Middle Park alumnus Sam Elliott is travelling to Leipzig, Germany, to compete with the CU Boulder Super Computing Team. Held as part of the annual International Super Computing Conference, the CU team will square off against American rivals from Purdue University, Harvard, and the Massachusetts Institute of Technology as well as an international field including Germany, South Africa, Australia, and the Chinese National Institute of Defense.
Super Computing involves “clusters” of high-powered computers linked together to process very large data sets for complex calculations beyond the ability of typical desktop machines. Applications include weather prediction and climate modeling, genetic mapping, astrophysics, and military. The CU team has assembled a rack of four R720 Servers contributed by Dell Computers. The teams are given problems to solve and are limited to a very small power budget (mostly for cooling fans) which keeps the hardware relatively small. The key, according to Elliott, is carving out logical parts of the massive data sets, then team members run different applications in an optimal order, ultimately pulling it all back together again for a solution. In last year’s challenge in Denver, where Elliott attended his first competition, the teams were given global weather data collected in 2012 one week before Hurricane Sandy hit the East Coast and asked to predict landfall.
“The first thing is to get the Apps to run, because the data sets are so massive and we can easily crash,” explains Elliott, “then you try to go faster than the other teams.” Elliott is an MPHS 2012 graduate majoring in engineering physics at CU Boulder, and is working this summer at the National Oceanic Atmospheric Administration.