Triton (programming language): Difference between revisions

Revision as of 04:01, 9 September 2022

Revision as of 13:40, 22 January 2023

Line 2:

{{multiple issues|{{single source|date=September 2021}}{{notability|date=September 2021}}}}

'''Triton''' is a programming language and compiler for writing highly-efficient, custom [[Primitives (computer graphics)|deep-Learninglearning primitives]] developed by [[OpenAI]].<ref>{{Cite web|date=2021-07-28|title=OpenAI releases Triton, a programming language for AI workload optimization|url=https://venturebeat.com/2021/07/28/openai-releases-triton-a-programming-language-for-ai-workload-optimization/|access-date=2021-09-19|website=VentureBeat|language=en-US}}</ref>

The aim of Triton is to provide an open-source environment to write fast code at higher productivity than [[CUDA]], but also with higher flexibility than other existing [[Digital subscriber line|DSLs]].{{cn|date=September 2021}}

Line 11:

===Supported hardware===

* [[NVIDIANvidia]] GPUs (Compute Capability 7.0+)

* Under development: AMD GPUs, CPUs