{"data":{"slug":"flashattention","etype":"paper","name":"FlashAttention (2022)","cat":"Systems","subtitle":"The kernel trick that made long context affordable.","summary":"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.","attrs":null,"kf":null,"aliases":["Flash Attention"],"ticker":null,"exchange":null,"hq":null,"founded":null,"website":null,"status":null,"region":null,"asof":"2026-01-15","sources":[{"url":"https://arxiv.org/abs/2205.14135","label":"arXiv 2205.14135"}]},"meta":{"found":true,"count":1,"served_from":"site cache (no live upstream calls)","as_of":"2026-10-14T06:34:05+00:00","disclaimer":"Research and information only — not investment advice. Structured records carry basis labels (vendor spec / reported / estimate); market quotes are delayed/indicative.","attribution":"Records compiled from vendor documentation, official filings and reputable public reporting; market quotes via financial data providers. See /data-sources."}}