Full course description
Term: Fall 2025
Date: October 24th, 2025
Time: 12:00 to 1:00 p.m.
Location: Online Only
Instructor: Chris Kuhlman
Presented By: Advanced Research Computing (ARC)
Description:
Prerequisites:
A beginner level understanding of computer programming concepts.
Basic experience with navigating a computer and running software.
Access to a VT network (e.g., on campus or connected to VPN).
An ARC account to follow along.
Preferred: Familiarity with working in a linux shell environment.
No additional software (e.g., compilers) is needed; all required software is on the ARC clusters.
Many jobs submitted by users to ARC clusters follow patterns (i.e., recurring themes). We address two of those in this workshop. One pattern involves submitting jobs for all combinations of user inputs—the so-called Cartesian product of variables. So if you have 14 values of one variable, 16 values of a second variable, and 23 values of a third variable, then this means that you will have 5152 (=14*16*23) different executions of your code. A tool, GNU/Parallel, elegantly assists with these types of jobs. The second pattern involves MPI—the Message Passing Interface standard—for distributed computing. There are several ways to control these types of code executions and we explore some of them here. Since these patterns are frequently occurring—meaning that other users find them useful/important—you might find them useful, too.
Topics covered are: (1) motivation and use cases for GNU parallel, (2) GNU/parallel example, (3) motivation for MPI program control, and (4) MPI examples.
These approaches can make both you and your code run more efficiently.