AMD And Intel Unite To Supercharge AI Software Across All x86 CPUs
AMD and Intel are teaming up to simplify AI development on x86 CPUs, consolidating their respective optimizations for PyTorch under the open-source oneDNN library. The move aims to give developers a common software path across both companies' processors, eliminating some of the fragmentation involved in optimizing AI workloads for different CPU architectures.
As part of the agreement, AMD will integrate its existing ZenDNN optimizations into Intel's oneDNN, which becomes the default open-source library for x86 acceleration within PyTorch. This continues the companies' recent collaborations, where they most recently partnered to develop the AI Compute Extensions (ACE), a next-gen x86 ISA extension designed to accelerate matrix multiplication, with future integration planned through oneDNN updates.
The timing of this collaboration is hard to ignore. NVIDIA's RTX Spark laptops began arriving last week, bringing Blackwell graphics, a Grace CPU, up to 128GB of unified memory, and critically, NVIDIA's extensive AI software...
Copyright of this story solely belongs to hothardware.com. To see the full text click HERE
