Over 1 Million Jobs Have Run on UA's Ocelote Supercomputer

When Ariella Gladstein submitted her genome simulation to Ocelote, the high performance computing system at UA's Research Data Center (RDC), she didn't think it was a big deal. But Gladstein, a PhD student in Ecology and Evolutionary Biology, marked an important milestone with her submission—it was the one millionth job for Ocelote.

Ocelote provides high-speed computation resources at no charge to all UA researchers on a first-available basis. With its 336 nodes of central processing units (CPUs), 46 nodes of graphics processing units (GPUs), and 2TB RAM Large Memory Node, Ocelote is helping researchers across a diverse range of disciplines, from astronomy to humanities.

"This milestone is significant for many reasons, but primarily because it demonstrates the demand for high performance computing and the value it provides to UA researchers," said Jeremy Frumkin, executive director of research technologies. Research computing staff marked the occasion by congratulating Gladstein with a UA baseball cap and technology prize package. 

Gladstein's research focuses on using genomes to infer population histories to determine how population sizes have changed over time.  Because Gladstein's research methods use complex simulations, high performance computing is necessary. "I do a lot of simulations of genomes and then I compare those simulations to real data. I've run at least one million jobs on Ocelote!"

UITS Senior Systems Administrator, Chris Reidy says Ariella is one of the "power users" of Ocelote. "We don't even count all of the jobs she does separately. When she runs an array of a thousand, we combine them as one."

Gladstein says that without a funding source, this type of research computing could be very expensive and prohibitive to getting a degree. "I don't have a research grant. There is no way I could have done any of my research without access to computing. You can buy computing power through Amazon Web Services, but when I run 6 million hours, which I have-over everything I have done, that would cost a lot without a research grant."

Gladstein's advice to students who have not yet realized the Research Data Center is available to them? "There is a lot of computing power here, so use it! You can basically do whatever computing you need here, for free."

Researchers interested in using the HPC and RDC resources can find out more at it.arizona.edu/service/high-performance-computing, or by sending an email to hpc-consult@list.arizona.edu.