flash-attention-rdna2

Exact FlashAttention for AMD RDNA2 (gfx1030 / RX 6800 XT), written in HIP. Drop-in replacement for flash-attn: 19.8 TFLOP/s forward, decode at 99% of HBM bandwidth.

amd

View on GitHub

3 Stars

0 Forks

0 Watchers

HIP Language

apache-2.0 License

42.1 SrcLog Score

Cost to Build

$4.9K

Market Value

$3.5K

How is this calculated?

Growth over time

2 data points · 2026-08-09 → 2026-09-06

Stars Forks Watchers

💬

How do you feel about this project?

Ask AI about flash-attention-rdna2

Question copied to clipboard

What is the 2kiss/flash-attention-rdna2 GitHub project? Description: "Exact FlashAttention for AMD RDNA2 (gfx1030 / RX 6800 XT), written in HIP. Drop-in replacement for flash-attn: 19.8 TFLOP/s forward, decode at 99% of HBM bandwidth.". Written in HIP. Explain what it does, its main use cases, key features, and who would benefit from using it.

Question is copied to clipboard — paste it after the AI opens.

How to clone flash-attention-rdna2

Clone via HTTPS

git clone https://github.com/2kiss/flash-attention-rdna2.git

Clone via SSH

[email protected]:2kiss/flash-attention-rdna2.git

Download ZIP

Download main.zip

Found an issue?

Report bugs or request features on the flash-attention-rdna2 issue tracker:

Open GitHub Issues

Similar to flash-attention-rdna2

webpack madge steal wufuc conditioner awesome-vulkan esl zydis staxrip pyopencl coriander lodjs easytimer.js ocl angular-async-loader node-dependency-tree purge-wrangler require-vuejs GatelessGateSharp SimpleSvm set-egpu parenchyma node-precinct TypeScript-AMD-Boilerplate ScanTree awesome-d3d12 Tensile generator-module-boilerplate akase romania-choropleth