Python on the Cluster: Glossary

Yes, you can run python code on a cluster

Make sure the effort needed to run on a cluster isn’t too high

Python is loaded on the cluster through modules

Virtual environments store a version of python and some packages

Package requirements ensure consistency for versions

Submit jobs to run long-running or computationally intensive Python code

Create virtual environments and install packages from the Alliance wheelhouse

Working with local disk in a job can provide performance benefits

You can use the scheduler to ask for a GPU to run code on

Array jobs allow you to run several jobs with a single job script

Interactive jobs are a useful way to set up or solve issues with python code on a cluster

Most Alliance clusters have a JupyterHub