0

Microsoft documents that GPU acceleration can be enabled on frameworks built on top of DirectX here

NVidia's base Docker image for CUDA is Linux based here

Question: How to enable GPU acceleration on Windows Containers on Docker that aren't based on DirectX ? Say for CUDA or NVidia based systems?

NOTE: The goal is to run AI/Neural Network code on a Windows container in Docker

Improve this question
3
  • Docker supporting CUDA (on Windows) would be something that the dockers would have to provide. All the Microsoft documentation indicates is that you can GPU accelerate a docker contain with DX.
    – Ramhound
    Commented Dec 24, 2019 at 3:28
  • @Ramhound "the dockers", do you mean Docker employees or the Windows docker container. If you meant the latter, then how does one go about doing that, and making sure that the host system is passing along its GPU resources to the Windows container outside of DX to run Caffe, pytorch or Tensorflow?
    – yippee
    Commented Dec 26, 2019 at 19:21
  • Grammatical mistake. The comment was supposed to be “Docker developers”.
    – Ramhound
    Commented Dec 26, 2019 at 20:13

0

Reset to default

You must log in to answer this question.

Browse other questions tagged .