SYSTEMS
Two of World's Top 500 Supercomputers Powered by Rackable Systems
SPECIAL COVERAGE FROM ISC2006-- Rackable Systems, a leading provider of servers and storage for large-scale data center deployments, today announced that its cluster deployment at the University of Florida's High Performance Computing (HPC) Center achieved a ranking of 280 in the latest TOP500 Supercomputing list, released this week at ISC2006. Based on the High-Performance Linpack (HPL) benchmark, the TOP500 list highlights the 500 most powerful computer systems in the world. The University of Florida deployed highly dense, thermally efficient rack-mount servers and storage from Rackable Systems in its HPC distributed-memory cluster, as well as at the university's Tier-2 Center, a member of a global grid computing infrastructure that supports the CMS (Compact Muon Solenoid) experiment and other computationally intensive scientific applications. The HPC Center's distributed-memory cluster earned an impressive 2.974 Teraflops and 81.3% parallel efficiency using the HPL benchmark to secure its spot as one of the world's fastest supercomputers. "Powering our facility with Rackable Systems' DC-powered servers ensures we maintain high levels of performance and reliability while running our demanding scientific and engineering applications," said Dr. Charles Taylor, senior HPC systems engineer at the University of Florida. "We packed three times the computational capacity of our previous cluster in half the space without significantly increasing our power consumption," he added. 
The University of Florida's distributed memory cluster consists of Rackable Systems Foundation Series 1U servers with over 800 AMD Opteron 275 processing cores. Additionally, an Infiniband interconnect fabric consisting of channel adapters, switches and software from Cisco Systems was used in conjunction with 40 Terabytes of OmniStor Fibre Channel RAID storage from Rackable Systems. "The TOP500 list serves as the benchmark of excellence among the supercomputing community, and we are thrilled to be recognized for powering the University of Florida's HPC Center," said Giovanni Coglitore, founder and chief technology officer of Rackable Systems. "Increasing computational requirements and power costs are driving demand for Rackable Systems' unmatched thermal and power efficient servers for HPC environments." Rackable Systems was also ranked 483 on the TOP500 list for its Emerald cluster at AMD's Developer Center in Sunnyvale, California. The cluster consists of 144 nodes, each leveraging two Dual-Core AMD Opteron Model 275 processors.
The University of Florida's distributed memory cluster consists of Rackable Systems Foundation Series 1U servers with over 800 AMD Opteron 275 processing cores. Additionally, an Infiniband interconnect fabric consisting of channel adapters, switches and software from Cisco Systems was used in conjunction with 40 Terabytes of OmniStor Fibre Channel RAID storage from Rackable Systems. "The TOP500 list serves as the benchmark of excellence among the supercomputing community, and we are thrilled to be recognized for powering the University of Florida's HPC Center," said Giovanni Coglitore, founder and chief technology officer of Rackable Systems. "Increasing computational requirements and power costs are driving demand for Rackable Systems' unmatched thermal and power efficient servers for HPC environments." Rackable Systems was also ranked 483 on the TOP500 list for its Emerald cluster at AMD's Developer Center in Sunnyvale, California. The cluster consists of 144 nodes, each leveraging two Dual-Core AMD Opteron Model 275 processors.