Remote computing for FMB

More details on these steps can be found in the CHESS user guide: https://www.chess.cornell.edu/user-guide

Getting started: checklist

1. Do you have a CLASSE ID?

This is the same ID you would use to access BeamPASS user portal.

• If you have a Cornell affiliation, your CLASSE ID is probably the same as your NetID

If you will be a CHESS user and do not have a CLASSE ID or if you forgot CLASSE ID password, visit BeamPASS and follow the appropriate link.

2. Is your CLASSE account active?

If you don't remember the last time you logged in to your CLASSE account: activate your account or reset your password.

3. Are you enrolled in CLASSE Duo?

Most CHESS computing resources require CLASSE Duo two-factor authentication. Please follow the instructions here: https://wiki.classe.cornell.edu/Computing/CLASSEDuo

4. Have you installed and tested your NoMachine connection?

If you haven't connected to CHESS resources via NoMachine before, please install the client and test your connection by following the instructions here: https://wiki.classe.cornell.edu/CHESS/RemoteUserGuide

Jupyter Hub

We recommend running many of our custom scripts using JupyterHub.

Setup part 1: On NoMachine

Steps to add FMB specific environments to your JupyterHub kernel list ("Adding new environments" on the CLASSE page):

1. Log in to lnx201 on NoMachine
2. Open a new terminal window
   • Make sure you're in your own user directory: cd /nfs/user/yourCLASSEID

3. Activate the SAXS/WAXS workflow environment:

   • source /nfs/chess/sw/miniforge3_chap/bin/activate; conda activate saxswaxsworkflow

4. Run:

python -m ipykernel install --user --name=saxswaxsworkflow --display-name “SAXSWAXSWorkflow”

Optional- these environments are not current as of 2024-1:

1. Activate the InstantPlot environment
   • Either run the shortcut activate_instantplot, OR run:
   • source /nfs/chess/opt/miniconda3_msnc/bin/activate; conda activate saxswaxs-viewer
2. Run:

   python -m ipykernel install --user --name=saxswaxs-viewer --display-name “InstantPlot”

3. Deactivate the environment: conda deactivate

Setup part 2: In your browser

Documentation from CLASSE IT on setting up and connecting to Jupyter Hub: https://wiki.classe.cornell.edu/Computing/JupyterHub

When you connect to JupyterHub, generally you'll want to connect to the CLASSE Compute Farm server, using the option with the 32 CPUs if you plan to do batch processing. There is also and FMB-specific queue which will become available to FMB users.

-- LouisaSmieska - 26 June 2023