Triton is a programming language and compiler for writing highly efficient custom Deep-Learning primitives developed by OpenAI^[1].

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 DSLs.

Compatibility

Supported Platforms:

Linux

Supported Hardware:

NVIDIA GPUs (Compute Capability 7.0+)
Under development: AMD GPUs, CPUs

References

^ "OpenAI releases Triton, a programming language for AI workload optimization". VentureBeat. 2021-07-28. Retrieved 2021-09-19.

[1] "OpenAI releases Triton, a programming language for AI workload optimization". VentureBeat. 2021-07-28. Retrieved 2021-09-19.

[1]