flash-attention/csrc/flash_attn_ck
rocking 88d1657a14
[AMD ROCm] Fix KVcache bug and improve performance (#1328)
* update ck

* update ck

* update ck again

* update ck

* use pointer as seed and offset

* update CK

* Remove useless "else"

* Fix page-attn block table read out-of-bound

---------

Co-authored-by: Po Yen, Chen <PoYen.Chen@amd.com>
2024-11-12 11:32:11 -08:00
..
flash_api.cpp Hotfix due to change of upstream api (#1239) 2024-09-20 12:45:25 -07:00
flash_common.cpp Support page kvcache in AMD ROCm (#1198) 2024-09-15 23:17:28 -07:00
flash_common.hpp [AMD ROCm] Fix KVcache bug and improve performance (#1328) 2024-11-12 11:32:11 -08:00
mha_bwd.cpp [AMD ROCm] Fix KVcache bug and improve performance (#1328) 2024-11-12 11:32:11 -08:00
mha_fwd_kvcache.cpp Support page kvcache in AMD ROCm (#1198) 2024-09-15 23:17:28 -07:00
mha_fwd.cpp [AMD ROCm] Fix KVcache bug and improve performance (#1328) 2024-11-12 11:32:11 -08:00
mha_varlen_bwd.cpp [AMD ROCm] Fix KVcache bug and improve performance (#1328) 2024-11-12 11:32:11 -08:00
mha_varlen_fwd.cpp [AMD ROCm] Fix KVcache bug and improve performance (#1328) 2024-11-12 11:32:11 -08:00