Skip to content

GitLab

Unverified Commit 7ca751ff authored by kk's avatar kk Committed by GitHub
Browse files

Fused moe triton cfg opt for rocm (#2612) 


Co-authored-by: default avatarwunhuang <wunhuang@amd.com>
parent c75adfec
Hide whitespace changes
Inline Side-by-side
Showing with 14 additions and 14 deletions
python/sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=4096,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8.json
View file @ 7ca751ff
......@@ -55,35 +55,35 @@
"kpack": 2
},
"128": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 256,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_warps": 8,
"num_stages": 0,
"waves_per_eu": 1,
"waves_per_eu": 0,
"matrix_instr_nonkdim": 16,
"kpack": 1
},
"256": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 256,
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_warps": 8,
"num_stages": 0,
"waves_per_eu": 1,
"waves_per_eu": 0,
"matrix_instr_nonkdim": 16,
"kpack": 1
},
"512": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 256,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_warps": 8,
"num_stages": 0,
"waves_per_eu": 2,
"waves_per_eu": 0,
"matrix_instr_nonkdim": 16,
"kpack": 2
},
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment