Acquisition of High-Memory Computation Hardware for UW–Madison Campus

“Big data” analysis problems — assembly of DNA sequences obtained from communities of many species, ecosystem simulation, the modeling of complex chemical reactions, as well as areas of computer science and mathematics — are key to research programs across the UW–Madison campus. They grow only more common as the campus adds more equipment like cutting-edge DNA sequencers, but they also require unusually powerful computing resources to process.

UW–Madison’s Center for High Throughput Computing (CHTC) can muster the power of more than 22,000 computing cores for use by campus labs, free of charge, but only two of CHTC’s roughly 600 servers have adequate main memory capacity (more than 128 gigabytes) for big data analysis. Access to such high-memory compute resources on national infrastructures, such as XSEDE’s Bridges cluster, is severely limited.

With UW2020 support, CHTC will add five state-of-the-art high-memory servers, each holding 3 terabytes of random access memory. This addition to the campus’s shared computing core will provide UW-Madison researchers with a unique, no-cost resource with which to carry out high-memory computation and free them of the need to outsource such work to national or commercial facilities off campus.