CCR/Ghub Jupyter Notebook Demo

This Jupyter Notebook provides a demonstration and template scripts for executing up to nineteen jobs in parallel on the University at Buffalo (UB)'s Center For Computational Research (CCR)'s high performance compute cluster, UB-HPC.

On CCR, sbatch and Slurm commands are used to execute the parallel jobs.

On Ghub, the Pegasus Workflow Management System (WMS) is used to execute the parallel jobs. Launch the Ghub Pegasus WMS Workflows Tutorial and Templates tool, GitHub Repository, Ghub_Pegasus_WMS_Tutorial_and_Templates, to learn more about running Pegasus WMS workflows on Ghub.

How to use this notebook:

To clone this GitHub repository, in a terminal window enter:

git clone https://github.com/GhubGateway/CCR_Ghub_Demo

CCR

• Login to CCR OnDemand.
• Launch a Quick Launch General-Compute Desktop Interactive App.
• Open a terminal window (Click the >_ button).
• In the terminal window, clone this GitHub repository.
• In the terminal window, change directory (cd) to the CCR_Ghub_Demo directory and install the ccrghubdemo_kernel Jupyter kernel:
  • Enter: source install_ccrghubdemo.2023.01_kernel.sh
• Launch the Quick Launch JupyterLab/Notebook Interactive App, check the JupyterLab check box and Connect to Jupyter.
• Open the ccrghubdemo.ipynb notebook.
• Select Kernel / Change Kernel and select the ccrghubdemo_kernel Kernel, the Python 3 (ipykernel) Kernel is set by default.
• Select Kernel, Restart Kernel and Run All Cells.

Ghub

• Login to Ghub.
• Launch the Workspace 10 Tool.
• In the terminal window, clone this GitHub repository.
• Launch the Jupyter Lab (202210) Tool.
• Open the ccrghubdemo.ipynb notebook.
• The Python 3 (ipykernel) Kernel is set by defaiult.
• Select Kernel, Restart & Run All.