Select a topic or year
ARMONK, NY - 08 Aug 2005: IBM today announced that New York University (NYU) has installed and begun to make use of a new IBM eServer™ BladeCenter®-based computing system capable of peak performance of 4.5 TeraFlops. According to the TOP500 List, a ranking of supercomputers published at www.top500.org, NYU's supercomputer is the fastest in New York City and the 117th fastest supercomputer in the world.
The new compact system will support NYU research with heavy computational requirements. Among the research enterprises that will be early beneficiaries of the new system are the Center for Atmosphere-Ocean Science in the Courant Institute for Mathematical Sciences, which is developing sophisticated models to study the behavior of the Earth's atmosphere and oceans, and a collaborative effort between the Center for Comparative Functional Genomics and the Courant Institute of Mathematical Sciences on genomics and bioinformatics.
By using IBM BladeCenter technology, NYU was able to create a supercomputing cluster with a footprint of only 30 square feet, allowing the University to best utilize valuable New York City real estate space.
"IBM's off-the-shelf blade server technology allowed us to easily connect a cluster of systems and solve the challenge of building a supercomputer in valuable space here in the city," said Marilyn McMillan, Associate Provost and Chief Information Technology Officer of New York University. "By leveraging IBM's Power Architecture technology with Linux, we are able to tap into a 64-bit computing platform that will greatly speed the research process and help our researchers to more rapidly turn scientific theory into real-world benefits to society. This supercomputer re-establishes NYU's position of leadership in high performance computing."
"The innovative design of IBM's BladeCenter has revolutionized high-performance computing, offering organizations such as NYU the option to scale-out its infrastructure in a space-saving manner without sacrificing power or performance," said Dave Turek, vice president of Deep Computing, IBM. "We're pleased to work with NYU on a powerful, state-of-the-art clustered supercomputing solution that will help the Courant Institute and other parts of the University maintain leadership in scientific research and help make IBM's JS20 systems the platform of choice for high performance computing applications."
One of the NYU research enterprises already designated to make use of the new supercomputing system is the Center for Atmosphere-Ocean Science in the Courant Institute for Mathematical Sciences (http://caos.cims.nyu.edu/page/home). The Center uses computational modeling to simulate the atmospheric circulation and oceanic currents. The acquisition of the supercomputer, supported in part by the Department of the Navy, Office of Naval Research, will help researchers combine theory, mathematical analysis and numerical models to better understand the Earth's climate, including the impacts of weather events such as thunderstorms and hurricanes, coupled atmospheric and oceanic oscillations such as El Nino, and oceanic currents like the Gulf Stream and the Antarctic Circumpolar Current.
The supercomputer will also help foster a research program in genomics at NYU, enhancing collaboration between the Center for Comparative Functional Genomics and the Courant Institute of Mathematical Sciences. Aspects of this research include advanced algorithm and software development for data integration, data mining and visualization, biological network simulation, and high-throughput microscopy. The acquisition of the supercomputer has been supported in part by the U.S. Army Medical Research and Materiel Command (USAMRMC).
IBM is also working with New York University to create a joint test center, where high performance computing application developers can enable, test and tune their applications and solutions using IBM JS20 systems. This center, comprised of a 64-node version of the supercomputer, will be available to developers and will help to promote innovation utilizing IBM's BladeCenter JS20 and Linux on Power platforms.
The supercomputer is capable of a peak performance of more than 4.5 TeraFlops, and consists of a cluster of 256 IBM eServer BladeCenter JS20 systems running Red Hat Enterprise Linux AS 4 (RHES AS 4), each with two, 2.2 GHz PowerPC 970 processors, or 512 processors in total. It is the first supercomputer in the United States that is using IPv6, the next generation Internet Protocol. IBM's joint test center project with NYU will consist of 64 IBM eServer BladeCenter JS20 systems (128 processors) running Linux, and capable of peak performance of more than 1 TeraFlop. Myrinet switch technology connects each BladeCenter cluster in the high-performance system.