Anav PrasadandJohannes Gäßler
88458164c7
* flash attention support for head dimension 512 added * FA D=512 - match 576 configs, limit ncols2, revert vec cap * fix HIP tile kernel build for D=512 * fix HIP tile kernel occupancy for D=512 on AMD * Apply suggestions from code review Co-authored-by: Johannes Gäßler <johannesg@5d6.de> * fix tile FA compilation --------- Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
13 lines
439 B
Plaintext
13 lines
439 B
Plaintext
// This file has been autogenerated by generate_cu_files.py, do not edit manually.
|
|
|
|
#include "../fattn-mma-f16.cuh"
|
|
|
|
DECL_FATTN_MMA_F16_CASE(64, 64, 8, 4);
|
|
DECL_FATTN_MMA_F16_CASE(80, 80, 8, 4);
|
|
DECL_FATTN_MMA_F16_CASE(96, 96, 8, 4);
|
|
DECL_FATTN_MMA_F16_CASE(112, 112, 8, 4);
|
|
DECL_FATTN_MMA_F16_CASE(128, 128, 8, 4);
|
|
DECL_FATTN_MMA_F16_CASE(256, 256, 8, 4);
|
|
DECL_FATTN_MMA_F16_CASE(512, 512, 8, 4);
|
|
DECL_FATTN_MMA_F16_CASE(576, 512, 8, 4);
|