DefTruth / ffpa-attn-mma

📚FFPA(Split-D): Yet another Faster Flash Prefill Attention with O(1) GPU SRAM complexity for headdim > 256, ~2x↑🎉vs SDPA EA.

Date Created 2024-11-29 (3 months ago)

Commits 246 (last one 5 days ago)

Stargazers 147 (5 this week)

Watchers 2 (0 this week)

Forks 6

License gpl-3.0

Ranking

RepositoryStats indexes 629,376 repositories, of these DefTruth/ffpa-attn-mma is ranked #233,888 (63rd percentile) for total stargazers, and #480,083 for total watchers. Github reports the primary language for this repository as Cuda, for repositories using this language it is ranked #155/391.

DefTruth/ffpa-attn-mma is also tagged with popular topics, for these it's ranked: cuda (#356/684)

Other Information

DefTruth/ffpa-attn-mma has Github issues enabled, there are 4 open issues and 1 closed issue.

There have been 9 releases, the latest one was published on 2025-03-17 (4 days ago) with the name v0.0.2.post4.

Homepage URL: https://zhuanlan.zhihu.com/p/13975660308

All Topics

mla cuda sdpa mlsys deepseek attention flash-mla fused-mla deepseek-r1 deepseek-v3 tensor-cores flash-attention

Star History

Github stargazers over time

Watcher History

Github watchers over time, collection started in '23

Recent Commit History

246 commits on the default branch (main) since jan '22

Yearly Commits

Commits to the default branch (main) per year

Issue History

Languages

The primary language is Cuda but there's also others...

Opengraph Image

updated: 2025-03-18 @ 09:11am, id: 896023715 / R_kgDONWg8ow