"Penguin's Scyld TaskMaster advanced workload manager provided the best value proposition versus other options for cluster management and job control", stated Dr. Greg Newby, chief scientist, Arctic Region Supercomputer Center. "Penguin was able to provide a turnkey system, with centralized support for all system components. From our experience with the first Penguin cluster there was no question that the track record of the Pacman cluster gave complete confidence in choosing Penguin again."
The Pacman - Pacific Area Climate Monitoring and Analysis Network - is a Penguin Computing cluster comprised of AMD Opteron processors, with 2,080 compute cores and 89 terabytes of storage blades provided by Panasas ActiveStor 12. The AMD Opteron processors provide the best price/performance and ease of adoption for the familiar Linux x86_64 environment. Because many applications on Pacman utilize MPI for high-speed communication, InfiniBand connectors were chosen. Penguin provided a memory footprint of 4GB/core, in order to accommodate a variety of user needs including single-node and multi-node.
"Penguin's priority for working with ARSC was to install the best performing supercomputing cluster for all of their computational needs", stated Charles Wuischpard, CEO Penguin Computing. "Pacman's cluster architecture is appropriate for ARSC solving large compute and memory intensive parallel jobs."
The Pacman supercomputing cluster was built specifically for the Arctic Region Supercomputing Center at the University of Alaska Fairbanks to support University of Alaska Fairbanks' academic and research community, and was purchased in part from a nationally competitive award from the National Science Foundation.