APPLICATIONS
Voltaire Continues To Lead InfiniBand Deployments in Top 500
- Written by: Writer
- Category: APPLICATIONS
Voltaire today announced that the company's InfiniBand-based Grid Backbone(TM) switching solutions are powering eleven of the world's fastest supercomputers as ranked on the Top500 list released this week. This includes NASA's Columbia supercomputer, which is the top ranked InfiniBand site and holds the number-four position on the list with performance of 51.87 teraflops. Voltaire Grid Backbone solutions deliver industry-leading performance and networking efficiency to enable highly scalable clusters and distributed computing architectures. The Top500 list, published twice a year prior to the Supercomputing Conference held this week in Seattle, and the International Supercomputer Conference held in June, ranks supercomputers worldwide according to their performance on the LINPACK benchmark. Voltaire continues to lead the growth of InfiniBand deployments on the prestigious list with eleven supercomputer clusters, more than any other InfiniBand systems vendor. The majority of the Voltaire installations leverage the Voltaire Grid Director ISR 9288, the industry's highest port count multi-service switch with InfiniBand, Gigabit Ethernet and Fibre Channel integrated into a single chassis. The Grid Director ISR 9288 has been commercially available since September 2004 and is used by businesses and research institutions around the globe to build high-performance clusters and grids ranging from hundreds to thousands of nodes. In addition to having the largest number of Top500 InfiniBand sites, the Voltaire installations highlight the company's strong OEM and reseller partnerships with many leading systems vendors and integrators. The Voltaire InfiniBand-based clustered supercomputers on the list were delivered through the broadest array of server and reseller partners including: HP, IBM, SGI, NEC, Dell, Appro and GraphStream. "We are extremely pleased to maintain our leadership position in delivering high-performance InfiniBand-based supercomputers to our Top500 customers. It is encouraging to see customers validate InfiniBand technology as Ethernet-based solutions drop out of the top 50 ranking," said Ronnie Kenneth, chairman and CEO, Voltaire. "This new Top500 list showcases the results of our commitment to work with our world-class OEM and reseller partners to deliver high performance, scalable and network efficient systems to our mutual customers." Voltaire congratulates and thanks the following customers for reporting their use of Voltaire InfiniBand-based solutions to the November 2005 Top500 list: Installation Site (Rank) NASA/Ames Research Center/NAS (4) NASA/Goddard Space Flight Center (55) Wright-Patterson Air Force Base/DoD ASC (61) Hewlett-Packard (101, 128) Los Alamos National Laboratory (132) Trinity College Dublin (226) Sandia National Laboratories (286) Institute of Genomics and Integrative Biology (293) HLRS/HWW/Universitaet Stuttgart (372) Lawrence Livermore National Laboratory (408) Voltaire Displays DDR at Supercomputing 2005 In a separate release this week, Voltaire announced the immediate availability of double data rate (DDR) InfiniBand for its Grid Director switching series. Voltaire's DDR solutions deliver double the performance of 10 Gbps interconnect offerings and allow for the use of fewer switch chassis, switching chips and cables, thereby minimizing cluster complexity and costs. Voltaire is showcasing the Grid Director switches with DDR in booth 1630 at Supercomputing 2005. Voltaire also announced this week that Japan's Tokyo Institute of Technology (Tokyo Tech) has selected Voltaire's InfiniBand-based Grid Backbone switching solutions as the high-performance interconnect for a powerful Sun Microsystems supercomputer that will be used for computational scientific research. Tokyo Tech's supercomputer initially will deliver over 85 trillion floating point operations per second (TFlops) surpassing Japan's Earth Simulator, currently the largest supercomputer in Japan at 40 TFlops. This is planned to be extended to more than 100 TFlops at the time of operation in Q1 2006. The Tokyo Tech system is expected to be one of the five largest supercomputers in the world as ranked by the Top500.