EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

Tune MI300X fused MoE Triton kernel JSON config. (#3492)

Browse Source
This commit is contained in:
Wen-Heng (Jack) Chung
2025-02-11 12:27:25 -06:00
committed by GitHub
parent bb418ced80
commit cadd5dbe6a
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
python/sglang/srt/layers/moe/fused_moe_triton/configs/E=256,N=256,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8,block_shape=[128, 128].json
View File

@@ -72,10 +72,10 @@
"waves_per_eu": 0
},
"64": {
"BLOCK_SIZE_M": 256,
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"GROUP_SIZE_M": 4,
"num_warps": 4,
"num_stages": 2,
"waves_per_eu": 0