CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-27 08:21:30 +00:00
Code Issues Packages Projects Releases Wiki Activity

ci : migrate ggml ci to self-hosted runners (#16116)

Browse Source 
* ci : migrate ggml ci to a self-hosted runners

* ci : add T4 runner

* ci : add instructions for adding self-hosted runners

* ci : disable test-backend-ops from debug builds due to slowness

* ci : add AMD V710 runner (vulkan)

* cont : add ROCM workflow

* ci : switch to qwen3 0.6b model

* cont : fix the context size
This commit is contained in:
Georgi Gerganov
2025-09-21 16:50:45 +03:00
committed by GitHub
parent 1eeb523c3e
commit 28baac9c9f
4 changed files with 295 additions and 372 deletions
Download Patch File Download Diff File Expand all files Collapse all files

192
.github/workflows/build.yml vendored
View File

@@ -1247,3 +1247,195 @@ jobs:
-DGGML_CANN=on \
-DSOC_TYPE=${{ matrix.device }}
cmake --build build -j $(nproc)
# TODO: simplify the following workflows using a matrix
# TODO: run lighter CI on PRs and the full CI only on master (if needed)
ggml-ci-x64-cpu-low-perf:
runs-on: [self-hosted, Linux, X64, CPU, low-perf]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
ggml-ci-arm64-cpu-low-perf:
runs-on: [self-hosted, Linux, ARM64, CPU, low-perf]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
ggml-ci-x64-cpu-high-perf:
runs-on: [self-hosted, Linux, X64, CPU, high-perf]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
ggml-ci-arm64-cpu-high-perf:
runs-on: [self-hosted, Linux, ARM64, CPU, high-perf]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
ggml-ci-x64-nvidia-v100-cuda:
runs-on: [self-hosted, Linux, X64, NVIDIA, V100]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
nvidia-smi
GG_BUILD_CUDA=1 bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
ggml-ci-x64-nvidia-v100-vulkan:
runs-on: [self-hosted, Linux, X64, NVIDIA, V100]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
vulkaninfo
GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
ggml-ci-x64-nvidia-t4-cuda:
runs-on: [self-hosted, Linux, X64, NVIDIA, T4]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
nvidia-smi
GG_BUILD_CUDA=1 bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
ggml-ci-x64-nvidia-t4-vulkan:
runs-on: [self-hosted, Linux, X64, NVIDIA, T4]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
vulkaninfo
GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
ggml-ci-x64-nvidia-t4-vulkan-coopmat1:
runs-on: [self-hosted, Linux, X64, NVIDIA, T4]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
vulkaninfo
GG_BUILD_VULKAN=1 GGML_VK_DISABLE_COOPMAT2=1 bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
ggml-ci-x64-cpu-amx:
runs-on: [self-hosted, Linux, X64, CPU, AMX]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
ggml-ci-x64-amd-v710-vulkan:
runs-on: [self-hosted, Linux, X64, AMD, V710]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
vulkaninfo
GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
ggml-ci-x64-amd-v710-rocm:
runs-on: [self-hosted, Linux, X64, AMD, V710]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
vulkaninfo
GG_BUILD_ROCM=1 GG_BUILD_AMDGPU_TARGETS="gfx1101" bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
ggml-ci-mac-metal:
runs-on: [self-hosted, macOS, ARM64]
steps:
- name: Clone
id: checkout
uses: actions/checkout@v4
- name: Test
id: ggml-ci
run: |
GG_BUILD_METAL=1 bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp
# TODO: install vulkan drivers
# ggml-ci-mac-vulkan:
# runs-on: [self-hosted, macOS, ARM64]
#
# steps:
# - name: Clone
# id: checkout
# uses: actions/checkout@v4
#
# - name: Test
# id: ggml-ci
# run: |
# GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp