flash-attention/csrc/flash_attn
Tri Dao 7fc39832e2 Use block_size=128 for headdim=128 on SM80
Previously we were using block_size=256.
2022-10-21 13:19:54 -07:00
..
cutlass@319a389f42 Add Cutlass as submodule 2022-06-02 09:54:16 -07:00
src Use block_size=128 for headdim=128 on SM80 2022-10-21 13:19:54 -07:00
fmha_api.cpp Split fwd on the seqlen_q dimension 2022-10-21 12:04:27 -07:00