Commit Graph
4 Commits
Author SHA1 Message Date
AesSedai 046e284437 Add flash attention MMA / Tiles to support MiMo-V2.5 (#22812)
* mimo-v2.5: add flash attention mma/tiles for for d_kq=192 d_v=128

* mimo-v2.5: follow (256, 256) fattn templates

* mimo-v2.5: cleanup comments

* mimo-v2.5: further comment cleanup

* mimo-v2.5: address PR feedback
fix GQA handling
check for other dangling 320/576 carveouts and mirror them for 192
Add to backend ops test so new paths are covered
2026-05-09 11:28:29 +08:00
Anav PrasadandJohannes Gäßler 88458164c7 CUDA: Add Flash Attention Support for Head Dimension 512 (#20998)
* 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>
2026-04-01 09:07:24 +02:00
Johannes Gäßler 0cf6725e9f CUDA: FA support for Deepseek (Ampere or newer) (#13306)
* CUDA: FA support for Deepseek (Ampere or newer)

* do loop unrolling via C++ template
2025-05-09 13:34:58 +02:00
Johannes Gäßler 5fa07c2f93 CUDA: optimize FA for GQA + large batches (#12014) 2025-02-22 12:20:17 +01:00