Location: Brendan Iribe Center, Room 1116
During this hands-on workshop, a MathWorks engineer will introduce parallel and distributed computing in MATLAB with a focus on speeding up application codes and offloading compute. By working through common scenarios and workflows using hands-on demos, you will gain a detailed understanding of the parallel constructs in MATLAB, their capabilities, and some of the common hurdles that you'll encounter when using them. In the afternoon session you will learn how to run your MATLAB code on local High Performance Computing (HPC) resources and how to submit jobs remotely from your laptop to run on Zaratan and/or Juggernaut.
The morning session is open to all MATLAB users on campus and will discuss the parallel constructs in MATLAB in general, which can be of use on multicore workstations and laptops. Practical exercises during this session will be performed using the MathWorks Cloud Center.
The afternoon session is aimed at users of the University's HPC clusters and will include instructions on using MATLAB Parallel Server to offload compute to the HPC cluster. For practical exercises during this session you will use the interactive MATLAB application on Zaratan using the onDemand Web Portal as well as your own laptop to remotely submit MATLAB jobs on Zaratan. Please have MATLAB version R2022b installed on your laptop!
Prerequisites
Some basic familiarity with MATLAB is assumed. If you are not familiar with MATLAB or a bit rusty, you might wish to take the free self-paced MATLAB Onramp course from MathWorks (about 2 hours).
System Requirements
This is a bring-your-own-laptop style event. As part of the workshop, you will be configuring your laptop to be able to offload MATLAB compute worker tasks to the UMD Zaratan HPC cluster, and it is better to do this on your own system than on a lab system.
So please bring a laptop with the following installed on it:
MATLAB R2022b including the Parallel Server Toolbox. This is available at no charge to members of the University of Maryland community via the campus Total Academic Headcount license;
an ssh client (on Windows preferably PuTTY).
All registered participants will get access to the Zaratan HPC cluster (for the duration of the workshop).