slurmfun

MATLAB tools for submitting jobs to the SLURM workload manager.

Summary

This repository provides tools for submitting MATLAB jobs to the SLURM scheduling system. The MATLAB function slurmfun can be used similar to the MATLAB function cellfun, i.e. it will apply a user-defined function to all elements of a cell array and return a cell array of output arguments. Each function call will be submitted as a separate job to the scheduler. See also help slurmfun for details.

Cluster Environment

Up to now slurmfun has been used and tested in the cluster environments of the Ernst Strüngmann Institute (ESI) gGmbH for Neuroscience in Cooperation with Max Planck Society and the Cooperative Brain Imaging Center comprising

• SLURM 17, 20.02, 20.11.9
• MATLAB 2014a up to 2023b
• Debian 8, RHEL 8.1-8.6

The default paths for log files and the SLURM working directory ('slurmWorkingDirectory', availableToolboxes in slurmfun.m), need to be adjusted for the specific cluster environment slurmfun is used in.

Installation

Clone this repository

git clone https://github.com/esi-neuroscience/slurmfun.git

and add it to your MATLAB path

addpath /path/to/slurmfun

Usage

Consider the function myfunction that generates a in1-by-in2 matrix of uniformly distributed random numbers in the interval (0,1):

function out = myfunction(in1, in2)
out = rand(in1,in2);

To generate 5 matrices in parallel with slurmfun use, e.g.,

nJobs = 5;
inc1 = num2cell(randi(20, nJobs, 1) + 60);
inc2 = num2cell(randi(20, nJobs, 1) + 60);
[out, jobs] = slurmfun(@myfunction, inc1, inc2, 'partition', 'partitionName', 'mem', '7500M', 'cpu', 1);

A more elaborate example can be found in Examples, for a full list of options, use help slurmfun.

Contact and Support

To report bugs or ask questions please use our GitHub issue tracker.

Project Status

This project is actively maintained and (sometimes) updated.