Issue link: http://catalog.e-digitaleditions.com/i/77272
New Windows-based Cluster Supercomputer Now Available to Purdue Researchers Forestry and natural resources Professor Bryan Pijanowski studies how people and land affect ecosystems, examining such issues as projected loss of prime farmland to urbanization and the potential effects of climate change on fish habitats. His research involves computer modeling, sometimes on a national scale and at resolutions as fine as every 30 meters. The work can involve hundreds of gigabytes to terabytes of data to begin with, trillions of data points and software that in some cases will only run in Microsoft Windows. In the past, that could mean running a model for months on a desktop computer. But with a new cluster supercomputer using the high-perfor- mance computing version of Windows, now available to Purdue research- ers through ITaP, Pijanowski can get results in just a few hours, along with more detailed and accurate modeling. "We could actually run the entire world at 30-meter resolution if we wanted," Pijanowski says. "To do cutting-edge research at this scale puts us at the head of the line." computing accessible to more campus researchers, along with additional research computing software. "We want to serve everybody, ITaP built the Windows HPC cluster to make high-performance Smith, manager of research support for the Rosen Center for Advanced Computing, ITaP's research computing unit. The new cluster runs Windows HPC Server 2008 R2, the latest version of " says Preston Microsoft's operating system for high-performance computing. Windows HPC is an alternative to Linux-based clustering software packages, with similar features such as cluster management tools, a job scheduler and a message passing interface (MPI) library. Through community clustering, faculty partners and ITaP make more Purdue's New Top-ranked Research Supercomputing Cluster Up and Running Fast Alina Alexeenko thinks Purdue's new Carter commu- nity cluster should give her lab a lot more computing power to study rarified gas flow problems — which can span topics from spacecraft exhaust plumes to tiny micro-electro-mechanical systems, or MEMS, devices — at half the cost of some high-end servers she uses. Carter is ready for full production and available to any Purdue researcher or research group. More details are available on the Carter cluster information website, www.rcac.purdue.edu/userinfo/resources/ carter. To order, visit the Carter cluster order website, www.rcac.purdue.edu/order. Purdue faculty researchers who have tested Carter report that it can speed up the time to results for many research applications and enable more complex simulations. Alexeenko, assistant professor of aeronautics and astronautics, and her students have tested Carter with simulations involving the effects of particles in the atmosphere after a large meteor strike. They also have studied the vacuum systems used to remove water from pharmaceuticals developed in solution to stabilize them for life on the shelf. computing power available for Purdue research projects than faculty and campus units could afford individually. Community clustering also maximizes resources by sharing computing power among faculty partners whenever it is idle. Researchers always have ready access to their share and potentially more if they need it. If you're interested in the Windows HPC cluster, contact rcac-cluster- purchase@purdue.edu for more information. "The results have been encouraging," says Alexeen- ko, who uses the Coates and Hansen community clusters in addition to her servers. "Basically, it's about three times faster than what we have now." Carter ranked 54th on the latest TOP500 list of the world's most powerful supercomputers and was among the half dozen most powerful machines at U.S. academic institutions. It was the most powerful on a U.S. campus where the research computing facilities are not part of a federally funded laboratory. Writer: Greg Kline is a science and technology writer with Information Technology at Purdue. Summer 2012 11 (Argonne National Laboratory CC)