CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-05 09:36:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

metal : fix comments [no ci]

Browse Source
This commit is contained in:
Georgi Gerganov
2025-03-25 14:36:22 +02:00
parent 24a9ea8b44
commit 982c82f1e6
2 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/src/ggml-metal/ggml-metal-impl.h
View File

@@ -3,7 +3,7 @@
// kernel parameters for mat-vec threadgroups
//
// N_R0: number of src0 rows to process per threadgroup
// N_R0: number of src0 rows to process per simdgroup
// N_SG: number of simdgroups per threadgroup
//
// TODO: for optimal performance, become function of the device and work size

2
ggml/src/ggml-metal/ggml-metal.m
View File

@@ -2883,7 +2883,7 @@ static void ggml_metal_encode_node(
id<MTLComputePipelineState> pipeline = nil;
int nsg = 0; // number of simdgroups
int nr0 = 0; // number of src0 rows per threadgroup
int nr0 = 0; // number of src0 rows per simdgroup
int nr1 = 1; // number of src1 rows per threadgroup
size_t smem = 0; // shared memory