Triton has made it possible for OpenAI researchers with no GPU experience to write screaming-fast GPU code. Makes it not-so-scary to try out ideas that are outside the bounds of what PyTorch provides natively:
OpenAI
 @OpenAI
28 Jul 2021
We’re releasing Triton 1.0, an open-source Python-like programming language for writing efficient GPU code. OpenAI researchers with no GPU programming experience have used Triton to produce kernels that are 2x faster than their PyTorch equivalents. openai.com/blog/triton/

Jul 28, 2021 · 4:28 PM UTC

3
46
1
339
Replying to @gdb
GJ!
Replying to @gdb
Happily awaiting the next tweet in 5 years saying OpenAI developed thought-to-text. Whatever neurons fired off in your brain as you thought about the problem space serve as the architecture network for the AI implementation
1
Replying to @gdb
I kinda prefer Triton X1.00
3