Install with Containers

If you need to run simulations on a computing cluster, or the built-in installer doesn't work for you, pre-built Docker and Apptainer images with all dependencies are available on the GitHub Container Registry.

If you're on a personal machine, go with Docker; if you're on a cluster, Apptainer is usually the better fit (and is often already installed).

Available Images

Tag	Description
:latest	CPU-only image
:latest-gpu	GPU-enabled image
:VERSION	Specific release

Using Docker

If you need help getting started with Docker, see the Docker documentation.

Here is an outline of the steps to set things up:

1. Pull the image

   sh docker pull ghcr.io/harvard-neutrino/prometheus:latest

2. Run the container and launch a shell

   sh docker run --rm -it ghcr.io/harvard-neutrino/prometheus:latest

3. Run an example

   sh docker run --rm -v "$PWD/output:/output" ghcr.io/harvard-neutrino/prometheus:latest python /opt/prometheus/examples/01_basic_water.py

   The -v "$PWD/output:/output" part of the command enables volume mounting to access output files on your local machine.

For more container running options, refer to the Docker containers documentation.

GPU Support

For GPU-accelerated simulations, you will need the NVIDIA Container Toolkit. Once installed, use the GPU image:

docker run --rm -it --gpus all ghcr.io/harvard-neutrino/prometheus:latest-gpu

If this fails, check that your NVIDIA drivers and container toolkit are correctly installed.

Using Apptainer / Singularity

Apptainer (formerly Singularity) is particularly useful for running simulations on computing clusters, where it is often pre-installed. If you need help getting started, see the Apptainer documentation.

1. Load the Apptainer module (if on a cluster)

   sh module load apptainer

2. Pull the container image

   sh apptainer pull docker://ghcr.io/harvard-neutrino/prometheus:latest

3. Run an example

   sh apptainer exec prometheus_latest.sif python /opt/prometheus/examples/01_basic_water.py