Purdue research computing resources for faculty among best in the nation

August 4, 2011

To gauge whether specific genes may play roles in diseases, Purdue statistics Professor Hyonho Chun often repeats her calculations thousands of times for each of thousands of genes.

That makes for a lot of heavyweight computation and it helps to have a heavy-duty computing system to do the work. Chun, a biological statistician with a focus on genomics, has one of the most powerful research supercomputers in the world at her disposal. She’s a faculty partner in Purdue’s Rossmann community cluster, which made the latest TOP500 supercomputers list.

The TOP500 Supercomputer Sites project has been ranking the 500 most powerful known systems twice a year since 1993 to track trends in high-performance computing. Purdue’s Coates cluster also made the latest list. The Steele cluster, built in 2008, made the list that year and in 2009.

Purdue now ranks fourth among U.S. academic institutions with systems listed and tops the Big Ten. Faculty partners and ITaP already are in the process of building another cluster, called Hansen, which is expected to be operational in September 2011 and should make the next TOP500 list.

Purdue faculty researchers wishing to see available hardware and to order capacity in the Hansen cluster can visit: http://www.rcac.purdue.edu/order/.

Like Rossmann, Coates and Steele, Hansen is part of Purdue’s 2010 Campus Technology Innovators Award-winning Community Cluster Program. The program is an ITaP partnership with faculty researchers — more than 100 to date, many of whom have invested in more than one of the clusters. The program has increased the research computing power available at Purdue by more than 10 times since it began in 2008.

Chun expects to need that kind of computing power even more as the research in her lab progresses. The standard now is to “divide and conquer,” to break genomes into manageable pieces for examination. But Chun and other researchers are looking to divide less and conquer more going forward.

“You want to see more genes at one time,” she says. “My research is building in that direction.”

Rossmann is used cooperatively by faculty in aeronautics, biology, chemistry, engineering, nanotechnology, physics and statistics, among other fields. All told, Purdue’s three community clusters had delivered more than 300 million research computing hours to faculty and their students as of summer 2011.

Through community clustering, ITaP pools internal and external funds into a group purchase to garner better prices from vendors and make more computing power available for Purdue research projects than faculty and campus units could afford individually.

ITaP then installs, administers and maintains the community cluster systems, including security and user support, so researchers can concentrate on doing research rather than on running a high-performance computing system.

Community clustering also maximizes the use of resources by sharing computing power when it is idle. Researchers always have access to the capacity they purchase, and potentially more if they need it.

Medicinal chemistry and molecular pharmacology Professor Markus Lill has bought into all three community clusters for his research in computer-aided drug discovery, which focuses on identifying compounds that may lead to more effective drugs and fewer side effects.

“What was promised — price, maintenance, sharing of nodes — everything, I think, works nicely,” says Lill. His Rossmann participation came in conjunction with a National Institutes of Health-funded project exploring how the body metabolizes a variety of compounds, a process integral to drug efficacy.

As with the other clusters, ITaP surveyed faculty about their needs in deciding on Hansen’s specifications, such as storage and memory options, including a large memory option benefitting certain kinds of calculations.

Writer: Greg Kline, science and technology writer, Information Technology at Purdue (ITaP), 765-494-8167 (office), 765-426-8545 (mobile), gkline@purdue.edu

Originally posted: August 4, 2011