Brown supercomputer lifespan extended, space still available for purchase

August 6, 2019

ITaP Research Computing recently announced that the Brown community cluster supercomputer will remain in service through 2023, providing researchers with an additional year of value beyond the clusters’ typical five-year lifespan.

“With the end of Moore’s Law, the pace of processor innovation has slowed, and the relative utility of our existing investment in supercomputers has lengthened. It now makes financial sense to run a system like Brown for longer than five years, rather than spending the university’s funds on a new system that is not significantly more capable than Brown,” says Preston Smith, ITaP’s director of research services and support

Nodes are still available for purchase through ITaP Research Computing’s website. Brown, built as part of Purdue’s Community Cluster Program, is well-suited to a variety of science and engineering research.

Carlo Scalo’s work in aerodynamics and vortex dynamics is one example. Scalo, an assistant professor of mechanical engineering, and aeronautical and astronautical engineering, by courtesy, studies the flow of air over everything from a comparatively low-speed commercial airliner to a high-speed missile.

The higher the speed, the more computational power required to capture the details of the flow. Scalo relies heavily on Brown to accomplish this computationally intensive work.

Scalo also has access to even more powerful supercomputers at the the Department of Defense (DOD) Supercomputing Resource Center (DSRC), but finds himself needing to use them less often as Purdue continues to invest in campus computing resources. “The scope of what on-campus computing is able to do for my research has really grown,” he says.

Community clustering makes more computing power available for Purdue researchers than faculty and campus units could individually afford. ITaP Research Computing installs, administers and maintains the community clusters, including security, software installation and expert user support.

“It just makes sense financially,” says Scalo, of the community cluster program. It was one of the primary reasons he chose Purdue over other schools when he was interviewing for faculty jobs.

Like Scalo, 38 percent of the faculty partners in Brown are assistant professors – all increasingly reliant on high-performance computing. These early career faculty weigh the value of the community cluster program in their decision to come to Purdue.

Scalo points out that the appropriate comparison for someone contemplating an investment in Brown is not just the cost of buying and maintaining comparable hardware, but also the cost of dedicated expert support.

“Are you getting the staff that reply to you in five minutes, or show up in your office and code with you? Can you meet them in coffee shops around campus? You won’t, if it’s an external resource.”

Community cluster partners always have ready access to the cluster capacity they purchase, but they also share capacity fellow researchers aren’t using. This offers users access to substantially more computational power if needed and keeps the machines busy.

The Community Cluster Program now has 196 faculty partners from all three physical Purdue campuses, all of Purdue’s primary colleges, and 57 different departments. ITaP Research Computing delivered 373 million computational hours to community cluster partners in calendar year 2018. In all, 66 percent of Purdue’s fiscal year 2018 research expenditures ($276 million) were spent by Research Computing faculty.

To learn more about Brown or the Community Cluster Program, contact Smith, psmith@purdue.edu or 49-49729.

Originally posted: August 6, 2019  11:47am