AMD and Intel unite PyTorch CPU optimization work in oneDNN
The collaboration gives developers a common software path for building and optimizing PyTorch applications across both companies’ x86 processors. AMD is incorporating its ZenDNN optimizations into oneDNN. According to Phoronix, ZenDNN has operated as an API-compatible fork of Intel’s library, tailored to AMD’s Zen processor designs.
- HPCwire describes oneDNN as an open-source deep learning library governed by the Unified Acceleration Foundation.
Covered by 3 publishers within 5 hours of the first report.
Reporting3
Phoronix Intel & AMD Now Collaborating On Software Optimizations With PyTorch/oneDNN · Michael Larabel
HPCwire AMD and Intel Simplify AI Development on x86 CPUs Using PyTorch · Andrew Jolly