This repository contains the Singularity recipe used to build containers for BRER simulations. A pre-built image is hosted on Sylabs Singularity library.
The main project for running these simulations is hosted at https://github.com/kassonlab/run_brer.
Pull directly from singularity library (recommended):
singularity pull --name singularity-brer.sif library://kassonlab/default/brer
or build it yourself:
sudo singularity build singularity-brer.sif deffile
where deffile is one of the recipe files in this repository.
The GROMACS build in this container is GPU-compatible (built with CUDA). In order to take advantage of this, use
the Singularity exec command with the --nv option:
singularity exec --nv singularity-brer.sif python3 my_run_script.py
--nv will bind the host nvidia drivers to the container, so be sure that your drivers are compatible with the CUDA version in the container (default CUDA 10.1).
An example run script is provided on the main project website)
Note: There is no Python 2 installed in the container (/usr/bin/python is actually Python3). Any Python scripts you write that you wish to run in the container
must be compatible with Python 3.X
Warning: Use singularity exec ..., not singularity run ....