IT Home October 21 news: AMD and Intel yesterday (October 20) jointly announced a collaboration on x86 CPU optimization for PyTorch:The two parties will integrate their respective optimization work into the open-source deep learning library oneDNN, providing developers with a unified software path across both companies' CPUs.
IT Home note: PyTorch is an open-source Python deep learning framework primarily used for building, training, and deploying neural networks. It is known for its flexibility, ease of debugging, and Python-like programming style, and is widely used in academic research and commercial AI applications.
According to the announcement, the two parties plan to consolidate their respective processor optimization work into the open-source library oneDNN, making it the default open-source library for x86 acceleration within PyTorch. AMD will integrate its ZenDNN optimizations into oneDNN and join the Unified Acceleration Foundation (UXL Foundation) as a steering committee member.
The rare cooperation between the two parties was facilitated by the x86 Ecosystem Advisory Group, "turning competition into collaboration" in the CPU software ecosystem. If it goes smoothly, developers, cloud service providers, and enterprise users may see a more consistent PyTorch performance experience on x86 servers, workstations, and AI PCs.
Beyond oneDNN, AMD and Intel are also advancing "AI Compute Extensions" (ACE). This is a next-generation instruction set architecture for x86 focused on accelerating matrix multiplication (a core operation in AI model training and inference), and the related capabilities are planned to be integrated into PyTorch via oneDNN updates in the future.
For developers, PyTorch workloads that previously might have needed to be adapted and tuned separately for AMD or Intel CPUs are expected in the future to gain optimized support through a single software path. The two companies say this will reduce fragmentation in the x86 AI ecosystem, cut duplicated development and maintenance costs, and make building, deploying, and migrating AI applications across platforms smoother.