Skip to content

Cluster setup

Santiago Castro Dau edited this page Feb 1, 2024 · 2 revisions

Your .bashrc

You can run the workflow either on a cluster setup or on you local machine. For local runs no further configuration is required except for running the snakemake commands with the --cores N flag, where N is the number of cores you want to allocate.

If you want to run the workflow on a cluster you need to configure the submission of the cluster runs in a profile. An example profile is provided in $HOME/graph-concept-learner/workflows/profile. Adding the following to your $HOME/.bashrc will enable cluster execution:

export SNAKEMAKE_PROFILE=$HOME/graph-concept-learner/workflows/profile

Adjust the main configuration file depending on the batch system

The workflow supports SLURM and LSF batch systems. Users must specify in $HOME/graph-concept-learner/workflows/profile/config.yaml which one will be used.

For LSF in $HOME/graph-concept-learner/workflows/profile/config.yaml:

# Specify custom job submission wrapper
cluster: "../scripts/LSF_cluster_job.py"

For SLURM:

# Specify custom job submission wrapper
cluster: "../scripts/SLURM_cluster_job.py"

# Specify custom job status command wrapper
cluster-status: "../scripts/SLURM_cluster_status.py"

Clone this wiki locally