and have expanded their x86 Ecosystem Advisory Group collaboration to streamline AI development on x86 CPUs using . Following a recommendation from the advisory group, the companies have established oneDNN as the default open-source library.
As part of this initiative, AMD is integrating its ZenDNN optimizations into the oneDNN library to enhance performance. AMD has also joined the UXL Foundation as a Steering Member, further supporting the framework intended to enable PyTorch applications.
This strategic alignment simplifies the software stack for developers by providing a unified acceleration path across x86 hardware. By standardizing on oneDNN for PyTorch, both chipmakers are ensuring more efficient AI deployment across the x86 ecosystem.