Widen Triton TBE forward gather issue (#4490) - #4490

Closed

stashuk-olek wants to merge 7 commits into

meta-pytorch:mainfrom

stashuk-olek:export-D114280947

Commits on Aug 8, 2026

Select CUDA v2 bounds checking for Triton TBE (#4484)

stashuk-olek
authored and
facebook-github-bot
committed
67db5dd
Fuse Triton TBE forward bounds checking (#4485)

stashuk-olek
authored and
facebook-github-bot
committed
0702e0b
Optimize Triton TBE forward for small long-bag tables (#4486)

stashuk-olek
authored and
facebook-github-bot
committed
2893e5b
Reduce Triton TBE backward reduction pressure (#4487)

stashuk-olek
authored and
facebook-github-bot
committed
a90ee76
Reduce Triton TBE unweighted backward reduction width (#4488)

stashuk-olek
authored and
facebook-github-bot
committed
b1e0972
Coarsen Triton TBE forward bag scheduling (#4489)

stashuk-olek
authored and
facebook-github-bot
committed
9023ffe
Widen Triton TBE forward gather issue (#4490)

stashuk-olek
authored and
facebook-github-bot
committed
d9882fa