# Intel and AMD collaborate on oneDNN optimizations for PyTorch

Digest AI · Hardware & Compute · published 2026-10-20T20:02:30Z

Canonical: https://digestai.news/story/intel-and-amd-collaborate-on-onednn-optimizations-for-pytorch

## Summary

Intel and AMD announced a joint effort to improve the oneDNN neural‑network library for use with PyTorch. The two rivals will combine their x86 CPU optimization work, with AMD’s ZenDNN enhancements being merged into the upstream oneDNN code base. The collaboration follows earlier joint activity on the x86 Ecosystem Advisory Group and aims to make oneDNN the default open‑source library for accelerating PyTorch on x86 processors.

In addition to the software work, AMD said it will join the UXL Foundation as a steering member. The foundation, created in 2023, brings together former oneAPI projects under a consortium that includes Intel, Samsung, Qualcomm, Arm and others. AMD’s participation signals a broader industry push to standardise CPU‑based AI acceleration across competing silicon vendors.

## Key points

- Intel and AMD announced they will jointly develop oneDNN optimizations for PyTorch.
- AMD’s ZenDNN improvements will be contributed to the upstream oneDNN library.
- AMD joins the UXL Foundation as a steering member to help guide oneDNN development.

## Why it matters

Uniting rival CPU vendors on a common AI acceleration library could boost PyTorch performance on x86 CPUs and simplify deployment for developers.

## Sources

1. [Intel & AMD now collaborating on software optimizations with PyTorch/oneDNN](https://phoronix.com/news/Intel-AMD-oneDNN-Optimize) (phoronix.com, 2026-10-20)

## Cite

Digest AI, "Intel and AMD collaborate on oneDNN optimizations for PyTorch", 20 October 2026, https://digestai.news/story/intel-and-amd-collaborate-on-onednn-optimizations-for-pytorch

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/intel-and-amd-collaborate-on-onednn-optimizations-for-pytorch.json