Skip to content

MXFP4 x BF16 CUTLASS MoE backend perf and profiling improvement on Hopper#12451

Closed
StudyingShao wants to merge 6 commits into
NVIDIA:mainfrom
StudyingShao:jiangs/1.3.0rc3/opt_hopper_mix_dtype_moe
Closed

MXFP4 x BF16 CUTLASS MoE backend perf and profiling improvement on Hopper#12451
StudyingShao wants to merge 6 commits into
NVIDIA:mainfrom
StudyingShao:jiangs/1.3.0rc3/opt_hopper_mix_dtype_moe

Add 4-bit weights interleave functions

79315f6
Select commit
Loading
Failed to load commit list.
DCO / DCO succeeded Mar 23, 2026 in 1s

DCO

All commits are signed off!

View more details on DCO