Introducing Triton: Open-source GPU programming for neural networks
OpenAI Blog
Triton, an open-source Python-like programming language, has been released to allow researchers without CUDA expertise to write efficient GPU code. The language produces performance on par with expert-written code in most cases. This removes a barrier to GPU programming by eliminating the need to learn low-level CUDA before writing neural network computations.
Why it matters
We’re releasing Triton 1.0, an open-source Python-like programming language which enables researchers with no CUDA experience to write highly efficient GPU code—most of the time on par with what an expert would be able to produce.
