Files
composable_kernel/include/ck_tile/ops/fmha
ltqin 92653168c2 flashattention fwd add (80, 96) instance (#3415)
* add hdim (96,96) instance

* change to (80,96)

* format py

* remove 96 in optdim

* when N=6 change to llvm_amdgcn_raw_buffer_load_i32x3
2025-12-17 09:16:11 -08:00
..