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

[AMD] Minimax-M3 : unblock mxfp8 block convert on gfx950 (#32036)

Browse Source
This commit is contained in:
YC Yen-Ching Tseng
2026-07-30 02:43:27 -07:00
committed by GitHub
parent 4ba7d5ad93
commit 04d6fb4d6c
1 changed files with 7 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/sglang/srt/layers/quantization/fp8.py
+7 -1
View File
@@ -117,7 +117,13 @@ _is_fp8_fnuz = is_fp8_fnuz()
_is_gfx95_supported = is_gfx95_supported()
# gfx942 (MI300) has no MX matmul HW; MXFP8 checkpoints are converted to
# block-fp8 [128,128] at load and run through the native block-fp8 kernels.
_mxfp8_to_block_fp8_required = mxfp8_block_convert_required()
# SGLANG_FORCE_MXFP8_BLOCK_CONVERT=1 opts into that same block-fp8 path on
# gfx950 (MI35x): it routes the fp8 GEMMs / fused MoE through the mature aiter
# block-scale kernels instead of the native MX dot_scaled path (measured +20%
# throughput at equal accuracy on MiniMax-M3, GSM8K 0.9719 vs 0.9689).
_mxfp8_to_block_fp8_required = mxfp8_block_convert_required() or get_bool_env_var(
"SGLANG_FORCE_MXFP8_BLOCK_CONVERT"
)
_use_hip_int4 = get_bool_env_var("SGLANG_INT4_WEIGHT") and _is_hip
_use_aiter = envs.SGLANG_USE_AITER.get() and _is_hip
_is_shuffle_moe_mxfp4 = is_gfx95_supported()