Overview
Jupyter Notebooks provide an interactive environment for writing, running, and visualizing Python code in your browser — launched through Open OnDemand, no SSH required.
Getting started
- Open the OnDemand portal and authenticate with NetID + DUO
- Launch via the Jupyter Notebook icon or Interactive Apps → Jupyter Notebook
- Configure the session (see below) and click Launch
- When status reaches Running, click Connect to Jupyter
Session configuration
| Setting | Guidance |
|---|---|
| Queue | e.g. short-40core-shared for light work |
| Number of hours | Expected session duration |
| Number of cores | 1–4 typically sufficient |
| Memory | Use 'G' notation for gigabytes |
| Email notification | Optional alert when the session starts |
Tip: Request only the resources you need to help reduce queue wait times.
File access
Your home directory loads automatically. Reach other spaces with a symlink:
ln -s /gpfs/scratch/NetID /gpfs/home/NetID
Best practices
- Use conda or virtual environments for reproducibility
- Track resource consumption during extended computations
- Terminate inactive notebooks and delete finished sessions to conserve cluster resources
Screenshots
Applies to