CHPC Cluster Drawing



Media Stuff

Network Links

Data Center Weather


SC12 Student Competition

SC12 will be in Salt Lake City this year. In addition to sponsoring the largest booth CHPC has had at the annual event, CHPC staff are also mentoring the University's team in the student cluster competition. The team is made up of six computer science students: Leif Andersen, Bruce Bolick, Ian King, Tom Robertson, Kathryn Rodgers, and Tyler Sorenson. Professor Mary Hall, School of Computing, is the team's faculty advisor. The two-day competition requires that each team build a small cluster from component parts they will receive on the first day of the competition. This cluster, known as the LittleFe, is a complete 6-node Beowulf style cluster. Each node has two 64bit Intel CPU cores (4 hyper-threaded cores) and a cuda capable nvidia geforce embeded GPU. Brian Haymore, a CHPC hardware expert, has guided the students as they practiced building their cluster with parts that replicate what they will get at the competition and as they learn how best to exploit the hardware's capabilities.

Once their machine is built, the students have the remainder of the first day to install the operating system and open-source software of their choice, test, tune and run HPCC benchmarks. Brian, along with Martin Cuma and Wim Cardoen, CHPC software specialists, are helping the students as they choose software and learn how to optimize and compile code for best results.

At the end of the first day, the teams will receive the data sets. This year the challenge is the computationally complex ''Traveling Salesman Problem'' (TSP). The data include three sets of cities and coordinates. The students must construct an optimal algorithm to find the shortest route that includes a visit to each city once and a return to the city of origin for each of the data sets. The teams will be judged on how frequently they have the current best score and the visualization of their results. The judges will also interview the teams to access their knowledge of HPC, parallelization and the TSP.

The teams will be working at their tasks in full view of conference attendees. If you are attending SC12, find the University's team on the convention floor and cheer them on.