MATLAB

MATLAB Individual Edition and MATLAB Parallel Server are available on all nodes of the cluster.

Using MATLAB

MATLAB can be used through the following methods:

Using MATLAB Graphically through the ‘MatlabInJupyterLab’ LiCO Cloud Tool
Using a Graphical User Interface through X11 forwarding
Using a command-line/nodisplay MATLAB session
Executing MATLAB scripts as batch jobs
Using MATLAB Parallel Server to distribute computations

‘MatlabInJupyterLab’ LiCO Cloud Tool

The ‘MatlabInJupyterLab’ LiCO Cloud Tool provides Jupyter and the MATLAB Integration for Jupyter. This allows you to either get a full MATLAB interface from the browser or to use MATLAB code within a Jupyter notebook. Pleae login and access the internal page Cloud Tools for instructions on how to do this.

Graphical User Interface Session

To initiate a Graphical User Interface session, connect to the login node or a node which is allocated to you with X11 Forwarding enabled (please Login and access the internal site https://docs.lovelace.plymouth.ac.uk/private/ for instructions on this). Assuming that you have set up an ssh host named lovelace referring to the login node and you are on a supported platform, simply run:

ssh -XY lovelace
module load matlab_parallel_server
matlab

Command-Line Session

To load MATLAB and initiate a command-line session on a compute node, run:

module load matlab_parallel_server
srun --pty matlab -nodisplay

MATLAB Script as Batch Job

To run a MATLAB script as a batch job you may, for example, create a script called hello.m containing the following text (or any other script you might write):

disp('Hello, World!')

You may then run this script on a compute node as follows:

module load matlab_parallel_server
srun matlab -nodisplay < hello.m