项目核验档案

tile-ai/tilelang

tilelang 是一个用 Python 语法编写 GPU/CPU/NPU 高性能算子（如 GEMM、FlashAttention）的 DSL，构建在 Apache TVM 编译器基础设施之上，支持 CUDA、HIP、Metal 和昇腾后端。

项目概览

它针对算子开发中生产力与底层优化难以兼得的痛点，提供 Pythonic DSL 加 TVM 编译栈的方案，让开发者以 Python 库形式（pip 安装即用）编写可编译到多种后端的算子。

通过 PyPI 单条命令 pip install tilelang 安装，用 python -c "import tilelang; print(tilelang.__version__)" 一行验证，随后可运行 Quick Start 的 GEMM+ReLU 示例。

GitHub 项目简介: Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels
README: Tile Language (**tile-lang**) is a concise domain-specific language designed to streamline the development of high-performance GPU/CPU/NPU kernels (e.g., GEMM, Dequant GEMM, Flash…
README: The default `auto` target detects CUDA, HIP, Metal, and Ascend devices; select an explicit target when compiling for another backend or architecture.
README: The following example defines, compiles, runs, and verifies an FP16 GEMM kernel with FP32 accumulation and a fused ReLU epilogue. It uses PyTorch CUDA tensors
README: **2025-02-12 — [TileLang v0.1.0](https://github.com/tile-ai/tilelang/releases/tag/v0.1.0):** published the first v0.1 public release.

来源更新时间：2026-10-06T18:47:07Z