PyTorch 2.4 Launches to Speed Up AI Tasks with Initial Support for Intel GPUs
PyTorch 2.4 supports the Intel Data Centre GPU Max Series, which includes Intel GPUs and the SYCL software stack
To support streaming devices, this version generalizes the PyTorch device and runtime
The SYCL programming language is now utilized to implement popular Aten operators in the eager mode
OneMKL and oneDNN optimise the most performance-critical graphs and operators
PyTorch Profiler which is built on an integration between Kineto and oneMKL is being actively worked on in front of the forthcoming PyTorch 2.5 release
PyTorch 2.4's first support release includes the first Intel Data Centre GPU Max Series GPU for AI task acceleration
For Python programs, the extension can be loaded as a module, and for C++ projects, it can be linked as a library
It can be dynamically enabled in Python programs by importing intel_extension_for_pytorch