started · updated
Intel and AMD to unify x86 AI software via oneDNN library
Intel and AMD have announced a collaboration to consolidate their separate x86 CPU optimization paths for the PyTorch AI framework into a single open-source library, oneDNN. This move aims to reduce the development burden by providing a unified software path for AI workloads, regardless of whether the underlying hardware is Intel or AMD silicon.
Under the agreement, recommended by the x86 Ecosystem Advisory Group (EAG), AMD will integrate its proprietary ZenDNN optimizations into oneDNN. Additionally, AMD will join the UXL Foundation as a steering member to participate in the governance and future direction of the library.
The partnership focuses on simplifying the deployment of AI applications, particularly for AI inference, edge computing, and latency-sensitive tasks. The companies also intend to collaborate through the EAG on next-generation x86 instruction set extensions, such as AI Compute Extensions (ACE), to ensure these capabilities reach developers through standardized oneDNN updates.
Entities
AMD · Intel · PyTorch · UXL Foundation · oneDNN
Claims
What the coverage asserts, and how many sources carry each claim.
- [● 3 SOURCES] AMD will integrate its ZenDNN optimizations into oneDNN. www.tokenpost.kr · pcekspert.com · cryptobriefing.com
- [● 3 SOURCES] AMD is joining the UXL Foundation as a steering member. www.tokenpost.kr · pcekspert.com · cryptobriefing.com
- [● 3 SOURCES] Intel and AMD will consolidate their x86 CPU optimization paths for PyTorch into the oneDNN library. www.tokenpost.kr · pcekspert.com · cryptobriefing.com
- [● 3 SOURCES] The decision was made following recommendations from the x86 Ecosystem Advisory Group (EAG). www.tokenpost.kr · pcekspert.com · cryptobriefing.com
- [● 2 SOURCES] The companies are collaborating on AI Compute Extensions (ACE) through the EAG. www.tokenpost.kr · cryptobriefing.com