How to Select Compute Resources for a Job on ARC Clusters

Term: Summer 2026

Date: July 14th, 2026

Time: 2:45 to 3:45 p.m.

Location: Online Only

Instructor:  Matt Brown

Presented By: Advanced Research Computing (ARC)

Description:

Prerequisites:
You will need an ARC account to follow along. 
Access to a VT network (e.g., on campus [eduroam] or connected to VPN).
Familiar with Unix shell/command line environment.
Preferred: Attended one or more previous ARC workshops or have a familiarity with ARC systems.
Preferred:  Understand the basics of slurm jobs.

Description:
ARC clusters have different computing resources (compute node architectures, types of processors, memory, numbers of compute nodes of different types), making the collection of ARC clusters purposely heterogeneous.  The clusters provide different resources with different strengths to better handle particular types of computations.  Similarly, computing needs across VT users vary.  Consequently, a goal of every user should be to assign your (slurm) jobs to resources that are (1) most efficient in performing your computations, and (2) most available to start your job sooner.  We address concepts and examples for how to select ARC resources for your jobs.