© 2008 by the Rector and Visitors of the University of Virginia.

The information contained on the University of Virginia’s Department of Information Technology and Communication (ITC) website is provided as a public service with the understanding that ITC makes no representations or warranties, either expressed or implied, concerning the accuracy, completeness, reliability or suitability of the information, including warrantees of title, non-infringement of copyright or patent rights of others. These pages are expected to represent the University of Virginia community and the State of Virginia in a professional manner in accordance with the University of Virginia’s Computing Policies.

ITCWeb: Matlab Support

Matlab Distributed Computing Toolbox

ITC has obtained a multi-user network license for the Matlab Distributed Computing Toolbox. The Distributed Computing Toolbox (DCT) enables you to solve computationally and data-intensive problems using MATLAB in a multiprocessor computing environment. You can use the toolbox to solve problems that comprise either several independent units of work or a single large computation by harnessing multiple processors.

For more information, see the Web page at
     http://www.mathworks.com/products/distribtb/

The DCT is installed on ITC's Linux clusters and is available for Linux machines that mount the Matlab installation from the UNIX server jeeves.itc. Using the Matlab Distributed Computing Engine on the Linux clusters allows The DCT to run up to 128 separate processes. Users are encouraged not to use more than 16 processes in any one job.

The following archive (.tar) file contains examples of how to submit jobs using the DCT to the Linux clusters through PBS. It should be downloaded to your cluster account and untarred (tar -xf dct_examples.tar).
     dct_examples.tar

Three things to note are:

1. DCT jobs have to be submitted from within Matlab, which is invoked using the command matlab. The example scripts show how to setup and submit the jobs.

2. The examples currently have the parallel example jobs being submitted to the cluster testq by adding '-q testq' to the qsub command in pbsParallelSubmitFcn.m so that you can get quick turnaround. To submit parallel jobs to the general workq, you would have to remove this from the qsub command in that file.

3. Use parallel_submit2.m and parallel_retrieve2.m as templates for your jobs that you would submit and then log off the cluster.

4. The standard PBS directives can be inserted in the pbsWrapper.sh and pbsParallelWrapper.sh shell scripts, e.g. #PBS -M user_id@virginia.edu.

The pdf version of the Mathworks presentation on the DCT from 4/23/07 can be found here.
(Accessible only from on grounds or using UVA-Anywhere)

General information about the availability of Matlab at U.Va. can be found on the Web page at
     http://www.itc.virginia.edu/research/matlab/

Specific questions about the Distributed Computing toolbox as well as general questions concerning the use of Matlab should contact ITC Research Computing Support; email us at Res-Consult@virginia.edu or by telephoning or visiting either the Research Computing Lab (243-8799) in the Brown Science and Engineering Library, Clark Hall or the Scholars' Lab (243-8800) in Alderman Library, 4th floor.