CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-31 08:51:55 +00:00
Code Issues Packages Projects Releases Wiki Activity

metal : fix compute pass descriptor autorelease crash (#9718)

Browse Source
This commit is contained in:
Jack Mousseau
2024-10-03 11:01:46 -07:00
committed by GitHub
parent a7ad553513
commit 5d5ab1e5cc
1 changed files with 1 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
ggml/src/ggml-metal.m
View File

@@ -217,8 +217,6 @@ struct ggml_backend_metal_context {
id<MTLDevice> device; id<MTLDevice> device;
id<MTLCommandQueue> queue; id<MTLCommandQueue> queue;
MTLComputePassDescriptor * edesc;
dispatch_queue_t d_queue; dispatch_queue_t d_queue;
struct ggml_metal_kernel kernels[GGML_METAL_KERNEL_TYPE_COUNT]; struct ggml_metal_kernel kernels[GGML_METAL_KERNEL_TYPE_COUNT];
@@ -304,8 +302,6 @@ static struct ggml_backend_metal_context * ggml_metal_init(void) {
struct ggml_backend_metal_context * ctx = calloc(1, sizeof(struct ggml_backend_metal_context)); struct ggml_backend_metal_context * ctx = calloc(1, sizeof(struct ggml_backend_metal_context));
ctx->device = device; ctx->device = device;
ctx->queue = [ctx->device newCommandQueue]; ctx->queue = [ctx->device newCommandQueue];
ctx->edesc = MTLComputePassDescriptor.computePassDescriptor;
ctx->edesc.dispatchType = MTLDispatchTypeSerial;
ctx->d_queue = dispatch_queue_create("ggml-metal", DISPATCH_QUEUE_CONCURRENT); ctx->d_queue = dispatch_queue_create("ggml-metal", DISPATCH_QUEUE_CONCURRENT);
id<MTLLibrary> metal_library; id<MTLLibrary> metal_library;
@@ -3016,7 +3012,7 @@ static enum ggml_status ggml_metal_graph_compute(
const int n_nodes_per_cb = ctx->n_nodes_per_cb; const int n_nodes_per_cb = ctx->n_nodes_per_cb;
id<MTLCommandBuffer> command_buffer = ctx->command_buffers[cb_idx]; id<MTLCommandBuffer> command_buffer = ctx->command_buffers[cb_idx];
id<MTLComputeCommandEncoder> encoder = [command_buffer computeCommandEncoderWithDescriptor: ctx->edesc]; id<MTLComputeCommandEncoder> encoder = [command_buffer computeCommandEncoder];
int node_start = 0; int node_start = 0;
int node_end = n_nodes_0; int node_end = n_nodes_0;