Matlab – Common FAQs

Table of Contents

How do I install Matlab?

How do I install Matlab to make Matlab check out licenses from the UNC license servers?

If I have an existing installation of Matlab and receive a new PLP code, how do I apply this new code?

Will Matlab run on Mac OS?

Why do I receive a java.log error when launching MATLAB 6.0 (and later versions) on Linux?

Why do my Matlab jobs immediately fail when running on the compute cluster?

How do I improve the performance of the Matlab GUI when running it on the cluster?

Additional help

How do I install Matlab?

MathWorks provides installation instructions for all operating systems on their site.

Installation Instructions

How do I install Matlab to make Matlab check out licenses from the UNC license servers?

The file license.dat, or network.lic, should be changed to contain:
SERVER license0.its.unc.edu 782bcb6f9d7d 1712
SERVER license1.its.unc.edu 782bcb6fa65f 1712
SERVER license2.its.unc.edu 782bcb6dc354 1712
USE_SERVER

If I have an existing installation of Matlab and receive a new PLP code, how do I apply this new code?

To apply a new PLP code, run the Matlab installation again. We suggest that you always say Yes when asked to check for Updates for products unless you know of specific reasons to revert to the CD version of a component. You may find that from one PLP code to the next you have access to different sets of toolboxes; this is intentional. The latest PLP code corresponds directly with the current valid license; if there is a change in the set of features made available in a license, those features will not appear as available for installation unless you are using the corresponding PLP code.

Will Matlab run on Mac OS?

If you have Matlab installed locally on your MacOS machine, you may want to make sure that the patch from MathWorks is applied.

MathWorks released the patches for MacOS X. The instructions in applying the patches are documented in detail by MathWorks as the following:

1. Is Matlab supported on Mac OS X 10.2 (Jaguar)?

2. Is Matlab supported on Mac OS X 10.3 (Panther)?

3. Is Matlab supported on the Mac OS X 10.4 (Tiger)?

Why do I receive a java.log error when launching MATLAB 6.0 (and later versions) on Linux?

When invoking Matlab on a Linux 2.4 machine, Matlab aborts when starting up and creates a java.log file in your home directory. You would see something like the following.

Opening Log file: /afs/isis/home/c/d/cdpoon/java.log.23456 Aborted

Most likely, your Linux machine is missing the en_US locale which the Blackdown Java VM requires for some character conversion operations. Normally, the file is located in /usr/share/locale/en_US directory, though it may vary from one glibc install to another.

Further information on this error can be found on MathWorks Technical Support page.

Why do my Matlab jobs immediately fail when running on the compute cluster?

If you find that your Matlab job dies after having accumulated very little CPU time and your LSF job output looks similar to

Exited with exit code 1.
Resource usage summary:
    CPU time   :      0.91 sec.
    Max Processes  :         1
    Max Threads    :         1
The output (if any) follows:

then the following may resolve the issue.

The easy solution is to use the bmatlab script to submit your Matlab job.

If you do not want to use the bmatlab script, then first modify your environment by executing the following command at the server prompt

unset DISPLAY

and then resubmit your Matlab job as before.

Also, if you want to reset your DISPLAY environment variable, then simply log off and log back into the server.

How do I improve performance of the Matlab GUI when running it on the cluster?

If you find that your Matlab session is running too slowly, then you can try to speed things up by adding a file called “java.opts” to the working directory of your Matlab session, containing the line (without the double quotes) “-Dsun.java2d.pmoffscreen=false”. This has been found to speed up the performance of Matlab on the cluster.

Additional help

More on Matlab

Research Computing home page