ACADEMIA

High-performance computing clusters (HPCC) continue to gain momentum at the world's leading research institutions, as reflected by three universities that are deploying new or expanding their current standards-based Dell clusters. "Standards-based high-performance computing clusters offer customers major computing performance for lower cost that traditional proprietary systems, enabling universities to conduct research on their own campuses, instead of having to buy time at the nation's largest supercomputing centers," said John Mullen, vice president of sales, Dell Higher Education. Mullen said that Dell's ability to drive supercomputing costs down is an example of the "Dell Effect," defined as the drop in prices and resulting increase in value that occurs when Dell enters a product category or new market. University of Tennessee -- Chattanooga The University of Tennessee at Chattanooga (UTC) has completed Phase 1 of its three-part cluster strategy, having installed 137 Dell PowerEdge 1850 servers running the Linux operating system. Housed at UTC's Computational Simulation and Design Center (SimCenter), the Dell cluster is focused on computational field simulation, including solid and fluid mechanics, electromagnetics, energy and mass transport, chemical reactions, and materials science. Wally Edmondson, systems administrator at UTC's SimCenter, expects to expand the Dell cluster to more than 430 nodes by the end of 2004. "The Dell cluster enables us to do calculations faster and more accurately than we could before, and it offers the price for performance we need," Edmondson said. Duke University Duke University is using a 264-node cluster comprised of Dell PowerEdge 1750 dual-processor servers to perform research in computational biology, bioinformatics, genomic studies, statistics and computational chemistry. Duke has established a participatory model for the cluster where the university provides systems support, infrastructure and facilities while research groups supply hardware. These groups are granted access to the full cluster. "This shared cluster model allows researchers to have the power of a large cluster while only contributing a fraction of the cost," said Dr. William Rankin, director of the cluster and grid technology group within the Center for Computational Science, Engineering and Medicine at Duke. "With so many users depending on the cluster, we must ensure that we maintain a high percentage of uptime. Our Dell team offers us the right support at the right time." Pennsylvania State University Penn State has expanded its two Dell clusters to 176-nodes each. They're used by more than 500 researchers in the fields of engineering, meteorology, chemistry, biology, statistics and more. Known as Lion-XL and Lion-XM, these clusters were built using Dell PowerEdge 2650 servers and Dell PowerEdge 1750 servers, respectively. Both contain Intel(R) Xeon dual processors and run the Linux operating system. An additional 40 Dell servers and PowerVault storage systems make up another cluster at Penn State named Pleiades. This cluster makes available nearly 40 terabytes of data to an international community of gravitational physicists. "The standards-based clusters offer us a very attractive and cost-effective solution for solving large computational problems in many disciplines," said Vijay Agarwala, Penn State's director of Graduate Education and Research Services. "It provides supercomputing-class performance at a fraction of the cost." Dell HPCC experts in its custom solutions lab, as well as Dell Services work closely with many of these universities to plan, build and deploy these large-scale systems. The Dell HPCC program uses standards-based building blocks in the clusters to simplify the platforms and drive out cost. Dell engineers work with best of breed partners to integrate the latest standardized, high-performance technologies and drive standards where they are needed. For more information on Dell supercomputing clusters, please visit www.dell.com/hpcc.