Welcome to Triton Docs

Before you start cooking.

Make GPUs Go Brrrr! 🔥
Triton lets you write fast GPU kernels with Python. Say goodbye to complex CUDA and hello to simplicity without sacrificing performance.

Before We Start

  • To ensure a smooth learning experience, we highly recommend going through the kernels in increasing order of complexity.
  • Each kernel builds upon concepts introduced in the previous ones, and some explanations are provided only once to avoid redundancy.
  • By following this structured approach, you'll gain a deeper understanding of Triton and GPU programming step by step. 🚀

Why Triton?

  • Speed: Write GPU code that rivals CUDA.
  • Ease: Pythonic APIs make everything simpler.
  • Power: Ideal for deep learning, scientific computing, and more.

Install Triton

Get Triton up and running in seconds:

  • Install with pip:
    pip install triton
    

Ready to make your GPUs work for you? Let's get coding! 🚀