llama.cpp
CUDA: generalized (mma) FA, add Volta support
#17505
Merged

CUDA: generalized (mma) FA, add Volta support #17505

JohannesGaessler
JohannesGaessler JohannesGaessler requested a review from am17an am17an 38 days ago
JohannesGaessler JohannesGaessler requested a review from ggerganov ggerganov 38 days ago
github-actions github-actions added Nvidia GPU
github-actions github-actions added ggml
JohannesGaessler JohannesGaessler changed the title CUDA: ganeralized (mma) FA, add Volta support CUDA: generalized (mma) FA, add Volta support 38 days ago
JohannesGaessler JohannesGaessler force pushed 38 days ago
JohannesGaessler JohannesGaessler force pushed to 2ef0c5f6 38 days ago
zhang-hui-yulo
JohannesGaessler CUDA: generalized (mma) FA, add Volta support
17f191e9
JohannesGaessler fix const correctness
e2c50b1d
JohannesGaessler fix turing config lookup
301ae300
JohannesGaessler JohannesGaessler force pushed from b92e6f86 to 301ae300 35 days ago
JohannesGaessler refactor template parameters
13500e8c
Hedede
JohannesGaessler adjust kernel selection logic
394ced5f
JohannesGaessler
JohannesGaessler fix trailing whitespace
3e1ca0c6
Hedede
Hedede
JohannesGaessler fix kernel selection logic
ec176eef
ggerganov
am17an
am17an commented on 2025-12-02
JohannesGaessler
am17an
ggerganov
ggerganov approved these changes on 2025-12-02
am17an
am17an approved these changes on 2025-12-02
use struct for MMA FA kernel config
d861a34e
JohannesGaessler JohannesGaessler force pushed from 323c6830 to d861a34e 30 days ago
JohannesGaessler JohannesGaessler merged 2e1c9cd8 into master 30 days ago
LostRuins
JohannesGaessler
LostRuins
LostRuins commented on 2025-12-06

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone