---
title: 'FlashAttention-3: Fast and Accurate Attention with Asynchrony and Low-precision'
license: CC-BY-4.0
license_url: https://creativecommons.org/licenses/by/4.0/
license_note: 'paper is CC BY 4.0 (verified on arXiv abs page 2026-11-12; icon by-4.0.png, http://creativecommons.org/licenses/by/4.0/); note is our own summary'
source: https://iinventedlight.com/library/notes/2407-08608-flashattention-3.md
source_license: 'CC BY 4.0'
attribution: 'Summary by iinventedlight (AI-assisted). Source: Jay Shah, Ganesh Bikshandi, Ying Zhang, Vijay Thakkar, Pradeep Ramani, Tri Dao, FlashAttention-3: Fast and Accurate Attention with Asynchrony and Low-precision (CC BY 4.0).'
author: 'iinventedlight (AI-assisted)'
date: 2026-11-12
status: published
approved_by: "Cole Withington, 2026-11-12 14:44 PT, Librarian chat widget t51s316: \"Yes, publish all 100 second-list arXiv notes\""
checked_on: 2026-11-12
as_of: 2026-11-12
confidence: medium
untested: 'Not reproduced. Abs page and HTML key sections read 2026-11-12. Named code URLs checked for HTTP 200; code not run.'
category: arxiv
topic_area: 'ai-research: efficiency'
tags: arxiv, efficiency, attention, gpu
arxiv_id: 2407.08608
alias: 2407.08608-flashattention-3
sources:
  - id: S1
    title: 'FlashAttention-3: Fast and Accurate Attention with Asynchrony and Low-precision'
    publisher: arXiv
    authors: 'Jay Shah, Ganesh Bikshandi, Ying Zhang, Vijay Thakkar, Pradeep Ramani, Tri Dao'
    url: https://arxiv.org/abs/2407.08608
    version_read: 'v2 (12 Jul 2024)'
    accessed: 2026-11-12
    license: 'CC BY 4.0'
---
# FlashAttention-3: Fast and Accurate Attention with Asynchrony and Low-precision

## Summary
FlashAttention-3 speeds up attention on Hopper GPUs with three main techniques: overlapping computation and data movement via warp-specialization, interleaving block-wise matmul and softmax, and block quantization plus incoherent processing that uses hardware support for FP8. They report only 35% utilization on the H100 GPU. They report that FlashAttention-3 achieves a 1.5-2.0x speedup on H100 GPUs, with FP16 reaching up to 740 TFLOPs/s (75% utilization) and FP8 reaching close to 1.2 PFLOPs/s. They also report that FP8 FlashAttention-3 achieves 2.6x lower numerical error than a baseline FP8 attention. [S1]

## Claims
- id: C1
  claim: 'The authors say FlashAttention-2 achieved only 35% utilization on the H100 GPU.'
  source: S1
  evidence: "FlashAttention-2 achieving only 35% utilization on the H100 GPU"
- id: C2
  claim: 'They report a 1.5-2.0x speedup on H100, with FP16 reaching up to 740 TFLOPs/s (75% utilization).'
  source: S1
  evidence: "with FP16 reaching up to 740 TFLOPs/s (75% utilization)"
- id: C3
  claim: 'They report that FP8 FlashAttention-3 reaches close to 1.2 PFLOPs/s.'
  source: S1
  evidence: "with FP8 reaching close to 1.2 PFLOPs/s"

## Verification
- Code available: Paper: https://github.com/Dao-AILab/flash-attention, which returned HTTP 200 on 2026-11-12. Not run.
- Reproduced: not reproduced (2026-11-12). Code not run.
- Related work: Live notes [2205-14135-flashattention](https://iinventedlight.com/library/notes/2205-14135-flashattention/) and [2307-08691-flashattention-2](https://iinventedlight.com/library/notes/2307-08691-flashattention-2/). This is the paper-verification note for FlashAttention-3 itself.
- Limits: Speed and utilization figures are author-reported on H100 GPUs (v2, 12 Jul 2024). This note does not re-benchmark the kernel. The abstract describes a 1.5-2.0x speedup range for the H100 results.
- License: CC BY 4.0 on the abs page (2026-11-12; icon by-4.0.png; http://creativecommons.org/licenses/by/4.0/). Submitted 11 Jul 2024 (v1); last revised 12 Jul 2024 (v2). Sources: [S1] https://arxiv.org/abs/2407.08608, v2 (12 Jul 2024). checked_on: 2026-11-12.

## When to apply
Use when asking how FlashAttention was adapted to Hopper asynchrony and FP8, or when comparing FA-2 versus FA-3 utilization claims.

## Facts
- id: F1
  claim: 'FlashAttention-3 targets Hopper GPUs with asynchrony, warp-specialization, and FP8 low-precision techniques.'
  confidence: medium
  date: 2026-11-12
  sources: [S1]
  license: CC BY 4.0
- id: F2
  claim: 'The authors report that FlashAttention-2 reached only 35% utilization on H100, versus up to 75% for FP16 FlashAttention-3.'
  confidence: medium
  date: 2026-11-12
  sources: [S1]
  license: CC BY 4.0
- id: F3
  claim: 'They report up to 740 TFLOPs/s in FP16 and close to 1.2 PFLOPs/s in FP8, with 2.6x lower numerical error than a baseline FP8 attention.'
  confidence: medium
  date: 2026-11-12
  sources: [S1]
  license: CC BY 4.0

## Sources
- [S1] arXiv 2407.08608, https://arxiv.org/abs/2407.08608. License: CC BY 4.0. v2 (12 Jul 2024). Accessed 2026-11-12.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Summary by iinventedlight (AI-assisted). Paper is CC BY 4.0.

---

Agents: [/agents.md](/agents.md) | [/.well-known/agent-library.json](/.well-known/agent-library.json).
