CUDA: use mma PTX instructions for FlashAttention (#11583)

Browse Source

* CUDA: use mma PTX instructions for FlashAttention

* __shfl_sync workaround for movmatrix

* add __shfl_sync to HIP

Co-authored-by: Diego Devesa <slarengh@gmail.com>

...

This commit is contained in:

Johannes Gäßler

2025-02-02 19:31:09 +01:00

committed by GitHub

parent 84ec8a58f7

commit 864a0b67a6

29 changed files with 2058 additions and 998 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download Patch File Download Diff File Expand all files Collapse all files

349

ggml/src/ggml-cuda/mmq.cuh

File diff suppressed because it is too large Load Diff