JupyterHub is a web-based platform that provides faculty and students with secure, personal workspaces on our HPC cluster. It acts as a management layer, giving each user their own instance of JupyterLab for coding, data analysis, and visualization. This centralized, multi-user environment simplifies access to powerful computing tools for research and coursework.
Why use JupyterLab?
JupyterLab offers a more visual, interactive experience compared to the command line, making it easier for users to write code, analyze data, and visualize results in one place.
Are JupyterLab and Jupyter Notebook the same thing?
No. JupyterLab is a more advanced, flexible interface that builds on Jupyter Notebook by allowing multiple tabs, drag-and-drop functionality, and the ability to view and edit multiple files side-by-side. Essentially, JupyterLab offers a richer, more customizable workspace, while Jupyter Notebook is simpler, more focused environment.
Accessing JupyterLab
- Open a web browser and go to: https://n1.hpc.at.sfsu.edu.
-
Sign in with your SF State login (if off campus, you will need to use VPN):
-
Once signed in, the JupyterLab interface will display as follows:
Submitting a job from JupyterLab to cluster
- In JupyterLab, select Terminal.
- Create a script (e.g. vi nvidia-sml.test.sh):
- Once the script is saved, it will appear on the list of files in the Home directory (pane on left side of the page).
- Navigate to Launcher tab and select Slurm Queue Manager. Through this interface you will be able to view any jobs currently running (Slurm Queue), submit jobs, and view any notifications (e.g. job submitted).
- In the Submit Jobs tab, under Choose a slurm script from your current directory, select your script, then select Submit Job:
- Navigate to the Job Notifications tab to confirm that the job was submitted.
-
The output file will appear in the list of files in your home directory (left side panel) when the job is finished:
Shutting Down Your Server and Logging Out in Jupyter Lab
When you are ready to stop working on JupyterLab, we kindly ask that you shut down your server and log out. In some cases, you may need to leave your server running. For instance, if you are performing a very time intensive analysis that may take hours to complete, it is appropriate to keep your server running. Unless you are planning to run your server for a long period of time, we ask that you shut it down before you leave.
If you are running time intensive analysis that may take hours:
If your job is going to take hours to complete so you will need to come back to it later. Simply log out of JupyterLab. Please note that logging out will not stop the job that you submitted to run. Below are instructions on how to logout:
-
In the top right corner, select File, and scroll down and select Logout:
If you are running short runtime jobs and are done using the JupyterHub:
When you are done running your analysis and are ready to stop working on JupyterLab, we kindly ask that you shut down your server and log out. Below are instructions on how to do so:
- In the top right corner, select File, and scroll down and select Hub Control Panel:
- Select the Stop My Server button:
- Then select the Logout button on the left-hand corner. You will be logged out of JupyterLab.
Comments
0 comments
Article is closed for comments.