Hansung Kim 068d48534e flash: Swap S1/S0 to avoid GEMM II - softmax bank conflict ...

+ remove spurrious fences to better overlap GEMM I and DMA

2024-09-11 00:55:36 -07:00

..

.gitignore

flash: gitignore

2024-08-15 21:04:59 -07:00

common.h

flash: Change kernel arg to contain qkv; strip stimulus gen from host code

2024-08-15 21:03:02 -07:00

flash_impl.hpp

flash: Add flag in SMEM for dependency check on O

2024-09-10 13:42:47 -07:00

half.hpp

Add flash attention kernel skeleton

2024-08-14 20:46:09 -07:00

kernel.cpp

flash: Add Gemmini-accelerated kernel

2024-09-07 22:40:58 -07:00

kernel.gemmini.cpp

flash: Swap S1/S0 to avoid GEMM II - softmax bank conflict

2024-09-11 00:55:36 -07:00

main.cpp

flash: Fix load addr for V tile; test with seqlen=128

2024-08-20 14:34:09 -07:00

Makefile

flash: Fix online softmax for DMA layout

2024-09-07 23:21:28 -07:00