[ROCm] Initial AITER Integration for mha_bwd asm kernels (#152630)

Generates AITER plumbing via cmake. Calls into fav3 asm bwd CK kernels.

Update submodule composable kernel for this change

Pull Request resolved: https://github.com/pytorch/pytorch/pull/152630
Approved by: https://github.com/xw285cornell, https://github.com/yoyoyocmu
This commit is contained in:
Andy Lugo
2025-07-01 02:53:22 +00:00
committed by PyTorch MergeBot
parent f40efde2a4
commit b5ce77c1f5
8 changed files with 80 additions and 21 deletions