These docs are under construction.

Interactive Computing

The Sol cluster is designed for asynchronous batch calculation, however we have reserved some resources that make interactive computing easier. These are intended for prototyping new codes and developing new workflows.

Jupyter

If you visit our web portal at hpcportal.cc.lehigh.edu (you must be on the Lehigh network or the VPN), you can connect to a Jupyter Notebook server, listed under the interactive apps in this instanced of Open OnDemand.

This environment provides a default Anaconda environment which you can also access on the cluster with these commands:

module load miniconda3
conda activate cenv-2025.01

We have included all of the basics, including NumPy compiled with MKL, and a number of other popular data science tools. Users can install their own environments and use them in Jupyter by building a custom kernel.

Custom kernels in Jupyter

# create a python virtual environment or conda environment
python -m venv ./venv
# access the environment
source ./venv/bin/activate
# install the kernel software with pip
pip install ipykernel
# register your kernel so Jupyter can find it
python -m ipykernel install --user --name project-alpha --display-name "Project Alpha, 2025-Feb"

The installation command above sets two names. The --name flag is used in the path to the settings file. In this example, we can find this file at:

~/.local/share/jupyter/kernels/cenv-alpha-2025.02/kernel.json

This file encodes the --display-name that we set above, and this name is visible inside the Jupyter Notebook. To use this kernel, and any associated software that you’ve already installed, you can change the kernel using the menu system in the upper right.