Updated 2022-12-05

Run Jupyter Notebooks Interactively

Overview

  • The Jupyter Notebook is an open-source web application that allows you to create and share documents that contain live code, equations, visualizations and narrative text.

  • Uses include

    • Data cleaning and transformation
    • Numerical simulation
    • Statistical modeling
    • Data visualization
    • Machine learning
    • Much more!

  • You can run Jupyter on PACE clusters via Open OnDemand, entirely within a browser window.

Run Jupyter Interactively

Set up Interactive (Open OnDemand) Session

  • Open the Open OnDemand Guide and select the version of Open OnDemand you want to use. Note: You must be connected to the Georgia Tech VPN to access Open OnDemand
  • Open the link and navigate to the top of the page and select Slurm Interactive Apps > Jupyter
  • Enter your username and choose whether you want to use inferno (paid) or embers
  • Adjust the rest of the settings to your needs and then click Launch
  • For a more in-depth guide on how to run Jupyter Interactively open the [Jupyter Guide] (https://docs.pace.gatech.edu/ood/guide/#jupyter-notebook)

Using Jupyter with Your Own Conda Environment

You can build your own conda environments and use them in Jupyter. If you install the ipykernel package in a conda env you have created, it will appear as a choice in the list of Jupyter kernels.

Complete these steps on the command line before launching Jupyter to set up the environment:

  • Complete the One-Time Setup for Anaconda on PACE
  • Load Anaconda: module load anaconda3
  • Create a conda environment with the name of your choice (e.g., "myenv"): conda create --name myenv
  • Activate the environment: conda activate myenv
  • Install the ipykernel package to support Jupyter: conda install ipykernel
  • Install any other packages you would like to use in your environment

After you have set up your environment, you can use it any time you launch a Jupyter notebook. Be sure to create the environment and install the ipykernel before submitting the Jupyter job. This setup needs to be completed only once per environment.

Inside Jupyter, you will now see a choice of kernel named Python [conda env:.conda-myenv]. Select it to run the notebook's code in your environment.