.. highlight:: rst

Sample R Jobs
++++++++++++++

Installing R Packages
*********************

While users do not have administrative permissions on the rcfcluster, R packages can still be installed into your home directory. However, some R packages have dependencies that require administrative priveleges. If you need to install one of these such packages, please send a software request to rcfsupport@groups.umass.edu

Otherwise, to install a package: ::

    ssh rcfcluster
    module load R
    R
    install.packages("package_name")
    q()


Running R in Interactive Mode with Slurm
*************

After logging in to rcfcluster, enter an interactive bash environment (you can add additional arguments to specify resource requests, as described in the Getting Started section): ::

        srun --pty bash

This command will automatically log you in to the compute node that best fits your resource requests. Launch R with: ::

        module load R
        R


Submitting a Non-Interactive (Batch) Job to Slurm
*************

Below is an example of an sbatch script that calls an Rscript: ::

        #!/bin/bash
        #SBATCH --job-name=R-example

        # Log standard output & standard error to files. %j is just a placeholder for the unique slurm jobID:
        #SBATCH --output=/home/username/%j.out
        #SBATCH --error=/home/username/%j.err

        # Resource Requests: Running 1 job ("ntasks") on a single node ("nodes") with 1 cpu dedicated to that job ("cpus-per-task").
        #SBATCH --time=1:00
        #SBATCH --mem=500
        #SBATCH --nodes=1
        #SBATCH --ntasks=1
        #SBATCH --cpus-per-task=1
        #SBATCH --mail-user=username@umass.edu
        # Send email when job begins:
        #SBATCH --mail-type=BEGIN
        # Send email if job fails:
        #SBATCH --mail-type=FAIL
        # Send email when job ends successfully
        #SBATCH --mail-type=END

        # Load the R module in order to use R, then can use Rscript to launch job:
        module load R
        Rscript /home/username/myScript.R


Utilizing Batch Scripts to Submit Multiple Jobs in Parallel
*************

While the above example demonstrates how to launch a single R job on a compute node, you will receive the most benefit by running multiple jobs in parallel (perhaps each with a different set of parameters). An example of this is provided in the Getting Started section.