These instructions are intended to help you set up a Julia kernel to be used with Jupyter notebooks on the cluster.
The installation of Julia packages could take significant time and therefore we recommend that the Julia packages setup is done on a compute node via an interactive session.
Interactive session on the FAS cluster are initiated with the salloc
command as illustrated below:
[user@holylogin01 ~]$ salloc -pty -p test --mem=4G -t 120
salloc: Pending job allocation 31172193
salloc: job 31172193 queued and waiting for resources
salloc: job 31172193 has been allocated resources
salloc: Granted job allocation 31172193
salloc: Waiting for resource configuration
salloc: Nodes holy7c26601 are ready for job
[user@holy7c26601 ~]$
After your interactive session has started, you need to load the required software modules - a Julia module and a Python module. Please refer to the RC user portal for searching software modules.
[user@holy7c26601 ~]$ export PATH=$PATH:/n/holylabs/LABS/jharvard_lab/Users/jharvard/software/julia-1.9.3/bin
[user@holy7c26601 ~]$ module load python/3.10.12-fasrc01
The next step is to start Julia and install the IJulia package, which binds the Julia kernel with Jupyter.
julia> using Pkg
julia> Pkg.add("IJulia")
julia> Pkg.build("IJulia")
To learn how to schedule a Jupyter notebook or Jupyter Lab session via our interactive computing portal (VDI) follow these instructions.
From the the Interactive Apps
dropdown menu in the VDI portal select the Jupyter notebook / Jupyterlab
app. Choose the parameters of your Jupyter job and launch the interactive session. Once the Jupyterlab interface opens, the available kernels will be displayed.
NOTE: The available Notebook kernels may differ in your environment depending on the actual
conda
environments and Julia versions installed in your user space. When you select the desired Julia kernel, the Julia notebook will open in a new tab in your browser.