AI Potluck
Infrastructure / Core ML frameworks & libraries

Triton

Triton (triton-lang)

Triton is a Python-embedded language and compiler for writing highly efficient custom GPU kernels (deep-learning primitives), aiming for higher productivity than CUDA with more flexibility than other DSLs. It compiles via MLIR and targets NVIDIA GPUs and AMD GPUs, with CPU support in progress. Originally created by Philippe Tillet at OpenAI, it is now developed under the triton-lang organization. It is widely used as the kernel backend for projects like PyTorch's inductor.

Verified live 2026-06-22 via primary sources. MIT license body confirmed (Tillet/OpenAI copyright); full source public.

Openness

5 high confidence
5.0
license
MIT(OSI)
source
public
core-gated
ungated

MIT license body confirmed (Tillet/OpenAI copyright); full source public.

Adoption

5 high confidence
5.0

PyPI last_month downloads ~64.7M.

Capability

5 high confidence
5.0

Foundational kernel-authoring layer (e.g. PyTorch inductor backend) across NVIDIA/AMD GPUs.

Unchanged since 2026-07-30 (last edited, not re-checked)