search

AndreWeiner / of_pytorch_docker

Dockerfile to create an image with OpenFOAM-plus and PyTorch support
MIT License
37 stars 14 forks source link

adding LIGGGHTS(R)-PUBLIC to Singularity container #2

Closed SomePersonSomeWhereInTheWorld closed 2 years ago

SomePersonSomeWhereInTheWorld commented 2 years ago

This is a feature request to add LIGGGHTS(R)-PUBLIC to the Singularity container. I'm happy to troubleshoot any steps.

AndreWeiner commented 2 years ago

Dear @RobbieTheK, thanks for opening this issue. I don't see how this feature could be of general interest to OpenFOAM+ML users. Could you elaborate? If you want to use the DEM package in conjunction with OpenFOAM, feel free to extend the Dockerfile provided in this repository. If you don't need the ML part, start with the official container. Best, Andre

SomePersonSomeWhereInTheWorld commented 2 years ago

Dear @RobbieTheK, thanks for opening this issue. I don't see how this feature could be of general interest to OpenFOAM+ML users. Could you elaborate? If you want to use the DEM package in conjunction with OpenFOAM, feel free to extend the Dockerfile provided in this repository. If you don't need the ML part, start with the official container.

We have some data scientists that use them together and being on RHEL makes it difficult to get all of these working natively. We also don't use Docker for security reasons, so I thought it'd be a nice feature to have a Singularity container with the DEM package included.

AndreWeiner commented 2 years ago

I see. Nonetheless, I would start with an OpenFOAM-only container since the container provided here is meant to be used for CFD + machine learning applications (PyTorch). Singularity supports converting Docker into Singularity images (you don't even need Docker; Singularity allows building images from a Docker registry). Best, Andre