Skip to content

Kimi K3 CPU decode optimizations: MXFP4 kernels, parallel expert streaming, NUMA placement - #2

Open
neilcouture wants to merge 1 commit into
ferrumox:mainfrom
neilcouture:optimize-performance
Open

Kimi K3 CPU decode optimizations: MXFP4 kernels, parallel expert streaming, NUMA placement#2
neilcouture wants to merge 1 commit into
ferrumox:mainfrom
neilcouture:optimize-performance

Commits

Commits on Aug 2, 2026