Home › Research Papers › FlashAttention (2022)
FlashAttention (2022)
Research Paper
Systems
The kernel trick that made long context affordable.
You're looking at a research-paper record: a landmark paper and what it introduced to the field.
Tri Dao's FlashAttention reordered attention computation around GPU memory hierarchy — exact results, several times faster, memory linear instead of quadratic. Versions 2 and 3 chase each hardware generation; virtually every serious training and inference stack ships it. Systems research as leverage: one algorithm changed the economics of every cluster.
How this connects to the rest of the stack
This record's links at a glance — click a group to jump to it.
Related
Sources
Record reviewed as of 2026-01-15 — see Methodology.