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

metal: somewhat faster f16 x f32 matrix multiply kernel (#2951)

Browse Source 
* Somewhat faster f16 x f32 matrix multiply kernel

* Better use 32 thread groups for f16 x f32

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow
2023-09-01 11:15:57 +03:00
committed by GitHub
parent bce1fef328
commit e8d9158925
2 changed files with 29 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ggml-metal.m
View File

@@ -840,7 +840,7 @@ void ggml_metal_graph_compute(
switch (src0t) {
case GGML_TYPE_F16:
{
nth0 = 64;
nth0 = 32;
nth1 = 1;
[encoder setComputePipelineState:ctx->pipeline_mul_mat_f16_f32];
} break;