News › ai-infrastructure
By Zayden R., July 18, 2026
The UXL Foundation has released oneDNN 3.13, adding support for Intel's Nova Lake with AVX10.2 and further Intel optimizations. This update follows AMD's recent ZenDNN 6.0 rollout.
The UXL Foundation has just rolled out oneDNN 3.13, marking the neural network library's latest feature release. This update is particularly noteworthy for its future-proofing measures, as it brings support for Intel's upcoming Nova Lake architecture with AVX10.2. While oneDNN was originally developed by Intel as part of the oneAPI initiative, it now finds itself under the stewardship of the UXL Foundation. Despite this transition, the library remains heavily focused on Intel optimizations, making it an essential tool for those working with Intel hardware.
The addition of AVX10.2 support signals a clear commitment to staying ahead of the curve in terms of hardware compatibility. This is a solid improvement for developers who are preparing for the next generation of Intel processors. It's a move that aligns well with the broader industry trend of optimizing for specific hardware capabilities to maximize performance.
Technical details aside, the release demonstrates a continuous dedication to enhancing machine learning capabilities. Version 3.13 is not just about keeping pace with hardware; it also includes several other Intel optimizations aimed at improving overall efficiency. For engineers and developers, this means more robust support and better performance for neural network tasks.
This update follows closely on the heels of AMD's ZenDNN 6.0 release earlier this month. Both companies seem to be in a race to provide enhanced performance and support for their respective hardware lines, which can only be good news for the end-users.
In terms of practical impact, sysadmins and developers working with Intel hardware will find this update particularly beneficial. The new features and optimizations make oneDNN 3.13 a compelling choice for those looking to leverage Intel's latest technology in their AI and machine learning projects.