4C is available pre-built in a docker image

[ppraegla]

With PR #96 we introduced a docker image with a pre-built 4C executable. Anyone can test and use 4C with just a few simple commands without configuring any environment and waiting for the build to complete. You can just pull the image docker pull ghcr.io/4c-multiphysics/4c:latest.

Usage

Run the docker image

docker run --interactive --tty ghcr.io/4c-multiphysics/4c:latest

and, e.g., run the minimal tests

cd /home/user/4C/build
ctest -L minimal

Run your input files

Run the docker image and mount a directory (e.g., the current directory ./)

docker run --interactive --tty --mount type=bind,source=./,target=/workspace ghcr.io/4c-multiphysics/4c:latest

Set the correct owner of the mount (Unfortunately, we need to do that because we are the user user inside the docker image and the mount point is owned by root)

sudo chown user:user -R /workspace/

Run 4C with an input file from the host system and the output is also written to the host system

mpirun -np 2 /home/user/4C/build/4C /workspace/particle_dem_1d_radius_from_input.dat /workspace/out

Visualize the results in Paraview on your PC.

Disclaimer This whole process is still a little awkward to use. But there will certainly be improvements in the future.

Additional information

At the moment, the docker image includes all dependencies required to run and test 4C, as well as the source. We will work on reducing the size of the docker image and only include the essentials to run 4C.

Share your feedback

Please give feedback on how well the docker image works and what can be improved for your workflow.

[gilrrei]

This works quite nicely! What about introducing an alias for 4C within the container?

[gilrrei]

Or alternatively using entry points in docker 🤔

[ppraegla]

Thanks for the suggestion. I will look into it at some point :)