[sgl-kernel][Feat][B200][1/N] Support MXFP8 Grouped GEMM in Blackwell (#13731)

Co-authored-by: Yineng Zhang <me@zhyncs.com>
This commit is contained in:
Qi Yuhang
2025-12-04 10:09:09 +08:00
committed by GitHub
parent df026bb110
commit 16ff892c18
12 changed files with 1174 additions and 1 deletions

View File

@@ -32,7 +32,11 @@ from sgl_kernel.elementwise import (
rmsnorm,
silu_and_mul,
)
from sgl_kernel.expert_specialization import es_fp8_blockwise_scaled_grouped_mm
from sgl_kernel.expert_specialization import (
es_fp8_blockwise_scaled_grouped_mm,
es_sm100_mxfp8_blockscaled_grouped_mm,
es_sm100_mxfp8_blockscaled_grouped_quant,
)
from sgl_kernel.fused_moe import moe_wna16_marlin_gemm
from sgl_kernel.gemm import (
awq_dequantize,