Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

opencl: add q8_0 mm support (#16469)

Browse Source 
* opencl: add mm_q8_0_f32

* opencl: fix data loading for incomplete tile

* opencl: use q8_0 mm for larger matrix

* opencl: add some tests to cover the path
This commit is contained in:
lhez
2025-10-15 10:51:04 -07:00
committed by GitHub
parent d93f8439b0
commit 0cb7a0683b
6 changed files with 271 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-opencl/CMakeLists.txt
+1
View File
@@ -93,6 +93,7 @@ set(GGML_OPENCL_KERNELS
mul_mv_id_mxfp4_f32_flat
mul_mm_f32_f32_l4_lm
mul_mm_f16_f32_l4_lm
mul_mm_q8_0_f32_l4_lm
mul
norm
relu