intel/intel-npu-acceleration-library

DirectML vs This Package

DavidMartinezGonzalez opened this issue · 0 comments

Hi,

I'd like to ask about differences between using this package and DirectML (https://github.com/microsoft/DirectML/tree/master).

Both seems to support NPU on Windows (https://devblogs.microsoft.com/directx/introducing-neural-processor-unit-npu-support-in-directml-developer-preview/) . So, my questions, applicable at least for Windows:

  • are they compatible? How?
  • if not, why should we use one or the other? Are there situations where one is better than the other?

Thanks!