PyTorch-Triton 3.7 Adds Runtime Extensions
PyTorch · x · 2026-07-16
PyTorch-Triton 3.7 has been released, introducing the Triton Plugin Extensions system.
This mechanism allows for the dynamic loading of custom compiler passes, dialect ops, and DSL extensions at runtime, extending upstream Triton directly without the need for forking or recompiling.
Official technical deep-dive articles and getting started guides are also provided. This update is highly useful for teams needing compiler or operator-level customization on Triton.
More from Infra
- Strangeworks launches Aura to turn enterprise ops into production optimization systems — whurley · 2026-07-22
- Graph workload 854.graph500 enters SPEC CPU 2026 as a new CPU benchmark — Prof_DavidBader · 2026-07-22
- HilbertRaum open-sources a fully local AI chat and document analysis app for private use — Vladowski · 2026-07-22
- Hybrid and local inference are emerging as a response to AI energy and token costs — dmitry140 · 2026-07-22
- NVIDIA details Vera CPU with 2x performance claims and a 22,000-core rack — ryanshrout · 2026-07-22
- NVIDIA says Vera Rubin NVL72 delivers 10x more tokens per megawatt than Blackwell — nvidia · 2026-07-22