Skip to main content

Singularity Containers

Singularity and Apptainer bring operating system-level virtualization to scientific and high-performance computing, allowing you to package complete software environments --- including operating systems, software applications, libraries, and data --- in a simple, portable, and reproducible way, which can then be run almost anywhere.

If you have a container that you would like use to run your Jupyter notebook session, you can use the --sif option in your launch command followed by either:

  • a URI to where the container is stored in a registry (e.g., Docker Hub), or
  • a path to where the container image is stored on a local filesystem.

In general, we recommend the use of containers served from common registries, which will be cached to your local filesystem for reuse by default.

For example, let's say you need an R environment for your Jupyter notebook session. Try the latest r-notebook container from the Jupyter Docker Stacks project on Quay.io.

galyleo launch --account abc123 --partition debug --cpus 4 --memory 8 --time-limit 00:30:00 --sif docker://quay.io/jupyter/r-notebook:latest --quiet

Or, if you want to work on your latest AI project, then go ahead and launch a GPU-accelerated PyTorch container from the NGC Catalog on an NVIDIA V100 GPU available in Expanse's gpu-debug partition.

galyleo launch --account abc123 --partition gpu-debug --cpus 10 --memory 92 --gpus 1 --time-limit 00:30:00 --sif docker://nvcr.io/nvidia/pytorch:24.12-py3 --bind /expanse,/scratch --nv --quiet

Here, the user-defined --bind mount option enables access to the /expanse filesystems (e.g., /expanse/lustre) and the node-local NVMe /scratch disk available on each compute node from within the container. By default, only your $HOME directory is accessible from within the container.

The --nv flag enables NVIDIA GPU support.