# tile-ai/tilelang

Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels

Repository: https://github.com/tile-ai/tilelang
Canonical: https://ross.abutalabs.com/products/tilelang
Homepage: https://tilelang.com/
Language: Python
License: NOASSERTION
License Family: other
Last push: 2026-08-26T08:13:12+00:00

## Health v2 (maintenance only)
Score: 88/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 96, longevity 49
- inputs: {"age_days": 699, "days_push": 7, "days_rel": 30, "gap_med": 25, "n_releases_24m": 20}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 7288, forks 714 (observed 2026-08-28T04:09:58.042834+00:00)

## What it is
TileLang is a Pythonic domain-specific language and compiler (built on TVM) for writing high-performance GPU/CPU/accelerator kernels such as GEMM, FlashAttention, and linear attention. It lets developers express tiled compute kernels concisely while the compiler handles low-level optimizations across CUDA, Metal, and other backends.

## Use cases
- write high-performance CUDA kernels in Python
- implement FlashAttention or FlashMLA efficiently
- generate optimized GEMM and dequant GEMM kernels
- compile kernels for NVIDIA, AMD, and Apple Metal GPUs
- autotune tile sizes and layouts for peak performance
- prototype deep learning operators without writing raw CUDA

## When to choose
- you need near-SOTA kernel performance with Pythonic productivity
- you target multiple GPU backends (CUDA, Metal, ROCm) from one codebase
- you are building custom attention or matrix-multiply operators
- you want TVM-based compilation with autotuning

## When to avoid
- you only need off-the-shelf operators from PyTorch or cuBLAS
- you want a general-purpose programming language rather than a kernel DSL
- your workloads are CPU-only and simple
- you need long-term API stability - the project removes legacy APIs between releases

## Facets
- artifact type: library
- maturity: active
- function: compiler, programming-language, gpu-computing, machine-learning, deep-learning, benchmarking
- domain: compilers, programming-languages, gpu-computing, deep-learning, machine-learning, performance
- platform: python, windows, cpp
- tags: dsl, tvm, kernel-authoring, gemm, flashattention, cuda, metal, autotuning, code-generation, gpu, linux, macos

## Member repositories
- tile-ai/tilelang (main) score 88

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:58.042834+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 2026-08-29T17:38:52.014729+00:00, confidence not recorded.
  - readme: https://github.com/tile-ai/tilelang (fetched 2026-08-28T04:09:58.042834+00:00, sha df1a73c79a76)
  - homepage: https://tilelang.com/ (fetched 2026-08-29T08:34:18.840125+00:00, sha 6f019fc32dac)
  - site_page: https://tilelang.com/get_started/Installation.html (fetched 2026-08-29T08:34:18.843041+00:00, sha 33a54eeae427)
- Data as of 2026-08-30T08:39:29.467469+00:00.