AMD and Intel team up to build one PyTorch path for CPU AI inference

Tom's Hardware reports AMD and Intel plan a single PyTorch code path for AI inference on CPUs, folding AMD's ZenDNN library into Intel's oneDNN. HotHardware says the goal is to make AI workloads easier to optimize across the companies' x86 processors. It reports that oneDNN 3.14 already includes initial AMD Zen optimizations, though they require explicit opt-in.