Specify Your Desired Version, Account, Hours, Cores, Partition
After choosing Jupyter Lab or Jupyter Notebook from the Interactive Apps menu you’ll need to specify your desired python distribution, SLURM account, hours, cores, partition (standard or largemem), and memory (2GB minimum). You’ll also need to specify the Anaconda Python module you want to use.
Upon selecting “Launch”, your job will be queued on one of your nodes and shown on the “My Interactive Sessions” screen. As soon as the job’s status is “Running”, you can click on “Connect to Jupyter”.
For instructions on using Jupyter Notebook, see the official documentation.
Choosing a Python Distribution
For optimal dependency resolution and environment management, mamba is faster and more reliable than Anaconda. mamba uses libmamba as its dependency solver, which is more efficient than the default solver for Anaconda. Several versions of both Anaconda and mamba are installed on ARC HPC's. Please refer to the software catalog for a complete list of available versions. The Anaconda installation in Great Lakes, Armis2, and Lighthouse are configured to allow the use of Anaconda's default repository, which has a limited free license. In contrast, the mamba installation uses conda-forge which is open-source, and free for all types of academic research.
Using a Conda Environment in Jupyter
Setup
# Load the conda/mamba module
ml mamba/py3.12
# Following 2 commands only required once to initialize mamba/conda
conda init
source ~/.bashrc
# Create the environment with desired packages and ipykernel
conda create -n {myenv} {packages} ipykernel
conda activate {myenv}
# Make the environment available to Juptyer
python -m ipykernel install --user --name {myenv} --display-name conda:{myenv}
Running the notebook
- https://greatlakes.arc-ts.umich.edu
- Select Interactive Apps -> Jupyter Notebook/Lab
- Choose the desired job configuration, and python distribution. Remember to choose the same distribution which was used to create the Kernel. In this example, that would be
mamba/py3.12 - Connect when available
- Choose the Kernel in the dropdown as named above
conda:myenv