neuralmagic / deepsparse

Sparsity-aware deep learning inference runtime for CPUs
https://neuralmagic.com/deepsparse/
Other
2.98k stars 172 forks source link

Build instructions for other CPU architectures #1077

Closed mgiessing closed 1 year ago

mgiessing commented 1 year ago

Describe the bug I'm trying to install/build for currently unsupprted CPU architecture - do you have instructions on how to achieve that?

Environment Include all relevant environment information:

  1. OS: RHEL9.1
  2. Python version [e.g. 3.8]: 3.10
  3. DeepSparse version: 1.5.0

Additional context I'm trying to build for ppc64le arch which is CPU-only, but it has Matrix-math accelerator engines on the core which is aimed for ML & DL inference.

mgoin commented 1 year ago

Hi @mgiessing we're currently working on aarch64 support and are in alpha for that. We keep the core implementation of the engine closed-source so we have to support those new platforms ourselves. We aren't considering PowerPC at the moment since it isn't as widely used as ARM. Thanks for asking!