Add cutlass decode kernel to TritonBench#4853

Open

Aya-ZIbra wants to merge 3 commits intopytorch:mainfrom

Aya-ZIbra:export-D80041532

Commits on Sep 22, 2025

Disable SWA related decode kernel tests since they are not supported

Gefei Zuo
authored and
facebook-github-bot
committed
Support bf16 in blackwell cutlass decode attention kernel

Gefei Zuo
authored and
facebook-github-bot
committed

Commits on Sep 23, 2025

Add cutlass decode kernel to TritonBench (#4853)

Aya-ZIbra
authored and
facebook-github-bot
committed