minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

perf: add H200 Triton MoE configs for E256 N512 (#34150)

Browse Source 
Co-authored-by: Vedant Jhaveri <vjhaveri@linkedin.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This commit is contained in:
Vedant V Jhaveri
2026-08-14 14:37:58 -07:00
committed by GitHub
co-authored by Vedant Jhaveri Copilot
parent 22dde1dd5b
commit 03c1d58112
2 changed files with 50 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_7_1/E=256,N=512,device_name=NVIDIA_H200.json
+24
View File
@@ -0,0 +1,24 @@
{
"256": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1
},
"512": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 8,
"num_stages": 3
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3
}
}
python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_7_1/E=256,N=512,device_name=NVIDIA_H200_down.json
+26
View File
@@ -0,0 +1,26 @@
{
"256": {
"BLOCK_SIZE_M": 16,
"BLOCK_SIZE_N": 32,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1
},
"512": {
"BLOCK_SIZE_M": 32,
"BLOCK_SIZE_N": 64,
"BLOCK_SIZE_K": 128,
"GROUP_SIZE_M": 32,
"num_warps": 4,
"num_stages": 3,
"USE_TMA": true
},
"1024": {
"BLOCK_SIZE_M": 64,
"BLOCK_SIZE_N": 128,
"BLOCK_SIZE_K": 64,
"GROUP_SIZE_M": 1,
"num_warps": 4,
"num_stages": 3,
"USE_TMA": true
}
}