Miniconda3 is not officially supported on this cluster.
However, there has been significant demand for its use and so we will show how to set up and work with Miniconda3, including creating and configuring Conda environments, and installing necessary packages.
By following this guide, you can take advantage of Conda’s powerful environment management features while working within the cluster environment.
First, start an interactive session with enough memory (e.g., 20GB) to work with Conda like this: srun --mem 20G --pty bash
To access Miniconda3, load the appropriate Spack module: module load spack miniconda3
Once Miniconda is loaded, set up the Conda environment configuration to customize the location where Conda stores packages and environments:
source /usr/local/spack/opt/spack/linux-debian12-x86_64/gcc-12.2.0/miniconda3-24.3.0-iqeknetqo7ngpr57d6gmu3dg4rzlcgk6/etc/profile.d/conda.sh
Then, configure Conda to use specific directories for storing packages and environments. Replace PILOGIN and LOGIN with your actual login details:
conda config --prepend pkgs_dirs /sci/labs/<PILOGIN>/<LOGIN>/.conda/pkgsconda config --prepend envs_dirs /sci/labs/<PILOGIN>/<LOGIN>/.conda/envs
To create a new Conda environment, use the following command: conda create -n <environment name>
To activate the newly created environment: conda activate <environment name>
Once inside your environment, add important Conda channels for your needs. The following channels are commonly used:
conda config --env --add channels defaultsconda config --env --add channels biocondaconda config --env --add channels conda-forge
Once your environment is activated and configured, you can install packages using the conda install command. Replace <package-name> with the desired package: conda install <package-name>
For example:
conda install numpyconda install pandas
List installed environments: conda env list
Deactivate the current environment: conda deactivate
Remove an environment: conda remove --name env --all