# HazyResearch/ThunderKittens

Tile primitives for speedy kernels

Repository: https://github.com/HazyResearch/ThunderKittens
Canonical: https://ross.abutalabs.com/products/thunderkittens
Language: Cuda
License: MIT
License Family: permissive
Last push: 2026-08-25T00:26:53+00:00

## Health v2 (maintenance only)
Score: 70/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 35, longevity 65
- inputs: {"age_days": 912, "days_push": 9, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3659, forks 318 (observed 2026-08-28T04:08:13.269345+00:00)

## What it is
ThunderKittens is a C++/CUDA framework of tile-based primitives for writing fast deep learning GPU kernels. It embeds natively into CUDA so developers can build high-performance kernels like FlashAttention-3 with simple, extensible code.

## Use cases
- write fast CUDA kernels for deep learning
- implement attention kernels like FlashAttention
- optimize matrix multiply operations on NVIDIA GPUs
- build custom GPU kernels for LLM training and inference
- prototype high-performance kernels without writing raw CUDA
- use low-precision formats like MXFP8 and NVFP4 on Blackwell GPUs

## When to choose
- you need to write custom high-performance CUDA kernels for NVIDIA GPUs
- you want kernel performance comparable to hand-tuned implementations with simpler code
- you target modern NVIDIA architectures like Hopper or Blackwell
- you need tile-level primitives for matrix operations in deep learning workloads

## When to avoid
- you target AMD GPUs (use HipKittens instead)
- you rely on Ampere GPU support, which is no longer actively developed
- you want a Python package with pip installation, since kernels must now be compiled individually
- you need a general-purpose linear algebra library rather than kernel primitives

## Facets
- artifact type: framework
- maturity: active
- function: gpu-computing, deep-learning, machine-learning, llm-inference, llm-training
- domain: deep-learning, gpu-computing, machine-learning, large-language-models, developer-tools
- platform: cpp, cross-platform
- tags: cuda, kernels, tile-primitives, flash-attention, nvidia, blackwell, hopper, performance, dsl, gpu, linux

## Member repositories
- HazyResearch/ThunderKittens (main) score 70

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:13.269345+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:31:36.058468+00:00, confidence not recorded.
  - readme: https://github.com/HazyResearch/ThunderKittens (fetched 2026-08-28T04:08:13.269345+00:00, sha aa092135d61f)
- Data as of 2026-08-30T08:39:29.467469+00:00.