "devOptional": true, - "license": "MIT" - }, - "node_modules/@jridgewell/trace-mapping": { - "version": "0.3.25", - "resolved": "https://registry.npmjs.org/@jridgewell/trace-mapping/-/trace-mapping-0.3.25.tgz", - "integrity": "sha512-vNk6aEwybGtawWmy/PzwnGDOjCkLWSD2wqvjGGAgOAwCGWySYXfYoxt00IJkTF+8Lb57DwOb3Aa0o9CApepiYQ==", - "devOptional": true, - "license": "MIT", - "dependencies": { - "@jridgewell/resolve-uri": "^3.1.0", - "@jridgewell/sourcemap-codec": "^1.4.14" - } - }, - "node_modules/@napi-rs/canvas": { - "version": "0.1.70", - "resolved": "https://registry.npmjs.org/@napi-rs/canvas/-/canvas-0.1.70.tgz", - "integrity": "sha512-nD6NGa4JbNYSZYsTnLGrqe9Kn/lCkA4ybXt8sx5ojDqZjr2i0TWAHxx/vhgfjX+i3hCdKWufxYwi7CfXqtITSA==", - "license": "MIT", - "optional": true, - "engines": { - "node": ">= 10" - }, - "optionalDependencies": { - "@napi-rs/canvas-android-arm64": "0.1.70", - "@napi-rs/canvas-darwin-arm64": "0.1.70", - "@napi-rs/canvas-darwin-x64": "0.1.70", - "@napi-rs/canvas-linux-arm-gnueabihf": "0.1.70", - "@napi-rs/canvas-linux-arm64-gnu": "0.1.70", - "@napi-rs/canvas-linux-arm64-musl": "0.1.70", - "@napi-rs/canvas-linux-riscv64-gnu": "0.1.70", - "@napi-rs/canvas-linux-x64-gnu": "0.1.70", - "@napi-rs/canvas-linux-x64-musl": "0.1.70", - "@napi-rs/canvas-win32-x64-msvc": "0.1.70" - } - }, - "node_modules/@napi-rs/canvas-android-arm64": { - "version": "0.1.70", - "resolved": "https://registry.npmjs.org/@napi-rs/canvas-android-arm64/-/canvas-android-arm64-0.1.70.tgz", - "integrity": "sha512-I/YOuQ0wbkVYxVaYtCgN42WKTYxNqFA0gTcTrHIGG1jfpDSyZWII/uHcjOo4nzd19io6Y4+/BqP8E5hJgf9OmQ==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "android" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@napi-rs/canvas-darwin-arm64": { - "version": "0.1.70", - "resolved": "https://registry.npmjs.org/@napi-rs/canvas-darwin-arm64/-/canvas-darwin-arm64-0.1.70.tgz", - "integrity": "sha512-4pPGyXetHIHkw2TOJHujt3mkCP8LdDu8+CT15ld9Id39c752RcI0amDHSuMLMQfAjvusA9B5kKxazwjMGjEJpQ==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "darwin" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@napi-rs/canvas-darwin-x64": { - "version": "0.1.70", - "resolved": "https://registry.npmjs.org/@napi-rs/canvas-darwin-x64/-/canvas-darwin-x64-0.1.70.tgz", - "integrity": "sha512-+2N6Os9LbkmDMHL+raknrUcLQhsXzc5CSXRbXws9C3pv/mjHRVszQ9dhFUUe9FjfPhCJznO6USVdwOtu7pOrzQ==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "darwin" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@napi-rs/canvas-linux-arm-gnueabihf": { - "version": "0.1.70", - "resolved": "https://registry.npmjs.org/@napi-rs/canvas-linux-arm-gnueabihf/-/canvas-linux-arm-gnueabihf-0.1.70.tgz", - "integrity": "sha512-QjscX9OaKq/990sVhSMj581xuqLgiaPVMjjYvWaCmAJRkNQ004QfoSMEm3FoTqM4DRoquP8jvuEXScVJsc1rqQ==", - "cpu": [ - "arm" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@napi-rs/canvas-linux-arm64-gnu": { - "version": "0.1.70", - "resolved": "https://registry.npmjs.org/@napi-rs/canvas-linux-arm64-gnu/-/canvas-linux-arm64-gnu-0.1.70.tgz", - "integrity": "sha512-LNakMOwwqwiHIwMpnMAbFRczQMQ7TkkMyATqFCOtUJNlE6LPP/QiUj/mlFrNbUn/hctqShJ60gWEb52ZTALbVw==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@napi-rs/canvas-linux-arm64-musl": { - "version": "0.1.70", - "resolved": "https://registry.npmjs.org/@napi-rs/canvas-linux-arm64-musl/-/canvas-linux-arm64-musl-0.1.70.tgz", - "integrity": "sha512-wBTOllEYNfJCHOdZj9v8gLzZ4oY3oyPX8MSRvaxPm/s7RfEXxCyZ8OhJ5xAyicsDdbE5YBZqdmaaeP5+xKxvtg==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@napi-rs/canvas-linux-riscv64-gnu": { - "version": "0.1.70", - "resolved": "https://registry.npmjs.org/@napi-rs/canvas-linux-riscv64-gnu/-/canvas-linux-riscv64-gnu-0.1.70.tgz", - "integrity": "sha512-GVUUPC8TuuFqHip0rxHkUqArQnlzmlXmTEBuXAWdgCv85zTCFH8nOHk/YCF5yo0Z2eOm8nOi90aWs0leJ4OE5Q==", - "cpu": [ - "riscv64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@napi-rs/canvas-linux-x64-gnu": { - "version": "0.1.70", - "resolved": "https://registry.npmjs.org/@napi-rs/canvas-linux-x64-gnu/-/canvas-linux-x64-gnu-0.1.70.tgz", - "integrity": "sha512-/kvUa2lZRwGNyfznSn5t1ShWJnr/m5acSlhTV3eXECafObjl0VBuA1HJw0QrilLpb4Fe0VLywkpD1NsMoVDROQ==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@napi-rs/canvas-linux-x64-musl": { - "version": "0.1.70", - "resolved": "https://registry.npmjs.org/@napi-rs/canvas-linux-x64-musl/-/canvas-linux-x64-musl-0.1.70.tgz", - "integrity": "sha512-aqlv8MLpycoMKRmds7JWCfVwNf1fiZxaU7JwJs9/ExjTD8lX2KjsO7CTeAj5Cl4aEuzxUWbJPUUE2Qu9cZ1vfg==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@napi-rs/canvas-win32-x64-msvc": { - "version": "0.1.70", - "resolved": "https://registry.npmjs.org/@napi-rs/canvas-win32-x64-msvc/-/canvas-win32-x64-msvc-0.1.70.tgz", - "integrity": "sha512-Q9QU3WIpwBTVHk4cPfBjGHGU4U0llQYRXgJtFtYqqGNEOKVN4OT6PQ+ve63xwIPODMpZ0HHyj/KLGc9CWc3EtQ==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "win32" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@nodelib/fs.scandir": { - "version": "2.1.5", - "resolved": "https://registry.npmjs.org/@nodelib/fs.scandir/-/fs.scandir-2.1.5.tgz", - "integrity": "sha512-vq24Bq3ym5HEQm2NKCr3yXDwjc7vTsEThRDnkp2DK9p1uqLR+DHurm/NOTo0KG7HYHU7eppKZj3MyqYuMBf62g==", - "dev": true, - "license": "MIT", - "dependencies": { - "@nodelib/fs.stat": "2.0.5", - "run-parallel": "^1.1.9" - }, - "engines": { - "node": ">= 8" - } - }, - "node_modules/@nodelib/fs.stat": { - "version": "2.0.5", - "resolved": "https://registry.npmjs.org/@nodelib/fs.stat/-/fs.stat-2.0.5.tgz", - "integrity": "sha512-RkhPPp2zrqDAQA/2jNhnztcPAlv64XdhIp7a7454A5ovI7Bukxgt7MX7udwAu3zg1DcpPU0rz3VV1SeaqvY4+A==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">= 8" - } - }, - "node_modules/@nodelib/fs.walk": { - "version": "1.2.8", - "resolved": "https://registry.npmjs.org/@nodelib/fs.walk/-/fs.walk-1.2.8.tgz", - "integrity": "sha512-oGB+UxlgWcgQkgwo8GcEGwemoTFt3FIO9ababBmaGwXIoBKZ+GTy0pP185beGg7Llih/NSHSV2XAs1lnznocSg==", - "dev": true, - "license": "MIT", - "dependencies": { - "@nodelib/fs.scandir": "2.1.5", - "fastq": "^1.6.0" - }, - "engines": { - "node": ">= 8" - } - }, - "node_modules/@rollup/rollup-android-arm-eabi": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-android-arm-eabi/-/rollup-android-arm-eabi-4.34.2.tgz", - "integrity": "sha512-6Fyg9yQbwJR+ykVdT9sid1oc2ewejS6h4wzQltmJfSW53N60G/ah9pngXGANdy9/aaE/TcUFpWosdm7JXS1WTQ==", - "cpu": [ - "arm" - ], - "license": "MIT", - "optional": true, - "os": [ - "android" - ] - }, - "node_modules/@rollup/rollup-android-arm64": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-android-arm64/-/rollup-android-arm64-4.34.2.tgz", - "integrity": "sha512-K5GfWe+vtQ3kyEbihrimM38UgX57UqHp+oME7X/EX9Im6suwZfa7Hsr8AtzbJvukTpwMGs+4s29YMSO3rwWtsw==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "android" - ] - }, - "node_modules/@rollup/rollup-darwin-arm64": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-darwin-arm64/-/rollup-darwin-arm64-4.34.2.tgz", - "integrity": "sha512-PSN58XG/V/tzqDb9kDGutUruycgylMlUE59f40ny6QIRNsTEIZsrNQTJKUN2keMMSmlzgunMFqyaGLmly39sug==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "darwin" - ] - }, - "node_modules/@rollup/rollup-darwin-x64": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-darwin-x64/-/rollup-darwin-x64-4.34.2.tgz", - "integrity": "sha512-gQhK788rQJm9pzmXyfBB84VHViDERhAhzGafw+E5mUpnGKuxZGkMVDa3wgDFKT6ukLC5V7QTifzsUKdNVxp5qQ==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "darwin" - ] - }, - "node_modules/@rollup/rollup-freebsd-arm64": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-freebsd-arm64/-/rollup-freebsd-arm64-4.34.2.tgz", - "integrity": "sha512-eiaHgQwGPpxLC3+zTAcdKl4VsBl3r0AiJOd1Um/ArEzAjN/dbPK1nROHrVkdnoE6p7Svvn04w3f/jEZSTVHunA==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "freebsd" - ] - }, - "node_modules/@rollup/rollup-freebsd-x64": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-freebsd-x64/-/rollup-freebsd-x64-4.34.2.tgz", - "integrity": "sha512-lhdiwQ+jf8pewYOTG4bag0Qd68Jn1v2gO1i0mTuiD+Qkt5vNfHVK/jrT7uVvycV8ZchlzXp5HDVmhpzjC6mh0g==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "freebsd" - ] - }, - "node_modules/@rollup/rollup-linux-arm-gnueabihf": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm-gnueabihf/-/rollup-linux-arm-gnueabihf-4.34.2.tgz", - "integrity": "sha512-lfqTpWjSvbgQP1vqGTXdv+/kxIznKXZlI109WkIFPbud41bjigjNmOAAKoazmRGx+k9e3rtIdbq2pQZPV1pMig==", - "cpu": [ - "arm" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ] - }, - "node_modules/@rollup/rollup-linux-arm-musleabihf": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm-musleabihf/-/rollup-linux-arm-musleabihf-4.34.2.tgz", - "integrity": "sha512-RGjqULqIurqqv+NJTyuPgdZhka8ImMLB32YwUle2BPTDqDoXNgwFjdjQC59FbSk08z0IqlRJjrJ0AvDQ5W5lpw==", - "cpu": [ - "arm" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ] - }, - "node_modules/@rollup/rollup-linux-arm64-gnu": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm64-gnu/-/rollup-linux-arm64-gnu-4.34.2.tgz", - "integrity": "sha512-ZvkPiheyXtXlFqHpsdgscx+tZ7hoR59vOettvArinEspq5fxSDSgfF+L5wqqJ9R4t+n53nyn0sKxeXlik7AY9Q==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ] - }, - "node_modules/@rollup/rollup-linux-arm64-musl": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm64-musl/-/rollup-linux-arm64-musl-4.34.2.tgz", - "integrity": "sha512-UlFk+E46TZEoxD9ufLKDBzfSG7Ki03fo6hsNRRRHF+KuvNZ5vd1RRVQm8YZlGsjcJG8R252XFK0xNPay+4WV7w==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ] - }, - "node_modules/@rollup/rollup-linux-loongarch64-gnu": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-loongarch64-gnu/-/rollup-linux-loongarch64-gnu-4.34.2.tgz", - "integrity": "sha512-hJhfsD9ykx59jZuuoQgYT1GEcNNi3RCoEmbo5OGfG8RlHOiVS7iVNev9rhLKh7UBYq409f4uEw0cclTXx8nh8Q==", - "cpu": [ - "loong64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ] - }, - "node_modules/@rollup/rollup-linux-powerpc64le-gnu": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-powerpc64le-gnu/-/rollup-linux-powerpc64le-gnu-4.34.2.tgz", - "integrity": "sha512-g/O5IpgtrQqPegvqopvmdCF9vneLE7eqYfdPWW8yjPS8f63DNam3U4ARL1PNNB64XHZDHKpvO2Giftf43puB8Q==", - "cpu": [ - "ppc64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ] - }, - "node_modules/@rollup/rollup-linux-riscv64-gnu": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-riscv64-gnu/-/rollup-linux-riscv64-gnu-4.34.2.tgz", - "integrity": "sha512-bSQijDC96M6PuooOuXHpvXUYiIwsnDmqGU8+br2U7iPoykNi9JtMUpN7K6xml29e0evK0/g0D1qbAUzWZFHY5Q==", - "cpu": [ - "riscv64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ] - }, - "node_modules/@rollup/rollup-linux-s390x-gnu": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-s390x-gnu/-/rollup-linux-s390x-gnu-4.34.2.tgz", - "integrity": "sha512-49TtdeVAsdRuiUHXPrFVucaP4SivazetGUVH8CIxVsNsaPHV4PFkpLmH9LeqU/R4Nbgky9lzX5Xe1NrzLyraVA==", - "cpu": [ - "s390x" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ] - }, - "node_modules/@rollup/rollup-linux-x64-gnu": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-x64-gnu/-/rollup-linux-x64-gnu-4.34.2.tgz", - "integrity": "sha512-j+jFdfOycLIQ7FWKka9Zd3qvsIyugg5LeZuHF6kFlXo6MSOc6R1w37YUVy8VpAKd81LMWGi5g9J25P09M0SSIw==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ] - }, - "node_modules/@rollup/rollup-linux-x64-musl": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-x64-musl/-/rollup-linux-x64-musl-4.34.2.tgz", - "integrity": "sha512-aDPHyM/D2SpXfSNCVWCxyHmOqN9qb7SWkY1+vaXqMNMXslZYnwh9V/UCudl6psyG0v6Ukj7pXanIpfZwCOEMUg==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ] - }, - "node_modules/@rollup/rollup-win32-arm64-msvc": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-arm64-msvc/-/rollup-win32-arm64-msvc-4.34.2.tgz", - "integrity": "sha512-LQRkCyUBnAo7r8dbEdtNU08EKLCJMgAk2oP5H3R7BnUlKLqgR3dUjrLBVirmc1RK6U6qhtDw29Dimeer8d5hzQ==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "win32" - ] - }, - "node_modules/@rollup/rollup-win32-ia32-msvc": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-ia32-msvc/-/rollup-win32-ia32-msvc-4.34.2.tgz", - "integrity": "sha512-wt8OhpQUi6JuPFkm1wbVi1BByeag87LDFzeKSXzIdGcX4bMLqORTtKxLoCbV57BHYNSUSOKlSL4BYYUghainYA==", - "cpu": [ - "ia32" - ], - "license": "MIT", - "optional": true, - "os": [ - "win32" - ] - }, - "node_modules/@rollup/rollup-win32-x64-msvc": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-x64-msvc/-/rollup-win32-x64-msvc-4.34.2.tgz", - "integrity": "sha512-rUrqINax0TvrPBXrFKg0YbQx18NpPN3NNrgmaao9xRNbTwek7lOXObhx8tQy8gelmQ/gLaGy1WptpU2eKJZImg==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "win32" - ] - }, - "node_modules/@sec-ant/readable-stream": { - "version": "0.6.0", - "resolved": "https://registry.npmjs.org/@sec-ant/readable-stream/-/readable-stream-0.6.0.tgz", - "integrity": "sha512-uiBh8DrB5FN35gP6/o8JEhEQ7/ci1jUsOZO/VMUjyvTpjtV54VstOXVj1TvTj/wsT23pfX6butxxh3qufsW3+g==", - "license": "MIT" - }, - "node_modules/@tailwindcss/node": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/node/-/node-4.1.1.tgz", - "integrity": "sha512-xvlh4pvfG/bkv0fEtJDABAm1tjtSmSyi2QmS4zyj1EKNI1UiOYiUq1IphSwDsNJ5vJ9cWEGs4rJXpUdCN2kujQ==", - "license": "MIT", - "dependencies": { - "enhanced-resolve": "^5.18.1", - "jiti": "^2.4.2", - "lightningcss": "1.29.2", - "tailwindcss": "4.1.1" - } - }, - "node_modules/@tailwindcss/oxide": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/oxide/-/oxide-4.1.1.tgz", - "integrity": "sha512-7+YBgnPQ4+jv6B6WVOerJ6WOzDzNJXrRKDts674v6TKAqFqYRr9+EBtSziO7nNcwQ8JtoZNMeqA+WJDjtCM/7w==", - "license": "MIT", - "engines": { - "node": ">= 10" - }, - "optionalDependencies": { - "@tailwindcss/oxide-android-arm64": "4.1.1", - "@tailwindcss/oxide-darwin-arm64": "4.1.1", - "@tailwindcss/oxide-darwin-x64": "4.1.1", - "@tailwindcss/oxide-freebsd-x64": "4.1.1", - "@tailwindcss/oxide-linux-arm-gnueabihf": "4.1.1", - "@tailwindcss/oxide-linux-arm64-gnu": "4.1.1", - "@tailwindcss/oxide-linux-arm64-musl": "4.1.1", - "@tailwindcss/oxide-linux-x64-gnu": "4.1.1", - "@tailwindcss/oxide-linux-x64-musl": "4.1.1", - "@tailwindcss/oxide-win32-arm64-msvc": "4.1.1", - "@tailwindcss/oxide-win32-x64-msvc": "4.1.1" - } - }, - "node_modules/@tailwindcss/oxide-android-arm64": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-android-arm64/-/oxide-android-arm64-4.1.1.tgz", - "integrity": "sha512-gTyRzfdParpoCU1yyUC/iN6XK6T0Ra4bDlF8Aeul5NP9cLzKEZDogdNVNGv5WZmCDkVol7qlex7TMmcfytMmmw==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "android" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@tailwindcss/oxide-darwin-arm64": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-darwin-arm64/-/oxide-darwin-arm64-4.1.1.tgz", - "integrity": "sha512-dI0QbdMWBvLB3MtaTKetzUKG9CUUQow8JSP4Nm+OxVokeZ+N+f1OmZW/hW1LzMxpx9RQCBgSRL+IIvKRat5Wdg==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "darwin" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@tailwindcss/oxide-darwin-x64": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-darwin-x64/-/oxide-darwin-x64-4.1.1.tgz", - "integrity": "sha512-2Y+NPQOTRBCItshPgY/CWg4bKi7E9evMg4bgdb6h9iZObCZLOe3doPcuSxGS3DB0dKyMFKE8pTdWtFUbxZBMSA==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "darwin" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@tailwindcss/oxide-freebsd-x64": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-freebsd-x64/-/oxide-freebsd-x64-4.1.1.tgz", - "integrity": "sha512-N97NGMsB/7CHShbc5ube4dcsW/bYENkBrg8yWi8ieN9boYVRdw3cZviVryV/Nfu9bKbBV9kUvduFF2qBI7rEqg==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "freebsd" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@tailwindcss/oxide-linux-arm-gnueabihf": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-linux-arm-gnueabihf/-/oxide-linux-arm-gnueabihf-4.1.1.tgz", - "integrity": "sha512-33Lk6KbHnUZbXqza6RWNFo9wqPQ4+H5BAn1CkUUfC1RZ1vYbyDN6+iJPj53wmnWJ3mhRI8jWt3Jt1fO02IVdUQ==", - "cpu": [ - "arm" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@tailwindcss/oxide-linux-arm64-gnu": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-linux-arm64-gnu/-/oxide-linux-arm64-gnu-4.1.1.tgz", - "integrity": "sha512-LyW35RzSUy+80WYScv03HKasAUmMFDaSbNpWfk1gG5gEE9kuRGnDzSrqMoLAmY/kzMCYP/1kqmUiAx8EFLkI2A==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@tailwindcss/oxide-linux-arm64-musl": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-linux-arm64-musl/-/oxide-linux-arm64-musl-4.1.1.tgz", - "integrity": "sha512-1KPnDMlHdqjPTUSFjx55pafvs8RZXRgxfeRgUrukwDKkuj7gFk28vW3Mx65YdiugAc9NWs3VgueZWaM1Po6uGw==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@tailwindcss/oxide-linux-x64-gnu": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-linux-x64-gnu/-/oxide-linux-x64-gnu-4.1.1.tgz", - "integrity": "sha512-4WdzA+MRlsinEEE6yxNMLJxpw0kE9XVipbAKdTL8BeUpyC2TdA3TL46lBulXzKp3BIxh3nqyR/UCqzl5o+3waQ==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@tailwindcss/oxide-linux-x64-musl": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-linux-x64-musl/-/oxide-linux-x64-musl-4.1.1.tgz", - "integrity": "sha512-q7Ugbw3ARcjCW2VMUYrcMbJ6aMQuWPArBBE2EqC/swPZTdGADvMQSlvR0VKusUM4HoSsO7ZbvcZ53YwR57+AKw==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@tailwindcss/oxide-win32-arm64-msvc": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-win32-arm64-msvc/-/oxide-win32-arm64-msvc-4.1.1.tgz", - "integrity": "sha512-0KpqsovgHcIzm7eAGzzEZsEs0/nPYXnRBv+aPq/GehpNQuE/NAQu+YgZXIIof+VflDFuyXOEnaFr7T5MZ1INhA==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "win32" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@tailwindcss/oxide-win32-x64-msvc": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-win32-x64-msvc/-/oxide-win32-x64-msvc-4.1.1.tgz", - "integrity": "sha512-B1mjeXNS26kBOHv5sXARf6Wd0PWHV9x1TDlW0ummrBUOUAxAy5wcy4Nii1wzNvCdvC448hgiL06ylhwAbNthmg==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "win32" - ], - "engines": { - "node": ">= 10" - } - }, - "node_modules/@tailwindcss/postcss": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/postcss/-/postcss-4.1.1.tgz", - "integrity": "sha512-GX9AEM+msH0i2Yh1b6CuDRaZRo3kmbvIrLbSfvJ53C3uaAgsQ//fTQAh9HMQ6t1a9zvoUptlYqG//plWsBQTCw==", - "license": "MIT", - "dependencies": { - "@alloc/quick-lru": "^5.2.0", - "@tailwindcss/node": "4.1.1", - "@tailwindcss/oxide": "4.1.1", - "postcss": "^8.4.41", - "tailwindcss": "4.1.1" - } - }, - "node_modules/@tailwindcss/vite": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/@tailwindcss/vite/-/vite-4.1.1.tgz", - "integrity": "sha512-tFTkRZwXq4XKr3S2dUZBxy80wbWYHdDSsu4QOB1yE1HJFKjfxKVpXtup4dyTVdQcLInoHC9lZXFPHnjoBP774g==", - "license": "MIT", - "dependencies": { - "@tailwindcss/node": "4.1.1", - "@tailwindcss/oxide": "4.1.1", - "tailwindcss": "4.1.1" - }, - "peerDependencies": { - "vite": "^5.2.0 || ^6" - } - }, - "node_modules/@types/babel__core": { - "version": "7.20.5", - "resolved": "https://registry.npmjs.org/@types/babel__core/-/babel__core-7.20.5.tgz", - "integrity": "sha512-qoQprZvz5wQFJwMDqeseRXWv3rqMvhgpbXFfVyWhbx9X47POIA6i/+dXefEmZKoAgOaTdaIgNSMqMIU61yRyzA==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/parser": "^7.20.7", - "@babel/types": "^7.20.7", - "@types/babel__generator": "*", - "@types/babel__template": "*", - "@types/babel__traverse": "*" - } - }, - "node_modules/@types/babel__generator": { - "version": "7.6.8", - "resolved": "https://registry.npmjs.org/@types/babel__generator/-/babel__generator-7.6.8.tgz", - "integrity": "sha512-ASsj+tpEDsEiFr1arWrlN6V3mdfjRMZt6LtK/Vp/kreFLnr5QH5+DhvD5nINYZXzwJvXeGq+05iUXcAzVrqWtw==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/types": "^7.0.0" - } - }, - "node_modules/@types/babel__template": { - "version": "7.4.4", - "resolved": "https://registry.npmjs.org/@types/babel__template/-/babel__template-7.4.4.tgz", - "integrity": "sha512-h/NUaSyG5EyxBIp8YRxo4RMe2/qQgvyowRwVMzhYhBCONbW8PUsg4lkFMrhgZhUe5z3L3MiLDuvyJ/CaPa2A8A==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/parser": "^7.1.0", - "@babel/types": "^7.0.0" - } - }, - "node_modules/@types/babel__traverse": { - "version": "7.20.6", - "resolved": "https://registry.npmjs.org/@types/babel__traverse/-/babel__traverse-7.20.6.tgz", - "integrity": "sha512-r1bzfrm0tomOI8g1SzvCaQHo6Lcv6zu0EA+W2kHrt8dyrHQxGzBBL4kdkzIS+jBMV+EYcMAEAqXqYaLJq5rOZg==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/types": "^7.20.7" - } - }, - "node_modules/@types/cookie": { - "version": "0.6.0", - "resolved": "https://registry.npmjs.org/@types/cookie/-/cookie-0.6.0.tgz", - "integrity": "sha512-4Kh9a6B2bQciAhf7FSuMRRkUWecJgJu9nPnx3yzpsfXX/c50REIqpHY4C82bXP90qrLtXtkDxTZosYO3UpOwlA==", - "license": "MIT" - }, - "node_modules/@types/debug": { - "version": "4.1.12", - "resolved": "https://registry.npmjs.org/@types/debug/-/debug-4.1.12.tgz", - "integrity": "sha512-vIChWdVG3LG1SMxEvI/AK+FWJthlrqlTu7fbrlywTkkaONwk/UAGaULXRlf8vkzFBLVm0zkMdCquhL5aOjhXPQ==", - "license": "MIT", - "dependencies": { - "@types/ms": "*" - } - }, - "node_modules/@types/estree": { - "version": "1.0.6", - "resolved": "https://registry.npmjs.org/@types/estree/-/estree-1.0.6.tgz", - "integrity": "sha512-AYnb1nQyY49te+VRAVgmzfcgjYS91mY5P0TKUDCLEM+gNnA+3T6rWITXRLYCpahpqSQbN5cE+gHpnPyXjHWxcw==", - "license": "MIT" - }, - "node_modules/@types/estree-jsx": { - "version": "1.0.5", - "resolved": "https://registry.npmjs.org/@types/estree-jsx/-/estree-jsx-1.0.5.tgz", - "integrity": "sha512-52CcUVNFyfb1A2ALocQw/Dd1BQFNmSdkuC3BkZ6iqhdMfQz7JWOFRuJFloOzjk+6WijU56m9oKXFAXc7o3Towg==", - "license": "MIT", - "dependencies": { - "@types/estree": "*" - } - }, - "node_modules/@types/hast": { - "version": "3.0.4", - "resolved": "https://registry.npmjs.org/@types/hast/-/hast-3.0.4.tgz", - "integrity": "sha512-WPs+bbQw5aCj+x6laNGWLH3wviHtoCv/P3+otBhbOhJgG8qtpdAMlTCxLtsTWA7LH1Oh/bFCHsBn0TPS5m30EQ==", - "license": "MIT", - "dependencies": { - "@types/unist": "*" - } - }, - "node_modules/@types/json-schema": { - "version": "7.0.15", - "resolved": "https://registry.npmjs.org/@types/json-schema/-/json-schema-7.0.15.tgz", - "integrity": "sha512-5+fP8P8MFNC+AyZCDxrB2pkZFPGzqQWUzpSeuuVLvm8VMcorNYavBqoFcxK8bQz4Qsbn4oUEEem4wDLfcysGHA==", - "dev": true, - "license": "MIT" - }, - "node_modules/@types/katex": { - "version": "0.16.7", - "resolved": "https://registry.npmjs.org/@types/katex/-/katex-0.16.7.tgz", - "integrity": "sha512-HMwFiRujE5PjrgwHQ25+bsLJgowjGjm5Z8FVSf0N6PwgJrwxH0QxzHYDcKsTfV3wva0vzrpqMTJS2jXPr5BMEQ==", - "license": "MIT" - }, - "node_modules/@types/linkify-it": { - "version": "5.0.0", - "resolved": "https://registry.npmjs.org/@types/linkify-it/-/linkify-it-5.0.0.tgz", - "integrity": "sha512-sVDA58zAw4eWAffKOaQH5/5j3XeayukzDk+ewSsnv3p4yJEZHCCzMDiZM8e0OUrRvmpGZ85jf4yDHkHsgBNr9Q==", - "dev": true, - "license": "MIT" - }, - "node_modules/@types/markdown-it": { - "version": "14.1.2", - "resolved": "https://registry.npmjs.org/@types/markdown-it/-/markdown-it-14.1.2.tgz", - "integrity": "sha512-promo4eFwuiW+TfGxhi+0x3czqTYJkG8qB17ZUJiVF10Xm7NLVRSLUsfRTU/6h1e24VvRnXCx+hG7li58lkzog==", - "dev": true, - "license": "MIT", - "dependencies": { - "@types/linkify-it": "^5", - "@types/mdurl": "^2" - } - }, - "node_modules/@types/mdast": { - "version": "4.0.4", - "resolved": "https://registry.npmjs.org/@types/mdast/-/mdast-4.0.4.tgz", - "integrity": "sha512-kGaNbPh1k7AFzgpud/gMdvIm5xuECykRR+JnWKQno9TAXVa6WIVCGTPvYGekIDL4uwCZQSYbUxNBSb1aUo79oA==", - "license": "MIT", - "dependencies": { - "@types/unist": "*" - } - }, - "node_modules/@types/mdurl": { - "version": "2.0.0", - "resolved": "https://registry.npmjs.org/@types/mdurl/-/mdurl-2.0.0.tgz", - "integrity": "sha512-RGdgjQUZba5p6QEFAVx2OGb8rQDL/cPRG7GiedRzMcJ1tYnUANBncjbSB1NRGwbvjcPeikRABz2nshyPk1bhWg==", - "dev": true, - "license": "MIT" - }, - "node_modules/@types/ms": { - "version": "2.1.0", - "resolved": "https://registry.npmjs.org/@types/ms/-/ms-2.1.0.tgz", - "integrity": "sha512-GsCCIZDE/p3i96vtEqx+7dBUGXrc7zeSK3wwPHIaRThS+9OhWIXRqzs4d6k1SVU8g91DrNRWxWUGhp5KXQb2VA==", - "license": "MIT" - }, - "node_modules/@types/node": { - "version": "22.13.1", - "resolved": "https://registry.npmjs.org/@types/node/-/node-22.13.1.tgz", - "integrity": "sha512-jK8uzQlrvXqEU91UxiK5J7pKHyzgnI1Qnl0QDHIgVGuolJhRb9EEl28Cj9b3rGR8B2lhFCtvIm5os8lFnO/1Ew==", - "devOptional": true, - "license": "MIT", - "dependencies": { - "undici-types": "~6.20.0" - } - }, - "node_modules/@types/prop-types": { - "version": "15.7.14", - "resolved": "https://registry.npmjs.org/@types/prop-types/-/prop-types-15.7.14.tgz", - "integrity": "sha512-gNMvNH49DJ7OJYv+KAKn0Xp45p8PLl6zo2YnvDIbTd4J6MER2BmWN49TG7n9LvkyihINxeKW8+3bfS2yDC9dzQ==", - "license": "MIT" - }, - "node_modules/@types/react": { - "version": "18.3.18", - "resolved": "https://registry.npmjs.org/@types/react/-/react-18.3.18.tgz", - "integrity": "sha512-t4yC+vtgnkYjNSKlFx1jkAhH8LgTo2N/7Qvi83kdEaUtMDiwpbLAktKDaAMlRcJ5eSxZkH74eEGt1ky31d7kfQ==", - "license": "MIT", - "dependencies": { - "@types/prop-types": "*", - "csstype": "^3.0.2" - } - }, - "node_modules/@types/react-dom": { - "version": "18.3.5", - "resolved": "https://registry.npmjs.org/@types/react-dom/-/react-dom-18.3.5.tgz", - "integrity": "sha512-P4t6saawp+b/dFrUr2cvkVsfvPguwsxtH6dNIYRllMsefqFzkZk5UIjzyDOv5g1dXIPdG4Sp1yCR4Z6RCUsG/Q==", - "dev": true, - "license": "MIT", - "peerDependencies": { - "@types/react": "^18.0.0" - } - }, - "node_modules/@types/unist": { - "version": "3.0.3", - "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", - "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", - "license": "MIT" - }, - "node_modules/@typescript-eslint/eslint-plugin": { - "version": "8.23.0", - "resolved": "https://registry.npmjs.org/@typescript-eslint/eslint-plugin/-/eslint-plugin-8.23.0.tgz", - "integrity": "sha512-vBz65tJgRrA1Q5gWlRfvoH+w943dq9K1p1yDBY2pc+a1nbBLZp7fB9+Hk8DaALUbzjqlMfgaqlVPT1REJdkt/w==", - "dev": true, - "license": "MIT", - "dependencies": { - "@eslint-community/regexpp": "^4.10.0", - "@typescript-eslint/scope-manager": "8.23.0", - "@typescript-eslint/type-utils": "8.23.0", - "@typescript-eslint/utils": "8.23.0", - "@typescript-eslint/visitor-keys": "8.23.0", - "graphemer": "^1.4.0", - "ignore": "^5.3.1", - "natural-compare": "^1.4.0", - "ts-api-utils": "^2.0.1" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/typescript-eslint" - }, - "peerDependencies": { - "@typescript-eslint/parser": "^8.0.0 || ^8.0.0-alpha.0", - "eslint": "^8.57.0 || ^9.0.0", - "typescript": ">=4.8.4 <5.8.0" - } - }, - "node_modules/@typescript-eslint/parser": { - "version": "8.23.0", - "resolved": "https://registry.npmjs.org/@typescript-eslint/parser/-/parser-8.23.0.tgz", - "integrity": "sha512-h2lUByouOXFAlMec2mILeELUbME5SZRN/7R9Cw2RD2lRQQY08MWMM+PmVVKKJNK1aIwqTo9t/0CvOxwPbRIE2Q==", - "dev": true, - "license": "MIT", - "dependencies": { - "@typescript-eslint/scope-manager": "8.23.0", - "@typescript-eslint/types": "8.23.0", - "@typescript-eslint/typescript-estree": "8.23.0", - "@typescript-eslint/visitor-keys": "8.23.0", - "debug": "^4.3.4" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/typescript-eslint" - }, - "peerDependencies": { - "eslint": "^8.57.0 || ^9.0.0", - "typescript": ">=4.8.4 <5.8.0" - } - }, - "node_modules/@typescript-eslint/scope-manager": { - "version": "8.23.0", - "resolved": "https://registry.npmjs.org/@typescript-eslint/scope-manager/-/scope-manager-8.23.0.tgz", - "integrity": "sha512-OGqo7+dXHqI7Hfm+WqkZjKjsiRtFUQHPdGMXzk5mYXhJUedO7e/Y7i8AK3MyLMgZR93TX4bIzYrfyVjLC+0VSw==", - "dev": true, - "license": "MIT", - "dependencies": { - "@typescript-eslint/types": "8.23.0", - "@typescript-eslint/visitor-keys": "8.23.0" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/typescript-eslint" - } - }, - "node_modules/@typescript-eslint/type-utils": { - "version": "8.23.0", - "resolved": "https://registry.npmjs.org/@typescript-eslint/type-utils/-/type-utils-8.23.0.tgz", - "integrity": "sha512-iIuLdYpQWZKbiH+RkCGc6iu+VwscP5rCtQ1lyQ7TYuKLrcZoeJVpcLiG8DliXVkUxirW/PWlmS+d6yD51L9jvA==", - "dev": true, - "license": "MIT", - "dependencies": { - "@typescript-eslint/typescript-estree": "8.23.0", - "@typescript-eslint/utils": "8.23.0", - "debug": "^4.3.4", - "ts-api-utils": "^2.0.1" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/typescript-eslint" - }, - "peerDependencies": { - "eslint": "^8.57.0 || ^9.0.0", - "typescript": ">=4.8.4 <5.8.0" - } - }, - "node_modules/@typescript-eslint/types": { - "version": "8.23.0", - "resolved": "https://registry.npmjs.org/@typescript-eslint/types/-/types-8.23.0.tgz", - "integrity": "sha512-1sK4ILJbCmZOTt9k4vkoulT6/y5CHJ1qUYxqpF1K/DBAd8+ZUL4LlSCxOssuH5m4rUaaN0uS0HlVPvd45zjduQ==", - "dev": true, - "license": "MIT", - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/typescript-eslint" - } - }, - "node_modules/@typescript-eslint/typescript-estree": { - "version": "8.23.0", - "resolved": "https://registry.npmjs.org/@typescript-eslint/typescript-estree/-/typescript-estree-8.23.0.tgz", - "integrity": "sha512-LcqzfipsB8RTvH8FX24W4UUFk1bl+0yTOf9ZA08XngFwMg4Kj8A+9hwz8Cr/ZS4KwHrmo9PJiLZkOt49vPnuvQ==", - "dev": true, - "license": "MIT", - "dependencies": { - "@typescript-eslint/types": "8.23.0", - "@typescript-eslint/visitor-keys": "8.23.0", - "debug": "^4.3.4", - "fast-glob": "^3.3.2", - "is-glob": "^4.0.3", - "minimatch": "^9.0.4", - "semver": "^7.6.0", - "ts-api-utils": "^2.0.1" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/typescript-eslint" - }, - "peerDependencies": { - "typescript": ">=4.8.4 <5.8.0" - } - }, - "node_modules/@typescript-eslint/typescript-estree/node_modules/brace-expansion": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-2.0.1.tgz", - "integrity": "sha512-XnAIvQ8eM+kC6aULx6wuQiwVsnzsi9d3WxzV3FpWTGA19F621kwdbsAcFKXgKUHZWsy+mY6iL1sHTxWEFCytDA==", - "dev": true, - "license": "MIT", - "dependencies": { - "balanced-match": "^1.0.0" - } - }, - "node_modules/@typescript-eslint/typescript-estree/node_modules/minimatch": { - "version": "9.0.5", - "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-9.0.5.tgz", - "integrity": "sha512-G6T0ZX48xgozx7587koeX9Ys2NYy6Gmv//P89sEte9V9whIapMNF4idKxnW2QtCcLiTWlb/wfCabAtAFWhhBow==", - "dev": true, - "license": "ISC", - "dependencies": { - "brace-expansion": "^2.0.1" - }, - "engines": { - "node": ">=16 || 14 >=14.17" - }, - "funding": { - "url": "https://github.com/sponsors/isaacs" - } - }, - "node_modules/@typescript-eslint/typescript-estree/node_modules/semver": { - "version": "7.7.1", - "resolved": "https://registry.npmjs.org/semver/-/semver-7.7.1.tgz", - "integrity": "sha512-hlq8tAfn0m/61p4BVRcPzIGr6LKiMwo4VM6dGi6pt4qcRkmNzTcWq6eCEjEh+qXjkMDvPlOFFSGwQjoEa6gyMA==", - "dev": true, - "license": "ISC", - "bin": { - "semver": "bin/semver.js" - }, - "engines": { - "node": ">=10" - } - }, - "node_modules/@typescript-eslint/utils": { - "version": "8.23.0", - "resolved": "https://registry.npmjs.org/@typescript-eslint/utils/-/utils-8.23.0.tgz", - "integrity": "sha512-uB/+PSo6Exu02b5ZEiVtmY6RVYO7YU5xqgzTIVZwTHvvK3HsL8tZZHFaTLFtRG3CsV4A5mhOv+NZx5BlhXPyIA==", - "dev": true, - "license": "MIT", - "dependencies": { - "@eslint-community/eslint-utils": "^4.4.0", - "@typescript-eslint/scope-manager": "8.23.0", - "@typescript-eslint/types": "8.23.0", - "@typescript-eslint/typescript-estree": "8.23.0" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/typescript-eslint" - }, - "peerDependencies": { - "eslint": "^8.57.0 || ^9.0.0", - "typescript": ">=4.8.4 <5.8.0" - } - }, - "node_modules/@typescript-eslint/visitor-keys": { - "version": "8.23.0", - "resolved": "https://registry.npmjs.org/@typescript-eslint/visitor-keys/-/visitor-keys-8.23.0.tgz", - "integrity": "sha512-oWWhcWDLwDfu++BGTZcmXWqpwtkwb5o7fxUIGksMQQDSdPW9prsSnfIOZMlsj4vBOSrcnjIUZMiIjODgGosFhQ==", - "dev": true, - "license": "MIT", - "dependencies": { - "@typescript-eslint/types": "8.23.0", - "eslint-visitor-keys": "^4.2.0" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/typescript-eslint" - } - }, - "node_modules/@ungap/structured-clone": { - "version": "1.3.0", - "resolved": "https://registry.npmjs.org/@ungap/structured-clone/-/structured-clone-1.3.0.tgz", - "integrity": "sha512-WmoN8qaIAo7WTYWbAZuG8PYEhn5fkz7dZrqTBZ7dtt//lL2Gwms1IcnQ5yHqjDfX8Ft5j4YzDM23f87zBfDe9g==", - "license": "ISC" - }, - "node_modules/@vitejs/plugin-react": { - "version": "4.3.4", - "resolved": "https://registry.npmjs.org/@vitejs/plugin-react/-/plugin-react-4.3.4.tgz", - "integrity": "sha512-SCCPBJtYLdE8PX/7ZQAs1QAZ8Jqwih+0VBLum1EGqmCCQal+MIUqLCzj3ZUy8ufbC0cAM4LRlSTm7IQJwWT4ug==", - "dev": true, - "license": "MIT", - "dependencies": { - "@babel/core": "^7.26.0", - "@babel/plugin-transform-react-jsx-self": "^7.25.9", - "@babel/plugin-transform-react-jsx-source": "^7.25.9", - "@types/babel__core": "^7.20.5", - "react-refresh": "^0.14.2" - }, - "engines": { - "node": "^14.18.0 || >=16.0.0" - }, - "peerDependencies": { - "vite": "^4.2.0 || ^5.0.0 || ^6.0.0" - } - }, - "node_modules/@vscode/markdown-it-katex": { - "version": "1.1.1", - "resolved": "https://registry.npmjs.org/@vscode/markdown-it-katex/-/markdown-it-katex-1.1.1.tgz", - "integrity": "sha512-3KTlbsRBPJQLE2YmLL7K6nunTlU+W9T5+FjfNdWuIUKgxSS6HWLQHaO3L4MkJi7z7MpIPpY+g4N+cWNBPE/MSA==", - "license": "MIT", - "dependencies": { - "katex": "^0.16.4" - } - }, - "node_modules/acorn": { - "version": "8.14.0", - "resolved": "https://registry.npmjs.org/acorn/-/acorn-8.14.0.tgz", - "integrity": "sha512-cl669nCJTZBsL97OF4kUQm5g5hC2uihk0NxY3WENAC0TYdILVkAyHymAntgxGkl7K+t0cXIrH5siy5S4XkFycA==", - "devOptional": true, - "license": "MIT", - "bin": { - "acorn": "bin/acorn" - }, - "engines": { - "node": ">=0.4.0" - } - }, - "node_modules/acorn-jsx": { - "version": "5.3.2", - "resolved": "https://registry.npmjs.org/acorn-jsx/-/acorn-jsx-5.3.2.tgz", - "integrity": "sha512-rq9s+JNhf0IChjtDXxllJ7g41oZk5SlXtp0LHwyA5cejwn7vKmKp4pPri6YEePv2PU65sAsegbXtIinmDFDXgQ==", - "dev": true, - "license": "MIT", - "peerDependencies": { - "acorn": "^6.0.0 || ^7.0.0 || ^8.0.0" - } - }, - "node_modules/ajv": { - "version": "6.12.6", - "resolved": "https://registry.npmjs.org/ajv/-/ajv-6.12.6.tgz", - "integrity": "sha512-j3fVLgvTo527anyYyJOGTYJbG+vnnQYvE0m5mmkc1TK+nxAppkCLMIL0aZ4dblVCNoGShhm+kzE4ZUykBoMg4g==", - "dev": true, - "license": "MIT", - "dependencies": { - "fast-deep-equal": "^3.1.1", - "fast-json-stable-stringify": "^2.0.0", - "json-schema-traverse": "^0.4.1", - "uri-js": "^4.2.2" - }, - "funding": { - "type": "github", - "url": "https://github.com/sponsors/epoberezkin" - } - }, - "node_modules/ansi-styles": { - "version": "4.3.0", - "resolved": "https://registry.npmjs.org/ansi-styles/-/ansi-styles-4.3.0.tgz", - "integrity": "sha512-zbB9rCJAT1rbjiVDb2hqKFHNYLxgtk8NURxZ3IZwD3F6NtxbXZQCnnSi1Lkx+IDohdPlFp222wVALIheZJQSEg==", - "dev": true, - "license": "MIT", - "dependencies": { - "color-convert": "^2.0.1" - }, - "engines": { - "node": ">=8" - }, - "funding": { - "url": "https://github.com/chalk/ansi-styles?sponsor=1" - } - }, - "node_modules/argparse": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/argparse/-/argparse-2.0.1.tgz", - "integrity": "sha512-8+9WqebbFzpX9OR+Wa6O29asIogeRMzcGtAINdpMHHyAg10f05aSFVBbcEqGf/PXw1EjAZ+q2/bEBg3DvurK3Q==", - "dev": true, - "license": "Python-2.0" - }, - "node_modules/attr-accept": { - "version": "2.2.5", - "resolved": "https://registry.npmjs.org/attr-accept/-/attr-accept-2.2.5.tgz", - "integrity": "sha512-0bDNnY/u6pPwHDMoF0FieU354oBi0a8rD9FcsLwzcGWbc8KS8KPIi7y+s13OlVY+gMWc/9xEMUgNE6Qm8ZllYQ==", - "license": "MIT", - "engines": { - "node": ">=4" - } - }, - "node_modules/autoprefixer": { - "version": "10.4.20", - "resolved": "https://registry.npmjs.org/autoprefixer/-/autoprefixer-10.4.20.tgz", - "integrity": "sha512-XY25y5xSv/wEoqzDyXXME4AFfkZI0P23z6Fs3YgymDnKJkCGOnkL0iTxCa85UTqaSgfcqyf3UA6+c7wUvx/16g==", - "funding": [ - { - "type": "opencollective", - "url": "https://opencollective.com/postcss/" - }, - { - "type": "tidelift", - "url": "https://tidelift.com/funding/github/npm/autoprefixer" - }, - { - "type": "github", - "url": "https://github.com/sponsors/ai" - } - ], - "license": "MIT", - "dependencies": { - "browserslist": "^4.23.3", - "caniuse-lite": "^1.0.30001646", - "fraction.js": "^4.3.7", - "normalize-range": "^0.1.2", - "picocolors": "^1.0.1", - "postcss-value-parser": "^4.2.0" - }, - "bin": { - "autoprefixer": "bin/autoprefixer" - }, - "engines": { - "node": "^10 || ^12 || >=14" - }, - "peerDependencies": { - "postcss": "^8.1.0" - } - }, - "node_modules/bail": { - "version": "2.0.2", - "resolved": "https://registry.npmjs.org/bail/-/bail-2.0.2.tgz", - "integrity": "sha512-0xO6mYd7JB2YesxDKplafRpsiOzPt9V02ddPCLbY1xYGPOX24NTyN50qnUxgCPcSoYMhKpAuBTjQoRZCAkUDRw==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/balanced-match": { - "version": "1.0.2", - "resolved": "https://registry.npmjs.org/balanced-match/-/balanced-match-1.0.2.tgz", - "integrity": "sha512-3oSeUO0TMV67hN1AmbXsK4yaqU7tjiHlbxRDZOpH0KW9+CeX4bRAaX0Anxt0tx2MrpRpWwQaPwIlISEJhYU5Pw==", - "dev": true, - "license": "MIT" - }, - "node_modules/brace-expansion": { - "version": "1.1.11", - "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-1.1.11.tgz", - "integrity": "sha512-iCuPHDFgrHX7H2vEI/5xpz07zSHB00TpugqhmYtVmMO6518mCuRMoOYFldEBl0g187ufozdaHgWKcYFb61qGiA==", - "dev": true, - "license": "MIT", - "dependencies": { - "balanced-match": "^1.0.0", - "concat-map": "0.0.1" - } - }, - "node_modules/braces": { - "version": "3.0.3", - "resolved": "https://registry.npmjs.org/braces/-/braces-3.0.3.tgz", - "integrity": "sha512-yQbXgO/OSZVD2IsiLlro+7Hf6Q18EJrKSEsdoMzKePKXct3gvD8oLcOQdIzGupr5Fj+EDe8gO/lxc1BzfMpxvA==", - "license": "MIT", - "dependencies": { - "fill-range": "^7.1.1" - }, - "engines": { - "node": ">=8" - } - }, - "node_modules/browserslist": { - "version": "4.24.4", - "resolved": "https://registry.npmjs.org/browserslist/-/browserslist-4.24.4.tgz", - "integrity": "sha512-KDi1Ny1gSePi1vm0q4oxSF8b4DR44GF4BbmS2YdhPLOEqd8pDviZOGH/GsmRwoWJ2+5Lr085X7naowMwKHDG1A==", - "funding": [ - { - "type": "opencollective", - "url": "https://opencollective.com/browserslist" - }, - { - "type": "tidelift", - "url": "https://tidelift.com/funding/github/npm/browserslist" - }, - { - "type": "github", - "url": "https://github.com/sponsors/ai" - } - ], - "license": "MIT", - "dependencies": { - "caniuse-lite": "^1.0.30001688", - "electron-to-chromium": "^1.5.73", - "node-releases": "^2.0.19", - "update-browserslist-db": "^1.1.1" - }, - "bin": { - "browserslist": "cli.js" - }, - "engines": { - "node": "^6 || ^7 || ^8 || ^9 || ^10 || ^11 || ^12 || >=13.7" - } - }, - "node_modules/buffer-builder": { - "version": "0.2.0", - "resolved": "https://registry.npmjs.org/buffer-builder/-/buffer-builder-0.2.0.tgz", - "integrity": "sha512-7VPMEPuYznPSoR21NE1zvd2Xna6c/CloiZCfcMXR1Jny6PjX0N4Nsa38zcBFo/FMK+BlA+FLKbJCQ0i2yxp+Xg==", - "devOptional": true, - "license": "MIT/X11" - }, - "node_modules/buffer-from": { - "version": "1.1.2", - "resolved": "https://registry.npmjs.org/buffer-from/-/buffer-from-1.1.2.tgz", - "integrity": "sha512-E+XQCRwSbaaiChtv6k6Dwgc+bx+Bs6vuKJHHl5kox/BaKbhiXzqQOwK4cO22yElGp2OCmjwVhT3HmxgyPGnJfQ==", - "license": "MIT", - "optional": true, - "peer": true - }, - "node_modules/callsites": { - "version": "3.1.0", - "resolved": "https://registry.npmjs.org/callsites/-/callsites-3.1.0.tgz", - "integrity": "sha512-P8BjAsXvZS+VIDUI11hHCQEv74YT67YUi5JJFNWIqL235sBmjX4+qx9Muvls5ivyNENctx46xQLQ3aTuE7ssaQ==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=6" - } - }, - "node_modules/caniuse-lite": { - "version": "1.0.30001697", - "resolved": "https://registry.npmjs.org/caniuse-lite/-/caniuse-lite-1.0.30001697.tgz", - "integrity": "sha512-GwNPlWJin8E+d7Gxq96jxM6w0w+VFeyyXRsjU58emtkYqnbwHqXm5uT2uCmO0RQE9htWknOP4xtBlLmM/gWxvQ==", - "funding": [ - { - "type": "opencollective", - "url": "https://opencollective.com/browserslist" - }, - { - "type": "tidelift", - "url": "https://tidelift.com/funding/github/npm/caniuse-lite" - }, - { - "type": "github", - "url": "https://github.com/sponsors/ai" - } - ], - "license": "CC-BY-4.0" - }, - "node_modules/ccount": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/ccount/-/ccount-2.0.1.tgz", - "integrity": "sha512-eyrF0jiFpY+3drT6383f1qhkbGsLSifNAjA61IUjZjmLCWjItY6LB9ft9YhoDgwfmclB2zhu51Lc7+95b8NRAg==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/chalk": { - "version": "4.1.2", - "resolved": "https://registry.npmjs.org/chalk/-/chalk-4.1.2.tgz", - "integrity": "sha512-oKnbhFyRIXpUuez8iBMmyEa4nbj4IOQyuhc/wy9kY7/WVPcwIO9VA668Pu8RkO7+0G76SLROeyw9CpQ061i4mA==", - "dev": true, - "license": "MIT", - "dependencies": { - "ansi-styles": "^4.1.0", - "supports-color": "^7.1.0" - }, - "engines": { - "node": ">=10" - }, - "funding": { - "url": "https://github.com/chalk/chalk?sponsor=1" - } - }, - "node_modules/character-entities": { - "version": "2.0.2", - "resolved": "https://registry.npmjs.org/character-entities/-/character-entities-2.0.2.tgz", - "integrity": "sha512-shx7oQ0Awen/BRIdkjkvz54PnEEI/EjwXDSIZp86/KKdbafHh1Df/RYGBhn4hbe2+uKC9FnT5UCEdyPz3ai9hQ==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/character-entities-html4": { - "version": "2.1.0", - "resolved": "https://registry.npmjs.org/character-entities-html4/-/character-entities-html4-2.1.0.tgz", - "integrity": "sha512-1v7fgQRj6hnSwFpq1Eu0ynr/CDEw0rXo2B61qXrLNdHZmPKgb7fqS1a2JwF0rISo9q77jDI8VMEHoApn8qDoZA==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/character-entities-legacy": { - "version": "3.0.0", - "resolved": "https://registry.npmjs.org/character-entities-legacy/-/character-entities-legacy-3.0.0.tgz", - "integrity": "sha512-RpPp0asT/6ufRm//AJVwpViZbGM/MkjQFxJccQRHmISF/22NBtsHqAWmL+/pmkPWoIUJdWyeVleTl1wydHATVQ==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/character-reference-invalid": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/character-reference-invalid/-/character-reference-invalid-2.0.1.tgz", - "integrity": "sha512-iBZ4F4wRbyORVsu0jPV7gXkOsGYjGHPmAyv+HiHG8gi5PtC9KI2j1+v8/tlibRvjoWX027ypmG/n0HtO5t7unw==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/color-convert": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/color-convert/-/color-convert-2.0.1.tgz", - "integrity": "sha512-RRECPsj7iu/xb5oKYcsFHSppFNnsj/52OVTRKb4zP5onXwVF3zVmmToNcOfGC+CRDpfK/U584fMg38ZHCaElKQ==", - "dev": true, - "license": "MIT", - "dependencies": { - "color-name": "~1.1.4" - }, - "engines": { - "node": ">=7.0.0" - } - }, - "node_modules/color-name": { - "version": "1.1.4", - "resolved": "https://registry.npmjs.org/color-name/-/color-name-1.1.4.tgz", - "integrity": "sha512-dOy+3AuW3a2wNbZHIuMZpTcgjGuLU/uBL/ubcZF9OXbDo8ff4O8yVp5Bf0efS8uEoYo5q4Fx7dY9OgQGXgAsQA==", - "dev": true, - "license": "MIT" - }, - "node_modules/colorjs.io": { - "version": "0.5.2", - "resolved": "https://registry.npmjs.org/colorjs.io/-/colorjs.io-0.5.2.tgz", - "integrity": "sha512-twmVoizEW7ylZSN32OgKdXRmo1qg+wT5/6C3xu5b9QsWzSFAhHLn2xd8ro0diCsKfCj1RdaTP/nrcW+vAoQPIw==", - "devOptional": true, - "license": "MIT" - }, - "node_modules/comma-separated-tokens": { - "version": "2.0.3", - "resolved": "https://registry.npmjs.org/comma-separated-tokens/-/comma-separated-tokens-2.0.3.tgz", - "integrity": "sha512-Fu4hJdvzeylCfQPp9SGWidpzrMs7tTrlu6Vb8XGaRGck8QSNZJJp538Wrb60Lax4fPwR64ViY468OIUTbRlGZg==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/commander": { - "version": "8.3.0", - "resolved": "https://registry.npmjs.org/commander/-/commander-8.3.0.tgz", - "integrity": "sha512-OkTL9umf+He2DZkUq8f8J9of7yL6RJKI24dVITBmNfZBmri9zYZQrKkuXiKhyfPSu8tUhnVBB1iKXevvnlR4Ww==", - "license": "MIT", - "engines": { - "node": ">= 12" - } - }, - "node_modules/concat-map": { - "version": "0.0.1", - "resolved": "https://registry.npmjs.org/concat-map/-/concat-map-0.0.1.tgz", - "integrity": "sha512-/Srv4dswyQNBfohGpz9o6Yb3Gz3SrUDqBH5rTuhGR7ahtlbYKnVxw2bCFMRljaA7EXHaXZ8wsHdodFvbkhKmqg==", - "dev": true, - "license": "MIT" - }, - "node_modules/convert-source-map": { - "version": "2.0.0", - "resolved": "https://registry.npmjs.org/convert-source-map/-/convert-source-map-2.0.0.tgz", - "integrity": "sha512-Kvp459HrV2FEJ1CAsi1Ku+MY3kasH19TFykTz2xWmMeq6bk2NU3XXvfJ+Q61m0xktWwt+1HSYf3JZsTms3aRJg==", - "dev": true, - "license": "MIT" - }, - "node_modules/cookie": { - "version": "1.0.2", - "resolved": "https://registry.npmjs.org/cookie/-/cookie-1.0.2.tgz", - "integrity": "sha512-9Kr/j4O16ISv8zBBhJoi4bXOYNTkFLOqSL3UDB0njXxCXNezjeyVrJyGOWtgfs/q2km1gwBcfH8q1yEGoMYunA==", - "license": "MIT", - "engines": { - "node": ">=18" - } - }, - "node_modules/cross-spawn": { - "version": "7.0.6", - "resolved": "https://registry.npmjs.org/cross-spawn/-/cross-spawn-7.0.6.tgz", - "integrity": "sha512-uV2QOWP2nWzsy2aMp8aRibhi9dlzF5Hgh5SHaB9OiTGEyDTiJJyx0uy51QXdyWbtAHNua4XJzUKca3OzKUd3vA==", - "dev": true, - "license": "MIT", - "dependencies": { - "path-key": "^3.1.0", - "shebang-command": "^2.0.0", - "which": "^2.0.1" - }, - "engines": { - "node": ">= 8" - } - }, - "node_modules/csstype": { - "version": "3.1.3", - "resolved": "https://registry.npmjs.org/csstype/-/csstype-3.1.3.tgz", - "integrity": "sha512-M1uQkMl8rQK/szD0LNhtqxIPLpimGm8sOBwU7lLnCpSbTyY3yeU1Vc7l4KT5zT4s/yOxHH5O7tIuuLOCnLADRw==", - "license": "MIT" - }, - "node_modules/daisyui": { - "version": "5.0.12", - "resolved": "https://registry.npmjs.org/daisyui/-/daisyui-5.0.12.tgz", - "integrity": "sha512-01DU0eYBcHgPtuf5fxcrkGkIN6/Uyaqmkle5Yo3ZyW9YVAu036ALZbjv2KH5euvUbeQ4r9q3gAarGcf7Tywhng==", - "license": "MIT", - "funding": { - "url": "https://github.com/saadeghi/daisyui?sponsor=1" - } - }, - "node_modules/debug": { - "version": "4.4.0", - "resolved": "https://registry.npmjs.org/debug/-/debug-4.4.0.tgz", - "integrity": "sha512-6WTZ/IxCY/T6BALoZHaE4ctp9xm+Z5kY/pzYaCHRFeyVhojxlrm+46y68HA6hr0TcwEssoxNiDEUJQjfPZ/RYA==", - "license": "MIT", - "dependencies": { - "ms": "^2.1.3" - }, - "engines": { - "node": ">=6.0" - }, - "peerDependenciesMeta": { - "supports-color": { - "optional": true - } - } - }, - "node_modules/decode-named-character-reference": { - "version": "1.0.2", - "resolved": "https://registry.npmjs.org/decode-named-character-reference/-/decode-named-character-reference-1.0.2.tgz", - "integrity": "sha512-O8x12RzrUF8xyVcY0KJowWsmaJxQbmy0/EtnNtHRpsOcT7dFk5W598coHqBVpmWo1oQQfsCqfCmkZN5DJrZVdg==", - "license": "MIT", - "dependencies": { - "character-entities": "^2.0.0" - }, - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/deep-is": { - "version": "0.1.4", - "resolved": "https://registry.npmjs.org/deep-is/-/deep-is-0.1.4.tgz", - "integrity": "sha512-oIPzksmTg4/MriiaYGO+okXDT7ztn/w3Eptv/+gSIdMdKsJo0u4CfYNFJPy+4SKMuCqGw2wxnA+URMg3t8a/bQ==", - "dev": true, - "license": "MIT" - }, - "node_modules/dequal": { - "version": "2.0.3", - "resolved": "https://registry.npmjs.org/dequal/-/dequal-2.0.3.tgz", - "integrity": "sha512-0je+qPKHEMohvfRTCEo3CrPG6cAzAYgmzKyxRiYSSDkS6eGJdyVJm7WaYA5ECaAD9wLB2T4EEeymA5aFVcYXCA==", - "license": "MIT", - "engines": { - "node": ">=6" - } - }, - "node_modules/detect-libc": { - "version": "2.0.3", - "resolved": "https://registry.npmjs.org/detect-libc/-/detect-libc-2.0.3.tgz", - "integrity": "sha512-bwy0MGW55bG41VqxxypOsdSdGqLwXPI/focwgTYCFMbdUiBAxLg9CFzG08sz2aqzknwiX7Hkl0bQENjg8iLByw==", - "license": "Apache-2.0", - "engines": { - "node": ">=8" - } - }, - "node_modules/devlop": { - "version": "1.1.0", - "resolved": "https://registry.npmjs.org/devlop/-/devlop-1.1.0.tgz", - "integrity": "sha512-RWmIqhcFf1lRYBvNmr7qTNuyCt/7/ns2jbpp1+PalgE/rDQcBT0fioSMUpJ93irlUhC5hrg4cYqe6U+0ImW0rA==", - "license": "MIT", - "dependencies": { - "dequal": "^2.0.0" - }, - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/dexie": { - "version": "4.0.11", - "resolved": "https://registry.npmjs.org/dexie/-/dexie-4.0.11.tgz", - "integrity": "sha512-SOKO002EqlvBYYKQSew3iymBoN2EQ4BDw/3yprjh7kAfFzjBYkaMNa/pZvcA7HSWlcKSQb9XhPe3wKyQ0x4A8A==", - "license": "Apache-2.0" - }, - "node_modules/electron-to-chromium": { - "version": "1.5.91", - "resolved": "https://registry.npmjs.org/electron-to-chromium/-/electron-to-chromium-1.5.91.tgz", - "integrity": "sha512-sNSHHyq048PFmZY4S90ax61q+gLCs0X0YmcOII9wG9S2XwbVr+h4VW2wWhnbp/Eys3cCwTxVF292W3qPaxIapQ==", - "license": "ISC" - }, - "node_modules/enhanced-resolve": { - "version": "5.18.1", - "resolved": "https://registry.npmjs.org/enhanced-resolve/-/enhanced-resolve-5.18.1.tgz", - "integrity": "sha512-ZSW3ma5GkcQBIpwZTSRAI8N71Uuwgs93IezB7mf7R60tC8ZbJideoDNKjHn2O9KIlx6rkGTTEk1xUCK2E1Y2Yg==", - "license": "MIT", - "dependencies": { - "graceful-fs": "^4.2.4", - "tapable": "^2.2.0" - }, - "engines": { - "node": ">=10.13.0" - } - }, - "node_modules/entities": { - "version": "4.5.0", - "resolved": "https://registry.npmjs.org/entities/-/entities-4.5.0.tgz", - "integrity": "sha512-V0hjH4dGPh9Ao5p0MoRY6BVqtwCjhz6vI5LT8AJ55H+4g9/4vbHx1I54fS0XuclLhDHArPQCiMjDxjaL8fPxhw==", - "license": "BSD-2-Clause", - "engines": { - "node": ">=0.12" - }, - "funding": { - "url": "https://github.com/fb55/entities?sponsor=1" - } - }, - "node_modules/esbuild": { - "version": "0.24.2", - "resolved": "https://registry.npmjs.org/esbuild/-/esbuild-0.24.2.tgz", - "integrity": "sha512-+9egpBW8I3CD5XPe0n6BfT5fxLzxrlDzqydF3aviG+9ni1lDC/OvMHcxqEFV0+LANZG5R1bFMWfUrjVsdwxJvA==", - "hasInstallScript": true, - "license": "MIT", - "bin": { - "esbuild": "bin/esbuild" - }, - "engines": { - "node": ">=18" - }, - "optionalDependencies": { - "@esbuild/aix-ppc64": "0.24.2", - "@esbuild/android-arm": "0.24.2", - "@esbuild/android-arm64": "0.24.2", - "@esbuild/android-x64": "0.24.2", - "@esbuild/darwin-arm64": "0.24.2", - "@esbuild/darwin-x64": "0.24.2", - "@esbuild/freebsd-arm64": "0.24.2", - "@esbuild/freebsd-x64": "0.24.2", - "@esbuild/linux-arm": "0.24.2", - "@esbuild/linux-arm64": "0.24.2", - "@esbuild/linux-ia32": "0.24.2", - "@esbuild/linux-loong64": "0.24.2", - "@esbuild/linux-mips64el": "0.24.2", - "@esbuild/linux-ppc64": "0.24.2", - "@esbuild/linux-riscv64": "0.24.2", - "@esbuild/linux-s390x": "0.24.2", - "@esbuild/linux-x64": "0.24.2", - "@esbuild/netbsd-arm64": "0.24.2", - "@esbuild/netbsd-x64": "0.24.2", - "@esbuild/openbsd-arm64": "0.24.2", - "@esbuild/openbsd-x64": "0.24.2", - "@esbuild/sunos-x64": "0.24.2", - "@esbuild/win32-arm64": "0.24.2", - "@esbuild/win32-ia32": "0.24.2", - "@esbuild/win32-x64": "0.24.2" - } - }, - "node_modules/escalade": { - "version": "3.2.0", - "resolved": "https://registry.npmjs.org/escalade/-/escalade-3.2.0.tgz", - "integrity": "sha512-WUj2qlxaQtO4g6Pq5c29GTcWGDyd8itL8zTlipgECz3JesAiiOKotd8JU6otB3PACgG6xkJUyVhboMS+bje/jA==", - "license": "MIT", - "engines": { - "node": ">=6" - } - }, - "node_modules/escape-string-regexp": { - "version": "4.0.0", - "resolved": "https://registry.npmjs.org/escape-string-regexp/-/escape-string-regexp-4.0.0.tgz", - "integrity": "sha512-TtpcNJ3XAzx3Gq8sWRzJaVajRs0uVxA2YAkdb1jm2YkPz4G6egUFAyA3n5vtEIZefPk5Wa4UXbKuS5fKkJWdgA==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=10" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/eslint": { - "version": "9.19.0", - "resolved": "https://registry.npmjs.org/eslint/-/eslint-9.19.0.tgz", - "integrity": "sha512-ug92j0LepKlbbEv6hD911THhoRHmbdXt2gX+VDABAW/Ir7D3nqKdv5Pf5vtlyY6HQMTEP2skXY43ueqTCWssEA==", - "dev": true, - "license": "MIT", - "dependencies": { - "@eslint-community/eslint-utils": "^4.2.0", - "@eslint-community/regexpp": "^4.12.1", - "@eslint/config-array": "^0.19.0", - "@eslint/core": "^0.10.0", - "@eslint/eslintrc": "^3.2.0", - "@eslint/js": "9.19.0", - "@eslint/plugin-kit": "^0.2.5", - "@humanfs/node": "^0.16.6", - "@humanwhocodes/module-importer": "^1.0.1", - "@humanwhocodes/retry": "^0.4.1", - "@types/estree": "^1.0.6", - "@types/json-schema": "^7.0.15", - "ajv": "^6.12.4", - "chalk": "^4.0.0", - "cross-spawn": "^7.0.6", - "debug": "^4.3.2", - "escape-string-regexp": "^4.0.0", - "eslint-scope": "^8.2.0", - "eslint-visitor-keys": "^4.2.0", - "espree": "^10.3.0", - "esquery": "^1.5.0", - "esutils": "^2.0.2", - "fast-deep-equal": "^3.1.3", - "file-entry-cache": "^8.0.0", - "find-up": "^5.0.0", - "glob-parent": "^6.0.2", - "ignore": "^5.2.0", - "imurmurhash": "^0.1.4", - "is-glob": "^4.0.0", - "json-stable-stringify-without-jsonify": "^1.0.1", - "lodash.merge": "^4.6.2", - "minimatch": "^3.1.2", - "natural-compare": "^1.4.0", - "optionator": "^0.9.3" - }, - "bin": { - "eslint": "bin/eslint.js" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "url": "https://eslint.org/donate" - }, - "peerDependencies": { - "jiti": "*" - }, - "peerDependenciesMeta": { - "jiti": { - "optional": true - } - } - }, - "node_modules/eslint-plugin-react-hooks": { - "version": "5.1.0", - "resolved": "https://registry.npmjs.org/eslint-plugin-react-hooks/-/eslint-plugin-react-hooks-5.1.0.tgz", - "integrity": "sha512-mpJRtPgHN2tNAvZ35AMfqeB3Xqeo273QxrHJsbBEPWODRM4r0yB6jfoROqKEYrOn27UtRPpcpHc2UqyBSuUNTw==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=10" - }, - "peerDependencies": { - "eslint": "^3.0.0 || ^4.0.0 || ^5.0.0 || ^6.0.0 || ^7.0.0 || ^8.0.0-0 || ^9.0.0" - } - }, - "node_modules/eslint-plugin-react-refresh": { - "version": "0.4.18", - "resolved": "https://registry.npmjs.org/eslint-plugin-react-refresh/-/eslint-plugin-react-refresh-0.4.18.tgz", - "integrity": "sha512-IRGEoFn3OKalm3hjfolEWGqoF/jPqeEYFp+C8B0WMzwGwBMvlRDQd06kghDhF0C61uJ6WfSDhEZE/sAQjduKgw==", - "dev": true, - "license": "MIT", - "peerDependencies": { - "eslint": ">=8.40" - } - }, - "node_modules/eslint-scope": { - "version": "8.2.0", - "resolved": "https://registry.npmjs.org/eslint-scope/-/eslint-scope-8.2.0.tgz", - "integrity": "sha512-PHlWUfG6lvPc3yvP5A4PNyBL1W8fkDUccmI21JUu/+GKZBoH/W5u6usENXUrWFRsyoW5ACUjFGgAFQp5gUlb/A==", - "dev": true, - "license": "BSD-2-Clause", - "dependencies": { - "esrecurse": "^4.3.0", - "estraverse": "^5.2.0" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "url": "https://opencollective.com/eslint" - } - }, - "node_modules/eslint-visitor-keys": { - "version": "4.2.0", - "resolved": "https://registry.npmjs.org/eslint-visitor-keys/-/eslint-visitor-keys-4.2.0.tgz", - "integrity": "sha512-UyLnSehNt62FFhSwjZlHmeokpRK59rcz29j+F1/aDgbkbRTk7wIc9XzdoasMUbRNKDM0qQt/+BJ4BrpFeABemw==", - "dev": true, - "license": "Apache-2.0", - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "url": "https://opencollective.com/eslint" - } - }, - "node_modules/espree": { - "version": "10.3.0", - "resolved": "https://registry.npmjs.org/espree/-/espree-10.3.0.tgz", - "integrity": "sha512-0QYC8b24HWY8zjRnDTL6RiHfDbAWn63qb4LMj1Z4b076A4une81+z03Kg7l7mn/48PUTqoLptSXez8oknU8Clg==", - "dev": true, - "license": "BSD-2-Clause", - "dependencies": { - "acorn": "^8.14.0", - "acorn-jsx": "^5.3.2", - "eslint-visitor-keys": "^4.2.0" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "url": "https://opencollective.com/eslint" - } - }, - "node_modules/esquery": { - "version": "1.6.0", - "resolved": "https://registry.npmjs.org/esquery/-/esquery-1.6.0.tgz", - "integrity": "sha512-ca9pw9fomFcKPvFLXhBKUK90ZvGibiGOvRJNbjljY7s7uq/5YO4BOzcYtJqExdx99rF6aAcnRxHmcUHcz6sQsg==", - "dev": true, - "license": "BSD-3-Clause", - "dependencies": { - "estraverse": "^5.1.0" - }, - "engines": { - "node": ">=0.10" - } - }, - "node_modules/esrecurse": { - "version": "4.3.0", - "resolved": "https://registry.npmjs.org/esrecurse/-/esrecurse-4.3.0.tgz", - "integrity": "sha512-KmfKL3b6G+RXvP8N1vr3Tq1kL/oCFgn2NYXEtqP8/L3pKapUA4G8cFVaoF3SU323CD4XypR/ffioHmkti6/Tag==", - "dev": true, - "license": "BSD-2-Clause", - "dependencies": { - "estraverse": "^5.2.0" - }, - "engines": { - "node": ">=4.0" - } - }, - "node_modules/estraverse": { - "version": "5.3.0", - "resolved": "https://registry.npmjs.org/estraverse/-/estraverse-5.3.0.tgz", - "integrity": "sha512-MMdARuVEQziNTeJD8DgMqmhwR11BRQ/cBP+pLtYdSTnf3MIO8fFeiINEbX36ZdNlfU/7A9f3gUw49B3oQsvwBA==", - "dev": true, - "license": "BSD-2-Clause", - "engines": { - "node": ">=4.0" - } - }, - "node_modules/estree-util-is-identifier-name": { - "version": "3.0.0", - "resolved": "https://registry.npmjs.org/estree-util-is-identifier-name/-/estree-util-is-identifier-name-3.0.0.tgz", - "integrity": "sha512-hFtqIDZTIUZ9BXLb8y4pYGyk6+wekIivNVTcmvk8NoOh+VeRn5y6cEHzbURrWbfp1fIqdVipilzj+lfaadNZmg==", - "license": "MIT", - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/esutils": { - "version": "2.0.3", - "resolved": "https://registry.npmjs.org/esutils/-/esutils-2.0.3.tgz", - "integrity": "sha512-kVscqXk4OCp68SZ0dkgEKVi6/8ij300KBWTJq32P/dYeWTSwK41WyTxalN1eRmA5Z9UU/LX9D7FWSmV9SAYx6g==", - "dev": true, - "license": "BSD-2-Clause", - "engines": { - "node": ">=0.10.0" - } - }, - "node_modules/extend": { - "version": "3.0.2", - "resolved": "https://registry.npmjs.org/extend/-/extend-3.0.2.tgz", - "integrity": "sha512-fjquC59cD7CyW6urNXK0FBufkZcoiGG80wTuPujX590cB5Ttln20E2UB4S/WARVqhXffZl2LNgS+gQdPIIim/g==", - "license": "MIT" - }, - "node_modules/fast-deep-equal": { - "version": "3.1.3", - "resolved": "https://registry.npmjs.org/fast-deep-equal/-/fast-deep-equal-3.1.3.tgz", - "integrity": "sha512-f3qQ9oQy9j2AhBe/H9VC91wLmKBCCU/gDOnKNAYG5hswO7BLKj09Hc5HYNz9cGI++xlpDCIgDaitVs03ATR84Q==", - "dev": true, - "license": "MIT" - }, - "node_modules/fast-glob": { - "version": "3.3.3", - "resolved": "https://registry.npmjs.org/fast-glob/-/fast-glob-3.3.3.tgz", - "integrity": "sha512-7MptL8U0cqcFdzIzwOTHoilX9x5BrNqye7Z/LuC7kCMRio1EMSyqRK3BEAUD7sXRq4iT4AzTVuZdhgQ2TCvYLg==", - "dev": true, - "license": "MIT", - "dependencies": { - "@nodelib/fs.stat": "^2.0.2", - "@nodelib/fs.walk": "^1.2.3", - "glob-parent": "^5.1.2", - "merge2": "^1.3.0", - "micromatch": "^4.0.8" - }, - "engines": { - "node": ">=8.6.0" - } - }, - "node_modules/fast-glob/node_modules/glob-parent": { - "version": "5.1.2", - "resolved": "https://registry.npmjs.org/glob-parent/-/glob-parent-5.1.2.tgz", - "integrity": "sha512-AOIgSQCepiJYwP3ARnGx+5VnTu2HBYdzbGP45eLw1vr3zB3vZLeyed1sC9hnbcOc9/SrMyM5RPQrkGz4aS9Zow==", - "dev": true, - "license": "ISC", - "dependencies": { - "is-glob": "^4.0.1" - }, - "engines": { - "node": ">= 6" - } - }, - "node_modules/fast-json-stable-stringify": { - "version": "2.1.0", - "resolved": "https://registry.npmjs.org/fast-json-stable-stringify/-/fast-json-stable-stringify-2.1.0.tgz", - "integrity": "sha512-lhd/wF+Lk98HZoTCtlVraHtfh5XYijIjalXck7saUtuanSDyLMxnHhSXEDJqHxD7msR8D0uCmqlkwjCV8xvwHw==", - "dev": true, - "license": "MIT" - }, - "node_modules/fast-levenshtein": { - "version": "2.0.6", - "resolved": "https://registry.npmjs.org/fast-levenshtein/-/fast-levenshtein-2.0.6.tgz", - "integrity": "sha512-DCXu6Ifhqcks7TZKY3Hxp3y6qphY5SJZmrWMDrKcERSOXWQdMhU9Ig/PYrzyw/ul9jOIyh0N4M0tbC5hodg8dw==", - "dev": true, - "license": "MIT" - }, - "node_modules/fastq": { - "version": "1.19.0", - "resolved": "https://registry.npmjs.org/fastq/-/fastq-1.19.0.tgz", - "integrity": "sha512-7SFSRCNjBQIZH/xZR3iy5iQYR8aGBE0h3VG6/cwlbrpdciNYBMotQav8c1XI3HjHH+NikUpP53nPdlZSdWmFzA==", - "dev": true, - "license": "ISC", - "dependencies": { - "reusify": "^1.0.4" - } - }, - "node_modules/fflate": { - "version": "0.8.2", - "resolved": "https://registry.npmjs.org/fflate/-/fflate-0.8.2.tgz", - "integrity": "sha512-cPJU47OaAoCbg0pBvzsgpTPhmhqI5eJjh/JIu8tPj5q+T7iLvW/JAYUqmE7KOB4R1ZyEhzBaIQpQpardBF5z8A==", - "dev": true, - "license": "MIT" - }, - "node_modules/file-entry-cache": { - "version": "8.0.0", - "resolved": "https://registry.npmjs.org/file-entry-cache/-/file-entry-cache-8.0.0.tgz", - "integrity": "sha512-XXTUwCvisa5oacNGRP9SfNtYBNAMi+RPwBFmblZEF7N7swHYQS6/Zfk7SRwx4D5j3CH211YNRco1DEMNVfZCnQ==", - "dev": true, - "license": "MIT", - "dependencies": { - "flat-cache": "^4.0.0" - }, - "engines": { - "node": ">=16.0.0" - } - }, - "node_modules/file-selector": { - "version": "2.1.2", - "resolved": "https://registry.npmjs.org/file-selector/-/file-selector-2.1.2.tgz", - "integrity": "sha512-QgXo+mXTe8ljeqUFaX3QVHc5osSItJ/Km+xpocx0aSqWGMSCf6qYs/VnzZgS864Pjn5iceMRFigeAV7AfTlaig==", - "license": "MIT", - "dependencies": { - "tslib": "^2.7.0" - }, - "engines": { - "node": ">= 12" - } - }, - "node_modules/fill-range": { - "version": "7.1.1", - "resolved": "https://registry.npmjs.org/fill-range/-/fill-range-7.1.1.tgz", - "integrity": "sha512-YsGpe3WHLK8ZYi4tWDg2Jy3ebRz2rXowDxnld4bkQB00cc/1Zw9AWnC0i9ztDJitivtQvaI9KaLyKrc+hBW0yg==", - "license": "MIT", - "dependencies": { - "to-regex-range": "^5.0.1" - }, - "engines": { - "node": ">=8" - } - }, - "node_modules/find-up": { - "version": "5.0.0", - "resolved": "https://registry.npmjs.org/find-up/-/find-up-5.0.0.tgz", - "integrity": "sha512-78/PXT1wlLLDgTzDs7sjq9hzz0vXD+zn+7wypEe4fXQxCmdmqfGsEPQxmiCSQI3ajFV91bVSsvNtrJRiW6nGng==", - "dev": true, - "license": "MIT", - "dependencies": { - "locate-path": "^6.0.0", - "path-exists": "^4.0.0" - }, - "engines": { - "node": ">=10" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/flat-cache": { - "version": "4.0.1", - "resolved": "https://registry.npmjs.org/flat-cache/-/flat-cache-4.0.1.tgz", - "integrity": "sha512-f7ccFPK3SXFHpx15UIGyRJ/FJQctuKZ0zVuN3frBo4HnK3cay9VEW0R6yPYFHC0AgqhukPzKjq22t5DmAyqGyw==", - "dev": true, - "license": "MIT", - "dependencies": { - "flatted": "^3.2.9", - "keyv": "^4.5.4" - }, - "engines": { - "node": ">=16" - } - }, - "node_modules/flatted": { - "version": "3.3.2", - "resolved": "https://registry.npmjs.org/flatted/-/flatted-3.3.2.tgz", - "integrity": "sha512-AiwGJM8YcNOaobumgtng+6NHuOqC3A7MixFeDafM3X9cIUM+xUXoS5Vfgf+OihAYe20fxqNM9yPBXJzRtZ/4eA==", - "dev": true, - "license": "ISC" - }, - "node_modules/fraction.js": { - "version": "4.3.7", - "resolved": "https://registry.npmjs.org/fraction.js/-/fraction.js-4.3.7.tgz", - "integrity": "sha512-ZsDfxO51wGAXREY55a7la9LScWpwv9RxIrYABrlvOFBlH/ShPnrtsXeuUIfXKKOVicNxQ+o8JTbJvjS4M89yew==", - "license": "MIT", - "engines": { - "node": "*" - }, - "funding": { - "type": "patreon", - "url": "https://github.com/sponsors/rawify" - } - }, - "node_modules/fsevents": { - "version": "2.3.3", - "resolved": "https://registry.npmjs.org/fsevents/-/fsevents-2.3.3.tgz", - "integrity": "sha512-5xoDfX+fL7faATnagmWPpbFtwh/R77WmMMqqHGS65C3vvB0YHrgF+B1YmZ3441tMj5n63k0212XNoJwzlhffQw==", - "hasInstallScript": true, - "license": "MIT", - "optional": true, - "os": [ - "darwin" - ], - "engines": { - "node": "^8.16.0 || ^10.6.0 || >=11.0.0" - } - }, - "node_modules/gensync": { - "version": "1.0.0-beta.2", - "resolved": "https://registry.npmjs.org/gensync/-/gensync-1.0.0-beta.2.tgz", - "integrity": "sha512-3hN7NaskYvMDLQY55gnW3NQ+mesEAepTqlg+VEbj7zzqEMBVNhzcGYYeqFo/TlYz6eQiFcp1HcsCZO+nGgS8zg==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=6.9.0" - } - }, - "node_modules/glob-parent": { - "version": "6.0.2", - "resolved": "https://registry.npmjs.org/glob-parent/-/glob-parent-6.0.2.tgz", - "integrity": "sha512-XxwI8EOhVQgWp6iDL+3b0r86f4d6AX6zSU55HfB4ydCEuXLXc5FcYeOu+nnGftS4TEju/11rt4KJPTMgbfmv4A==", - "dev": true, - "license": "ISC", - "dependencies": { - "is-glob": "^4.0.3" - }, - "engines": { - "node": ">=10.13.0" - } - }, - "node_modules/globals": { - "version": "15.14.0", - "resolved": "https://registry.npmjs.org/globals/-/globals-15.14.0.tgz", - "integrity": "sha512-OkToC372DtlQeje9/zHIo5CT8lRP/FUgEOKBEhU4e0abL7J7CD24fD9ohiLN5hagG/kWCYj4K5oaxxtj2Z0Dig==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=18" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/goober": { - "version": "2.1.16", - "resolved": "https://registry.npmjs.org/goober/-/goober-2.1.16.tgz", - "integrity": "sha512-erjk19y1U33+XAMe1VTvIONHYoSqE4iS7BYUZfHaqeohLmnC0FdxEh7rQU+6MZ4OajItzjZFSRtVANrQwNq6/g==", - "license": "MIT", - "peerDependencies": { - "csstype": "^3.0.10" - } - }, - "node_modules/graceful-fs": { - "version": "4.2.11", - "resolved": "https://registry.npmjs.org/graceful-fs/-/graceful-fs-4.2.11.tgz", - "integrity": "sha512-RbJ5/jmFcNNCcDV5o9eTnBLJ/HszWV0P73bc+Ff4nS/rJj+YaS6IGyiOL0VoBYX+l1Wrl3k63h/KrH+nhJ0XvQ==", - "license": "ISC" - }, - "node_modules/graphemer": { - "version": "1.4.0", - "resolved": "https://registry.npmjs.org/graphemer/-/graphemer-1.4.0.tgz", - "integrity": "sha512-EtKwoO6kxCL9WO5xipiHTZlSzBm7WLT627TqC/uVRd0HKmq8NXyebnNYxDoBi7wt8eTWrUrKXCOVaFq9x1kgag==", - "dev": true, - "license": "MIT" - }, - "node_modules/has-flag": { - "version": "4.0.0", - "resolved": "https://registry.npmjs.org/has-flag/-/has-flag-4.0.0.tgz", - "integrity": "sha512-EykJT/Q1KjTWctppgIAgfSO0tKVuZUjhgMr17kqTumMl6Afv3EISleU7qZUzoXDFTAHTDC4NOoG/ZxU3EvlMPQ==", - "devOptional": true, - "license": "MIT", - "engines": { - "node": ">=8" - } - }, - "node_modules/hast-util-from-dom": { - "version": "5.0.1", - "resolved": "https://registry.npmjs.org/hast-util-from-dom/-/hast-util-from-dom-5.0.1.tgz", - "integrity": "sha512-N+LqofjR2zuzTjCPzyDUdSshy4Ma6li7p/c3pA78uTwzFgENbgbUrm2ugwsOdcjI1muO+o6Dgzp9p8WHtn/39Q==", - "license": "ISC", - "dependencies": { - "@types/hast": "^3.0.0", - "hastscript": "^9.0.0", - "web-namespaces": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/hast-util-from-html": { - "version": "2.0.3", - "resolved": "https://registry.npmjs.org/hast-util-from-html/-/hast-util-from-html-2.0.3.tgz", - "integrity": "sha512-CUSRHXyKjzHov8yKsQjGOElXy/3EKpyX56ELnkHH34vDVw1N1XSQ1ZcAvTyAPtGqLTuKP/uxM+aLkSPqF/EtMw==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0", - "devlop": "^1.1.0", - "hast-util-from-parse5": "^8.0.0", - "parse5": "^7.0.0", - "vfile": "^6.0.0", - "vfile-message": "^4.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/hast-util-from-html-isomorphic": { - "version": "2.0.0", - "resolved": "https://registry.npmjs.org/hast-util-from-html-isomorphic/-/hast-util-from-html-isomorphic-2.0.0.tgz", - "integrity": "sha512-zJfpXq44yff2hmE0XmwEOzdWin5xwH+QIhMLOScpX91e/NSGPsAzNCvLQDIEPyO2TXi+lBmU6hjLIhV8MwP2kw==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0", - "hast-util-from-dom": "^5.0.0", - "hast-util-from-html": "^2.0.0", - "unist-util-remove-position": "^5.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/hast-util-from-parse5": { - "version": "8.0.2", - "resolved": "https://registry.npmjs.org/hast-util-from-parse5/-/hast-util-from-parse5-8.0.2.tgz", - "integrity": "sha512-SfMzfdAi/zAoZ1KkFEyyeXBn7u/ShQrfd675ZEE9M3qj+PMFX05xubzRyF76CCSJu8au9jgVxDV1+okFvgZU4A==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0", - "@types/unist": "^3.0.0", - "devlop": "^1.0.0", - "hastscript": "^9.0.0", - "property-information": "^6.0.0", - "vfile": "^6.0.0", - "vfile-location": "^5.0.0", - "web-namespaces": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/hast-util-is-element": { - "version": "3.0.0", - "resolved": "https://registry.npmjs.org/hast-util-is-element/-/hast-util-is-element-3.0.0.tgz", - "integrity": "sha512-Val9mnv2IWpLbNPqc/pUem+a7Ipj2aHacCwgNfTiK0vJKl0LF+4Ba4+v1oPHFpf3bLYmreq0/l3Gud9S5OH42g==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/hast-util-parse-selector": { - "version": "4.0.0", - "resolved": "https://registry.npmjs.org/hast-util-parse-selector/-/hast-util-parse-selector-4.0.0.tgz", - "integrity": "sha512-wkQCkSYoOGCRKERFWcxMVMOcYE2K1AaNLU8DXS9arxnLOUEWbOXKXiJUNzEpqZ3JOKpnha3jkFrumEjVliDe7A==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/hast-util-to-jsx-runtime": { - "version": "2.3.2", - "resolved": "https://registry.npmjs.org/hast-util-to-jsx-runtime/-/hast-util-to-jsx-runtime-2.3.2.tgz", - "integrity": "sha512-1ngXYb+V9UT5h+PxNRa1O1FYguZK/XL+gkeqvp7EdHlB9oHUG0eYRo/vY5inBdcqo3RkPMC58/H94HvkbfGdyg==", - "license": "MIT", - "dependencies": { - "@types/estree": "^1.0.0", - "@types/hast": "^3.0.0", - "@types/unist": "^3.0.0", - "comma-separated-tokens": "^2.0.0", - "devlop": "^1.0.0", - "estree-util-is-identifier-name": "^3.0.0", - "hast-util-whitespace": "^3.0.0", - "mdast-util-mdx-expression": "^2.0.0", - "mdast-util-mdx-jsx": "^3.0.0", - "mdast-util-mdxjs-esm": "^2.0.0", - "property-information": "^6.0.0", - "space-separated-tokens": "^2.0.0", - "style-to-object": "^1.0.0", - "unist-util-position": "^5.0.0", - "vfile-message": "^4.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/hast-util-to-text": { - "version": "4.0.2", - "resolved": "https://registry.npmjs.org/hast-util-to-text/-/hast-util-to-text-4.0.2.tgz", - "integrity": "sha512-KK6y/BN8lbaq654j7JgBydev7wuNMcID54lkRav1P0CaE1e47P72AWWPiGKXTJU271ooYzcvTAn/Zt0REnvc7A==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0", - "@types/unist": "^3.0.0", - "hast-util-is-element": "^3.0.0", - "unist-util-find-after": "^5.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/hast-util-whitespace": { - "version": "3.0.0", - "resolved": "https://registry.npmjs.org/hast-util-whitespace/-/hast-util-whitespace-3.0.0.tgz", - "integrity": "sha512-88JUN06ipLwsnv+dVn+OIYOvAuvBMy/Qoi6O7mQHxdPXpjy+Cd6xRkWwux7DKO+4sYILtLBRIKgsdpS2gQc7qw==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/hastscript": { - "version": "9.0.0", - "resolved": "https://registry.npmjs.org/hastscript/-/hastscript-9.0.0.tgz", - "integrity": "sha512-jzaLBGavEDKHrc5EfFImKN7nZKKBdSLIdGvCwDZ9TfzbF2ffXiov8CKE445L2Z1Ek2t/m4SKQ2j6Ipv7NyUolw==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0", - "comma-separated-tokens": "^2.0.0", - "hast-util-parse-selector": "^4.0.0", - "property-information": "^6.0.0", - "space-separated-tokens": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/highlight.js": { - "version": "11.11.1", - "resolved": "https://registry.npmjs.org/highlight.js/-/highlight.js-11.11.1.tgz", - "integrity": "sha512-Xwwo44whKBVCYoliBQwaPvtd/2tYFkRQtXDWj1nackaV2JPXx3L0+Jvd8/qCJ2p+ML0/XVkJ2q+Mr+UVdpJK5w==", - "license": "BSD-3-Clause", - "engines": { - "node": ">=12.0.0" - } - }, - "node_modules/html-url-attributes": { - "version": "3.0.1", - "resolved": "https://registry.npmjs.org/html-url-attributes/-/html-url-attributes-3.0.1.tgz", - "integrity": "sha512-ol6UPyBWqsrO6EJySPz2O7ZSr856WDrEzM5zMqp+FJJLGMW35cLYmmZnl0vztAZxRUoNZJFTCohfjuIJ8I4QBQ==", - "license": "MIT", - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/ignore": { - "version": "5.3.2", - "resolved": "https://registry.npmjs.org/ignore/-/ignore-5.3.2.tgz", - "integrity": "sha512-hsBTNUqQTDwkWtcdYI2i06Y/nUBEsNEDJKjWdigLvegy8kDuJAS8uRlpkkcQpyEXL0Z/pjDy5HBmMjRCJ2gq+g==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">= 4" - } - }, - "node_modules/immutable": { - "version": "5.0.3", - "resolved": "https://registry.npmjs.org/immutable/-/immutable-5.0.3.tgz", - "integrity": "sha512-P8IdPQHq3lA1xVeBRi5VPqUm5HDgKnx0Ru51wZz5mjxHr5n3RWhjIpOFU7ybkUxfB+5IToy+OLaHYDBIWsv+uw==", - "devOptional": true, - "license": "MIT" - }, - "node_modules/import-fresh": { - "version": "3.3.1", - "resolved": "https://registry.npmjs.org/import-fresh/-/import-fresh-3.3.1.tgz", - "integrity": "sha512-TR3KfrTZTYLPB6jUjfx6MF9WcWrHL9su5TObK4ZkYgBdWKPOFoSoQIdEuTuR82pmtxH2spWG9h6etwfr1pLBqQ==", - "dev": true, - "license": "MIT", - "dependencies": { - "parent-module": "^1.0.0", - "resolve-from": "^4.0.0" - }, - "engines": { - "node": ">=6" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/imurmurhash": { - "version": "0.1.4", - "resolved": "https://registry.npmjs.org/imurmurhash/-/imurmurhash-0.1.4.tgz", - "integrity": "sha512-JmXMZ6wuvDmLiHEml9ykzqO6lwFbof0GG4IkcGaENdCRDDmMVnny7s5HsIgHCbaq0w2MyPhDqkhTUgS2LU2PHA==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=0.8.19" - } - }, - "node_modules/inline-style-parser": { - "version": "0.2.4", - "resolved": "https://registry.npmjs.org/inline-style-parser/-/inline-style-parser-0.2.4.tgz", - "integrity": "sha512-0aO8FkhNZlj/ZIbNi7Lxxr12obT7cL1moPfE4tg1LkX7LlLfC6DeX4l2ZEud1ukP9jNQyNnfzQVqwbwmAATY4Q==", - "license": "MIT" - }, - "node_modules/is-alphabetical": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/is-alphabetical/-/is-alphabetical-2.0.1.tgz", - "integrity": "sha512-FWyyY60MeTNyeSRpkM2Iry0G9hpr7/9kD40mD/cGQEuilcZYS4okz8SN2Q6rLCJ8gbCt6fN+rC+6tMGS99LaxQ==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/is-alphanumerical": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/is-alphanumerical/-/is-alphanumerical-2.0.1.tgz", - "integrity": "sha512-hmbYhX/9MUMF5uh7tOXyK/n0ZvWpad5caBA17GsC6vyuCqaWliRG5K1qS9inmUhEMaOBIW7/whAnSwveW/LtZw==", - "license": "MIT", - "dependencies": { - "is-alphabetical": "^2.0.0", - "is-decimal": "^2.0.0" - }, - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/is-decimal": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/is-decimal/-/is-decimal-2.0.1.tgz", - "integrity": "sha512-AAB9hiomQs5DXWcRB1rqsxGUstbRroFOPPVAomNk/3XHR5JyEZChOyTWe2oayKnsSsr/kcGqF+z6yuH6HHpN0A==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/is-extglob": { - "version": "2.1.1", - "resolved": "https://registry.npmjs.org/is-extglob/-/is-extglob-2.1.1.tgz", - "integrity": "sha512-SbKbANkN603Vi4jEZv49LeVJMn4yGwsbzZworEoyEiutsN3nJYdbO36zfhGJ6QEDpOZIFkDtnq5JRxmvl3jsoQ==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=0.10.0" - } - }, - "node_modules/is-glob": { - "version": "4.0.3", - "resolved": "https://registry.npmjs.org/is-glob/-/is-glob-4.0.3.tgz", - "integrity": "sha512-xelSayHH36ZgE7ZWhli7pW34hNbNl8Ojv5KVmkJD4hBdD3th8Tfk9vYasLM+mXWOZhFkgZfxhLSnrwRr4elSSg==", - "dev": true, - "license": "MIT", - "dependencies": { - "is-extglob": "^2.1.1" - }, - "engines": { - "node": ">=0.10.0" - } - }, - "node_modules/is-hexadecimal": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/is-hexadecimal/-/is-hexadecimal-2.0.1.tgz", - "integrity": "sha512-DgZQp241c8oO6cA1SbTEWiXeoxV42vlcJxgH+B3hi1AiqqKruZR3ZGF8In3fj4+/y/7rHvlOZLZtgJ/4ttYGZg==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/is-number": { - "version": "7.0.0", - "resolved": "https://registry.npmjs.org/is-number/-/is-number-7.0.0.tgz", - "integrity": "sha512-41Cifkg6e8TylSpdtTpeLVMqvSBEVzTttHvERD741+pnZ8ANv0004MRL43QKPDlK9cGvNp6NZWZUBlbGXYxxng==", - "license": "MIT", - "engines": { - "node": ">=0.12.0" - } - }, - "node_modules/is-plain-obj": { - "version": "4.1.0", - "resolved": "https://registry.npmjs.org/is-plain-obj/-/is-plain-obj-4.1.0.tgz", - "integrity": "sha512-+Pgi+vMuUNkJyExiMBt5IlFoMyKnr5zhJ4Uspz58WOhBF5QoIZkFyNHIbBAtHwzVAgk5RtndVNsDRN61/mmDqg==", - "license": "MIT", - "engines": { - "node": ">=12" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/isexe": { - "version": "2.0.0", - "resolved": "https://registry.npmjs.org/isexe/-/isexe-2.0.0.tgz", - "integrity": "sha512-RHxMLp9lnKHGHRng9QFhRCMbYAcVpn69smSGcq3f36xjgVVWThj4qqLbTLlq7Ssj8B+fIQ1EuCEGI2lKsyQeIw==", - "dev": true, - "license": "ISC" - }, - "node_modules/jiti": { - "version": "2.4.2", - "resolved": "https://registry.npmjs.org/jiti/-/jiti-2.4.2.tgz", - "integrity": "sha512-rg9zJN+G4n2nfJl5MW3BMygZX56zKPNVEYYqq7adpmMh4Jn2QNEwhvQlFy6jPVdcod7txZtKHWnyZiA3a0zP7A==", - "license": "MIT", - "bin": { - "jiti": "lib/jiti-cli.mjs" - } - }, - "node_modules/js-tokens": { - "version": "4.0.0", - "resolved": "https://registry.npmjs.org/js-tokens/-/js-tokens-4.0.0.tgz", - "integrity": "sha512-RdJUflcE3cUzKiMqQgsCu06FPu9UdIJO0beYbPhHN4k6apgJtifcoCtT9bcxOpYBtpD2kCM6Sbzg4CausW/PKQ==", - "license": "MIT" - }, - "node_modules/js-yaml": { - "version": "4.1.0", - "resolved": "https://registry.npmjs.org/js-yaml/-/js-yaml-4.1.0.tgz", - "integrity": "sha512-wpxZs9NoxZaJESJGIZTyDEaYpl0FKSA+FB9aJiyemKhMwkxQg63h4T1KJgUGHpTqPDNRcmmYLugrRjJlBtWvRA==", - "dev": true, - "license": "MIT", - "dependencies": { - "argparse": "^2.0.1" - }, - "bin": { - "js-yaml": "bin/js-yaml.js" - } - }, - "node_modules/jsesc": { - "version": "3.1.0", - "resolved": "https://registry.npmjs.org/jsesc/-/jsesc-3.1.0.tgz", - "integrity": "sha512-/sM3dO2FOzXjKQhJuo0Q173wf2KOo8t4I8vHy6lF9poUp7bKT0/NHE8fPX23PwfhnykfqnC2xRxOnVw5XuGIaA==", - "dev": true, - "license": "MIT", - "bin": { - "jsesc": "bin/jsesc" - }, - "engines": { - "node": ">=6" - } - }, - "node_modules/json-buffer": { - "version": "3.0.1", - "resolved": "https://registry.npmjs.org/json-buffer/-/json-buffer-3.0.1.tgz", - "integrity": "sha512-4bV5BfR2mqfQTJm+V5tPPdf+ZpuhiIvTuAB5g8kcrXOZpTT/QwwVRWBywX1ozr6lEuPdbHxwaJlm9G6mI2sfSQ==", - "dev": true, - "license": "MIT" - }, - "node_modules/json-schema-traverse": { - "version": "0.4.1", - "resolved": "https://registry.npmjs.org/json-schema-traverse/-/json-schema-traverse-0.4.1.tgz", - "integrity": "sha512-xbbCH5dCYU5T8LcEhhuh7HJ88HXuW3qsI3Y0zOZFKfZEHcpWiHU/Jxzk629Brsab/mMiHQti9wMP+845RPe3Vg==", - "dev": true, - "license": "MIT" - }, - "node_modules/json-stable-stringify-without-jsonify": { - "version": "1.0.1", - "resolved": "https://registry.npmjs.org/json-stable-stringify-without-jsonify/-/json-stable-stringify-without-jsonify-1.0.1.tgz", - "integrity": "sha512-Bdboy+l7tA3OGW6FjyFHWkP5LuByj1Tk33Ljyq0axyzdk9//JSi2u3fP1QSmd1KNwq6VOKYGlAu87CisVir6Pw==", - "dev": true, - "license": "MIT" - }, - "node_modules/json5": { - "version": "2.2.3", - "resolved": "https://registry.npmjs.org/json5/-/json5-2.2.3.tgz", - "integrity": "sha512-XmOWe7eyHYH14cLdVPoyg+GOH3rYX++KpzrylJwSW98t3Nk+U8XOl8FWKOgwtzdb8lXGf6zYwDUzeHMWfxasyg==", - "dev": true, - "license": "MIT", - "bin": { - "json5": "lib/cli.js" - }, - "engines": { - "node": ">=6" - } - }, - "node_modules/katex": { - "version": "0.16.21", - "resolved": "https://registry.npmjs.org/katex/-/katex-0.16.21.tgz", - "integrity": "sha512-XvqR7FgOHtWupfMiigNzmh+MgUVmDGU2kXZm899ZkPfcuoPuFxyHmXsgATDpFZDAXCI8tvinaVcDo8PIIJSo4A==", - "funding": [ - "https://opencollective.com/katex", - "https://github.com/sponsors/katex" - ], - "license": "MIT", - "dependencies": { - "commander": "^8.3.0" - }, - "bin": { - "katex": "cli.js" - } - }, - "node_modules/keyv": { - "version": "4.5.4", - "resolved": "https://registry.npmjs.org/keyv/-/keyv-4.5.4.tgz", - "integrity": "sha512-oxVHkHR/EJf2CNXnWxRLW6mg7JyCCUcG0DtEGmL2ctUo1PNTin1PUil+r/+4r5MpVgC/fn1kjsx7mjSujKqIpw==", - "dev": true, - "license": "MIT", - "dependencies": { - "json-buffer": "3.0.1" - } - }, - "node_modules/levn": { - "version": "0.4.1", - "resolved": "https://registry.npmjs.org/levn/-/levn-0.4.1.tgz", - "integrity": "sha512-+bT2uH4E5LGE7h/n3evcS/sQlJXCpIp6ym8OWJ5eV6+67Dsql/LaaT7qJBAt2rzfoa/5QBGBhxDix1dMt2kQKQ==", - "dev": true, - "license": "MIT", - "dependencies": { - "prelude-ls": "^1.2.1", - "type-check": "~0.4.0" - }, - "engines": { - "node": ">= 0.8.0" - } - }, - "node_modules/lightningcss": { - "version": "1.29.2", - "resolved": "https://registry.npmjs.org/lightningcss/-/lightningcss-1.29.2.tgz", - "integrity": "sha512-6b6gd/RUXKaw5keVdSEtqFVdzWnU5jMxTUjA2bVcMNPLwSQ08Sv/UodBVtETLCn7k4S1Ibxwh7k68IwLZPgKaA==", - "license": "MPL-2.0", - "dependencies": { - "detect-libc": "^2.0.3" - }, - "engines": { - "node": ">= 12.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/parcel" - }, - "optionalDependencies": { - "lightningcss-darwin-arm64": "1.29.2", - "lightningcss-darwin-x64": "1.29.2", - "lightningcss-freebsd-x64": "1.29.2", - "lightningcss-linux-arm-gnueabihf": "1.29.2", - "lightningcss-linux-arm64-gnu": "1.29.2", - "lightningcss-linux-arm64-musl": "1.29.2", - "lightningcss-linux-x64-gnu": "1.29.2", - "lightningcss-linux-x64-musl": "1.29.2", - "lightningcss-win32-arm64-msvc": "1.29.2", - "lightningcss-win32-x64-msvc": "1.29.2" - } - }, - "node_modules/lightningcss-darwin-arm64": { - "version": "1.29.2", - "resolved": "https://registry.npmjs.org/lightningcss-darwin-arm64/-/lightningcss-darwin-arm64-1.29.2.tgz", - "integrity": "sha512-cK/eMabSViKn/PG8U/a7aCorpeKLMlK0bQeNHmdb7qUnBkNPnL+oV5DjJUo0kqWsJUapZsM4jCfYItbqBDvlcA==", - "cpu": [ - "arm64" - ], - "license": "MPL-2.0", - "optional": true, - "os": [ - "darwin" - ], - "engines": { - "node": ">= 12.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/parcel" - } - }, - "node_modules/lightningcss-darwin-x64": { - "version": "1.29.2", - "resolved": "https://registry.npmjs.org/lightningcss-darwin-x64/-/lightningcss-darwin-x64-1.29.2.tgz", - "integrity": "sha512-j5qYxamyQw4kDXX5hnnCKMf3mLlHvG44f24Qyi2965/Ycz829MYqjrVg2H8BidybHBp9kom4D7DR5VqCKDXS0w==", - "cpu": [ - "x64" - ], - "license": "MPL-2.0", - "optional": true, - "os": [ - "darwin" - ], - "engines": { - "node": ">= 12.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/parcel" - } - }, - "node_modules/lightningcss-freebsd-x64": { - "version": "1.29.2", - "resolved": "https://registry.npmjs.org/lightningcss-freebsd-x64/-/lightningcss-freebsd-x64-1.29.2.tgz", - "integrity": "sha512-wDk7M2tM78Ii8ek9YjnY8MjV5f5JN2qNVO+/0BAGZRvXKtQrBC4/cn4ssQIpKIPP44YXw6gFdpUF+Ps+RGsCwg==", - "cpu": [ - "x64" - ], - "license": "MPL-2.0", - "optional": true, - "os": [ - "freebsd" - ], - "engines": { - "node": ">= 12.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/parcel" - } - }, - "node_modules/lightningcss-linux-arm-gnueabihf": { - "version": "1.29.2", - "resolved": "https://registry.npmjs.org/lightningcss-linux-arm-gnueabihf/-/lightningcss-linux-arm-gnueabihf-1.29.2.tgz", - "integrity": "sha512-IRUrOrAF2Z+KExdExe3Rz7NSTuuJ2HvCGlMKoquK5pjvo2JY4Rybr+NrKnq0U0hZnx5AnGsuFHjGnNT14w26sg==", - "cpu": [ - "arm" - ], - "license": "MPL-2.0", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 12.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/parcel" - } - }, - "node_modules/lightningcss-linux-arm64-gnu": { - "version": "1.29.2", - "resolved": "https://registry.npmjs.org/lightningcss-linux-arm64-gnu/-/lightningcss-linux-arm64-gnu-1.29.2.tgz", - "integrity": "sha512-KKCpOlmhdjvUTX/mBuaKemp0oeDIBBLFiU5Fnqxh1/DZ4JPZi4evEH7TKoSBFOSOV3J7iEmmBaw/8dpiUvRKlQ==", - "cpu": [ - "arm64" - ], - "license": "MPL-2.0", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 12.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/parcel" - } - }, - "node_modules/lightningcss-linux-arm64-musl": { - "version": "1.29.2", - "resolved": "https://registry.npmjs.org/lightningcss-linux-arm64-musl/-/lightningcss-linux-arm64-musl-1.29.2.tgz", - "integrity": "sha512-Q64eM1bPlOOUgxFmoPUefqzY1yV3ctFPE6d/Vt7WzLW4rKTv7MyYNky+FWxRpLkNASTnKQUaiMJ87zNODIrrKQ==", - "cpu": [ - "arm64" - ], - "license": "MPL-2.0", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 12.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/parcel" - } - }, - "node_modules/lightningcss-linux-x64-gnu": { - "version": "1.29.2", - "resolved": "https://registry.npmjs.org/lightningcss-linux-x64-gnu/-/lightningcss-linux-x64-gnu-1.29.2.tgz", - "integrity": "sha512-0v6idDCPG6epLXtBH/RPkHvYx74CVziHo6TMYga8O2EiQApnUPZsbR9nFNrg2cgBzk1AYqEd95TlrsL7nYABQg==", - "cpu": [ - "x64" - ], - "license": "MPL-2.0", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 12.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/parcel" - } - }, - "node_modules/lightningcss-linux-x64-musl": { - "version": "1.29.2", - "resolved": "https://registry.npmjs.org/lightningcss-linux-x64-musl/-/lightningcss-linux-x64-musl-1.29.2.tgz", - "integrity": "sha512-rMpz2yawkgGT8RULc5S4WiZopVMOFWjiItBT7aSfDX4NQav6M44rhn5hjtkKzB+wMTRlLLqxkeYEtQ3dd9696w==", - "cpu": [ - "x64" - ], - "license": "MPL-2.0", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">= 12.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/parcel" - } - }, - "node_modules/lightningcss-win32-arm64-msvc": { - "version": "1.29.2", - "resolved": "https://registry.npmjs.org/lightningcss-win32-arm64-msvc/-/lightningcss-win32-arm64-msvc-1.29.2.tgz", - "integrity": "sha512-nL7zRW6evGQqYVu/bKGK+zShyz8OVzsCotFgc7judbt6wnB2KbiKKJwBE4SGoDBQ1O94RjW4asrCjQL4i8Fhbw==", - "cpu": [ - "arm64" - ], - "license": "MPL-2.0", - "optional": true, - "os": [ - "win32" - ], - "engines": { - "node": ">= 12.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/parcel" - } - }, - "node_modules/lightningcss-win32-x64-msvc": { - "version": "1.29.2", - "resolved": "https://registry.npmjs.org/lightningcss-win32-x64-msvc/-/lightningcss-win32-x64-msvc-1.29.2.tgz", - "integrity": "sha512-EdIUW3B2vLuHmv7urfzMI/h2fmlnOQBk1xlsDxkN1tCWKjNFjfLhGxYk8C8mzpSfr+A6jFFIi8fU6LbQGsRWjA==", - "cpu": [ - "x64" - ], - "license": "MPL-2.0", - "optional": true, - "os": [ - "win32" - ], - "engines": { - "node": ">= 12.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/parcel" - } - }, - "node_modules/locate-path": { - "version": "6.0.0", - "resolved": "https://registry.npmjs.org/locate-path/-/locate-path-6.0.0.tgz", - "integrity": "sha512-iPZK6eYjbxRu3uB4/WZ3EsEIMJFMqAoopl3R+zuq0UjcAm/MO6KCweDgPfP3elTztoKP3KtnVHxTn2NHBSDVUw==", - "dev": true, - "license": "MIT", - "dependencies": { - "p-locate": "^5.0.0" - }, - "engines": { - "node": ">=10" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/lodash.merge": { - "version": "4.6.2", - "resolved": "https://registry.npmjs.org/lodash.merge/-/lodash.merge-4.6.2.tgz", - "integrity": "sha512-0KpjqXRVvrYyCsX1swR/XTK0va6VQkQM6MNo7PqW77ByjAhoARA8EfrP1N4+KlKj8YS0ZUCtRT/YUuhyYDujIQ==", - "dev": true, - "license": "MIT" - }, - "node_modules/longest-streak": { - "version": "3.1.0", - "resolved": "https://registry.npmjs.org/longest-streak/-/longest-streak-3.1.0.tgz", - "integrity": "sha512-9Ri+o0JYgehTaVBBDoMqIl8GXtbWg711O3srftcHhZ0dqnETqLaoIK0x17fUw9rFSlK/0NlsKe0Ahhyl5pXE2g==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/loose-envify": { - "version": "1.4.0", - "resolved": "https://registry.npmjs.org/loose-envify/-/loose-envify-1.4.0.tgz", - "integrity": "sha512-lyuxPGr/Wfhrlem2CL/UcnUc1zcqKAImBDzukY7Y5F/yQiNdko6+fRLevlw1HgMySw7f611UIY408EtxRSoK3Q==", - "license": "MIT", - "dependencies": { - "js-tokens": "^3.0.0 || ^4.0.0" - }, - "bin": { - "loose-envify": "cli.js" - } - }, - "node_modules/lowlight": { - "version": "3.3.0", - "resolved": "https://registry.npmjs.org/lowlight/-/lowlight-3.3.0.tgz", - "integrity": "sha512-0JNhgFoPvP6U6lE/UdVsSq99tn6DhjjpAj5MxG49ewd2mOBVtwWYIT8ClyABhq198aXXODMU6Ox8DrGy/CpTZQ==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0", - "devlop": "^1.0.0", - "highlight.js": "~11.11.0" - }, - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/lru-cache": { - "version": "5.1.1", - "resolved": "https://registry.npmjs.org/lru-cache/-/lru-cache-5.1.1.tgz", - "integrity": "sha512-KpNARQA3Iwv+jTA0utUVVbrh+Jlrr1Fv0e56GGzAFOXN7dk/FviaDW8LHmK52DlcH4WP2n6gI8vN1aesBFgo9w==", - "dev": true, - "license": "ISC", - "dependencies": { - "yallist": "^3.0.2" - } - }, - "node_modules/markdown-table": { - "version": "3.0.4", - "resolved": "https://registry.npmjs.org/markdown-table/-/markdown-table-3.0.4.tgz", - "integrity": "sha512-wiYz4+JrLyb/DqW2hkFJxP7Vd7JuTDm77fvbM8VfEQdmSMqcImWeeRbHwZjBjIFki/VaMK2BhFi7oUUZeM5bqw==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/mdast-util-find-and-replace": { - "version": "3.0.2", - "resolved": "https://registry.npmjs.org/mdast-util-find-and-replace/-/mdast-util-find-and-replace-3.0.2.tgz", - "integrity": "sha512-Tmd1Vg/m3Xz43afeNxDIhWRtFZgM2VLyaf4vSTYwudTyeuTneoL3qtWMA5jeLyz/O1vDJmmV4QuScFCA2tBPwg==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "escape-string-regexp": "^5.0.0", - "unist-util-is": "^6.0.0", - "unist-util-visit-parents": "^6.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-find-and-replace/node_modules/escape-string-regexp": { - "version": "5.0.0", - "resolved": "https://registry.npmjs.org/escape-string-regexp/-/escape-string-regexp-5.0.0.tgz", - "integrity": "sha512-/veY75JbMK4j1yjvuUxuVsiS/hr/4iHs9FTT6cgTexxdE0Ly/glccBAkloH/DofkjRbZU3bnoj38mOmhkZ0lHw==", - "license": "MIT", - "engines": { - "node": ">=12" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/mdast-util-from-markdown": { - "version": "2.0.2", - "resolved": "https://registry.npmjs.org/mdast-util-from-markdown/-/mdast-util-from-markdown-2.0.2.tgz", - "integrity": "sha512-uZhTV/8NBuw0WHkPTrCqDOl0zVe1BIng5ZtHoDk49ME1qqcjYmmLmOf0gELgcRMxN4w2iuIeVso5/6QymSrgmA==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "@types/unist": "^3.0.0", - "decode-named-character-reference": "^1.0.0", - "devlop": "^1.0.0", - "mdast-util-to-string": "^4.0.0", - "micromark": "^4.0.0", - "micromark-util-decode-numeric-character-reference": "^2.0.0", - "micromark-util-decode-string": "^2.0.0", - "micromark-util-normalize-identifier": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0", - "unist-util-stringify-position": "^4.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-gfm": { - "version": "3.0.0", - "resolved": "https://registry.npmjs.org/mdast-util-gfm/-/mdast-util-gfm-3.0.0.tgz", - "integrity": "sha512-dgQEX5Amaq+DuUqf26jJqSK9qgixgd6rYDHAv4aTBuA92cTknZlKpPfa86Z/s8Dj8xsAQpFfBmPUHWJBWqS4Bw==", - "license": "MIT", - "dependencies": { - "mdast-util-from-markdown": "^2.0.0", - "mdast-util-gfm-autolink-literal": "^2.0.0", - "mdast-util-gfm-footnote": "^2.0.0", - "mdast-util-gfm-strikethrough": "^2.0.0", - "mdast-util-gfm-table": "^2.0.0", - "mdast-util-gfm-task-list-item": "^2.0.0", - "mdast-util-to-markdown": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-gfm-autolink-literal": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/mdast-util-gfm-autolink-literal/-/mdast-util-gfm-autolink-literal-2.0.1.tgz", - "integrity": "sha512-5HVP2MKaP6L+G6YaxPNjuL0BPrq9orG3TsrZ9YXbA3vDw/ACI4MEsnoDpn6ZNm7GnZgtAcONJyPhOP8tNJQavQ==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "ccount": "^2.0.0", - "devlop": "^1.0.0", - "mdast-util-find-and-replace": "^3.0.0", - "micromark-util-character": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-gfm-footnote": { - "version": "2.0.0", - "resolved": "https://registry.npmjs.org/mdast-util-gfm-footnote/-/mdast-util-gfm-footnote-2.0.0.tgz", - "integrity": "sha512-5jOT2boTSVkMnQ7LTrd6n/18kqwjmuYqo7JUPe+tRCY6O7dAuTFMtTPauYYrMPpox9hlN0uOx/FL8XvEfG9/mQ==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "devlop": "^1.1.0", - "mdast-util-from-markdown": "^2.0.0", - "mdast-util-to-markdown": "^2.0.0", - "micromark-util-normalize-identifier": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-gfm-strikethrough": { - "version": "2.0.0", - "resolved": "https://registry.npmjs.org/mdast-util-gfm-strikethrough/-/mdast-util-gfm-strikethrough-2.0.0.tgz", - "integrity": "sha512-mKKb915TF+OC5ptj5bJ7WFRPdYtuHv0yTRxK2tJvi+BDqbkiG7h7u/9SI89nRAYcmap2xHQL9D+QG/6wSrTtXg==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "mdast-util-from-markdown": "^2.0.0", - "mdast-util-to-markdown": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-gfm-table": { - "version": "2.0.0", - "resolved": "https://registry.npmjs.org/mdast-util-gfm-table/-/mdast-util-gfm-table-2.0.0.tgz", - "integrity": "sha512-78UEvebzz/rJIxLvE7ZtDd/vIQ0RHv+3Mh5DR96p7cS7HsBhYIICDBCu8csTNWNO6tBWfqXPWekRuj2FNOGOZg==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "devlop": "^1.0.0", - "markdown-table": "^3.0.0", - "mdast-util-from-markdown": "^2.0.0", - "mdast-util-to-markdown": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-gfm-task-list-item": { - "version": "2.0.0", - "resolved": "https://registry.npmjs.org/mdast-util-gfm-task-list-item/-/mdast-util-gfm-task-list-item-2.0.0.tgz", - "integrity": "sha512-IrtvNvjxC1o06taBAVJznEnkiHxLFTzgonUdy8hzFVeDun0uTjxxrRGVaNFqkU1wJR3RBPEfsxmU6jDWPofrTQ==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "devlop": "^1.0.0", - "mdast-util-from-markdown": "^2.0.0", - "mdast-util-to-markdown": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-math": { - "version": "3.0.0", - "resolved": "https://registry.npmjs.org/mdast-util-math/-/mdast-util-math-3.0.0.tgz", - "integrity": "sha512-Tl9GBNeG/AhJnQM221bJR2HPvLOSnLE/T9cJI9tlc6zwQk2nPk/4f0cHkOdEixQPC/j8UtKDdITswvLAy1OZ1w==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0", - "@types/mdast": "^4.0.0", - "devlop": "^1.0.0", - "longest-streak": "^3.0.0", - "mdast-util-from-markdown": "^2.0.0", - "mdast-util-to-markdown": "^2.1.0", - "unist-util-remove-position": "^5.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-mdx-expression": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/mdast-util-mdx-expression/-/mdast-util-mdx-expression-2.0.1.tgz", - "integrity": "sha512-J6f+9hUp+ldTZqKRSg7Vw5V6MqjATc+3E4gf3CFNcuZNWD8XdyI6zQ8GqH7f8169MM6P7hMBRDVGnn7oHB9kXQ==", - "license": "MIT", - "dependencies": { - "@types/estree-jsx": "^1.0.0", - "@types/hast": "^3.0.0", - "@types/mdast": "^4.0.0", - "devlop": "^1.0.0", - "mdast-util-from-markdown": "^2.0.0", - "mdast-util-to-markdown": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-mdx-jsx": { - "version": "3.2.0", - "resolved": "https://registry.npmjs.org/mdast-util-mdx-jsx/-/mdast-util-mdx-jsx-3.2.0.tgz", - "integrity": "sha512-lj/z8v0r6ZtsN/cGNNtemmmfoLAFZnjMbNyLzBafjzikOM+glrjNHPlf6lQDOTccj9n5b0PPihEBbhneMyGs1Q==", - "license": "MIT", - "dependencies": { - "@types/estree-jsx": "^1.0.0", - "@types/hast": "^3.0.0", - "@types/mdast": "^4.0.0", - "@types/unist": "^3.0.0", - "ccount": "^2.0.0", - "devlop": "^1.1.0", - "mdast-util-from-markdown": "^2.0.0", - "mdast-util-to-markdown": "^2.0.0", - "parse-entities": "^4.0.0", - "stringify-entities": "^4.0.0", - "unist-util-stringify-position": "^4.0.0", - "vfile-message": "^4.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-mdxjs-esm": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/mdast-util-mdxjs-esm/-/mdast-util-mdxjs-esm-2.0.1.tgz", - "integrity": "sha512-EcmOpxsZ96CvlP03NghtH1EsLtr0n9Tm4lPUJUBccV9RwUOneqSycg19n5HGzCf+10LozMRSObtVr3ee1WoHtg==", - "license": "MIT", - "dependencies": { - "@types/estree-jsx": "^1.0.0", - "@types/hast": "^3.0.0", - "@types/mdast": "^4.0.0", - "devlop": "^1.0.0", - "mdast-util-from-markdown": "^2.0.0", - "mdast-util-to-markdown": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-newline-to-break": { - "version": "2.0.0", - "resolved": "https://registry.npmjs.org/mdast-util-newline-to-break/-/mdast-util-newline-to-break-2.0.0.tgz", - "integrity": "sha512-MbgeFca0hLYIEx/2zGsszCSEJJ1JSCdiY5xQxRcLDDGa8EPvlLPupJ4DSajbMPAnC0je8jfb9TiUATnxxrHUog==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "mdast-util-find-and-replace": "^3.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-phrasing": { - "version": "4.1.0", - "resolved": "https://registry.npmjs.org/mdast-util-phrasing/-/mdast-util-phrasing-4.1.0.tgz", - "integrity": "sha512-TqICwyvJJpBwvGAMZjj4J2n0X8QWp21b9l0o7eXyVJ25YNWYbJDVIyD1bZXE6WtV6RmKJVYmQAKWa0zWOABz2w==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "unist-util-is": "^6.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-to-hast": { - "version": "13.2.0", - "resolved": "https://registry.npmjs.org/mdast-util-to-hast/-/mdast-util-to-hast-13.2.0.tgz", - "integrity": "sha512-QGYKEuUsYT9ykKBCMOEDLsU5JRObWQusAolFMeko/tYPufNkRffBAQjIE+99jbA87xv6FgmjLtwjh9wBWajwAA==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0", - "@types/mdast": "^4.0.0", - "@ungap/structured-clone": "^1.0.0", - "devlop": "^1.0.0", - "micromark-util-sanitize-uri": "^2.0.0", - "trim-lines": "^3.0.0", - "unist-util-position": "^5.0.0", - "unist-util-visit": "^5.0.0", - "vfile": "^6.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-to-markdown": { - "version": "2.1.2", - "resolved": "https://registry.npmjs.org/mdast-util-to-markdown/-/mdast-util-to-markdown-2.1.2.tgz", - "integrity": "sha512-xj68wMTvGXVOKonmog6LwyJKrYXZPvlwabaryTjLh9LuvovB/KAH+kvi8Gjj+7rJjsFi23nkUxRQv1KqSroMqA==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "@types/unist": "^3.0.0", - "longest-streak": "^3.0.0", - "mdast-util-phrasing": "^4.0.0", - "mdast-util-to-string": "^4.0.0", - "micromark-util-classify-character": "^2.0.0", - "micromark-util-decode-string": "^2.0.0", - "unist-util-visit": "^5.0.0", - "zwitch": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/mdast-util-to-string": { - "version": "4.0.0", - "resolved": "https://registry.npmjs.org/mdast-util-to-string/-/mdast-util-to-string-4.0.0.tgz", - "integrity": "sha512-0H44vDimn51F0YwvxSJSm0eCDOJTRlmN0R1yBh4HLj9wiV1Dn0QoXGbvFAWj2hSItVTlCmBF1hqKlIyUBVFLPg==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/merge2": { - "version": "1.4.1", - "resolved": "https://registry.npmjs.org/merge2/-/merge2-1.4.1.tgz", - "integrity": "sha512-8q7VEgMJW4J8tcfVPy8g09NcQwZdbwFEqhe/WZkoIzjn/3TGDwtOCYtXGxA3O8tPzpczCCDgv+P2P5y00ZJOOg==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">= 8" - } - }, - "node_modules/micromark": { - "version": "4.0.1", - "resolved": "https://registry.npmjs.org/micromark/-/micromark-4.0.1.tgz", - "integrity": "sha512-eBPdkcoCNvYcxQOAKAlceo5SNdzZWfF+FcSupREAzdAh9rRmE239CEQAiTwIgblwnoM8zzj35sZ5ZwvSEOF6Kw==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "@types/debug": "^4.0.0", - "debug": "^4.0.0", - "decode-named-character-reference": "^1.0.0", - "devlop": "^1.0.0", - "micromark-core-commonmark": "^2.0.0", - "micromark-factory-space": "^2.0.0", - "micromark-util-character": "^2.0.0", - "micromark-util-chunked": "^2.0.0", - "micromark-util-combine-extensions": "^2.0.0", - "micromark-util-decode-numeric-character-reference": "^2.0.0", - "micromark-util-encode": "^2.0.0", - "micromark-util-normalize-identifier": "^2.0.0", - "micromark-util-resolve-all": "^2.0.0", - "micromark-util-sanitize-uri": "^2.0.0", - "micromark-util-subtokenize": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - } - }, - "node_modules/micromark-core-commonmark": { - "version": "2.0.2", - "resolved": "https://registry.npmjs.org/micromark-core-commonmark/-/micromark-core-commonmark-2.0.2.tgz", - "integrity": "sha512-FKjQKbxd1cibWMM1P9N+H8TwlgGgSkWZMmfuVucLCHaYqeSvJ0hFeHsIa65pA2nYbes0f8LDHPMrd9X7Ujxg9w==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "decode-named-character-reference": "^1.0.0", - "devlop": "^1.0.0", - "micromark-factory-destination": "^2.0.0", - "micromark-factory-label": "^2.0.0", - "micromark-factory-space": "^2.0.0", - "micromark-factory-title": "^2.0.0", - "micromark-factory-whitespace": "^2.0.0", - "micromark-util-character": "^2.0.0", - "micromark-util-chunked": "^2.0.0", - "micromark-util-classify-character": "^2.0.0", - "micromark-util-html-tag-name": "^2.0.0", - "micromark-util-normalize-identifier": "^2.0.0", - "micromark-util-resolve-all": "^2.0.0", - "micromark-util-subtokenize": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - } - }, - "node_modules/micromark-extension-gfm": { - "version": "3.0.0", - "resolved": "https://registry.npmjs.org/micromark-extension-gfm/-/micromark-extension-gfm-3.0.0.tgz", - "integrity": "sha512-vsKArQsicm7t0z2GugkCKtZehqUm31oeGBV/KVSorWSy8ZlNAv7ytjFhvaryUiCUJYqs+NoE6AFhpQvBTM6Q4w==", - "license": "MIT", - "dependencies": { - "micromark-extension-gfm-autolink-literal": "^2.0.0", - "micromark-extension-gfm-footnote": "^2.0.0", - "micromark-extension-gfm-strikethrough": "^2.0.0", - "micromark-extension-gfm-table": "^2.0.0", - "micromark-extension-gfm-tagfilter": "^2.0.0", - "micromark-extension-gfm-task-list-item": "^2.0.0", - "micromark-util-combine-extensions": "^2.0.0", - "micromark-util-types": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/micromark-extension-gfm-autolink-literal": { - "version": "2.1.0", - "resolved": "https://registry.npmjs.org/micromark-extension-gfm-autolink-literal/-/micromark-extension-gfm-autolink-literal-2.1.0.tgz", - "integrity": "sha512-oOg7knzhicgQ3t4QCjCWgTmfNhvQbDDnJeVu9v81r7NltNCVmhPy1fJRX27pISafdjL+SVc4d3l48Gb6pbRypw==", - "license": "MIT", - "dependencies": { - "micromark-util-character": "^2.0.0", - "micromark-util-sanitize-uri": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/micromark-extension-gfm-footnote": { - "version": "2.1.0", - "resolved": "https://registry.npmjs.org/micromark-extension-gfm-footnote/-/micromark-extension-gfm-footnote-2.1.0.tgz", - "integrity": "sha512-/yPhxI1ntnDNsiHtzLKYnE3vf9JZ6cAisqVDauhp4CEHxlb4uoOTxOCJ+9s51bIB8U1N1FJ1RXOKTIlD5B/gqw==", - "license": "MIT", - "dependencies": { - "devlop": "^1.0.0", - "micromark-core-commonmark": "^2.0.0", - "micromark-factory-space": "^2.0.0", - "micromark-util-character": "^2.0.0", - "micromark-util-normalize-identifier": "^2.0.0", - "micromark-util-sanitize-uri": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/micromark-extension-gfm-strikethrough": { - "version": "2.1.0", - "resolved": "https://registry.npmjs.org/micromark-extension-gfm-strikethrough/-/micromark-extension-gfm-strikethrough-2.1.0.tgz", - "integrity": "sha512-ADVjpOOkjz1hhkZLlBiYA9cR2Anf8F4HqZUO6e5eDcPQd0Txw5fxLzzxnEkSkfnD0wziSGiv7sYhk/ktvbf1uw==", - "license": "MIT", - "dependencies": { - "devlop": "^1.0.0", - "micromark-util-chunked": "^2.0.0", - "micromark-util-classify-character": "^2.0.0", - "micromark-util-resolve-all": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/micromark-extension-gfm-table": { - "version": "2.1.1", - "resolved": "https://registry.npmjs.org/micromark-extension-gfm-table/-/micromark-extension-gfm-table-2.1.1.tgz", - "integrity": "sha512-t2OU/dXXioARrC6yWfJ4hqB7rct14e8f7m0cbI5hUmDyyIlwv5vEtooptH8INkbLzOatzKuVbQmAYcbWoyz6Dg==", - "license": "MIT", - "dependencies": { - "devlop": "^1.0.0", - "micromark-factory-space": "^2.0.0", - "micromark-util-character": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/micromark-extension-gfm-tagfilter": { - "version": "2.0.0", - "resolved": "https://registry.npmjs.org/micromark-extension-gfm-tagfilter/-/micromark-extension-gfm-tagfilter-2.0.0.tgz", - "integrity": "sha512-xHlTOmuCSotIA8TW1mDIM6X2O1SiX5P9IuDtqGonFhEK0qgRI4yeC6vMxEV2dgyr2TiD+2PQ10o+cOhdVAcwfg==", - "license": "MIT", - "dependencies": { - "micromark-util-types": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/micromark-extension-gfm-task-list-item": { - "version": "2.1.0", - "resolved": "https://registry.npmjs.org/micromark-extension-gfm-task-list-item/-/micromark-extension-gfm-task-list-item-2.1.0.tgz", - "integrity": "sha512-qIBZhqxqI6fjLDYFTBIa4eivDMnP+OZqsNwmQ3xNLE4Cxwc+zfQEfbs6tzAo2Hjq+bh6q5F+Z8/cksrLFYWQQw==", - "license": "MIT", - "dependencies": { - "devlop": "^1.0.0", - "micromark-factory-space": "^2.0.0", - "micromark-util-character": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/micromark-extension-math": { - "version": "3.1.0", - "resolved": "https://registry.npmjs.org/micromark-extension-math/-/micromark-extension-math-3.1.0.tgz", - "integrity": "sha512-lvEqd+fHjATVs+2v/8kg9i5Q0AP2k85H0WUOwpIVvUML8BapsMvh1XAogmQjOCsLpoKRCVQqEkQBB3NhVBcsOg==", - "license": "MIT", - "dependencies": { - "@types/katex": "^0.16.0", - "devlop": "^1.0.0", - "katex": "^0.16.0", - "micromark-factory-space": "^2.0.0", - "micromark-util-character": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/micromark-factory-destination": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-factory-destination/-/micromark-factory-destination-2.0.1.tgz", - "integrity": "sha512-Xe6rDdJlkmbFRExpTOmRj9N3MaWmbAgdpSrBQvCFqhezUn4AHqJHbaEnfbVYYiexVSs//tqOdY/DxhjdCiJnIA==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "micromark-util-character": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - } - }, - "node_modules/micromark-factory-label": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-factory-label/-/micromark-factory-label-2.0.1.tgz", - "integrity": "sha512-VFMekyQExqIW7xIChcXn4ok29YE3rnuyveW3wZQWWqF4Nv9Wk5rgJ99KzPvHjkmPXF93FXIbBp6YdW3t71/7Vg==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "devlop": "^1.0.0", - "micromark-util-character": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - } - }, - "node_modules/micromark-factory-space": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-factory-space/-/micromark-factory-space-2.0.1.tgz", - "integrity": "sha512-zRkxjtBxxLd2Sc0d+fbnEunsTj46SWXgXciZmHq0kDYGnck/ZSGj9/wULTV95uoeYiK5hRXP2mJ98Uo4cq/LQg==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "micromark-util-character": "^2.0.0", - "micromark-util-types": "^2.0.0" - } - }, - "node_modules/micromark-factory-title": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-factory-title/-/micromark-factory-title-2.0.1.tgz", - "integrity": "sha512-5bZ+3CjhAd9eChYTHsjy6TGxpOFSKgKKJPJxr293jTbfry2KDoWkhBb6TcPVB4NmzaPhMs1Frm9AZH7OD4Cjzw==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "micromark-factory-space": "^2.0.0", - "micromark-util-character": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - } - }, - "node_modules/micromark-factory-whitespace": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-factory-whitespace/-/micromark-factory-whitespace-2.0.1.tgz", - "integrity": "sha512-Ob0nuZ3PKt/n0hORHyvoD9uZhr+Za8sFoP+OnMcnWK5lngSzALgQYKMr9RJVOWLqQYuyn6ulqGWSXdwf6F80lQ==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "micromark-factory-space": "^2.0.0", - "micromark-util-character": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - } - }, - "node_modules/micromark-util-character": { - "version": "2.1.1", - "resolved": "https://registry.npmjs.org/micromark-util-character/-/micromark-util-character-2.1.1.tgz", - "integrity": "sha512-wv8tdUTJ3thSFFFJKtpYKOYiGP2+v96Hvk4Tu8KpCAsTMs6yi+nVmGh1syvSCsaxz45J6Jbw+9DD6g97+NV67Q==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - } - }, - "node_modules/micromark-util-chunked": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-util-chunked/-/micromark-util-chunked-2.0.1.tgz", - "integrity": "sha512-QUNFEOPELfmvv+4xiNg2sRYeS/P84pTW0TCgP5zc9FpXetHY0ab7SxKyAQCNCc1eK0459uoLI1y5oO5Vc1dbhA==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "micromark-util-symbol": "^2.0.0" - } - }, - "node_modules/micromark-util-classify-character": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-util-classify-character/-/micromark-util-classify-character-2.0.1.tgz", - "integrity": "sha512-K0kHzM6afW/MbeWYWLjoHQv1sgg2Q9EccHEDzSkxiP/EaagNzCm7T/WMKZ3rjMbvIpvBiZgwR3dKMygtA4mG1Q==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "micromark-util-character": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - } - }, - "node_modules/micromark-util-combine-extensions": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-util-combine-extensions/-/micromark-util-combine-extensions-2.0.1.tgz", - "integrity": "sha512-OnAnH8Ujmy59JcyZw8JSbK9cGpdVY44NKgSM7E9Eh7DiLS2E9RNQf0dONaGDzEG9yjEl5hcqeIsj4hfRkLH/Bg==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "micromark-util-chunked": "^2.0.0", - "micromark-util-types": "^2.0.0" - } - }, - "node_modules/micromark-util-decode-numeric-character-reference": { - "version": "2.0.2", - "resolved": "https://registry.npmjs.org/micromark-util-decode-numeric-character-reference/-/micromark-util-decode-numeric-character-reference-2.0.2.tgz", - "integrity": "sha512-ccUbYk6CwVdkmCQMyr64dXz42EfHGkPQlBj5p7YVGzq8I7CtjXZJrubAYezf7Rp+bjPseiROqe7G6foFd+lEuw==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "micromark-util-symbol": "^2.0.0" - } - }, - "node_modules/micromark-util-decode-string": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-util-decode-string/-/micromark-util-decode-string-2.0.1.tgz", - "integrity": "sha512-nDV/77Fj6eH1ynwscYTOsbK7rR//Uj0bZXBwJZRfaLEJ1iGBR6kIfNmlNqaqJf649EP0F3NWNdeJi03elllNUQ==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "decode-named-character-reference": "^1.0.0", - "micromark-util-character": "^2.0.0", - "micromark-util-decode-numeric-character-reference": "^2.0.0", - "micromark-util-symbol": "^2.0.0" - } - }, - "node_modules/micromark-util-encode": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-util-encode/-/micromark-util-encode-2.0.1.tgz", - "integrity": "sha512-c3cVx2y4KqUnwopcO9b/SCdo2O67LwJJ/UyqGfbigahfegL9myoEFoDYZgkT7f36T0bLrM9hZTAaAyH+PCAXjw==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT" - }, - "node_modules/micromark-util-html-tag-name": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-util-html-tag-name/-/micromark-util-html-tag-name-2.0.1.tgz", - "integrity": "sha512-2cNEiYDhCWKI+Gs9T0Tiysk136SnR13hhO8yW6BGNyhOC4qYFnwF1nKfD3HFAIXA5c45RrIG1ub11GiXeYd1xA==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT" - }, - "node_modules/micromark-util-normalize-identifier": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-util-normalize-identifier/-/micromark-util-normalize-identifier-2.0.1.tgz", - "integrity": "sha512-sxPqmo70LyARJs0w2UclACPUUEqltCkJ6PhKdMIDuJ3gSf/Q+/GIe3WKl0Ijb/GyH9lOpUkRAO2wp0GVkLvS9Q==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "micromark-util-symbol": "^2.0.0" - } - }, - "node_modules/micromark-util-resolve-all": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-util-resolve-all/-/micromark-util-resolve-all-2.0.1.tgz", - "integrity": "sha512-VdQyxFWFT2/FGJgwQnJYbe1jjQoNTS4RjglmSjTUlpUMa95Htx9NHeYW4rGDJzbjvCsl9eLjMQwGeElsqmzcHg==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "micromark-util-types": "^2.0.0" - } - }, - "node_modules/micromark-util-sanitize-uri": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-util-sanitize-uri/-/micromark-util-sanitize-uri-2.0.1.tgz", - "integrity": "sha512-9N9IomZ/YuGGZZmQec1MbgxtlgougxTodVwDzzEouPKo3qFWvymFHWcnDi2vzV1ff6kas9ucW+o3yzJK9YB1AQ==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "micromark-util-character": "^2.0.0", - "micromark-util-encode": "^2.0.0", - "micromark-util-symbol": "^2.0.0" - } - }, - "node_modules/micromark-util-subtokenize": { - "version": "2.0.4", - "resolved": "https://registry.npmjs.org/micromark-util-subtokenize/-/micromark-util-subtokenize-2.0.4.tgz", - "integrity": "sha512-N6hXjrin2GTJDe3MVjf5FuXpm12PGm80BrUAeub9XFXca8JZbP+oIwY4LJSVwFUCL1IPm/WwSVUN7goFHmSGGQ==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT", - "dependencies": { - "devlop": "^1.0.0", - "micromark-util-chunked": "^2.0.0", - "micromark-util-symbol": "^2.0.0", - "micromark-util-types": "^2.0.0" - } - }, - "node_modules/micromark-util-symbol": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-util-symbol/-/micromark-util-symbol-2.0.1.tgz", - "integrity": "sha512-vs5t8Apaud9N28kgCrRUdEed4UJ+wWNvicHLPxCa9ENlYuAY31M0ETy5y1vA33YoNPDFTghEbnh6efaE8h4x0Q==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT" - }, - "node_modules/micromark-util-types": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/micromark-util-types/-/micromark-util-types-2.0.1.tgz", - "integrity": "sha512-534m2WhVTddrcKVepwmVEVnUAmtrx9bfIjNoQHRqfnvdaHQiFytEhJoTgpWJvDEXCO5gLTQh3wYC1PgOJA4NSQ==", - "funding": [ - { - "type": "GitHub Sponsors", - "url": "https://github.com/sponsors/unifiedjs" - }, - { - "type": "OpenCollective", - "url": "https://opencollective.com/unified" - } - ], - "license": "MIT" - }, - "node_modules/micromatch": { - "version": "4.0.8", - "resolved": "https://registry.npmjs.org/micromatch/-/micromatch-4.0.8.tgz", - "integrity": "sha512-PXwfBhYu0hBCPw8Dn0E+WDYb7af3dSLVWKi3HGv84IdF4TyFoC0ysxFd0Goxw7nSv4T/PzEJQxsYsEiFCKo2BA==", - "license": "MIT", - "dependencies": { - "braces": "^3.0.3", - "picomatch": "^2.3.1" - }, - "engines": { - "node": ">=8.6" - } - }, - "node_modules/minimatch": { - "version": "3.1.2", - "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-3.1.2.tgz", - "integrity": "sha512-J7p63hRiAjw1NDEww1W7i37+ByIrOWO5XQQAzZ3VOcL0PNybwpfmV/N05zFAzwQ9USyEcX6t3UO+K5aqBQOIHw==", - "dev": true, - "license": "ISC", - "dependencies": { - "brace-expansion": "^1.1.7" - }, - "engines": { - "node": "*" - } - }, - "node_modules/ms": { - "version": "2.1.3", - "resolved": "https://registry.npmjs.org/ms/-/ms-2.1.3.tgz", - "integrity": "sha512-6FlzubTLZG3J2a/NVCAleEhjzq5oxgHyaCU9yYXvcLsvoVaHJq/s5xXI6/XXP6tz7R9xAOtHnSO/tXtF3WRTlA==", - "license": "MIT" - }, - "node_modules/nanoid": { - "version": "3.3.8", - "resolved": "https://registry.npmjs.org/nanoid/-/nanoid-3.3.8.tgz", - "integrity": "sha512-WNLf5Sd8oZxOm+TzppcYk8gVOgP+l58xNy58D0nbUnOxOWRWvlcCV4kUF7ltmI6PsrLl/BgKEyS4mqsGChFN0w==", - "funding": [ - { - "type": "github", - "url": "https://github.com/sponsors/ai" - } - ], - "license": "MIT", - "bin": { - "nanoid": "bin/nanoid.cjs" - }, - "engines": { - "node": "^10 || ^12 || ^13.7 || ^14 || >=15.0.1" - } - }, - "node_modules/natural-compare": { - "version": "1.4.0", - "resolved": "https://registry.npmjs.org/natural-compare/-/natural-compare-1.4.0.tgz", - "integrity": "sha512-OWND8ei3VtNC9h7V60qff3SVobHr996CTwgxubgyQYEpg290h9J0buyECNNJexkFm5sOajh5G116RYA1c8ZMSw==", - "dev": true, - "license": "MIT" - }, - "node_modules/node-releases": { - "version": "2.0.19", - "resolved": "https://registry.npmjs.org/node-releases/-/node-releases-2.0.19.tgz", - "integrity": "sha512-xxOWJsBKtzAq7DY0J+DTzuz58K8e7sJbdgwkbMWQe8UYB6ekmsQ45q0M/tJDsGaZmbC+l7n57UV8Hl5tHxO9uw==", - "license": "MIT" - }, - "node_modules/normalize-range": { - "version": "0.1.2", - "resolved": "https://registry.npmjs.org/normalize-range/-/normalize-range-0.1.2.tgz", - "integrity": "sha512-bdok/XvKII3nUpklnV6P2hxtMNrCboOjAcyBuQnWEhO665FwrSNRxU+AqpsyvO6LgGYPspN+lu5CLtw4jPRKNA==", - "license": "MIT", - "engines": { - "node": ">=0.10.0" - } - }, - "node_modules/object-assign": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/object-assign/-/object-assign-4.1.1.tgz", - "integrity": "sha512-rJgTQnkUnH1sFw8yT6VSU3zD3sWmu6sZhIseY8VX+GRu3P6F7Fu+JNDoXfklElbLJSnc3FUQHVe4cU5hj+BcUg==", - "license": "MIT", - "engines": { - "node": ">=0.10.0" - } - }, - "node_modules/optionator": { - "version": "0.9.4", - "resolved": "https://registry.npmjs.org/optionator/-/optionator-0.9.4.tgz", - "integrity": "sha512-6IpQ7mKUxRcZNLIObR0hz7lxsapSSIYNZJwXPGeF0mTVqGKFIXj1DQcMoT22S3ROcLyY/rz0PWaWZ9ayWmad9g==", - "dev": true, - "license": "MIT", - "dependencies": { - "deep-is": "^0.1.3", - "fast-levenshtein": "^2.0.6", - "levn": "^0.4.1", - "prelude-ls": "^1.2.1", - "type-check": "^0.4.0", - "word-wrap": "^1.2.5" - }, - "engines": { - "node": ">= 0.8.0" - } - }, - "node_modules/p-limit": { - "version": "3.1.0", - "resolved": "https://registry.npmjs.org/p-limit/-/p-limit-3.1.0.tgz", - "integrity": "sha512-TYOanM3wGwNGsZN2cVTYPArw454xnXj5qmWF1bEoAc4+cU/ol7GVh7odevjp1FNHduHc3KZMcFduxU5Xc6uJRQ==", - "dev": true, - "license": "MIT", - "dependencies": { - "yocto-queue": "^0.1.0" - }, - "engines": { - "node": ">=10" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/p-locate": { - "version": "5.0.0", - "resolved": "https://registry.npmjs.org/p-locate/-/p-locate-5.0.0.tgz", - "integrity": "sha512-LaNjtRWUBY++zB5nE/NwcaoMylSPk+S+ZHNB1TzdbMJMny6dynpAGt7X/tl/QYq3TIeE6nxHppbo2LGymrG5Pw==", - "dev": true, - "license": "MIT", - "dependencies": { - "p-limit": "^3.0.2" - }, - "engines": { - "node": ">=10" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/parent-module": { - "version": "1.0.1", - "resolved": "https://registry.npmjs.org/parent-module/-/parent-module-1.0.1.tgz", - "integrity": "sha512-GQ2EWRpQV8/o+Aw8YqtfZZPfNRWZYkbidE9k5rpl/hC3vtHHBfGm2Ifi6qWV+coDGkrUKZAxE3Lot5kcsRlh+g==", - "dev": true, - "license": "MIT", - "dependencies": { - "callsites": "^3.0.0" - }, - "engines": { - "node": ">=6" - } - }, - "node_modules/parse-entities": { - "version": "4.0.2", - "resolved": "https://registry.npmjs.org/parse-entities/-/parse-entities-4.0.2.tgz", - "integrity": "sha512-GG2AQYWoLgL877gQIKeRPGO1xF9+eG1ujIb5soS5gPvLQ1y2o8FL90w2QWNdf9I361Mpp7726c+lj3U0qK1uGw==", - "license": "MIT", - "dependencies": { - "@types/unist": "^2.0.0", - "character-entities-legacy": "^3.0.0", - "character-reference-invalid": "^2.0.0", - "decode-named-character-reference": "^1.0.0", - "is-alphanumerical": "^2.0.0", - "is-decimal": "^2.0.0", - "is-hexadecimal": "^2.0.0" - }, - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/parse-entities/node_modules/@types/unist": { - "version": "2.0.11", - "resolved": "https://registry.npmjs.org/@types/unist/-/unist-2.0.11.tgz", - "integrity": "sha512-CmBKiL6NNo/OqgmMn95Fk9Whlp2mtvIv+KNpQKN2F4SjvrEesubTRWGYSg+BnWZOnlCaSTU1sMpsBOzgbYhnsA==", - "license": "MIT" - }, - "node_modules/parse5": { - "version": "7.2.1", - "resolved": "https://registry.npmjs.org/parse5/-/parse5-7.2.1.tgz", - "integrity": "sha512-BuBYQYlv1ckiPdQi/ohiivi9Sagc9JG+Ozs0r7b/0iK3sKmrb0b9FdWdBbOdx6hBCM/F9Ir82ofnBhtZOjCRPQ==", - "license": "MIT", - "dependencies": { - "entities": "^4.5.0" - }, - "funding": { - "url": "https://github.com/inikulin/parse5?sponsor=1" - } - }, - "node_modules/path-exists": { - "version": "4.0.0", - "resolved": "https://registry.npmjs.org/path-exists/-/path-exists-4.0.0.tgz", - "integrity": "sha512-ak9Qy5Q7jYb2Wwcey5Fpvg2KoAc/ZIhLSLOSBmRmygPsGwkVVt0fZa0qrtMz+m6tJTAHfZQ8FnmB4MG4LWy7/w==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=8" - } - }, - "node_modules/path-key": { - "version": "3.1.1", - "resolved": "https://registry.npmjs.org/path-key/-/path-key-3.1.1.tgz", - "integrity": "sha512-ojmeN0qd+y0jszEtoY48r0Peq5dwMEkIlCOu6Q5f41lfkswXuKtYrhgoTpLnyIcHm24Uhqx+5Tqm2InSwLhE6Q==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=8" - } - }, - "node_modules/pdfjs-dist": { - "version": "5.2.133", - "resolved": "https://registry.npmjs.org/pdfjs-dist/-/pdfjs-dist-5.2.133.tgz", - "integrity": "sha512-abE6ZWDxztt+gGFzfm4bX2ggfxUk9wsDEoFzIJm9LozaY3JdXR7jyLK4Bjs+XLXplCduuWS1wGhPC4tgTn/kzg==", - "license": "Apache-2.0", - "engines": { - "node": ">=20.16.0 || >=22.3.0" - }, - "optionalDependencies": { - "@napi-rs/canvas": "^0.1.67" - } - }, - "node_modules/picocolors": { - "version": "1.1.1", - "resolved": "https://registry.npmjs.org/picocolors/-/picocolors-1.1.1.tgz", - "integrity": "sha512-xceH2snhtb5M9liqDsmEw56le376mTZkEX/jEb/RxNFyegNul7eNslCXP9FDj/Lcu0X8KEyMceP2ntpaHrDEVA==", - "license": "ISC" - }, - "node_modules/picomatch": { - "version": "2.3.1", - "resolved": "https://registry.npmjs.org/picomatch/-/picomatch-2.3.1.tgz", - "integrity": "sha512-JU3teHTNjmE2VCGFzuY8EXzCDVwEqB2a8fsIvwaStHhAWJEeVd1o1QD80CU6+ZdEXXSLbSsuLwJjkCBWqRQUVA==", - "license": "MIT", - "engines": { - "node": ">=8.6" - }, - "funding": { - "url": "https://github.com/sponsors/jonschlinkert" - } - }, - "node_modules/postcss": { - "version": "8.5.1", - "resolved": "https://registry.npmjs.org/postcss/-/postcss-8.5.1.tgz", - "integrity": "sha512-6oz2beyjc5VMn/KV1pPw8fliQkhBXrVn1Z3TVyqZxU8kZpzEKhBdmCFqI6ZbmGtamQvQGuU1sgPTk8ZrXDD7jQ==", - "funding": [ - { - "type": "opencollective", - "url": "https://opencollective.com/postcss/" - }, - { - "type": "tidelift", - "url": "https://tidelift.com/funding/github/npm/postcss" - }, - { - "type": "github", - "url": "https://github.com/sponsors/ai" - } - ], - "license": "MIT", - "dependencies": { - "nanoid": "^3.3.8", - "picocolors": "^1.1.1", - "source-map-js": "^1.2.1" - }, - "engines": { - "node": "^10 || ^12 || >=14" - } - }, - "node_modules/postcss-value-parser": { - "version": "4.2.0", - "resolved": "https://registry.npmjs.org/postcss-value-parser/-/postcss-value-parser-4.2.0.tgz", - "integrity": "sha512-1NNCs6uurfkVbeXG4S8JFT9t19m45ICnif8zWLd5oPSZ50QnwMfK+H3jv408d4jw/7Bttv5axS5IiHoLaVNHeQ==", - "license": "MIT" - }, - "node_modules/prelude-ls": { - "version": "1.2.1", - "resolved": "https://registry.npmjs.org/prelude-ls/-/prelude-ls-1.2.1.tgz", - "integrity": "sha512-vkcDPrRZo1QZLbn5RLGPpg/WmIQ65qoWWhcGKf/b5eplkkarX0m9z8ppCat4mlOqUsWpyNuYgO3VRyrYHSzX5g==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">= 0.8.0" - } - }, - "node_modules/prettier": { - "version": "3.4.2", - "resolved": "https://registry.npmjs.org/prettier/-/prettier-3.4.2.tgz", - "integrity": "sha512-e9MewbtFo+Fevyuxn/4rrcDAaq0IYxPGLvObpQjiZBMAzB9IGmzlnG9RZy3FFas+eBMu2vA0CszMeduow5dIuQ==", - "dev": true, - "license": "MIT", - "bin": { - "prettier": "bin/prettier.cjs" - }, - "engines": { - "node": ">=14" - }, - "funding": { - "url": "https://github.com/prettier/prettier?sponsor=1" - } - }, - "node_modules/prop-types": { - "version": "15.8.1", - "resolved": "https://registry.npmjs.org/prop-types/-/prop-types-15.8.1.tgz", - "integrity": "sha512-oj87CgZICdulUohogVAR7AjlC0327U4el4L6eAvOqCeudMDVU0NThNaV+b9Df4dXgSP1gXMTnPdhfe/2qDH5cg==", - "license": "MIT", - "dependencies": { - "loose-envify": "^1.4.0", - "object-assign": "^4.1.1", - "react-is": "^16.13.1" - } - }, - "node_modules/property-information": { - "version": "6.5.0", - "resolved": "https://registry.npmjs.org/property-information/-/property-information-6.5.0.tgz", - "integrity": "sha512-PgTgs/BlvHxOu8QuEN7wi5A0OmXaBcHpmCSTehcs6Uuu9IkDIEo13Hy7n898RHfrQ49vKCoGeWZSaAK01nwVig==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/punycode": { - "version": "2.3.1", - "resolved": "https://registry.npmjs.org/punycode/-/punycode-2.3.1.tgz", - "integrity": "sha512-vYt7UD1U9Wg6138shLtLOvdAu+8DsC/ilFtEVHcH+wydcSpNE20AfSOduf6MkRFahL5FY7X1oU7nKVZFtfq8Fg==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=6" - } - }, - "node_modules/queue-microtask": { - "version": "1.2.3", - "resolved": "https://registry.npmjs.org/queue-microtask/-/queue-microtask-1.2.3.tgz", - "integrity": "sha512-NuaNSa6flKT5JaSYQzJok04JzTL1CA6aGhv5rfLW3PgqA+M2ChpZQnAC8h8i4ZFkBS8X5RqkDBHA7r4hej3K9A==", - "dev": true, - "funding": [ - { - "type": "github", - "url": "https://github.com/sponsors/feross" - }, - { - "type": "patreon", - "url": "https://www.patreon.com/feross" - }, - { - "type": "consulting", - "url": "https://feross.org/support" - } - ], - "license": "MIT" - }, - "node_modules/react": { - "version": "18.3.1", - "resolved": "https://registry.npmjs.org/react/-/react-18.3.1.tgz", - "integrity": "sha512-wS+hAgJShR0KhEvPJArfuPVN1+Hz1t0Y6n5jLrGQbkb4urgPE/0Rve+1kMB1v/oWgHgm4WIcV+i7F2pTVj+2iQ==", - "license": "MIT", - "dependencies": { - "loose-envify": "^1.1.0" - }, - "engines": { - "node": ">=0.10.0" - } - }, - "node_modules/react-dom": { - "version": "18.3.1", - "resolved": "https://registry.npmjs.org/react-dom/-/react-dom-18.3.1.tgz", - "integrity": "sha512-5m4nQKp+rZRb09LNH59GM4BxTh9251/ylbKIbpe7TpGxfJ+9kv6BLkLBXIjjspbgbnIBNqlI23tRnTWT0snUIw==", - "license": "MIT", - "dependencies": { - "loose-envify": "^1.1.0", - "scheduler": "^0.23.2" - }, - "peerDependencies": { - "react": "^18.3.1" - } - }, - "node_modules/react-dropzone": { - "version": "14.3.8", - "resolved": "https://registry.npmjs.org/react-dropzone/-/react-dropzone-14.3.8.tgz", - "integrity": "sha512-sBgODnq+lcA4P296DY4wacOZz3JFpD99fp+hb//iBO2HHnyeZU3FwWyXJ6salNpqQdsZrgMrotuko/BdJMV8Ug==", - "license": "MIT", - "dependencies": { - "attr-accept": "^2.2.4", - "file-selector": "^2.1.0", - "prop-types": "^15.8.1" - }, - "engines": { - "node": ">= 10.13" - }, - "peerDependencies": { - "react": ">= 16.8 || 18.0.0" - } - }, - "node_modules/react-hot-toast": { - "version": "2.5.2", - "resolved": "https://registry.npmjs.org/react-hot-toast/-/react-hot-toast-2.5.2.tgz", - "integrity": "sha512-Tun3BbCxzmXXM7C+NI4qiv6lT0uwGh4oAfeJyNOjYUejTsm35mK9iCaYLGv8cBz9L5YxZLx/2ii7zsIwPtPUdw==", - "license": "MIT", - "dependencies": { - "csstype": "^3.1.3", - "goober": "^2.1.16" - }, - "engines": { - "node": ">=10" - }, - "peerDependencies": { - "react": ">=16", - "react-dom": ">=16" - } - }, - "node_modules/react-is": { - "version": "16.13.1", - "resolved": "https://registry.npmjs.org/react-is/-/react-is-16.13.1.tgz", - "integrity": "sha512-24e6ynE2H+OKt4kqsOvNd8kBpV65zoxbA4BVsEOB3ARVWQki/DHzaUoC5KuON/BiccDaCCTZBuOcfZs70kR8bQ==", - "license": "MIT" - }, - "node_modules/react-markdown": { - "version": "9.0.3", - "resolved": "https://registry.npmjs.org/react-markdown/-/react-markdown-9.0.3.tgz", - "integrity": "sha512-Yk7Z94dbgYTOrdk41Z74GoKA7rThnsbbqBTRYuxoe08qvfQ9tJVhmAKw6BJS/ZORG7kTy/s1QvYzSuaoBA1qfw==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0", - "devlop": "^1.0.0", - "hast-util-to-jsx-runtime": "^2.0.0", - "html-url-attributes": "^3.0.0", - "mdast-util-to-hast": "^13.0.0", - "remark-parse": "^11.0.0", - "remark-rehype": "^11.0.0", - "unified": "^11.0.0", - "unist-util-visit": "^5.0.0", - "vfile": "^6.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - }, - "peerDependencies": { - "@types/react": ">=18", - "react": ">=18" - } - }, - "node_modules/react-refresh": { - "version": "0.14.2", - "resolved": "https://registry.npmjs.org/react-refresh/-/react-refresh-0.14.2.tgz", - "integrity": "sha512-jCvmsr+1IUSMUyzOkRcvnVbX3ZYC6g9TDrDbFuFmRDq7PD4yaGbLKNQL6k2jnArV8hjYxh7hVhAZB6s9HDGpZA==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=0.10.0" - } - }, - "node_modules/react-router": { - "version": "7.1.5", - "resolved": "https://registry.npmjs.org/react-router/-/react-router-7.1.5.tgz", - "integrity": "sha512-8BUF+hZEU4/z/JD201yK6S+UYhsf58bzYIDq2NS1iGpwxSXDu7F+DeGSkIXMFBuHZB21FSiCzEcUb18cQNdRkA==", - "license": "MIT", - "dependencies": { - "@types/cookie": "^0.6.0", - "cookie": "^1.0.1", - "set-cookie-parser": "^2.6.0", - "turbo-stream": "2.4.0" - }, - "engines": { - "node": ">=20.0.0" - }, - "peerDependencies": { - "react": ">=18", - "react-dom": ">=18" - }, - "peerDependenciesMeta": { - "react-dom": { - "optional": true - } - } - }, - "node_modules/rehype-highlight": { - "version": "7.0.2", - "resolved": "https://registry.npmjs.org/rehype-highlight/-/rehype-highlight-7.0.2.tgz", - "integrity": "sha512-k158pK7wdC2qL3M5NcZROZ2tR/l7zOzjxXd5VGdcfIyoijjQqpHd3JKtYSBDpDZ38UI2WJWuFAtkMDxmx5kstA==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0", - "hast-util-to-text": "^4.0.0", - "lowlight": "^3.0.0", - "unist-util-visit": "^5.0.0", - "vfile": "^6.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/rehype-katex": { - "version": "7.0.1", - "resolved": "https://registry.npmjs.org/rehype-katex/-/rehype-katex-7.0.1.tgz", - "integrity": "sha512-OiM2wrZ/wuhKkigASodFoo8wimG3H12LWQaH8qSPVJn9apWKFSH3YOCtbKpBorTVw/eI7cuT21XBbvwEswbIOA==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0", - "@types/katex": "^0.16.0", - "hast-util-from-html-isomorphic": "^2.0.0", - "hast-util-to-text": "^4.0.0", - "katex": "^0.16.0", - "unist-util-visit-parents": "^6.0.0", - "vfile": "^6.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/remark-breaks": { - "version": "4.0.0", - "resolved": "https://registry.npmjs.org/remark-breaks/-/remark-breaks-4.0.0.tgz", - "integrity": "sha512-IjEjJOkH4FuJvHZVIW0QCDWxcG96kCq7An/KVH2NfJe6rKZU2AsHeB3OEjPNRxi4QC34Xdx7I2KGYn6IpT7gxQ==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "mdast-util-newline-to-break": "^2.0.0", - "unified": "^11.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/remark-gfm": { - "version": "4.0.0", - "resolved": "https://registry.npmjs.org/remark-gfm/-/remark-gfm-4.0.0.tgz", - "integrity": "sha512-U92vJgBPkbw4Zfu/IiW2oTZLSL3Zpv+uI7My2eq8JxKgqraFdU8YUGicEJCEgSbeaG+QDFqIcwwfMTOEelPxuA==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "mdast-util-gfm": "^3.0.0", - "micromark-extension-gfm": "^3.0.0", - "remark-parse": "^11.0.0", - "remark-stringify": "^11.0.0", - "unified": "^11.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/remark-math": { - "version": "6.0.0", - "resolved": "https://registry.npmjs.org/remark-math/-/remark-math-6.0.0.tgz", - "integrity": "sha512-MMqgnP74Igy+S3WwnhQ7kqGlEerTETXMvJhrUzDikVZ2/uogJCb+WHUg97hK9/jcfc0dkD73s3LN8zU49cTEtA==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "mdast-util-math": "^3.0.0", - "micromark-extension-math": "^3.0.0", - "unified": "^11.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/remark-parse": { - "version": "11.0.0", - "resolved": "https://registry.npmjs.org/remark-parse/-/remark-parse-11.0.0.tgz", - "integrity": "sha512-FCxlKLNGknS5ba/1lmpYijMUzX2esxW5xQqjWxw2eHFfS2MSdaHVINFmhjo+qN1WhZhNimq0dZATN9pH0IDrpA==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "mdast-util-from-markdown": "^2.0.0", - "micromark-util-types": "^2.0.0", - "unified": "^11.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/remark-rehype": { - "version": "11.1.1", - "resolved": "https://registry.npmjs.org/remark-rehype/-/remark-rehype-11.1.1.tgz", - "integrity": "sha512-g/osARvjkBXb6Wo0XvAeXQohVta8i84ACbenPpoSsxTOQH/Ae0/RGP4WZgnMH5pMLpsj4FG7OHmcIcXxpza8eQ==", - "license": "MIT", - "dependencies": { - "@types/hast": "^3.0.0", - "@types/mdast": "^4.0.0", - "mdast-util-to-hast": "^13.0.0", - "unified": "^11.0.0", - "vfile": "^6.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/remark-stringify": { - "version": "11.0.0", - "resolved": "https://registry.npmjs.org/remark-stringify/-/remark-stringify-11.0.0.tgz", - "integrity": "sha512-1OSmLd3awB/t8qdoEOMazZkNsfVTeY4fTsgzcQFdXNq8ToTN4ZGwrMnlda4K6smTFKD+GRV6O48i6Z4iKgPPpw==", - "license": "MIT", - "dependencies": { - "@types/mdast": "^4.0.0", - "mdast-util-to-markdown": "^2.0.0", - "unified": "^11.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/resolve-from": { - "version": "4.0.0", - "resolved": "https://registry.npmjs.org/resolve-from/-/resolve-from-4.0.0.tgz", - "integrity": "sha512-pb/MYmXstAkysRFx8piNI1tGFNQIFA3vkE3Gq4EuA1dF6gHp/+vgZqsCGJapvy8N3Q+4o7FwvquPJcnZ7RYy4g==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=4" - } - }, - "node_modules/reusify": { - "version": "1.0.4", - "resolved": "https://registry.npmjs.org/reusify/-/reusify-1.0.4.tgz", - "integrity": "sha512-U9nH88a3fc/ekCF1l0/UP1IosiuIjyTh7hBvXVMHYgVcfGvt897Xguj2UOLDeI5BG2m7/uwyaLVT6fbtCwTyzw==", - "dev": true, - "license": "MIT", - "engines": { - "iojs": ">=1.0.0", - "node": ">=0.10.0" - } - }, - "node_modules/rollup": { - "version": "4.34.2", - "resolved": "https://registry.npmjs.org/rollup/-/rollup-4.34.2.tgz", - "integrity": "sha512-sBDUoxZEaqLu9QeNalL8v3jw6WjPku4wfZGyTU7l7m1oC+rpRihXc/n/H+4148ZkGz5Xli8CHMns//fFGKvpIQ==", - "license": "MIT", - "dependencies": { - "@types/estree": "1.0.6" - }, - "bin": { - "rollup": "dist/bin/rollup" - }, - "engines": { - "node": ">=18.0.0", - "npm": ">=8.0.0" - }, - "optionalDependencies": { - "@rollup/rollup-android-arm-eabi": "4.34.2", - "@rollup/rollup-android-arm64": "4.34.2", - "@rollup/rollup-darwin-arm64": "4.34.2", - "@rollup/rollup-darwin-x64": "4.34.2", - "@rollup/rollup-freebsd-arm64": "4.34.2", - "@rollup/rollup-freebsd-x64": "4.34.2", - "@rollup/rollup-linux-arm-gnueabihf": "4.34.2", - "@rollup/rollup-linux-arm-musleabihf": "4.34.2", - "@rollup/rollup-linux-arm64-gnu": "4.34.2", - "@rollup/rollup-linux-arm64-musl": "4.34.2", - "@rollup/rollup-linux-loongarch64-gnu": "4.34.2", - "@rollup/rollup-linux-powerpc64le-gnu": "4.34.2", - "@rollup/rollup-linux-riscv64-gnu": "4.34.2", - "@rollup/rollup-linux-s390x-gnu": "4.34.2", - "@rollup/rollup-linux-x64-gnu": "4.34.2", - "@rollup/rollup-linux-x64-musl": "4.34.2", - "@rollup/rollup-win32-arm64-msvc": "4.34.2", - "@rollup/rollup-win32-ia32-msvc": "4.34.2", - "@rollup/rollup-win32-x64-msvc": "4.34.2", - "fsevents": "~2.3.2" - } - }, - "node_modules/run-parallel": { - "version": "1.2.0", - "resolved": "https://registry.npmjs.org/run-parallel/-/run-parallel-1.2.0.tgz", - "integrity": "sha512-5l4VyZR86LZ/lDxZTR6jqL8AFE2S0IFLMP26AbjsLVADxHdhB/c0GUsH+y39UfCi3dzz8OlQuPmnaJOMoDHQBA==", - "dev": true, - "funding": [ - { - "type": "github", - "url": "https://github.com/sponsors/feross" - }, - { - "type": "patreon", - "url": "https://www.patreon.com/feross" - }, - { - "type": "consulting", - "url": "https://feross.org/support" - } - ], - "license": "MIT", - "dependencies": { - "queue-microtask": "^1.2.2" - } - }, - "node_modules/rxjs": { - "version": "7.8.1", - "resolved": "https://registry.npmjs.org/rxjs/-/rxjs-7.8.1.tgz", - "integrity": "sha512-AA3TVj+0A2iuIoQkWEK/tqFjBq2j+6PO6Y0zJcvzLAFhEFIO3HL0vls9hWLncZbAAbK0mar7oZ4V079I/qPMxg==", - "devOptional": true, - "license": "Apache-2.0", - "dependencies": { - "tslib": "^2.1.0" - } - }, - "node_modules/sass-embedded": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded/-/sass-embedded-1.83.4.tgz", - "integrity": "sha512-Hf2burRA/y5PGxsg6jB9UpoK/xZ6g/pgrkOcdl6j+rRg1Zj8XhGKZ1MTysZGtTPUUmiiErqzkP5+Kzp95yv9GQ==", - "devOptional": true, - "license": "MIT", - "dependencies": { - "@bufbuild/protobuf": "^2.0.0", - "buffer-builder": "^0.2.0", - "colorjs.io": "^0.5.0", - "immutable": "^5.0.2", - "rxjs": "^7.4.0", - "supports-color": "^8.1.1", - "sync-child-process": "^1.0.2", - "varint": "^6.0.0" - }, - "bin": { - "sass": "dist/bin/sass.js" - }, - "engines": { - "node": ">=16.0.0" - }, - "optionalDependencies": { - "sass-embedded-android-arm": "1.83.4", - "sass-embedded-android-arm64": "1.83.4", - "sass-embedded-android-ia32": "1.83.4", - "sass-embedded-android-riscv64": "1.83.4", - "sass-embedded-android-x64": "1.83.4", - "sass-embedded-darwin-arm64": "1.83.4", - "sass-embedded-darwin-x64": "1.83.4", - "sass-embedded-linux-arm": "1.83.4", - "sass-embedded-linux-arm64": "1.83.4", - "sass-embedded-linux-ia32": "1.83.4", - "sass-embedded-linux-musl-arm": "1.83.4", - "sass-embedded-linux-musl-arm64": "1.83.4", - "sass-embedded-linux-musl-ia32": "1.83.4", - "sass-embedded-linux-musl-riscv64": "1.83.4", - "sass-embedded-linux-musl-x64": "1.83.4", - "sass-embedded-linux-riscv64": "1.83.4", - "sass-embedded-linux-x64": "1.83.4", - "sass-embedded-win32-arm64": "1.83.4", - "sass-embedded-win32-ia32": "1.83.4", - "sass-embedded-win32-x64": "1.83.4" - } - }, - "node_modules/sass-embedded-android-arm": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-android-arm/-/sass-embedded-android-arm-1.83.4.tgz", - "integrity": "sha512-9Z4pJAOgEkXa3VDY/o+U6l5XvV0mZTJcSl0l/mSPHihjAHSpLYnOW6+KOWeM8dxqrsqTYcd6COzhanI/a++5Gw==", - "cpu": [ - "arm" - ], - "license": "MIT", - "optional": true, - "os": [ - "android" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-android-arm64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-android-arm64/-/sass-embedded-android-arm64-1.83.4.tgz", - "integrity": "sha512-tgX4FzmbVqnQmD67ZxQDvI+qFNABrboOQgwsG05E5bA/US42zGajW9AxpECJYiMXVOHmg+d81ICbjb0fsVHskw==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "android" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-android-ia32": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-android-ia32/-/sass-embedded-android-ia32-1.83.4.tgz", - "integrity": "sha512-RsFOziFqPcfZXdFRULC4Ayzy9aK6R6FwQ411broCjlOBX+b0gurjRadkue3cfUEUR5mmy0KeCbp7zVKPLTK+5Q==", - "cpu": [ - "ia32" - ], - "license": "MIT", - "optional": true, - "os": [ - "android" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-android-riscv64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-android-riscv64/-/sass-embedded-android-riscv64-1.83.4.tgz", - "integrity": "sha512-EHwh0nmQarBBrMRU928eTZkFGx19k/XW2YwbPR4gBVdWLkbTgCA5aGe8hTE6/1zStyx++3nDGvTZ78+b/VvvLg==", - "cpu": [ - "riscv64" - ], - "license": "MIT", - "optional": true, - "os": [ - "android" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-android-x64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-android-x64/-/sass-embedded-android-x64-1.83.4.tgz", - "integrity": "sha512-0PgQNuPWYy1jEOEPDVsV89KfqOsMLIp9CSbjBY7jRcwRhyVAcigqrUG6bDeNtojHUYKA1kU+Eh/85WxOHUOgBw==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "android" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-darwin-arm64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-darwin-arm64/-/sass-embedded-darwin-arm64-1.83.4.tgz", - "integrity": "sha512-rp2ywymWc3nymnSnAFG5R/8hvxWCsuhK3wOnD10IDlmNB7o4rzKby1c+2ZfpQGowlYGWsWWTgz8FW2qzmZsQRw==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "darwin" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-darwin-x64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-darwin-x64/-/sass-embedded-darwin-x64-1.83.4.tgz", - "integrity": "sha512-kLkN2lXz9PCgGfDS8Ev5YVcl/V2173L6379en/CaFuJJi7WiyPgBymW7hOmfCt4uO4R1y7CP2Uc08DRtZsBlAA==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "darwin" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-linux-arm": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-linux-arm/-/sass-embedded-linux-arm-1.83.4.tgz", - "integrity": "sha512-nL90ryxX2lNmFucr9jYUyHHx21AoAgdCL1O5Ltx2rKg2xTdytAGHYo2MT5S0LIeKLa/yKP/hjuSvrbICYNDvtA==", - "cpu": [ - "arm" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-linux-arm64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-linux-arm64/-/sass-embedded-linux-arm64-1.83.4.tgz", - "integrity": "sha512-E0zjsZX2HgESwyqw31EHtI39DKa7RgK7nvIhIRco1d0QEw227WnoR9pjH3M/ZQy4gQj3GKilOFHM5Krs/omeIA==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-linux-ia32": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-linux-ia32/-/sass-embedded-linux-ia32-1.83.4.tgz", - "integrity": "sha512-ew5HpchSzgAYbQoriRh8QhlWn5Kw2nQ2jHoV9YLwGKe3fwwOWA0KDedssvDv7FWnY/FCqXyymhLd6Bxae4Xquw==", - "cpu": [ - "ia32" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-linux-musl-arm": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-linux-musl-arm/-/sass-embedded-linux-musl-arm-1.83.4.tgz", - "integrity": "sha512-0RrJRwMrmm+gG0VOB5b5Cjs7Sd+lhqpQJa6EJNEaZHljJokEfpE5GejZsGMRMIQLxEvVphZnnxl6sonCGFE/QQ==", - "cpu": [ - "arm" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-linux-musl-arm64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-linux-musl-arm64/-/sass-embedded-linux-musl-arm64-1.83.4.tgz", - "integrity": "sha512-IzMgalf6MZOxgp4AVCgsaWAFDP/IVWOrgVXxkyhw29fyAEoSWBJH4k87wyPhEtxSuzVHLxKNbc8k3UzdWmlBFg==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-linux-musl-ia32": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-linux-musl-ia32/-/sass-embedded-linux-musl-ia32-1.83.4.tgz", - "integrity": "sha512-LLb4lYbcxPzX4UaJymYXC+WwokxUlfTJEFUv5VF0OTuSsHAGNRs/rslPtzVBTvMeG9TtlOQDhku1F7G6iaDotA==", - "cpu": [ - "ia32" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-linux-musl-riscv64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-linux-musl-riscv64/-/sass-embedded-linux-musl-riscv64-1.83.4.tgz", - "integrity": "sha512-zoKlPzD5Z13HKin1UGR74QkEy+kZEk2AkGX5RelRG494mi+IWwRuWCppXIovor9+BQb9eDWPYPoMVahwN5F7VA==", - "cpu": [ - "riscv64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-linux-musl-x64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-linux-musl-x64/-/sass-embedded-linux-musl-x64-1.83.4.tgz", - "integrity": "sha512-hB8+/PYhfEf2zTIcidO5Bpof9trK6WJjZ4T8g2MrxQh8REVtdPcgIkoxczRynqybf9+fbqbUwzXtiUao2GV+vQ==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-linux-riscv64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-linux-riscv64/-/sass-embedded-linux-riscv64-1.83.4.tgz", - "integrity": "sha512-83fL4n+oeDJ0Y4KjASmZ9jHS1Vl9ESVQYHMhJE0i4xDi/P3BNarm2rsKljq/QtrwGpbqwn8ujzOu7DsNCMDSHA==", - "cpu": [ - "riscv64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-linux-x64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-linux-x64/-/sass-embedded-linux-x64-1.83.4.tgz", - "integrity": "sha512-NlnGdvCmTD5PK+LKXlK3sAuxOgbRIEoZfnHvxd157imCm/s2SYF/R28D0DAAjEViyI8DovIWghgbcqwuertXsA==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "linux" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-win32-arm64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-win32-arm64/-/sass-embedded-win32-arm64-1.83.4.tgz", - "integrity": "sha512-J2BFKrEaeSrVazU2qTjyQdAk+MvbzJeTuCET0uAJEXSKtvQ3AzxvzndS7LqkDPbF32eXAHLw8GVpwcBwKbB3Uw==", - "cpu": [ - "arm64" - ], - "license": "MIT", - "optional": true, - "os": [ - "win32" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-win32-ia32": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-win32-ia32/-/sass-embedded-win32-ia32-1.83.4.tgz", - "integrity": "sha512-uPAe9T/5sANFhJS5dcfAOhOJy8/l2TRYG4r+UO3Wp4yhqbN7bggPvY9c7zMYS0OC8tU/bCvfYUDFHYMCl91FgA==", - "cpu": [ - "ia32" - ], - "license": "MIT", - "optional": true, - "os": [ - "win32" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded-win32-x64": { - "version": "1.83.4", - "resolved": "https://registry.npmjs.org/sass-embedded-win32-x64/-/sass-embedded-win32-x64-1.83.4.tgz", - "integrity": "sha512-C9fkDY0jKITdJFij4UbfPFswxoXN9O/Dr79v17fJnstVwtUojzVJWKHUXvF0Zg2LIR7TCc4ju3adejKFxj7ueA==", - "cpu": [ - "x64" - ], - "license": "MIT", - "optional": true, - "os": [ - "win32" - ], - "engines": { - "node": ">=14.0.0" - } - }, - "node_modules/sass-embedded/node_modules/supports-color": { - "version": "8.1.1", - "resolved": "https://registry.npmjs.org/supports-color/-/supports-color-8.1.1.tgz", - "integrity": "sha512-MpUEN2OodtUzxvKQl72cUF7RQ5EiHsGvSsVG0ia9c5RbWGL2CI4C7EpPS8UTBIplnlzZiNuV56w+FuNxy3ty2Q==", - "devOptional": true, - "license": "MIT", - "dependencies": { - "has-flag": "^4.0.0" - }, - "engines": { - "node": ">=10" - }, - "funding": { - "url": "https://github.com/chalk/supports-color?sponsor=1" - } - }, - "node_modules/scheduler": { - "version": "0.23.2", - "resolved": "https://registry.npmjs.org/scheduler/-/scheduler-0.23.2.tgz", - "integrity": "sha512-UOShsPwz7NrMUqhR6t0hWjFduvOzbtv7toDH1/hIrfRNIDBnnBWd0CwJTGvTpngVlmwGCdP9/Zl/tVrDqcuYzQ==", - "license": "MIT", - "dependencies": { - "loose-envify": "^1.1.0" - } - }, - "node_modules/semver": { - "version": "6.3.1", - "resolved": "https://registry.npmjs.org/semver/-/semver-6.3.1.tgz", - "integrity": "sha512-BR7VvDCVHO+q2xBEWskxS6DJE1qRnb7DxzUrogb71CWoSficBxYsiAGd+Kl0mmq/MprG9yArRkyrQxTO6XjMzA==", - "dev": true, - "license": "ISC", - "bin": { - "semver": "bin/semver.js" - } - }, - "node_modules/set-cookie-parser": { - "version": "2.7.1", - "resolved": "https://registry.npmjs.org/set-cookie-parser/-/set-cookie-parser-2.7.1.tgz", - "integrity": "sha512-IOc8uWeOZgnb3ptbCURJWNjWUPcO3ZnTTdzsurqERrP6nPyv+paC55vJM0LpOlT2ne+Ix+9+CRG1MNLlyZ4GjQ==", - "license": "MIT" - }, - "node_modules/shebang-command": { - "version": "2.0.0", - "resolved": "https://registry.npmjs.org/shebang-command/-/shebang-command-2.0.0.tgz", - "integrity": "sha512-kHxr2zZpYtdmrN1qDjrrX/Z1rR1kG8Dx+gkpK1G4eXmvXswmcE1hTWBWYUzlraYw1/yZp6YuDY77YtvbN0dmDA==", - "dev": true, - "license": "MIT", - "dependencies": { - "shebang-regex": "^3.0.0" - }, - "engines": { - "node": ">=8" - } - }, - "node_modules/shebang-regex": { - "version": "3.0.0", - "resolved": "https://registry.npmjs.org/shebang-regex/-/shebang-regex-3.0.0.tgz", - "integrity": "sha512-7++dFhtcx3353uBaq8DDR4NuxBetBzC7ZQOhmTQInHEd6bSrXdiEyzCvG07Z44UYdLShWUyXt5M/yhz8ekcb1A==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=8" - } - }, - "node_modules/source-map": { - "version": "0.6.1", - "resolved": "https://registry.npmjs.org/source-map/-/source-map-0.6.1.tgz", - "integrity": "sha512-UjgapumWlbMhkBgzT7Ykc5YXUT46F0iKu8SGXq0bcwP5dz/h0Plj6enJqjz1Zbq2l5WaqYnrVbwWOWMyF3F47g==", - "license": "BSD-3-Clause", - "optional": true, - "peer": true, - "engines": { - "node": ">=0.10.0" - } - }, - "node_modules/source-map-js": { - "version": "1.2.1", - "resolved": "https://registry.npmjs.org/source-map-js/-/source-map-js-1.2.1.tgz", - "integrity": "sha512-UXWMKhLOwVKb728IUtQPXxfYU+usdybtUrK/8uGE8CQMvrhOpwvzDBwj0QhSL7MQc7vIsISBG8VQ8+IDQxpfQA==", - "license": "BSD-3-Clause", - "engines": { - "node": ">=0.10.0" - } - }, - "node_modules/source-map-support": { - "version": "0.5.21", - "resolved": "https://registry.npmjs.org/source-map-support/-/source-map-support-0.5.21.tgz", - "integrity": "sha512-uBHU3L3czsIyYXKX88fdrGovxdSCoTGDRZ6SYXtSRxLZUzHg5P/66Ht6uoUlHu9EZod+inXhKo3qQgwXUT/y1w==", - "license": "MIT", - "optional": true, - "peer": true, - "dependencies": { - "buffer-from": "^1.0.0", - "source-map": "^0.6.0" - } - }, - "node_modules/space-separated-tokens": { - "version": "2.0.2", - "resolved": "https://registry.npmjs.org/space-separated-tokens/-/space-separated-tokens-2.0.2.tgz", - "integrity": "sha512-PEGlAwrG8yXGXRjW32fGbg66JAlOAwbObuqVoJpv/mRgoWDQfgH1wDPvtzWyUSNAXBGSk8h755YDbbcEy3SH2Q==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/stringify-entities": { - "version": "4.0.4", - "resolved": "https://registry.npmjs.org/stringify-entities/-/stringify-entities-4.0.4.tgz", - "integrity": "sha512-IwfBptatlO+QCJUo19AqvrPNqlVMpW9YEL2LIVY+Rpv2qsjCGxaDLNRgeGsQWJhfItebuJhsGSLjaBbNSQ+ieg==", - "license": "MIT", - "dependencies": { - "character-entities-html4": "^2.0.0", - "character-entities-legacy": "^3.0.0" - }, - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/strip-json-comments": { - "version": "3.1.1", - "resolved": "https://registry.npmjs.org/strip-json-comments/-/strip-json-comments-3.1.1.tgz", - "integrity": "sha512-6fPc+R4ihwqP6N/aIv2f1gMH8lOVtWQHoqC4yK6oSDVVocumAsfCqjkXnqiYMhmMwS/mEHLp7Vehlt3ql6lEig==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=8" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/style-to-object": { - "version": "1.0.8", - "resolved": "https://registry.npmjs.org/style-to-object/-/style-to-object-1.0.8.tgz", - "integrity": "sha512-xT47I/Eo0rwJmaXC4oilDGDWLohVhR6o/xAQcPQN8q6QBuZVL8qMYL85kLmST5cPjAorwvqIA4qXTRQoYHaL6g==", - "license": "MIT", - "dependencies": { - "inline-style-parser": "0.2.4" - } - }, - "node_modules/supports-color": { - "version": "7.2.0", - "resolved": "https://registry.npmjs.org/supports-color/-/supports-color-7.2.0.tgz", - "integrity": "sha512-qpCAvRl9stuOHveKsn7HncJRvv501qIacKzQlO/+Lwxc9+0q2wLyv4Dfvt80/DPn2pqOBsJdDiogXGR9+OvwRw==", - "dev": true, - "license": "MIT", - "dependencies": { - "has-flag": "^4.0.0" - }, - "engines": { - "node": ">=8" - } - }, - "node_modules/sync-child-process": { - "version": "1.0.2", - "resolved": "https://registry.npmjs.org/sync-child-process/-/sync-child-process-1.0.2.tgz", - "integrity": "sha512-8lD+t2KrrScJ/7KXCSyfhT3/hRq78rC0wBFqNJXv3mZyn6hW2ypM05JmlSvtqRbeq6jqA94oHbxAr2vYsJ8vDA==", - "devOptional": true, - "license": "MIT", - "dependencies": { - "sync-message-port": "^1.0.0" - }, - "engines": { - "node": ">=16.0.0" - } - }, - "node_modules/sync-message-port": { - "version": "1.1.3", - "resolved": "https://registry.npmjs.org/sync-message-port/-/sync-message-port-1.1.3.tgz", - "integrity": "sha512-GTt8rSKje5FilG+wEdfCkOcLL7LWqpMlr2c3LRuKt/YXxcJ52aGSbGBAdI4L3aaqfrBt6y711El53ItyH1NWzg==", - "devOptional": true, - "license": "MIT", - "engines": { - "node": ">=16.0.0" - } - }, - "node_modules/tailwindcss": { - "version": "4.1.1", - "resolved": "https://registry.npmjs.org/tailwindcss/-/tailwindcss-4.1.1.tgz", - "integrity": "sha512-QNbdmeS979Efzim2g/bEvfuh+fTcIdp1y7gA+sb6OYSW74rt7Cr7M78AKdf6HqWT3d5AiTb7SwTT3sLQxr4/qw==", - "license": "MIT" - }, - "node_modules/tapable": { - "version": "2.2.1", - "resolved": "https://registry.npmjs.org/tapable/-/tapable-2.2.1.tgz", - "integrity": "sha512-GNzQvQTOIP6RyTfE2Qxb8ZVlNmw0n88vp1szwWRimP02mnTsx3Wtn5qRdqY9w2XduFNUgvOwhNnQsjwCp+kqaQ==", - "license": "MIT", - "engines": { - "node": ">=6" - } - }, - "node_modules/terser": { - "version": "5.39.1", - "resolved": "https://registry.npmjs.org/terser/-/terser-5.39.1.tgz", - "integrity": "sha512-Mm6+uad0ZuDtcV8/4uOZQDQ8RuiC5Pu+iZRedJtF7yA/27sPL7d++In/AJKpWZlU3SYMPPkVfwetn6sgZ66pUA==", - "license": "BSD-2-Clause", - "optional": true, - "peer": true, - "dependencies": { - "@jridgewell/source-map": "^0.3.3", - "acorn": "^8.8.2", - "commander": "^2.20.0", - "source-map-support": "~0.5.20" - }, - "bin": { - "terser": "bin/terser" - }, - "engines": { - "node": ">=10" - } - }, - "node_modules/terser/node_modules/commander": { - "version": "2.20.3", - "resolved": "https://registry.npmjs.org/commander/-/commander-2.20.3.tgz", - "integrity": "sha512-GpVkmM8vF2vQUkj2LvZmD35JxeJOLCwJ9cUkugyk2nuhbv3+mJvpLYYt+0+USMxE+oj+ey/lJEnhZw75x/OMcQ==", - "license": "MIT", - "optional": true, - "peer": true - }, - "node_modules/textlinestream": { - "version": "1.1.1", - "resolved": "https://registry.npmjs.org/textlinestream/-/textlinestream-1.1.1.tgz", - "integrity": "sha512-iBHbi7BQxrFmwZUQJsT0SjNzlLLsXhvW/kg7EyOMVMBIrlnj/qYofwo1LVLZi+3GbUEo96Iu2eqToI2+lZoAEQ==", - "license": "MIT" - }, - "node_modules/to-regex-range": { - "version": "5.0.1", - "resolved": "https://registry.npmjs.org/to-regex-range/-/to-regex-range-5.0.1.tgz", - "integrity": "sha512-65P7iz6X5yEr1cwcgvQxbbIw7Uk3gOy5dIdtZ4rDveLqhrdJP+Li/Hx6tyK0NEb+2GCyneCMJiGqrADCSNk8sQ==", - "license": "MIT", - "dependencies": { - "is-number": "^7.0.0" - }, - "engines": { - "node": ">=8.0" - } - }, - "node_modules/trim-lines": { - "version": "3.0.1", - "resolved": "https://registry.npmjs.org/trim-lines/-/trim-lines-3.0.1.tgz", - "integrity": "sha512-kRj8B+YHZCc9kQYdWfJB2/oUl9rA99qbowYYBtr4ui4mZyAQ2JpvVBd/6U2YloATfqBhBTSMhTpgBHtU0Mf3Rg==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/trough": { - "version": "2.2.0", - "resolved": "https://registry.npmjs.org/trough/-/trough-2.2.0.tgz", - "integrity": "sha512-tmMpK00BjZiUyVyvrBK7knerNgmgvcV/KLVyuma/SC+TQN167GrMRciANTz09+k3zW8L8t60jWO1GpfkZdjTaw==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/ts-api-utils": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/ts-api-utils/-/ts-api-utils-2.0.1.tgz", - "integrity": "sha512-dnlgjFSVetynI8nzgJ+qF62efpglpWRk8isUEWZGWlJYySCTD6aKvbUDu+zbPeDakk3bg5H4XpitHukgfL1m9w==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=18.12" - }, - "peerDependencies": { - "typescript": ">=4.8.4" - } - }, - "node_modules/tslib": { - "version": "2.8.1", - "resolved": "https://registry.npmjs.org/tslib/-/tslib-2.8.1.tgz", - "integrity": "sha512-oJFu94HQb+KVduSUQL7wnpmqnfmLsOA/nAh6b6EH0wCEoK0/mPeXU6c3wKDV83MkOuHPRHtSXKKU99IBazS/2w==", - "license": "0BSD" - }, - "node_modules/turbo-stream": { - "version": "2.4.0", - "resolved": "https://registry.npmjs.org/turbo-stream/-/turbo-stream-2.4.0.tgz", - "integrity": "sha512-FHncC10WpBd2eOmGwpmQsWLDoK4cqsA/UT/GqNoaKOQnT8uzhtCbg3EoUDMvqpOSAI0S26mr0rkjzbOO6S3v1g==", - "license": "ISC" - }, - "node_modules/type-check": { - "version": "0.4.0", - "resolved": "https://registry.npmjs.org/type-check/-/type-check-0.4.0.tgz", - "integrity": "sha512-XleUoc9uwGXqjWwXaUTZAmzMcFZ5858QA2vvx1Ur5xIcixXIP+8LnFDgRplU30us6teqdlskFfu+ae4K79Ooew==", - "dev": true, - "license": "MIT", - "dependencies": { - "prelude-ls": "^1.2.1" - }, - "engines": { - "node": ">= 0.8.0" - } - }, - "node_modules/typescript": { - "version": "5.6.3", - "resolved": "https://registry.npmjs.org/typescript/-/typescript-5.6.3.tgz", - "integrity": "sha512-hjcS1mhfuyi4WW8IWtjP7brDrG2cuDZukyrYrSauoXGNgx0S7zceP07adYkJycEr56BOUTNPzbInooiN3fn1qw==", - "dev": true, - "license": "Apache-2.0", - "bin": { - "tsc": "bin/tsc", - "tsserver": "bin/tsserver" - }, - "engines": { - "node": ">=14.17" - } - }, - "node_modules/typescript-eslint": { - "version": "8.23.0", - "resolved": "https://registry.npmjs.org/typescript-eslint/-/typescript-eslint-8.23.0.tgz", - "integrity": "sha512-/LBRo3HrXr5LxmrdYSOCvoAMm7p2jNizNfbIpCgvG4HMsnoprRUOce/+8VJ9BDYWW68rqIENE/haVLWPeFZBVQ==", - "dev": true, - "license": "MIT", - "dependencies": { - "@typescript-eslint/eslint-plugin": "8.23.0", - "@typescript-eslint/parser": "8.23.0", - "@typescript-eslint/utils": "8.23.0" - }, - "engines": { - "node": "^18.18.0 || ^20.9.0 || >=21.1.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/typescript-eslint" - }, - "peerDependencies": { - "eslint": "^8.57.0 || ^9.0.0", - "typescript": ">=4.8.4 <5.8.0" - } - }, - "node_modules/undici-types": { - "version": "6.20.0", - "resolved": "https://registry.npmjs.org/undici-types/-/undici-types-6.20.0.tgz", - "integrity": "sha512-Ny6QZ2Nju20vw1SRHe3d9jVu6gJ+4e3+MMpqu7pqE5HT6WsTSlce++GQmK5UXS8mzV8DSYHrQH+Xrf2jVcuKNg==", - "devOptional": true, - "license": "MIT" - }, - "node_modules/unified": { - "version": "11.0.5", - "resolved": "https://registry.npmjs.org/unified/-/unified-11.0.5.tgz", - "integrity": "sha512-xKvGhPWw3k84Qjh8bI3ZeJjqnyadK+GEFtazSfZv/rKeTkTjOJho6mFqh2SM96iIcZokxiOpg78GazTSg8+KHA==", - "license": "MIT", - "dependencies": { - "@types/unist": "^3.0.0", - "bail": "^2.0.0", - "devlop": "^1.0.0", - "extend": "^3.0.0", - "is-plain-obj": "^4.0.0", - "trough": "^2.0.0", - "vfile": "^6.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/unist-util-find-after": { - "version": "5.0.0", - "resolved": "https://registry.npmjs.org/unist-util-find-after/-/unist-util-find-after-5.0.0.tgz", - "integrity": "sha512-amQa0Ep2m6hE2g72AugUItjbuM8X8cGQnFoHk0pGfrFeT9GZhzN5SW8nRsiGKK7Aif4CrACPENkA6P/Lw6fHGQ==", - "license": "MIT", - "dependencies": { - "@types/unist": "^3.0.0", - "unist-util-is": "^6.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/unist-util-is": { - "version": "6.0.0", - "resolved": "https://registry.npmjs.org/unist-util-is/-/unist-util-is-6.0.0.tgz", - "integrity": "sha512-2qCTHimwdxLfz+YzdGfkqNlH0tLi9xjTnHddPmJwtIG9MGsdbutfTc4P+haPD7l7Cjxf/WZj+we5qfVPvvxfYw==", - "license": "MIT", - "dependencies": { - "@types/unist": "^3.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/unist-util-position": { - "version": "5.0.0", - "resolved": "https://registry.npmjs.org/unist-util-position/-/unist-util-position-5.0.0.tgz", - "integrity": "sha512-fucsC7HjXvkB5R3kTCO7kUjRdrS0BJt3M/FPxmHMBOm8JQi2BsHAHFsy27E0EolP8rp0NzXsJ+jNPyDWvOJZPA==", - "license": "MIT", - "dependencies": { - "@types/unist": "^3.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/unist-util-remove-position": { - "version": "5.0.0", - "resolved": "https://registry.npmjs.org/unist-util-remove-position/-/unist-util-remove-position-5.0.0.tgz", - "integrity": "sha512-Hp5Kh3wLxv0PHj9m2yZhhLt58KzPtEYKQQ4yxfYFEO7EvHwzyDYnduhHnY1mDxoqr7VUwVuHXk9RXKIiYS1N8Q==", - "license": "MIT", - "dependencies": { - "@types/unist": "^3.0.0", - "unist-util-visit": "^5.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/unist-util-stringify-position": { - "version": "4.0.0", - "resolved": "https://registry.npmjs.org/unist-util-stringify-position/-/unist-util-stringify-position-4.0.0.tgz", - "integrity": "sha512-0ASV06AAoKCDkS2+xw5RXJywruurpbC4JZSm7nr7MOt1ojAzvyyaO+UxZf18j8FCF6kmzCZKcAgN/yu2gm2XgQ==", - "license": "MIT", - "dependencies": { - "@types/unist": "^3.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/unist-util-visit": { - "version": "5.0.0", - "resolved": "https://registry.npmjs.org/unist-util-visit/-/unist-util-visit-5.0.0.tgz", - "integrity": "sha512-MR04uvD+07cwl/yhVuVWAtw+3GOR/knlL55Nd/wAdblk27GCVt3lqpTivy/tkJcZoNPzTwS1Y+KMojlLDhoTzg==", - "license": "MIT", - "dependencies": { - "@types/unist": "^3.0.0", - "unist-util-is": "^6.0.0", - "unist-util-visit-parents": "^6.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/unist-util-visit-parents": { - "version": "6.0.1", - "resolved": "https://registry.npmjs.org/unist-util-visit-parents/-/unist-util-visit-parents-6.0.1.tgz", - "integrity": "sha512-L/PqWzfTP9lzzEa6CKs0k2nARxTdZduw3zyh8d2NVBnsyvHjSX4TWse388YrrQKbvI8w20fGjGlhgT96WwKykw==", - "license": "MIT", - "dependencies": { - "@types/unist": "^3.0.0", - "unist-util-is": "^6.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/update-browserslist-db": { - "version": "1.1.2", - "resolved": "https://registry.npmjs.org/update-browserslist-db/-/update-browserslist-db-1.1.2.tgz", - "integrity": "sha512-PPypAm5qvlD7XMZC3BujecnaOxwhrtoFR+Dqkk5Aa/6DssiH0ibKoketaj9w8LP7Bont1rYeoV5plxD7RTEPRg==", - "funding": [ - { - "type": "opencollective", - "url": "https://opencollective.com/browserslist" - }, - { - "type": "tidelift", - "url": "https://tidelift.com/funding/github/npm/browserslist" - }, - { - "type": "github", - "url": "https://github.com/sponsors/ai" - } - ], - "license": "MIT", - "dependencies": { - "escalade": "^3.2.0", - "picocolors": "^1.1.1" - }, - "bin": { - "update-browserslist-db": "cli.js" - }, - "peerDependencies": { - "browserslist": ">= 4.21.0" - } - }, - "node_modules/uri-js": { - "version": "4.4.1", - "resolved": "https://registry.npmjs.org/uri-js/-/uri-js-4.4.1.tgz", - "integrity": "sha512-7rKUyy33Q1yc98pQ1DAmLtwX109F7TIfWlW1Ydo8Wl1ii1SeHieeh0HHfPeL2fMXK6z0s8ecKs9frCuLJvndBg==", - "dev": true, - "license": "BSD-2-Clause", - "dependencies": { - "punycode": "^2.1.0" - } - }, - "node_modules/varint": { - "version": "6.0.0", - "resolved": "https://registry.npmjs.org/varint/-/varint-6.0.0.tgz", - "integrity": "sha512-cXEIW6cfr15lFv563k4GuVuW/fiwjknytD37jIOLSdSWuOI6WnO/oKwmP2FQTU2l01LP8/M5TSAJpzUaGe3uWg==", - "devOptional": true, - "license": "MIT" - }, - "node_modules/vfile": { - "version": "6.0.3", - "resolved": "https://registry.npmjs.org/vfile/-/vfile-6.0.3.tgz", - "integrity": "sha512-KzIbH/9tXat2u30jf+smMwFCsno4wHVdNmzFyL+T/L3UGqqk6JKfVqOFOZEpZSHADH1k40ab6NUIXZq422ov3Q==", - "license": "MIT", - "dependencies": { - "@types/unist": "^3.0.0", - "vfile-message": "^4.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/vfile-location": { - "version": "5.0.3", - "resolved": "https://registry.npmjs.org/vfile-location/-/vfile-location-5.0.3.tgz", - "integrity": "sha512-5yXvWDEgqeiYiBe1lbxYF7UMAIm/IcopxMHrMQDq3nvKcjPKIhZklUKL+AE7J7uApI4kwe2snsK+eI6UTj9EHg==", - "license": "MIT", - "dependencies": { - "@types/unist": "^3.0.0", - "vfile": "^6.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/vfile-message": { - "version": "4.0.2", - "resolved": "https://registry.npmjs.org/vfile-message/-/vfile-message-4.0.2.tgz", - "integrity": "sha512-jRDZ1IMLttGj41KcZvlrYAaI3CfqpLpfpf+Mfig13viT6NKvRzWZ+lXz0Y5D60w6uJIBAOGq9mSHf0gktF0duw==", - "license": "MIT", - "dependencies": { - "@types/unist": "^3.0.0", - "unist-util-stringify-position": "^4.0.0" - }, - "funding": { - "type": "opencollective", - "url": "https://opencollective.com/unified" - } - }, - "node_modules/vite": { - "version": "6.0.11", - "resolved": "https://registry.npmjs.org/vite/-/vite-6.0.11.tgz", - "integrity": "sha512-4VL9mQPKoHy4+FE0NnRE/kbY51TOfaknxAjt3fJbGJxhIpBZiqVzlZDEesWWsuREXHwNdAoOFZ9MkPEVXczHwg==", - "license": "MIT", - "dependencies": { - "esbuild": "^0.24.2", - "postcss": "^8.4.49", - "rollup": "^4.23.0" - }, - "bin": { - "vite": "bin/vite.js" - }, - "engines": { - "node": "^18.0.0 || ^20.0.0 || >=22.0.0" - }, - "funding": { - "url": "https://github.com/vitejs/vite?sponsor=1" - }, - "optionalDependencies": { - "fsevents": "~2.3.3" - }, - "peerDependencies": { - "@types/node": "^18.0.0 || ^20.0.0 || >=22.0.0", - "jiti": ">=1.21.0", - "less": "*", - "lightningcss": "^1.21.0", - "sass": "*", - "sass-embedded": "*", - "stylus": "*", - "sugarss": "*", - "terser": "^5.16.0", - "tsx": "^4.8.1", - "yaml": "^2.4.2" - }, - "peerDependenciesMeta": { - "@types/node": { - "optional": true - }, - "jiti": { - "optional": true - }, - "less": { - "optional": true - }, - "lightningcss": { - "optional": true - }, - "sass": { - "optional": true - }, - "sass-embedded": { - "optional": true - }, - "stylus": { - "optional": true - }, - "sugarss": { - "optional": true - }, - "terser": { - "optional": true - }, - "tsx": { - "optional": true - }, - "yaml": { - "optional": true - } - } - }, - "node_modules/vite-plugin-singlefile": { - "version": "2.1.0", - "resolved": "https://registry.npmjs.org/vite-plugin-singlefile/-/vite-plugin-singlefile-2.1.0.tgz", - "integrity": "sha512-7tJo+UgZABlKpY/nubth/wxJ4+pUGREPnEwNOknxwl2MM0zTvF14KTU4Ln1lc140gjLLV5mjDrvuoquU7OZqCg==", - "license": "MIT", - "dependencies": { - "micromatch": "^4.0.8" - }, - "engines": { - "node": ">18.0.0" - }, - "peerDependencies": { - "rollup": "^4.28.1", - "vite": "^5.4.11 || ^6.0.0" - } - }, - "node_modules/web-namespaces": { - "version": "2.0.1", - "resolved": "https://registry.npmjs.org/web-namespaces/-/web-namespaces-2.0.1.tgz", - "integrity": "sha512-bKr1DkiNa2krS7qxNtdrtHAmzuYGFQLiQ13TsorsdT6ULTkPLKuu5+GsFpDlg6JFjUTwX2DyhMPG2be8uPrqsQ==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - }, - "node_modules/which": { - "version": "2.0.2", - "resolved": "https://registry.npmjs.org/which/-/which-2.0.2.tgz", - "integrity": "sha512-BLI3Tl1TW3Pvl70l3yq3Y64i+awpwXqsGBYWkkqMtnbXgrMD+yj7rhW0kuEDxzJaYXGjEW5ogapKNMEKNMjibA==", - "dev": true, - "license": "ISC", - "dependencies": { - "isexe": "^2.0.0" - }, - "bin": { - "node-which": "bin/node-which" - }, - "engines": { - "node": ">= 8" - } - }, - "node_modules/word-wrap": { - "version": "1.2.5", - "resolved": "https://registry.npmjs.org/word-wrap/-/word-wrap-1.2.5.tgz", - "integrity": "sha512-BN22B5eaMMI9UMtjrGd5g5eCYPpCPDUy0FJXbYsaT5zYxjFOckS53SQDE3pWkVoWpHXVb3BrYcEN4Twa55B5cA==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=0.10.0" - } - }, - "node_modules/yallist": { - "version": "3.1.1", - "resolved": "https://registry.npmjs.org/yallist/-/yallist-3.1.1.tgz", - "integrity": "sha512-a4UGQaWPH59mOXUYnAG2ewncQS4i4F43Tv3JoAM+s2VDAmS9NsK8GpDMLrCHPksFT7h3K6TOoUNn2pb7RoXx4g==", - "dev": true, - "license": "ISC" - }, - "node_modules/yaml": { - "version": "2.7.0", - "resolved": "https://registry.npmjs.org/yaml/-/yaml-2.7.0.tgz", - "integrity": "sha512-+hSoy/QHluxmC9kCIJyL/uyFmLmc+e5CFR5Wa+bpIhIj85LVb9ZH2nVnqrHoSvKogwODv0ClqZkmiSSaIH5LTA==", - "license": "ISC", - "optional": true, - "peer": true, - "bin": { - "yaml": "bin.mjs" - }, - "engines": { - "node": ">= 14" - } - }, - "node_modules/yocto-queue": { - "version": "0.1.0", - "resolved": "https://registry.npmjs.org/yocto-queue/-/yocto-queue-0.1.0.tgz", - "integrity": "sha512-rVksvsnNCdJ/ohGc6xgPwyN8eheCxsiLM8mxuE/t/mOVqJewPuO1miLpTHQiRgTKCLexL4MeAFVagts7HmNZ2Q==", - "dev": true, - "license": "MIT", - "engines": { - "node": ">=10" - }, - "funding": { - "url": "https://github.com/sponsors/sindresorhus" - } - }, - "node_modules/zwitch": { - "version": "2.0.4", - "resolved": "https://registry.npmjs.org/zwitch/-/zwitch-2.0.4.tgz", - "integrity": "sha512-bXE4cR/kVZhKZX/RjPEflHaKVhUVl85noU3v6b8apfQEc1x4A+zBxjZ4lN8LqGd6WZ3dl98pY4o717VFmoPp+A==", - "license": "MIT", - "funding": { - "type": "github", - "url": "https://github.com/sponsors/wooorm" - } - } - } + "name": "webui", + "version": "1.0.0", + "lockfileVersion": 3, + "requires": true, + "packages": { + "": { + "name": "webui", + "version": "1.0.0", + "dependencies": { + "highlight.js": "^11.11.1", + "mode-watcher": "^1.1.0", + "pdfjs-dist": "^5.4.54", + "rehype-highlight": "^7.0.2", + "rehype-stringify": "^10.0.1", + "remark": "^15.0.1", + "remark-breaks": "^4.0.0", + "remark-gfm": "^4.0.1", + "remark-html": "^16.0.1", + "remark-rehype": "^11.1.2", + "svelte-sonner": "^1.0.5", + "unist-util-visit": "^5.0.0" + }, + "devDependencies": { + "@chromatic-com/storybook": "^4.0.1", + "@eslint/compat": "^1.2.5", + "@eslint/js": "^9.18.0", + "@internationalized/date": "^3.8.2", + "@lucide/svelte": "^0.515.0", + "@playwright/test": "^1.49.1", + "@storybook/addon-a11y": "^9.0.17", + "@storybook/addon-docs": "^9.0.17", + "@storybook/addon-svelte-csf": "^5.0.7", + "@storybook/addon-vitest": "^9.0.17", + "@storybook/sveltekit": "^9.0.17", + "@sveltejs/adapter-static": "^3.0.8", + "@sveltejs/kit": "^2.22.0", + "@sveltejs/vite-plugin-svelte": "^6.0.0", + "@tailwindcss/forms": "^0.5.9", + "@tailwindcss/typography": "^0.5.15", + "@tailwindcss/vite": "^4.0.0", + "@types/node": "^22", + "@vitest/browser": "^3.2.3", + "bits-ui": "^2.8.11", + "clsx": "^2.1.1", + "dexie": "^4.0.11", + "eslint": "^9.18.0", + "eslint-config-prettier": "^10.0.1", + "eslint-plugin-storybook": "^9.0.17", + "eslint-plugin-svelte": "^3.0.0", + "fflate": "^0.8.2", + "globals": "^16.0.0", + "mdsvex": "^0.12.3", + "playwright": "^1.53.0", + "prettier": "^3.4.2", + "prettier-plugin-svelte": "^3.3.3", + "prettier-plugin-tailwindcss": "^0.6.11", + "rehype-katex": "^7.0.1", + "remark-math": "^6.0.0", + "storybook": "^9.0.17", + "svelte": "^5.0.0", + "svelte-check": "^4.0.0", + "tailwind-merge": "^3.3.1", + "tailwind-variants": "^1.0.0", + "tailwindcss": "^4.0.0", + "tw-animate-css": "^1.3.5", + "typescript": "^5.0.0", + "typescript-eslint": "^8.20.0", + "uuid": "^13.0.0", + "vite": "^7.0.4", + "vite-plugin-devtools-json": "^0.2.0", + "vitest": "^3.2.3", + "vitest-browser-svelte": "^0.1.0" + } + }, + "node_modules/@adobe/css-tools": { + "version": "4.4.3", + "resolved": "https://registry.npmjs.org/@adobe/css-tools/-/css-tools-4.4.3.tgz", + "integrity": "sha512-VQKMkwriZbaOgVCby1UDY/LDk5fIjhQicCvVPFqfe+69fWaPWydbWJ3wRt59/YzIwda1I81loas3oCoHxnqvdA==", + "dev": true, + "license": "MIT" + }, + "node_modules/@ampproject/remapping": { + "version": "2.3.0", + "resolved": "https://registry.npmjs.org/@ampproject/remapping/-/remapping-2.3.0.tgz", + "integrity": "sha512-30iZtAPgz+LTIYoeivqYo853f02jBYSd5uGnGpkFV0M3xOt9aN73erkgYAmZU43x4VfqcnLxW9Kpg3R5LC4YYw==", + "license": "Apache-2.0", + "dependencies": { + "@jridgewell/gen-mapping": "^0.3.5", + "@jridgewell/trace-mapping": "^0.3.24" + }, + "engines": { + "node": ">=6.0.0" + } + }, + "node_modules/@babel/code-frame": { + "version": "7.27.1", + "resolved": "https://registry.npmjs.org/@babel/code-frame/-/code-frame-7.27.1.tgz", + "integrity": "sha512-cjQ7ZlQ0Mv3b47hABuTevyTuYN4i+loJKGeV9flcCgIK37cCXRh+L1bd3iBHlynerhQ7BhCkn2BPbQUL+rGqFg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@babel/helper-validator-identifier": "^7.27.1", + "js-tokens": "^4.0.0", + "picocolors": "^1.1.1" + }, + "engines": { + "node": ">=6.9.0" + } + }, + "node_modules/@babel/helper-validator-identifier": { + "version": "7.27.1", + "resolved": "https://registry.npmjs.org/@babel/helper-validator-identifier/-/helper-validator-identifier-7.27.1.tgz", + "integrity": "sha512-D2hP9eA+Sqx1kBZgzxZh0y1trbuU+JoDkiEwqhQ36nodYqJwyEIhPSdMNd7lOm/4io72luTPWH20Yda0xOuUow==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=6.9.0" + } + }, + "node_modules/@babel/runtime": { + "version": "7.27.6", + "resolved": "https://registry.npmjs.org/@babel/runtime/-/runtime-7.27.6.tgz", + "integrity": "sha512-vbavdySgbTTrmFE+EsiqUTzlOr5bzlnJtUv9PynGCAKvfQqjIXbvFdumPM/GxMDfyuGMJaJAU6TO4zc1Jf1i8Q==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=6.9.0" + } + }, + "node_modules/@chromatic-com/storybook": { + "version": "4.0.1", + "resolved": "https://registry.npmjs.org/@chromatic-com/storybook/-/storybook-4.0.1.tgz", + "integrity": "sha512-GQXe5lyZl3yLewLJQyFXEpOp2h+mfN2bPrzYaOFNCJjO4Js9deKbRHTOSaiP2FRwZqDLdQwy2+SEGeXPZ94yYw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@neoconfetti/react": "^1.0.0", + "chromatic": "^12.0.0", + "filesize": "^10.0.12", + "jsonfile": "^6.1.0", + "strip-ansi": "^7.1.0" + }, + "engines": { + "node": ">=20.0.0", + "yarn": ">=1.22.18" + }, + "peerDependencies": { + "storybook": "^0.0.0-0 || ^9.0.0 || ^9.1.0-0" + } + }, + "node_modules/@esbuild/aix-ppc64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/aix-ppc64/-/aix-ppc64-0.25.8.tgz", + "integrity": "sha512-urAvrUedIqEiFR3FYSLTWQgLu5tb+m0qZw0NBEasUeo6wuqatkMDaRT+1uABiGXEu5vqgPd7FGE1BhsAIy9QVA==", + "cpu": [ + "ppc64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "aix" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/android-arm": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/android-arm/-/android-arm-0.25.8.tgz", + "integrity": "sha512-RONsAvGCz5oWyePVnLdZY/HHwA++nxYWIX1atInlaW6SEkwq6XkP3+cb825EUcRs5Vss/lGh/2YxAb5xqc07Uw==", + "cpu": [ + "arm" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/android-arm64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/android-arm64/-/android-arm64-0.25.8.tgz", + "integrity": "sha512-OD3p7LYzWpLhZEyATcTSJ67qB5D+20vbtr6vHlHWSQYhKtzUYrETuWThmzFpZtFsBIxRvhO07+UgVA9m0i/O1w==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/android-x64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/android-x64/-/android-x64-0.25.8.tgz", + "integrity": "sha512-yJAVPklM5+4+9dTeKwHOaA+LQkmrKFX96BM0A/2zQrbS6ENCmxc4OVoBs5dPkCCak2roAD+jKCdnmOqKszPkjA==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/darwin-arm64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/darwin-arm64/-/darwin-arm64-0.25.8.tgz", + "integrity": "sha512-Jw0mxgIaYX6R8ODrdkLLPwBqHTtYHJSmzzd+QeytSugzQ0Vg4c5rDky5VgkoowbZQahCbsv1rT1KW72MPIkevw==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/darwin-x64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/darwin-x64/-/darwin-x64-0.25.8.tgz", + "integrity": "sha512-Vh2gLxxHnuoQ+GjPNvDSDRpoBCUzY4Pu0kBqMBDlK4fuWbKgGtmDIeEC081xi26PPjn+1tct+Bh8FjyLlw1Zlg==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/freebsd-arm64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/freebsd-arm64/-/freebsd-arm64-0.25.8.tgz", + "integrity": "sha512-YPJ7hDQ9DnNe5vxOm6jaie9QsTwcKedPvizTVlqWG9GBSq+BuyWEDazlGaDTC5NGU4QJd666V0yqCBL2oWKPfA==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "freebsd" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/freebsd-x64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/freebsd-x64/-/freebsd-x64-0.25.8.tgz", + "integrity": "sha512-MmaEXxQRdXNFsRN/KcIimLnSJrk2r5H8v+WVafRWz5xdSVmWLoITZQXcgehI2ZE6gioE6HirAEToM/RvFBeuhw==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "freebsd" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/linux-arm": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/linux-arm/-/linux-arm-0.25.8.tgz", + "integrity": "sha512-FuzEP9BixzZohl1kLf76KEVOsxtIBFwCaLupVuk4eFVnOZfU+Wsn+x5Ryam7nILV2pkq2TqQM9EZPsOBuMC+kg==", + "cpu": [ + "arm" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/linux-arm64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/linux-arm64/-/linux-arm64-0.25.8.tgz", + "integrity": "sha512-WIgg00ARWv/uYLU7lsuDK00d/hHSfES5BzdWAdAig1ioV5kaFNrtK8EqGcUBJhYqotlUByUKz5Qo6u8tt7iD/w==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/linux-ia32": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/linux-ia32/-/linux-ia32-0.25.8.tgz", + "integrity": "sha512-A1D9YzRX1i+1AJZuFFUMP1E9fMaYY+GnSQil9Tlw05utlE86EKTUA7RjwHDkEitmLYiFsRd9HwKBPEftNdBfjg==", + "cpu": [ + "ia32" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/linux-loong64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/linux-loong64/-/linux-loong64-0.25.8.tgz", + "integrity": "sha512-O7k1J/dwHkY1RMVvglFHl1HzutGEFFZ3kNiDMSOyUrB7WcoHGf96Sh+64nTRT26l3GMbCW01Ekh/ThKM5iI7hQ==", + "cpu": [ + "loong64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/linux-mips64el": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/linux-mips64el/-/linux-mips64el-0.25.8.tgz", + "integrity": "sha512-uv+dqfRazte3BzfMp8PAQXmdGHQt2oC/y2ovwpTteqrMx2lwaksiFZ/bdkXJC19ttTvNXBuWH53zy/aTj1FgGw==", + "cpu": [ + "mips64el" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/linux-ppc64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/linux-ppc64/-/linux-ppc64-0.25.8.tgz", + "integrity": "sha512-GyG0KcMi1GBavP5JgAkkstMGyMholMDybAf8wF5A70CALlDM2p/f7YFE7H92eDeH/VBtFJA5MT4nRPDGg4JuzQ==", + "cpu": [ + "ppc64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/linux-riscv64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/linux-riscv64/-/linux-riscv64-0.25.8.tgz", + "integrity": "sha512-rAqDYFv3yzMrq7GIcen3XP7TUEG/4LK86LUPMIz6RT8A6pRIDn0sDcvjudVZBiiTcZCY9y2SgYX2lgK3AF+1eg==", + "cpu": [ + "riscv64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/linux-s390x": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/linux-s390x/-/linux-s390x-0.25.8.tgz", + "integrity": "sha512-Xutvh6VjlbcHpsIIbwY8GVRbwoviWT19tFhgdA7DlenLGC/mbc3lBoVb7jxj9Z+eyGqvcnSyIltYUrkKzWqSvg==", + "cpu": [ + "s390x" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/linux-x64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/linux-x64/-/linux-x64-0.25.8.tgz", + "integrity": "sha512-ASFQhgY4ElXh3nDcOMTkQero4b1lgubskNlhIfJrsH5OKZXDpUAKBlNS0Kx81jwOBp+HCeZqmoJuihTv57/jvQ==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/netbsd-arm64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/netbsd-arm64/-/netbsd-arm64-0.25.8.tgz", + "integrity": "sha512-d1KfruIeohqAi6SA+gENMuObDbEjn22olAR7egqnkCD9DGBG0wsEARotkLgXDu6c4ncgWTZJtN5vcgxzWRMzcw==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "netbsd" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/netbsd-x64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/netbsd-x64/-/netbsd-x64-0.25.8.tgz", + "integrity": "sha512-nVDCkrvx2ua+XQNyfrujIG38+YGyuy2Ru9kKVNyh5jAys6n+l44tTtToqHjino2My8VAY6Lw9H7RI73XFi66Cg==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "netbsd" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/openbsd-arm64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/openbsd-arm64/-/openbsd-arm64-0.25.8.tgz", + "integrity": "sha512-j8HgrDuSJFAujkivSMSfPQSAa5Fxbvk4rgNAS5i3K+r8s1X0p1uOO2Hl2xNsGFppOeHOLAVgYwDVlmxhq5h+SQ==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "openbsd" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/openbsd-x64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/openbsd-x64/-/openbsd-x64-0.25.8.tgz", + "integrity": "sha512-1h8MUAwa0VhNCDp6Af0HToI2TJFAn1uqT9Al6DJVzdIBAd21m/G0Yfc77KDM3uF3T/YaOgQq3qTJHPbTOInaIQ==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "openbsd" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/openharmony-arm64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/openharmony-arm64/-/openharmony-arm64-0.25.8.tgz", + "integrity": "sha512-r2nVa5SIK9tSWd0kJd9HCffnDHKchTGikb//9c7HX+r+wHYCpQrSgxhlY6KWV1nFo1l4KFbsMlHk+L6fekLsUg==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "openharmony" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/sunos-x64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/sunos-x64/-/sunos-x64-0.25.8.tgz", + "integrity": "sha512-zUlaP2S12YhQ2UzUfcCuMDHQFJyKABkAjvO5YSndMiIkMimPmxA+BYSBikWgsRpvyxuRnow4nS5NPnf9fpv41w==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "sunos" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/win32-arm64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/win32-arm64/-/win32-arm64-0.25.8.tgz", + "integrity": "sha512-YEGFFWESlPva8hGL+zvj2z/SaK+pH0SwOM0Nc/d+rVnW7GSTFlLBGzZkuSU9kFIGIo8q9X3ucpZhu8PDN5A2sQ==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/win32-ia32": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/win32-ia32/-/win32-ia32-0.25.8.tgz", + "integrity": "sha512-hiGgGC6KZ5LZz58OL/+qVVoZiuZlUYlYHNAmczOm7bs2oE1XriPFi5ZHHrS8ACpV5EjySrnoCKmcbQMN+ojnHg==", + "cpu": [ + "ia32" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@esbuild/win32-x64": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/@esbuild/win32-x64/-/win32-x64-0.25.8.tgz", + "integrity": "sha512-cn3Yr7+OaaZq1c+2pe+8yxC8E144SReCQjN6/2ynubzYjvyqZjTXfQJpAcQpsdJq3My7XADANiYGHoFC69pLQw==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">=18" + } + }, + "node_modules/@eslint-community/eslint-utils": { + "version": "4.7.0", + "resolved": "https://registry.npmjs.org/@eslint-community/eslint-utils/-/eslint-utils-4.7.0.tgz", + "integrity": "sha512-dyybb3AcajC7uha6CvhdVRJqaKyn7w2YKqKyAN37NKYgZT36w+iRb0Dymmc5qEJ549c/S31cMMSFd75bteCpCw==", + "dev": true, + "license": "MIT", + "dependencies": { + "eslint-visitor-keys": "^3.4.3" + }, + "engines": { + "node": "^12.22.0 || ^14.17.0 || >=16.0.0" + }, + "funding": { + "url": "https://opencollective.com/eslint" + }, + "peerDependencies": { + "eslint": "^6.0.0 || ^7.0.0 || >=8.0.0" + } + }, + "node_modules/@eslint-community/eslint-utils/node_modules/eslint-visitor-keys": { + "version": "3.4.3", + "resolved": "https://registry.npmjs.org/eslint-visitor-keys/-/eslint-visitor-keys-3.4.3.tgz", + "integrity": "sha512-wpc+LXeiyiisxPlEkUzU6svyS1frIO3Mgxj1fdy7Pm8Ygzguax2N3Fa/D/ag1WqbOprdI+uY6wMUl8/a2G+iag==", + "dev": true, + "license": "Apache-2.0", + "engines": { + "node": "^12.22.0 || ^14.17.0 || >=16.0.0" + }, + "funding": { + "url": "https://opencollective.com/eslint" + } + }, + "node_modules/@eslint-community/regexpp": { + "version": "4.12.1", + "resolved": "https://registry.npmjs.org/@eslint-community/regexpp/-/regexpp-4.12.1.tgz", + "integrity": "sha512-CCZCDJuduB9OUkFkY2IgppNZMi2lBQgD2qzwXkEia16cge2pijY/aXi96CJMquDMn3nJdlPV1A5KrJEXwfLNzQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": "^12.0.0 || ^14.0.0 || >=16.0.0" + } + }, + "node_modules/@eslint/compat": { + "version": "1.3.1", + "resolved": "https://registry.npmjs.org/@eslint/compat/-/compat-1.3.1.tgz", + "integrity": "sha512-k8MHony59I5EPic6EQTCNOuPoVBnoYXkP+20xvwFjN7t0qI3ImyvyBgg+hIVPwC8JaxVjjUZld+cLfBLFDLucg==", + "dev": true, + "license": "Apache-2.0", + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "peerDependencies": { + "eslint": "^8.40 || 9" + }, + "peerDependenciesMeta": { + "eslint": { + "optional": true + } + } + }, + "node_modules/@eslint/config-array": { + "version": "0.21.0", + "resolved": "https://registry.npmjs.org/@eslint/config-array/-/config-array-0.21.0.tgz", + "integrity": "sha512-ENIdc4iLu0d93HeYirvKmrzshzofPw6VkZRKQGe9Nv46ZnWUzcF1xV01dcvEg/1wXUR61OmmlSfyeyO7EvjLxQ==", + "dev": true, + "license": "Apache-2.0", + "dependencies": { + "@eslint/object-schema": "^2.1.6", + "debug": "^4.3.1", + "minimatch": "^3.1.2" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + } + }, + "node_modules/@eslint/config-helpers": { + "version": "0.3.0", + "resolved": "https://registry.npmjs.org/@eslint/config-helpers/-/config-helpers-0.3.0.tgz", + "integrity": "sha512-ViuymvFmcJi04qdZeDc2whTHryouGcDlaxPqarTD0ZE10ISpxGUVZGZDx4w01upyIynL3iu6IXH2bS1NhclQMw==", + "dev": true, + "license": "Apache-2.0", + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + } + }, + "node_modules/@eslint/core": { + "version": "0.15.2", + "resolved": "https://registry.npmjs.org/@eslint/core/-/core-0.15.2.tgz", + "integrity": "sha512-78Md3/Rrxh83gCxoUc0EiciuOHsIITzLy53m3d9UyiW8y9Dj2D29FeETqyKA+BRK76tnTp6RXWb3pCay8Oyomg==", + "dev": true, + "license": "Apache-2.0", + "dependencies": { + "@types/json-schema": "^7.0.15" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + } + }, + "node_modules/@eslint/eslintrc": { + "version": "3.3.1", + "resolved": "https://registry.npmjs.org/@eslint/eslintrc/-/eslintrc-3.3.1.tgz", + "integrity": "sha512-gtF186CXhIl1p4pJNGZw8Yc6RlshoePRvE0X91oPGb3vZ8pM3qOS9W9NGPat9LziaBV7XrJWGylNQXkGcnM3IQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "ajv": "^6.12.4", + "debug": "^4.3.2", + "espree": "^10.0.1", + "globals": "^14.0.0", + "ignore": "^5.2.0", + "import-fresh": "^3.2.1", + "js-yaml": "^4.1.0", + "minimatch": "^3.1.2", + "strip-json-comments": "^3.1.1" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "url": "https://opencollective.com/eslint" + } + }, + "node_modules/@eslint/eslintrc/node_modules/globals": { + "version": "14.0.0", + "resolved": "https://registry.npmjs.org/globals/-/globals-14.0.0.tgz", + "integrity": "sha512-oahGvuMGQlPw/ivIYBjVSrWAfWLBeku5tpPE2fOPLi+WHffIWbuh2tCjhyQhTBPMf5E9jDEH4FOmTYgYwbKwtQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=18" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/@eslint/js": { + "version": "9.31.0", + "resolved": "https://registry.npmjs.org/@eslint/js/-/js-9.31.0.tgz", + "integrity": "sha512-LOm5OVt7D4qiKCqoiPbA7LWmI+tbw1VbTUowBcUMgQSuM6poJufkFkYDcQpo5KfgD39TnNySV26QjOh7VFpSyw==", + "dev": true, + "license": "MIT", + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "url": "https://eslint.org/donate" + } + }, + "node_modules/@eslint/object-schema": { + "version": "2.1.6", + "resolved": "https://registry.npmjs.org/@eslint/object-schema/-/object-schema-2.1.6.tgz", + "integrity": "sha512-RBMg5FRL0I0gs51M/guSAj5/e14VQ4tpZnQNWwuDT66P14I43ItmPfIZRhO9fUVIPOAQXU47atlywZ/czoqFPA==", + "dev": true, + "license": "Apache-2.0", + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + } + }, + "node_modules/@eslint/plugin-kit": { + "version": "0.3.5", + "resolved": "https://registry.npmjs.org/@eslint/plugin-kit/-/plugin-kit-0.3.5.tgz", + "integrity": "sha512-Z5kJ+wU3oA7MMIqVR9tyZRtjYPr4OC004Q4Rw7pgOKUOKkJfZ3O24nz3WYfGRpMDNmcOi3TwQOmgm7B7Tpii0w==", + "dev": true, + "license": "Apache-2.0", + "dependencies": { + "@eslint/core": "^0.15.2", + "levn": "^0.4.1" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + } + }, + "node_modules/@floating-ui/core": { + "version": "1.7.2", + "resolved": "https://registry.npmjs.org/@floating-ui/core/-/core-1.7.2.tgz", + "integrity": "sha512-wNB5ooIKHQc+Kui96jE/n69rHFWAVoxn5CAzL1Xdd8FG03cgY3MLO+GF9U3W737fYDSgPWA6MReKhBQBop6Pcw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@floating-ui/utils": "^0.2.10" + } + }, + "node_modules/@floating-ui/dom": { + "version": "1.7.2", + "resolved": "https://registry.npmjs.org/@floating-ui/dom/-/dom-1.7.2.tgz", + "integrity": "sha512-7cfaOQuCS27HD7DX+6ib2OrnW+b4ZBwDNnCcT0uTyidcmyWb03FnQqJybDBoCnpdxwBSfA94UAYlRCt7mV+TbA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@floating-ui/core": "^1.7.2", + "@floating-ui/utils": "^0.2.10" + } + }, + "node_modules/@floating-ui/utils": { + "version": "0.2.10", + "resolved": "https://registry.npmjs.org/@floating-ui/utils/-/utils-0.2.10.tgz", + "integrity": "sha512-aGTxbpbg8/b5JfU1HXSrbH3wXZuLPJcNEcZQFMxLs3oSzgtVu6nFPkbbGGUvBcUjKV2YyB9Wxxabo+HEH9tcRQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/@humanfs/core": { + "version": "0.19.1", + "resolved": "https://registry.npmjs.org/@humanfs/core/-/core-0.19.1.tgz", + "integrity": "sha512-5DyQ4+1JEUzejeK1JGICcideyfUbGixgS9jNgex5nqkW+cY7WZhxBigmieN5Qnw9ZosSNVC9KQKyb+GUaGyKUA==", + "dev": true, + "license": "Apache-2.0", + "engines": { + "node": ">=18.18.0" + } + }, + "node_modules/@humanfs/node": { + "version": "0.16.6", + "resolved": "https://registry.npmjs.org/@humanfs/node/-/node-0.16.6.tgz", + "integrity": "sha512-YuI2ZHQL78Q5HbhDiBA1X4LmYdXCKCMQIfw0pw7piHJwyREFebJUvrQN4cMssyES6x+vfUbx1CIpaQUKYdQZOw==", + "dev": true, + "license": "Apache-2.0", + "dependencies": { + "@humanfs/core": "^0.19.1", + "@humanwhocodes/retry": "^0.3.0" + }, + "engines": { + "node": ">=18.18.0" + } + }, + "node_modules/@humanfs/node/node_modules/@humanwhocodes/retry": { + "version": "0.3.1", + "resolved": "https://registry.npmjs.org/@humanwhocodes/retry/-/retry-0.3.1.tgz", + "integrity": "sha512-JBxkERygn7Bv/GbN5Rv8Ul6LVknS+5Bp6RgDC/O8gEBU/yeH5Ui5C/OlWrTb6qct7LjjfT6Re2NxB0ln0yYybA==", + "dev": true, + "license": "Apache-2.0", + "engines": { + "node": ">=18.18" + }, + "funding": { + "type": "github", + "url": "https://github.com/sponsors/nzakas" + } + }, + "node_modules/@humanwhocodes/module-importer": { + "version": "1.0.1", + "resolved": "https://registry.npmjs.org/@humanwhocodes/module-importer/-/module-importer-1.0.1.tgz", + "integrity": "sha512-bxveV4V8v5Yb4ncFTT3rPSgZBOpCkjfK0y4oVVVJwIuDVBRMDXrPyXRL988i5ap9m9bnyEEjWfm5WkBmtffLfA==", + "dev": true, + "license": "Apache-2.0", + "engines": { + "node": ">=12.22" + }, + "funding": { + "type": "github", + "url": "https://github.com/sponsors/nzakas" + } + }, + "node_modules/@humanwhocodes/retry": { + "version": "0.4.3", + "resolved": "https://registry.npmjs.org/@humanwhocodes/retry/-/retry-0.4.3.tgz", + "integrity": "sha512-bV0Tgo9K4hfPCek+aMAn81RppFKv2ySDQeMoSZuvTASywNTnVJCArCZE2FWqpvIatKu7VMRLWlR1EazvVhDyhQ==", + "dev": true, + "license": "Apache-2.0", + "engines": { + "node": ">=18.18" + }, + "funding": { + "type": "github", + "url": "https://github.com/sponsors/nzakas" + } + }, + "node_modules/@internationalized/date": { + "version": "3.8.2", + "resolved": "https://registry.npmjs.org/@internationalized/date/-/date-3.8.2.tgz", + "integrity": "sha512-/wENk7CbvLbkUvX1tu0mwq49CVkkWpkXubGel6birjRPyo6uQ4nQpnq5xZu823zRCwwn82zgHrvgF1vZyvmVgA==", + "dev": true, + "license": "Apache-2.0", + "dependencies": { + "@swc/helpers": "^0.5.0" + } + }, + "node_modules/@isaacs/fs-minipass": { + "version": "4.0.1", + "resolved": "https://registry.npmjs.org/@isaacs/fs-minipass/-/fs-minipass-4.0.1.tgz", + "integrity": "sha512-wgm9Ehl2jpeqP3zw/7mo3kRHFp5MEDhqAdwy1fTGkHAwnkGOVsgpvQhL8B5n1qlb01jV3n/bI0ZfZp5lWA1k4w==", + "dev": true, + "license": "ISC", + "dependencies": { + "minipass": "^7.0.4" + }, + "engines": { + "node": ">=18.0.0" + } + }, + "node_modules/@jridgewell/gen-mapping": { + "version": "0.3.12", + "resolved": "https://registry.npmjs.org/@jridgewell/gen-mapping/-/gen-mapping-0.3.12.tgz", + "integrity": "sha512-OuLGC46TjB5BbN1dH8JULVVZY4WTdkF7tV9Ys6wLL1rubZnCMstOhNHueU5bLCrnRuDhKPDM4g6sw4Bel5Gzqg==", + "license": "MIT", + "dependencies": { + "@jridgewell/sourcemap-codec": "^1.5.0", + "@jridgewell/trace-mapping": "^0.3.24" + } + }, + "node_modules/@jridgewell/resolve-uri": { + "version": "3.1.2", + "resolved": "https://registry.npmjs.org/@jridgewell/resolve-uri/-/resolve-uri-3.1.2.tgz", + "integrity": "sha512-bRISgCIjP20/tbWSPWMEi54QVPRZExkuD9lJL+UIxUKtwVJA8wW1Trb1jMs1RFXo1CBTNZ/5hpC9QvmKWdopKw==", + "license": "MIT", + "engines": { + "node": ">=6.0.0" + } + }, + "node_modules/@jridgewell/sourcemap-codec": { + "version": "1.5.4", + "resolved": "https://registry.npmjs.org/@jridgewell/sourcemap-codec/-/sourcemap-codec-1.5.4.tgz", + "integrity": "sha512-VT2+G1VQs/9oz078bLrYbecdZKs912zQlkelYpuf+SXF+QvZDYJlbx/LSx+meSAwdDFnF8FVXW92AVjjkVmgFw==", + "license": "MIT" + }, + "node_modules/@jridgewell/trace-mapping": { + "version": "0.3.29", + "resolved": "https://registry.npmjs.org/@jridgewell/trace-mapping/-/trace-mapping-0.3.29.tgz", + "integrity": "sha512-uw6guiW/gcAGPDhLmd77/6lW8QLeiV5RUTsAX46Db6oLhGaVj4lhnPwb184s1bkc8kdVg/+h988dro8GRDpmYQ==", + "license": "MIT", + "dependencies": { + "@jridgewell/resolve-uri": "^3.1.0", + "@jridgewell/sourcemap-codec": "^1.4.14" + } + }, + "node_modules/@lucide/svelte": { + "version": "0.515.0", + "resolved": "https://registry.npmjs.org/@lucide/svelte/-/svelte-0.515.0.tgz", + "integrity": "sha512-CEAyqcZmNBfYzVgaRmK2RFJP5tnbXxekRyDk0XX/eZQRfsJmkDvmQwXNX8C869BgNeryzmrRyjHhUL6g9ZOHNA==", + "dev": true, + "license": "ISC", + "peerDependencies": { + "svelte": "^5" + } + }, + "node_modules/@mdx-js/react": { + "version": "3.1.0", + "resolved": "https://registry.npmjs.org/@mdx-js/react/-/react-3.1.0.tgz", + "integrity": "sha512-QjHtSaoameoalGnKDT3FoIl4+9RwyTmo9ZJGBdLOks/YOiWHoRDI3PUwEzOE7kEmGcV3AFcp9K6dYu9rEuKLAQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/mdx": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + }, + "peerDependencies": { + "@types/react": ">=16", + "react": ">=16" + } + }, + "node_modules/@napi-rs/canvas": { + "version": "0.1.76", + "resolved": "https://registry.npmjs.org/@napi-rs/canvas/-/canvas-0.1.76.tgz", + "integrity": "sha512-YIk5okeNN53GzjvWmAyCQFE9xrLeQXzYpudX4TiLvqaz9SqXgIgxIuKPe4DKyB5nccsQMIev7JGKTzZaN5rFdw==", + "license": "MIT", + "optional": true, + "workspaces": [ + "e2e/*" + ], + "engines": { + "node": ">= 10" + }, + "optionalDependencies": { + "@napi-rs/canvas-android-arm64": "0.1.76", + "@napi-rs/canvas-darwin-arm64": "0.1.76", + "@napi-rs/canvas-darwin-x64": "0.1.76", + "@napi-rs/canvas-linux-arm-gnueabihf": "0.1.76", + "@napi-rs/canvas-linux-arm64-gnu": "0.1.76", + "@napi-rs/canvas-linux-arm64-musl": "0.1.76", + "@napi-rs/canvas-linux-riscv64-gnu": "0.1.76", + "@napi-rs/canvas-linux-x64-gnu": "0.1.76", + "@napi-rs/canvas-linux-x64-musl": "0.1.76", + "@napi-rs/canvas-win32-x64-msvc": "0.1.76" + } + }, + "node_modules/@napi-rs/canvas-android-arm64": { + "version": "0.1.76", + "resolved": "https://registry.npmjs.org/@napi-rs/canvas-android-arm64/-/canvas-android-arm64-0.1.76.tgz", + "integrity": "sha512-7EAfkLBQo2QoEzpHdInFbfEUYTXsiO2hvtFo1D9zfTzcQM8n5piZdOpJ3EIkmpe8yLoSV8HLyUQtq4bv11x6Tg==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@napi-rs/canvas-darwin-arm64": { + "version": "0.1.76", + "resolved": "https://registry.npmjs.org/@napi-rs/canvas-darwin-arm64/-/canvas-darwin-arm64-0.1.76.tgz", + "integrity": "sha512-Cs8WRMzaWSJWeWY8tvnCe+TuduHUbB0xFhZ0FmOrNy2prPxT4A6aU3FQu8hR9XJw8kKZ7v902wzaDmy9SdhG8A==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@napi-rs/canvas-darwin-x64": { + "version": "0.1.76", + "resolved": "https://registry.npmjs.org/@napi-rs/canvas-darwin-x64/-/canvas-darwin-x64-0.1.76.tgz", + "integrity": "sha512-ya+T6gV9XAq7YAnMa2fKhWXAuRR5cpRny2IoHacoMxgtOARnUkJO/k3hIb52FtMoq7UxLi5+IFGVHU6ZiMu4Ag==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@napi-rs/canvas-linux-arm-gnueabihf": { + "version": "0.1.76", + "resolved": "https://registry.npmjs.org/@napi-rs/canvas-linux-arm-gnueabihf/-/canvas-linux-arm-gnueabihf-0.1.76.tgz", + "integrity": "sha512-fgnPb+FKVuixACvkHGldJqYXExORBwvqGgL0K80uE6SGH2t0UKD2auHw2CtBy14DUzfg82PkupO2ix2w7kB+Xw==", + "cpu": [ + "arm" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@napi-rs/canvas-linux-arm64-gnu": { + "version": "0.1.76", + "resolved": "https://registry.npmjs.org/@napi-rs/canvas-linux-arm64-gnu/-/canvas-linux-arm64-gnu-0.1.76.tgz", + "integrity": "sha512-r8OxIenvBPOa4I014k1ZWTCz2dB0ZTsxMP7+ovMOKO7jkl1Z+YZo2OTAqxArpMhN0wdEeI3Lw9zUcn2HgwEgDA==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@napi-rs/canvas-linux-arm64-musl": { + "version": "0.1.76", + "resolved": "https://registry.npmjs.org/@napi-rs/canvas-linux-arm64-musl/-/canvas-linux-arm64-musl-0.1.76.tgz", + "integrity": "sha512-smxwzKfHYaOYG7QXUuDPrFEC7WqjL3Lx4AM6mk8/FxDAS+8o0eoZJwSu+zXsaBLimEQUozEYgEGtJ2JJ0RdL4A==", + "cpu": [ + "arm64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@napi-rs/canvas-linux-riscv64-gnu": { + "version": "0.1.76", + "resolved": "https://registry.npmjs.org/@napi-rs/canvas-linux-riscv64-gnu/-/canvas-linux-riscv64-gnu-0.1.76.tgz", + "integrity": "sha512-G2PsFwsP+r4syEoNLStV3n1wtNAClwf8s/qB57bexG08R4f4WaiBd+x+d4iYS0Y5o90YIEm8/ewZn4bLIa0wNQ==", + "cpu": [ + "riscv64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@napi-rs/canvas-linux-x64-gnu": { + "version": "0.1.76", + "resolved": "https://registry.npmjs.org/@napi-rs/canvas-linux-x64-gnu/-/canvas-linux-x64-gnu-0.1.76.tgz", + "integrity": "sha512-SNK+vgge4DnuONYdYE3Y09LivGgUiUPQDU+PdGNZJIzIi0hRDLcA59eag8LGeQfPmJW84c1aZD04voihybKFog==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@napi-rs/canvas-linux-x64-musl": { + "version": "0.1.76", + "resolved": "https://registry.npmjs.org/@napi-rs/canvas-linux-x64-musl/-/canvas-linux-x64-musl-0.1.76.tgz", + "integrity": "sha512-tWHLBI9iVoR1NsfpHz1MGERTkqcca8akbH/CzX6JQUNC+lJOeYYXeRuK8hKqMIg1LI+4QOMAtHNVeZu8NvjEug==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@napi-rs/canvas-win32-x64-msvc": { + "version": "0.1.76", + "resolved": "https://registry.npmjs.org/@napi-rs/canvas-win32-x64-msvc/-/canvas-win32-x64-msvc-0.1.76.tgz", + "integrity": "sha512-ifM5HOGw2hP5QLQzCB41Riw3Pq5yKAAjZpn+lJC0sYBmyS2s/Kq6KpTOKxf0CuptkI1wMcRcYQfhLRdeWiYvIg==", + "cpu": [ + "x64" + ], + "license": "MIT", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@neoconfetti/react": { + "version": "1.0.0", + "resolved": "https://registry.npmjs.org/@neoconfetti/react/-/react-1.0.0.tgz", + "integrity": "sha512-klcSooChXXOzIm+SE5IISIAn3bYzYfPjbX7D7HoqZL84oAfgREeSg5vSIaSFH+DaGzzvImTyWe1OyrJ67vik4A==", + "dev": true, + "license": "MIT" + }, + "node_modules/@nodelib/fs.scandir": { + "version": "2.1.5", + "resolved": "https://registry.npmjs.org/@nodelib/fs.scandir/-/fs.scandir-2.1.5.tgz", + "integrity": "sha512-vq24Bq3ym5HEQm2NKCr3yXDwjc7vTsEThRDnkp2DK9p1uqLR+DHurm/NOTo0KG7HYHU7eppKZj3MyqYuMBf62g==", + "dev": true, + "license": "MIT", + "dependencies": { + "@nodelib/fs.stat": "2.0.5", + "run-parallel": "^1.1.9" + }, + "engines": { + "node": ">= 8" + } + }, + "node_modules/@nodelib/fs.stat": { + "version": "2.0.5", + "resolved": "https://registry.npmjs.org/@nodelib/fs.stat/-/fs.stat-2.0.5.tgz", + "integrity": "sha512-RkhPPp2zrqDAQA/2jNhnztcPAlv64XdhIp7a7454A5ovI7Bukxgt7MX7udwAu3zg1DcpPU0rz3VV1SeaqvY4+A==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">= 8" + } + }, + "node_modules/@nodelib/fs.walk": { + "version": "1.2.8", + "resolved": "https://registry.npmjs.org/@nodelib/fs.walk/-/fs.walk-1.2.8.tgz", + "integrity": "sha512-oGB+UxlgWcgQkgwo8GcEGwemoTFt3FIO9ababBmaGwXIoBKZ+GTy0pP185beGg7Llih/NSHSV2XAs1lnznocSg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@nodelib/fs.scandir": "2.1.5", + "fastq": "^1.6.0" + }, + "engines": { + "node": ">= 8" + } + }, + "node_modules/@playwright/test": { + "version": "1.54.1", + "resolved": "https://registry.npmjs.org/@playwright/test/-/test-1.54.1.tgz", + "integrity": "sha512-FS8hQ12acieG2dYSksmLOF7BNxnVf2afRJdCuM1eMSxj6QTSE6G4InGF7oApGgDb65MX7AwMVlIkpru0yZA4Xw==", + "dev": true, + "license": "Apache-2.0", + "dependencies": { + "playwright": "1.54.1" + }, + "bin": { + "playwright": "cli.js" + }, + "engines": { + "node": ">=18" + } + }, + "node_modules/@polka/url": { + "version": "1.0.0-next.29", + "resolved": "https://registry.npmjs.org/@polka/url/-/url-1.0.0-next.29.tgz", + "integrity": "sha512-wwQAWhWSuHaag8c4q/KN/vCoeOJYshAIvMQwD4GpSb3OiZklFfvAgmj0VCBBImRpuF/aFgIRzllXlVX93Jevww==", + "dev": true, + "license": "MIT" + }, + "node_modules/@rollup/rollup-android-arm-eabi": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-android-arm-eabi/-/rollup-android-arm-eabi-4.45.1.tgz", + "integrity": "sha512-NEySIFvMY0ZQO+utJkgoMiCAjMrGvnbDLHvcmlA33UXJpYBCvlBEbMMtV837uCkS+plG2umfhn0T5mMAxGrlRA==", + "cpu": [ + "arm" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "android" + ] + }, + "node_modules/@rollup/rollup-android-arm64": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-android-arm64/-/rollup-android-arm64-4.45.1.tgz", + "integrity": "sha512-ujQ+sMXJkg4LRJaYreaVx7Z/VMgBBd89wGS4qMrdtfUFZ+TSY5Rs9asgjitLwzeIbhwdEhyj29zhst3L1lKsRQ==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "android" + ] + }, + "node_modules/@rollup/rollup-darwin-arm64": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-darwin-arm64/-/rollup-darwin-arm64-4.45.1.tgz", + "integrity": "sha512-FSncqHvqTm3lC6Y13xncsdOYfxGSLnP+73k815EfNmpewPs+EyM49haPS105Rh4aF5mJKywk9X0ogzLXZzN9lA==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ] + }, + "node_modules/@rollup/rollup-darwin-x64": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-darwin-x64/-/rollup-darwin-x64-4.45.1.tgz", + "integrity": "sha512-2/vVn/husP5XI7Fsf/RlhDaQJ7x9zjvC81anIVbr4b/f0xtSmXQTFcGIQ/B1cXIYM6h2nAhJkdMHTnD7OtQ9Og==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ] + }, + "node_modules/@rollup/rollup-freebsd-arm64": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-freebsd-arm64/-/rollup-freebsd-arm64-4.45.1.tgz", + "integrity": "sha512-4g1kaDxQItZsrkVTdYQ0bxu4ZIQ32cotoQbmsAnW1jAE4XCMbcBPDirX5fyUzdhVCKgPcrwWuucI8yrVRBw2+g==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "freebsd" + ] + }, + "node_modules/@rollup/rollup-freebsd-x64": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-freebsd-x64/-/rollup-freebsd-x64-4.45.1.tgz", + "integrity": "sha512-L/6JsfiL74i3uK1Ti2ZFSNsp5NMiM4/kbbGEcOCps99aZx3g8SJMO1/9Y0n/qKlWZfn6sScf98lEOUe2mBvW9A==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "freebsd" + ] + }, + "node_modules/@rollup/rollup-linux-arm-gnueabihf": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm-gnueabihf/-/rollup-linux-arm-gnueabihf-4.45.1.tgz", + "integrity": "sha512-RkdOTu2jK7brlu+ZwjMIZfdV2sSYHK2qR08FUWcIoqJC2eywHbXr0L8T/pONFwkGukQqERDheaGTeedG+rra6Q==", + "cpu": [ + "arm" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-arm-musleabihf": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm-musleabihf/-/rollup-linux-arm-musleabihf-4.45.1.tgz", + "integrity": "sha512-3kJ8pgfBt6CIIr1o+HQA7OZ9mp/zDk3ctekGl9qn/pRBgrRgfwiffaUmqioUGN9hv0OHv2gxmvdKOkARCtRb8Q==", + "cpu": [ + "arm" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-arm64-gnu": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm64-gnu/-/rollup-linux-arm64-gnu-4.45.1.tgz", + "integrity": "sha512-k3dOKCfIVixWjG7OXTCOmDfJj3vbdhN0QYEqB+OuGArOChek22hn7Uy5A/gTDNAcCy5v2YcXRJ/Qcnm4/ma1xw==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-arm64-musl": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm64-musl/-/rollup-linux-arm64-musl-4.45.1.tgz", + "integrity": "sha512-PmI1vxQetnM58ZmDFl9/Uk2lpBBby6B6rF4muJc65uZbxCs0EA7hhKCk2PKlmZKuyVSHAyIw3+/SiuMLxKxWog==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-loongarch64-gnu": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-loongarch64-gnu/-/rollup-linux-loongarch64-gnu-4.45.1.tgz", + "integrity": "sha512-9UmI0VzGmNJ28ibHW2GpE2nF0PBQqsyiS4kcJ5vK+wuwGnV5RlqdczVocDSUfGX/Na7/XINRVoUgJyFIgipoRg==", + "cpu": [ + "loong64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-powerpc64le-gnu": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-powerpc64le-gnu/-/rollup-linux-powerpc64le-gnu-4.45.1.tgz", + "integrity": "sha512-7nR2KY8oEOUTD3pBAxIBBbZr0U7U+R9HDTPNy+5nVVHDXI4ikYniH1oxQz9VoB5PbBU1CZuDGHkLJkd3zLMWsg==", + "cpu": [ + "ppc64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-riscv64-gnu": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-riscv64-gnu/-/rollup-linux-riscv64-gnu-4.45.1.tgz", + "integrity": "sha512-nlcl3jgUultKROfZijKjRQLUu9Ma0PeNv/VFHkZiKbXTBQXhpytS8CIj5/NfBeECZtY2FJQubm6ltIxm/ftxpw==", + "cpu": [ + "riscv64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-riscv64-musl": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-riscv64-musl/-/rollup-linux-riscv64-musl-4.45.1.tgz", + "integrity": "sha512-HJV65KLS51rW0VY6rvZkiieiBnurSzpzore1bMKAhunQiECPuxsROvyeaot/tcK3A3aGnI+qTHqisrpSgQrpgA==", + "cpu": [ + "riscv64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-s390x-gnu": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-s390x-gnu/-/rollup-linux-s390x-gnu-4.45.1.tgz", + "integrity": "sha512-NITBOCv3Qqc6hhwFt7jLV78VEO/il4YcBzoMGGNxznLgRQf43VQDae0aAzKiBeEPIxnDrACiMgbqjuihx08OOw==", + "cpu": [ + "s390x" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-x64-gnu": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-x64-gnu/-/rollup-linux-x64-gnu-4.45.1.tgz", + "integrity": "sha512-+E/lYl6qu1zqgPEnTrs4WysQtvc/Sh4fC2nByfFExqgYrqkKWp1tWIbe+ELhixnenSpBbLXNi6vbEEJ8M7fiHw==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-linux-x64-musl": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-x64-musl/-/rollup-linux-x64-musl-4.45.1.tgz", + "integrity": "sha512-a6WIAp89p3kpNoYStITT9RbTbTnqarU7D8N8F2CV+4Cl9fwCOZraLVuVFvlpsW0SbIiYtEnhCZBPLoNdRkjQFw==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ] + }, + "node_modules/@rollup/rollup-win32-arm64-msvc": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-arm64-msvc/-/rollup-win32-arm64-msvc-4.45.1.tgz", + "integrity": "sha512-T5Bi/NS3fQiJeYdGvRpTAP5P02kqSOpqiopwhj0uaXB6nzs5JVi2XMJb18JUSKhCOX8+UE1UKQufyD6Or48dJg==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "win32" + ] + }, + "node_modules/@rollup/rollup-win32-ia32-msvc": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-ia32-msvc/-/rollup-win32-ia32-msvc-4.45.1.tgz", + "integrity": "sha512-lxV2Pako3ujjuUe9jiU3/s7KSrDfH6IgTSQOnDWr9aJ92YsFd7EurmClK0ly/t8dzMkDtd04g60WX6yl0sGfdw==", + "cpu": [ + "ia32" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "win32" + ] + }, + "node_modules/@rollup/rollup-win32-x64-msvc": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-x64-msvc/-/rollup-win32-x64-msvc-4.45.1.tgz", + "integrity": "sha512-M/fKi4sasCdM8i0aWJjCSFm2qEnYRR8AMLG2kxp6wD13+tMGA4Z1tVAuHkNRjud5SW2EM3naLuK35w9twvf6aA==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "win32" + ] + }, + "node_modules/@standard-schema/spec": { + "version": "1.0.0", + "resolved": "https://registry.npmjs.org/@standard-schema/spec/-/spec-1.0.0.tgz", + "integrity": "sha512-m2bOd0f2RT9k8QJx1JN85cZYyH1RqFBdlwtkSlf4tBDYLCiiZnv1fIIwacK6cqwXavOydf0NPToMQgpKq+dVlA==", + "dev": true, + "license": "MIT" + }, + "node_modules/@storybook/addon-a11y": { + "version": "9.0.17", + "resolved": "https://registry.npmjs.org/@storybook/addon-a11y/-/addon-a11y-9.0.17.tgz", + "integrity": "sha512-9cXNK3q/atx3hwJAt9HkJbd9vUxCXfKKiNNuSACbf8h9/j6u3jktulKOf6Xjc3B8lwn6ZpdK/x1HHZN2kTqsvg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@storybook/global": "^5.0.0", + "axe-core": "^4.2.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/storybook" + }, + "peerDependencies": { + "storybook": "^9.0.17" + } + }, + "node_modules/@storybook/addon-docs": { + "version": "9.0.17", + "resolved": "https://registry.npmjs.org/@storybook/addon-docs/-/addon-docs-9.0.17.tgz", + "integrity": "sha512-LOX/kKgQGnyulrqZHsvf77+ZoH/nSUaplGr5hvZglW/U6ak6fO9seJyXAzVKEnC6p+F8n02kFBZbi3s+znQhSg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@mdx-js/react": "^3.0.0", + "@storybook/csf-plugin": "9.0.17", + "@storybook/icons": "^1.2.12", + "@storybook/react-dom-shim": "9.0.17", + "react": "^16.8.0 || ^17.0.0 || ^18.0.0 || ^19.0.0", + "react-dom": "^16.8.0 || ^17.0.0 || ^18.0.0 || ^19.0.0", + "ts-dedent": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/storybook" + }, + "peerDependencies": { + "storybook": "^9.0.17" + } + }, + "node_modules/@storybook/addon-svelte-csf": { + "version": "5.0.7", + "resolved": "https://registry.npmjs.org/@storybook/addon-svelte-csf/-/addon-svelte-csf-5.0.7.tgz", + "integrity": "sha512-6Zmy5HjOlrrG6OoKRTGDr9LR6zRK4/Sa7raFzQRKHGASgMlfKsMdNTNO0sxnMUWCu2JMS6HsuoLtB3Ma8SlYtg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@storybook/csf": "^0.1.13", + "dedent": "^1.5.3", + "es-toolkit": "^1.26.1", + "esrap": "^1.2.2", + "magic-string": "^0.30.12", + "svelte-ast-print": "^0.4.0", + "zimmerframe": "^1.1.2" + }, + "peerDependencies": { + "@storybook/svelte": "^0.0.0-0 || ^8.2.0 || ^9.0.0 || ^9.1.0-0", + "@sveltejs/vite-plugin-svelte": "^4.0.0 || ^5.0.0 || ^6.0.0", + "storybook": "^0.0.0-0 || ^8.2.0 || ^9.0.0 || ^9.1.0-0", + "svelte": "^5.0.0", + "vite": "^5.0.0 || ^6.0.0 || ^7.0.0" + } + }, + "node_modules/@storybook/addon-vitest": { + "version": "9.0.17", + "resolved": "https://registry.npmjs.org/@storybook/addon-vitest/-/addon-vitest-9.0.17.tgz", + "integrity": "sha512-eogqcGbACR1sTedBSE2SP/4QV1ruicHYEhYjBtoPIjvYgymN1g5KSuQNysLx4f0SvAzczrcNjX2WVVLX2DVyzA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@storybook/global": "^5.0.0", + "@storybook/icons": "^1.4.0", + "prompts": "^2.4.0", + "ts-dedent": "^2.2.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/storybook" + }, + "peerDependencies": { + "@vitest/browser": "^3.0.0", + "@vitest/runner": "^3.0.0", + "storybook": "^9.0.17", + "vitest": "^3.0.0" + }, + "peerDependenciesMeta": { + "@vitest/browser": { + "optional": true + }, + "@vitest/runner": { + "optional": true + }, + "vitest": { + "optional": true + } + } + }, + "node_modules/@storybook/builder-vite": { + "version": "9.0.17", + "resolved": "https://registry.npmjs.org/@storybook/builder-vite/-/builder-vite-9.0.17.tgz", + "integrity": "sha512-lyuvgGhb0NaVk1tdB4xwzky6+YXQfxlxfNQqENYZ9uYQZdPfErMa4ZTXVQTV+CQHAa2NL+p/dG2JPAeu39e9UA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@storybook/csf-plugin": "9.0.17", + "ts-dedent": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/storybook" + }, + "peerDependencies": { + "storybook": "^9.0.17", + "vite": "^5.0.0 || ^6.0.0 || ^7.0.0" + } + }, + "node_modules/@storybook/csf": { + "version": "0.1.13", + "resolved": "https://registry.npmjs.org/@storybook/csf/-/csf-0.1.13.tgz", + "integrity": "sha512-7xOOwCLGB3ebM87eemep89MYRFTko+D8qE7EdAAq74lgdqRR5cOUtYWJLjO2dLtP94nqoOdHJo6MdLLKzg412Q==", + "dev": true, + "license": "MIT", + "dependencies": { + "type-fest": "^2.19.0" + } + }, + "node_modules/@storybook/csf-plugin": { + "version": "9.0.17", + "resolved": "https://registry.npmjs.org/@storybook/csf-plugin/-/csf-plugin-9.0.17.tgz", + "integrity": "sha512-6Q4eo1ObrLlsnB6bIt6T8+45XAb4to2pQGNrI7QPkLQRLrZinrJcNbLY7AGkyIoCOEsEbq08n09/nClQUbu8HA==", + "dev": true, + "license": "MIT", + "dependencies": { + "unplugin": "^1.3.1" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/storybook" + }, + "peerDependencies": { + "storybook": "^9.0.17" + } + }, + "node_modules/@storybook/global": { + "version": "5.0.0", + "resolved": "https://registry.npmjs.org/@storybook/global/-/global-5.0.0.tgz", + "integrity": "sha512-FcOqPAXACP0I3oJ/ws6/rrPT9WGhu915Cg8D02a9YxLo0DE9zI+a9A5gRGvmQ09fiWPukqI8ZAEoQEdWUKMQdQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/@storybook/icons": { + "version": "1.4.0", + "resolved": "https://registry.npmjs.org/@storybook/icons/-/icons-1.4.0.tgz", + "integrity": "sha512-Td73IeJxOyalzvjQL+JXx72jlIYHgs+REaHiREOqfpo3A2AYYG71AUbcv+lg7mEDIweKVCxsMQ0UKo634c8XeA==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=14.0.0" + }, + "peerDependencies": { + "react": "^16.8.0 || ^17.0.0 || ^18.0.0 || ^19.0.0-beta", + "react-dom": "^16.8.0 || ^17.0.0 || ^18.0.0 || ^19.0.0-beta" + } + }, + "node_modules/@storybook/react-dom-shim": { + "version": "9.0.17", + "resolved": "https://registry.npmjs.org/@storybook/react-dom-shim/-/react-dom-shim-9.0.17.tgz", + "integrity": "sha512-ak/x/m6MDDxdE6rCDymTltaiQF3oiKrPHSwfM+YPgQR6MVmzTTs4+qaPfeev7FZEHq23IkfDMTmSTTJtX7Vs9A==", + "dev": true, + "license": "MIT", + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/storybook" + }, + "peerDependencies": { + "react": "^16.8.0 || ^17.0.0 || ^18.0.0 || ^19.0.0-beta", + "react-dom": "^16.8.0 || ^17.0.0 || ^18.0.0 || ^19.0.0-beta", + "storybook": "^9.0.17" + } + }, + "node_modules/@storybook/svelte": { + "version": "9.0.17", + "resolved": "https://registry.npmjs.org/@storybook/svelte/-/svelte-9.0.17.tgz", + "integrity": "sha512-RwOswdq7S3+ZOuoM/oRrcmlsKdjcd/3wMHbuirzYoAhdwsjubSuRepMV64O9RnlXd3x7rZw4fXpq1M/SVo5XiQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "ts-dedent": "^2.0.0", + "type-fest": "~2.19" + }, + "engines": { + "node": ">=20.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/storybook" + }, + "peerDependencies": { + "storybook": "^9.0.17", + "svelte": "^5.0.0" + } + }, + "node_modules/@storybook/sveltekit": { + "version": "9.0.17", + "resolved": "https://registry.npmjs.org/@storybook/sveltekit/-/sveltekit-9.0.17.tgz", + "integrity": "sha512-CUOATuW5Qk3SjNvmjH+wyx2GCsMF1cvw3gwkujV9kehPebzV20NhgHpbzSoepvwF7+Bj6jl8V6UxiMWk0jJFmA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@storybook/builder-vite": "9.0.17", + "@storybook/svelte": "9.0.17", + "@storybook/svelte-vite": "9.0.17" + }, + "engines": { + "node": ">=20.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/storybook" + }, + "peerDependencies": { + "storybook": "^9.0.17", + "svelte": "^5.0.0", + "vite": "^5.0.0 || ^6.0.0 || ^7.0.0" + } + }, + "node_modules/@storybook/sveltekit/node_modules/@storybook/svelte-vite": { + "version": "9.0.17", + "resolved": "https://registry.npmjs.org/@storybook/svelte-vite/-/svelte-vite-9.0.17.tgz", + "integrity": "sha512-fRIxOZy9IRI6BfL1LgFn+B+IckGOlT1SstD01y9ddO4pVKWih/l+vb44bnZs+Z0faJZbrG/LgfnXTOPj052Z8g==", + "dev": true, + "license": "MIT", + "dependencies": { + "@storybook/builder-vite": "9.0.17", + "@storybook/svelte": "9.0.17", + "magic-string": "^0.30.0", + "svelte2tsx": "^0.7.35", + "typescript": "^4.9.4 || ^5.0.0" + }, + "engines": { + "node": ">=20.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/storybook" + }, + "peerDependencies": { + "@sveltejs/vite-plugin-svelte": "^2.0.0 || ^3.0.0 || ^4.0.0 || ^5.0.0", + "storybook": "^9.0.17", + "svelte": "^5.0.0", + "vite": "^5.0.0 || ^6.0.0 || ^7.0.0" + } + }, + "node_modules/@storybook/sveltekit/node_modules/@sveltejs/vite-plugin-svelte": { + "version": "5.1.1", + "resolved": "https://registry.npmjs.org/@sveltejs/vite-plugin-svelte/-/vite-plugin-svelte-5.1.1.tgz", + "integrity": "sha512-Y1Cs7hhTc+a5E9Va/xwKlAJoariQyHY+5zBgCZg4PFWNYQ1nMN9sjK1zhw1gK69DuqVP++sht/1GZg1aRwmAXQ==", + "dev": true, + "license": "MIT", + "peer": true, + "dependencies": { + "@sveltejs/vite-plugin-svelte-inspector": "^4.0.1", + "debug": "^4.4.1", + "deepmerge": "^4.3.1", + "kleur": "^4.1.5", + "magic-string": "^0.30.17", + "vitefu": "^1.0.6" + }, + "engines": { + "node": "^18.0.0 || ^20.0.0 || >=22" + }, + "peerDependencies": { + "svelte": "^5.0.0", + "vite": "^6.0.0" + } + }, + "node_modules/@storybook/sveltekit/node_modules/@sveltejs/vite-plugin-svelte/node_modules/@sveltejs/vite-plugin-svelte-inspector": { + "version": "4.0.1", + "resolved": "https://registry.npmjs.org/@sveltejs/vite-plugin-svelte-inspector/-/vite-plugin-svelte-inspector-4.0.1.tgz", + "integrity": "sha512-J/Nmb2Q2y7mck2hyCX4ckVHcR5tu2J+MtBEQqpDrrgELZ2uvraQcK/ioCV61AqkdXFgriksOKIceDcQmqnGhVw==", + "dev": true, + "license": "MIT", + "peer": true, + "dependencies": { + "debug": "^4.3.7" + }, + "engines": { + "node": "^18.0.0 || ^20.0.0 || >=22" + }, + "peerDependencies": { + "@sveltejs/vite-plugin-svelte": "^5.0.0", + "svelte": "^5.0.0", + "vite": "^6.0.0" + } + }, + "node_modules/@sveltejs/acorn-typescript": { + "version": "1.0.5", + "resolved": "https://registry.npmjs.org/@sveltejs/acorn-typescript/-/acorn-typescript-1.0.5.tgz", + "integrity": "sha512-IwQk4yfwLdibDlrXVE04jTZYlLnwsTT2PIOQQGNLWfjavGifnk1JD1LcZjZaBTRcxZu2FfPfNLOE04DSu9lqtQ==", + "license": "MIT", + "peerDependencies": { + "acorn": "^8.9.0" + } + }, + "node_modules/@sveltejs/adapter-static": { + "version": "3.0.9", + "resolved": "https://registry.npmjs.org/@sveltejs/adapter-static/-/adapter-static-3.0.9.tgz", + "integrity": "sha512-aytHXcMi7lb9ljsWUzXYQ0p5X1z9oWud2olu/EpmH7aCu4m84h7QLvb5Wp+CFirKcwoNnYvYWhyP/L8Vh1ztdw==", + "dev": true, + "license": "MIT", + "peerDependencies": { + "@sveltejs/kit": "^2.0.0" + } + }, + "node_modules/@sveltejs/kit": { + "version": "2.37.0", + "resolved": "https://registry.npmjs.org/@sveltejs/kit/-/kit-2.37.0.tgz", + "integrity": "sha512-xgKtpjQ6Ry4mdShd01ht5AODUsW7+K1iValPDq7QX8zI1hWOKREH9GjG8SRCN5tC4K7UXmMhuQam7gbLByVcnw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@standard-schema/spec": "^1.0.0", + "@sveltejs/acorn-typescript": "^1.0.5", + "@types/cookie": "^0.6.0", + "acorn": "^8.14.1", + "cookie": "^0.6.0", + "devalue": "^5.3.2", + "esm-env": "^1.2.2", + "kleur": "^4.1.5", + "magic-string": "^0.30.5", + "mrmime": "^2.0.0", + "sade": "^1.8.1", + "set-cookie-parser": "^2.6.0", + "sirv": "^3.0.0" + }, + "bin": { + "svelte-kit": "svelte-kit.js" + }, + "engines": { + "node": ">=18.13" + }, + "peerDependencies": { + "@opentelemetry/api": "^1.0.0", + "@sveltejs/vite-plugin-svelte": "^3.0.0 || ^4.0.0-next.1 || ^5.0.0 || ^6.0.0-next.0", + "svelte": "^4.0.0 || ^5.0.0-next.0", + "vite": "^5.0.3 || ^6.0.0 || ^7.0.0-beta.0" + }, + "peerDependenciesMeta": { + "@opentelemetry/api": { + "optional": true + } + } + }, + "node_modules/@sveltejs/vite-plugin-svelte": { + "version": "6.1.0", + "resolved": "https://registry.npmjs.org/@sveltejs/vite-plugin-svelte/-/vite-plugin-svelte-6.1.0.tgz", + "integrity": "sha512-+U6lz1wvGEG/BvQyL4z/flyNdQ9xDNv5vrh+vWBWTHaebqT0c9RNggpZTo/XSPoHsSCWBlYaTlRX8pZ9GATXCw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@sveltejs/vite-plugin-svelte-inspector": "^5.0.0-next.1", + "debug": "^4.4.1", + "deepmerge": "^4.3.1", + "kleur": "^4.1.5", + "magic-string": "^0.30.17", + "vitefu": "^1.1.1" + }, + "engines": { + "node": "^20.19 || ^22.12 || >=24" + }, + "peerDependencies": { + "svelte": "^5.0.0", + "vite": "^6.3.0 || ^7.0.0" + } + }, + "node_modules/@sveltejs/vite-plugin-svelte-inspector": { + "version": "5.0.0", + "resolved": "https://registry.npmjs.org/@sveltejs/vite-plugin-svelte-inspector/-/vite-plugin-svelte-inspector-5.0.0.tgz", + "integrity": "sha512-iwQ8Z4ET6ZFSt/gC+tVfcsSBHwsqc6RumSaiLUkAurW3BCpJam65cmHw0oOlDMTO0u+PZi9hilBRYN+LZNHTUQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "debug": "^4.4.1" + }, + "engines": { + "node": "^20.19 || ^22.12 || >=24" + }, + "peerDependencies": { + "@sveltejs/vite-plugin-svelte": "^6.0.0-next.0", + "svelte": "^5.0.0", + "vite": "^6.3.0 || ^7.0.0" + } + }, + "node_modules/@swc/helpers": { + "version": "0.5.17", + "resolved": "https://registry.npmjs.org/@swc/helpers/-/helpers-0.5.17.tgz", + "integrity": "sha512-5IKx/Y13RsYd+sauPb2x+U/xZikHjolzfuDgTAl/Tdf3Q8rslRvC19NKDLgAJQ6wsqADk10ntlv08nPFw/gO/A==", + "dev": true, + "license": "Apache-2.0", + "dependencies": { + "tslib": "^2.8.0" + } + }, + "node_modules/@tailwindcss/forms": { + "version": "0.5.10", + "resolved": "https://registry.npmjs.org/@tailwindcss/forms/-/forms-0.5.10.tgz", + "integrity": "sha512-utI1ONF6uf/pPNO68kmN1b8rEwNXv3czukalo8VtJH8ksIkZXr3Q3VYudZLkCsDd4Wku120uF02hYK25XGPorw==", + "dev": true, + "license": "MIT", + "dependencies": { + "mini-svg-data-uri": "^1.2.3" + }, + "peerDependencies": { + "tailwindcss": ">=3.0.0 || >= 3.0.0-alpha.1 || >= 4.0.0-alpha.20 || >= 4.0.0-beta.1" + } + }, + "node_modules/@tailwindcss/node": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/node/-/node-4.1.11.tgz", + "integrity": "sha512-yzhzuGRmv5QyU9qLNg4GTlYI6STedBWRE7NjxP45CsFYYq9taI0zJXZBMqIC/c8fViNLhmrbpSFS57EoxUmD6Q==", + "dev": true, + "license": "MIT", + "dependencies": { + "@ampproject/remapping": "^2.3.0", + "enhanced-resolve": "^5.18.1", + "jiti": "^2.4.2", + "lightningcss": "1.30.1", + "magic-string": "^0.30.17", + "source-map-js": "^1.2.1", + "tailwindcss": "4.1.11" + } + }, + "node_modules/@tailwindcss/oxide": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide/-/oxide-4.1.11.tgz", + "integrity": "sha512-Q69XzrtAhuyfHo+5/HMgr1lAiPP/G40OMFAnws7xcFEYqcypZmdW8eGXaOUIeOl1dzPJBPENXgbjsOyhg2nkrg==", + "dev": true, + "hasInstallScript": true, + "license": "MIT", + "dependencies": { + "detect-libc": "^2.0.4", + "tar": "^7.4.3" + }, + "engines": { + "node": ">= 10" + }, + "optionalDependencies": { + "@tailwindcss/oxide-android-arm64": "4.1.11", + "@tailwindcss/oxide-darwin-arm64": "4.1.11", + "@tailwindcss/oxide-darwin-x64": "4.1.11", + "@tailwindcss/oxide-freebsd-x64": "4.1.11", + "@tailwindcss/oxide-linux-arm-gnueabihf": "4.1.11", + "@tailwindcss/oxide-linux-arm64-gnu": "4.1.11", + "@tailwindcss/oxide-linux-arm64-musl": "4.1.11", + "@tailwindcss/oxide-linux-x64-gnu": "4.1.11", + "@tailwindcss/oxide-linux-x64-musl": "4.1.11", + "@tailwindcss/oxide-wasm32-wasi": "4.1.11", + "@tailwindcss/oxide-win32-arm64-msvc": "4.1.11", + "@tailwindcss/oxide-win32-x64-msvc": "4.1.11" + } + }, + "node_modules/@tailwindcss/oxide-android-arm64": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-android-arm64/-/oxide-android-arm64-4.1.11.tgz", + "integrity": "sha512-3IfFuATVRUMZZprEIx9OGDjG3Ou3jG4xQzNTvjDoKmU9JdmoCohQJ83MYd0GPnQIu89YoJqvMM0G3uqLRFtetg==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "android" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@tailwindcss/oxide-darwin-arm64": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-darwin-arm64/-/oxide-darwin-arm64-4.1.11.tgz", + "integrity": "sha512-ESgStEOEsyg8J5YcMb1xl8WFOXfeBmrhAwGsFxxB2CxY9evy63+AtpbDLAyRkJnxLy2WsD1qF13E97uQyP1lfQ==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@tailwindcss/oxide-darwin-x64": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-darwin-x64/-/oxide-darwin-x64-4.1.11.tgz", + "integrity": "sha512-EgnK8kRchgmgzG6jE10UQNaH9Mwi2n+yw1jWmof9Vyg2lpKNX2ioe7CJdf9M5f8V9uaQxInenZkOxnTVL3fhAw==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@tailwindcss/oxide-freebsd-x64": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-freebsd-x64/-/oxide-freebsd-x64-4.1.11.tgz", + "integrity": "sha512-xdqKtbpHs7pQhIKmqVpxStnY1skuNh4CtbcyOHeX1YBE0hArj2romsFGb6yUmzkq/6M24nkxDqU8GYrKrz+UcA==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "freebsd" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@tailwindcss/oxide-linux-arm-gnueabihf": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-linux-arm-gnueabihf/-/oxide-linux-arm-gnueabihf-4.1.11.tgz", + "integrity": "sha512-ryHQK2eyDYYMwB5wZL46uoxz2zzDZsFBwfjssgB7pzytAeCCa6glsiJGjhTEddq/4OsIjsLNMAiMlHNYnkEEeg==", + "cpu": [ + "arm" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@tailwindcss/oxide-linux-arm64-gnu": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-linux-arm64-gnu/-/oxide-linux-arm64-gnu-4.1.11.tgz", + "integrity": "sha512-mYwqheq4BXF83j/w75ewkPJmPZIqqP1nhoghS9D57CLjsh3Nfq0m4ftTotRYtGnZd3eCztgbSPJ9QhfC91gDZQ==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@tailwindcss/oxide-linux-arm64-musl": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-linux-arm64-musl/-/oxide-linux-arm64-musl-4.1.11.tgz", + "integrity": "sha512-m/NVRFNGlEHJrNVk3O6I9ggVuNjXHIPoD6bqay/pubtYC9QIdAMpS+cswZQPBLvVvEF6GtSNONbDkZrjWZXYNQ==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@tailwindcss/oxide-linux-x64-gnu": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-linux-x64-gnu/-/oxide-linux-x64-gnu-4.1.11.tgz", + "integrity": "sha512-YW6sblI7xukSD2TdbbaeQVDysIm/UPJtObHJHKxDEcW2exAtY47j52f8jZXkqE1krdnkhCMGqP3dbniu1Te2Fg==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@tailwindcss/oxide-linux-x64-musl": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-linux-x64-musl/-/oxide-linux-x64-musl-4.1.11.tgz", + "integrity": "sha512-e3C/RRhGunWYNC3aSF7exsQkdXzQ/M+aYuZHKnw4U7KQwTJotnWsGOIVih0s2qQzmEzOFIJ3+xt7iq67K/p56Q==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@tailwindcss/oxide-wasm32-wasi": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-wasm32-wasi/-/oxide-wasm32-wasi-4.1.11.tgz", + "integrity": "sha512-Xo1+/GU0JEN/C/dvcammKHzeM6NqKovG+6921MR6oadee5XPBaKOumrJCXvopJ/Qb5TH7LX/UAywbqrP4lax0g==", + "bundleDependencies": [ + "@napi-rs/wasm-runtime", + "@emnapi/core", + "@emnapi/runtime", + "@tybys/wasm-util", + "@emnapi/wasi-threads", + "tslib" + ], + "cpu": [ + "wasm32" + ], + "dev": true, + "license": "MIT", + "optional": true, + "dependencies": { + "@emnapi/core": "^1.4.3", + "@emnapi/runtime": "^1.4.3", + "@emnapi/wasi-threads": "^1.0.2", + "@napi-rs/wasm-runtime": "^0.2.11", + "@tybys/wasm-util": "^0.9.0", + "tslib": "^2.8.0" + }, + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/@tailwindcss/oxide-win32-arm64-msvc": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-win32-arm64-msvc/-/oxide-win32-arm64-msvc-4.1.11.tgz", + "integrity": "sha512-UgKYx5PwEKrac3GPNPf6HVMNhUIGuUh4wlDFR2jYYdkX6pL/rn73zTq/4pzUm8fOjAn5L8zDeHp9iXmUGOXZ+w==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@tailwindcss/oxide-win32-x64-msvc": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/oxide-win32-x64-msvc/-/oxide-win32-x64-msvc-4.1.11.tgz", + "integrity": "sha512-YfHoggn1j0LK7wR82TOucWc5LDCguHnoS879idHekmmiR7g9HUtMw9MI0NHatS28u/Xlkfi9w5RJWgz2Dl+5Qg==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MIT", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">= 10" + } + }, + "node_modules/@tailwindcss/typography": { + "version": "0.5.16", + "resolved": "https://registry.npmjs.org/@tailwindcss/typography/-/typography-0.5.16.tgz", + "integrity": "sha512-0wDLwCVF5V3x3b1SGXPCDcdsbDHMBe+lkFzBRaHeLvNi+nrrnZ1lA18u+OTWO8iSWU2GxUOCvlXtDuqftc1oiA==", + "dev": true, + "license": "MIT", + "dependencies": { + "lodash.castarray": "^4.4.0", + "lodash.isplainobject": "^4.0.6", + "lodash.merge": "^4.6.2", + "postcss-selector-parser": "6.0.10" + }, + "peerDependencies": { + "tailwindcss": ">=3.0.0 || insiders || >=4.0.0-alpha.20 || >=4.0.0-beta.1" + } + }, + "node_modules/@tailwindcss/vite": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/@tailwindcss/vite/-/vite-4.1.11.tgz", + "integrity": "sha512-RHYhrR3hku0MJFRV+fN2gNbDNEh3dwKvY8XJvTxCSXeMOsCRSr+uKvDWQcbizrHgjML6ZmTE5OwMrl5wKcujCw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@tailwindcss/node": "4.1.11", + "@tailwindcss/oxide": "4.1.11", + "tailwindcss": "4.1.11" + }, + "peerDependencies": { + "vite": "^5.2.0 || ^6 || ^7" + } + }, + "node_modules/@testing-library/dom": { + "version": "10.4.0", + "resolved": "https://registry.npmjs.org/@testing-library/dom/-/dom-10.4.0.tgz", + "integrity": "sha512-pemlzrSESWbdAloYml3bAJMEfNh1Z7EduzqPKprCH5S341frlpYnUEW0H72dLxa6IsYr+mPno20GiSm+h9dEdQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "@babel/code-frame": "^7.10.4", + "@babel/runtime": "^7.12.5", + "@types/aria-query": "^5.0.1", + "aria-query": "5.3.0", + "chalk": "^4.1.0", + "dom-accessibility-api": "^0.5.9", + "lz-string": "^1.5.0", + "pretty-format": "^27.0.2" + }, + "engines": { + "node": ">=18" + } + }, + "node_modules/@testing-library/jest-dom": { + "version": "6.6.3", + "resolved": "https://registry.npmjs.org/@testing-library/jest-dom/-/jest-dom-6.6.3.tgz", + "integrity": "sha512-IteBhl4XqYNkM54f4ejhLRJiZNqcSCoXUOG2CPK7qbD322KjQozM4kHQOfkG2oln9b9HTYqs+Sae8vBATubxxA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@adobe/css-tools": "^4.4.0", + "aria-query": "^5.0.0", + "chalk": "^3.0.0", + "css.escape": "^1.5.1", + "dom-accessibility-api": "^0.6.3", + "lodash": "^4.17.21", + "redent": "^3.0.0" + }, + "engines": { + "node": ">=14", + "npm": ">=6", + "yarn": ">=1" + } + }, + "node_modules/@testing-library/jest-dom/node_modules/chalk": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/chalk/-/chalk-3.0.0.tgz", + "integrity": "sha512-4D3B6Wf41KOYRFdszmDqMCGq5VV/uMAB273JILmO+3jAlh8X4qDtdtgCR3fxtbLEMzSx22QdhnDcJvu2u1fVwg==", + "dev": true, + "license": "MIT", + "dependencies": { + "ansi-styles": "^4.1.0", + "supports-color": "^7.1.0" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/@testing-library/jest-dom/node_modules/dom-accessibility-api": { + "version": "0.6.3", + "resolved": "https://registry.npmjs.org/dom-accessibility-api/-/dom-accessibility-api-0.6.3.tgz", + "integrity": "sha512-7ZgogeTnjuHbo+ct10G9Ffp0mif17idi0IyWNVA/wcwcm7NPOD/WEHVP3n7n3MhXqxoIYm8d6MuZohYWIZ4T3w==", + "dev": true, + "license": "MIT" + }, + "node_modules/@testing-library/user-event": { + "version": "14.6.1", + "resolved": "https://registry.npmjs.org/@testing-library/user-event/-/user-event-14.6.1.tgz", + "integrity": "sha512-vq7fv0rnt+QTXgPxr5Hjc210p6YKq2kmdziLgnsZGgLJ9e6VAShx1pACLuRjd/AS/sr7phAR58OIIpf0LlmQNw==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=12", + "npm": ">=6" + }, + "peerDependencies": { + "@testing-library/dom": ">=7.21.4" + } + }, + "node_modules/@types/aria-query": { + "version": "5.0.4", + "resolved": "https://registry.npmjs.org/@types/aria-query/-/aria-query-5.0.4.tgz", + "integrity": "sha512-rfT93uj5s0PRL7EzccGMs3brplhcrghnDoV26NqKhCAS1hVo+WdNsPvE/yb6ilfr5hi2MEk6d5EWJTKdxg8jVw==", + "dev": true, + "license": "MIT" + }, + "node_modules/@types/chai": { + "version": "5.2.2", + "resolved": "https://registry.npmjs.org/@types/chai/-/chai-5.2.2.tgz", + "integrity": "sha512-8kB30R7Hwqf40JPiKhVzodJs2Qc1ZJ5zuT3uzw5Hq/dhNCl3G3l83jfpdI1e20BP348+fV7VIL/+FxaXkqBmWg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/deep-eql": "*" + } + }, + "node_modules/@types/cookie": { + "version": "0.6.0", + "resolved": "https://registry.npmjs.org/@types/cookie/-/cookie-0.6.0.tgz", + "integrity": "sha512-4Kh9a6B2bQciAhf7FSuMRRkUWecJgJu9nPnx3yzpsfXX/c50REIqpHY4C82bXP90qrLtXtkDxTZosYO3UpOwlA==", + "dev": true, + "license": "MIT" + }, + "node_modules/@types/debug": { + "version": "4.1.12", + "resolved": "https://registry.npmjs.org/@types/debug/-/debug-4.1.12.tgz", + "integrity": "sha512-vIChWdVG3LG1SMxEvI/AK+FWJthlrqlTu7fbrlywTkkaONwk/UAGaULXRlf8vkzFBLVm0zkMdCquhL5aOjhXPQ==", + "license": "MIT", + "dependencies": { + "@types/ms": "*" + } + }, + "node_modules/@types/deep-eql": { + "version": "4.0.2", + "resolved": "https://registry.npmjs.org/@types/deep-eql/-/deep-eql-4.0.2.tgz", + "integrity": "sha512-c9h9dVVMigMPc4bwTvC5dxqtqJZwQPePsWjPlpSOnojbor6pGqdk541lfA7AqFQr5pB1BRdq0juY9db81BwyFw==", + "dev": true, + "license": "MIT" + }, + "node_modules/@types/estree": { + "version": "1.0.8", + "resolved": "https://registry.npmjs.org/@types/estree/-/estree-1.0.8.tgz", + "integrity": "sha512-dWHzHa2WqEXI/O1E9OjrocMTKJl2mSrEolh1Iomrv6U+JuNwaHXsXx9bLu5gG7BUWFIN0skIQJQ/L1rIex4X6w==", + "license": "MIT" + }, + "node_modules/@types/hast": { + "version": "3.0.4", + "resolved": "https://registry.npmjs.org/@types/hast/-/hast-3.0.4.tgz", + "integrity": "sha512-WPs+bbQw5aCj+x6laNGWLH3wviHtoCv/P3+otBhbOhJgG8qtpdAMlTCxLtsTWA7LH1Oh/bFCHsBn0TPS5m30EQ==", + "license": "MIT", + "dependencies": { + "@types/unist": "*" + } + }, + "node_modules/@types/json-schema": { + "version": "7.0.15", + "resolved": "https://registry.npmjs.org/@types/json-schema/-/json-schema-7.0.15.tgz", + "integrity": "sha512-5+fP8P8MFNC+AyZCDxrB2pkZFPGzqQWUzpSeuuVLvm8VMcorNYavBqoFcxK8bQz4Qsbn4oUEEem4wDLfcysGHA==", + "dev": true, + "license": "MIT" + }, + "node_modules/@types/katex": { + "version": "0.16.7", + "resolved": "https://registry.npmjs.org/@types/katex/-/katex-0.16.7.tgz", + "integrity": "sha512-HMwFiRujE5PjrgwHQ25+bsLJgowjGjm5Z8FVSf0N6PwgJrwxH0QxzHYDcKsTfV3wva0vzrpqMTJS2jXPr5BMEQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/@types/mdast": { + "version": "4.0.4", + "resolved": "https://registry.npmjs.org/@types/mdast/-/mdast-4.0.4.tgz", + "integrity": "sha512-kGaNbPh1k7AFzgpud/gMdvIm5xuECykRR+JnWKQno9TAXVa6WIVCGTPvYGekIDL4uwCZQSYbUxNBSb1aUo79oA==", + "license": "MIT", + "dependencies": { + "@types/unist": "*" + } + }, + "node_modules/@types/mdx": { + "version": "2.0.13", + "resolved": "https://registry.npmjs.org/@types/mdx/-/mdx-2.0.13.tgz", + "integrity": "sha512-+OWZQfAYyio6YkJb3HLxDrvnx6SWWDbC0zVPfBRzUk0/nqoDyf6dNxQi3eArPe8rJ473nobTMQ/8Zk+LxJ+Yuw==", + "dev": true, + "license": "MIT" + }, + "node_modules/@types/ms": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/@types/ms/-/ms-2.1.0.tgz", + "integrity": "sha512-GsCCIZDE/p3i96vtEqx+7dBUGXrc7zeSK3wwPHIaRThS+9OhWIXRqzs4d6k1SVU8g91DrNRWxWUGhp5KXQb2VA==", + "license": "MIT" + }, + "node_modules/@types/node": { + "version": "22.16.5", + "resolved": "https://registry.npmjs.org/@types/node/-/node-22.16.5.tgz", + "integrity": "sha512-bJFoMATwIGaxxx8VJPeM8TonI8t579oRvgAuT8zFugJsJZgzqv0Fu8Mhp68iecjzG7cnN3mO2dJQ5uUM2EFrgQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "undici-types": "~6.21.0" + } + }, + "node_modules/@types/react": { + "version": "19.1.8", + "resolved": "https://registry.npmjs.org/@types/react/-/react-19.1.8.tgz", + "integrity": "sha512-AwAfQ2Wa5bCx9WP8nZL2uMZWod7J7/JSplxbTmBQ5ms6QpqNYm672H0Vu9ZVKVngQ+ii4R/byguVEUZQyeg44g==", + "dev": true, + "license": "MIT", + "peer": true, + "dependencies": { + "csstype": "^3.0.2" + } + }, + "node_modules/@types/unist": { + "version": "2.0.11", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-2.0.11.tgz", + "integrity": "sha512-CmBKiL6NNo/OqgmMn95Fk9Whlp2mtvIv+KNpQKN2F4SjvrEesubTRWGYSg+BnWZOnlCaSTU1sMpsBOzgbYhnsA==", + "license": "MIT" + }, + "node_modules/@typescript-eslint/eslint-plugin": { + "version": "8.37.0", + "resolved": "https://registry.npmjs.org/@typescript-eslint/eslint-plugin/-/eslint-plugin-8.37.0.tgz", + "integrity": "sha512-jsuVWeIkb6ggzB+wPCsR4e6loj+rM72ohW6IBn2C+5NCvfUVY8s33iFPySSVXqtm5Hu29Ne/9bnA0JmyLmgenA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@eslint-community/regexpp": "^4.10.0", + "@typescript-eslint/scope-manager": "8.37.0", + "@typescript-eslint/type-utils": "8.37.0", + "@typescript-eslint/utils": "8.37.0", + "@typescript-eslint/visitor-keys": "8.37.0", + "graphemer": "^1.4.0", + "ignore": "^7.0.0", + "natural-compare": "^1.4.0", + "ts-api-utils": "^2.1.0" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/typescript-eslint" + }, + "peerDependencies": { + "@typescript-eslint/parser": "^8.37.0", + "eslint": "^8.57.0 || ^9.0.0", + "typescript": ">=4.8.4 <5.9.0" + } + }, + "node_modules/@typescript-eslint/eslint-plugin/node_modules/ignore": { + "version": "7.0.5", + "resolved": "https://registry.npmjs.org/ignore/-/ignore-7.0.5.tgz", + "integrity": "sha512-Hs59xBNfUIunMFgWAbGX5cq6893IbWg4KnrjbYwX3tx0ztorVgTDA6B2sxf8ejHJ4wz8BqGUMYlnzNBer5NvGg==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">= 4" + } + }, + "node_modules/@typescript-eslint/parser": { + "version": "8.37.0", + "resolved": "https://registry.npmjs.org/@typescript-eslint/parser/-/parser-8.37.0.tgz", + "integrity": "sha512-kVIaQE9vrN9RLCQMQ3iyRlVJpTiDUY6woHGb30JDkfJErqrQEmtdWH3gV0PBAfGZgQXoqzXOO0T3K6ioApbbAA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@typescript-eslint/scope-manager": "8.37.0", + "@typescript-eslint/types": "8.37.0", + "@typescript-eslint/typescript-estree": "8.37.0", + "@typescript-eslint/visitor-keys": "8.37.0", + "debug": "^4.3.4" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/typescript-eslint" + }, + "peerDependencies": { + "eslint": "^8.57.0 || ^9.0.0", + "typescript": ">=4.8.4 <5.9.0" + } + }, + "node_modules/@typescript-eslint/project-service": { + "version": "8.37.0", + "resolved": "https://registry.npmjs.org/@typescript-eslint/project-service/-/project-service-8.37.0.tgz", + "integrity": "sha512-BIUXYsbkl5A1aJDdYJCBAo8rCEbAvdquQ8AnLb6z5Lp1u3x5PNgSSx9A/zqYc++Xnr/0DVpls8iQ2cJs/izTXA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@typescript-eslint/tsconfig-utils": "^8.37.0", + "@typescript-eslint/types": "^8.37.0", + "debug": "^4.3.4" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/typescript-eslint" + }, + "peerDependencies": { + "typescript": ">=4.8.4 <5.9.0" + } + }, + "node_modules/@typescript-eslint/scope-manager": { + "version": "8.37.0", + "resolved": "https://registry.npmjs.org/@typescript-eslint/scope-manager/-/scope-manager-8.37.0.tgz", + "integrity": "sha512-0vGq0yiU1gbjKob2q691ybTg9JX6ShiVXAAfm2jGf3q0hdP6/BruaFjL/ManAR/lj05AvYCH+5bbVo0VtzmjOA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@typescript-eslint/types": "8.37.0", + "@typescript-eslint/visitor-keys": "8.37.0" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/typescript-eslint" + } + }, + "node_modules/@typescript-eslint/tsconfig-utils": { + "version": "8.37.0", + "resolved": "https://registry.npmjs.org/@typescript-eslint/tsconfig-utils/-/tsconfig-utils-8.37.0.tgz", + "integrity": "sha512-1/YHvAVTimMM9mmlPvTec9NP4bobA1RkDbMydxG8omqwJJLEW/Iy2C4adsAESIXU3WGLXFHSZUU+C9EoFWl4Zg==", + "dev": true, + "license": "MIT", + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/typescript-eslint" + }, + "peerDependencies": { + "typescript": ">=4.8.4 <5.9.0" + } + }, + "node_modules/@typescript-eslint/type-utils": { + "version": "8.37.0", + "resolved": "https://registry.npmjs.org/@typescript-eslint/type-utils/-/type-utils-8.37.0.tgz", + "integrity": "sha512-SPkXWIkVZxhgwSwVq9rqj/4VFo7MnWwVaRNznfQDc/xPYHjXnPfLWn+4L6FF1cAz6e7dsqBeMawgl7QjUMj4Ow==", + "dev": true, + "license": "MIT", + "dependencies": { + "@typescript-eslint/types": "8.37.0", + "@typescript-eslint/typescript-estree": "8.37.0", + "@typescript-eslint/utils": "8.37.0", + "debug": "^4.3.4", + "ts-api-utils": "^2.1.0" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/typescript-eslint" + }, + "peerDependencies": { + "eslint": "^8.57.0 || ^9.0.0", + "typescript": ">=4.8.4 <5.9.0" + } + }, + "node_modules/@typescript-eslint/types": { + "version": "8.37.0", + "resolved": "https://registry.npmjs.org/@typescript-eslint/types/-/types-8.37.0.tgz", + "integrity": "sha512-ax0nv7PUF9NOVPs+lmQ7yIE7IQmAf8LGcXbMvHX5Gm+YJUYNAl340XkGnrimxZ0elXyoQJuN5sbg6C4evKA4SQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/typescript-eslint" + } + }, + "node_modules/@typescript-eslint/typescript-estree": { + "version": "8.37.0", + "resolved": "https://registry.npmjs.org/@typescript-eslint/typescript-estree/-/typescript-estree-8.37.0.tgz", + "integrity": "sha512-zuWDMDuzMRbQOM+bHyU4/slw27bAUEcKSKKs3hcv2aNnc/tvE/h7w60dwVw8vnal2Pub6RT1T7BI8tFZ1fE+yg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@typescript-eslint/project-service": "8.37.0", + "@typescript-eslint/tsconfig-utils": "8.37.0", + "@typescript-eslint/types": "8.37.0", + "@typescript-eslint/visitor-keys": "8.37.0", + "debug": "^4.3.4", + "fast-glob": "^3.3.2", + "is-glob": "^4.0.3", + "minimatch": "^9.0.4", + "semver": "^7.6.0", + "ts-api-utils": "^2.1.0" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/typescript-eslint" + }, + "peerDependencies": { + "typescript": ">=4.8.4 <5.9.0" + } + }, + "node_modules/@typescript-eslint/typescript-estree/node_modules/brace-expansion": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-2.0.2.tgz", + "integrity": "sha512-Jt0vHyM+jmUBqojB7E1NIYadt0vI0Qxjxd2TErW94wDz+E2LAm5vKMXXwg6ZZBTHPuUlDgQHKXvjGBdfcF1ZDQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "balanced-match": "^1.0.0" + } + }, + "node_modules/@typescript-eslint/typescript-estree/node_modules/minimatch": { + "version": "9.0.5", + "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-9.0.5.tgz", + "integrity": "sha512-G6T0ZX48xgozx7587koeX9Ys2NYy6Gmv//P89sEte9V9whIapMNF4idKxnW2QtCcLiTWlb/wfCabAtAFWhhBow==", + "dev": true, + "license": "ISC", + "dependencies": { + "brace-expansion": "^2.0.1" + }, + "engines": { + "node": ">=16 || 14 >=14.17" + }, + "funding": { + "url": "https://github.com/sponsors/isaacs" + } + }, + "node_modules/@typescript-eslint/utils": { + "version": "8.37.0", + "resolved": "https://registry.npmjs.org/@typescript-eslint/utils/-/utils-8.37.0.tgz", + "integrity": "sha512-TSFvkIW6gGjN2p6zbXo20FzCABbyUAuq6tBvNRGsKdsSQ6a7rnV6ADfZ7f4iI3lIiXc4F4WWvtUfDw9CJ9pO5A==", + "dev": true, + "license": "MIT", + "dependencies": { + "@eslint-community/eslint-utils": "^4.7.0", + "@typescript-eslint/scope-manager": "8.37.0", + "@typescript-eslint/types": "8.37.0", + "@typescript-eslint/typescript-estree": "8.37.0" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/typescript-eslint" + }, + "peerDependencies": { + "eslint": "^8.57.0 || ^9.0.0", + "typescript": ">=4.8.4 <5.9.0" + } + }, + "node_modules/@typescript-eslint/visitor-keys": { + "version": "8.37.0", + "resolved": "https://registry.npmjs.org/@typescript-eslint/visitor-keys/-/visitor-keys-8.37.0.tgz", + "integrity": "sha512-YzfhzcTnZVPiLfP/oeKtDp2evwvHLMe0LOy7oe+hb9KKIumLNohYS9Hgp1ifwpu42YWxhZE8yieggz6JpqO/1w==", + "dev": true, + "license": "MIT", + "dependencies": { + "@typescript-eslint/types": "8.37.0", + "eslint-visitor-keys": "^4.2.1" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/typescript-eslint" + } + }, + "node_modules/@ungap/structured-clone": { + "version": "1.3.0", + "resolved": "https://registry.npmjs.org/@ungap/structured-clone/-/structured-clone-1.3.0.tgz", + "integrity": "sha512-WmoN8qaIAo7WTYWbAZuG8PYEhn5fkz7dZrqTBZ7dtt//lL2Gwms1IcnQ5yHqjDfX8Ft5j4YzDM23f87zBfDe9g==", + "license": "ISC" + }, + "node_modules/@vitest/browser": { + "version": "3.2.4", + "resolved": "https://registry.npmjs.org/@vitest/browser/-/browser-3.2.4.tgz", + "integrity": "sha512-tJxiPrWmzH8a+w9nLKlQMzAKX/7VjFs50MWgcAj7p9XQ7AQ9/35fByFYptgPELyLw+0aixTnC4pUWV+APcZ/kw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@testing-library/dom": "^10.4.0", + "@testing-library/user-event": "^14.6.1", + "@vitest/mocker": "3.2.4", + "@vitest/utils": "3.2.4", + "magic-string": "^0.30.17", + "sirv": "^3.0.1", + "tinyrainbow": "^2.0.0", + "ws": "^8.18.2" + }, + "funding": { + "url": "https://opencollective.com/vitest" + }, + "peerDependencies": { + "playwright": "*", + "vitest": "3.2.4", + "webdriverio": "^7.0.0 || ^8.0.0 || ^9.0.0" + }, + "peerDependenciesMeta": { + "playwright": { + "optional": true + }, + "safaridriver": { + "optional": true + }, + "webdriverio": { + "optional": true + } + } + }, + "node_modules/@vitest/expect": { + "version": "3.2.4", + "resolved": "https://registry.npmjs.org/@vitest/expect/-/expect-3.2.4.tgz", + "integrity": "sha512-Io0yyORnB6sikFlt8QW5K7slY4OjqNX9jmJQ02QDda8lyM6B5oNgVWoSoKPac8/kgnCUzuHQKrSLtu/uOqqrig==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/chai": "^5.2.2", + "@vitest/spy": "3.2.4", + "@vitest/utils": "3.2.4", + "chai": "^5.2.0", + "tinyrainbow": "^2.0.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/@vitest/mocker": { + "version": "3.2.4", + "resolved": "https://registry.npmjs.org/@vitest/mocker/-/mocker-3.2.4.tgz", + "integrity": "sha512-46ryTE9RZO/rfDd7pEqFl7etuyzekzEhUbTW3BvmeO/BcCMEgq59BKhek3dXDWgAj4oMK6OZi+vRr1wPW6qjEQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "@vitest/spy": "3.2.4", + "estree-walker": "^3.0.3", + "magic-string": "^0.30.17" + }, + "funding": { + "url": "https://opencollective.com/vitest" + }, + "peerDependencies": { + "msw": "^2.4.9", + "vite": "^5.0.0 || ^6.0.0 || ^7.0.0-0" + }, + "peerDependenciesMeta": { + "msw": { + "optional": true + }, + "vite": { + "optional": true + } + } + }, + "node_modules/@vitest/mocker/node_modules/estree-walker": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/estree-walker/-/estree-walker-3.0.3.tgz", + "integrity": "sha512-7RUKfXgSMMkzt6ZuXmqapOurLGPPfgj6l9uRZ7lRGolvk0y2yocc35LdcxKC5PQZdn2DMqioAQ2NoWcrTKmm6g==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/estree": "^1.0.0" + } + }, + "node_modules/@vitest/pretty-format": { + "version": "3.2.4", + "resolved": "https://registry.npmjs.org/@vitest/pretty-format/-/pretty-format-3.2.4.tgz", + "integrity": "sha512-IVNZik8IVRJRTr9fxlitMKeJeXFFFN0JaB9PHPGQ8NKQbGpfjlTx9zO4RefN8gp7eqjNy8nyK3NZmBzOPeIxtA==", + "dev": true, + "license": "MIT", + "dependencies": { + "tinyrainbow": "^2.0.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/@vitest/runner": { + "version": "3.2.4", + "resolved": "https://registry.npmjs.org/@vitest/runner/-/runner-3.2.4.tgz", + "integrity": "sha512-oukfKT9Mk41LreEW09vt45f8wx7DordoWUZMYdY/cyAk7w5TWkTRCNZYF7sX7n2wB7jyGAl74OxgwhPgKaqDMQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "@vitest/utils": "3.2.4", + "pathe": "^2.0.3", + "strip-literal": "^3.0.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/@vitest/snapshot": { + "version": "3.2.4", + "resolved": "https://registry.npmjs.org/@vitest/snapshot/-/snapshot-3.2.4.tgz", + "integrity": "sha512-dEYtS7qQP2CjU27QBC5oUOxLE/v5eLkGqPE0ZKEIDGMs4vKWe7IjgLOeauHsR0D5YuuycGRO5oSRXnwnmA78fQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "@vitest/pretty-format": "3.2.4", + "magic-string": "^0.30.17", + "pathe": "^2.0.3" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/@vitest/spy": { + "version": "3.2.4", + "resolved": "https://registry.npmjs.org/@vitest/spy/-/spy-3.2.4.tgz", + "integrity": "sha512-vAfasCOe6AIK70iP5UD11Ac4siNUNJ9i/9PZ3NKx07sG6sUxeag1LWdNrMWeKKYBLlzuK+Gn65Yd5nyL6ds+nw==", + "dev": true, + "license": "MIT", + "dependencies": { + "tinyspy": "^4.0.3" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/@vitest/utils": { + "version": "3.2.4", + "resolved": "https://registry.npmjs.org/@vitest/utils/-/utils-3.2.4.tgz", + "integrity": "sha512-fB2V0JFrQSMsCo9HiSq3Ezpdv4iYaXRG1Sx8edX3MwxfyNn83mKiGzOcH+Fkxt4MHxr3y42fQi1oeAInqgX2QA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@vitest/pretty-format": "3.2.4", + "loupe": "^3.1.4", + "tinyrainbow": "^2.0.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/acorn": { + "version": "8.15.0", + "resolved": "https://registry.npmjs.org/acorn/-/acorn-8.15.0.tgz", + "integrity": "sha512-NZyJarBfL7nWwIq+FDL6Zp/yHEhePMNnnJ0y3qfieCrmNvYct8uvtiV41UvlSe6apAfk0fY1FbWx+NwfmpvtTg==", + "license": "MIT", + "bin": { + "acorn": "bin/acorn" + }, + "engines": { + "node": ">=0.4.0" + } + }, + "node_modules/acorn-jsx": { + "version": "5.3.2", + "resolved": "https://registry.npmjs.org/acorn-jsx/-/acorn-jsx-5.3.2.tgz", + "integrity": "sha512-rq9s+JNhf0IChjtDXxllJ7g41oZk5SlXtp0LHwyA5cejwn7vKmKp4pPri6YEePv2PU65sAsegbXtIinmDFDXgQ==", + "dev": true, + "license": "MIT", + "peerDependencies": { + "acorn": "^6.0.0 || ^7.0.0 || ^8.0.0" + } + }, + "node_modules/ajv": { + "version": "6.12.6", + "resolved": "https://registry.npmjs.org/ajv/-/ajv-6.12.6.tgz", + "integrity": "sha512-j3fVLgvTo527anyYyJOGTYJbG+vnnQYvE0m5mmkc1TK+nxAppkCLMIL0aZ4dblVCNoGShhm+kzE4ZUykBoMg4g==", + "dev": true, + "license": "MIT", + "dependencies": { + "fast-deep-equal": "^3.1.1", + "fast-json-stable-stringify": "^2.0.0", + "json-schema-traverse": "^0.4.1", + "uri-js": "^4.2.2" + }, + "funding": { + "type": "github", + "url": "https://github.com/sponsors/epoberezkin" + } + }, + "node_modules/ansi-regex": { + "version": "5.0.1", + "resolved": "https://registry.npmjs.org/ansi-regex/-/ansi-regex-5.0.1.tgz", + "integrity": "sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/ansi-styles": { + "version": "4.3.0", + "resolved": "https://registry.npmjs.org/ansi-styles/-/ansi-styles-4.3.0.tgz", + "integrity": "sha512-zbB9rCJAT1rbjiVDb2hqKFHNYLxgtk8NURxZ3IZwD3F6NtxbXZQCnnSi1Lkx+IDohdPlFp222wVALIheZJQSEg==", + "dev": true, + "license": "MIT", + "dependencies": { + "color-convert": "^2.0.1" + }, + "engines": { + "node": ">=8" + }, + "funding": { + "url": "https://github.com/chalk/ansi-styles?sponsor=1" + } + }, + "node_modules/argparse": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/argparse/-/argparse-2.0.1.tgz", + "integrity": "sha512-8+9WqebbFzpX9OR+Wa6O29asIogeRMzcGtAINdpMHHyAg10f05aSFVBbcEqGf/PXw1EjAZ+q2/bEBg3DvurK3Q==", + "dev": true, + "license": "Python-2.0" + }, + "node_modules/aria-query": { + "version": "5.3.0", + "resolved": "https://registry.npmjs.org/aria-query/-/aria-query-5.3.0.tgz", + "integrity": "sha512-b0P0sZPKtyu8HkeRAfCq0IfURZK+SuwMjY1UXGBU27wpAiTwQAIlq56IbIO+ytk/JjS1fMR14ee5WBBfKi5J6A==", + "dev": true, + "license": "Apache-2.0", + "dependencies": { + "dequal": "^2.0.3" + } + }, + "node_modules/assertion-error": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/assertion-error/-/assertion-error-2.0.1.tgz", + "integrity": "sha512-Izi8RQcffqCeNVgFigKli1ssklIbpHnCYc6AknXGYoB6grJqyeby7jv12JUQgmTAnIDnbck1uxksT4dzN3PWBA==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=12" + } + }, + "node_modules/ast-types": { + "version": "0.16.1", + "resolved": "https://registry.npmjs.org/ast-types/-/ast-types-0.16.1.tgz", + "integrity": "sha512-6t10qk83GOG8p0vKmaCr8eiilZwO171AvbROMtvvNiwrTly62t+7XkA8RdIIVbpMhCASAsxgAzdRSwh6nw/5Dg==", + "dev": true, + "license": "MIT", + "dependencies": { + "tslib": "^2.0.1" + }, + "engines": { + "node": ">=4" + } + }, + "node_modules/axe-core": { + "version": "4.10.3", + "resolved": "https://registry.npmjs.org/axe-core/-/axe-core-4.10.3.tgz", + "integrity": "sha512-Xm7bpRXnDSX2YE2YFfBk2FnF0ep6tmG7xPh8iHee8MIcrgq762Nkce856dYtJYLkuIoYZvGfTs/PbZhideTcEg==", + "dev": true, + "license": "MPL-2.0", + "engines": { + "node": ">=4" + } + }, + "node_modules/axobject-query": { + "version": "4.1.0", + "resolved": "https://registry.npmjs.org/axobject-query/-/axobject-query-4.1.0.tgz", + "integrity": "sha512-qIj0G9wZbMGNLjLmg1PT6v2mE9AH2zlnADJD/2tC6E00hgmhUOfEB6greHPAfLRSufHqROIUTkw6E+M3lH0PTQ==", + "license": "Apache-2.0", + "engines": { + "node": ">= 0.4" + } + }, + "node_modules/bail": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/bail/-/bail-2.0.2.tgz", + "integrity": "sha512-0xO6mYd7JB2YesxDKplafRpsiOzPt9V02ddPCLbY1xYGPOX24NTyN50qnUxgCPcSoYMhKpAuBTjQoRZCAkUDRw==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/balanced-match": { + "version": "1.0.2", + "resolved": "https://registry.npmjs.org/balanced-match/-/balanced-match-1.0.2.tgz", + "integrity": "sha512-3oSeUO0TMV67hN1AmbXsK4yaqU7tjiHlbxRDZOpH0KW9+CeX4bRAaX0Anxt0tx2MrpRpWwQaPwIlISEJhYU5Pw==", + "dev": true, + "license": "MIT" + }, + "node_modules/better-opn": { + "version": "3.0.2", + "resolved": "https://registry.npmjs.org/better-opn/-/better-opn-3.0.2.tgz", + "integrity": "sha512-aVNobHnJqLiUelTaHat9DZ1qM2w0C0Eym4LPI/3JxOnSokGVdsl1T1kN7TFvsEAD8G47A6VKQ0TVHqbBnYMJlQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "open": "^8.0.4" + }, + "engines": { + "node": ">=12.0.0" + } + }, + "node_modules/bits-ui": { + "version": "2.8.11", + "resolved": "https://registry.npmjs.org/bits-ui/-/bits-ui-2.8.11.tgz", + "integrity": "sha512-lKN9rAk69my6j7H1D4B87r8LrHuEtfEsf1xCixBj9yViql2BdI3f04HyyyT7T1GOCpgb9+8b0B+nm3LN81Konw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@floating-ui/core": "^1.7.1", + "@floating-ui/dom": "^1.7.1", + "esm-env": "^1.1.2", + "runed": "^0.29.1", + "svelte-toolbelt": "^0.9.3", + "tabbable": "^6.2.0" + }, + "engines": { + "node": ">=20" + }, + "funding": { + "url": "https://github.com/sponsors/huntabyte" + }, + "peerDependencies": { + "@internationalized/date": "^3.8.1", + "svelte": "^5.33.0" + } + }, + "node_modules/bits-ui/node_modules/runed": { + "version": "0.29.2", + "resolved": "https://registry.npmjs.org/runed/-/runed-0.29.2.tgz", + "integrity": "sha512-0cq6cA6sYGZwl/FvVqjx9YN+1xEBu9sDDyuWdDW1yWX7JF2wmvmVKfH+hVCZs+csW+P3ARH92MjI3H9QTagOQA==", + "dev": true, + "funding": [ + "https://github.com/sponsors/huntabyte", + "https://github.com/sponsors/tglide" + ], + "license": "MIT", + "dependencies": { + "esm-env": "^1.0.0" + }, + "peerDependencies": { + "svelte": "^5.7.0" + } + }, + "node_modules/bits-ui/node_modules/svelte-toolbelt": { + "version": "0.9.3", + "resolved": "https://registry.npmjs.org/svelte-toolbelt/-/svelte-toolbelt-0.9.3.tgz", + "integrity": "sha512-HCSWxCtVmv+c6g1ACb8LTwHVbDqLKJvHpo6J8TaqwUme2hj9ATJCpjCPNISR1OCq2Q4U1KT41if9ON0isINQZw==", + "dev": true, + "funding": [ + "https://github.com/sponsors/huntabyte" + ], + "dependencies": { + "clsx": "^2.1.1", + "runed": "^0.29.0", + "style-to-object": "^1.0.8" + }, + "engines": { + "node": ">=18", + "pnpm": ">=8.7.0" + }, + "peerDependencies": { + "svelte": "^5.30.2" + } + }, + "node_modules/brace-expansion": { + "version": "1.1.12", + "resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-1.1.12.tgz", + "integrity": "sha512-9T9UjW3r0UW5c1Q7GTwllptXwhvYmEzFhzMfZ9H7FQWt+uZePjZPjBP/W1ZEyZ1twGWom5/56TF4lPcqjnDHcg==", + "dev": true, + "license": "MIT", + "dependencies": { + "balanced-match": "^1.0.0", + "concat-map": "0.0.1" + } + }, + "node_modules/braces": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/braces/-/braces-3.0.3.tgz", + "integrity": "sha512-yQbXgO/OSZVD2IsiLlro+7Hf6Q18EJrKSEsdoMzKePKXct3gvD8oLcOQdIzGupr5Fj+EDe8gO/lxc1BzfMpxvA==", + "dev": true, + "license": "MIT", + "dependencies": { + "fill-range": "^7.1.1" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/cac": { + "version": "6.7.14", + "resolved": "https://registry.npmjs.org/cac/-/cac-6.7.14.tgz", + "integrity": "sha512-b6Ilus+c3RrdDk+JhLKUAQfzzgLEPy6wcXqS7f/xe1EETvsDP6GORG7SFuOs6cID5YkqchW/LXZbX5bc8j7ZcQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/callsites": { + "version": "3.1.0", + "resolved": "https://registry.npmjs.org/callsites/-/callsites-3.1.0.tgz", + "integrity": "sha512-P8BjAsXvZS+VIDUI11hHCQEv74YT67YUi5JJFNWIqL235sBmjX4+qx9Muvls5ivyNENctx46xQLQ3aTuE7ssaQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=6" + } + }, + "node_modules/ccount": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/ccount/-/ccount-2.0.1.tgz", + "integrity": "sha512-eyrF0jiFpY+3drT6383f1qhkbGsLSifNAjA61IUjZjmLCWjItY6LB9ft9YhoDgwfmclB2zhu51Lc7+95b8NRAg==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/chai": { + "version": "5.2.1", + "resolved": "https://registry.npmjs.org/chai/-/chai-5.2.1.tgz", + "integrity": "sha512-5nFxhUrX0PqtyogoYOA8IPswy5sZFTOsBFl/9bNsmDLgsxYTzSZQJDPppDnZPTQbzSEm0hqGjWPzRemQCYbD6A==", + "dev": true, + "license": "MIT", + "dependencies": { + "assertion-error": "^2.0.1", + "check-error": "^2.1.1", + "deep-eql": "^5.0.1", + "loupe": "^3.1.0", + "pathval": "^2.0.0" + }, + "engines": { + "node": ">=18" + } + }, + "node_modules/chalk": { + "version": "4.1.2", + "resolved": "https://registry.npmjs.org/chalk/-/chalk-4.1.2.tgz", + "integrity": "sha512-oKnbhFyRIXpUuez8iBMmyEa4nbj4IOQyuhc/wy9kY7/WVPcwIO9VA668Pu8RkO7+0G76SLROeyw9CpQ061i4mA==", + "dev": true, + "license": "MIT", + "dependencies": { + "ansi-styles": "^4.1.0", + "supports-color": "^7.1.0" + }, + "engines": { + "node": ">=10" + }, + "funding": { + "url": "https://github.com/chalk/chalk?sponsor=1" + } + }, + "node_modules/character-entities": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/character-entities/-/character-entities-2.0.2.tgz", + "integrity": "sha512-shx7oQ0Awen/BRIdkjkvz54PnEEI/EjwXDSIZp86/KKdbafHh1Df/RYGBhn4hbe2+uKC9FnT5UCEdyPz3ai9hQ==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/character-entities-html4": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/character-entities-html4/-/character-entities-html4-2.1.0.tgz", + "integrity": "sha512-1v7fgQRj6hnSwFpq1Eu0ynr/CDEw0rXo2B61qXrLNdHZmPKgb7fqS1a2JwF0rISo9q77jDI8VMEHoApn8qDoZA==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/character-entities-legacy": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/character-entities-legacy/-/character-entities-legacy-3.0.0.tgz", + "integrity": "sha512-RpPp0asT/6ufRm//AJVwpViZbGM/MkjQFxJccQRHmISF/22NBtsHqAWmL+/pmkPWoIUJdWyeVleTl1wydHATVQ==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/check-error": { + "version": "2.1.1", + "resolved": "https://registry.npmjs.org/check-error/-/check-error-2.1.1.tgz", + "integrity": "sha512-OAlb+T7V4Op9OwdkjmguYRqncdlx5JiofwOAUkmTF+jNdHwzTaTs4sRAGpzLF3oOz5xAyDGrPgeIDFQmDOTiJw==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">= 16" + } + }, + "node_modules/chokidar": { + "version": "4.0.3", + "resolved": "https://registry.npmjs.org/chokidar/-/chokidar-4.0.3.tgz", + "integrity": "sha512-Qgzu8kfBvo+cA4962jnP1KkS6Dop5NS6g7R5LFYJr4b8Ub94PPQXUksCw9PvXoeXPRRddRNC5C1JQUR2SMGtnA==", + "dev": true, + "license": "MIT", + "dependencies": { + "readdirp": "^4.0.1" + }, + "engines": { + "node": ">= 14.16.0" + }, + "funding": { + "url": "https://paulmillr.com/funding/" + } + }, + "node_modules/chownr": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/chownr/-/chownr-3.0.0.tgz", + "integrity": "sha512-+IxzY9BZOQd/XuYPRmrvEVjF/nqj5kgT4kEq7VofrDoM1MxoRjEWkrCC3EtLi59TVawxTAn+orJwFQcrqEN1+g==", + "dev": true, + "license": "BlueOak-1.0.0", + "engines": { + "node": ">=18" + } + }, + "node_modules/chromatic": { + "version": "12.2.0", + "resolved": "https://registry.npmjs.org/chromatic/-/chromatic-12.2.0.tgz", + "integrity": "sha512-GswmBW9ZptAoTns1BMyjbm55Z7EsIJnUvYKdQqXIBZIKbGErmpA+p4c0BYA+nzw5B0M+rb3Iqp1IaH8TFwIQew==", + "dev": true, + "license": "MIT", + "bin": { + "chroma": "dist/bin.js", + "chromatic": "dist/bin.js", + "chromatic-cli": "dist/bin.js" + }, + "peerDependencies": { + "@chromatic-com/cypress": "^0.*.* || ^1.0.0", + "@chromatic-com/playwright": "^0.*.* || ^1.0.0" + }, + "peerDependenciesMeta": { + "@chromatic-com/cypress": { + "optional": true + }, + "@chromatic-com/playwright": { + "optional": true + } + } + }, + "node_modules/clsx": { + "version": "2.1.1", + "resolved": "https://registry.npmjs.org/clsx/-/clsx-2.1.1.tgz", + "integrity": "sha512-eYm0QWBtUrBWZWG0d386OGAw16Z995PiOVo2B7bjWSbHedGl5e0ZWaq65kOGgUSNesEIDkB9ISbTg/JK9dhCZA==", + "license": "MIT", + "engines": { + "node": ">=6" + } + }, + "node_modules/color-convert": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/color-convert/-/color-convert-2.0.1.tgz", + "integrity": "sha512-RRECPsj7iu/xb5oKYcsFHSppFNnsj/52OVTRKb4zP5onXwVF3zVmmToNcOfGC+CRDpfK/U584fMg38ZHCaElKQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "color-name": "~1.1.4" + }, + "engines": { + "node": ">=7.0.0" + } + }, + "node_modules/color-name": { + "version": "1.1.4", + "resolved": "https://registry.npmjs.org/color-name/-/color-name-1.1.4.tgz", + "integrity": "sha512-dOy+3AuW3a2wNbZHIuMZpTcgjGuLU/uBL/ubcZF9OXbDo8ff4O8yVp5Bf0efS8uEoYo5q4Fx7dY9OgQGXgAsQA==", + "dev": true, + "license": "MIT" + }, + "node_modules/comma-separated-tokens": { + "version": "2.0.3", + "resolved": "https://registry.npmjs.org/comma-separated-tokens/-/comma-separated-tokens-2.0.3.tgz", + "integrity": "sha512-Fu4hJdvzeylCfQPp9SGWidpzrMs7tTrlu6Vb8XGaRGck8QSNZJJp538Wrb60Lax4fPwR64ViY468OIUTbRlGZg==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/commander": { + "version": "8.3.0", + "resolved": "https://registry.npmjs.org/commander/-/commander-8.3.0.tgz", + "integrity": "sha512-OkTL9umf+He2DZkUq8f8J9of7yL6RJKI24dVITBmNfZBmri9zYZQrKkuXiKhyfPSu8tUhnVBB1iKXevvnlR4Ww==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">= 12" + } + }, + "node_modules/concat-map": { + "version": "0.0.1", + "resolved": "https://registry.npmjs.org/concat-map/-/concat-map-0.0.1.tgz", + "integrity": "sha512-/Srv4dswyQNBfohGpz9o6Yb3Gz3SrUDqBH5rTuhGR7ahtlbYKnVxw2bCFMRljaA7EXHaXZ8wsHdodFvbkhKmqg==", + "dev": true, + "license": "MIT" + }, + "node_modules/cookie": { + "version": "0.6.0", + "resolved": "https://registry.npmjs.org/cookie/-/cookie-0.6.0.tgz", + "integrity": "sha512-U71cyTamuh1CRNCfpGY6to28lxvNwPG4Guz/EVjgf3Jmzv0vlDp1atT9eS5dDjMYHucpHbWns6Lwf3BKz6svdw==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">= 0.6" + } + }, + "node_modules/cross-spawn": { + "version": "7.0.6", + "resolved": "https://registry.npmjs.org/cross-spawn/-/cross-spawn-7.0.6.tgz", + "integrity": "sha512-uV2QOWP2nWzsy2aMp8aRibhi9dlzF5Hgh5SHaB9OiTGEyDTiJJyx0uy51QXdyWbtAHNua4XJzUKca3OzKUd3vA==", + "dev": true, + "license": "MIT", + "dependencies": { + "path-key": "^3.1.0", + "shebang-command": "^2.0.0", + "which": "^2.0.1" + }, + "engines": { + "node": ">= 8" + } + }, + "node_modules/css.escape": { + "version": "1.5.1", + "resolved": "https://registry.npmjs.org/css.escape/-/css.escape-1.5.1.tgz", + "integrity": "sha512-YUifsXXuknHlUsmlgyY0PKzgPOr7/FjCePfHNt0jxm83wHZi44VDMQ7/fGNkjY3/jV1MC+1CmZbaHzugyeRtpg==", + "dev": true, + "license": "MIT" + }, + "node_modules/cssesc": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/cssesc/-/cssesc-3.0.0.tgz", + "integrity": "sha512-/Tb/JcjK111nNScGob5MNtsntNM1aCNUDipB/TkwZFhyDrrE47SOx/18wF2bbjgc3ZzCSKW1T5nt5EbFoAz/Vg==", + "dev": true, + "license": "MIT", + "bin": { + "cssesc": "bin/cssesc" + }, + "engines": { + "node": ">=4" + } + }, + "node_modules/csstype": { + "version": "3.1.3", + "resolved": "https://registry.npmjs.org/csstype/-/csstype-3.1.3.tgz", + "integrity": "sha512-M1uQkMl8rQK/szD0LNhtqxIPLpimGm8sOBwU7lLnCpSbTyY3yeU1Vc7l4KT5zT4s/yOxHH5O7tIuuLOCnLADRw==", + "dev": true, + "license": "MIT", + "peer": true + }, + "node_modules/debug": { + "version": "4.4.1", + "resolved": "https://registry.npmjs.org/debug/-/debug-4.4.1.tgz", + "integrity": "sha512-KcKCqiftBJcZr++7ykoDIEwSa3XWowTfNPo92BYxjXiyYEVrUQh2aLyhxBCwww+heortUFxEJYcRzosstTEBYQ==", + "license": "MIT", + "dependencies": { + "ms": "^2.1.3" + }, + "engines": { + "node": ">=6.0" + }, + "peerDependenciesMeta": { + "supports-color": { + "optional": true + } + } + }, + "node_modules/decode-named-character-reference": { + "version": "1.2.0", + "resolved": "https://registry.npmjs.org/decode-named-character-reference/-/decode-named-character-reference-1.2.0.tgz", + "integrity": "sha512-c6fcElNV6ShtZXmsgNgFFV5tVX2PaV4g+MOAkb8eXHvn6sryJBrZa9r0zV6+dtTyoCKxtDy5tyQ5ZwQuidtd+Q==", + "license": "MIT", + "dependencies": { + "character-entities": "^2.0.0" + }, + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/dedent": { + "version": "1.6.0", + "resolved": "https://registry.npmjs.org/dedent/-/dedent-1.6.0.tgz", + "integrity": "sha512-F1Z+5UCFpmQUzJa11agbyPVMbpgT/qA3/SKyJ1jyBgm7dUcUEa8v9JwDkerSQXfakBwFljIxhOJqGkjUwZ9FSA==", + "dev": true, + "license": "MIT", + "peerDependencies": { + "babel-plugin-macros": "^3.1.0" + }, + "peerDependenciesMeta": { + "babel-plugin-macros": { + "optional": true + } + } + }, + "node_modules/dedent-js": { + "version": "1.0.1", + "resolved": "https://registry.npmjs.org/dedent-js/-/dedent-js-1.0.1.tgz", + "integrity": "sha512-OUepMozQULMLUmhxS95Vudo0jb0UchLimi3+pQ2plj61Fcy8axbP9hbiD4Sz6DPqn6XG3kfmziVfQ1rSys5AJQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/deep-eql": { + "version": "5.0.2", + "resolved": "https://registry.npmjs.org/deep-eql/-/deep-eql-5.0.2.tgz", + "integrity": "sha512-h5k/5U50IJJFpzfL6nO9jaaumfjO/f2NjK/oYB2Djzm4p9L+3T9qWpZqZ2hAbLPuuYq9wrU08WQyBTL5GbPk5Q==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=6" + } + }, + "node_modules/deep-is": { + "version": "0.1.4", + "resolved": "https://registry.npmjs.org/deep-is/-/deep-is-0.1.4.tgz", + "integrity": "sha512-oIPzksmTg4/MriiaYGO+okXDT7ztn/w3Eptv/+gSIdMdKsJo0u4CfYNFJPy+4SKMuCqGw2wxnA+URMg3t8a/bQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/deepmerge": { + "version": "4.3.1", + "resolved": "https://registry.npmjs.org/deepmerge/-/deepmerge-4.3.1.tgz", + "integrity": "sha512-3sUqbMEc77XqpdNO7FRyRog+eW3ph+GYCbj+rK+uYyRMuwsVy0rMiVtPn+QJlKFvWP/1PYpapqYn0Me2knFn+A==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/define-lazy-prop": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/define-lazy-prop/-/define-lazy-prop-2.0.0.tgz", + "integrity": "sha512-Ds09qNh8yw3khSjiJjiUInaGX9xlqZDY7JVryGxdxV7NPeuqQfplOpQ66yJFZut3jLa5zOwkXw1g9EI2uKh4Og==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/dequal": { + "version": "2.0.3", + "resolved": "https://registry.npmjs.org/dequal/-/dequal-2.0.3.tgz", + "integrity": "sha512-0je+qPKHEMohvfRTCEo3CrPG6cAzAYgmzKyxRiYSSDkS6eGJdyVJm7WaYA5ECaAD9wLB2T4EEeymA5aFVcYXCA==", + "license": "MIT", + "engines": { + "node": ">=6" + } + }, + "node_modules/detect-libc": { + "version": "2.0.4", + "resolved": "https://registry.npmjs.org/detect-libc/-/detect-libc-2.0.4.tgz", + "integrity": "sha512-3UDv+G9CsCKO1WKMGw9fwq/SWJYbI0c5Y7LU1AXYoDdbhE2AHQ6N6Nb34sG8Fj7T5APy8qXDCKuuIHd1BR0tVA==", + "dev": true, + "license": "Apache-2.0", + "engines": { + "node": ">=8" + } + }, + "node_modules/devalue": { + "version": "5.3.2", + "resolved": "https://registry.npmjs.org/devalue/-/devalue-5.3.2.tgz", + "integrity": "sha512-UDsjUbpQn9kvm68slnrs+mfxwFkIflOhkanmyabZ8zOYk8SMEIbJ3TK+88g70hSIeytu4y18f0z/hYHMTrXIWw==", + "dev": true, + "license": "MIT" + }, + "node_modules/devlop": { + "version": "1.1.0", + "resolved": "https://registry.npmjs.org/devlop/-/devlop-1.1.0.tgz", + "integrity": "sha512-RWmIqhcFf1lRYBvNmr7qTNuyCt/7/ns2jbpp1+PalgE/rDQcBT0fioSMUpJ93irlUhC5hrg4cYqe6U+0ImW0rA==", + "license": "MIT", + "dependencies": { + "dequal": "^2.0.0" + }, + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/dexie": { + "version": "4.0.11", + "resolved": "https://registry.npmjs.org/dexie/-/dexie-4.0.11.tgz", + "integrity": "sha512-SOKO002EqlvBYYKQSew3iymBoN2EQ4BDw/3yprjh7kAfFzjBYkaMNa/pZvcA7HSWlcKSQb9XhPe3wKyQ0x4A8A==", + "dev": true, + "license": "Apache-2.0" + }, + "node_modules/dom-accessibility-api": { + "version": "0.5.16", + "resolved": "https://registry.npmjs.org/dom-accessibility-api/-/dom-accessibility-api-0.5.16.tgz", + "integrity": "sha512-X7BJ2yElsnOJ30pZF4uIIDfBEVgF4XEBxL9Bxhy6dnrm5hkzqmsWHGTiHqRiITNhMyFLyAiWndIJP7Z1NTteDg==", + "dev": true, + "license": "MIT" + }, + "node_modules/enhanced-resolve": { + "version": "5.18.2", + "resolved": "https://registry.npmjs.org/enhanced-resolve/-/enhanced-resolve-5.18.2.tgz", + "integrity": "sha512-6Jw4sE1maoRJo3q8MsSIn2onJFbLTOjY9hlx4DZXmOKvLRd1Ok2kXmAGXaafL2+ijsJZ1ClYbl/pmqr9+k4iUQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "graceful-fs": "^4.2.4", + "tapable": "^2.2.0" + }, + "engines": { + "node": ">=10.13.0" + } + }, + "node_modules/entities": { + "version": "6.0.1", + "resolved": "https://registry.npmjs.org/entities/-/entities-6.0.1.tgz", + "integrity": "sha512-aN97NXWF6AWBTahfVOIrB/NShkzi5H7F9r1s9mD3cDj4Ko5f2qhhVoYMibXF7GlLveb/D2ioWay8lxI97Ven3g==", + "dev": true, + "license": "BSD-2-Clause", + "engines": { + "node": ">=0.12" + }, + "funding": { + "url": "https://github.com/fb55/entities?sponsor=1" + } + }, + "node_modules/es-module-lexer": { + "version": "1.7.0", + "resolved": "https://registry.npmjs.org/es-module-lexer/-/es-module-lexer-1.7.0.tgz", + "integrity": "sha512-jEQoCwk8hyb2AZziIOLhDqpm5+2ww5uIE6lkO/6jcOCusfk6LhMHpXXfBLXTZ7Ydyt0j4VoUQv6uGNYbdW+kBA==", + "dev": true, + "license": "MIT" + }, + "node_modules/es-toolkit": { + "version": "1.39.7", + "resolved": "https://registry.npmjs.org/es-toolkit/-/es-toolkit-1.39.7.tgz", + "integrity": "sha512-ek/wWryKouBrZIjkwW2BFf91CWOIMvoy2AE5YYgUrfWsJQM2Su1LoLtrw8uusEpN9RfqLlV/0FVNjT0WMv8Bxw==", + "dev": true, + "license": "MIT", + "workspaces": [ + "docs", + "benchmarks" + ] + }, + "node_modules/esbuild": { + "version": "0.25.8", + "resolved": "https://registry.npmjs.org/esbuild/-/esbuild-0.25.8.tgz", + "integrity": "sha512-vVC0USHGtMi8+R4Kz8rt6JhEWLxsv9Rnu/lGYbPR8u47B+DCBksq9JarW0zOO7bs37hyOK1l2/oqtbciutL5+Q==", + "dev": true, + "hasInstallScript": true, + "license": "MIT", + "bin": { + "esbuild": "bin/esbuild" + }, + "engines": { + "node": ">=18" + }, + "optionalDependencies": { + "@esbuild/aix-ppc64": "0.25.8", + "@esbuild/android-arm": "0.25.8", + "@esbuild/android-arm64": "0.25.8", + "@esbuild/android-x64": "0.25.8", + "@esbuild/darwin-arm64": "0.25.8", + "@esbuild/darwin-x64": "0.25.8", + "@esbuild/freebsd-arm64": "0.25.8", + "@esbuild/freebsd-x64": "0.25.8", + "@esbuild/linux-arm": "0.25.8", + "@esbuild/linux-arm64": "0.25.8", + "@esbuild/linux-ia32": "0.25.8", + "@esbuild/linux-loong64": "0.25.8", + "@esbuild/linux-mips64el": "0.25.8", + "@esbuild/linux-ppc64": "0.25.8", + "@esbuild/linux-riscv64": "0.25.8", + "@esbuild/linux-s390x": "0.25.8", + "@esbuild/linux-x64": "0.25.8", + "@esbuild/netbsd-arm64": "0.25.8", + "@esbuild/netbsd-x64": "0.25.8", + "@esbuild/openbsd-arm64": "0.25.8", + "@esbuild/openbsd-x64": "0.25.8", + "@esbuild/openharmony-arm64": "0.25.8", + "@esbuild/sunos-x64": "0.25.8", + "@esbuild/win32-arm64": "0.25.8", + "@esbuild/win32-ia32": "0.25.8", + "@esbuild/win32-x64": "0.25.8" + } + }, + "node_modules/esbuild-register": { + "version": "3.6.0", + "resolved": "https://registry.npmjs.org/esbuild-register/-/esbuild-register-3.6.0.tgz", + "integrity": "sha512-H2/S7Pm8a9CL1uhp9OvjwrBh5Pvx0H8qVOxNu8Wed9Y7qv56MPtq+GGM8RJpq6glYJn9Wspr8uw7l55uyinNeg==", + "dev": true, + "license": "MIT", + "dependencies": { + "debug": "^4.3.4" + }, + "peerDependencies": { + "esbuild": ">=0.12 <1" + } + }, + "node_modules/escape-string-regexp": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/escape-string-regexp/-/escape-string-regexp-4.0.0.tgz", + "integrity": "sha512-TtpcNJ3XAzx3Gq8sWRzJaVajRs0uVxA2YAkdb1jm2YkPz4G6egUFAyA3n5vtEIZefPk5Wa4UXbKuS5fKkJWdgA==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=10" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/eslint": { + "version": "9.31.0", + "resolved": "https://registry.npmjs.org/eslint/-/eslint-9.31.0.tgz", + "integrity": "sha512-QldCVh/ztyKJJZLr4jXNUByx3gR+TDYZCRXEktiZoUR3PGy4qCmSbkxcIle8GEwGpb5JBZazlaJ/CxLidXdEbQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "@eslint-community/eslint-utils": "^4.2.0", + "@eslint-community/regexpp": "^4.12.1", + "@eslint/config-array": "^0.21.0", + "@eslint/config-helpers": "^0.3.0", + "@eslint/core": "^0.15.0", + "@eslint/eslintrc": "^3.3.1", + "@eslint/js": "9.31.0", + "@eslint/plugin-kit": "^0.3.1", + "@humanfs/node": "^0.16.6", + "@humanwhocodes/module-importer": "^1.0.1", + "@humanwhocodes/retry": "^0.4.2", + "@types/estree": "^1.0.6", + "@types/json-schema": "^7.0.15", + "ajv": "^6.12.4", + "chalk": "^4.0.0", + "cross-spawn": "^7.0.6", + "debug": "^4.3.2", + "escape-string-regexp": "^4.0.0", + "eslint-scope": "^8.4.0", + "eslint-visitor-keys": "^4.2.1", + "espree": "^10.4.0", + "esquery": "^1.5.0", + "esutils": "^2.0.2", + "fast-deep-equal": "^3.1.3", + "file-entry-cache": "^8.0.0", + "find-up": "^5.0.0", + "glob-parent": "^6.0.2", + "ignore": "^5.2.0", + "imurmurhash": "^0.1.4", + "is-glob": "^4.0.0", + "json-stable-stringify-without-jsonify": "^1.0.1", + "lodash.merge": "^4.6.2", + "minimatch": "^3.1.2", + "natural-compare": "^1.4.0", + "optionator": "^0.9.3" + }, + "bin": { + "eslint": "bin/eslint.js" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "url": "https://eslint.org/donate" + }, + "peerDependencies": { + "jiti": "*" + }, + "peerDependenciesMeta": { + "jiti": { + "optional": true + } + } + }, + "node_modules/eslint-config-prettier": { + "version": "10.1.8", + "resolved": "https://registry.npmjs.org/eslint-config-prettier/-/eslint-config-prettier-10.1.8.tgz", + "integrity": "sha512-82GZUjRS0p/jganf6q1rEO25VSoHH0hKPCTrgillPjdI/3bgBhAE1QzHrHTizjpRvy6pGAvKjDJtk2pF9NDq8w==", + "dev": true, + "license": "MIT", + "bin": { + "eslint-config-prettier": "bin/cli.js" + }, + "funding": { + "url": "https://opencollective.com/eslint-config-prettier" + }, + "peerDependencies": { + "eslint": ">=7.0.0" + } + }, + "node_modules/eslint-plugin-storybook": { + "version": "9.0.17", + "resolved": "https://registry.npmjs.org/eslint-plugin-storybook/-/eslint-plugin-storybook-9.0.17.tgz", + "integrity": "sha512-IuTdlwCEwoDNobdygRCxNhlKXHmsDfPtPvHGcsY35x2Bx8KItrjfekO19gJrjc1VT2CMfcZMYF8OBKaxHELupw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@typescript-eslint/utils": "^8.8.1" + }, + "engines": { + "node": ">=20.0.0" + }, + "peerDependencies": { + "eslint": ">=8", + "storybook": "^9.0.17" + } + }, + "node_modules/eslint-plugin-svelte": { + "version": "3.11.0", + "resolved": "https://registry.npmjs.org/eslint-plugin-svelte/-/eslint-plugin-svelte-3.11.0.tgz", + "integrity": "sha512-KliWlkieHyEa65aQIkRwUFfHzT5Cn4u3BQQsu3KlkJOs7c1u7ryn84EWaOjEzilbKgttT4OfBURA8Uc4JBSQIw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@eslint-community/eslint-utils": "^4.6.1", + "@jridgewell/sourcemap-codec": "^1.5.0", + "esutils": "^2.0.3", + "globals": "^16.0.0", + "known-css-properties": "^0.37.0", + "postcss": "^8.4.49", + "postcss-load-config": "^3.1.4", + "postcss-safe-parser": "^7.0.0", + "semver": "^7.6.3", + "svelte-eslint-parser": "^1.3.0" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "url": "https://github.com/sponsors/ota-meshi" + }, + "peerDependencies": { + "eslint": "^8.57.1 || ^9.0.0", + "svelte": "^3.37.0 || ^4.0.0 || ^5.0.0" + }, + "peerDependenciesMeta": { + "svelte": { + "optional": true + } + } + }, + "node_modules/eslint-scope": { + "version": "8.4.0", + "resolved": "https://registry.npmjs.org/eslint-scope/-/eslint-scope-8.4.0.tgz", + "integrity": "sha512-sNXOfKCn74rt8RICKMvJS7XKV/Xk9kA7DyJr8mJik3S7Cwgy3qlkkmyS2uQB3jiJg6VNdZd/pDBJu0nvG2NlTg==", + "dev": true, + "license": "BSD-2-Clause", + "dependencies": { + "esrecurse": "^4.3.0", + "estraverse": "^5.2.0" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "url": "https://opencollective.com/eslint" + } + }, + "node_modules/eslint-visitor-keys": { + "version": "4.2.1", + "resolved": "https://registry.npmjs.org/eslint-visitor-keys/-/eslint-visitor-keys-4.2.1.tgz", + "integrity": "sha512-Uhdk5sfqcee/9H/rCOJikYz67o0a2Tw2hGRPOG2Y1R2dg7brRe1uG0yaNQDHu+TO/uQPF/5eCapvYSmHUjt7JQ==", + "dev": true, + "license": "Apache-2.0", + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "url": "https://opencollective.com/eslint" + } + }, + "node_modules/esm-env": { + "version": "1.2.2", + "resolved": "https://registry.npmjs.org/esm-env/-/esm-env-1.2.2.tgz", + "integrity": "sha512-Epxrv+Nr/CaL4ZcFGPJIYLWFom+YeV1DqMLHJoEd9SYRxNbaFruBwfEX/kkHUJf55j2+TUbmDcmuilbP1TmXHA==", + "license": "MIT" + }, + "node_modules/espree": { + "version": "10.4.0", + "resolved": "https://registry.npmjs.org/espree/-/espree-10.4.0.tgz", + "integrity": "sha512-j6PAQ2uUr79PZhBjP5C5fhl8e39FmRnOjsD5lGnWrFU8i2G776tBK7+nP8KuQUTTyAZUwfQqXAgrVH5MbH9CYQ==", + "dev": true, + "license": "BSD-2-Clause", + "dependencies": { + "acorn": "^8.15.0", + "acorn-jsx": "^5.3.2", + "eslint-visitor-keys": "^4.2.1" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "url": "https://opencollective.com/eslint" + } + }, + "node_modules/esprima": { + "version": "4.0.1", + "resolved": "https://registry.npmjs.org/esprima/-/esprima-4.0.1.tgz", + "integrity": "sha512-eGuFFw7Upda+g4p+QHvnW0RyTX/SVeJBDM/gCtMARO0cLuT2HcEKnTPvhjV6aGeqrCB/sbNop0Kszm0jsaWU4A==", + "dev": true, + "license": "BSD-2-Clause", + "bin": { + "esparse": "bin/esparse.js", + "esvalidate": "bin/esvalidate.js" + }, + "engines": { + "node": ">=4" + } + }, + "node_modules/esquery": { + "version": "1.6.0", + "resolved": "https://registry.npmjs.org/esquery/-/esquery-1.6.0.tgz", + "integrity": "sha512-ca9pw9fomFcKPvFLXhBKUK90ZvGibiGOvRJNbjljY7s7uq/5YO4BOzcYtJqExdx99rF6aAcnRxHmcUHcz6sQsg==", + "dev": true, + "license": "BSD-3-Clause", + "dependencies": { + "estraverse": "^5.1.0" + }, + "engines": { + "node": ">=0.10" + } + }, + "node_modules/esrap": { + "version": "1.4.9", + "resolved": "https://registry.npmjs.org/esrap/-/esrap-1.4.9.tgz", + "integrity": "sha512-3OMlcd0a03UGuZpPeUC1HxR3nA23l+HEyCiZw3b3FumJIN9KphoGzDJKMXI1S72jVS1dsenDyQC0kJlO1U9E1g==", + "dev": true, + "license": "MIT", + "dependencies": { + "@jridgewell/sourcemap-codec": "^1.4.15" + } + }, + "node_modules/esrecurse": { + "version": "4.3.0", + "resolved": "https://registry.npmjs.org/esrecurse/-/esrecurse-4.3.0.tgz", + "integrity": "sha512-KmfKL3b6G+RXvP8N1vr3Tq1kL/oCFgn2NYXEtqP8/L3pKapUA4G8cFVaoF3SU323CD4XypR/ffioHmkti6/Tag==", + "dev": true, + "license": "BSD-2-Clause", + "dependencies": { + "estraverse": "^5.2.0" + }, + "engines": { + "node": ">=4.0" + } + }, + "node_modules/estraverse": { + "version": "5.3.0", + "resolved": "https://registry.npmjs.org/estraverse/-/estraverse-5.3.0.tgz", + "integrity": "sha512-MMdARuVEQziNTeJD8DgMqmhwR11BRQ/cBP+pLtYdSTnf3MIO8fFeiINEbX36ZdNlfU/7A9f3gUw49B3oQsvwBA==", + "dev": true, + "license": "BSD-2-Clause", + "engines": { + "node": ">=4.0" + } + }, + "node_modules/esutils": { + "version": "2.0.3", + "resolved": "https://registry.npmjs.org/esutils/-/esutils-2.0.3.tgz", + "integrity": "sha512-kVscqXk4OCp68SZ0dkgEKVi6/8ij300KBWTJq32P/dYeWTSwK41WyTxalN1eRmA5Z9UU/LX9D7FWSmV9SAYx6g==", + "dev": true, + "license": "BSD-2-Clause", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/expect-type": { + "version": "1.2.2", + "resolved": "https://registry.npmjs.org/expect-type/-/expect-type-1.2.2.tgz", + "integrity": "sha512-JhFGDVJ7tmDJItKhYgJCGLOWjuK9vPxiXoUFLwLDc99NlmklilbiQJwoctZtt13+xMw91MCk/REan6MWHqDjyA==", + "dev": true, + "license": "Apache-2.0", + "engines": { + "node": ">=12.0.0" + } + }, + "node_modules/extend": { + "version": "3.0.2", + "resolved": "https://registry.npmjs.org/extend/-/extend-3.0.2.tgz", + "integrity": "sha512-fjquC59cD7CyW6urNXK0FBufkZcoiGG80wTuPujX590cB5Ttln20E2UB4S/WARVqhXffZl2LNgS+gQdPIIim/g==", + "license": "MIT" + }, + "node_modules/fast-deep-equal": { + "version": "3.1.3", + "resolved": "https://registry.npmjs.org/fast-deep-equal/-/fast-deep-equal-3.1.3.tgz", + "integrity": "sha512-f3qQ9oQy9j2AhBe/H9VC91wLmKBCCU/gDOnKNAYG5hswO7BLKj09Hc5HYNz9cGI++xlpDCIgDaitVs03ATR84Q==", + "dev": true, + "license": "MIT" + }, + "node_modules/fast-glob": { + "version": "3.3.3", + "resolved": "https://registry.npmjs.org/fast-glob/-/fast-glob-3.3.3.tgz", + "integrity": "sha512-7MptL8U0cqcFdzIzwOTHoilX9x5BrNqye7Z/LuC7kCMRio1EMSyqRK3BEAUD7sXRq4iT4AzTVuZdhgQ2TCvYLg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@nodelib/fs.stat": "^2.0.2", + "@nodelib/fs.walk": "^1.2.3", + "glob-parent": "^5.1.2", + "merge2": "^1.3.0", + "micromatch": "^4.0.8" + }, + "engines": { + "node": ">=8.6.0" + } + }, + "node_modules/fast-glob/node_modules/glob-parent": { + "version": "5.1.2", + "resolved": "https://registry.npmjs.org/glob-parent/-/glob-parent-5.1.2.tgz", + "integrity": "sha512-AOIgSQCepiJYwP3ARnGx+5VnTu2HBYdzbGP45eLw1vr3zB3vZLeyed1sC9hnbcOc9/SrMyM5RPQrkGz4aS9Zow==", + "dev": true, + "license": "ISC", + "dependencies": { + "is-glob": "^4.0.1" + }, + "engines": { + "node": ">= 6" + } + }, + "node_modules/fast-json-stable-stringify": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/fast-json-stable-stringify/-/fast-json-stable-stringify-2.1.0.tgz", + "integrity": "sha512-lhd/wF+Lk98HZoTCtlVraHtfh5XYijIjalXck7saUtuanSDyLMxnHhSXEDJqHxD7msR8D0uCmqlkwjCV8xvwHw==", + "dev": true, + "license": "MIT" + }, + "node_modules/fast-levenshtein": { + "version": "2.0.6", + "resolved": "https://registry.npmjs.org/fast-levenshtein/-/fast-levenshtein-2.0.6.tgz", + "integrity": "sha512-DCXu6Ifhqcks7TZKY3Hxp3y6qphY5SJZmrWMDrKcERSOXWQdMhU9Ig/PYrzyw/ul9jOIyh0N4M0tbC5hodg8dw==", + "dev": true, + "license": "MIT" + }, + "node_modules/fastq": { + "version": "1.19.1", + "resolved": "https://registry.npmjs.org/fastq/-/fastq-1.19.1.tgz", + "integrity": "sha512-GwLTyxkCXjXbxqIhTsMI2Nui8huMPtnxg7krajPJAjnEG/iiOS7i+zCtWGZR9G0NBKbXKh6X9m9UIsYX/N6vvQ==", + "dev": true, + "license": "ISC", + "dependencies": { + "reusify": "^1.0.4" + } + }, + "node_modules/fdir": { + "version": "6.4.6", + "resolved": "https://registry.npmjs.org/fdir/-/fdir-6.4.6.tgz", + "integrity": "sha512-hiFoqpyZcfNm1yc4u8oWCf9A2c4D3QjCrks3zmoVKVxpQRzmPNar1hUJcBG2RQHvEVGDN+Jm81ZheVLAQMK6+w==", + "dev": true, + "license": "MIT", + "peerDependencies": { + "picomatch": "^3 || ^4" + }, + "peerDependenciesMeta": { + "picomatch": { + "optional": true + } + } + }, + "node_modules/fflate": { + "version": "0.8.2", + "resolved": "https://registry.npmjs.org/fflate/-/fflate-0.8.2.tgz", + "integrity": "sha512-cPJU47OaAoCbg0pBvzsgpTPhmhqI5eJjh/JIu8tPj5q+T7iLvW/JAYUqmE7KOB4R1ZyEhzBaIQpQpardBF5z8A==", + "dev": true, + "license": "MIT" + }, + "node_modules/file-entry-cache": { + "version": "8.0.0", + "resolved": "https://registry.npmjs.org/file-entry-cache/-/file-entry-cache-8.0.0.tgz", + "integrity": "sha512-XXTUwCvisa5oacNGRP9SfNtYBNAMi+RPwBFmblZEF7N7swHYQS6/Zfk7SRwx4D5j3CH211YNRco1DEMNVfZCnQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "flat-cache": "^4.0.0" + }, + "engines": { + "node": ">=16.0.0" + } + }, + "node_modules/filesize": { + "version": "10.1.6", + "resolved": "https://registry.npmjs.org/filesize/-/filesize-10.1.6.tgz", + "integrity": "sha512-sJslQKU2uM33qH5nqewAwVB2QgR6w1aMNsYUp3aN5rMRyXEwJGmZvaWzeJFNTOXWlHQyBFCWrdj3fV/fsTOX8w==", + "dev": true, + "license": "BSD-3-Clause", + "engines": { + "node": ">= 10.4.0" + } + }, + "node_modules/fill-range": { + "version": "7.1.1", + "resolved": "https://registry.npmjs.org/fill-range/-/fill-range-7.1.1.tgz", + "integrity": "sha512-YsGpe3WHLK8ZYi4tWDg2Jy3ebRz2rXowDxnld4bkQB00cc/1Zw9AWnC0i9ztDJitivtQvaI9KaLyKrc+hBW0yg==", + "dev": true, + "license": "MIT", + "dependencies": { + "to-regex-range": "^5.0.1" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/find-up": { + "version": "5.0.0", + "resolved": "https://registry.npmjs.org/find-up/-/find-up-5.0.0.tgz", + "integrity": "sha512-78/PXT1wlLLDgTzDs7sjq9hzz0vXD+zn+7wypEe4fXQxCmdmqfGsEPQxmiCSQI3ajFV91bVSsvNtrJRiW6nGng==", + "dev": true, + "license": "MIT", + "dependencies": { + "locate-path": "^6.0.0", + "path-exists": "^4.0.0" + }, + "engines": { + "node": ">=10" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/flat-cache": { + "version": "4.0.1", + "resolved": "https://registry.npmjs.org/flat-cache/-/flat-cache-4.0.1.tgz", + "integrity": "sha512-f7ccFPK3SXFHpx15UIGyRJ/FJQctuKZ0zVuN3frBo4HnK3cay9VEW0R6yPYFHC0AgqhukPzKjq22t5DmAyqGyw==", + "dev": true, + "license": "MIT", + "dependencies": { + "flatted": "^3.2.9", + "keyv": "^4.5.4" + }, + "engines": { + "node": ">=16" + } + }, + "node_modules/flatted": { + "version": "3.3.3", + "resolved": "https://registry.npmjs.org/flatted/-/flatted-3.3.3.tgz", + "integrity": "sha512-GX+ysw4PBCz0PzosHDepZGANEuFCMLrnRTiEy9McGjmkCQYwRq4A/X786G/fjM/+OjsWSU1ZrY5qyARZmO/uwg==", + "dev": true, + "license": "ISC" + }, + "node_modules/fsevents": { + "version": "2.3.2", + "resolved": "https://registry.npmjs.org/fsevents/-/fsevents-2.3.2.tgz", + "integrity": "sha512-xiqMQR4xAeHTuB9uWm+fFRcIOgKBMiOBP+eXiyT7jsgVCq1bkVygt00oASowB7EdtpOHaaPgKt812P9ab+DDKA==", + "dev": true, + "hasInstallScript": true, + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": "^8.16.0 || ^10.6.0 || >=11.0.0" + } + }, + "node_modules/glob-parent": { + "version": "6.0.2", + "resolved": "https://registry.npmjs.org/glob-parent/-/glob-parent-6.0.2.tgz", + "integrity": "sha512-XxwI8EOhVQgWp6iDL+3b0r86f4d6AX6zSU55HfB4ydCEuXLXc5FcYeOu+nnGftS4TEju/11rt4KJPTMgbfmv4A==", + "dev": true, + "license": "ISC", + "dependencies": { + "is-glob": "^4.0.3" + }, + "engines": { + "node": ">=10.13.0" + } + }, + "node_modules/globals": { + "version": "16.3.0", + "resolved": "https://registry.npmjs.org/globals/-/globals-16.3.0.tgz", + "integrity": "sha512-bqWEnJ1Nt3neqx2q5SFfGS8r/ahumIakg3HcwtNlrVlwXIeNumWn/c7Pn/wKzGhf6SaW6H6uWXLqC30STCMchQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=18" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/graceful-fs": { + "version": "4.2.11", + "resolved": "https://registry.npmjs.org/graceful-fs/-/graceful-fs-4.2.11.tgz", + "integrity": "sha512-RbJ5/jmFcNNCcDV5o9eTnBLJ/HszWV0P73bc+Ff4nS/rJj+YaS6IGyiOL0VoBYX+l1Wrl3k63h/KrH+nhJ0XvQ==", + "dev": true, + "license": "ISC" + }, + "node_modules/graphemer": { + "version": "1.4.0", + "resolved": "https://registry.npmjs.org/graphemer/-/graphemer-1.4.0.tgz", + "integrity": "sha512-EtKwoO6kxCL9WO5xipiHTZlSzBm7WLT627TqC/uVRd0HKmq8NXyebnNYxDoBi7wt8eTWrUrKXCOVaFq9x1kgag==", + "dev": true, + "license": "MIT" + }, + "node_modules/has-flag": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/has-flag/-/has-flag-4.0.0.tgz", + "integrity": "sha512-EykJT/Q1KjTWctppgIAgfSO0tKVuZUjhgMr17kqTumMl6Afv3EISleU7qZUzoXDFTAHTDC4NOoG/ZxU3EvlMPQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/hast-util-from-dom": { + "version": "5.0.1", + "resolved": "https://registry.npmjs.org/hast-util-from-dom/-/hast-util-from-dom-5.0.1.tgz", + "integrity": "sha512-N+LqofjR2zuzTjCPzyDUdSshy4Ma6li7p/c3pA78uTwzFgENbgbUrm2ugwsOdcjI1muO+o6Dgzp9p8WHtn/39Q==", + "dev": true, + "license": "ISC", + "dependencies": { + "@types/hast": "^3.0.0", + "hastscript": "^9.0.0", + "web-namespaces": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/hast-util-from-html": { + "version": "2.0.3", + "resolved": "https://registry.npmjs.org/hast-util-from-html/-/hast-util-from-html-2.0.3.tgz", + "integrity": "sha512-CUSRHXyKjzHov8yKsQjGOElXy/3EKpyX56ELnkHH34vDVw1N1XSQ1ZcAvTyAPtGqLTuKP/uxM+aLkSPqF/EtMw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "devlop": "^1.1.0", + "hast-util-from-parse5": "^8.0.0", + "parse5": "^7.0.0", + "vfile": "^6.0.0", + "vfile-message": "^4.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/hast-util-from-html-isomorphic": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/hast-util-from-html-isomorphic/-/hast-util-from-html-isomorphic-2.0.0.tgz", + "integrity": "sha512-zJfpXq44yff2hmE0XmwEOzdWin5xwH+QIhMLOScpX91e/NSGPsAzNCvLQDIEPyO2TXi+lBmU6hjLIhV8MwP2kw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "hast-util-from-dom": "^5.0.0", + "hast-util-from-html": "^2.0.0", + "unist-util-remove-position": "^5.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/hast-util-from-html/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, + "license": "MIT" + }, + "node_modules/hast-util-from-html/node_modules/unist-util-stringify-position": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/unist-util-stringify-position/-/unist-util-stringify-position-4.0.0.tgz", + "integrity": "sha512-0ASV06AAoKCDkS2+xw5RXJywruurpbC4JZSm7nr7MOt1ojAzvyyaO+UxZf18j8FCF6kmzCZKcAgN/yu2gm2XgQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/hast-util-from-html/node_modules/vfile-message": { + "version": "4.0.3", + "resolved": "https://registry.npmjs.org/vfile-message/-/vfile-message-4.0.3.tgz", + "integrity": "sha512-QTHzsGd1EhbZs4AsQ20JX1rC3cOlt/IWJruk893DfLRr57lcnOeMaWG4K0JrRta4mIJZKth2Au3mM3u03/JWKw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0", + "unist-util-stringify-position": "^4.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/hast-util-from-parse5": { + "version": "8.0.3", + "resolved": "https://registry.npmjs.org/hast-util-from-parse5/-/hast-util-from-parse5-8.0.3.tgz", + "integrity": "sha512-3kxEVkEKt0zvcZ3hCRYI8rqrgwtlIOFMWkbclACvjlDw8Li9S2hk/d51OI0nr/gIpdMHNepwgOKqZ/sy0Clpyg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "@types/unist": "^3.0.0", + "devlop": "^1.0.0", + "hastscript": "^9.0.0", + "property-information": "^7.0.0", + "vfile": "^6.0.0", + "vfile-location": "^5.0.0", + "web-namespaces": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/hast-util-from-parse5/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, + "license": "MIT" + }, + "node_modules/hast-util-is-element": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/hast-util-is-element/-/hast-util-is-element-3.0.0.tgz", + "integrity": "sha512-Val9mnv2IWpLbNPqc/pUem+a7Ipj2aHacCwgNfTiK0vJKl0LF+4Ba4+v1oPHFpf3bLYmreq0/l3Gud9S5OH42g==", + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/hast-util-parse-selector": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/hast-util-parse-selector/-/hast-util-parse-selector-4.0.0.tgz", + "integrity": "sha512-wkQCkSYoOGCRKERFWcxMVMOcYE2K1AaNLU8DXS9arxnLOUEWbOXKXiJUNzEpqZ3JOKpnha3jkFrumEjVliDe7A==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/hast-util-sanitize": { + "version": "5.0.2", + "resolved": "https://registry.npmjs.org/hast-util-sanitize/-/hast-util-sanitize-5.0.2.tgz", + "integrity": "sha512-3yTWghByc50aGS7JlGhk61SPenfE/p1oaFeNwkOOyrscaOkMGrcW9+Cy/QAIOBpZxP1yqDIzFMR0+Np0i0+usg==", + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "@ungap/structured-clone": "^1.0.0", + "unist-util-position": "^5.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/hast-util-to-html": { + "version": "9.0.5", + "resolved": "https://registry.npmjs.org/hast-util-to-html/-/hast-util-to-html-9.0.5.tgz", + "integrity": "sha512-OguPdidb+fbHQSU4Q4ZiLKnzWo8Wwsf5bZfbvu7//a9oTYoqD/fWpe96NuHkoS9h0ccGOTe0C4NGXdtS0iObOw==", + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "@types/unist": "^3.0.0", + "ccount": "^2.0.0", + "comma-separated-tokens": "^2.0.0", + "hast-util-whitespace": "^3.0.0", + "html-void-elements": "^3.0.0", + "mdast-util-to-hast": "^13.0.0", + "property-information": "^7.0.0", + "space-separated-tokens": "^2.0.0", + "stringify-entities": "^4.0.0", + "zwitch": "^2.0.4" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/hast-util-to-html/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "license": "MIT" + }, + "node_modules/hast-util-to-text": { + "version": "4.0.2", + "resolved": "https://registry.npmjs.org/hast-util-to-text/-/hast-util-to-text-4.0.2.tgz", + "integrity": "sha512-KK6y/BN8lbaq654j7JgBydev7wuNMcID54lkRav1P0CaE1e47P72AWWPiGKXTJU271ooYzcvTAn/Zt0REnvc7A==", + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "@types/unist": "^3.0.0", + "hast-util-is-element": "^3.0.0", + "unist-util-find-after": "^5.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/hast-util-to-text/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "license": "MIT" + }, + "node_modules/hast-util-whitespace": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/hast-util-whitespace/-/hast-util-whitespace-3.0.0.tgz", + "integrity": "sha512-88JUN06ipLwsnv+dVn+OIYOvAuvBMy/Qoi6O7mQHxdPXpjy+Cd6xRkWwux7DKO+4sYILtLBRIKgsdpS2gQc7qw==", + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/hastscript": { + "version": "9.0.1", + "resolved": "https://registry.npmjs.org/hastscript/-/hastscript-9.0.1.tgz", + "integrity": "sha512-g7df9rMFX/SPi34tyGCyUBREQoKkapwdY/T04Qn9TDWfHhAYt4/I0gMVirzK5wEzeUqIjEB+LXC/ypb7Aqno5w==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "comma-separated-tokens": "^2.0.0", + "hast-util-parse-selector": "^4.0.0", + "property-information": "^7.0.0", + "space-separated-tokens": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/highlight.js": { + "version": "11.11.1", + "resolved": "https://registry.npmjs.org/highlight.js/-/highlight.js-11.11.1.tgz", + "integrity": "sha512-Xwwo44whKBVCYoliBQwaPvtd/2tYFkRQtXDWj1nackaV2JPXx3L0+Jvd8/qCJ2p+ML0/XVkJ2q+Mr+UVdpJK5w==", + "license": "BSD-3-Clause", + "engines": { + "node": ">=12.0.0" + } + }, + "node_modules/html-void-elements": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/html-void-elements/-/html-void-elements-3.0.0.tgz", + "integrity": "sha512-bEqo66MRXsUGxWHV5IP0PUiAWwoEjba4VCzg0LjFJBpchPaTfyfCKTG6bc5F8ucKec3q5y6qOdGyYTSBEvhCrg==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/ignore": { + "version": "5.3.2", + "resolved": "https://registry.npmjs.org/ignore/-/ignore-5.3.2.tgz", + "integrity": "sha512-hsBTNUqQTDwkWtcdYI2i06Y/nUBEsNEDJKjWdigLvegy8kDuJAS8uRlpkkcQpyEXL0Z/pjDy5HBmMjRCJ2gq+g==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">= 4" + } + }, + "node_modules/import-fresh": { + "version": "3.3.1", + "resolved": "https://registry.npmjs.org/import-fresh/-/import-fresh-3.3.1.tgz", + "integrity": "sha512-TR3KfrTZTYLPB6jUjfx6MF9WcWrHL9su5TObK4ZkYgBdWKPOFoSoQIdEuTuR82pmtxH2spWG9h6etwfr1pLBqQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "parent-module": "^1.0.0", + "resolve-from": "^4.0.0" + }, + "engines": { + "node": ">=6" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/imurmurhash": { + "version": "0.1.4", + "resolved": "https://registry.npmjs.org/imurmurhash/-/imurmurhash-0.1.4.tgz", + "integrity": "sha512-JmXMZ6wuvDmLiHEml9ykzqO6lwFbof0GG4IkcGaENdCRDDmMVnny7s5HsIgHCbaq0w2MyPhDqkhTUgS2LU2PHA==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=0.8.19" + } + }, + "node_modules/indent-string": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/indent-string/-/indent-string-4.0.0.tgz", + "integrity": "sha512-EdDDZu4A2OyIK7Lr/2zG+w5jmbuk1DVBnEwREQvBzspBJkCEbRa8GxU1lghYcaGJCnRWibjDXlq779X1/y5xwg==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/inline-style-parser": { + "version": "0.2.4", + "resolved": "https://registry.npmjs.org/inline-style-parser/-/inline-style-parser-0.2.4.tgz", + "integrity": "sha512-0aO8FkhNZlj/ZIbNi7Lxxr12obT7cL1moPfE4tg1LkX7LlLfC6DeX4l2ZEud1ukP9jNQyNnfzQVqwbwmAATY4Q==", + "license": "MIT" + }, + "node_modules/is-docker": { + "version": "2.2.1", + "resolved": "https://registry.npmjs.org/is-docker/-/is-docker-2.2.1.tgz", + "integrity": "sha512-F+i2BKsFrH66iaUFc0woD8sLy8getkwTwtOBjvs56Cx4CgJDeKQeqfz8wAYiSb8JOprWhHH5p77PbmYCvvUuXQ==", + "dev": true, + "license": "MIT", + "bin": { + "is-docker": "cli.js" + }, + "engines": { + "node": ">=8" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/is-extglob": { + "version": "2.1.1", + "resolved": "https://registry.npmjs.org/is-extglob/-/is-extglob-2.1.1.tgz", + "integrity": "sha512-SbKbANkN603Vi4jEZv49LeVJMn4yGwsbzZworEoyEiutsN3nJYdbO36zfhGJ6QEDpOZIFkDtnq5JRxmvl3jsoQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/is-glob": { + "version": "4.0.3", + "resolved": "https://registry.npmjs.org/is-glob/-/is-glob-4.0.3.tgz", + "integrity": "sha512-xelSayHH36ZgE7ZWhli7pW34hNbNl8Ojv5KVmkJD4hBdD3th8Tfk9vYasLM+mXWOZhFkgZfxhLSnrwRr4elSSg==", + "dev": true, + "license": "MIT", + "dependencies": { + "is-extglob": "^2.1.1" + }, + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/is-number": { + "version": "7.0.0", + "resolved": "https://registry.npmjs.org/is-number/-/is-number-7.0.0.tgz", + "integrity": "sha512-41Cifkg6e8TylSpdtTpeLVMqvSBEVzTttHvERD741+pnZ8ANv0004MRL43QKPDlK9cGvNp6NZWZUBlbGXYxxng==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=0.12.0" + } + }, + "node_modules/is-plain-obj": { + "version": "4.1.0", + "resolved": "https://registry.npmjs.org/is-plain-obj/-/is-plain-obj-4.1.0.tgz", + "integrity": "sha512-+Pgi+vMuUNkJyExiMBt5IlFoMyKnr5zhJ4Uspz58WOhBF5QoIZkFyNHIbBAtHwzVAgk5RtndVNsDRN61/mmDqg==", + "license": "MIT", + "engines": { + "node": ">=12" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/is-wsl": { + "version": "2.2.0", + "resolved": "https://registry.npmjs.org/is-wsl/-/is-wsl-2.2.0.tgz", + "integrity": "sha512-fKzAra0rGJUUBwGBgNkHZuToZcn+TtXHpeCgmkMJMMYx1sQDYaCSyjJBSCa2nH1DGm7s3n1oBnohoVTBaN7Lww==", + "dev": true, + "license": "MIT", + "dependencies": { + "is-docker": "^2.0.0" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/isexe": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/isexe/-/isexe-2.0.0.tgz", + "integrity": "sha512-RHxMLp9lnKHGHRng9QFhRCMbYAcVpn69smSGcq3f36xjgVVWThj4qqLbTLlq7Ssj8B+fIQ1EuCEGI2lKsyQeIw==", + "dev": true, + "license": "ISC" + }, + "node_modules/jiti": { + "version": "2.4.2", + "resolved": "https://registry.npmjs.org/jiti/-/jiti-2.4.2.tgz", + "integrity": "sha512-rg9zJN+G4n2nfJl5MW3BMygZX56zKPNVEYYqq7adpmMh4Jn2QNEwhvQlFy6jPVdcod7txZtKHWnyZiA3a0zP7A==", + "dev": true, + "license": "MIT", + "bin": { + "jiti": "lib/jiti-cli.mjs" + } + }, + "node_modules/js-tokens": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/js-tokens/-/js-tokens-4.0.0.tgz", + "integrity": "sha512-RdJUflcE3cUzKiMqQgsCu06FPu9UdIJO0beYbPhHN4k6apgJtifcoCtT9bcxOpYBtpD2kCM6Sbzg4CausW/PKQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/js-yaml": { + "version": "4.1.0", + "resolved": "https://registry.npmjs.org/js-yaml/-/js-yaml-4.1.0.tgz", + "integrity": "sha512-wpxZs9NoxZaJESJGIZTyDEaYpl0FKSA+FB9aJiyemKhMwkxQg63h4T1KJgUGHpTqPDNRcmmYLugrRjJlBtWvRA==", + "dev": true, + "license": "MIT", + "dependencies": { + "argparse": "^2.0.1" + }, + "bin": { + "js-yaml": "bin/js-yaml.js" + } + }, + "node_modules/json-buffer": { + "version": "3.0.1", + "resolved": "https://registry.npmjs.org/json-buffer/-/json-buffer-3.0.1.tgz", + "integrity": "sha512-4bV5BfR2mqfQTJm+V5tPPdf+ZpuhiIvTuAB5g8kcrXOZpTT/QwwVRWBywX1ozr6lEuPdbHxwaJlm9G6mI2sfSQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/json-schema-traverse": { + "version": "0.4.1", + "resolved": "https://registry.npmjs.org/json-schema-traverse/-/json-schema-traverse-0.4.1.tgz", + "integrity": "sha512-xbbCH5dCYU5T8LcEhhuh7HJ88HXuW3qsI3Y0zOZFKfZEHcpWiHU/Jxzk629Brsab/mMiHQti9wMP+845RPe3Vg==", + "dev": true, + "license": "MIT" + }, + "node_modules/json-stable-stringify-without-jsonify": { + "version": "1.0.1", + "resolved": "https://registry.npmjs.org/json-stable-stringify-without-jsonify/-/json-stable-stringify-without-jsonify-1.0.1.tgz", + "integrity": "sha512-Bdboy+l7tA3OGW6FjyFHWkP5LuByj1Tk33Ljyq0axyzdk9//JSi2u3fP1QSmd1KNwq6VOKYGlAu87CisVir6Pw==", + "dev": true, + "license": "MIT" + }, + "node_modules/jsonfile": { + "version": "6.1.0", + "resolved": "https://registry.npmjs.org/jsonfile/-/jsonfile-6.1.0.tgz", + "integrity": "sha512-5dgndWOriYSm5cnYaJNhalLNDKOqFwyDB/rr1E9ZsGciGvKPs8R2xYGCacuf3z6K1YKDz182fd+fY3cn3pMqXQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "universalify": "^2.0.0" + }, + "optionalDependencies": { + "graceful-fs": "^4.1.6" + } + }, + "node_modules/katex": { + "version": "0.16.22", + "resolved": "https://registry.npmjs.org/katex/-/katex-0.16.22.tgz", + "integrity": "sha512-XCHRdUw4lf3SKBaJe4EvgqIuWwkPSo9XoeO8GjQW94Bp7TWv9hNhzZjZ+OH9yf1UmLygb7DIT5GSFQiyt16zYg==", + "dev": true, + "funding": [ + "https://opencollective.com/katex", + "https://github.com/sponsors/katex" + ], + "license": "MIT", + "dependencies": { + "commander": "^8.3.0" + }, + "bin": { + "katex": "cli.js" + } + }, + "node_modules/keyv": { + "version": "4.5.4", + "resolved": "https://registry.npmjs.org/keyv/-/keyv-4.5.4.tgz", + "integrity": "sha512-oxVHkHR/EJf2CNXnWxRLW6mg7JyCCUcG0DtEGmL2ctUo1PNTin1PUil+r/+4r5MpVgC/fn1kjsx7mjSujKqIpw==", + "dev": true, + "license": "MIT", + "dependencies": { + "json-buffer": "3.0.1" + } + }, + "node_modules/kleur": { + "version": "4.1.5", + "resolved": "https://registry.npmjs.org/kleur/-/kleur-4.1.5.tgz", + "integrity": "sha512-o+NO+8WrRiQEE4/7nwRJhN1HWpVmJm511pBHUxPLtp0BUISzlBplORYSmTclCnJvQq2tKu/sgl3xVpkc7ZWuQQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=6" + } + }, + "node_modules/known-css-properties": { + "version": "0.37.0", + "resolved": "https://registry.npmjs.org/known-css-properties/-/known-css-properties-0.37.0.tgz", + "integrity": "sha512-JCDrsP4Z1Sb9JwG0aJ8Eo2r7k4Ou5MwmThS/6lcIe1ICyb7UBJKGRIUUdqc2ASdE/42lgz6zFUnzAIhtXnBVrQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/levn": { + "version": "0.4.1", + "resolved": "https://registry.npmjs.org/levn/-/levn-0.4.1.tgz", + "integrity": "sha512-+bT2uH4E5LGE7h/n3evcS/sQlJXCpIp6ym8OWJ5eV6+67Dsql/LaaT7qJBAt2rzfoa/5QBGBhxDix1dMt2kQKQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "prelude-ls": "^1.2.1", + "type-check": "~0.4.0" + }, + "engines": { + "node": ">= 0.8.0" + } + }, + "node_modules/lightningcss": { + "version": "1.30.1", + "resolved": "https://registry.npmjs.org/lightningcss/-/lightningcss-1.30.1.tgz", + "integrity": "sha512-xi6IyHML+c9+Q3W0S4fCQJOym42pyurFiJUHEcEyHS0CeKzia4yZDEsLlqOFykxOdHpNy0NmvVO31vcSqAxJCg==", + "dev": true, + "license": "MPL-2.0", + "dependencies": { + "detect-libc": "^2.0.3" + }, + "engines": { + "node": ">= 12.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/parcel" + }, + "optionalDependencies": { + "lightningcss-darwin-arm64": "1.30.1", + "lightningcss-darwin-x64": "1.30.1", + "lightningcss-freebsd-x64": "1.30.1", + "lightningcss-linux-arm-gnueabihf": "1.30.1", + "lightningcss-linux-arm64-gnu": "1.30.1", + "lightningcss-linux-arm64-musl": "1.30.1", + "lightningcss-linux-x64-gnu": "1.30.1", + "lightningcss-linux-x64-musl": "1.30.1", + "lightningcss-win32-arm64-msvc": "1.30.1", + "lightningcss-win32-x64-msvc": "1.30.1" + } + }, + "node_modules/lightningcss-darwin-arm64": { + "version": "1.30.1", + "resolved": "https://registry.npmjs.org/lightningcss-darwin-arm64/-/lightningcss-darwin-arm64-1.30.1.tgz", + "integrity": "sha512-c8JK7hyE65X1MHMN+Viq9n11RRC7hgin3HhYKhrMyaXflk5GVplZ60IxyoVtzILeKr+xAJwg6zK6sjTBJ0FKYQ==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MPL-2.0", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": ">= 12.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/parcel" + } + }, + "node_modules/lightningcss-darwin-x64": { + "version": "1.30.1", + "resolved": "https://registry.npmjs.org/lightningcss-darwin-x64/-/lightningcss-darwin-x64-1.30.1.tgz", + "integrity": "sha512-k1EvjakfumAQoTfcXUcHQZhSpLlkAuEkdMBsI/ivWw9hL+7FtilQc0Cy3hrx0AAQrVtQAbMI7YjCgYgvn37PzA==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MPL-2.0", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": ">= 12.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/parcel" + } + }, + "node_modules/lightningcss-freebsd-x64": { + "version": "1.30.1", + "resolved": "https://registry.npmjs.org/lightningcss-freebsd-x64/-/lightningcss-freebsd-x64-1.30.1.tgz", + "integrity": "sha512-kmW6UGCGg2PcyUE59K5r0kWfKPAVy4SltVeut+umLCFoJ53RdCUWxcRDzO1eTaxf/7Q2H7LTquFHPL5R+Gjyig==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MPL-2.0", + "optional": true, + "os": [ + "freebsd" + ], + "engines": { + "node": ">= 12.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/parcel" + } + }, + "node_modules/lightningcss-linux-arm-gnueabihf": { + "version": "1.30.1", + "resolved": "https://registry.npmjs.org/lightningcss-linux-arm-gnueabihf/-/lightningcss-linux-arm-gnueabihf-1.30.1.tgz", + "integrity": "sha512-MjxUShl1v8pit+6D/zSPq9S9dQ2NPFSQwGvxBCYaBYLPlCWuPh9/t1MRS8iUaR8i+a6w7aps+B4N0S1TYP/R+Q==", + "cpu": [ + "arm" + ], + "dev": true, + "license": "MPL-2.0", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 12.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/parcel" + } + }, + "node_modules/lightningcss-linux-arm64-gnu": { + "version": "1.30.1", + "resolved": "https://registry.npmjs.org/lightningcss-linux-arm64-gnu/-/lightningcss-linux-arm64-gnu-1.30.1.tgz", + "integrity": "sha512-gB72maP8rmrKsnKYy8XUuXi/4OctJiuQjcuqWNlJQ6jZiWqtPvqFziskH3hnajfvKB27ynbVCucKSm2rkQp4Bw==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MPL-2.0", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 12.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/parcel" + } + }, + "node_modules/lightningcss-linux-arm64-musl": { + "version": "1.30.1", + "resolved": "https://registry.npmjs.org/lightningcss-linux-arm64-musl/-/lightningcss-linux-arm64-musl-1.30.1.tgz", + "integrity": "sha512-jmUQVx4331m6LIX+0wUhBbmMX7TCfjF5FoOH6SD1CttzuYlGNVpA7QnrmLxrsub43ClTINfGSYyHe2HWeLl5CQ==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MPL-2.0", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 12.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/parcel" + } + }, + "node_modules/lightningcss-linux-x64-gnu": { + "version": "1.30.1", + "resolved": "https://registry.npmjs.org/lightningcss-linux-x64-gnu/-/lightningcss-linux-x64-gnu-1.30.1.tgz", + "integrity": "sha512-piWx3z4wN8J8z3+O5kO74+yr6ze/dKmPnI7vLqfSqI8bccaTGY5xiSGVIJBDd5K5BHlvVLpUB3S2YCfelyJ1bw==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MPL-2.0", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 12.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/parcel" + } + }, + "node_modules/lightningcss-linux-x64-musl": { + "version": "1.30.1", + "resolved": "https://registry.npmjs.org/lightningcss-linux-x64-musl/-/lightningcss-linux-x64-musl-1.30.1.tgz", + "integrity": "sha512-rRomAK7eIkL+tHY0YPxbc5Dra2gXlI63HL+v1Pdi1a3sC+tJTcFrHX+E86sulgAXeI7rSzDYhPSeHHjqFhqfeQ==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MPL-2.0", + "optional": true, + "os": [ + "linux" + ], + "engines": { + "node": ">= 12.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/parcel" + } + }, + "node_modules/lightningcss-win32-arm64-msvc": { + "version": "1.30.1", + "resolved": "https://registry.npmjs.org/lightningcss-win32-arm64-msvc/-/lightningcss-win32-arm64-msvc-1.30.1.tgz", + "integrity": "sha512-mSL4rqPi4iXq5YVqzSsJgMVFENoa4nGTT/GjO2c0Yl9OuQfPsIfncvLrEW6RbbB24WtZ3xP/2CCmI3tNkNV4oA==", + "cpu": [ + "arm64" + ], + "dev": true, + "license": "MPL-2.0", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">= 12.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/parcel" + } + }, + "node_modules/lightningcss-win32-x64-msvc": { + "version": "1.30.1", + "resolved": "https://registry.npmjs.org/lightningcss-win32-x64-msvc/-/lightningcss-win32-x64-msvc-1.30.1.tgz", + "integrity": "sha512-PVqXh48wh4T53F/1CCu8PIPCxLzWyCnn/9T5W1Jpmdy5h9Cwd+0YQS6/LwhHXSafuc61/xg9Lv5OrCby6a++jg==", + "cpu": [ + "x64" + ], + "dev": true, + "license": "MPL-2.0", + "optional": true, + "os": [ + "win32" + ], + "engines": { + "node": ">= 12.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/parcel" + } + }, + "node_modules/lilconfig": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/lilconfig/-/lilconfig-2.1.0.tgz", + "integrity": "sha512-utWOt/GHzuUxnLKxB6dk81RoOeoNeHgbrXiuGk4yyF5qlRz+iIVWu56E2fqGHFrXz0QNUhLB/8nKqvRH66JKGQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=10" + } + }, + "node_modules/locate-character": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/locate-character/-/locate-character-3.0.0.tgz", + "integrity": "sha512-SW13ws7BjaeJ6p7Q6CO2nchbYEc3X3J6WrmTTDto7yMPqVSZTUyY5Tjbid+Ab8gLnATtygYtiDIJGQRRn2ZOiA==", + "license": "MIT" + }, + "node_modules/locate-path": { + "version": "6.0.0", + "resolved": "https://registry.npmjs.org/locate-path/-/locate-path-6.0.0.tgz", + "integrity": "sha512-iPZK6eYjbxRu3uB4/WZ3EsEIMJFMqAoopl3R+zuq0UjcAm/MO6KCweDgPfP3elTztoKP3KtnVHxTn2NHBSDVUw==", + "dev": true, + "license": "MIT", + "dependencies": { + "p-locate": "^5.0.0" + }, + "engines": { + "node": ">=10" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/lodash": { + "version": "4.17.21", + "resolved": "https://registry.npmjs.org/lodash/-/lodash-4.17.21.tgz", + "integrity": "sha512-v2kDEe57lecTulaDIuNTPy3Ry4gLGJ6Z1O3vE1krgXZNrsQ+LFTGHVxVjcXPs17LhbZVGedAJv8XZ1tvj5FvSg==", + "dev": true, + "license": "MIT" + }, + "node_modules/lodash.castarray": { + "version": "4.4.0", + "resolved": "https://registry.npmjs.org/lodash.castarray/-/lodash.castarray-4.4.0.tgz", + "integrity": "sha512-aVx8ztPv7/2ULbArGJ2Y42bG1mEQ5mGjpdvrbJcJFU3TbYybe+QlLS4pst9zV52ymy2in1KpFPiZnAOATxD4+Q==", + "dev": true, + "license": "MIT" + }, + "node_modules/lodash.isplainobject": { + "version": "4.0.6", + "resolved": "https://registry.npmjs.org/lodash.isplainobject/-/lodash.isplainobject-4.0.6.tgz", + "integrity": "sha512-oSXzaWypCMHkPC3NvBEaPHf0KsA5mvPrOPgQWDsbg8n7orZ290M0BmC/jgRZ4vcJ6DTAhjrsSYgdsW/F+MFOBA==", + "dev": true, + "license": "MIT" + }, + "node_modules/lodash.merge": { + "version": "4.6.2", + "resolved": "https://registry.npmjs.org/lodash.merge/-/lodash.merge-4.6.2.tgz", + "integrity": "sha512-0KpjqXRVvrYyCsX1swR/XTK0va6VQkQM6MNo7PqW77ByjAhoARA8EfrP1N4+KlKj8YS0ZUCtRT/YUuhyYDujIQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/longest-streak": { + "version": "3.1.0", + "resolved": "https://registry.npmjs.org/longest-streak/-/longest-streak-3.1.0.tgz", + "integrity": "sha512-9Ri+o0JYgehTaVBBDoMqIl8GXtbWg711O3srftcHhZ0dqnETqLaoIK0x17fUw9rFSlK/0NlsKe0Ahhyl5pXE2g==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/loupe": { + "version": "3.1.4", + "resolved": "https://registry.npmjs.org/loupe/-/loupe-3.1.4.tgz", + "integrity": "sha512-wJzkKwJrheKtknCOKNEtDK4iqg/MxmZheEMtSTYvnzRdEYaZzmgH976nenp8WdJRdx5Vc1X/9MO0Oszl6ezeXg==", + "dev": true, + "license": "MIT" + }, + "node_modules/lower-case": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/lower-case/-/lower-case-2.0.2.tgz", + "integrity": "sha512-7fm3l3NAF9WfN6W3JOmf5drwpVqX78JtoGJ3A6W0a6ZnldM41w2fV5D490psKFTpMds8TJse/eHLFFsNHHjHgg==", + "dev": true, + "license": "MIT", + "dependencies": { + "tslib": "^2.0.3" + } + }, + "node_modules/lowlight": { + "version": "3.3.0", + "resolved": "https://registry.npmjs.org/lowlight/-/lowlight-3.3.0.tgz", + "integrity": "sha512-0JNhgFoPvP6U6lE/UdVsSq99tn6DhjjpAj5MxG49ewd2mOBVtwWYIT8ClyABhq198aXXODMU6Ox8DrGy/CpTZQ==", + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "devlop": "^1.0.0", + "highlight.js": "~11.11.0" + }, + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/lz-string": { + "version": "1.5.0", + "resolved": "https://registry.npmjs.org/lz-string/-/lz-string-1.5.0.tgz", + "integrity": "sha512-h5bgJWpxJNswbU7qCrV0tIKQCaS3blPDrqKWx+QxzuzL1zGUzij9XCWLrSLsJPu5t+eWA/ycetzYAO5IOMcWAQ==", + "dev": true, + "license": "MIT", + "bin": { + "lz-string": "bin/bin.js" + } + }, + "node_modules/magic-string": { + "version": "0.30.17", + "resolved": "https://registry.npmjs.org/magic-string/-/magic-string-0.30.17.tgz", + "integrity": "sha512-sNPKHvyjVf7gyjwS4xGTaW/mCnF8wnjtifKBEhxfZ7E/S8tQ0rssrwGNn6q8JH/ohItJfSQp9mBtQYuTlH5QnA==", + "license": "MIT", + "dependencies": { + "@jridgewell/sourcemap-codec": "^1.5.0" + } + }, + "node_modules/markdown-table": { + "version": "3.0.4", + "resolved": "https://registry.npmjs.org/markdown-table/-/markdown-table-3.0.4.tgz", + "integrity": "sha512-wiYz4+JrLyb/DqW2hkFJxP7Vd7JuTDm77fvbM8VfEQdmSMqcImWeeRbHwZjBjIFki/VaMK2BhFi7oUUZeM5bqw==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/mdast-util-find-and-replace": { + "version": "3.0.2", + "resolved": "https://registry.npmjs.org/mdast-util-find-and-replace/-/mdast-util-find-and-replace-3.0.2.tgz", + "integrity": "sha512-Tmd1Vg/m3Xz43afeNxDIhWRtFZgM2VLyaf4vSTYwudTyeuTneoL3qtWMA5jeLyz/O1vDJmmV4QuScFCA2tBPwg==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "escape-string-regexp": "^5.0.0", + "unist-util-is": "^6.0.0", + "unist-util-visit-parents": "^6.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-find-and-replace/node_modules/escape-string-regexp": { + "version": "5.0.0", + "resolved": "https://registry.npmjs.org/escape-string-regexp/-/escape-string-regexp-5.0.0.tgz", + "integrity": "sha512-/veY75JbMK4j1yjvuUxuVsiS/hr/4iHs9FTT6cgTexxdE0Ly/glccBAkloH/DofkjRbZU3bnoj38mOmhkZ0lHw==", + "license": "MIT", + "engines": { + "node": ">=12" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/mdast-util-from-markdown": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/mdast-util-from-markdown/-/mdast-util-from-markdown-2.0.2.tgz", + "integrity": "sha512-uZhTV/8NBuw0WHkPTrCqDOl0zVe1BIng5ZtHoDk49ME1qqcjYmmLmOf0gELgcRMxN4w2iuIeVso5/6QymSrgmA==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "@types/unist": "^3.0.0", + "decode-named-character-reference": "^1.0.0", + "devlop": "^1.0.0", + "mdast-util-to-string": "^4.0.0", + "micromark": "^4.0.0", + "micromark-util-decode-numeric-character-reference": "^2.0.0", + "micromark-util-decode-string": "^2.0.0", + "micromark-util-normalize-identifier": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0", + "unist-util-stringify-position": "^4.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-from-markdown/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "license": "MIT" + }, + "node_modules/mdast-util-from-markdown/node_modules/unist-util-stringify-position": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/unist-util-stringify-position/-/unist-util-stringify-position-4.0.0.tgz", + "integrity": "sha512-0ASV06AAoKCDkS2+xw5RXJywruurpbC4JZSm7nr7MOt1ojAzvyyaO+UxZf18j8FCF6kmzCZKcAgN/yu2gm2XgQ==", + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-gfm": { + "version": "3.1.0", + "resolved": "https://registry.npmjs.org/mdast-util-gfm/-/mdast-util-gfm-3.1.0.tgz", + "integrity": "sha512-0ulfdQOM3ysHhCJ1p06l0b0VKlhU0wuQs3thxZQagjcjPrlFRqY215uZGHHJan9GEAXd9MbfPjFJz+qMkVR6zQ==", + "license": "MIT", + "dependencies": { + "mdast-util-from-markdown": "^2.0.0", + "mdast-util-gfm-autolink-literal": "^2.0.0", + "mdast-util-gfm-footnote": "^2.0.0", + "mdast-util-gfm-strikethrough": "^2.0.0", + "mdast-util-gfm-table": "^2.0.0", + "mdast-util-gfm-task-list-item": "^2.0.0", + "mdast-util-to-markdown": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-gfm-autolink-literal": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/mdast-util-gfm-autolink-literal/-/mdast-util-gfm-autolink-literal-2.0.1.tgz", + "integrity": "sha512-5HVP2MKaP6L+G6YaxPNjuL0BPrq9orG3TsrZ9YXbA3vDw/ACI4MEsnoDpn6ZNm7GnZgtAcONJyPhOP8tNJQavQ==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "ccount": "^2.0.0", + "devlop": "^1.0.0", + "mdast-util-find-and-replace": "^3.0.0", + "micromark-util-character": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-gfm-footnote": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/mdast-util-gfm-footnote/-/mdast-util-gfm-footnote-2.1.0.tgz", + "integrity": "sha512-sqpDWlsHn7Ac9GNZQMeUzPQSMzR6Wv0WKRNvQRg0KqHh02fpTz69Qc1QSseNX29bhz1ROIyNyxExfawVKTm1GQ==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "devlop": "^1.1.0", + "mdast-util-from-markdown": "^2.0.0", + "mdast-util-to-markdown": "^2.0.0", + "micromark-util-normalize-identifier": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-gfm-strikethrough": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/mdast-util-gfm-strikethrough/-/mdast-util-gfm-strikethrough-2.0.0.tgz", + "integrity": "sha512-mKKb915TF+OC5ptj5bJ7WFRPdYtuHv0yTRxK2tJvi+BDqbkiG7h7u/9SI89nRAYcmap2xHQL9D+QG/6wSrTtXg==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "mdast-util-from-markdown": "^2.0.0", + "mdast-util-to-markdown": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-gfm-table": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/mdast-util-gfm-table/-/mdast-util-gfm-table-2.0.0.tgz", + "integrity": "sha512-78UEvebzz/rJIxLvE7ZtDd/vIQ0RHv+3Mh5DR96p7cS7HsBhYIICDBCu8csTNWNO6tBWfqXPWekRuj2FNOGOZg==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "devlop": "^1.0.0", + "markdown-table": "^3.0.0", + "mdast-util-from-markdown": "^2.0.0", + "mdast-util-to-markdown": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-gfm-task-list-item": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/mdast-util-gfm-task-list-item/-/mdast-util-gfm-task-list-item-2.0.0.tgz", + "integrity": "sha512-IrtvNvjxC1o06taBAVJznEnkiHxLFTzgonUdy8hzFVeDun0uTjxxrRGVaNFqkU1wJR3RBPEfsxmU6jDWPofrTQ==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "devlop": "^1.0.0", + "mdast-util-from-markdown": "^2.0.0", + "mdast-util-to-markdown": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-math": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/mdast-util-math/-/mdast-util-math-3.0.0.tgz", + "integrity": "sha512-Tl9GBNeG/AhJnQM221bJR2HPvLOSnLE/T9cJI9tlc6zwQk2nPk/4f0cHkOdEixQPC/j8UtKDdITswvLAy1OZ1w==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "@types/mdast": "^4.0.0", + "devlop": "^1.0.0", + "longest-streak": "^3.0.0", + "mdast-util-from-markdown": "^2.0.0", + "mdast-util-to-markdown": "^2.1.0", + "unist-util-remove-position": "^5.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-newline-to-break": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/mdast-util-newline-to-break/-/mdast-util-newline-to-break-2.0.0.tgz", + "integrity": "sha512-MbgeFca0hLYIEx/2zGsszCSEJJ1JSCdiY5xQxRcLDDGa8EPvlLPupJ4DSajbMPAnC0je8jfb9TiUATnxxrHUog==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "mdast-util-find-and-replace": "^3.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-phrasing": { + "version": "4.1.0", + "resolved": "https://registry.npmjs.org/mdast-util-phrasing/-/mdast-util-phrasing-4.1.0.tgz", + "integrity": "sha512-TqICwyvJJpBwvGAMZjj4J2n0X8QWp21b9l0o7eXyVJ25YNWYbJDVIyD1bZXE6WtV6RmKJVYmQAKWa0zWOABz2w==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "unist-util-is": "^6.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-to-hast": { + "version": "13.2.0", + "resolved": "https://registry.npmjs.org/mdast-util-to-hast/-/mdast-util-to-hast-13.2.0.tgz", + "integrity": "sha512-QGYKEuUsYT9ykKBCMOEDLsU5JRObWQusAolFMeko/tYPufNkRffBAQjIE+99jbA87xv6FgmjLtwjh9wBWajwAA==", + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "@types/mdast": "^4.0.0", + "@ungap/structured-clone": "^1.0.0", + "devlop": "^1.0.0", + "micromark-util-sanitize-uri": "^2.0.0", + "trim-lines": "^3.0.0", + "unist-util-position": "^5.0.0", + "unist-util-visit": "^5.0.0", + "vfile": "^6.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-to-markdown": { + "version": "2.1.2", + "resolved": "https://registry.npmjs.org/mdast-util-to-markdown/-/mdast-util-to-markdown-2.1.2.tgz", + "integrity": "sha512-xj68wMTvGXVOKonmog6LwyJKrYXZPvlwabaryTjLh9LuvovB/KAH+kvi8Gjj+7rJjsFi23nkUxRQv1KqSroMqA==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "@types/unist": "^3.0.0", + "longest-streak": "^3.0.0", + "mdast-util-phrasing": "^4.0.0", + "mdast-util-to-string": "^4.0.0", + "micromark-util-classify-character": "^2.0.0", + "micromark-util-decode-string": "^2.0.0", + "unist-util-visit": "^5.0.0", + "zwitch": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdast-util-to-markdown/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "license": "MIT" + }, + "node_modules/mdast-util-to-string": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/mdast-util-to-string/-/mdast-util-to-string-4.0.0.tgz", + "integrity": "sha512-0H44vDimn51F0YwvxSJSm0eCDOJTRlmN0R1yBh4HLj9wiV1Dn0QoXGbvFAWj2hSItVTlCmBF1hqKlIyUBVFLPg==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdsvex": { + "version": "0.12.6", + "resolved": "https://registry.npmjs.org/mdsvex/-/mdsvex-0.12.6.tgz", + "integrity": "sha512-pupx2gzWh3hDtm/iDW4WuCpljmyHbHi34r7ktOqpPGvyiM4MyfNgdJ3qMizXdgCErmvYC9Nn/qyjePy+4ss9Wg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.4", + "@types/unist": "^2.0.3", + "prism-svelte": "^0.4.7", + "prismjs": "^1.17.1", + "unist-util-visit": "^2.0.1", + "vfile-message": "^2.0.4" + }, + "peerDependencies": { + "svelte": "^3.56.0 || ^4.0.0 || ^5.0.0-next.120" + } + }, + "node_modules/mdsvex/node_modules/unist-util-is": { + "version": "4.1.0", + "resolved": "https://registry.npmjs.org/unist-util-is/-/unist-util-is-4.1.0.tgz", + "integrity": "sha512-ZOQSsnce92GrxSqlnEEseX0gi7GH9zTJZ0p9dtu87WRb/37mMPO2Ilx1s/t9vBHrFhbgweUwb+t7cIn5dxPhZg==", + "dev": true, + "license": "MIT", + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdsvex/node_modules/unist-util-visit": { + "version": "2.0.3", + "resolved": "https://registry.npmjs.org/unist-util-visit/-/unist-util-visit-2.0.3.tgz", + "integrity": "sha512-iJ4/RczbJMkD0712mGktuGpm/U4By4FfDonL7N/9tATGIF4imikjOuagyMY53tnZq3NP6BcmlrHhEKAfGWjh7Q==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/unist": "^2.0.0", + "unist-util-is": "^4.0.0", + "unist-util-visit-parents": "^3.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/mdsvex/node_modules/unist-util-visit-parents": { + "version": "3.1.1", + "resolved": "https://registry.npmjs.org/unist-util-visit-parents/-/unist-util-visit-parents-3.1.1.tgz", + "integrity": "sha512-1KROIZWo6bcMrZEwiH2UrXDyalAa0uqzWCxCJj6lPOvTve2WkfgCytoDTPaMnodXh1WrXOq0haVYHj99ynJlsg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/unist": "^2.0.0", + "unist-util-is": "^4.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/merge2": { + "version": "1.4.1", + "resolved": "https://registry.npmjs.org/merge2/-/merge2-1.4.1.tgz", + "integrity": "sha512-8q7VEgMJW4J8tcfVPy8g09NcQwZdbwFEqhe/WZkoIzjn/3TGDwtOCYtXGxA3O8tPzpczCCDgv+P2P5y00ZJOOg==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">= 8" + } + }, + "node_modules/micromark": { + "version": "4.0.2", + "resolved": "https://registry.npmjs.org/micromark/-/micromark-4.0.2.tgz", + "integrity": "sha512-zpe98Q6kvavpCr1NPVSCMebCKfD7CA2NqZ+rykeNhONIJBpc1tFKt9hucLGwha3jNTNI8lHpctWJWoimVF4PfA==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "@types/debug": "^4.0.0", + "debug": "^4.0.0", + "decode-named-character-reference": "^1.0.0", + "devlop": "^1.0.0", + "micromark-core-commonmark": "^2.0.0", + "micromark-factory-space": "^2.0.0", + "micromark-util-character": "^2.0.0", + "micromark-util-chunked": "^2.0.0", + "micromark-util-combine-extensions": "^2.0.0", + "micromark-util-decode-numeric-character-reference": "^2.0.0", + "micromark-util-encode": "^2.0.0", + "micromark-util-normalize-identifier": "^2.0.0", + "micromark-util-resolve-all": "^2.0.0", + "micromark-util-sanitize-uri": "^2.0.0", + "micromark-util-subtokenize": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + } + }, + "node_modules/micromark-core-commonmark": { + "version": "2.0.3", + "resolved": "https://registry.npmjs.org/micromark-core-commonmark/-/micromark-core-commonmark-2.0.3.tgz", + "integrity": "sha512-RDBrHEMSxVFLg6xvnXmb1Ayr2WzLAWjeSATAoxwKYJV94TeNavgoIdA0a9ytzDSVzBy2YKFK+emCPOEibLeCrg==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "decode-named-character-reference": "^1.0.0", + "devlop": "^1.0.0", + "micromark-factory-destination": "^2.0.0", + "micromark-factory-label": "^2.0.0", + "micromark-factory-space": "^2.0.0", + "micromark-factory-title": "^2.0.0", + "micromark-factory-whitespace": "^2.0.0", + "micromark-util-character": "^2.0.0", + "micromark-util-chunked": "^2.0.0", + "micromark-util-classify-character": "^2.0.0", + "micromark-util-html-tag-name": "^2.0.0", + "micromark-util-normalize-identifier": "^2.0.0", + "micromark-util-resolve-all": "^2.0.0", + "micromark-util-subtokenize": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + } + }, + "node_modules/micromark-extension-gfm": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/micromark-extension-gfm/-/micromark-extension-gfm-3.0.0.tgz", + "integrity": "sha512-vsKArQsicm7t0z2GugkCKtZehqUm31oeGBV/KVSorWSy8ZlNAv7ytjFhvaryUiCUJYqs+NoE6AFhpQvBTM6Q4w==", + "license": "MIT", + "dependencies": { + "micromark-extension-gfm-autolink-literal": "^2.0.0", + "micromark-extension-gfm-footnote": "^2.0.0", + "micromark-extension-gfm-strikethrough": "^2.0.0", + "micromark-extension-gfm-table": "^2.0.0", + "micromark-extension-gfm-tagfilter": "^2.0.0", + "micromark-extension-gfm-task-list-item": "^2.0.0", + "micromark-util-combine-extensions": "^2.0.0", + "micromark-util-types": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/micromark-extension-gfm-autolink-literal": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/micromark-extension-gfm-autolink-literal/-/micromark-extension-gfm-autolink-literal-2.1.0.tgz", + "integrity": "sha512-oOg7knzhicgQ3t4QCjCWgTmfNhvQbDDnJeVu9v81r7NltNCVmhPy1fJRX27pISafdjL+SVc4d3l48Gb6pbRypw==", + "license": "MIT", + "dependencies": { + "micromark-util-character": "^2.0.0", + "micromark-util-sanitize-uri": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/micromark-extension-gfm-footnote": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/micromark-extension-gfm-footnote/-/micromark-extension-gfm-footnote-2.1.0.tgz", + "integrity": "sha512-/yPhxI1ntnDNsiHtzLKYnE3vf9JZ6cAisqVDauhp4CEHxlb4uoOTxOCJ+9s51bIB8U1N1FJ1RXOKTIlD5B/gqw==", + "license": "MIT", + "dependencies": { + "devlop": "^1.0.0", + "micromark-core-commonmark": "^2.0.0", + "micromark-factory-space": "^2.0.0", + "micromark-util-character": "^2.0.0", + "micromark-util-normalize-identifier": "^2.0.0", + "micromark-util-sanitize-uri": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/micromark-extension-gfm-strikethrough": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/micromark-extension-gfm-strikethrough/-/micromark-extension-gfm-strikethrough-2.1.0.tgz", + "integrity": "sha512-ADVjpOOkjz1hhkZLlBiYA9cR2Anf8F4HqZUO6e5eDcPQd0Txw5fxLzzxnEkSkfnD0wziSGiv7sYhk/ktvbf1uw==", + "license": "MIT", + "dependencies": { + "devlop": "^1.0.0", + "micromark-util-chunked": "^2.0.0", + "micromark-util-classify-character": "^2.0.0", + "micromark-util-resolve-all": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/micromark-extension-gfm-table": { + "version": "2.1.1", + "resolved": "https://registry.npmjs.org/micromark-extension-gfm-table/-/micromark-extension-gfm-table-2.1.1.tgz", + "integrity": "sha512-t2OU/dXXioARrC6yWfJ4hqB7rct14e8f7m0cbI5hUmDyyIlwv5vEtooptH8INkbLzOatzKuVbQmAYcbWoyz6Dg==", + "license": "MIT", + "dependencies": { + "devlop": "^1.0.0", + "micromark-factory-space": "^2.0.0", + "micromark-util-character": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/micromark-extension-gfm-tagfilter": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/micromark-extension-gfm-tagfilter/-/micromark-extension-gfm-tagfilter-2.0.0.tgz", + "integrity": "sha512-xHlTOmuCSotIA8TW1mDIM6X2O1SiX5P9IuDtqGonFhEK0qgRI4yeC6vMxEV2dgyr2TiD+2PQ10o+cOhdVAcwfg==", + "license": "MIT", + "dependencies": { + "micromark-util-types": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/micromark-extension-gfm-task-list-item": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/micromark-extension-gfm-task-list-item/-/micromark-extension-gfm-task-list-item-2.1.0.tgz", + "integrity": "sha512-qIBZhqxqI6fjLDYFTBIa4eivDMnP+OZqsNwmQ3xNLE4Cxwc+zfQEfbs6tzAo2Hjq+bh6q5F+Z8/cksrLFYWQQw==", + "license": "MIT", + "dependencies": { + "devlop": "^1.0.0", + "micromark-factory-space": "^2.0.0", + "micromark-util-character": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/micromark-extension-math": { + "version": "3.1.0", + "resolved": "https://registry.npmjs.org/micromark-extension-math/-/micromark-extension-math-3.1.0.tgz", + "integrity": "sha512-lvEqd+fHjATVs+2v/8kg9i5Q0AP2k85H0WUOwpIVvUML8BapsMvh1XAogmQjOCsLpoKRCVQqEkQBB3NhVBcsOg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/katex": "^0.16.0", + "devlop": "^1.0.0", + "katex": "^0.16.0", + "micromark-factory-space": "^2.0.0", + "micromark-util-character": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/micromark-factory-destination": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-factory-destination/-/micromark-factory-destination-2.0.1.tgz", + "integrity": "sha512-Xe6rDdJlkmbFRExpTOmRj9N3MaWmbAgdpSrBQvCFqhezUn4AHqJHbaEnfbVYYiexVSs//tqOdY/DxhjdCiJnIA==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "micromark-util-character": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + } + }, + "node_modules/micromark-factory-label": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-factory-label/-/micromark-factory-label-2.0.1.tgz", + "integrity": "sha512-VFMekyQExqIW7xIChcXn4ok29YE3rnuyveW3wZQWWqF4Nv9Wk5rgJ99KzPvHjkmPXF93FXIbBp6YdW3t71/7Vg==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "devlop": "^1.0.0", + "micromark-util-character": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + } + }, + "node_modules/micromark-factory-space": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-factory-space/-/micromark-factory-space-2.0.1.tgz", + "integrity": "sha512-zRkxjtBxxLd2Sc0d+fbnEunsTj46SWXgXciZmHq0kDYGnck/ZSGj9/wULTV95uoeYiK5hRXP2mJ98Uo4cq/LQg==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "micromark-util-character": "^2.0.0", + "micromark-util-types": "^2.0.0" + } + }, + "node_modules/micromark-factory-title": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-factory-title/-/micromark-factory-title-2.0.1.tgz", + "integrity": "sha512-5bZ+3CjhAd9eChYTHsjy6TGxpOFSKgKKJPJxr293jTbfry2KDoWkhBb6TcPVB4NmzaPhMs1Frm9AZH7OD4Cjzw==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "micromark-factory-space": "^2.0.0", + "micromark-util-character": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + } + }, + "node_modules/micromark-factory-whitespace": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-factory-whitespace/-/micromark-factory-whitespace-2.0.1.tgz", + "integrity": "sha512-Ob0nuZ3PKt/n0hORHyvoD9uZhr+Za8sFoP+OnMcnWK5lngSzALgQYKMr9RJVOWLqQYuyn6ulqGWSXdwf6F80lQ==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "micromark-factory-space": "^2.0.0", + "micromark-util-character": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + } + }, + "node_modules/micromark-util-character": { + "version": "2.1.1", + "resolved": "https://registry.npmjs.org/micromark-util-character/-/micromark-util-character-2.1.1.tgz", + "integrity": "sha512-wv8tdUTJ3thSFFFJKtpYKOYiGP2+v96Hvk4Tu8KpCAsTMs6yi+nVmGh1syvSCsaxz45J6Jbw+9DD6g97+NV67Q==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + } + }, + "node_modules/micromark-util-chunked": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-util-chunked/-/micromark-util-chunked-2.0.1.tgz", + "integrity": "sha512-QUNFEOPELfmvv+4xiNg2sRYeS/P84pTW0TCgP5zc9FpXetHY0ab7SxKyAQCNCc1eK0459uoLI1y5oO5Vc1dbhA==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "micromark-util-symbol": "^2.0.0" + } + }, + "node_modules/micromark-util-classify-character": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-util-classify-character/-/micromark-util-classify-character-2.0.1.tgz", + "integrity": "sha512-K0kHzM6afW/MbeWYWLjoHQv1sgg2Q9EccHEDzSkxiP/EaagNzCm7T/WMKZ3rjMbvIpvBiZgwR3dKMygtA4mG1Q==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "micromark-util-character": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + } + }, + "node_modules/micromark-util-combine-extensions": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-util-combine-extensions/-/micromark-util-combine-extensions-2.0.1.tgz", + "integrity": "sha512-OnAnH8Ujmy59JcyZw8JSbK9cGpdVY44NKgSM7E9Eh7DiLS2E9RNQf0dONaGDzEG9yjEl5hcqeIsj4hfRkLH/Bg==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "micromark-util-chunked": "^2.0.0", + "micromark-util-types": "^2.0.0" + } + }, + "node_modules/micromark-util-decode-numeric-character-reference": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/micromark-util-decode-numeric-character-reference/-/micromark-util-decode-numeric-character-reference-2.0.2.tgz", + "integrity": "sha512-ccUbYk6CwVdkmCQMyr64dXz42EfHGkPQlBj5p7YVGzq8I7CtjXZJrubAYezf7Rp+bjPseiROqe7G6foFd+lEuw==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "micromark-util-symbol": "^2.0.0" + } + }, + "node_modules/micromark-util-decode-string": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-util-decode-string/-/micromark-util-decode-string-2.0.1.tgz", + "integrity": "sha512-nDV/77Fj6eH1ynwscYTOsbK7rR//Uj0bZXBwJZRfaLEJ1iGBR6kIfNmlNqaqJf649EP0F3NWNdeJi03elllNUQ==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "decode-named-character-reference": "^1.0.0", + "micromark-util-character": "^2.0.0", + "micromark-util-decode-numeric-character-reference": "^2.0.0", + "micromark-util-symbol": "^2.0.0" + } + }, + "node_modules/micromark-util-encode": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-util-encode/-/micromark-util-encode-2.0.1.tgz", + "integrity": "sha512-c3cVx2y4KqUnwopcO9b/SCdo2O67LwJJ/UyqGfbigahfegL9myoEFoDYZgkT7f36T0bLrM9hZTAaAyH+PCAXjw==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT" + }, + "node_modules/micromark-util-html-tag-name": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-util-html-tag-name/-/micromark-util-html-tag-name-2.0.1.tgz", + "integrity": "sha512-2cNEiYDhCWKI+Gs9T0Tiysk136SnR13hhO8yW6BGNyhOC4qYFnwF1nKfD3HFAIXA5c45RrIG1ub11GiXeYd1xA==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT" + }, + "node_modules/micromark-util-normalize-identifier": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-util-normalize-identifier/-/micromark-util-normalize-identifier-2.0.1.tgz", + "integrity": "sha512-sxPqmo70LyARJs0w2UclACPUUEqltCkJ6PhKdMIDuJ3gSf/Q+/GIe3WKl0Ijb/GyH9lOpUkRAO2wp0GVkLvS9Q==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "micromark-util-symbol": "^2.0.0" + } + }, + "node_modules/micromark-util-resolve-all": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-util-resolve-all/-/micromark-util-resolve-all-2.0.1.tgz", + "integrity": "sha512-VdQyxFWFT2/FGJgwQnJYbe1jjQoNTS4RjglmSjTUlpUMa95Htx9NHeYW4rGDJzbjvCsl9eLjMQwGeElsqmzcHg==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "micromark-util-types": "^2.0.0" + } + }, + "node_modules/micromark-util-sanitize-uri": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-util-sanitize-uri/-/micromark-util-sanitize-uri-2.0.1.tgz", + "integrity": "sha512-9N9IomZ/YuGGZZmQec1MbgxtlgougxTodVwDzzEouPKo3qFWvymFHWcnDi2vzV1ff6kas9ucW+o3yzJK9YB1AQ==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "micromark-util-character": "^2.0.0", + "micromark-util-encode": "^2.0.0", + "micromark-util-symbol": "^2.0.0" + } + }, + "node_modules/micromark-util-subtokenize": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/micromark-util-subtokenize/-/micromark-util-subtokenize-2.1.0.tgz", + "integrity": "sha512-XQLu552iSctvnEcgXw6+Sx75GflAPNED1qx7eBJ+wydBb2KCbRZe+NwvIEEMM83uml1+2WSXpBAcp9IUCgCYWA==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT", + "dependencies": { + "devlop": "^1.0.0", + "micromark-util-chunked": "^2.0.0", + "micromark-util-symbol": "^2.0.0", + "micromark-util-types": "^2.0.0" + } + }, + "node_modules/micromark-util-symbol": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/micromark-util-symbol/-/micromark-util-symbol-2.0.1.tgz", + "integrity": "sha512-vs5t8Apaud9N28kgCrRUdEed4UJ+wWNvicHLPxCa9ENlYuAY31M0ETy5y1vA33YoNPDFTghEbnh6efaE8h4x0Q==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT" + }, + "node_modules/micromark-util-types": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/micromark-util-types/-/micromark-util-types-2.0.2.tgz", + "integrity": "sha512-Yw0ECSpJoViF1qTU4DC6NwtC4aWGt1EkzaQB8KPPyCRR8z9TWeV0HbEFGTO+ZY1wB22zmxnJqhPyTpOVCpeHTA==", + "funding": [ + { + "type": "GitHub Sponsors", + "url": "https://github.com/sponsors/unifiedjs" + }, + { + "type": "OpenCollective", + "url": "https://opencollective.com/unified" + } + ], + "license": "MIT" + }, + "node_modules/micromatch": { + "version": "4.0.8", + "resolved": "https://registry.npmjs.org/micromatch/-/micromatch-4.0.8.tgz", + "integrity": "sha512-PXwfBhYu0hBCPw8Dn0E+WDYb7af3dSLVWKi3HGv84IdF4TyFoC0ysxFd0Goxw7nSv4T/PzEJQxsYsEiFCKo2BA==", + "dev": true, + "license": "MIT", + "dependencies": { + "braces": "^3.0.3", + "picomatch": "^2.3.1" + }, + "engines": { + "node": ">=8.6" + } + }, + "node_modules/micromatch/node_modules/picomatch": { + "version": "2.3.1", + "resolved": "https://registry.npmjs.org/picomatch/-/picomatch-2.3.1.tgz", + "integrity": "sha512-JU3teHTNjmE2VCGFzuY8EXzCDVwEqB2a8fsIvwaStHhAWJEeVd1o1QD80CU6+ZdEXXSLbSsuLwJjkCBWqRQUVA==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=8.6" + }, + "funding": { + "url": "https://github.com/sponsors/jonschlinkert" + } + }, + "node_modules/min-indent": { + "version": "1.0.1", + "resolved": "https://registry.npmjs.org/min-indent/-/min-indent-1.0.1.tgz", + "integrity": "sha512-I9jwMn07Sy/IwOj3zVkVik2JTvgpaykDZEigL6Rx6N9LbMywwUSMtxET+7lVoDLLd3O3IXwJwvuuns8UB/HeAg==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=4" + } + }, + "node_modules/mini-svg-data-uri": { + "version": "1.4.4", + "resolved": "https://registry.npmjs.org/mini-svg-data-uri/-/mini-svg-data-uri-1.4.4.tgz", + "integrity": "sha512-r9deDe9p5FJUPZAk3A59wGH7Ii9YrjjWw0jmw/liSbHl2CHiyXj6FcDXDu2K3TjVAXqiJdaw3xxwlZZr9E6nHg==", + "dev": true, + "license": "MIT", + "bin": { + "mini-svg-data-uri": "cli.js" + } + }, + "node_modules/minimatch": { + "version": "3.1.2", + "resolved": "https://registry.npmjs.org/minimatch/-/minimatch-3.1.2.tgz", + "integrity": "sha512-J7p63hRiAjw1NDEww1W7i37+ByIrOWO5XQQAzZ3VOcL0PNybwpfmV/N05zFAzwQ9USyEcX6t3UO+K5aqBQOIHw==", + "dev": true, + "license": "ISC", + "dependencies": { + "brace-expansion": "^1.1.7" + }, + "engines": { + "node": "*" + } + }, + "node_modules/minipass": { + "version": "7.1.2", + "resolved": "https://registry.npmjs.org/minipass/-/minipass-7.1.2.tgz", + "integrity": "sha512-qOOzS1cBTWYF4BH8fVePDBOO9iptMnGUEZwNc/cMWnTV2nVLZ7VoNWEPHkYczZA0pdoA7dl6e7FL659nX9S2aw==", + "dev": true, + "license": "ISC", + "engines": { + "node": ">=16 || 14 >=14.17" + } + }, + "node_modules/minizlib": { + "version": "3.0.2", + "resolved": "https://registry.npmjs.org/minizlib/-/minizlib-3.0.2.tgz", + "integrity": "sha512-oG62iEk+CYt5Xj2YqI5Xi9xWUeZhDI8jjQmC5oThVH5JGCTgIjr7ciJDzC7MBzYd//WvR1OTmP5Q38Q8ShQtVA==", + "dev": true, + "license": "MIT", + "dependencies": { + "minipass": "^7.1.2" + }, + "engines": { + "node": ">= 18" + } + }, + "node_modules/mkdirp": { + "version": "3.0.1", + "resolved": "https://registry.npmjs.org/mkdirp/-/mkdirp-3.0.1.tgz", + "integrity": "sha512-+NsyUUAZDmo6YVHzL/stxSu3t9YS1iljliy3BSDrXJ/dkn1KYdmtZODGGjLcc9XLgVVpH4KshHB8XmZgMhaBXg==", + "dev": true, + "license": "MIT", + "bin": { + "mkdirp": "dist/cjs/src/bin.js" + }, + "engines": { + "node": ">=10" + }, + "funding": { + "url": "https://github.com/sponsors/isaacs" + } + }, + "node_modules/mode-watcher": { + "version": "1.1.0", + "resolved": "https://registry.npmjs.org/mode-watcher/-/mode-watcher-1.1.0.tgz", + "integrity": "sha512-mUT9RRGPDYenk59qJauN1rhsIMKBmWA3xMF+uRwE8MW/tjhaDSCCARqkSuDTq8vr4/2KcAxIGVjACxTjdk5C3g==", + "license": "MIT", + "dependencies": { + "runed": "^0.25.0", + "svelte-toolbelt": "^0.7.1" + }, + "peerDependencies": { + "svelte": "^5.27.0" + } + }, + "node_modules/mri": { + "version": "1.2.0", + "resolved": "https://registry.npmjs.org/mri/-/mri-1.2.0.tgz", + "integrity": "sha512-tzzskb3bG8LvYGFF/mDTpq3jpI6Q9wc3LEmBaghu+DdCssd1FakN7Bc0hVNmEyGq1bq3RgfkCb3cmQLpNPOroA==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=4" + } + }, + "node_modules/mrmime": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/mrmime/-/mrmime-2.0.1.tgz", + "integrity": "sha512-Y3wQdFg2Va6etvQ5I82yUhGdsKrcYox6p7FfL1LbK2J4V01F9TGlepTIhnK24t7koZibmg82KGglhA1XK5IsLQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=10" + } + }, + "node_modules/ms": { + "version": "2.1.3", + "resolved": "https://registry.npmjs.org/ms/-/ms-2.1.3.tgz", + "integrity": "sha512-6FlzubTLZG3J2a/NVCAleEhjzq5oxgHyaCU9yYXvcLsvoVaHJq/s5xXI6/XXP6tz7R9xAOtHnSO/tXtF3WRTlA==", + "license": "MIT" + }, + "node_modules/nanoid": { + "version": "3.3.11", + "resolved": "https://registry.npmjs.org/nanoid/-/nanoid-3.3.11.tgz", + "integrity": "sha512-N8SpfPUnUp1bK+PMYW8qSWdl9U+wwNWI4QKxOYDy9JAro3WMX7p2OeVRF9v+347pnakNevPmiHhNmZ2HbFA76w==", + "dev": true, + "funding": [ + { + "type": "github", + "url": "https://github.com/sponsors/ai" + } + ], + "license": "MIT", + "bin": { + "nanoid": "bin/nanoid.cjs" + }, + "engines": { + "node": "^10 || ^12 || ^13.7 || ^14 || >=15.0.1" + } + }, + "node_modules/natural-compare": { + "version": "1.4.0", + "resolved": "https://registry.npmjs.org/natural-compare/-/natural-compare-1.4.0.tgz", + "integrity": "sha512-OWND8ei3VtNC9h7V60qff3SVobHr996CTwgxubgyQYEpg290h9J0buyECNNJexkFm5sOajh5G116RYA1c8ZMSw==", + "dev": true, + "license": "MIT" + }, + "node_modules/no-case": { + "version": "3.0.4", + "resolved": "https://registry.npmjs.org/no-case/-/no-case-3.0.4.tgz", + "integrity": "sha512-fgAN3jGAh+RoxUGZHTSOLJIqUc2wmoBwGR4tbpNAKmmovFoWq0OdRkb0VkldReO2a2iBT/OEulG9XSUc10r3zg==", + "dev": true, + "license": "MIT", + "dependencies": { + "lower-case": "^2.0.2", + "tslib": "^2.0.3" + } + }, + "node_modules/open": { + "version": "8.4.2", + "resolved": "https://registry.npmjs.org/open/-/open-8.4.2.tgz", + "integrity": "sha512-7x81NCL719oNbsq/3mh+hVrAWmFuEYUqrq/Iw3kUzH8ReypT9QQ0BLoJS7/G9k6N81XjW4qHWtjWwe/9eLy1EQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "define-lazy-prop": "^2.0.0", + "is-docker": "^2.1.1", + "is-wsl": "^2.2.0" + }, + "engines": { + "node": ">=12" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/optionator": { + "version": "0.9.4", + "resolved": "https://registry.npmjs.org/optionator/-/optionator-0.9.4.tgz", + "integrity": "sha512-6IpQ7mKUxRcZNLIObR0hz7lxsapSSIYNZJwXPGeF0mTVqGKFIXj1DQcMoT22S3ROcLyY/rz0PWaWZ9ayWmad9g==", + "dev": true, + "license": "MIT", + "dependencies": { + "deep-is": "^0.1.3", + "fast-levenshtein": "^2.0.6", + "levn": "^0.4.1", + "prelude-ls": "^1.2.1", + "type-check": "^0.4.0", + "word-wrap": "^1.2.5" + }, + "engines": { + "node": ">= 0.8.0" + } + }, + "node_modules/p-limit": { + "version": "3.1.0", + "resolved": "https://registry.npmjs.org/p-limit/-/p-limit-3.1.0.tgz", + "integrity": "sha512-TYOanM3wGwNGsZN2cVTYPArw454xnXj5qmWF1bEoAc4+cU/ol7GVh7odevjp1FNHduHc3KZMcFduxU5Xc6uJRQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "yocto-queue": "^0.1.0" + }, + "engines": { + "node": ">=10" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/p-locate": { + "version": "5.0.0", + "resolved": "https://registry.npmjs.org/p-locate/-/p-locate-5.0.0.tgz", + "integrity": "sha512-LaNjtRWUBY++zB5nE/NwcaoMylSPk+S+ZHNB1TzdbMJMny6dynpAGt7X/tl/QYq3TIeE6nxHppbo2LGymrG5Pw==", + "dev": true, + "license": "MIT", + "dependencies": { + "p-limit": "^3.0.2" + }, + "engines": { + "node": ">=10" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/parent-module": { + "version": "1.0.1", + "resolved": "https://registry.npmjs.org/parent-module/-/parent-module-1.0.1.tgz", + "integrity": "sha512-GQ2EWRpQV8/o+Aw8YqtfZZPfNRWZYkbidE9k5rpl/hC3vtHHBfGm2Ifi6qWV+coDGkrUKZAxE3Lot5kcsRlh+g==", + "dev": true, + "license": "MIT", + "dependencies": { + "callsites": "^3.0.0" + }, + "engines": { + "node": ">=6" + } + }, + "node_modules/parse5": { + "version": "7.3.0", + "resolved": "https://registry.npmjs.org/parse5/-/parse5-7.3.0.tgz", + "integrity": "sha512-IInvU7fabl34qmi9gY8XOVxhYyMyuH2xUNpb2q8/Y+7552KlejkRvqvD19nMoUW/uQGGbqNpA6Tufu5FL5BZgw==", + "dev": true, + "license": "MIT", + "dependencies": { + "entities": "^6.0.0" + }, + "funding": { + "url": "https://github.com/inikulin/parse5?sponsor=1" + } + }, + "node_modules/pascal-case": { + "version": "3.1.2", + "resolved": "https://registry.npmjs.org/pascal-case/-/pascal-case-3.1.2.tgz", + "integrity": "sha512-uWlGT3YSnK9x3BQJaOdcZwrnV6hPpd8jFH1/ucpiLRPh/2zCVJKS19E4GvYHvaCcACn3foXZ0cLB9Wrx1KGe5g==", + "dev": true, + "license": "MIT", + "dependencies": { + "no-case": "^3.0.4", + "tslib": "^2.0.3" + } + }, + "node_modules/path-exists": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/path-exists/-/path-exists-4.0.0.tgz", + "integrity": "sha512-ak9Qy5Q7jYb2Wwcey5Fpvg2KoAc/ZIhLSLOSBmRmygPsGwkVVt0fZa0qrtMz+m6tJTAHfZQ8FnmB4MG4LWy7/w==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/path-key": { + "version": "3.1.1", + "resolved": "https://registry.npmjs.org/path-key/-/path-key-3.1.1.tgz", + "integrity": "sha512-ojmeN0qd+y0jszEtoY48r0Peq5dwMEkIlCOu6Q5f41lfkswXuKtYrhgoTpLnyIcHm24Uhqx+5Tqm2InSwLhE6Q==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/pathe": { + "version": "2.0.3", + "resolved": "https://registry.npmjs.org/pathe/-/pathe-2.0.3.tgz", + "integrity": "sha512-WUjGcAqP1gQacoQe+OBJsFA7Ld4DyXuUIjZ5cc75cLHvJ7dtNsTugphxIADwspS+AraAUePCKrSVtPLFj/F88w==", + "dev": true, + "license": "MIT" + }, + "node_modules/pathval": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/pathval/-/pathval-2.0.1.tgz", + "integrity": "sha512-//nshmD55c46FuFw26xV/xFAaB5HF9Xdap7HJBBnrKdAd6/GxDBaNA1870O79+9ueg61cZLSVc+OaFlfmObYVQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">= 14.16" + } + }, + "node_modules/pdfjs-dist": { + "version": "5.4.54", + "resolved": "https://registry.npmjs.org/pdfjs-dist/-/pdfjs-dist-5.4.54.tgz", + "integrity": "sha512-TBAiTfQw89gU/Z4LW98Vahzd2/LoCFprVGvGbTgFt+QCB1F+woyOPmNNVgLa6djX9Z9GGTnj7qE1UzpOVJiINw==", + "license": "Apache-2.0", + "engines": { + "node": ">=20.16.0 || >=22.3.0" + }, + "optionalDependencies": { + "@napi-rs/canvas": "^0.1.74" + } + }, + "node_modules/picocolors": { + "version": "1.1.1", + "resolved": "https://registry.npmjs.org/picocolors/-/picocolors-1.1.1.tgz", + "integrity": "sha512-xceH2snhtb5M9liqDsmEw56le376mTZkEX/jEb/RxNFyegNul7eNslCXP9FDj/Lcu0X8KEyMceP2ntpaHrDEVA==", + "dev": true, + "license": "ISC" + }, + "node_modules/picomatch": { + "version": "4.0.3", + "resolved": "https://registry.npmjs.org/picomatch/-/picomatch-4.0.3.tgz", + "integrity": "sha512-5gTmgEY/sqK6gFXLIsQNH19lWb4ebPDLA4SdLP7dsWkIXHWlG66oPuVvXSGFPppYZz8ZDZq0dYYrbHfBCVUb1Q==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=12" + }, + "funding": { + "url": "https://github.com/sponsors/jonschlinkert" + } + }, + "node_modules/playwright": { + "version": "1.54.1", + "resolved": "https://registry.npmjs.org/playwright/-/playwright-1.54.1.tgz", + "integrity": "sha512-peWpSwIBmSLi6aW2auvrUtf2DqY16YYcCMO8rTVx486jKmDTJg7UAhyrraP98GB8BoPURZP8+nxO7TSd4cPr5g==", + "dev": true, + "license": "Apache-2.0", + "dependencies": { + "playwright-core": "1.54.1" + }, + "bin": { + "playwright": "cli.js" + }, + "engines": { + "node": ">=18" + }, + "optionalDependencies": { + "fsevents": "2.3.2" + } + }, + "node_modules/playwright-core": { + "version": "1.54.1", + "resolved": "https://registry.npmjs.org/playwright-core/-/playwright-core-1.54.1.tgz", + "integrity": "sha512-Nbjs2zjj0htNhzgiy5wu+3w09YetDx5pkrpI/kZotDlDUaYk0HVA5xrBVPdow4SAUIlhgKcJeJg4GRKW6xHusA==", + "dev": true, + "license": "Apache-2.0", + "bin": { + "playwright-core": "cli.js" + }, + "engines": { + "node": ">=18" + } + }, + "node_modules/postcss": { + "version": "8.5.6", + "resolved": "https://registry.npmjs.org/postcss/-/postcss-8.5.6.tgz", + "integrity": "sha512-3Ybi1tAuwAP9s0r1UQ2J4n5Y0G05bJkpUIO0/bI9MhwmD70S5aTWbXGBwxHrelT+XM1k6dM0pk+SwNkpTRN7Pg==", + "dev": true, + "funding": [ + { + "type": "opencollective", + "url": "https://opencollective.com/postcss/" + }, + { + "type": "tidelift", + "url": "https://tidelift.com/funding/github/npm/postcss" + }, + { + "type": "github", + "url": "https://github.com/sponsors/ai" + } + ], + "license": "MIT", + "dependencies": { + "nanoid": "^3.3.11", + "picocolors": "^1.1.1", + "source-map-js": "^1.2.1" + }, + "engines": { + "node": "^10 || ^12 || >=14" + } + }, + "node_modules/postcss-load-config": { + "version": "3.1.4", + "resolved": "https://registry.npmjs.org/postcss-load-config/-/postcss-load-config-3.1.4.tgz", + "integrity": "sha512-6DiM4E7v4coTE4uzA8U//WhtPwyhiim3eyjEMFCnUpzbrkK9wJHgKDT2mR+HbtSrd/NubVaYTOpSpjUl8NQeRg==", + "dev": true, + "license": "MIT", + "dependencies": { + "lilconfig": "^2.0.5", + "yaml": "^1.10.2" + }, + "engines": { + "node": ">= 10" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/postcss/" + }, + "peerDependencies": { + "postcss": ">=8.0.9", + "ts-node": ">=9.0.0" + }, + "peerDependenciesMeta": { + "postcss": { + "optional": true + }, + "ts-node": { + "optional": true + } + } + }, + "node_modules/postcss-load-config/node_modules/yaml": { + "version": "1.10.2", + "resolved": "https://registry.npmjs.org/yaml/-/yaml-1.10.2.tgz", + "integrity": "sha512-r3vXyErRCYJ7wg28yvBY5VSoAF8ZvlcW9/BwUzEtUsjvX/DKs24dIkuwjtuprwJJHsbyUbLApepYTR1BN4uHrg==", + "dev": true, + "license": "ISC", + "engines": { + "node": ">= 6" + } + }, + "node_modules/postcss-safe-parser": { + "version": "7.0.1", + "resolved": "https://registry.npmjs.org/postcss-safe-parser/-/postcss-safe-parser-7.0.1.tgz", + "integrity": "sha512-0AioNCJZ2DPYz5ABT6bddIqlhgwhpHZ/l65YAYo0BCIn0xiDpsnTHz0gnoTGk0OXZW0JRs+cDwL8u/teRdz+8A==", + "dev": true, + "funding": [ + { + "type": "opencollective", + "url": "https://opencollective.com/postcss/" + }, + { + "type": "tidelift", + "url": "https://tidelift.com/funding/github/npm/postcss-safe-parser" + }, + { + "type": "github", + "url": "https://github.com/sponsors/ai" + } + ], + "license": "MIT", + "engines": { + "node": ">=18.0" + }, + "peerDependencies": { + "postcss": "^8.4.31" + } + }, + "node_modules/postcss-scss": { + "version": "4.0.9", + "resolved": "https://registry.npmjs.org/postcss-scss/-/postcss-scss-4.0.9.tgz", + "integrity": "sha512-AjKOeiwAitL/MXxQW2DliT28EKukvvbEWx3LBmJIRN8KfBGZbRTxNYW0kSqi1COiTZ57nZ9NW06S6ux//N1c9A==", + "dev": true, + "funding": [ + { + "type": "opencollective", + "url": "https://opencollective.com/postcss/" + }, + { + "type": "tidelift", + "url": "https://tidelift.com/funding/github/npm/postcss-scss" + }, + { + "type": "github", + "url": "https://github.com/sponsors/ai" + } + ], + "license": "MIT", + "engines": { + "node": ">=12.0" + }, + "peerDependencies": { + "postcss": "^8.4.29" + } + }, + "node_modules/postcss-selector-parser": { + "version": "6.0.10", + "resolved": "https://registry.npmjs.org/postcss-selector-parser/-/postcss-selector-parser-6.0.10.tgz", + "integrity": "sha512-IQ7TZdoaqbT+LCpShg46jnZVlhWD2w6iQYAcYXfHARZ7X1t/UGhhceQDs5X0cGqKvYlHNOuv7Oa1xmb0oQuA3w==", + "dev": true, + "license": "MIT", + "dependencies": { + "cssesc": "^3.0.0", + "util-deprecate": "^1.0.2" + }, + "engines": { + "node": ">=4" + } + }, + "node_modules/prelude-ls": { + "version": "1.2.1", + "resolved": "https://registry.npmjs.org/prelude-ls/-/prelude-ls-1.2.1.tgz", + "integrity": "sha512-vkcDPrRZo1QZLbn5RLGPpg/WmIQ65qoWWhcGKf/b5eplkkarX0m9z8ppCat4mlOqUsWpyNuYgO3VRyrYHSzX5g==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">= 0.8.0" + } + }, + "node_modules/prettier": { + "version": "3.6.2", + "resolved": "https://registry.npmjs.org/prettier/-/prettier-3.6.2.tgz", + "integrity": "sha512-I7AIg5boAr5R0FFtJ6rCfD+LFsWHp81dolrFD8S79U9tb8Az2nGrJncnMSnys+bpQJfRUzqs9hnA81OAA3hCuQ==", + "dev": true, + "license": "MIT", + "bin": { + "prettier": "bin/prettier.cjs" + }, + "engines": { + "node": ">=14" + }, + "funding": { + "url": "https://github.com/prettier/prettier?sponsor=1" + } + }, + "node_modules/prettier-plugin-svelte": { + "version": "3.4.0", + "resolved": "https://registry.npmjs.org/prettier-plugin-svelte/-/prettier-plugin-svelte-3.4.0.tgz", + "integrity": "sha512-pn1ra/0mPObzqoIQn/vUTR3ZZI6UuZ0sHqMK5x2jMLGrs53h0sXhkVuDcrlssHwIMk7FYrMjHBPoUSyyEEDlBQ==", + "dev": true, + "license": "MIT", + "peerDependencies": { + "prettier": "^3.0.0", + "svelte": "^3.2.0 || ^4.0.0-next.0 || ^5.0.0-next.0" + } + }, + "node_modules/prettier-plugin-tailwindcss": { + "version": "0.6.14", + "resolved": "https://registry.npmjs.org/prettier-plugin-tailwindcss/-/prettier-plugin-tailwindcss-0.6.14.tgz", + "integrity": "sha512-pi2e/+ZygeIqntN+vC573BcW5Cve8zUB0SSAGxqpB4f96boZF4M3phPVoOFCeypwkpRYdi7+jQ5YJJUwrkGUAg==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=14.21.3" + }, + "peerDependencies": { + "@ianvs/prettier-plugin-sort-imports": "*", + "@prettier/plugin-hermes": "*", + "@prettier/plugin-oxc": "*", + "@prettier/plugin-pug": "*", + "@shopify/prettier-plugin-liquid": "*", + "@trivago/prettier-plugin-sort-imports": "*", + "@zackad/prettier-plugin-twig": "*", + "prettier": "^3.0", + "prettier-plugin-astro": "*", + "prettier-plugin-css-order": "*", + "prettier-plugin-import-sort": "*", + "prettier-plugin-jsdoc": "*", + "prettier-plugin-marko": "*", + "prettier-plugin-multiline-arrays": "*", + "prettier-plugin-organize-attributes": "*", + "prettier-plugin-organize-imports": "*", + "prettier-plugin-sort-imports": "*", + "prettier-plugin-style-order": "*", + "prettier-plugin-svelte": "*" + }, + "peerDependenciesMeta": { + "@ianvs/prettier-plugin-sort-imports": { + "optional": true + }, + "@prettier/plugin-hermes": { + "optional": true + }, + "@prettier/plugin-oxc": { + "optional": true + }, + "@prettier/plugin-pug": { + "optional": true + }, + "@shopify/prettier-plugin-liquid": { + "optional": true + }, + "@trivago/prettier-plugin-sort-imports": { + "optional": true + }, + "@zackad/prettier-plugin-twig": { + "optional": true + }, + "prettier-plugin-astro": { + "optional": true + }, + "prettier-plugin-css-order": { + "optional": true + }, + "prettier-plugin-import-sort": { + "optional": true + }, + "prettier-plugin-jsdoc": { + "optional": true + }, + "prettier-plugin-marko": { + "optional": true + }, + "prettier-plugin-multiline-arrays": { + "optional": true + }, + "prettier-plugin-organize-attributes": { + "optional": true + }, + "prettier-plugin-organize-imports": { + "optional": true + }, + "prettier-plugin-sort-imports": { + "optional": true + }, + "prettier-plugin-style-order": { + "optional": true + }, + "prettier-plugin-svelte": { + "optional": true + } + } + }, + "node_modules/pretty-format": { + "version": "27.5.1", + "resolved": "https://registry.npmjs.org/pretty-format/-/pretty-format-27.5.1.tgz", + "integrity": "sha512-Qb1gy5OrP5+zDf2Bvnzdl3jsTf1qXVMazbvCoKhtKqVs4/YK4ozX4gKQJJVyNe+cajNPn0KoC0MC3FUmaHWEmQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "ansi-regex": "^5.0.1", + "ansi-styles": "^5.0.0", + "react-is": "^17.0.1" + }, + "engines": { + "node": "^10.13.0 || ^12.13.0 || ^14.15.0 || >=15.0.0" + } + }, + "node_modules/pretty-format/node_modules/ansi-styles": { + "version": "5.2.0", + "resolved": "https://registry.npmjs.org/ansi-styles/-/ansi-styles-5.2.0.tgz", + "integrity": "sha512-Cxwpt2SfTzTtXcfOlzGEee8O+c+MmUgGrNiBcXnuWxuFJHe6a5Hz7qwhwe5OgaSYI0IJvkLqWX1ASG+cJOkEiA==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=10" + }, + "funding": { + "url": "https://github.com/chalk/ansi-styles?sponsor=1" + } + }, + "node_modules/prism-svelte": { + "version": "0.4.7", + "resolved": "https://registry.npmjs.org/prism-svelte/-/prism-svelte-0.4.7.tgz", + "integrity": "sha512-yABh19CYbM24V7aS7TuPYRNMqthxwbvx6FF/Rw920YbyBWO3tnyPIqRMgHuSVsLmuHkkBS1Akyof463FVdkeDQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/prismjs": { + "version": "1.30.0", + "resolved": "https://registry.npmjs.org/prismjs/-/prismjs-1.30.0.tgz", + "integrity": "sha512-DEvV2ZF2r2/63V+tK8hQvrR2ZGn10srHbXviTlcv7Kpzw8jWiNTqbVgjO3IY8RxrrOUF8VPMQQFysYYYv0YZxw==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=6" + } + }, + "node_modules/prompts": { + "version": "2.4.2", + "resolved": "https://registry.npmjs.org/prompts/-/prompts-2.4.2.tgz", + "integrity": "sha512-NxNv/kLguCA7p3jE8oL2aEBsrJWgAakBpgmgK6lpPWV+WuOmY6r2/zbAVnP+T8bQlA0nzHXSJSJW0Hq7ylaD2Q==", + "dev": true, + "license": "MIT", + "dependencies": { + "kleur": "^3.0.3", + "sisteransi": "^1.0.5" + }, + "engines": { + "node": ">= 6" + } + }, + "node_modules/prompts/node_modules/kleur": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/kleur/-/kleur-3.0.3.tgz", + "integrity": "sha512-eTIzlVOSUR+JxdDFepEYcBMtZ9Qqdef+rnzWdRZuMbOywu5tO2w2N7rqjoANZ5k9vywhL6Br1VRjUIgTQx4E8w==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=6" + } + }, + "node_modules/property-information": { + "version": "7.1.0", + "resolved": "https://registry.npmjs.org/property-information/-/property-information-7.1.0.tgz", + "integrity": "sha512-TwEZ+X+yCJmYfL7TPUOcvBZ4QfoT5YenQiJuX//0th53DE6w0xxLEtfK3iyryQFddXuvkIk51EEgrJQ0WJkOmQ==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/punycode": { + "version": "2.3.1", + "resolved": "https://registry.npmjs.org/punycode/-/punycode-2.3.1.tgz", + "integrity": "sha512-vYt7UD1U9Wg6138shLtLOvdAu+8DsC/ilFtEVHcH+wydcSpNE20AfSOduf6MkRFahL5FY7X1oU7nKVZFtfq8Fg==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=6" + } + }, + "node_modules/queue-microtask": { + "version": "1.2.3", + "resolved": "https://registry.npmjs.org/queue-microtask/-/queue-microtask-1.2.3.tgz", + "integrity": "sha512-NuaNSa6flKT5JaSYQzJok04JzTL1CA6aGhv5rfLW3PgqA+M2ChpZQnAC8h8i4ZFkBS8X5RqkDBHA7r4hej3K9A==", + "dev": true, + "funding": [ + { + "type": "github", + "url": "https://github.com/sponsors/feross" + }, + { + "type": "patreon", + "url": "https://www.patreon.com/feross" + }, + { + "type": "consulting", + "url": "https://feross.org/support" + } + ], + "license": "MIT" + }, + "node_modules/react": { + "version": "19.1.0", + "resolved": "https://registry.npmjs.org/react/-/react-19.1.0.tgz", + "integrity": "sha512-FS+XFBNvn3GTAWq26joslQgWNoFu08F4kl0J4CgdNKADkdSGXQyTCnKteIAJy96Br6YbpEU1LSzV5dYtjMkMDg==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/react-dom": { + "version": "19.1.0", + "resolved": "https://registry.npmjs.org/react-dom/-/react-dom-19.1.0.tgz", + "integrity": "sha512-Xs1hdnE+DyKgeHJeJznQmYMIBG3TKIHJJT95Q58nHLSrElKlGQqDTR2HQ9fx5CN/Gk6Vh/kupBTDLU11/nDk/g==", + "dev": true, + "license": "MIT", + "dependencies": { + "scheduler": "^0.26.0" + }, + "peerDependencies": { + "react": "^19.1.0" + } + }, + "node_modules/react-is": { + "version": "17.0.2", + "resolved": "https://registry.npmjs.org/react-is/-/react-is-17.0.2.tgz", + "integrity": "sha512-w2GsyukL62IJnlaff/nRegPQR94C/XXamvMWmSHRJ4y7Ts/4ocGRmTHvOs8PSE6pB3dWOrD/nueuU5sduBsQ4w==", + "dev": true, + "license": "MIT" + }, + "node_modules/readdirp": { + "version": "4.1.2", + "resolved": "https://registry.npmjs.org/readdirp/-/readdirp-4.1.2.tgz", + "integrity": "sha512-GDhwkLfywWL2s6vEjyhri+eXmfH6j1L7JE27WhqLeYzoh/A3DBaYGEj2H/HFZCn/kMfim73FXxEJTw06WtxQwg==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">= 14.18.0" + }, + "funding": { + "type": "individual", + "url": "https://paulmillr.com/funding/" + } + }, + "node_modules/recast": { + "version": "0.23.11", + "resolved": "https://registry.npmjs.org/recast/-/recast-0.23.11.tgz", + "integrity": "sha512-YTUo+Flmw4ZXiWfQKGcwwc11KnoRAYgzAE2E7mXKCjSviTKShtxBsN6YUUBB2gtaBzKzeKunxhUwNHQuRryhWA==", + "dev": true, + "license": "MIT", + "dependencies": { + "ast-types": "^0.16.1", + "esprima": "~4.0.0", + "source-map": "~0.6.1", + "tiny-invariant": "^1.3.3", + "tslib": "^2.0.1" + }, + "engines": { + "node": ">= 4" + } + }, + "node_modules/redent": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/redent/-/redent-3.0.0.tgz", + "integrity": "sha512-6tDA8g98We0zd0GvVeMT9arEOnTw9qM03L9cJXaCjrip1OO764RDBLBfrB4cwzNGDj5OA5ioymC9GkizgWJDUg==", + "dev": true, + "license": "MIT", + "dependencies": { + "indent-string": "^4.0.0", + "strip-indent": "^3.0.0" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/rehype-highlight": { + "version": "7.0.2", + "resolved": "https://registry.npmjs.org/rehype-highlight/-/rehype-highlight-7.0.2.tgz", + "integrity": "sha512-k158pK7wdC2qL3M5NcZROZ2tR/l7zOzjxXd5VGdcfIyoijjQqpHd3JKtYSBDpDZ38UI2WJWuFAtkMDxmx5kstA==", + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "hast-util-to-text": "^4.0.0", + "lowlight": "^3.0.0", + "unist-util-visit": "^5.0.0", + "vfile": "^6.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/rehype-katex": { + "version": "7.0.1", + "resolved": "https://registry.npmjs.org/rehype-katex/-/rehype-katex-7.0.1.tgz", + "integrity": "sha512-OiM2wrZ/wuhKkigASodFoo8wimG3H12LWQaH8qSPVJn9apWKFSH3YOCtbKpBorTVw/eI7cuT21XBbvwEswbIOA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "@types/katex": "^0.16.0", + "hast-util-from-html-isomorphic": "^2.0.0", + "hast-util-to-text": "^4.0.0", + "katex": "^0.16.0", + "unist-util-visit-parents": "^6.0.0", + "vfile": "^6.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/rehype-stringify": { + "version": "10.0.1", + "resolved": "https://registry.npmjs.org/rehype-stringify/-/rehype-stringify-10.0.1.tgz", + "integrity": "sha512-k9ecfXHmIPuFVI61B9DeLPN0qFHfawM6RsuX48hoqlaKSF61RskNjSm1lI8PhBEM0MRdLxVVm4WmTqJQccH9mA==", + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "hast-util-to-html": "^9.0.0", + "unified": "^11.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/remark": { + "version": "15.0.1", + "resolved": "https://registry.npmjs.org/remark/-/remark-15.0.1.tgz", + "integrity": "sha512-Eht5w30ruCXgFmxVUSlNWQ9iiimq07URKeFS3hNc8cUWy1llX4KDWfyEDZRycMc+znsN9Ux5/tJ/BFdgdOwA3A==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "remark-parse": "^11.0.0", + "remark-stringify": "^11.0.0", + "unified": "^11.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/remark-breaks": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/remark-breaks/-/remark-breaks-4.0.0.tgz", + "integrity": "sha512-IjEjJOkH4FuJvHZVIW0QCDWxcG96kCq7An/KVH2NfJe6rKZU2AsHeB3OEjPNRxi4QC34Xdx7I2KGYn6IpT7gxQ==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "mdast-util-newline-to-break": "^2.0.0", + "unified": "^11.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/remark-gfm": { + "version": "4.0.1", + "resolved": "https://registry.npmjs.org/remark-gfm/-/remark-gfm-4.0.1.tgz", + "integrity": "sha512-1quofZ2RQ9EWdeN34S79+KExV1764+wCUGop5CPL1WGdD0ocPpu91lzPGbwWMECpEpd42kJGQwzRfyov9j4yNg==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "mdast-util-gfm": "^3.0.0", + "micromark-extension-gfm": "^3.0.0", + "remark-parse": "^11.0.0", + "remark-stringify": "^11.0.0", + "unified": "^11.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/remark-html": { + "version": "16.0.1", + "resolved": "https://registry.npmjs.org/remark-html/-/remark-html-16.0.1.tgz", + "integrity": "sha512-B9JqA5i0qZe0Nsf49q3OXyGvyXuZFDzAP2iOFLEumymuYJITVpiH1IgsTEwTpdptDmZlMDMWeDmSawdaJIGCXQ==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "hast-util-sanitize": "^5.0.0", + "hast-util-to-html": "^9.0.0", + "mdast-util-to-hast": "^13.0.0", + "unified": "^11.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/remark-math": { + "version": "6.0.0", + "resolved": "https://registry.npmjs.org/remark-math/-/remark-math-6.0.0.tgz", + "integrity": "sha512-MMqgnP74Igy+S3WwnhQ7kqGlEerTETXMvJhrUzDikVZ2/uogJCb+WHUg97hK9/jcfc0dkD73s3LN8zU49cTEtA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "mdast-util-math": "^3.0.0", + "micromark-extension-math": "^3.0.0", + "unified": "^11.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/remark-parse": { + "version": "11.0.0", + "resolved": "https://registry.npmjs.org/remark-parse/-/remark-parse-11.0.0.tgz", + "integrity": "sha512-FCxlKLNGknS5ba/1lmpYijMUzX2esxW5xQqjWxw2eHFfS2MSdaHVINFmhjo+qN1WhZhNimq0dZATN9pH0IDrpA==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "mdast-util-from-markdown": "^2.0.0", + "micromark-util-types": "^2.0.0", + "unified": "^11.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/remark-rehype": { + "version": "11.1.2", + "resolved": "https://registry.npmjs.org/remark-rehype/-/remark-rehype-11.1.2.tgz", + "integrity": "sha512-Dh7l57ianaEoIpzbp0PC9UKAdCSVklD8E5Rpw7ETfbTl3FqcOOgq5q2LVDhgGCkaBv7p24JXikPdvhhmHvKMsw==", + "license": "MIT", + "dependencies": { + "@types/hast": "^3.0.0", + "@types/mdast": "^4.0.0", + "mdast-util-to-hast": "^13.0.0", + "unified": "^11.0.0", + "vfile": "^6.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/remark-stringify": { + "version": "11.0.0", + "resolved": "https://registry.npmjs.org/remark-stringify/-/remark-stringify-11.0.0.tgz", + "integrity": "sha512-1OSmLd3awB/t8qdoEOMazZkNsfVTeY4fTsgzcQFdXNq8ToTN4ZGwrMnlda4K6smTFKD+GRV6O48i6Z4iKgPPpw==", + "license": "MIT", + "dependencies": { + "@types/mdast": "^4.0.0", + "mdast-util-to-markdown": "^2.0.0", + "unified": "^11.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/resolve-from": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/resolve-from/-/resolve-from-4.0.0.tgz", + "integrity": "sha512-pb/MYmXstAkysRFx8piNI1tGFNQIFA3vkE3Gq4EuA1dF6gHp/+vgZqsCGJapvy8N3Q+4o7FwvquPJcnZ7RYy4g==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=4" + } + }, + "node_modules/reusify": { + "version": "1.1.0", + "resolved": "https://registry.npmjs.org/reusify/-/reusify-1.1.0.tgz", + "integrity": "sha512-g6QUff04oZpHs0eG5p83rFLhHeV00ug/Yf9nZM6fLeUrPguBTkTQOdpAWWspMh55TZfVQDPaN3NQJfbVRAxdIw==", + "dev": true, + "license": "MIT", + "engines": { + "iojs": ">=1.0.0", + "node": ">=0.10.0" + } + }, + "node_modules/rollup": { + "version": "4.45.1", + "resolved": "https://registry.npmjs.org/rollup/-/rollup-4.45.1.tgz", + "integrity": "sha512-4iya7Jb76fVpQyLoiVpzUrsjQ12r3dM7fIVz+4NwoYvZOShknRmiv+iu9CClZml5ZLGb0XMcYLutK6w9tgxHDw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/estree": "1.0.8" + }, + "bin": { + "rollup": "dist/bin/rollup" + }, + "engines": { + "node": ">=18.0.0", + "npm": ">=8.0.0" + }, + "optionalDependencies": { + "@rollup/rollup-android-arm-eabi": "4.45.1", + "@rollup/rollup-android-arm64": "4.45.1", + "@rollup/rollup-darwin-arm64": "4.45.1", + "@rollup/rollup-darwin-x64": "4.45.1", + "@rollup/rollup-freebsd-arm64": "4.45.1", + "@rollup/rollup-freebsd-x64": "4.45.1", + "@rollup/rollup-linux-arm-gnueabihf": "4.45.1", + "@rollup/rollup-linux-arm-musleabihf": "4.45.1", + "@rollup/rollup-linux-arm64-gnu": "4.45.1", + "@rollup/rollup-linux-arm64-musl": "4.45.1", + "@rollup/rollup-linux-loongarch64-gnu": "4.45.1", + "@rollup/rollup-linux-powerpc64le-gnu": "4.45.1", + "@rollup/rollup-linux-riscv64-gnu": "4.45.1", + "@rollup/rollup-linux-riscv64-musl": "4.45.1", + "@rollup/rollup-linux-s390x-gnu": "4.45.1", + "@rollup/rollup-linux-x64-gnu": "4.45.1", + "@rollup/rollup-linux-x64-musl": "4.45.1", + "@rollup/rollup-win32-arm64-msvc": "4.45.1", + "@rollup/rollup-win32-ia32-msvc": "4.45.1", + "@rollup/rollup-win32-x64-msvc": "4.45.1", + "fsevents": "~2.3.2" + } + }, + "node_modules/run-parallel": { + "version": "1.2.0", + "resolved": "https://registry.npmjs.org/run-parallel/-/run-parallel-1.2.0.tgz", + "integrity": "sha512-5l4VyZR86LZ/lDxZTR6jqL8AFE2S0IFLMP26AbjsLVADxHdhB/c0GUsH+y39UfCi3dzz8OlQuPmnaJOMoDHQBA==", + "dev": true, + "funding": [ + { + "type": "github", + "url": "https://github.com/sponsors/feross" + }, + { + "type": "patreon", + "url": "https://www.patreon.com/feross" + }, + { + "type": "consulting", + "url": "https://feross.org/support" + } + ], + "license": "MIT", + "dependencies": { + "queue-microtask": "^1.2.2" + } + }, + "node_modules/runed": { + "version": "0.25.0", + "resolved": "https://registry.npmjs.org/runed/-/runed-0.25.0.tgz", + "integrity": "sha512-7+ma4AG9FT2sWQEA0Egf6mb7PBT2vHyuHail1ie8ropfSjvZGtEAx8YTmUjv/APCsdRRxEVvArNjALk9zFSOrg==", + "funding": [ + "https://github.com/sponsors/huntabyte", + "https://github.com/sponsors/tglide" + ], + "dependencies": { + "esm-env": "^1.0.0" + }, + "peerDependencies": { + "svelte": "^5.7.0" + } + }, + "node_modules/sade": { + "version": "1.8.1", + "resolved": "https://registry.npmjs.org/sade/-/sade-1.8.1.tgz", + "integrity": "sha512-xal3CZX1Xlo/k4ApwCFrHVACi9fBqJ7V+mwhBsuf/1IOKbBy098Fex+Wa/5QMubw09pSZ/u8EY8PWgevJsXp1A==", + "dev": true, + "license": "MIT", + "dependencies": { + "mri": "^1.1.0" + }, + "engines": { + "node": ">=6" + } + }, + "node_modules/scheduler": { + "version": "0.26.0", + "resolved": "https://registry.npmjs.org/scheduler/-/scheduler-0.26.0.tgz", + "integrity": "sha512-NlHwttCI/l5gCPR3D1nNXtWABUmBwvZpEQiD4IXSbIDq8BzLIK/7Ir5gTFSGZDUu37K5cMNp0hFtzO38sC7gWA==", + "dev": true, + "license": "MIT" + }, + "node_modules/semver": { + "version": "7.7.2", + "resolved": "https://registry.npmjs.org/semver/-/semver-7.7.2.tgz", + "integrity": "sha512-RF0Fw+rO5AMf9MAyaRXI4AV0Ulj5lMHqVxxdSgiVbixSCXoEmmX/jk0CuJw4+3SqroYO9VoUh+HcuJivvtJemA==", + "dev": true, + "license": "ISC", + "bin": { + "semver": "bin/semver.js" + }, + "engines": { + "node": ">=10" + } + }, + "node_modules/set-cookie-parser": { + "version": "2.7.1", + "resolved": "https://registry.npmjs.org/set-cookie-parser/-/set-cookie-parser-2.7.1.tgz", + "integrity": "sha512-IOc8uWeOZgnb3ptbCURJWNjWUPcO3ZnTTdzsurqERrP6nPyv+paC55vJM0LpOlT2ne+Ix+9+CRG1MNLlyZ4GjQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/shebang-command": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/shebang-command/-/shebang-command-2.0.0.tgz", + "integrity": "sha512-kHxr2zZpYtdmrN1qDjrrX/Z1rR1kG8Dx+gkpK1G4eXmvXswmcE1hTWBWYUzlraYw1/yZp6YuDY77YtvbN0dmDA==", + "dev": true, + "license": "MIT", + "dependencies": { + "shebang-regex": "^3.0.0" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/shebang-regex": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/shebang-regex/-/shebang-regex-3.0.0.tgz", + "integrity": "sha512-7++dFhtcx3353uBaq8DDR4NuxBetBzC7ZQOhmTQInHEd6bSrXdiEyzCvG07Z44UYdLShWUyXt5M/yhz8ekcb1A==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=8" + } + }, + "node_modules/siginfo": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/siginfo/-/siginfo-2.0.0.tgz", + "integrity": "sha512-ybx0WO1/8bSBLEWXZvEd7gMW3Sn3JFlW3TvX1nREbDLRNQNaeNN8WK0meBwPdAaOI7TtRRRJn/Es1zhrrCHu7g==", + "dev": true, + "license": "ISC" + }, + "node_modules/sirv": { + "version": "3.0.1", + "resolved": "https://registry.npmjs.org/sirv/-/sirv-3.0.1.tgz", + "integrity": "sha512-FoqMu0NCGBLCcAkS1qA+XJIQTR6/JHfQXl+uGteNCQ76T91DMUjPa9xfmeqMY3z80nLSg9yQmNjK0Px6RWsH/A==", + "dev": true, + "license": "MIT", + "dependencies": { + "@polka/url": "^1.0.0-next.24", + "mrmime": "^2.0.0", + "totalist": "^3.0.0" + }, + "engines": { + "node": ">=18" + } + }, + "node_modules/sisteransi": { + "version": "1.0.5", + "resolved": "https://registry.npmjs.org/sisteransi/-/sisteransi-1.0.5.tgz", + "integrity": "sha512-bLGGlR1QxBcynn2d5YmDX4MGjlZvy2MRBDRNHLJ8VI6l6+9FUiyTFNJ0IveOSP0bcXgVDPRcfGqA0pjaqUpfVg==", + "dev": true, + "license": "MIT" + }, + "node_modules/source-map": { + "version": "0.6.1", + "resolved": "https://registry.npmjs.org/source-map/-/source-map-0.6.1.tgz", + "integrity": "sha512-UjgapumWlbMhkBgzT7Ykc5YXUT46F0iKu8SGXq0bcwP5dz/h0Plj6enJqjz1Zbq2l5WaqYnrVbwWOWMyF3F47g==", + "dev": true, + "license": "BSD-3-Clause", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/source-map-js": { + "version": "1.2.1", + "resolved": "https://registry.npmjs.org/source-map-js/-/source-map-js-1.2.1.tgz", + "integrity": "sha512-UXWMKhLOwVKb728IUtQPXxfYU+usdybtUrK/8uGE8CQMvrhOpwvzDBwj0QhSL7MQc7vIsISBG8VQ8+IDQxpfQA==", + "dev": true, + "license": "BSD-3-Clause", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/space-separated-tokens": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/space-separated-tokens/-/space-separated-tokens-2.0.2.tgz", + "integrity": "sha512-PEGlAwrG8yXGXRjW32fGbg66JAlOAwbObuqVoJpv/mRgoWDQfgH1wDPvtzWyUSNAXBGSk8h755YDbbcEy3SH2Q==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/stackback": { + "version": "0.0.2", + "resolved": "https://registry.npmjs.org/stackback/-/stackback-0.0.2.tgz", + "integrity": "sha512-1XMJE5fQo1jGH6Y/7ebnwPOBEkIEnT4QF32d5R1+VXdXveM0IBMJt8zfaxX1P3QhVwrYe+576+jkANtSS2mBbw==", + "dev": true, + "license": "MIT" + }, + "node_modules/std-env": { + "version": "3.9.0", + "resolved": "https://registry.npmjs.org/std-env/-/std-env-3.9.0.tgz", + "integrity": "sha512-UGvjygr6F6tpH7o2qyqR6QYpwraIjKSdtzyBdyytFOHmPZY917kwdwLG0RbOjWOnKmnm3PeHjaoLLMie7kPLQw==", + "dev": true, + "license": "MIT" + }, + "node_modules/storybook": { + "version": "9.0.17", + "resolved": "https://registry.npmjs.org/storybook/-/storybook-9.0.17.tgz", + "integrity": "sha512-O+9jgJ+Trlq9VGD1uY4OBLKQWHHDKM/A/pA8vMW6PVehhGHNvpzcIC1bngr6mL5gGHZP2nBv+9XG8pTMcggMmg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@storybook/global": "^5.0.0", + "@testing-library/jest-dom": "^6.6.3", + "@testing-library/user-event": "^14.6.1", + "@vitest/expect": "3.2.4", + "@vitest/spy": "3.2.4", + "better-opn": "^3.0.2", + "esbuild": "^0.18.0 || ^0.19.0 || ^0.20.0 || ^0.21.0 || ^0.22.0 || ^0.23.0 || ^0.24.0 || ^0.25.0", + "esbuild-register": "^3.5.0", + "recast": "^0.23.5", + "semver": "^7.6.2", + "ws": "^8.18.0" + }, + "bin": { + "storybook": "bin/index.cjs" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/storybook" + }, + "peerDependencies": { + "prettier": "^2 || ^3" + }, + "peerDependenciesMeta": { + "prettier": { + "optional": true + } + } + }, + "node_modules/stringify-entities": { + "version": "4.0.4", + "resolved": "https://registry.npmjs.org/stringify-entities/-/stringify-entities-4.0.4.tgz", + "integrity": "sha512-IwfBptatlO+QCJUo19AqvrPNqlVMpW9YEL2LIVY+Rpv2qsjCGxaDLNRgeGsQWJhfItebuJhsGSLjaBbNSQ+ieg==", + "license": "MIT", + "dependencies": { + "character-entities-html4": "^2.0.0", + "character-entities-legacy": "^3.0.0" + }, + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/strip-ansi": { + "version": "7.1.0", + "resolved": "https://registry.npmjs.org/strip-ansi/-/strip-ansi-7.1.0.tgz", + "integrity": "sha512-iq6eVVI64nQQTRYq2KtEg2d2uU7LElhTJwsH4YzIHZshxlgZms/wIc4VoDQTlG/IvVIrBKG06CrZnp0qv7hkcQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "ansi-regex": "^6.0.1" + }, + "engines": { + "node": ">=12" + }, + "funding": { + "url": "https://github.com/chalk/strip-ansi?sponsor=1" + } + }, + "node_modules/strip-ansi/node_modules/ansi-regex": { + "version": "6.1.0", + "resolved": "https://registry.npmjs.org/ansi-regex/-/ansi-regex-6.1.0.tgz", + "integrity": "sha512-7HSX4QQb4CspciLpVFwyRe79O3xsIZDDLER21kERQ71oaPodF8jL725AgJMFAYbooIqolJoRLuM81SpeUkpkvA==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=12" + }, + "funding": { + "url": "https://github.com/chalk/ansi-regex?sponsor=1" + } + }, + "node_modules/strip-indent": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/strip-indent/-/strip-indent-3.0.0.tgz", + "integrity": "sha512-laJTa3Jb+VQpaC6DseHhF7dXVqHTfJPCRDaEbid/drOhgitgYku/letMUqOXFoWV0zIIUbjpdH2t+tYj4bQMRQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "min-indent": "^1.0.0" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/strip-json-comments": { + "version": "3.1.1", + "resolved": "https://registry.npmjs.org/strip-json-comments/-/strip-json-comments-3.1.1.tgz", + "integrity": "sha512-6fPc+R4ihwqP6N/aIv2f1gMH8lOVtWQHoqC4yK6oSDVVocumAsfCqjkXnqiYMhmMwS/mEHLp7Vehlt3ql6lEig==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=8" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/strip-literal": { + "version": "3.0.0", + "resolved": "https://registry.npmjs.org/strip-literal/-/strip-literal-3.0.0.tgz", + "integrity": "sha512-TcccoMhJOM3OebGhSBEmp3UZ2SfDMZUEBdRA/9ynfLi8yYajyWX3JiXArcJt4Umh4vISpspkQIY8ZZoCqjbviA==", + "dev": true, + "license": "MIT", + "dependencies": { + "js-tokens": "^9.0.1" + }, + "funding": { + "url": "https://github.com/sponsors/antfu" + } + }, + "node_modules/strip-literal/node_modules/js-tokens": { + "version": "9.0.1", + "resolved": "https://registry.npmjs.org/js-tokens/-/js-tokens-9.0.1.tgz", + "integrity": "sha512-mxa9E9ITFOt0ban3j6L5MpjwegGz6lBQmM1IJkWeBZGcMxto50+eWdjC/52xDbS2vy0k7vIMK0Fe2wfL9OQSpQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/style-to-object": { + "version": "1.0.9", + "resolved": "https://registry.npmjs.org/style-to-object/-/style-to-object-1.0.9.tgz", + "integrity": "sha512-G4qppLgKu/k6FwRpHiGiKPaPTFcG3g4wNVX/Qsfu+RqQM30E7Tyu/TEgxcL9PNLF5pdRLwQdE3YKKf+KF2Dzlw==", + "license": "MIT", + "dependencies": { + "inline-style-parser": "0.2.4" + } + }, + "node_modules/supports-color": { + "version": "7.2.0", + "resolved": "https://registry.npmjs.org/supports-color/-/supports-color-7.2.0.tgz", + "integrity": "sha512-qpCAvRl9stuOHveKsn7HncJRvv501qIacKzQlO/+Lwxc9+0q2wLyv4Dfvt80/DPn2pqOBsJdDiogXGR9+OvwRw==", + "dev": true, + "license": "MIT", + "dependencies": { + "has-flag": "^4.0.0" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/svelte": { + "version": "5.36.12", + "resolved": "https://registry.npmjs.org/svelte/-/svelte-5.36.12.tgz", + "integrity": "sha512-c3mWT+b0yBLl3gPGSHiy4pdSQCsPNTjLC0tVoOhrGJ6PPfCzD/RQpAmAfJtQZ304CAae2ph+L3C4aqds3R3seQ==", + "license": "MIT", + "dependencies": { + "@ampproject/remapping": "^2.3.0", + "@jridgewell/sourcemap-codec": "^1.5.0", + "@sveltejs/acorn-typescript": "^1.0.5", + "@types/estree": "^1.0.5", + "acorn": "^8.12.1", + "aria-query": "^5.3.1", + "axobject-query": "^4.1.0", + "clsx": "^2.1.1", + "esm-env": "^1.2.1", + "esrap": "^2.1.0", + "is-reference": "^3.0.3", + "locate-character": "^3.0.0", + "magic-string": "^0.30.11", + "zimmerframe": "^1.1.2" + }, + "engines": { + "node": ">=18" + } + }, + "node_modules/svelte-ast-print": { + "version": "0.4.2", + "resolved": "https://registry.npmjs.org/svelte-ast-print/-/svelte-ast-print-0.4.2.tgz", + "integrity": "sha512-hRHHufbJoArFmDYQKCpCvc0xUuIEfwYksvyLYEQyH+1xb5LD5sM/IthfooCdXZQtOIqXz6xm7NmaqdfwG4kh6w==", + "dev": true, + "funding": [ + { + "type": "github", + "url": "https://github.com/sponsors/xeho91" + }, + { + "type": "opencollective", + "url": "https://opencollective.com/xeho91" + } + ], + "license": "MIT", + "dependencies": { + "esrap": "1.2.2", + "zimmerframe": "1.1.2" + }, + "engines": { + "node": ">=18" + }, + "peerDependencies": { + "svelte": "^5.0.0" + } + }, + "node_modules/svelte-ast-print/node_modules/esrap": { + "version": "1.2.2", + "resolved": "https://registry.npmjs.org/esrap/-/esrap-1.2.2.tgz", + "integrity": "sha512-F2pSJklxx1BlQIQgooczXCPHmcWpn6EsP5oo73LQfonG9fIlIENQ8vMmfGXeojP9MrkzUNAfyU5vdFlR9shHAw==", + "dev": true, + "license": "MIT", + "dependencies": { + "@jridgewell/sourcemap-codec": "^1.4.15", + "@types/estree": "^1.0.1" + } + }, + "node_modules/svelte-check": { + "version": "4.3.0", + "resolved": "https://registry.npmjs.org/svelte-check/-/svelte-check-4.3.0.tgz", + "integrity": "sha512-Iz8dFXzBNAM7XlEIsUjUGQhbEE+Pvv9odb9+0+ITTgFWZBGeJRRYqHUUglwe2EkLD5LIsQaAc4IUJyvtKuOO5w==", + "dev": true, + "license": "MIT", + "dependencies": { + "@jridgewell/trace-mapping": "^0.3.25", + "chokidar": "^4.0.1", + "fdir": "^6.2.0", + "picocolors": "^1.0.0", + "sade": "^1.7.4" + }, + "bin": { + "svelte-check": "bin/svelte-check" + }, + "engines": { + "node": ">= 18.0.0" + }, + "peerDependencies": { + "svelte": "^4.0.0 || ^5.0.0-next.0", + "typescript": ">=5.0.0" + } + }, + "node_modules/svelte-eslint-parser": { + "version": "1.3.0", + "resolved": "https://registry.npmjs.org/svelte-eslint-parser/-/svelte-eslint-parser-1.3.0.tgz", + "integrity": "sha512-VCgMHKV7UtOGcGLGNFSbmdm6kEKjtzo5nnpGU/mnx4OsFY6bZ7QwRF5DUx+Hokw5Lvdyo8dpk8B1m8mliomrNg==", + "dev": true, + "license": "MIT", + "dependencies": { + "eslint-scope": "^8.2.0", + "eslint-visitor-keys": "^4.0.0", + "espree": "^10.0.0", + "postcss": "^8.4.49", + "postcss-scss": "^4.0.9", + "postcss-selector-parser": "^7.0.0" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "url": "https://github.com/sponsors/ota-meshi" + }, + "peerDependencies": { + "svelte": "^3.37.0 || ^4.0.0 || ^5.0.0" + }, + "peerDependenciesMeta": { + "svelte": { + "optional": true + } + } + }, + "node_modules/svelte-eslint-parser/node_modules/postcss-selector-parser": { + "version": "7.1.0", + "resolved": "https://registry.npmjs.org/postcss-selector-parser/-/postcss-selector-parser-7.1.0.tgz", + "integrity": "sha512-8sLjZwK0R+JlxlYcTuVnyT2v+htpdrjDOKuMcOVdYjt52Lh8hWRYpxBPoKx/Zg+bcjc3wx6fmQevMmUztS/ccA==", + "dev": true, + "license": "MIT", + "dependencies": { + "cssesc": "^3.0.0", + "util-deprecate": "^1.0.2" + }, + "engines": { + "node": ">=4" + } + }, + "node_modules/svelte-sonner": { + "version": "1.0.5", + "resolved": "https://registry.npmjs.org/svelte-sonner/-/svelte-sonner-1.0.5.tgz", + "integrity": "sha512-9dpGPFqKb/QWudYqGnEz93vuY+NgCEvyNvxoCLMVGw6sDN/3oVeKV1xiEirW2E1N3vJEyj5imSBNOGltQHA7mg==", + "license": "MIT", + "dependencies": { + "runed": "^0.28.0" + }, + "peerDependencies": { + "svelte": "^5.0.0" + } + }, + "node_modules/svelte-sonner/node_modules/runed": { + "version": "0.28.0", + "resolved": "https://registry.npmjs.org/runed/-/runed-0.28.0.tgz", + "integrity": "sha512-k2xx7RuO9hWcdd9f+8JoBeqWtYrm5CALfgpkg2YDB80ds/QE4w0qqu34A7fqiAwiBBSBQOid7TLxwxVC27ymWQ==", + "funding": [ + "https://github.com/sponsors/huntabyte", + "https://github.com/sponsors/tglide" + ], + "license": "MIT", + "dependencies": { + "esm-env": "^1.0.0" + }, + "peerDependencies": { + "svelte": "^5.7.0" + } + }, + "node_modules/svelte-toolbelt": { + "version": "0.7.1", + "resolved": "https://registry.npmjs.org/svelte-toolbelt/-/svelte-toolbelt-0.7.1.tgz", + "integrity": "sha512-HcBOcR17Vx9bjaOceUvxkY3nGmbBmCBBbuWLLEWO6jtmWH8f/QoWmbyUfQZrpDINH39en1b8mptfPQT9VKQ1xQ==", + "funding": [ + "https://github.com/sponsors/huntabyte" + ], + "dependencies": { + "clsx": "^2.1.1", + "runed": "^0.23.2", + "style-to-object": "^1.0.8" + }, + "engines": { + "node": ">=18", + "pnpm": ">=8.7.0" + }, + "peerDependencies": { + "svelte": "^5.0.0" + } + }, + "node_modules/svelte-toolbelt/node_modules/runed": { + "version": "0.23.4", + "resolved": "https://registry.npmjs.org/runed/-/runed-0.23.4.tgz", + "integrity": "sha512-9q8oUiBYeXIDLWNK5DfCWlkL0EW3oGbk845VdKlPeia28l751VpfesaB/+7pI6rnbx1I6rqoZ2fZxptOJLxILA==", + "funding": [ + "https://github.com/sponsors/huntabyte", + "https://github.com/sponsors/tglide" + ], + "dependencies": { + "esm-env": "^1.0.0" + }, + "peerDependencies": { + "svelte": "^5.7.0" + } + }, + "node_modules/svelte/node_modules/aria-query": { + "version": "5.3.2", + "resolved": "https://registry.npmjs.org/aria-query/-/aria-query-5.3.2.tgz", + "integrity": "sha512-COROpnaoap1E2F000S62r6A60uHZnmlvomhfyT2DlTcrY1OrBKn2UhH7qn5wTC9zMvD0AY7csdPSNwKP+7WiQw==", + "license": "Apache-2.0", + "engines": { + "node": ">= 0.4" + } + }, + "node_modules/svelte/node_modules/esrap": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/esrap/-/esrap-2.1.0.tgz", + "integrity": "sha512-yzmPNpl7TBbMRC5Lj2JlJZNPml0tzqoqP5B1JXycNUwtqma9AKCO0M2wHrdgsHcy1WRW7S9rJknAMtByg3usgA==", + "license": "MIT", + "dependencies": { + "@jridgewell/sourcemap-codec": "^1.4.15" + } + }, + "node_modules/svelte/node_modules/is-reference": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/is-reference/-/is-reference-3.0.3.tgz", + "integrity": "sha512-ixkJoqQvAP88E6wLydLGGqCJsrFUnqoH6HnaczB8XmDH1oaWU+xxdptvikTgaEhtZ53Ky6YXiBuUI2WXLMCwjw==", + "license": "MIT", + "dependencies": { + "@types/estree": "^1.0.6" + } + }, + "node_modules/svelte2tsx": { + "version": "0.7.41", + "resolved": "https://registry.npmjs.org/svelte2tsx/-/svelte2tsx-0.7.41.tgz", + "integrity": "sha512-/TUwpyn/Qc1wcGuayf2GSwvZ7htdAOzpo0JFFm96srKnRXoTD0gy4n06g+XgH8w016S3lPtyFVtFAm+0yJ0BZw==", + "dev": true, + "license": "MIT", + "dependencies": { + "dedent-js": "^1.0.1", + "pascal-case": "^3.1.1" + }, + "peerDependencies": { + "svelte": "^3.55 || ^4.0.0-next.0 || ^4.0 || ^5.0.0-next.0", + "typescript": "^4.9.4 || ^5.0.0" + } + }, + "node_modules/tabbable": { + "version": "6.2.0", + "resolved": "https://registry.npmjs.org/tabbable/-/tabbable-6.2.0.tgz", + "integrity": "sha512-Cat63mxsVJlzYvN51JmVXIgNoUokrIaT2zLclCXjRd8boZ0004U4KCs/sToJ75C6sdlByWxpYnb5Boif1VSFew==", + "dev": true, + "license": "MIT" + }, + "node_modules/tailwind-merge": { + "version": "3.3.1", + "resolved": "https://registry.npmjs.org/tailwind-merge/-/tailwind-merge-3.3.1.tgz", + "integrity": "sha512-gBXpgUm/3rp1lMZZrM/w7D8GKqshif0zAymAhbCyIt8KMe+0v9DQ7cdYLR4FHH/cKpdTXb+A/tKKU3eolfsI+g==", + "dev": true, + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/dcastil" + } + }, + "node_modules/tailwind-variants": { + "version": "1.0.0", + "resolved": "https://registry.npmjs.org/tailwind-variants/-/tailwind-variants-1.0.0.tgz", + "integrity": "sha512-2WSbv4ulEEyuBKomOunut65D8UZwxrHoRfYnxGcQNnHqlSCp2+B7Yz2W+yrNDrxRodOXtGD/1oCcKGNBnUqMqA==", + "dev": true, + "license": "MIT", + "dependencies": { + "tailwind-merge": "3.0.2" + }, + "engines": { + "node": ">=16.x", + "pnpm": ">=7.x" + }, + "peerDependencies": { + "tailwindcss": "*" + } + }, + "node_modules/tailwind-variants/node_modules/tailwind-merge": { + "version": "3.0.2", + "resolved": "https://registry.npmjs.org/tailwind-merge/-/tailwind-merge-3.0.2.tgz", + "integrity": "sha512-l7z+OYZ7mu3DTqrL88RiKrKIqO3NcpEO8V/Od04bNpvk0kiIFndGEoqfuzvj4yuhRkHKjRkII2z+KS2HfPcSxw==", + "dev": true, + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/dcastil" + } + }, + "node_modules/tailwindcss": { + "version": "4.1.11", + "resolved": "https://registry.npmjs.org/tailwindcss/-/tailwindcss-4.1.11.tgz", + "integrity": "sha512-2E9TBm6MDD/xKYe+dvJZAmg3yxIEDNRc0jwlNyDg/4Fil2QcSLjFKGVff0lAf1jjeaArlG/M75Ey/EYr/OJtBA==", + "dev": true, + "license": "MIT" + }, + "node_modules/tapable": { + "version": "2.2.2", + "resolved": "https://registry.npmjs.org/tapable/-/tapable-2.2.2.tgz", + "integrity": "sha512-Re10+NauLTMCudc7T5WLFLAwDhQ0JWdrMK+9B2M8zR5hRExKmsRDCBA7/aV/pNJFltmBFO5BAMlQFi/vq3nKOg==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=6" + } + }, + "node_modules/tar": { + "version": "7.4.3", + "resolved": "https://registry.npmjs.org/tar/-/tar-7.4.3.tgz", + "integrity": "sha512-5S7Va8hKfV7W5U6g3aYxXmlPoZVAwUMy9AOKyF2fVuZa2UD3qZjg578OrLRt8PcNN1PleVaL/5/yYATNL0ICUw==", + "dev": true, + "license": "ISC", + "dependencies": { + "@isaacs/fs-minipass": "^4.0.0", + "chownr": "^3.0.0", + "minipass": "^7.1.2", + "minizlib": "^3.0.1", + "mkdirp": "^3.0.1", + "yallist": "^5.0.0" + }, + "engines": { + "node": ">=18" + } + }, + "node_modules/tiny-invariant": { + "version": "1.3.3", + "resolved": "https://registry.npmjs.org/tiny-invariant/-/tiny-invariant-1.3.3.tgz", + "integrity": "sha512-+FbBPE1o9QAYvviau/qC5SE3caw21q3xkvWKBtja5vgqOWIHHJ3ioaq1VPfn/Szqctz2bU/oYeKd9/z5BL+PVg==", + "dev": true, + "license": "MIT" + }, + "node_modules/tinybench": { + "version": "2.9.0", + "resolved": "https://registry.npmjs.org/tinybench/-/tinybench-2.9.0.tgz", + "integrity": "sha512-0+DUvqWMValLmha6lr4kD8iAMK1HzV0/aKnCtWb9v9641TnP/MFb7Pc2bxoxQjTXAErryXVgUOfv2YqNllqGeg==", + "dev": true, + "license": "MIT" + }, + "node_modules/tinyexec": { + "version": "0.3.2", + "resolved": "https://registry.npmjs.org/tinyexec/-/tinyexec-0.3.2.tgz", + "integrity": "sha512-KQQR9yN7R5+OSwaK0XQoj22pwHoTlgYqmUscPYoknOoWCWfj/5/ABTMRi69FrKU5ffPVh5QcFikpWJI/P1ocHA==", + "dev": true, + "license": "MIT" + }, + "node_modules/tinyglobby": { + "version": "0.2.14", + "resolved": "https://registry.npmjs.org/tinyglobby/-/tinyglobby-0.2.14.tgz", + "integrity": "sha512-tX5e7OM1HnYr2+a2C/4V0htOcSQcoSTH9KgJnVvNm5zm/cyEWKJ7j7YutsH9CxMdtOkkLFy2AHrMci9IM8IPZQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "fdir": "^6.4.4", + "picomatch": "^4.0.2" + }, + "engines": { + "node": ">=12.0.0" + }, + "funding": { + "url": "https://github.com/sponsors/SuperchupuDev" + } + }, + "node_modules/tinypool": { + "version": "1.1.1", + "resolved": "https://registry.npmjs.org/tinypool/-/tinypool-1.1.1.tgz", + "integrity": "sha512-Zba82s87IFq9A9XmjiX5uZA/ARWDrB03OHlq+Vw1fSdt0I+4/Kutwy8BP4Y/y/aORMo61FQ0vIb5j44vSo5Pkg==", + "dev": true, + "license": "MIT", + "engines": { + "node": "^18.0.0 || >=20.0.0" + } + }, + "node_modules/tinyrainbow": { + "version": "2.0.0", + "resolved": "https://registry.npmjs.org/tinyrainbow/-/tinyrainbow-2.0.0.tgz", + "integrity": "sha512-op4nsTR47R6p0vMUUoYl/a+ljLFVtlfaXkLQmqfLR1qHma1h/ysYk4hEXZ880bf2CYgTskvTa/e196Vd5dDQXw==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/tinyspy": { + "version": "4.0.3", + "resolved": "https://registry.npmjs.org/tinyspy/-/tinyspy-4.0.3.tgz", + "integrity": "sha512-t2T/WLB2WRgZ9EpE4jgPJ9w+i66UZfDc8wHh0xrwiRNN+UwH98GIJkTeZqX9rg0i0ptwzqW+uYeIF0T4F8LR7A==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/to-regex-range": { + "version": "5.0.1", + "resolved": "https://registry.npmjs.org/to-regex-range/-/to-regex-range-5.0.1.tgz", + "integrity": "sha512-65P7iz6X5yEr1cwcgvQxbbIw7Uk3gOy5dIdtZ4rDveLqhrdJP+Li/Hx6tyK0NEb+2GCyneCMJiGqrADCSNk8sQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "is-number": "^7.0.0" + }, + "engines": { + "node": ">=8.0" + } + }, + "node_modules/totalist": { + "version": "3.0.1", + "resolved": "https://registry.npmjs.org/totalist/-/totalist-3.0.1.tgz", + "integrity": "sha512-sf4i37nQ2LBx4m3wB74y+ubopq6W/dIzXg0FDGjsYnZHVa1Da8FH853wlL2gtUhg+xJXjfk3kUZS3BRoQeoQBQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=6" + } + }, + "node_modules/trim-lines": { + "version": "3.0.1", + "resolved": "https://registry.npmjs.org/trim-lines/-/trim-lines-3.0.1.tgz", + "integrity": "sha512-kRj8B+YHZCc9kQYdWfJB2/oUl9rA99qbowYYBtr4ui4mZyAQ2JpvVBd/6U2YloATfqBhBTSMhTpgBHtU0Mf3Rg==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/trough": { + "version": "2.2.0", + "resolved": "https://registry.npmjs.org/trough/-/trough-2.2.0.tgz", + "integrity": "sha512-tmMpK00BjZiUyVyvrBK7knerNgmgvcV/KLVyuma/SC+TQN167GrMRciANTz09+k3zW8L8t60jWO1GpfkZdjTaw==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/ts-api-utils": { + "version": "2.1.0", + "resolved": "https://registry.npmjs.org/ts-api-utils/-/ts-api-utils-2.1.0.tgz", + "integrity": "sha512-CUgTZL1irw8u29bzrOD/nH85jqyc74D6SshFgujOIA7osm2Rz7dYH77agkx7H4FBNxDq7Cjf+IjaX/8zwFW+ZQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=18.12" + }, + "peerDependencies": { + "typescript": ">=4.8.4" + } + }, + "node_modules/ts-dedent": { + "version": "2.2.0", + "resolved": "https://registry.npmjs.org/ts-dedent/-/ts-dedent-2.2.0.tgz", + "integrity": "sha512-q5W7tVM71e2xjHZTlgfTDoPF/SmqKG5hddq9SzR49CH2hayqRKJtQ4mtRlSxKaJlR/+9rEM+mnBHf7I2/BQcpQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=6.10" + } + }, + "node_modules/tslib": { + "version": "2.8.1", + "resolved": "https://registry.npmjs.org/tslib/-/tslib-2.8.1.tgz", + "integrity": "sha512-oJFu94HQb+KVduSUQL7wnpmqnfmLsOA/nAh6b6EH0wCEoK0/mPeXU6c3wKDV83MkOuHPRHtSXKKU99IBazS/2w==", + "dev": true, + "license": "0BSD" + }, + "node_modules/tw-animate-css": { + "version": "1.3.5", + "resolved": "https://registry.npmjs.org/tw-animate-css/-/tw-animate-css-1.3.5.tgz", + "integrity": "sha512-t3u+0YNoloIhj1mMXs779P6MO9q3p3mvGn4k1n3nJPqJw/glZcuijG2qTSN4z4mgNRfW5ZC3aXJFLwDtiipZXA==", + "dev": true, + "license": "MIT", + "funding": { + "url": "https://github.com/sponsors/Wombosvideo" + } + }, + "node_modules/type-check": { + "version": "0.4.0", + "resolved": "https://registry.npmjs.org/type-check/-/type-check-0.4.0.tgz", + "integrity": "sha512-XleUoc9uwGXqjWwXaUTZAmzMcFZ5858QA2vvx1Ur5xIcixXIP+8LnFDgRplU30us6teqdlskFfu+ae4K79Ooew==", + "dev": true, + "license": "MIT", + "dependencies": { + "prelude-ls": "^1.2.1" + }, + "engines": { + "node": ">= 0.8.0" + } + }, + "node_modules/type-fest": { + "version": "2.19.0", + "resolved": "https://registry.npmjs.org/type-fest/-/type-fest-2.19.0.tgz", + "integrity": "sha512-RAH822pAdBgcNMAfWnCBU3CFZcfZ/i1eZjwFU/dsLKumyuuP3niueg2UAukXYF0E2AAoc82ZSSf9J0WQBinzHA==", + "dev": true, + "license": "(MIT OR CC0-1.0)", + "engines": { + "node": ">=12.20" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/typescript": { + "version": "5.8.3", + "resolved": "https://registry.npmjs.org/typescript/-/typescript-5.8.3.tgz", + "integrity": "sha512-p1diW6TqL9L07nNxvRMM7hMMw4c5XOo/1ibL4aAIGmSAt9slTE1Xgw5KWuof2uTOvCg9BY7ZRi+GaF+7sfgPeQ==", + "dev": true, + "license": "Apache-2.0", + "bin": { + "tsc": "bin/tsc", + "tsserver": "bin/tsserver" + }, + "engines": { + "node": ">=14.17" + } + }, + "node_modules/typescript-eslint": { + "version": "8.37.0", + "resolved": "https://registry.npmjs.org/typescript-eslint/-/typescript-eslint-8.37.0.tgz", + "integrity": "sha512-TnbEjzkE9EmcO0Q2zM+GE8NQLItNAJpMmED1BdgoBMYNdqMhzlbqfdSwiRlAzEK2pA9UzVW0gzaaIzXWg2BjfA==", + "dev": true, + "license": "MIT", + "dependencies": { + "@typescript-eslint/eslint-plugin": "8.37.0", + "@typescript-eslint/parser": "8.37.0", + "@typescript-eslint/typescript-estree": "8.37.0", + "@typescript-eslint/utils": "8.37.0" + }, + "engines": { + "node": "^18.18.0 || ^20.9.0 || >=21.1.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/typescript-eslint" + }, + "peerDependencies": { + "eslint": "^8.57.0 || ^9.0.0", + "typescript": ">=4.8.4 <5.9.0" + } + }, + "node_modules/undici-types": { + "version": "6.21.0", + "resolved": "https://registry.npmjs.org/undici-types/-/undici-types-6.21.0.tgz", + "integrity": "sha512-iwDZqg0QAGrg9Rav5H4n0M64c3mkR59cJ6wQp+7C4nI0gsmExaedaYLNO44eT4AtBBwjbTiGPMlt2Md0T9H9JQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/unified": { + "version": "11.0.5", + "resolved": "https://registry.npmjs.org/unified/-/unified-11.0.5.tgz", + "integrity": "sha512-xKvGhPWw3k84Qjh8bI3ZeJjqnyadK+GEFtazSfZv/rKeTkTjOJho6mFqh2SM96iIcZokxiOpg78GazTSg8+KHA==", + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0", + "bail": "^2.0.0", + "devlop": "^1.0.0", + "extend": "^3.0.0", + "is-plain-obj": "^4.0.0", + "trough": "^2.0.0", + "vfile": "^6.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/unified/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "license": "MIT" + }, + "node_modules/unist-util-find-after": { + "version": "5.0.0", + "resolved": "https://registry.npmjs.org/unist-util-find-after/-/unist-util-find-after-5.0.0.tgz", + "integrity": "sha512-amQa0Ep2m6hE2g72AugUItjbuM8X8cGQnFoHk0pGfrFeT9GZhzN5SW8nRsiGKK7Aif4CrACPENkA6P/Lw6fHGQ==", + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0", + "unist-util-is": "^6.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/unist-util-find-after/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "license": "MIT" + }, + "node_modules/unist-util-is": { + "version": "6.0.0", + "resolved": "https://registry.npmjs.org/unist-util-is/-/unist-util-is-6.0.0.tgz", + "integrity": "sha512-2qCTHimwdxLfz+YzdGfkqNlH0tLi9xjTnHddPmJwtIG9MGsdbutfTc4P+haPD7l7Cjxf/WZj+we5qfVPvvxfYw==", + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/unist-util-is/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "license": "MIT" + }, + "node_modules/unist-util-position": { + "version": "5.0.0", + "resolved": "https://registry.npmjs.org/unist-util-position/-/unist-util-position-5.0.0.tgz", + "integrity": "sha512-fucsC7HjXvkB5R3kTCO7kUjRdrS0BJt3M/FPxmHMBOm8JQi2BsHAHFsy27E0EolP8rp0NzXsJ+jNPyDWvOJZPA==", + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/unist-util-position/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "license": "MIT" + }, + "node_modules/unist-util-remove-position": { + "version": "5.0.0", + "resolved": "https://registry.npmjs.org/unist-util-remove-position/-/unist-util-remove-position-5.0.0.tgz", + "integrity": "sha512-Hp5Kh3wLxv0PHj9m2yZhhLt58KzPtEYKQQ4yxfYFEO7EvHwzyDYnduhHnY1mDxoqr7VUwVuHXk9RXKIiYS1N8Q==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0", + "unist-util-visit": "^5.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/unist-util-remove-position/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, + "license": "MIT" + }, + "node_modules/unist-util-stringify-position": { + "version": "2.0.3", + "resolved": "https://registry.npmjs.org/unist-util-stringify-position/-/unist-util-stringify-position-2.0.3.tgz", + "integrity": "sha512-3faScn5I+hy9VleOq/qNbAd6pAx7iH5jYBMS9I1HgQVijz/4mv5Bvw5iw1sC/90CODiKo81G/ps8AJrISn687g==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/unist": "^2.0.2" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/unist-util-visit": { + "version": "5.0.0", + "resolved": "https://registry.npmjs.org/unist-util-visit/-/unist-util-visit-5.0.0.tgz", + "integrity": "sha512-MR04uvD+07cwl/yhVuVWAtw+3GOR/knlL55Nd/wAdblk27GCVt3lqpTivy/tkJcZoNPzTwS1Y+KMojlLDhoTzg==", + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0", + "unist-util-is": "^6.0.0", + "unist-util-visit-parents": "^6.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/unist-util-visit-parents": { + "version": "6.0.1", + "resolved": "https://registry.npmjs.org/unist-util-visit-parents/-/unist-util-visit-parents-6.0.1.tgz", + "integrity": "sha512-L/PqWzfTP9lzzEa6CKs0k2nARxTdZduw3zyh8d2NVBnsyvHjSX4TWse388YrrQKbvI8w20fGjGlhgT96WwKykw==", + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0", + "unist-util-is": "^6.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/unist-util-visit-parents/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "license": "MIT" + }, + "node_modules/unist-util-visit/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "license": "MIT" + }, + "node_modules/universalify": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/universalify/-/universalify-2.0.1.tgz", + "integrity": "sha512-gptHNQghINnc/vTGIk0SOFGFNXw7JVrlRUtConJRlvaw6DuX0wO5Jeko9sWrMBhh+PsYAZ7oXAiOnf/UKogyiw==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">= 10.0.0" + } + }, + "node_modules/unplugin": { + "version": "1.16.1", + "resolved": "https://registry.npmjs.org/unplugin/-/unplugin-1.16.1.tgz", + "integrity": "sha512-4/u/j4FrCKdi17jaxuJA0jClGxB1AvU2hw/IuayPc4ay1XGaJs/rbb4v5WKwAjNifjmXK9PIFyuPiaK8azyR9w==", + "dev": true, + "license": "MIT", + "dependencies": { + "acorn": "^8.14.0", + "webpack-virtual-modules": "^0.6.2" + }, + "engines": { + "node": ">=14.0.0" + } + }, + "node_modules/uri-js": { + "version": "4.4.1", + "resolved": "https://registry.npmjs.org/uri-js/-/uri-js-4.4.1.tgz", + "integrity": "sha512-7rKUyy33Q1yc98pQ1DAmLtwX109F7TIfWlW1Ydo8Wl1ii1SeHieeh0HHfPeL2fMXK6z0s8ecKs9frCuLJvndBg==", + "dev": true, + "license": "BSD-2-Clause", + "dependencies": { + "punycode": "^2.1.0" + } + }, + "node_modules/util-deprecate": { + "version": "1.0.2", + "resolved": "https://registry.npmjs.org/util-deprecate/-/util-deprecate-1.0.2.tgz", + "integrity": "sha512-EPD5q1uXyFxJpCrLnCc1nHnq3gOa6DZBocAIiI2TaSCA7VCJ1UJDMagCzIkXNsUYfD1daK//LTEQ8xiIbrHtcw==", + "dev": true, + "license": "MIT" + }, + "node_modules/uuid": { + "version": "13.0.0", + "resolved": "https://registry.npmjs.org/uuid/-/uuid-13.0.0.tgz", + "integrity": "sha512-XQegIaBTVUjSHliKqcnFqYypAd4S+WCYt5NIeRs6w/UAry7z8Y9j5ZwRRL4kzq9U3sD6v+85er9FvkEaBpji2w==", + "dev": true, + "funding": [ + "https://github.com/sponsors/broofa", + "https://github.com/sponsors/ctavan" + ], + "license": "MIT", + "bin": { + "uuid": "dist-node/bin/uuid" + } + }, + "node_modules/vfile": { + "version": "6.0.3", + "resolved": "https://registry.npmjs.org/vfile/-/vfile-6.0.3.tgz", + "integrity": "sha512-KzIbH/9tXat2u30jf+smMwFCsno4wHVdNmzFyL+T/L3UGqqk6JKfVqOFOZEpZSHADH1k40ab6NUIXZq422ov3Q==", + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0", + "vfile-message": "^4.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/vfile-location": { + "version": "5.0.3", + "resolved": "https://registry.npmjs.org/vfile-location/-/vfile-location-5.0.3.tgz", + "integrity": "sha512-5yXvWDEgqeiYiBe1lbxYF7UMAIm/IcopxMHrMQDq3nvKcjPKIhZklUKL+AE7J7uApI4kwe2snsK+eI6UTj9EHg==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0", + "vfile": "^6.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/vfile-location/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "dev": true, + "license": "MIT" + }, + "node_modules/vfile-message": { + "version": "2.0.4", + "resolved": "https://registry.npmjs.org/vfile-message/-/vfile-message-2.0.4.tgz", + "integrity": "sha512-DjssxRGkMvifUOJre00juHoP9DPWuzjxKuMDrhNbk2TdaYYBNMStsNhEOt3idrtI12VQYM/1+iM0KOzXi4pxwQ==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/unist": "^2.0.0", + "unist-util-stringify-position": "^2.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/vfile/node_modules/@types/unist": { + "version": "3.0.3", + "resolved": "https://registry.npmjs.org/@types/unist/-/unist-3.0.3.tgz", + "integrity": "sha512-ko/gIFJRv177XgZsZcBwnqJN5x/Gien8qNOn0D5bQU/zAzVf9Zt3BlcUiLqhV9y4ARk0GbT3tnUiPNgnTXzc/Q==", + "license": "MIT" + }, + "node_modules/vfile/node_modules/unist-util-stringify-position": { + "version": "4.0.0", + "resolved": "https://registry.npmjs.org/unist-util-stringify-position/-/unist-util-stringify-position-4.0.0.tgz", + "integrity": "sha512-0ASV06AAoKCDkS2+xw5RXJywruurpbC4JZSm7nr7MOt1ojAzvyyaO+UxZf18j8FCF6kmzCZKcAgN/yu2gm2XgQ==", + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/vfile/node_modules/vfile-message": { + "version": "4.0.2", + "resolved": "https://registry.npmjs.org/vfile-message/-/vfile-message-4.0.2.tgz", + "integrity": "sha512-jRDZ1IMLttGj41KcZvlrYAaI3CfqpLpfpf+Mfig13viT6NKvRzWZ+lXz0Y5D60w6uJIBAOGq9mSHf0gktF0duw==", + "license": "MIT", + "dependencies": { + "@types/unist": "^3.0.0", + "unist-util-stringify-position": "^4.0.0" + }, + "funding": { + "type": "opencollective", + "url": "https://opencollective.com/unified" + } + }, + "node_modules/vite": { + "version": "7.0.5", + "resolved": "https://registry.npmjs.org/vite/-/vite-7.0.5.tgz", + "integrity": "sha512-1mncVwJxy2C9ThLwz0+2GKZyEXuC3MyWtAAlNftlZZXZDP3AJt5FmwcMit/IGGaNZ8ZOB2BNO/HFUB+CpN0NQw==", + "dev": true, + "license": "MIT", + "dependencies": { + "esbuild": "^0.25.0", + "fdir": "^6.4.6", + "picomatch": "^4.0.2", + "postcss": "^8.5.6", + "rollup": "^4.40.0", + "tinyglobby": "^0.2.14" + }, + "bin": { + "vite": "bin/vite.js" + }, + "engines": { + "node": "^20.19.0 || >=22.12.0" + }, + "funding": { + "url": "https://github.com/vitejs/vite?sponsor=1" + }, + "optionalDependencies": { + "fsevents": "~2.3.3" + }, + "peerDependencies": { + "@types/node": "^20.19.0 || >=22.12.0", + "jiti": ">=1.21.0", + "less": "^4.0.0", + "lightningcss": "^1.21.0", + "sass": "^1.70.0", + "sass-embedded": "^1.70.0", + "stylus": ">=0.54.8", + "sugarss": "^5.0.0", + "terser": "^5.16.0", + "tsx": "^4.8.1", + "yaml": "^2.4.2" + }, + "peerDependenciesMeta": { + "@types/node": { + "optional": true + }, + "jiti": { + "optional": true + }, + "less": { + "optional": true + }, + "lightningcss": { + "optional": true + }, + "sass": { + "optional": true + }, + "sass-embedded": { + "optional": true + }, + "stylus": { + "optional": true + }, + "sugarss": { + "optional": true + }, + "terser": { + "optional": true + }, + "tsx": { + "optional": true + }, + "yaml": { + "optional": true + } + } + }, + "node_modules/vite-node": { + "version": "3.2.4", + "resolved": "https://registry.npmjs.org/vite-node/-/vite-node-3.2.4.tgz", + "integrity": "sha512-EbKSKh+bh1E1IFxeO0pg1n4dvoOTt0UDiXMd/qn++r98+jPO1xtJilvXldeuQ8giIB5IkpjCgMleHMNEsGH6pg==", + "dev": true, + "license": "MIT", + "dependencies": { + "cac": "^6.7.14", + "debug": "^4.4.1", + "es-module-lexer": "^1.7.0", + "pathe": "^2.0.3", + "vite": "^5.0.0 || ^6.0.0 || ^7.0.0-0" + }, + "bin": { + "vite-node": "vite-node.mjs" + }, + "engines": { + "node": "^18.0.0 || ^20.0.0 || >=22.0.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + } + }, + "node_modules/vite-plugin-devtools-json": { + "version": "0.2.1", + "resolved": "https://registry.npmjs.org/vite-plugin-devtools-json/-/vite-plugin-devtools-json-0.2.1.tgz", + "integrity": "sha512-5aiNvf/iLTuLR1dUqoI5CLLGgeK2hd6u+tA+RIp7GUZDyAcM6ECaUEWOOtGpidbcxbkKq++KtmSqA3jhMbPwMA==", + "dev": true, + "license": "MIT", + "dependencies": { + "uuid": "^11.1.0" + }, + "peerDependencies": { + "vite": "^2.7.0 || ^3.0.0 || ^4.0.0 || ^5.0.0 || ^6.0.0 || ^7.0.0" + } + }, + "node_modules/vite-plugin-devtools-json/node_modules/uuid": { + "version": "11.1.0", + "resolved": "https://registry.npmjs.org/uuid/-/uuid-11.1.0.tgz", + "integrity": "sha512-0/A9rDy9P7cJ+8w1c9WD9V//9Wj15Ce2MPz8Ri6032usz+NfePxx5AcN3bN+r6ZL6jEo066/yNYB3tn4pQEx+A==", + "dev": true, + "funding": [ + "https://github.com/sponsors/broofa", + "https://github.com/sponsors/ctavan" + ], + "license": "MIT", + "bin": { + "uuid": "dist/esm/bin/uuid" + } + }, + "node_modules/vite/node_modules/fsevents": { + "version": "2.3.3", + "resolved": "https://registry.npmjs.org/fsevents/-/fsevents-2.3.3.tgz", + "integrity": "sha512-5xoDfX+fL7faATnagmWPpbFtwh/R77WmMMqqHGS65C3vvB0YHrgF+B1YmZ3441tMj5n63k0212XNoJwzlhffQw==", + "dev": true, + "hasInstallScript": true, + "license": "MIT", + "optional": true, + "os": [ + "darwin" + ], + "engines": { + "node": "^8.16.0 || ^10.6.0 || >=11.0.0" + } + }, + "node_modules/vitefu": { + "version": "1.1.1", + "resolved": "https://registry.npmjs.org/vitefu/-/vitefu-1.1.1.tgz", + "integrity": "sha512-B/Fegf3i8zh0yFbpzZ21amWzHmuNlLlmJT6n7bu5e+pCHUKQIfXSYokrqOBGEMMe9UG2sostKQF9mml/vYaWJQ==", + "dev": true, + "license": "MIT", + "workspaces": [ + "tests/deps/*", + "tests/projects/*", + "tests/projects/workspace/packages/*" + ], + "peerDependencies": { + "vite": "^3.0.0 || ^4.0.0 || ^5.0.0 || ^6.0.0 || ^7.0.0-beta.0" + }, + "peerDependenciesMeta": { + "vite": { + "optional": true + } + } + }, + "node_modules/vitest": { + "version": "3.2.4", + "resolved": "https://registry.npmjs.org/vitest/-/vitest-3.2.4.tgz", + "integrity": "sha512-LUCP5ev3GURDysTWiP47wRRUpLKMOfPh+yKTx3kVIEiu5KOMeqzpnYNsKyOoVrULivR8tLcks4+lga33Whn90A==", + "dev": true, + "license": "MIT", + "dependencies": { + "@types/chai": "^5.2.2", + "@vitest/expect": "3.2.4", + "@vitest/mocker": "3.2.4", + "@vitest/pretty-format": "^3.2.4", + "@vitest/runner": "3.2.4", + "@vitest/snapshot": "3.2.4", + "@vitest/spy": "3.2.4", + "@vitest/utils": "3.2.4", + "chai": "^5.2.0", + "debug": "^4.4.1", + "expect-type": "^1.2.1", + "magic-string": "^0.30.17", + "pathe": "^2.0.3", + "picomatch": "^4.0.2", + "std-env": "^3.9.0", + "tinybench": "^2.9.0", + "tinyexec": "^0.3.2", + "tinyglobby": "^0.2.14", + "tinypool": "^1.1.1", + "tinyrainbow": "^2.0.0", + "vite": "^5.0.0 || ^6.0.0 || ^7.0.0-0", + "vite-node": "3.2.4", + "why-is-node-running": "^2.3.0" + }, + "bin": { + "vitest": "vitest.mjs" + }, + "engines": { + "node": "^18.0.0 || ^20.0.0 || >=22.0.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + }, + "peerDependencies": { + "@edge-runtime/vm": "*", + "@types/debug": "^4.1.12", + "@types/node": "^18.0.0 || ^20.0.0 || >=22.0.0", + "@vitest/browser": "3.2.4", + "@vitest/ui": "3.2.4", + "happy-dom": "*", + "jsdom": "*" + }, + "peerDependenciesMeta": { + "@edge-runtime/vm": { + "optional": true + }, + "@types/debug": { + "optional": true + }, + "@types/node": { + "optional": true + }, + "@vitest/browser": { + "optional": true + }, + "@vitest/ui": { + "optional": true + }, + "happy-dom": { + "optional": true + }, + "jsdom": { + "optional": true + } + } + }, + "node_modules/vitest-browser-svelte": { + "version": "0.1.0", + "resolved": "https://registry.npmjs.org/vitest-browser-svelte/-/vitest-browser-svelte-0.1.0.tgz", + "integrity": "sha512-YB6ZUZZQNqU1T9NzvTEDpwpPv35Ng1NZMPBh81zDrLEdOgROGE6nJb79NWb1Eu/a8VkHifqArpOZfJfALge6xQ==", + "dev": true, + "license": "MIT", + "engines": { + "node": "^18.0.0 || >=20.0.0" + }, + "funding": { + "url": "https://opencollective.com/vitest" + }, + "peerDependencies": { + "@vitest/browser": "^2.1.0 || ^3.0.0-0", + "svelte": ">3.0.0", + "vitest": "^2.1.0 || ^3.0.0-0" + } + }, + "node_modules/web-namespaces": { + "version": "2.0.1", + "resolved": "https://registry.npmjs.org/web-namespaces/-/web-namespaces-2.0.1.tgz", + "integrity": "sha512-bKr1DkiNa2krS7qxNtdrtHAmzuYGFQLiQ13TsorsdT6ULTkPLKuu5+GsFpDlg6JFjUTwX2DyhMPG2be8uPrqsQ==", + "dev": true, + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + }, + "node_modules/webpack-virtual-modules": { + "version": "0.6.2", + "resolved": "https://registry.npmjs.org/webpack-virtual-modules/-/webpack-virtual-modules-0.6.2.tgz", + "integrity": "sha512-66/V2i5hQanC51vBQKPH4aI8NMAcBW59FVBs+rC7eGHupMyfn34q7rZIE+ETlJ+XTevqfUhVVBgSUNSW2flEUQ==", + "dev": true, + "license": "MIT" + }, + "node_modules/which": { + "version": "2.0.2", + "resolved": "https://registry.npmjs.org/which/-/which-2.0.2.tgz", + "integrity": "sha512-BLI3Tl1TW3Pvl70l3yq3Y64i+awpwXqsGBYWkkqMtnbXgrMD+yj7rhW0kuEDxzJaYXGjEW5ogapKNMEKNMjibA==", + "dev": true, + "license": "ISC", + "dependencies": { + "isexe": "^2.0.0" + }, + "bin": { + "node-which": "bin/node-which" + }, + "engines": { + "node": ">= 8" + } + }, + "node_modules/why-is-node-running": { + "version": "2.3.0", + "resolved": "https://registry.npmjs.org/why-is-node-running/-/why-is-node-running-2.3.0.tgz", + "integrity": "sha512-hUrmaWBdVDcxvYqnyh09zunKzROWjbZTiNy8dBEjkS7ehEDQibXJ7XvlmtbwuTclUiIyN+CyXQD4Vmko8fNm8w==", + "dev": true, + "license": "MIT", + "dependencies": { + "siginfo": "^2.0.0", + "stackback": "0.0.2" + }, + "bin": { + "why-is-node-running": "cli.js" + }, + "engines": { + "node": ">=8" + } + }, + "node_modules/word-wrap": { + "version": "1.2.5", + "resolved": "https://registry.npmjs.org/word-wrap/-/word-wrap-1.2.5.tgz", + "integrity": "sha512-BN22B5eaMMI9UMtjrGd5g5eCYPpCPDUy0FJXbYsaT5zYxjFOckS53SQDE3pWkVoWpHXVb3BrYcEN4Twa55B5cA==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=0.10.0" + } + }, + "node_modules/ws": { + "version": "8.18.3", + "resolved": "https://registry.npmjs.org/ws/-/ws-8.18.3.tgz", + "integrity": "sha512-PEIGCY5tSlUt50cqyMXfCzX+oOPqN0vuGqWzbcJ2xvnkzkq46oOpz7dQaTDBdfICb4N14+GARUDw2XV2N4tvzg==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=10.0.0" + }, + "peerDependencies": { + "bufferutil": "^4.0.1", + "utf-8-validate": ">=5.0.2" + }, + "peerDependenciesMeta": { + "bufferutil": { + "optional": true + }, + "utf-8-validate": { + "optional": true + } + } + }, + "node_modules/yallist": { + "version": "5.0.0", + "resolved": "https://registry.npmjs.org/yallist/-/yallist-5.0.0.tgz", + "integrity": "sha512-YgvUTfwqyc7UXVMrB+SImsVYSmTS8X/tSrtdNZMImM+n7+QTriRXyXim0mBrTXNeqzVF0KWGgHPeiyViFFrNDw==", + "dev": true, + "license": "BlueOak-1.0.0", + "engines": { + "node": ">=18" + } + }, + "node_modules/yocto-queue": { + "version": "0.1.0", + "resolved": "https://registry.npmjs.org/yocto-queue/-/yocto-queue-0.1.0.tgz", + "integrity": "sha512-rVksvsnNCdJ/ohGc6xgPwyN8eheCxsiLM8mxuE/t/mOVqJewPuO1miLpTHQiRgTKCLexL4MeAFVagts7HmNZ2Q==", + "dev": true, + "license": "MIT", + "engines": { + "node": ">=10" + }, + "funding": { + "url": "https://github.com/sponsors/sindresorhus" + } + }, + "node_modules/zimmerframe": { + "version": "1.1.2", + "resolved": "https://registry.npmjs.org/zimmerframe/-/zimmerframe-1.1.2.tgz", + "integrity": "sha512-rAbqEGa8ovJy4pyBxZM70hg4pE6gDgaQ0Sl9M3enG3I0d6H4XSAM3GeNGLKnsBpuijUow064sf7ww1nutC5/3w==", + "license": "MIT" + }, + "node_modules/zwitch": { + "version": "2.0.4", + "resolved": "https://registry.npmjs.org/zwitch/-/zwitch-2.0.4.tgz", + "integrity": "sha512-bXE4cR/kVZhKZX/RjPEflHaKVhUVl85noU3v6b8apfQEc1x4A+zBxjZ4lN8LqGd6WZ3dl98pY4o717VFmoPp+A==", + "license": "MIT", + "funding": { + "type": "github", + "url": "https://github.com/sponsors/wooorm" + } + } + } } diff --git a/tools/server/webui/package.json b/tools/server/webui/package.json index 8076840324..30c7cd15c6 100644 --- a/tools/server/webui/package.json +++ b/tools/server/webui/package.json @@ -1,66 +1,90 @@ { - "name": "webui", - "private": true, - "version": "0.0.0", - "type": "module", - "scripts": { - "dev": "vite", - "build": "npm run format && tsc -b && vite build", - "format": "eslint . && prettier --write .", - "lint": "eslint .", - "preview": "vite preview" - }, - "dependencies": { - "@heroicons/react": "^2.2.0", - "@sec-ant/readable-stream": "^0.6.0", - "@tailwindcss/postcss": "^4.1.1", - "@tailwindcss/vite": "^4.1.1", - "@vscode/markdown-it-katex": "^1.1.1", - "autoprefixer": "^10.4.20", - "daisyui": "^5.0.12", - "dexie": "^4.0.11", - "highlight.js": "^11.10.0", - "katex": "^0.16.15", - "pdfjs-dist": "^5.2.133", - "postcss": "^8.4.49", - "react": "^18.3.1", - "react-dom": "^18.3.1", - "react-dropzone": "^14.3.8", - "react-hot-toast": "^2.5.2", - "react-markdown": "^9.0.3", - "react-router": "^7.1.5", - "rehype-highlight": "^7.0.2", - "rehype-katex": "^7.0.1", - "remark-breaks": "^4.0.0", - "remark-gfm": "^4.0.0", - "remark-math": "^6.0.0", - "tailwindcss": "^4.1.1", - "textlinestream": "^1.1.1", - "vite-plugin-singlefile": "^2.0.3" - }, - "devDependencies": { - "@eslint/js": "^9.17.0", - "@types/markdown-it": "^14.1.2", - "@types/node": "^22.13.1", - "@types/react": "^18.3.18", - "@types/react-dom": "^18.3.5", - "@vitejs/plugin-react": "^4.3.4", - "eslint": "^9.17.0", - "eslint-plugin-react-hooks": "^5.0.0", - "eslint-plugin-react-refresh": "^0.4.16", - "fflate": "^0.8.2", - "globals": "^15.14.0", - "prettier": "^3.4.2", - "sass-embedded": "^1.83.4", - "typescript": "~5.6.2", - "typescript-eslint": "^8.18.2", - "vite": "^6.0.5" - }, - "prettier": { - "trailingComma": "es5", - "tabWidth": 2, - "semi": true, - "singleQuote": true, - "bracketSameLine": false - } + "name": "webui", + "private": true, + "version": "1.0.0", + "type": "module", + "scripts": { + "dev": "vite dev --host 0.0.0.0 & storybook dev -p 6006 --ci", + "build": "vite build && ./scripts/post-build.sh", + "preview": "vite preview", + "prepare": "svelte-kit sync || echo ''", + "check": "svelte-kit sync && svelte-check --tsconfig ./tsconfig.json", + "check:watch": "svelte-kit sync && svelte-check --tsconfig ./tsconfig.json --watch", + "reset": "rm -rf .svelte-kit node_modules", + "format": "prettier --write .", + "lint": "prettier --check . && eslint .", + "test": "npm run test:ui -- --run && npm run test:client -- --run && npm run test:server -- --run && npm run test:e2e", + "test:e2e": "playwright test", + "test:client": "vitest --project=client", + "test:server": "vitest --project=server", + "test:ui": "vitest --project=ui", + "test:unit": "vitest", + "storybook": "storybook dev -p 6006", + "build-storybook": "storybook build" + }, + "devDependencies": { + "@chromatic-com/storybook": "^4.0.1", + "@eslint/compat": "^1.2.5", + "@eslint/js": "^9.18.0", + "@internationalized/date": "^3.8.2", + "@lucide/svelte": "^0.515.0", + "@playwright/test": "^1.49.1", + "@storybook/addon-a11y": "^9.0.17", + "@storybook/addon-docs": "^9.0.17", + "@storybook/addon-svelte-csf": "^5.0.7", + "@storybook/addon-vitest": "^9.0.17", + "@storybook/sveltekit": "^9.0.17", + "@sveltejs/adapter-static": "^3.0.8", + "@sveltejs/kit": "^2.22.0", + "@sveltejs/vite-plugin-svelte": "^6.0.0", + "@tailwindcss/forms": "^0.5.9", + "@tailwindcss/typography": "^0.5.15", + "@tailwindcss/vite": "^4.0.0", + "@types/node": "^22", + "@vitest/browser": "^3.2.3", + "bits-ui": "^2.8.11", + "clsx": "^2.1.1", + "dexie": "^4.0.11", + "eslint": "^9.18.0", + "eslint-config-prettier": "^10.0.1", + "eslint-plugin-storybook": "^9.0.17", + "eslint-plugin-svelte": "^3.0.0", + "fflate": "^0.8.2", + "globals": "^16.0.0", + "mdsvex": "^0.12.3", + "playwright": "^1.53.0", + "prettier": "^3.4.2", + "prettier-plugin-svelte": "^3.3.3", + "prettier-plugin-tailwindcss": "^0.6.11", + "rehype-katex": "^7.0.1", + "remark-math": "^6.0.0", + "storybook": "^9.0.17", + "svelte": "^5.0.0", + "svelte-check": "^4.0.0", + "tailwind-merge": "^3.3.1", + "tailwind-variants": "^1.0.0", + "tailwindcss": "^4.0.0", + "tw-animate-css": "^1.3.5", + "typescript": "^5.0.0", + "typescript-eslint": "^8.20.0", + "uuid": "^13.0.0", + "vite": "^7.0.4", + "vite-plugin-devtools-json": "^0.2.0", + "vitest": "^3.2.3", + "vitest-browser-svelte": "^0.1.0" + }, + "dependencies": { + "highlight.js": "^11.11.1", + "mode-watcher": "^1.1.0", + "pdfjs-dist": "^5.4.54", + "rehype-highlight": "^7.0.2", + "rehype-stringify": "^10.0.1", + "remark": "^15.0.1", + "remark-breaks": "^4.0.0", + "remark-gfm": "^4.0.1", + "remark-html": "^16.0.1", + "remark-rehype": "^11.1.2", + "svelte-sonner": "^1.0.5", + "unist-util-visit": "^5.0.0" + } } diff --git a/tools/server/webui/playwright.config.ts b/tools/server/webui/playwright.config.ts new file mode 100644 index 0000000000..90ca19b09f --- /dev/null +++ b/tools/server/webui/playwright.config.ts @@ -0,0 +1,9 @@ +import { defineConfig } from '@playwright/test'; + +export default defineConfig({ + webServer: { + command: 'npm run build && npx http-server ../public -p 8181', + port: 8181 + }, + testDir: 'e2e' +}); diff --git a/tools/server/webui/postcss.config.js b/tools/server/webui/postcss.config.js deleted file mode 100644 index fb05b5692b..0000000000 --- a/tools/server/webui/postcss.config.js +++ /dev/null @@ -1,5 +0,0 @@ -export default { - plugins: { - "@tailwindcss/postcss": {}, - }, -} diff --git a/tools/server/webui/public/demo-conversation.json b/tools/server/webui/public/demo-conversation.json deleted file mode 100644 index 338b4aea59..0000000000 --- a/tools/server/webui/public/demo-conversation.json +++ /dev/null @@ -1,33 +0,0 @@ -{ - "demo": true, - "id": "conv-1734086746930", - "lastModified": 1734087548943, - "messages": [ - { - "id": 1734086764521, - "role": "user", - "content": "this is a demo conversation, used in dev mode" - }, - { - "id": 1734087548327, - "role": "assistant", - "content": "This is the formula:\n\n$\\frac{e^{x_i}}{\\sum_{j=1}^{n}e^{x_j}}$\n\nGiven an input vector \\(\\mathbf{x} = [x_1, x_2, \\ldots, x_n]\\)\n\n\\[\ny_i = \\frac{e^{x_i}}{\\sum_{j=1}^n e^{x_j}}\n\\]\n\n$2x + y = z$\n\nCode block latex:\n```latex\n\\frac{e^{x_i}}{\\sum_{j=1}^{n}e^{x_j}}\n```\n\nTest dollar sign: $1234 $4567\n\nInvalid latex syntax: $E = mc^$ and $$E = mc^$$", - "timings": { - "prompt_n": 1, - "prompt_ms": 28.923, - "predicted_n": 25, - "predicted_ms": 573.016 - } - }, - { - "id": 1734087548328, - "role": "user", - "content": "this is a demo conversation, used in dev mode" - }, - { - "id": 1734087548329, - "role": "assistant", - "content": "Code block:\n```js\nconsole.log('hello world')\n```\n```sh\nls -la /dev\n```" - } - ] -} diff --git a/tools/server/webui/scripts/install-git-hooks.sh b/tools/server/webui/scripts/install-git-hooks.sh new file mode 100755 index 0000000000..740075e281 --- /dev/null +++ b/tools/server/webui/scripts/install-git-hooks.sh @@ -0,0 +1,123 @@ +#!/bin/bash + +# Script to install pre-commit and post-commit hooks for webui +# Pre-commit: formats code and builds, stashes unstaged changes +# Post-commit: automatically unstashes changes + +REPO_ROOT=$(git rev-parse --show-toplevel) +PRE_COMMIT_HOOK="$REPO_ROOT/.git/hooks/pre-commit" +POST_COMMIT_HOOK="$REPO_ROOT/.git/hooks/post-commit" + +echo "Installing pre-commit and post-commit hooks for webui..." + +# Create the pre-commit hook +cat > "$PRE_COMMIT_HOOK" << 'EOF' +#!/bin/bash + +# Check if there are any changes in the webui directory +if git diff --cached --name-only | grep -q "^tools/server/webui/"; then + echo "Formatting webui code..." + + # Change to webui directory and run format + cd tools/server/webui + + # Check if npm is available and package.json exists + if [ ! -f "package.json" ]; then + echo "Error: package.json not found in tools/server/webui" + exit 1 + fi + + # Stash any unstaged changes to avoid conflicts during format/build + echo "Stashing unstaged changes..." + git stash push --keep-index --include-untracked -m "Pre-commit hook: stashed unstaged changes" + STASH_CREATED=$? + + # Run the format command + npm run format + + # Check if format command succeeded + if [ $? -ne 0 ]; then + echo "Error: npm run format failed" + if [ $STASH_CREATED -eq 0 ]; then + echo "You can restore your unstaged changes with: git stash pop" + fi + exit 1 + fi + + # Run the check command + npm run check + + # Check if check command succeeded + if [ $? -ne 0 ]; then + echo "Error: npm run check failed" + if [ $STASH_CREATED -eq 0 ]; then + echo "You can restore your unstaged changes with: git stash pop" + fi + exit 1 + fi + + # Run the build command + npm run build + + # Check if build command succeeded + if [ $? -ne 0 ]; then + echo "Error: npm run build failed" + if [ $STASH_CREATED -eq 0 ]; then + echo "You can restore your unstaged changes with: git stash pop" + fi + exit 1 + fi + + # Go back to repo root to add build output + cd ../../.. + + # Add the build output to staging area + git add tools/server/public/index.html.gz + + if [ $STASH_CREATED -eq 0 ]; then + echo "βœ… Build completed. Your unstaged changes have been stashed." + echo "They will be automatically restored after the commit." + # Create a marker file to indicate stash was created by pre-commit hook + touch .git/WEBUI_STASH_MARKER + fi + + echo "Webui code formatted successfully" +fi + +exit 0 +EOF + +# Create the post-commit hook +cat > "$POST_COMMIT_HOOK" << 'EOF' +#!/bin/bash + +# Check if we have a stash marker from the pre-commit hook +if [ -f .git/WEBUI_STASH_MARKER ]; then + echo "Restoring your unstaged changes..." + git stash pop + rm -f .git/WEBUI_STASH_MARKER + echo "βœ… Your unstaged changes have been restored." +fi + +exit 0 +EOF + +# Make both hooks executable +chmod +x "$PRE_COMMIT_HOOK" +chmod +x "$POST_COMMIT_HOOK" + +if [ $? -eq 0 ]; then + echo "βœ… Pre-commit and post-commit hooks installed successfully!" + echo " Pre-commit: $PRE_COMMIT_HOOK" + echo " Post-commit: $POST_COMMIT_HOOK" + echo "" + echo "The hooks will automatically:" + echo " β€’ Format and build webui code before commits" + echo " β€’ Stash unstaged changes during the process" + echo " β€’ Restore your unstaged changes after the commit" + echo "" + echo "To test the hooks, make a change to a file in the webui directory and commit it." +else + echo "❌ Failed to make hooks executable" + exit 1 +fi diff --git a/tools/server/webui/scripts/post-build.sh b/tools/server/webui/scripts/post-build.sh new file mode 100755 index 0000000000..ae5c1f30be --- /dev/null +++ b/tools/server/webui/scripts/post-build.sh @@ -0,0 +1,3 @@ +rm -rf ../public/_app; +rm ../public/favicon.svg; +rm ../public/index.html; \ No newline at end of file diff --git a/tools/server/webui/src/App.tsx b/tools/server/webui/src/App.tsx deleted file mode 100644 index 8dfcf49075..0000000000 --- a/tools/server/webui/src/App.tsx +++ /dev/null @@ -1,52 +0,0 @@ -import { HashRouter, Outlet, Route, Routes } from 'react-router'; -import Header from './components/Header'; -import Sidebar from './components/Sidebar'; -import { AppContextProvider, useAppContext } from './utils/app.context'; -import ChatScreen from './components/ChatScreen'; -import SettingDialog from './components/SettingDialog'; -import { Toaster } from 'react-hot-toast'; -import { ModalProvider } from './components/ModalProvider'; - -function App() { - return ( - - -
- - - }> - } /> - } /> - - - -
- - - ); -} - -function AppLayout() { - const { showSettings, setShowSettings } = useAppContext(); - return ( - <> - -
-
- -
- { - setShowSettings(false)} - /> - } - - - ); -} - -export default App; diff --git a/tools/server/webui/src/Config.ts b/tools/server/webui/src/Config.ts deleted file mode 100644 index c03ac287f3..0000000000 --- a/tools/server/webui/src/Config.ts +++ /dev/null @@ -1,96 +0,0 @@ -import daisyuiThemes from 'daisyui/theme/object'; -import { isNumeric } from './utils/misc'; - -export const isDev = import.meta.env.MODE === 'development'; - -// constants -export const BASE_URL = new URL('.', document.baseURI).href - .toString() - .replace(/\/$/, ''); - -export const CONFIG_DEFAULT = { - // Note: in order not to introduce breaking changes, please keep the same data type (number, string, etc) if you want to change the default value. Do not use null or undefined for default value. - // Do not use nested objects, keep it single level. Prefix the key if you need to group them. - apiKey: '', - systemMessage: '', - showTokensPerSecond: false, - showThoughtInProgress: false, - excludeThoughtOnReq: true, - pasteLongTextToFileLen: 2500, - pdfAsImage: false, - // make sure these default values are in sync with `common.h` - samplers: 'edkypmxt', - temperature: 0.8, - dynatemp_range: 0.0, - dynatemp_exponent: 1.0, - top_k: 40, - top_p: 0.95, - min_p: 0.05, - xtc_probability: 0.0, - xtc_threshold: 0.1, - typical_p: 1.0, - repeat_last_n: 64, - repeat_penalty: 1.0, - presence_penalty: 0.0, - frequency_penalty: 0.0, - dry_multiplier: 0.0, - dry_base: 1.75, - dry_allowed_length: 2, - dry_penalty_last_n: -1, - max_tokens: -1, - custom: '', // custom json-stringified object - // experimental features - pyIntepreterEnabled: false, -}; -export const CONFIG_INFO: Record = { - apiKey: 'Set the API Key if you are using --api-key option for the server.', - systemMessage: 'The starting message that defines how model should behave.', - pasteLongTextToFileLen: - 'On pasting long text, it will be converted to a file. You can control the file length by setting the value of this parameter. Value 0 means disable.', - samplers: - 'The order at which samplers are applied, in simplified way. Default is "dkypmxt": dry->top_k->typ_p->top_p->min_p->xtc->temperature', - temperature: - 'Controls the randomness of the generated text by affecting the probability distribution of the output tokens. Higher = more random, lower = more focused.', - dynatemp_range: - 'Addon for the temperature sampler. The added value to the range of dynamic temperature, which adjusts probabilities by entropy of tokens.', - dynatemp_exponent: - 'Addon for the temperature sampler. Smoothes out the probability redistribution based on the most probable token.', - top_k: 'Keeps only k top tokens.', - top_p: - 'Limits tokens to those that together have a cumulative probability of at least p', - min_p: - 'Limits tokens based on the minimum probability for a token to be considered, relative to the probability of the most likely token.', - xtc_probability: - 'XTC sampler cuts out top tokens; this parameter controls the chance of cutting tokens at all. 0 disables XTC.', - xtc_threshold: - 'XTC sampler cuts out top tokens; this parameter controls the token probability that is required to cut that token.', - typical_p: - 'Sorts and limits tokens based on the difference between log-probability and entropy.', - repeat_last_n: 'Last n tokens to consider for penalizing repetition', - repeat_penalty: - 'Controls the repetition of token sequences in the generated text', - presence_penalty: - 'Limits tokens based on whether they appear in the output or not.', - frequency_penalty: - 'Limits tokens based on how often they appear in the output.', - dry_multiplier: - 'DRY sampling reduces repetition in generated text even across long contexts. This parameter sets the DRY sampling multiplier.', - dry_base: - 'DRY sampling reduces repetition in generated text even across long contexts. This parameter sets the DRY sampling base value.', - dry_allowed_length: - 'DRY sampling reduces repetition in generated text even across long contexts. This parameter sets the allowed length for DRY sampling.', - dry_penalty_last_n: - 'DRY sampling reduces repetition in generated text even across long contexts. This parameter sets DRY penalty for the last n tokens.', - max_tokens: 'The maximum number of token per output.', - custom: '', // custom json-stringified object -}; -// config keys having numeric value (i.e. temperature, top_k, top_p, etc) -export const CONFIG_NUMERIC_KEYS = Object.entries(CONFIG_DEFAULT) - .filter((e) => isNumeric(e[1])) - .map((e) => e[0]); -// list of themes supported by daisyui -export const THEMES = ['light', 'dark'] - // make sure light & dark are always at the beginning - .concat( - Object.keys(daisyuiThemes).filter((t) => t !== 'light' && t !== 'dark') - ); diff --git a/tools/server/webui/src/app.css b/tools/server/webui/src/app.css new file mode 100644 index 0000000000..d05a422f02 --- /dev/null +++ b/tools/server/webui/src/app.css @@ -0,0 +1,123 @@ +@import 'tailwindcss'; + +@import 'tw-animate-css'; + +@custom-variant dark (&:is(.dark *)); + +:root { + --radius: 0.625rem; + --background: oklch(1 0 0); + --foreground: oklch(0.145 0 0); + --card: oklch(1 0 0); + --card-foreground: oklch(0.145 0 0); + --popover: oklch(1 0 0); + --popover-foreground: oklch(0.145 0 0); + --primary: oklch(0.205 0 0); + --primary-foreground: oklch(0.985 0 0); + --secondary: oklch(0.97 0 0); + --secondary-foreground: oklch(0.205 0 0); + --muted: oklch(0.97 0 0); + --muted-foreground: oklch(0.556 0 0); + --accent: oklch(0.97 0 0); + --accent-foreground: oklch(0.205 0 0); + --destructive: oklch(0.577 0.245 27.325); + --border: oklch(0.875 0 0); + --input: oklch(0.92 0 0); + --ring: oklch(0.708 0 0); + --chart-1: oklch(0.646 0.222 41.116); + --chart-2: oklch(0.6 0.118 184.704); + --chart-3: oklch(0.398 0.07 227.392); + --chart-4: oklch(0.828 0.189 84.429); + --chart-5: oklch(0.769 0.188 70.08); + --sidebar: oklch(0.985 0 0); + --sidebar-foreground: oklch(0.145 0 0); + --sidebar-primary: oklch(0.205 0 0); + --sidebar-primary-foreground: oklch(0.985 0 0); + --sidebar-accent: oklch(0.97 0 0); + --sidebar-accent-foreground: oklch(0.205 0 0); + --sidebar-border: oklch(0.922 0 0); + --sidebar-ring: oklch(0.708 0 0); + --code-background: oklch(0.225 0 0); + --code-foreground: oklch(0.875 0 0); +} + +.dark { + --background: oklch(0.16 0 0); + --foreground: oklch(0.985 0 0); + --card: oklch(0.205 0 0); + --card-foreground: oklch(0.985 0 0); + --popover: oklch(0.205 0 0); + --popover-foreground: oklch(0.985 0 0); + --primary: oklch(0.922 0 0); + --primary-foreground: oklch(0.205 0 0); + --secondary: oklch(0.269 0 0); + --secondary-foreground: oklch(0.985 0 0); + --muted: oklch(0.269 0 0); + --muted-foreground: oklch(0.708 0 0); + --accent: oklch(0.269 0 0); + --accent-foreground: oklch(0.985 0 0); + --destructive: oklch(0.704 0.191 22.216); + --border: oklch(1 0 0 / 30%); + --input: oklch(1 0 0 / 30%); + --ring: oklch(0.556 0 0); + --chart-1: oklch(0.488 0.243 264.376); + --chart-2: oklch(0.696 0.17 162.48); + --chart-3: oklch(0.769 0.188 70.08); + --chart-4: oklch(0.627 0.265 303.9); + --chart-5: oklch(0.645 0.246 16.439); + --sidebar: oklch(0.205 0 0); + --sidebar-foreground: oklch(0.985 0 0); + --sidebar-primary: oklch(0.488 0.243 264.376); + --sidebar-primary-foreground: oklch(0.985 0 0); + --sidebar-accent: oklch(0.269 0 0); + --sidebar-accent-foreground: oklch(0.985 0 0); + --sidebar-border: oklch(1 0 0 / 10%); + --sidebar-ring: oklch(0.556 0 0); +} + +@theme inline { + --radius-sm: calc(var(--radius) - 4px); + --radius-md: calc(var(--radius) - 2px); + --radius-lg: var(--radius); + --radius-xl: calc(var(--radius) + 4px); + --color-background: var(--background); + --color-foreground: var(--foreground); + --color-card: var(--card); + --color-card-foreground: var(--card-foreground); + --color-popover: var(--popover); + --color-popover-foreground: var(--popover-foreground); + --color-primary: var(--primary); + --color-primary-foreground: var(--primary-foreground); + --color-secondary: var(--secondary); + --color-secondary-foreground: var(--secondary-foreground); + --color-muted: var(--muted); + --color-muted-foreground: var(--muted-foreground); + --color-accent: var(--accent); + --color-accent-foreground: var(--accent-foreground); + --color-destructive: var(--destructive); + --color-border: var(--border); + --color-input: var(--input); + --color-ring: var(--ring); + --color-chart-1: var(--chart-1); + --color-chart-2: var(--chart-2); + --color-chart-3: var(--chart-3); + --color-chart-4: var(--chart-4); + --color-chart-5: var(--chart-5); + --color-sidebar: var(--sidebar); + --color-sidebar-foreground: var(--sidebar-foreground); + --color-sidebar-primary: var(--sidebar-primary); + --color-sidebar-primary-foreground: var(--sidebar-primary-foreground); + --color-sidebar-accent: var(--sidebar-accent); + --color-sidebar-accent-foreground: var(--sidebar-accent-foreground); + --color-sidebar-border: var(--sidebar-border); + --color-sidebar-ring: var(--sidebar-ring); +} + +@layer base { + * { + @apply border-border outline-ring/50; + } + body { + @apply bg-background text-foreground; + } +} diff --git a/tools/server/webui/src/app.d.ts b/tools/server/webui/src/app.d.ts new file mode 100644 index 0000000000..e9bb140939 --- /dev/null +++ b/tools/server/webui/src/app.d.ts @@ -0,0 +1,81 @@ +// See https://svelte.dev/docs/kit/types#app.d.ts +// for information about these interfaces + +// Import chat types from dedicated module + +import type { + ApiChatCompletionRequest, + ApiChatCompletionResponse, + ApiChatCompletionStreamChunk, + ApiChatMessageData, + ApiChatMessageContentPart, + ApiContextSizeError, + ApiErrorResponse, + ApiLlamaCppServerProps, + ApiProcessingState +} from '$lib/types/api'; + +import type { + ChatMessageType, + ChatRole, + ChatUploadedFile, + ChatMessageSiblingInfo, + ChatMessagePromptProgress, + ChatMessageTimings +} from '$lib/types/chat'; + +import type { + DatabaseConversation, + DatabaseMessage, + DatabaseMessageExtra, + DatabaseMessageExtraAudioFile, + DatabaseMessageExtraImageFile, + DatabaseMessageExtraTextFile, + DatabaseMessageExtraPdfFile +} from '$lib/types/database'; + +import type { + SettingsConfigValue, + SettingsFieldConfig, + SettingsConfigType +} from '$lib/types/settings'; + +declare global { + // namespace App { + // interface Error {} + // interface Locals {} + // interface PageData {} + // interface PageState {} + // interface Platform {} + // } + + export { + ApiChatCompletionRequest, + ApiChatCompletionResponse, + ApiChatCompletionStreamChunk, + ApiChatMessageData, + ApiChatMessageContentPart, + ApiContextSizeError, + ApiErrorResponse, + ApiLlamaCppServerProps, + ApiProcessingState, + ChatMessageData, + ChatMessagePromptProgress, + ChatMessageSiblingInfo, + ChatMessageTimings, + ChatMessageType, + ChatRole, + ChatUploadedFile, + DatabaseConversation, + DatabaseMessage, + DatabaseMessageExtra, + DatabaseMessageExtraAudioFile, + DatabaseMessageExtraImageFile, + DatabaseMessageExtraTextFile, + DatabaseMessageExtraPdfFile, + SettingsConfigValue, + SettingsFieldConfig, + SettingsConfigType, + SettingsChatServiceOptions + }; +} diff --git a/tools/server/webui/src/app.html b/tools/server/webui/src/app.html new file mode 100644 index 0000000000..1391f88488 --- /dev/null +++ b/tools/server/webui/src/app.html @@ -0,0 +1,12 @@ + + + + + + + %sveltekit.head% + + +
%sveltekit.body%
+ + diff --git a/tools/server/webui/src/components/CanvasPyInterpreter.tsx b/tools/server/webui/src/components/CanvasPyInterpreter.tsx deleted file mode 100644 index c2707fe20f..0000000000 --- a/tools/server/webui/src/components/CanvasPyInterpreter.tsx +++ /dev/null @@ -1,195 +0,0 @@ -import { useEffect, useState } from 'react'; -import { useAppContext } from '../utils/app.context'; -import { OpenInNewTab, XCloseButton } from '../utils/common'; -import { CanvasType } from '../utils/types'; -import { PlayIcon, StopIcon } from '@heroicons/react/24/outline'; -import StorageUtils from '../utils/storage'; - -const canInterrupt = typeof SharedArrayBuffer === 'function'; - -// adapted from https://pyodide.org/en/stable/usage/webworker.html -const WORKER_CODE = ` -importScripts("https://cdn.jsdelivr.net/pyodide/v0.27.2/full/pyodide.js"); - -let stdOutAndErr = []; - -let pyodideReadyPromise = loadPyodide({ - stdout: (data) => stdOutAndErr.push(data), - stderr: (data) => stdOutAndErr.push(data), -}); - -let alreadySetBuff = false; - -self.onmessage = async (event) => { - stdOutAndErr = []; - - // make sure loading is done - const pyodide = await pyodideReadyPromise; - const { id, python, context, interruptBuffer } = event.data; - - if (interruptBuffer && !alreadySetBuff) { - pyodide.setInterruptBuffer(interruptBuffer); - alreadySetBuff = true; - } - - // Now load any packages we need, run the code, and send the result back. - await pyodide.loadPackagesFromImports(python); - - // make a Python dictionary with the data from content - const dict = pyodide.globals.get("dict"); - const globals = dict(Object.entries(context)); - try { - self.postMessage({ id, running: true }); - // Execute the python code in this context - const result = pyodide.runPython(python, { globals }); - self.postMessage({ result, id, stdOutAndErr }); - } catch (error) { - self.postMessage({ error: error.message, id }); - } - interruptBuffer[0] = 0; -}; -`; - -let worker: Worker; -const interruptBuffer = canInterrupt - ? new Uint8Array(new SharedArrayBuffer(1)) - : null; - -const startWorker = () => { - if (!worker) { - worker = new Worker( - URL.createObjectURL(new Blob([WORKER_CODE], { type: 'text/javascript' })) - ); - } -}; - -if (StorageUtils.getConfig().pyIntepreterEnabled) { - startWorker(); -} - -const runCodeInWorker = ( - pyCode: string, - callbackRunning: () => void -): { - donePromise: Promise; - interrupt: () => void; -} => { - startWorker(); - const id = Math.random() * 1e8; - const context = {}; - if (interruptBuffer) { - interruptBuffer[0] = 0; - } - - const donePromise = new Promise((resolve) => { - worker.onmessage = (event) => { - const { error, stdOutAndErr, running } = event.data; - if (id !== event.data.id) return; - if (running) { - callbackRunning(); - return; - } else if (error) { - resolve(error.toString()); - } else { - resolve(stdOutAndErr.join('\n')); - } - }; - worker.postMessage({ id, python: pyCode, context, interruptBuffer }); - }); - - const interrupt = () => { - console.log('Interrupting...'); - console.trace(); - if (interruptBuffer) { - interruptBuffer[0] = 2; - } - }; - - return { donePromise, interrupt }; -}; - -export default function CanvasPyInterpreter() { - const { canvasData, setCanvasData } = useAppContext(); - - const [code, setCode] = useState(canvasData?.content ?? ''); // copy to avoid direct mutation - const [running, setRunning] = useState(false); - const [output, setOutput] = useState(''); - const [interruptFn, setInterruptFn] = useState<() => void>(); - const [showStopBtn, setShowStopBtn] = useState(false); - - const runCode = async (pycode: string) => { - interruptFn?.(); - setRunning(true); - setOutput('Loading Pyodide...'); - const { donePromise, interrupt } = runCodeInWorker(pycode, () => { - setOutput('Running...'); - setShowStopBtn(canInterrupt); - }); - setInterruptFn(() => interrupt); - const out = await donePromise; - setOutput(out); - setRunning(false); - setShowStopBtn(false); - }; - - // run code on mount - useEffect(() => { - setCode(canvasData?.content ?? ''); - runCode(canvasData?.content ?? ''); - // eslint-disable-next-line react-hooks/exhaustive-deps - }, [canvasData?.content]); - - if (canvasData?.type !== CanvasType.PY_INTERPRETER) { - return null; - } - - return ( -
-
-
- Python Interpreter - setCanvasData(null)} - /> -
-
- -
-
- - {showStopBtn && ( - - )} - - - Report a bug - - -
- -
-
-
-
- ); -} diff --git a/tools/server/webui/src/components/ChatInputExtraContextItem.tsx b/tools/server/webui/src/components/ChatInputExtraContextItem.tsx deleted file mode 100644 index 2d4179ea47..0000000000 --- a/tools/server/webui/src/components/ChatInputExtraContextItem.tsx +++ /dev/null @@ -1,135 +0,0 @@ -import { - DocumentTextIcon, - SpeakerWaveIcon, - XMarkIcon, -} from '@heroicons/react/24/outline'; -import { MessageExtra } from '../utils/types'; -import { useState } from 'react'; -import { classNames } from '../utils/misc'; - -export default function ChatInputExtraContextItem({ - items, - removeItem, - clickToShow, -}: { - items?: MessageExtra[]; - removeItem?: (index: number) => void; - clickToShow?: boolean; -}) { - const [show, setShow] = useState(-1); - const showingItem = show >= 0 ? items?.[show] : undefined; - - if (!items) return null; - - return ( -
- {items.map((item, i) => ( -
clickToShow && setShow(i)} - tabIndex={0} - aria-description={ - clickToShow ? `Click to show: ${item.name}` : undefined - } - role={clickToShow ? 'button' : 'menuitem'} - > - {removeItem && ( -
- -
- )} - -
- {item.type === 'imageFile' ? ( - <> - {`Preview - - ) : ( - <> -
- {item.type === 'audioFile' ? ( - - ) : ( - - )} -
- -
- {item.name ?? 'Extra content'} -
- - )} -
-
- ))} - - {showingItem && ( - -
-
- {showingItem.name ?? 'Extra content'} - -
- {showingItem.type === 'imageFile' ? ( - {`Preview - ) : showingItem.type === 'audioFile' ? ( - - ) : ( -
- 
-                  {showingItem.content}
-
-
- )} -
-
setShow(-1)}>
- - )} -
- ); -} diff --git a/tools/server/webui/src/components/ChatMessage.tsx b/tools/server/webui/src/components/ChatMessage.tsx deleted file mode 100644 index a1a7d152f9..0000000000 --- a/tools/server/webui/src/components/ChatMessage.tsx +++ /dev/null @@ -1,320 +0,0 @@ -import { useMemo, useState } from 'react'; -import { useAppContext } from '../utils/app.context'; -import { Message, PendingMessage } from '../utils/types'; -import { classNames } from '../utils/misc'; -import MarkdownDisplay, { CopyButton } from './MarkdownDisplay'; -import { - ArrowPathIcon, - ChevronLeftIcon, - ChevronRightIcon, - PencilSquareIcon, -} from '@heroicons/react/24/outline'; -import ChatInputExtraContextItem from './ChatInputExtraContextItem'; -import { BtnWithTooltips } from '../utils/common'; - -interface SplitMessage { - content: PendingMessage['content']; - thought?: string; - isThinking?: boolean; -} - -export default function ChatMessage({ - msg, - siblingLeafNodeIds, - siblingCurrIdx, - id, - onRegenerateMessage, - onEditMessage, - onChangeSibling, - isPending, -}: { - msg: Message | PendingMessage; - siblingLeafNodeIds: Message['id'][]; - siblingCurrIdx: number; - id?: string; - onRegenerateMessage(msg: Message): void; - onEditMessage(msg: Message, content: string): void; - onChangeSibling(sibling: Message['id']): void; - isPending?: boolean; -}) { - const { viewingChat, config } = useAppContext(); - const [editingContent, setEditingContent] = useState(null); - const timings = useMemo( - () => - msg.timings - ? { - ...msg.timings, - prompt_per_second: - (msg.timings.prompt_n / msg.timings.prompt_ms) * 1000, - predicted_per_second: - (msg.timings.predicted_n / msg.timings.predicted_ms) * 1000, - } - : null, - [msg.timings] - ); - const nextSibling = siblingLeafNodeIds[siblingCurrIdx + 1]; - const prevSibling = siblingLeafNodeIds[siblingCurrIdx - 1]; - - // for reasoning model, we split the message into content and thought - // TODO: implement this as remark/rehype plugin in the future - const { content, thought, isThinking }: SplitMessage = useMemo(() => { - if (msg.content === null || msg.role !== 'assistant') { - return { content: msg.content }; - } - const REGEX_THINK_OPEN = /|<\|channel\|>analysis<\|message\|>/; - const REGEX_THINK_CLOSE = /<\/think>|<\|end\|>/; - let actualContent = ''; - let thought = ''; - let isThinking = false; - let thinkSplit = msg.content.split(REGEX_THINK_OPEN, 2); - actualContent += thinkSplit[0]; - while (thinkSplit[1] !== undefined) { - // tag found - thinkSplit = thinkSplit[1].split(REGEX_THINK_CLOSE, 2); - thought += thinkSplit[0]; - isThinking = true; - if (thinkSplit[1] !== undefined) { - // closing tag found - isThinking = false; - thinkSplit = thinkSplit[1].split(REGEX_THINK_OPEN, 2); - actualContent += thinkSplit[0]; - } - } - return { content: actualContent, thought, isThinking }; - }, [msg]); - - if (!viewingChat) return null; - - const isUser = msg.role === 'user'; - - return ( -
-
- {msg.extra && msg.extra.length > 0 && ( - - )} - -
- {/* textarea for editing message */} - {editingContent !== null && ( - <> - -
- - - - )} - {/* not editing content, render message */} - {editingContent === null && ( - <> - {content === null ? ( - <> - {/* show loading dots for pending message */} - - - ) : ( - <> - {/* render message as markdown */} -
- {thought && ( - - )} - - -
- - )} - {/* render timings if enabled */} - {timings && config.showTokensPerSecond && ( -
-
- Speed: {timings.predicted_per_second.toFixed(1)} t/s -
-
- Prompt -
- Tokens: {timings.prompt_n} -
- Time: {timings.prompt_ms} ms -
- Speed: {timings.prompt_per_second.toFixed(1)} t/s -
- Generation -
- Tokens: {timings.predicted_n} -
- Time: {timings.predicted_ms} ms -
- Speed: {timings.predicted_per_second.toFixed(1)} t/s -
-
-
- )} - - )} -
-
- - {/* actions for each message */} - {msg.content !== null && ( -
- {siblingLeafNodeIds && siblingLeafNodeIds.length > 1 && ( -
- - - {siblingCurrIdx + 1} / {siblingLeafNodeIds.length} - - -
- )} - {/* user message */} - {msg.role === 'user' && ( - setEditingContent(msg.content)} - disabled={msg.content === null} - tooltipsContent="Edit message" - > - - - )} - {/* assistant message */} - {msg.role === 'assistant' && ( - <> - {!isPending && ( - { - if (msg.content !== null) { - onRegenerateMessage(msg as Message); - } - }} - disabled={msg.content === null} - tooltipsContent="Regenerate response" - > - - - )} - - )} - -
- )} -
- ); -} - -function ThoughtProcess({ - isThinking, - content, - open, -}: { - isThinking: boolean; - content: string; - open: boolean; -}) { - return ( -
- -
-
- {isThinking ? ( - - - Thinking - - ) : ( - <>Thought Process - )} -
-
-
-
- -
-
-
- ); -} diff --git a/tools/server/webui/src/components/ChatScreen.tsx b/tools/server/webui/src/components/ChatScreen.tsx deleted file mode 100644 index c1a6691445..0000000000 --- a/tools/server/webui/src/components/ChatScreen.tsx +++ /dev/null @@ -1,459 +0,0 @@ -import { ClipboardEvent, useEffect, useMemo, useRef, useState } from 'react'; -import { CallbackGeneratedChunk, useAppContext } from '../utils/app.context'; -import ChatMessage from './ChatMessage'; -import { CanvasType, Message, PendingMessage } from '../utils/types'; -import { classNames, cleanCurrentUrl } from '../utils/misc'; -import CanvasPyInterpreter from './CanvasPyInterpreter'; -import StorageUtils from '../utils/storage'; -import { useVSCodeContext } from '../utils/llama-vscode'; -import { useChatTextarea, ChatTextareaApi } from './useChatTextarea.ts'; -import { - ArrowUpIcon, - StopIcon, - PaperClipIcon, -} from '@heroicons/react/24/solid'; -import { - ChatExtraContextApi, - useChatExtraContext, -} from './useChatExtraContext.tsx'; -import Dropzone from 'react-dropzone'; -import toast from 'react-hot-toast'; -import ChatInputExtraContextItem from './ChatInputExtraContextItem.tsx'; -import { scrollToBottom, useChatScroll } from './useChatScroll.tsx'; - -/** - * A message display is a message node with additional information for rendering. - * For example, siblings of the message node are stored as their last node (aka leaf node). - */ -export interface MessageDisplay { - msg: Message | PendingMessage; - siblingLeafNodeIds: Message['id'][]; - siblingCurrIdx: number; - isPending?: boolean; -} - -/** - * If the current URL contains "?m=...", prefill the message input with the value. - * If the current URL contains "?q=...", prefill and SEND the message. - */ -const prefilledMsg = { - content() { - const url = new URL(window.location.href); - return url.searchParams.get('m') ?? url.searchParams.get('q') ?? ''; - }, - shouldSend() { - const url = new URL(window.location.href); - return url.searchParams.has('q'); - }, - clear() { - cleanCurrentUrl(['m', 'q']); - }, -}; - -function getListMessageDisplay( - msgs: Readonly, - leafNodeId: Message['id'] -): MessageDisplay[] { - const currNodes = StorageUtils.filterByLeafNodeId(msgs, leafNodeId, true); - const res: MessageDisplay[] = []; - const nodeMap = new Map(); - for (const msg of msgs) { - nodeMap.set(msg.id, msg); - } - // find leaf node from a message node - const findLeafNode = (msgId: Message['id']): Message['id'] => { - let currNode: Message | undefined = nodeMap.get(msgId); - while (currNode) { - if (currNode.children.length === 0) break; - currNode = nodeMap.get(currNode.children.at(-1) ?? -1); - } - return currNode?.id ?? -1; - }; - // traverse the current nodes - for (const msg of currNodes) { - const parentNode = nodeMap.get(msg.parent ?? -1); - if (!parentNode) continue; - const siblings = parentNode.children; - if (msg.type !== 'root') { - res.push({ - msg, - siblingLeafNodeIds: siblings.map(findLeafNode), - siblingCurrIdx: siblings.indexOf(msg.id), - }); - } - } - return res; -} - -export default function ChatScreen() { - const { - viewingChat, - sendMessage, - isGenerating, - stopGenerating, - pendingMessages, - canvasData, - replaceMessageAndGenerate, - } = useAppContext(); - - const textarea: ChatTextareaApi = useChatTextarea(prefilledMsg.content()); - const extraContext = useChatExtraContext(); - useVSCodeContext(textarea, extraContext); - - const msgListRef = useRef(null); - useChatScroll(msgListRef); - - // keep track of leaf node for rendering - const [currNodeId, setCurrNodeId] = useState(-1); - const messages: MessageDisplay[] = useMemo(() => { - if (!viewingChat) return []; - else return getListMessageDisplay(viewingChat.messages, currNodeId); - }, [currNodeId, viewingChat]); - - const currConvId = viewingChat?.conv.id ?? null; - const pendingMsg: PendingMessage | undefined = - pendingMessages[currConvId ?? '']; - - useEffect(() => { - // reset to latest node when conversation changes - setCurrNodeId(-1); - // scroll to bottom when conversation changes - scrollToBottom(false, 1); - }, [currConvId]); - - const onChunk: CallbackGeneratedChunk = (currLeafNodeId?: Message['id']) => { - if (currLeafNodeId) { - setCurrNodeId(currLeafNodeId); - } - // useChatScroll will handle the auto scroll - }; - - const sendNewMessage = async () => { - const lastInpMsg = textarea.value(); - if (lastInpMsg.trim().length === 0 || isGenerating(currConvId ?? '')) { - toast.error('Please enter a message'); - return; - } - textarea.setValue(''); - scrollToBottom(false); - setCurrNodeId(-1); - // get the last message node - const lastMsgNodeId = messages.at(-1)?.msg.id ?? null; - if ( - !(await sendMessage( - currConvId, - lastMsgNodeId, - lastInpMsg, - extraContext.items, - onChunk - )) - ) { - // restore the input message if failed - textarea.setValue(lastInpMsg); - } - // OK - extraContext.clearItems(); - }; - - // for vscode context - textarea.refOnSubmit.current = sendNewMessage; - - const handleEditMessage = async (msg: Message, content: string) => { - if (!viewingChat) return; - setCurrNodeId(msg.id); - scrollToBottom(false); - await replaceMessageAndGenerate( - viewingChat.conv.id, - msg.parent, - content, - msg.extra, - onChunk - ); - setCurrNodeId(-1); - scrollToBottom(false); - }; - - const handleRegenerateMessage = async (msg: Message) => { - if (!viewingChat) return; - setCurrNodeId(msg.parent); - scrollToBottom(false); - await replaceMessageAndGenerate( - viewingChat.conv.id, - msg.parent, - null, - msg.extra, - onChunk - ); - setCurrNodeId(-1); - scrollToBottom(false); - }; - - const hasCanvas = !!canvasData; - - useEffect(() => { - if (prefilledMsg.shouldSend()) { - // send the prefilled message if needed - sendNewMessage(); - } else { - // otherwise, focus on the input - textarea.focus(); - } - prefilledMsg.clear(); - // no need to keep track of sendNewMessage - // eslint-disable-next-line react-hooks/exhaustive-deps - }, [textarea.ref]); - - // due to some timing issues of StorageUtils.appendMsg(), we need to make sure the pendingMsg is not duplicated upon rendering (i.e. appears once in the saved conversation and once in the pendingMsg) - const pendingMsgDisplay: MessageDisplay[] = - pendingMsg && messages.at(-1)?.msg.id !== pendingMsg.id - ? [ - { - msg: pendingMsg, - siblingLeafNodeIds: [], - siblingCurrIdx: 0, - isPending: true, - }, - ] - : []; - - return ( -
-
- {/* chat messages */} -
-
- {/* placeholder to shift the message to the bottom */} - {viewingChat ? ( - '' - ) : ( - <> -
Send a message to start
- - - )} -
- {[...messages, ...pendingMsgDisplay].map((msg) => ( - - ))} -
- - {/* chat input */} - stopGenerating(currConvId ?? '')} - isGenerating={isGenerating(currConvId ?? '')} - /> -
-
- {canvasData?.type === CanvasType.PY_INTERPRETER && ( - - )} -
-
- ); -} - -function ServerInfo() { - const { serverProps } = useAppContext(); - const modalities = []; - if (serverProps?.modalities?.audio) { - modalities.push('audio'); - } - if (serverProps?.modalities?.vision) { - modalities.push('vision'); - } - return ( -
-
- Server Info -

- Model: {serverProps?.model_path?.split(/(\\|\/)/).pop()} -
- Build: {serverProps?.build_info} -
- {modalities.length > 0 ? ( - <> - Supported modalities: {modalities.join(', ')} - - ) : ( - '' - )} -

-
-
- ); -} - -function ChatInput({ - textarea, - extraContext, - onSend, - onStop, - isGenerating, -}: { - textarea: ChatTextareaApi; - extraContext: ChatExtraContextApi; - onSend: () => void; - onStop: () => void; - isGenerating: boolean; -}) { - const { config } = useAppContext(); - const [isDrag, setIsDrag] = useState(false); - - return ( -
- { - setIsDrag(false); - extraContext.onFileAdded(files); - }} - onDragEnter={() => setIsDrag(true)} - onDragLeave={() => setIsDrag(false)} - multiple={true} - > - {({ getRootProps, getInputProps }) => ( -
) => { - const text = e.clipboardData.getData('text/plain'); - if ( - text.length > 0 && - config.pasteLongTextToFileLen > 0 && - text.length > config.pasteLongTextToFileLen - ) { - // if the text is too long, we will convert it to a file - extraContext.addItems([ - { - type: 'context', - name: 'Pasted Content', - content: text, - }, - ]); - e.preventDefault(); - return; - } - - // if a file is pasted, we will handle it here - const files = Array.from(e.clipboardData.items) - .filter((item) => item.kind === 'file') - .map((item) => item.getAsFile()) - .filter((file) => file !== null); - - if (files.length > 0) { - e.preventDefault(); - extraContext.onFileAdded(files); - } - }} - {...getRootProps()} - > - {!isGenerating && ( - - )} - -
- - - {/* buttons area */} -
- - - {isGenerating ? ( - - ) : ( - - )} -
-
-
- )} - -
- ); -} diff --git a/tools/server/webui/src/components/Header.tsx b/tools/server/webui/src/components/Header.tsx deleted file mode 100644 index ccddc21dda..0000000000 --- a/tools/server/webui/src/components/Header.tsx +++ /dev/null @@ -1,92 +0,0 @@ -import { useEffect, useState } from 'react'; -import StorageUtils from '../utils/storage'; -import { useAppContext } from '../utils/app.context'; -import { classNames } from '../utils/misc'; -import daisyuiThemes from 'daisyui/theme/object'; -import { THEMES } from '../Config'; -import { - Cog8ToothIcon, - MoonIcon, - Bars3Icon, -} from '@heroicons/react/24/outline'; - -export default function Header() { - const [selectedTheme, setSelectedTheme] = useState(StorageUtils.getTheme()); - const { setShowSettings } = useAppContext(); - - const setTheme = (theme: string) => { - StorageUtils.setTheme(theme); - setSelectedTheme(theme); - }; - - useEffect(() => { - document.body.setAttribute('data-theme', selectedTheme); - document.body.setAttribute( - 'data-color-scheme', - daisyuiThemes[selectedTheme]?.['color-scheme'] ?? 'auto' - ); - }, [selectedTheme]); - - return ( -
- {/* open sidebar button */} - - -
llama.cpp
- - {/* action buttons (top right) */} -
-
setShowSettings(true)} - > - -
- - {/* theme controller is copied from https://daisyui.com/components/theme-controller/ */} -
-
-
- -
-
    -
  • - -
    • - {THEMES.map((theme) => ( -
  • - e.target.checked && setTheme(theme)} - /> -
    • - ))} -
-
-
-
-
- ); -} diff --git a/tools/server/webui/src/components/MarkdownDisplay.tsx b/tools/server/webui/src/components/MarkdownDisplay.tsx deleted file mode 100644 index 380dbc570a..0000000000 --- a/tools/server/webui/src/components/MarkdownDisplay.tsx +++ /dev/null @@ -1,317 +0,0 @@ -import React, { useMemo, useState } from 'react'; -import Markdown, { ExtraProps } from 'react-markdown'; -import remarkGfm from 'remark-gfm'; -import rehypeHightlight from 'rehype-highlight'; -import rehypeKatex from 'rehype-katex'; -import remarkMath from 'remark-math'; -import remarkBreaks from 'remark-breaks'; -import 'katex/dist/katex.min.css'; -import { classNames, copyStr } from '../utils/misc'; -import { ElementContent, Root } from 'hast'; -import { visit } from 'unist-util-visit'; -import { useAppContext } from '../utils/app.context'; -import { CanvasType } from '../utils/types'; -import { BtnWithTooltips } from '../utils/common'; -import { DocumentDuplicateIcon, PlayIcon } from '@heroicons/react/24/outline'; - -export default function MarkdownDisplay({ - content, - isGenerating, -}: { - content: string; - isGenerating?: boolean; -}) { - const preprocessedContent = useMemo( - () => preprocessLaTeX(content), - [content] - ); - return ( - ( - - ), - // note: do not use "pre", "p" or other basic html elements here, it will cause the node to re-render when the message is being generated (this should be a bug with react-markdown, not sure how to fix it) - }} - > - {preprocessedContent} - - ); -} - -const CodeBlockButtons: React.ElementType< - React.ClassAttributes & - React.HTMLAttributes & - ExtraProps & { origContent: string; isGenerating?: boolean } -> = ({ node, origContent, isGenerating }) => { - const { config } = useAppContext(); - const startOffset = node?.position?.start.offset ?? 0; - const endOffset = node?.position?.end.offset ?? 0; - - const copiedContent = useMemo( - () => - origContent - .substring(startOffset, endOffset) - .replace(/^```[^\n]+\n/g, '') - .replace(/```$/g, ''), - [origContent, startOffset, endOffset] - ); - - const codeLanguage = useMemo( - () => - origContent - .substring(startOffset, startOffset + 10) - .match(/^```([^\n]+)\n/)?.[1] ?? '', - [origContent, startOffset] - ); - - const canRunCode = - !isGenerating && - config.pyIntepreterEnabled && - codeLanguage.startsWith('py'); - - return ( -
- - {canRunCode && ( - - )} -
- ); -}; - -export const CopyButton = ({ - content, - className, -}: { - content: string; - className?: string; -}) => { - const [copied, setCopied] = useState(false); - return ( - { - copyStr(content); - setCopied(true); - }} - onMouseLeave={() => setCopied(false)} - tooltipsContent={copied ? 'Copied!' : 'Copy'} - > - - - ); -}; - -export const RunPyCodeButton = ({ - content, - className, -}: { - content: string; - className?: string; -}) => { - const { setCanvasData } = useAppContext(); - return ( - <> - - setCanvasData({ - type: CanvasType.PY_INTERPRETER, - content, - }) - } - tooltipsContent="Run code" - > - - - - ); -}; - -/** - * This injects the "button" element before each "pre" element. - * The actual button will be replaced with a react component in the MarkdownDisplay. - * We don't replace "pre" node directly because it will cause the node to re-render, which causes this bug: https://github.com/ggerganov/llama.cpp/issues/9608 - */ -function rehypeCustomCopyButton() { - return function (tree: Root) { - visit(tree, 'element', function (node) { - if (node.tagName === 'pre' && !node.properties.visited) { - const preNode = { ...node }; - // replace current node - preNode.properties.visited = 'true'; - node.tagName = 'div'; - node.properties = {}; - // add node for button - const btnNode: ElementContent = { - type: 'element', - tagName: 'button', - properties: {}, - children: [], - position: node.position, - }; - node.children = [btnNode, preNode]; - } - }); - }; -} - -/** - * The part below is copied and adapted from: - * https://github.com/danny-avila/LibreChat/blob/main/client/src/utils/latex.ts - * (MIT License) - */ - -// Regex to check if the processed content contains any potential LaTeX patterns -const containsLatexRegex = - /\\\(.*?\\\)|\\\[.*?\\\]|\$.*?\$|\\begin\{equation\}.*?\\end\{equation\}/; - -// Regex for inline and block LaTeX expressions -const inlineLatex = new RegExp(/\\\((.+?)\\\)/, 'g'); -const blockLatex = new RegExp(/\\\[(.*?[^\\])\\\]/, 'gs'); - -// Function to restore code blocks -const restoreCodeBlocks = (content: string, codeBlocks: string[]) => { - return content.replace( - /<>/g, - (_, index) => codeBlocks[index] - ); -}; - -// Regex to identify code blocks and inline code -const codeBlockRegex = /(```[\s\S]*?```|`.*?`)/g; - -export const processLaTeX = (_content: string) => { - let content = _content; - // Temporarily replace code blocks and inline code with placeholders - const codeBlocks: string[] = []; - let index = 0; - content = content.replace(codeBlockRegex, (match) => { - codeBlocks[index] = match; - return `<>`; - }); - - // Escape dollar signs followed by a digit or space and digit - let processedContent = content.replace(/(\$)(?=\s?\d)/g, '\\$'); - - // If no LaTeX patterns are found, restore code blocks and return the processed content - if (!containsLatexRegex.test(processedContent)) { - return restoreCodeBlocks(processedContent, codeBlocks); - } - - // Convert LaTeX expressions to a markdown compatible format - processedContent = processedContent - .replace(inlineLatex, (_: string, equation: string) => `$${equation}$`) // Convert inline LaTeX - .replace(blockLatex, (_: string, equation: string) => `$$${equation}$$`); // Convert block LaTeX - - // Restore code blocks - return restoreCodeBlocks(processedContent, codeBlocks); -}; - -/** - * Preprocesses LaTeX content by replacing delimiters and escaping certain characters. - * - * @param content The input string containing LaTeX expressions. - * @returns The processed string with replaced delimiters and escaped characters. - */ -export function preprocessLaTeX(content: string): string { - // Step 1: Protect code blocks - const codeBlocks: string[] = []; - content = content.replace(/(```[\s\S]*?```|`[^`\n]+`)/g, (_, code) => { - codeBlocks.push(code); - return `<>`; - }); - - // Step 2: Protect existing LaTeX expressions - const latexExpressions: string[] = []; - - // Protect block math ($$...$$), \[...\], and \(...\) as before. - content = content.replace( - /(\$\$[\s\S]*?\$\$|\\\[[\s\S]*?\\\]|\\\(.*?\\\))/g, - (match) => { - latexExpressions.push(match); - return `<>`; - } - ); - - // Protect inline math ($...$) only if it does NOT match a currency pattern. - // We assume a currency pattern is one where the inner content is purely numeric (with optional decimals). - content = content.replace(/\$([^$]+)\$/g, (match, inner) => { - if (/^\s*\d+(?:\.\d+)?\s*$/.test(inner)) { - // This looks like a currency value (e.g. "$123" or "$12.34"), - // so don't protect it. - return match; - } else { - // Otherwise, treat it as a LaTeX expression. - latexExpressions.push(match); - return `<>`; - } - }); - - // Step 3: Escape dollar signs that are likely currency indicators. - // (Now that inline math is protected, this will only escape dollars not already protected) - content = content.replace(/\$(?=\d)/g, '\\$'); - - // Step 4: Restore LaTeX expressions - content = content.replace( - /<>/g, - (_, index) => latexExpressions[parseInt(index)] - ); - - // Step 5: Restore code blocks - content = content.replace( - /<>/g, - (_, index) => codeBlocks[parseInt(index)] - ); - - // Step 6: Apply additional escaping functions - content = escapeBrackets(content); - content = escapeMhchem(content); - - return content; -} - -export function escapeBrackets(text: string): string { - const pattern = - /(```[\S\s]*?```|`.*?`)|\\\[([\S\s]*?[^\\])\\]|\\\((.*?)\\\)/g; - return text.replace( - pattern, - ( - match: string, - codeBlock: string | undefined, - squareBracket: string | undefined, - roundBracket: string | undefined - ): string => { - if (codeBlock != null) { - return codeBlock; - } else if (squareBracket != null) { - return `$$${squareBracket}$$`; - } else if (roundBracket != null) { - return `$${roundBracket}$`; - } - return match; - } - ); -} - -export function escapeMhchem(text: string) { - return text.replaceAll('$\\ce{', '$\\\\ce{').replaceAll('$\\pu{', '$\\\\pu{'); -} diff --git a/tools/server/webui/src/components/ModalProvider.tsx b/tools/server/webui/src/components/ModalProvider.tsx deleted file mode 100644 index f2ebf8e0a7..0000000000 --- a/tools/server/webui/src/components/ModalProvider.tsx +++ /dev/null @@ -1,151 +0,0 @@ -import React, { createContext, useState, useContext } from 'react'; - -type ModalContextType = { - showConfirm: (message: string) => Promise; - showPrompt: ( - message: string, - defaultValue?: string - ) => Promise; - showAlert: (message: string) => Promise; -}; -const ModalContext = createContext(null!); - -interface ModalState { - isOpen: boolean; - message: string; - defaultValue?: string; - resolve: ((value: T) => void) | null; -} - -export function ModalProvider({ children }: { children: React.ReactNode }) { - const [confirmState, setConfirmState] = useState>({ - isOpen: false, - message: '', - resolve: null, - }); - const [promptState, setPromptState] = useState< - ModalState - >({ isOpen: false, message: '', resolve: null }); - const [alertState, setAlertState] = useState>({ - isOpen: false, - message: '', - resolve: null, - }); - const inputRef = React.useRef(null); - - const showConfirm = (message: string): Promise => { - return new Promise((resolve) => { - setConfirmState({ isOpen: true, message, resolve }); - }); - }; - - const showPrompt = ( - message: string, - defaultValue?: string - ): Promise => { - return new Promise((resolve) => { - setPromptState({ isOpen: true, message, defaultValue, resolve }); - }); - }; - - const showAlert = (message: string): Promise => { - return new Promise((resolve) => { - setAlertState({ isOpen: true, message, resolve }); - }); - }; - - const handleConfirm = (result: boolean) => { - confirmState.resolve?.(result); - setConfirmState({ isOpen: false, message: '', resolve: null }); - }; - - const handlePrompt = (result?: string) => { - promptState.resolve?.(result); - setPromptState({ isOpen: false, message: '', resolve: null }); - }; - - const handleAlertClose = () => { - alertState.resolve?.(); - setAlertState({ isOpen: false, message: '', resolve: null }); - }; - - return ( - - {children} - - {/* Confirm Modal */} - {confirmState.isOpen && ( - -
-

{confirmState.message}

-
- - -
-
- - )} - - {/* Prompt Modal */} - {promptState.isOpen && ( - -
-

{promptState.message}

- { - if (e.key === 'Enter') { - handlePrompt((e.target as HTMLInputElement).value); - } - }} - /> -
- - -
-
- - )} - - {/* Alert Modal */} - {alertState.isOpen && ( - -
-

{alertState.message}

-
- -
-
- - )} - - ); -} - -export function useModals() { - const context = useContext(ModalContext); - if (!context) throw new Error('useModals must be used within ModalProvider'); - return context; -} diff --git a/tools/server/webui/src/components/SettingDialog.tsx b/tools/server/webui/src/components/SettingDialog.tsx deleted file mode 100644 index 45a8d73b00..0000000000 --- a/tools/server/webui/src/components/SettingDialog.tsx +++ /dev/null @@ -1,553 +0,0 @@ -import { useState } from 'react'; -import { useAppContext } from '../utils/app.context'; -import { CONFIG_DEFAULT, CONFIG_INFO } from '../Config'; -import { isDev } from '../Config'; -import StorageUtils from '../utils/storage'; -import { classNames, isBoolean, isNumeric, isString } from '../utils/misc'; -import { - BeakerIcon, - ChatBubbleOvalLeftEllipsisIcon, - Cog6ToothIcon, - FunnelIcon, - HandRaisedIcon, - SquaresPlusIcon, -} from '@heroicons/react/24/outline'; -import { OpenInNewTab } from '../utils/common'; -import { useModals } from './ModalProvider'; - -type SettKey = keyof typeof CONFIG_DEFAULT; - -const BASIC_KEYS: SettKey[] = [ - 'temperature', - 'top_k', - 'top_p', - 'min_p', - 'max_tokens', -]; -const SAMPLER_KEYS: SettKey[] = [ - 'dynatemp_range', - 'dynatemp_exponent', - 'typical_p', - 'xtc_probability', - 'xtc_threshold', -]; -const PENALTY_KEYS: SettKey[] = [ - 'repeat_last_n', - 'repeat_penalty', - 'presence_penalty', - 'frequency_penalty', - 'dry_multiplier', - 'dry_base', - 'dry_allowed_length', - 'dry_penalty_last_n', -]; - -enum SettingInputType { - SHORT_INPUT, - LONG_INPUT, - CHECKBOX, - CUSTOM, -} - -interface SettingFieldInput { - type: Exclude; - label: string | React.ReactElement; - help?: string | React.ReactElement; - key: SettKey; -} - -interface SettingFieldCustom { - type: SettingInputType.CUSTOM; - key: SettKey; - component: - | string - | React.FC<{ - value: string | boolean | number; - onChange: (value: string) => void; - }>; -} - -interface SettingSection { - title: React.ReactElement; - fields: (SettingFieldInput | SettingFieldCustom)[]; -} - -const ICON_CLASSNAME = 'w-4 h-4 mr-1 inline'; - -const SETTING_SECTIONS: SettingSection[] = [ - { - title: ( - <> - - General - - ), - fields: [ - { - type: SettingInputType.SHORT_INPUT, - label: 'API Key', - key: 'apiKey', - }, - { - type: SettingInputType.LONG_INPUT, - label: 'System Message (will be disabled if left empty)', - key: 'systemMessage', - }, - ...BASIC_KEYS.map( - (key) => - ({ - type: SettingInputType.SHORT_INPUT, - label: key, - key, - }) as SettingFieldInput - ), - { - type: SettingInputType.SHORT_INPUT, - label: 'Paste length to file', - key: 'pasteLongTextToFileLen', - }, - { - type: SettingInputType.CHECKBOX, - label: 'Parse PDF as image instead of text', - key: 'pdfAsImage', - }, - ], - }, - { - title: ( - <> - - Samplers - - ), - fields: [ - { - type: SettingInputType.SHORT_INPUT, - label: 'Samplers queue', - key: 'samplers', - }, - ...SAMPLER_KEYS.map( - (key) => - ({ - type: SettingInputType.SHORT_INPUT, - label: key, - key, - }) as SettingFieldInput - ), - ], - }, - { - title: ( - <> - - Penalties - - ), - fields: PENALTY_KEYS.map((key) => ({ - type: SettingInputType.SHORT_INPUT, - label: key, - key, - })), - }, - { - title: ( - <> - - Reasoning - - ), - fields: [ - { - type: SettingInputType.CHECKBOX, - label: 'Expand thought process by default when generating messages', - key: 'showThoughtInProgress', - }, - { - type: SettingInputType.CHECKBOX, - label: - 'Exclude thought process when sending requests to API (Recommended for DeepSeek-R1)', - key: 'excludeThoughtOnReq', - }, - ], - }, - { - title: ( - <> - - Advanced - - ), - fields: [ - { - type: SettingInputType.CUSTOM, - key: 'custom', // dummy key, won't be used - component: () => { - const debugImportDemoConv = async () => { - const res = await fetch('/demo-conversation.json'); - const demoConv = await res.json(); - StorageUtils.remove(demoConv.id); - for (const msg of demoConv.messages) { - StorageUtils.appendMsg(demoConv.id, msg); - } - }; - return ( - - ); - }, - }, - { - type: SettingInputType.CHECKBOX, - label: 'Show tokens per second', - key: 'showTokensPerSecond', - }, - { - type: SettingInputType.LONG_INPUT, - label: ( - <> - Custom JSON config (For more info, refer to{' '} - - server documentation - - ) - - ), - key: 'custom', - }, - ], - }, - { - title: ( - <> - - Experimental - - ), - fields: [ - { - type: SettingInputType.CUSTOM, - key: 'custom', // dummy key, won't be used - component: () => ( - <> -

- Experimental features are not guaranteed to work correctly. -
-
- If you encounter any problems, create a{' '} - - Bug (misc.) - {' '} - report on Github. Please also specify webui/experimental on - the report title and include screenshots. -
-
- Some features may require packages downloaded from CDN, so they - need internet connection. -

- - ), - }, - { - type: SettingInputType.CHECKBOX, - label: ( - <> - Enable Python interpreter -
- - This feature uses{' '} - pyodide, - downloaded from CDN. To use this feature, ask the LLM to generate - Python code inside a Markdown code block. You will see a "Run" - button on the code block, near the "Copy" button. - - - ), - key: 'pyIntepreterEnabled', - }, - ], - }, -]; - -export default function SettingDialog({ - show, - onClose, -}: { - show: boolean; - onClose: () => void; -}) { - const { config, saveConfig } = useAppContext(); - const [sectionIdx, setSectionIdx] = useState(0); - - // clone the config object to prevent direct mutation - const [localConfig, setLocalConfig] = useState( - JSON.parse(JSON.stringify(config)) - ); - const { showConfirm, showAlert } = useModals(); - - const resetConfig = async () => { - if (await showConfirm('Are you sure you want to reset all settings?')) { - setLocalConfig(CONFIG_DEFAULT); - } - }; - - const handleSave = async () => { - // copy the local config to prevent direct mutation - const newConfig: typeof CONFIG_DEFAULT = JSON.parse( - JSON.stringify(localConfig) - ); - // validate the config - for (const key in newConfig) { - const value = newConfig[key as SettKey]; - const mustBeBoolean = isBoolean(CONFIG_DEFAULT[key as SettKey]); - const mustBeString = isString(CONFIG_DEFAULT[key as SettKey]); - const mustBeNumeric = isNumeric(CONFIG_DEFAULT[key as SettKey]); - if (mustBeString) { - if (!isString(value)) { - await showAlert(`Value for ${key} must be string`); - return; - } - } else if (mustBeNumeric) { - const trimmedValue = value.toString().trim(); - const numVal = Number(trimmedValue); - if (isNaN(numVal) || !isNumeric(numVal) || trimmedValue.length === 0) { - await showAlert(`Value for ${key} must be numeric`); - return; - } - // force conversion to number - // @ts-expect-error this is safe - newConfig[key] = numVal; - } else if (mustBeBoolean) { - if (!isBoolean(value)) { - await showAlert(`Value for ${key} must be boolean`); - return; - } - } else { - console.error(`Unknown default type for key ${key}`); - } - } - if (isDev) console.log('Saving config', newConfig); - saveConfig(newConfig); - onClose(); - }; - - const onChange = (key: SettKey) => (value: string | boolean) => { - // note: we do not perform validation here, because we may get incomplete value as user is still typing it - setLocalConfig({ ...localConfig, [key]: value }); - }; - - return ( - -
-

Settings

-
- {/* Left panel, showing sections - Desktop version */} -
- {SETTING_SECTIONS.map((section, idx) => ( - - ))} -
- - {/* Left panel, showing sections - Mobile version */} - {/* This menu is skipped on a11y, otherwise it's repeated the desktop version */} -
-
- - {SETTING_SECTIONS[sectionIdx].title} - -
    - {SETTING_SECTIONS.map((section, idx) => ( -
    setSectionIdx(idx)} - dir="auto" - > - {section.title} -
    - ))} -
-
-
- - {/* Right panel, showing setting fields */} -
- {SETTING_SECTIONS[sectionIdx].fields.map((field, idx) => { - const key = `${sectionIdx}-${idx}`; - if (field.type === SettingInputType.SHORT_INPUT) { - return ( - - ); - } else if (field.type === SettingInputType.LONG_INPUT) { - return ( - - ); - } else if (field.type === SettingInputType.CHECKBOX) { - return ( - - ); - } else if (field.type === SettingInputType.CUSTOM) { - return ( -
- {typeof field.component === 'string' - ? field.component - : field.component({ - value: localConfig[field.key], - onChange: onChange(field.key), - })} -
- ); - } - })} - -

- Settings are saved in browser's localStorage -

-
-
- -
- - - -
-
- - ); -} - -function SettingsModalLongInput({ - configKey, - value, - onChange, - label, -}: { - configKey: SettKey; - value: string; - onChange: (value: string) => void; - label?: string; -}) { - return ( -