The Purdue team in the competition, where teams of undergraduates from around the country pushed their student-run supercomputers to the maximum over three days, won the award for getting the most done on the least amount of power, organizers announced Thursday (Nov. 19). Purdue also won the power consumption award last year.

The team spent a lot of time during the setup phase of the event testing different configurations to see how they affected power consumption, said team member Max Hapner, a senior in computer and information technology from South Whitley.

In part, that was to make sure the Purdue entry stayed under surprise power limits imposed by the competition’s organizers. But the students also were conscious of the issue of data-center power consumption, which has become a major concern and was one focus of the SC09 conference.

Purdue team member David King, a senior in electrical and computer engineering technology from Lafayette, wrote a special program to monitor the power consumption of Purdue’s Cluster Challenge machine. Team member Charles Timko, a junior in computer science from Highland, visualized the data on a large screen monitor.

The other Purdue team members are; Alex Miller, a sophomore in earth and atmospheric sciences from Zionsville; Michael Niksa, a senior in computer engineering from Cedarburg, Wis.; Michael Wleklinski, a junior in chemistry and statistics from South Bend; and Alex Younts, a junior in computer science from West Lafayette.

Purdue was competing against teams from the University of Colorado, Arizona State University and State University of New York at Stony Brook this year. Purdue's team finished second to Stony Brook in the overall competition.

The mini refrigerator-sized 2009 Cluster Challenge entry from Purdue was like a small version of the Coates cluster that Information Technology at Purdue (ITaP), Purdue's central information technology organization, installed this summer. Coates is the fastest campus supercomputer in the Big Ten and one of the fastest in the world. The student cluster was made up of similar HP computers containing AMD microprocessors. It also boasted similar high-speed 10 Gigabit Ethernet (10 GbE) networking from Juniper Networks to link the individual computers into a single powerhouse machine.

The Cluster Challenge is part of the largest international conference for high-performance computing, networking, storage and analysis. Cluster Challenge teams must run a set battery of benchmarking programs and working scientific applications as efficiently as possible. The team members shared the load monitoring and adjusting their machines around the clock from Monday morning to Wednesday evening. The 2009 applications covered climate modeling, computational chemistry and data visualization, among other things.

ITaP and its research computing arm, the Rosen Center for Advanced Computing, help sponsor the Purdue team.