Skip to content

CUDA: Faster FlashAttention kernel #6374

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 10 commits into from
Apr 2, 2024
Merged

CUDA: Faster FlashAttention kernel #6374

Show file tree
Hide file tree
Changes from 7 commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
912a6aa
CUDA: faster FlashAttention, kernel for bs == 1
JohannesGaessler Mar 29, 2024
e3794ef
16 cols for Phi-2
JohannesGaessler Mar 30, 2024
3a44bfe
no vec for hs, no hs==256 ncols==32 for Volta
JohannesGaessler Mar 30, 2024
824adc9
adjust kernel selection logic
JohannesGaessler Mar 31, 2024
4a4e549
4 warps, 256 stride for all D
JohannesGaessler Mar 31, 2024
b3e66f1
no ncols == 64
JohannesGaessler Apr 1, 2024
822d338
Multiple parallel blocks for batch size 1
JohannesGaessler Apr 1, 2024
ef7282a
fix compile warnings
JohannesGaessler Apr 2, 2024
46968c9
fix excessive KQ_b loads
JohannesGaessler Apr 2, 2024
b446893
fix cmake build
JohannesGaessler Apr 2, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions ggml-cuda/common.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -141,6 +141,7 @@
#define CC_PASCAL 600
#define MIN_CC_DP4A 610 // minimum compute capability for __dp4a, an intrinsic for byte-wise dot products
#define CC_VOLTA 700
#define CC_AMPERE 800
#define CC_OFFSET_AMD 1000000
#define CC_RDNA1 (CC_OFFSET_AMD + 1010)
#define CC_RDNA2 (CC_OFFSET_AMD + 1030)
Expand Down
Loading