INDUSTRY

SGI today announced that Japan's Institute of Statistical Mathematics will embark on a new generation of scientific and sociological research with an SGI(R) Altix(TM) 3000 supercluster driven by 256 Intel(R) Itanium(R) 2 processors. The Linux(R) OS-based SGI Altix system enables the Institute of Statistical Mathematics to run unprecedented workloads through an enormous 1.9 terabytes of global shared-memory across four nodes. Part of Japan's Ministry of Education, Culture, Sports, Science and Technology, the Institute of Statistical Mathematics will use the system to power its "Modeling of Phenomena" initiative aimed at uncovering key trends from an ever-increasing body of statistical data on Japan's environment, culture, economy, molecular biology, earth science and even natural disaster patterns. With an emphasis on discovery and prediction, the government and academic research institute will pursue computational assumptions based on large and complex data sets that could not have fit into alternative solutions. "Our new system will enable researchers to develop statistical data analysis software and to analyze massive data sets without having to worry about memory spaces," said Professor Genshiro Kitagawa, director of the Institute of Statistical Mathematics. Founded in 1944 and reorganized in 1985, the institute has depended on Japanese mainframe computers and a vector supercomputer for its statistical science research. But as the complexity of statistical models increased, so did the requirement for memory and compute capabilities. As a result, the institute opted to push beyond the limits of vector computing by implementing a scalable 256-processor Altix system with 1.9 terabytes of global shared-memory. Japan's largest Altix system is scheduled to be fully operational in early 2004, and will interface other SGI(R) solutions, including 5.12 terabytes of storage capacity provided by an SGI(R) TP9100 disk array. "More and more in statistical science, researchers are handling huge amounts of data in the Monte Carlo random number generation method," said Professor Kitagawa. "Vectorizing these computational codes proves increasingly difficult, but parallel computing is capable of achieving much higher- performance. Our new integrated system, which combines the SGI Altix 3000 supercluster with the NEC SX-6, will provide us with an advanced computing environment that our existing installation could never have achieved. We expect this new deployment to contribute significantly to the advancement of statistical science research in Japan." "Japan's Institute of Statistical Mathematics is pursuing socially significant work as a world-leading institute of statistical science," said Norio Izumi, president of SGI Japan. "We are pleased to provide this esteemed institute with the world's largest Altix 3000 system comprised of 256CPU with 1.9TB memory to help them promote advanced studies." The SGI Altix family of servers combines industry-standard 64-bit Linux with the Intel Itanium 2 processor family and SGI(R) NUMAflex(TM) architecture to enable global shared memory systems from a few to hundreds of processors with up to 4 terabytes of shared memory, which is a first for Linux OS-based computing. Powered by the third-generation NUMAflex supercomputing architecture, even the largest data sets can be handled and analyzed with ease and in record time for production workflows and the most demanding stability. Only the SGI Altix family of servers is designed around this scalable shared-memory architecture that analyzes data sets as whole entities, without breaking them up into smaller segments to be handled by individual processors through I/O bottlenecks. The Altix architecture has proven ideal both for complex shared-memory applications running on a large single system image, and for communication-intensive applications optimized for clustering in throughput workflows.