|
|||||||||||||||||
|
|||||||||||||||||
 |
NCSA NEWS |
||||||||||||||||
|
NCSA's Itanium™-based Linux Cluster Will Aid Scientific Discoveryreleased May 29, 2001
Contact
Note: for more news about NCSA's Itanium-based clusters, see A range of scientific codes already running on new 64-bit systemCHAMPAIGN, IL — The first large Itanium™-based Linux cluster for the national science community will be delivered to the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign later this summer. The 160-node cluster will have a peak performance of 1 teraflop—or 1 trillion calculations per second. The cluster of 160 IBM dual-processor IntelliStation Z Pro workstations will be one of two terascale-level Linux clusters at NCSA, and the two clusters will be the fastest Linux clusters in academia. NCSA's other terascale cluster, which was delivered in March, is based on the Intel® Pentium® III processor and consists of 512 dual-processor IBM eServer xSeries servers. In addition, NCSA's experimental 32-node cluster of IntelliStation Z Pro workstations using the Intel Itanium processor already shows that Itanium-based clusters will be a top platform for high-performance computing in the years to come. PPM (for Piecewise Parabolic Method), a that code solves two and three-dimensional hydrodynamics problems, has already achieved in excess of 670 megaflops on a single Intel Itanium processor in the experimental cluster. A code used by the MILC (for MIMD Lattice Computation) collaboration has achieved single processor performance of 1.2 gigaflops. The MILC code is used to study quantum chromodynamics (QCD), a theory used to describe the relationships among the most basic known particles in the universe. "Scientists are excited about using our Itanium-based cluster and are already moving codes to this new platform," said Dan Reed, director of NCSA and the National Computational Science Alliance (Alliance). "To the scientific user, the high performance of 64-bit Linux clusters provides the opportunity to do breakthrough scientific calculations and offers scalability down to the desktop level." Delivery of NCSA's Linux cluster using Itanium-based platforms is still about a month away, but a variety of scientific codes—ranging from astrophysics applications to MILC and PPM—are running successfully on the experimental cluster. Although the primary purpose of this small cluster is to get codes up and running smoothly on Itanium-based systems in preparation for delivery of the terascale-level cluster, it has already proved to be a useful research tool. A team of scientists from the University of Minnesota's Laboratory for Computational Science and Engineering (LCSE) is using the 32-node cluster to create detailed computer simulations that show the behavior of turbulent gas flows. This data is now being visualized on a 1-billion-cell grid. The LCSE team expects to perform even more detailed simulations consisting of an 8-billion-cell grid. "This shows what kind of exciting work can be done on even a moderately-sized Linux cluster using Itanium-based components," said Rob Pennington, head of NCSA's computing and communications division. "The large, terascale-level machine that we expect to have running and available to the research community later this summer will have very significant benefits for the scientific community." The National Center for Supercomputing Applications is the leading-edge site for the National Computational Science Alliance. NCSA is a leader in the development and deployment of cutting-edge high-performance computing, networking, and information technologies. The National Science Foundation, the state of Illinois, the University of Illinois, industrial partners, and other federal agencies fund NCSA. The National Computational Science Alliance is a partnership to prototype an advanced computational infrastructure for the 21st century and includes more than 50 academic, government and industry research partners from across the United States. The Alliance is one of two partnerships funded by the National Science Foundation's Partnerships for Advanced Computational Infrastructure (PACI) program, and receives cost-sharing at partner institutions. NSF also supports the National Partnership for Advanced Computational Infrastructure (NPACI), led by the San Diego Supercomputer Center.
|
||||||||||||||||
|
|
|||||||||||||||||
