SDSC Offers Users a Bit of the World's Fastest Computer

The San Diego Supercomputer Center (SDSC) announced today that it is the first academic center to offer users allocated time on the new IBM Blue Gene, which provides the compute power for the world's fastest supercomputer. Users will now have access to more than 18 million processor hours available on a system that is optimized for data-intensive computing. Nicknamed Intimidata, this Blue Gene system is a 2048-processor, single-rack server that packs 5.7 teraflops of compute power and 20 terabytes of general parallel file system storage into a compact footprint. The system has a ratio of 1:8 I/O nodes to compute nodes, which is the maximum ratio available to support data-intensive computing. This gives Intimidata an aggregate peak speed of 5.7 teraflops and a total memory of 512 GB. "Intimidata is a unique configuration for the Blue Gene series. A Blue Gene configured for data-oriented users," said SDSC Director Fran Berman. "This rack provides a distinctive resource for our users which complements SDSC's other nationally allocated HPC resources - DataStar and the TeraGrid cluster. All are available to the community as TeraGrid resources." "The SDSC enables some of the most important research in science today. With the power of Blue Gene, SDSC can take its research to the next level," said Dave Turek, vice president of Deep Computing, IBM. "The collaboration between IBM and the SDSC continues to broaden, applying even deeper levels of computational power to some of the most important problems in science today." To ensure effective parallel processing, all compute nodes are connected by two high-speed networks: a 3-D torus for point-to-point message passing and a global tree for collective message passing. All I/O nodes are connected internally to the global tree and externally via gigabit Ethernet. This gives an aggregate I/O rate of 16 GBps in SDSC's data-optimized configuration. Offering enormous computing power in a small package, SDSC's BlueGene server enables reductions in power consumption, cooling and space requirements for institutions requiring immense computing power. The new architecture's ability to produce cost-effective compute power in such a compact package provides a glimpse into the future of supercomputing. Funding is being provided by the National Science Foundation and the Department of Energy.