ljcc0930 / taichi

Productive & portable programming language for high-performance, sparse & differentiable computing on CPUs & GPUs

Tutorial | Examples | Forum

Documentation | 简体中文文档 | Contributor Guidelines

Overview

Taichi (太极) is a parallel programming language for high-performance numerical computations. It is embedded in Python, and its just-in-time compiler offloads compute-intensive tasks to multi-core CPUs and massively parallel GPUs.

Advanced features of Taichi include spatially sparse computing, differentiable programming [examples], and quantized computation.

Please check out our SIGGRAPH 2020 course on Taichi basics: YouTube, Bilibili, slides (pdf).

中文视频教程: [哔哩哔哩], [幻灯片]

Examples (More...)

Installation

python3 -m pip install taichi

Supported OS: Windows, Linux, Mac OS X; Python: 3.6-3.9 (64-bit only); Backends: x64 CPUs, CUDA, Apple Metal, OpenGL Compute Shaders.

Please build from source for other configurations (e.g., your CPU is ARM, or you want to try out our experimental C backend).

Note:

• The PyPI package supports x64 CPU, CUDA 10/11, Metal, and OpenGL Compute Shader backends.
• On Ubuntu 19.04+, please sudo apt install libtinfo5.
• On Windows, please install Microsoft Visual C++ Redistributable if you haven't.
• [All releases]

	Linux (CUDA)	OS X (10.14+)	Windows	Documentation
Build
PyPI

Developer Installation

Please follow this doc to learn how to build Taichi from source. Note that Taichi requires LLVM-10.0.0, and it is recommneded to use our prebuilt LLVM libraries for each platform.

Contributors

Note: contributor avatars above are randomly shuffled.

---

We welcome feedback and comments. If you would like to contribute to Taichi, please check out our Contributor Guidelines.

If you use Taichi in your research, please cite related papers:

• (SIGGRAPH Asia 2019) Taichi: High-Performance Computation on Sparse Data Structures [Video] [BibTex] [Code]
• (ICLR 2020) DiffTaichi: Differentiable Programming for Physical Simulation [Video] [BibTex] [Code]
• (SIGGRAPH 2021) QuanTaichi: A Compiler for Quantized Simulations [Video] [BibTex] [Code]

Links

• TaichiCon: Taichi developer conferences.
• GAMES 201 Lectures: (Chinese) A hands-on course on building advanced physics engines, based on Taichi.
• TaichiZoo: Running Taichi code in your browser ¹.
• 加入太极图形.
• 太极图形课.
• More...

Security

Please disclose security issues responsibly to [e-mail].

---

1. TaichiZoo is still in its Beta version. If you've encountered any issue, please do not hesitate to file a bug.