THE TRITON LANGUAGE | PHILIPPE TILLET

Triton is a new language that 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. Since its inception as a graduate research project, Triton has helped provide fast and maintainable kernels across several organizations (e.g., OpenAI, Facebook, Microsoft). In this talk, Phil Tillet (Technical Staff, OpenAI) will share the history of this project, how it works, its performance benefits and its integration into the PyTorch ecosystem. No CUDA experience needed.

[TheAIEpiphany]

Closely following how Triton evolves with PyTorch 2.0 - amazing work Phil!

[DataPastor]

Amazing work! Now I only have to figure it out, what to use it in a “normal” data science job (i.e. not deep learning research) for.

[TheCrmagic]

Thank you for your work. Are there more video tutorials available for triton?

[amortalbeing]

Thanks, but I'd like a few more examples from the ground to the final usable form.
That would greatly help people like me.
Great job by the way .

[SurajKumar-seoc]

hay this is nice and interesting. and I want to learn. how should I start. I have install triton and LLVM form code and I face some problem. Thank You.