Jupyter Notebooks on Supercomputers

A guide to using Jupyter Notebooks on various HPC systems including Perlmutter (NERSC) and Polaris (ALCF).

Author

Affiliation

Sam Foreman

ALCF

NERSC Instructions

1. Project Setup:

   • Login to Perlmutter:

     ssh <your_nersc_username>@perlmutter.nersc.gov
     mkdir -p /global/cfs/cdirs/m4388/$USER/
     cd /global/cfs/cdirs/m4388/$USER

2. Using Jupyter on Perlmutter @ NERSC

   • jupyter.nersc.gov

ALCF Instructions

1. Logging In: https://jupyter.alcf.anl.gov/

   Select “Login Polaris” and use your ALCF credentials and Multi-Factor Authentication.

   Note📝 Note

   Some of the images below show “ThetaGPU” being used, however, similar instructions apply to “Polaris” or “Sophia”.

   

   Login GIF

2. Server settings & start up

   

   server options

   You’ll want to set your server options to the following:

   • job profile: ThetaGPU Compute Node
   • Sophia Queue: by-gpu
   • Project List: ALCFAITP
   • Runtime: can be 5 to 60 minutes

   When you click “start” you are submitting a job to the batch queue and waiting for the job to begin.

3. Server shutdown & logout

   ALCF’s guiding policy is:

   “We ask that all users follow good etiquette and be excellent to one another.”

   If you simply close your browser window, or logout without shutting down the jupyter server, your job will continue to occupy the worker node. Be considerate and shutdown your job when you finish.

   

   Shutdown GIF