Pittsburgh Unveils Big Ben the Supercomputer
The Pittsburgh Supercomputing Center (PSC) has switched sports, but its newest and most powerful system, the Cray XT3, is another black-and-gold superstar, say officials. With a nod to the Pittsburgh Steelers quarterback, Big Ben is Cray XT3 serial #1 — the newest stage in the evolution of high-performance computing technology and a major boost for computational science in the United States. Acquired via a $9.7 million grant from the National Science Foundation (NSF) in September 2004, Big Ben — the first XT3 system to ship from Cray — comprises
2,090 processors with an overall peak performance of 10 teraflops: 10 trillion calculations per second. If every person on Earth, about 6.5 billion people, held a calculator and did one calculation per second, they would altogether still be 1,500 times slower than Big Ben. Big Ben, whose name also refers to Ben Franklin, will serve as a leading-edge computing resource on the TeraGrid. Built by NSF over the past four years, the TeraGrid is the world’s largest, most comprehensive cyberinfrastructure for open scientific research. With the availability of Big Ben, the TeraGrid will provide more than 50 teraflops of capability. The current lead system at PSC, LeMieux (evoking the Pittsburgh Penguins’ owner and star player), has been one of the most productive TeraGrid systems, and although LeMieux remains lively and available for the foreseeable future, Big Ben is expected to take over LeMieux’s role as the TeraGrid resource best suited for very large-scale, demanding projects. On a per processor basis, Big Ben is 2.4 times faster than LeMieux, which has 3,000 processors and provides six teraflops of peak performance. More than sheer processor speed, however, the most significant technological advance that Big Ben brings as a boost to U.S. research is “inter-processor bandwidth” — the speed at which processors share information with each other. Because of exceptional inter-processor bandwidth, Big Ben has already demonstrated nearly 13 times better performance than LeMieux on key applications when 1,000 or more processors are used. Many areas of research will benefit from this superior ability to communicate among processors, including nanotechnology, design of new materials, protein dynamics studies that lead to new therapeutic drugs, modeling of earthquake soil-vibration, and severe storm forecasting. “With this system,” said Arden Bement, director of the National Science Foundation, “we are fulfilling an important national goal — providing one of the fastest computing capabilities to the U.S. research community. We celebrate a significant leap in science and engineering research and education capacity. Richness of data, combined with powerful computing facilities and innovative people, promises a multitude of scientific breakthroughs. The Pittsburgh Supercomputing Center has all of these.” “PSC is a world-class computational center with proven success at putting high-performance computing systems to work running advanced applications to solve the toughest scientific problems. Cray is proud of its long and productive partnership with PSC,” said Jim Rottsolk, Cray Chairman and CEO. “We are dedicated to working with PSC to ensure the unqualified success of this machine and are looking forward to demonstrating the capabilities of the Cray XT3 that make it the ideal leadership-class computational system to serve scientists and engineers worldwide.” Established in 1986, PSC is a joint effort of Carnegie Mellon University and the University of Pittsburgh together with the Westinghouse Electric Company. “Computational science is a key to U.S. economic strength, and PSC’s leadership in this field for nearly 20 years is a national resource,” said Jared L. Cohon, president of Carnegie Mellon University. “Big Ben is the newest in a long line of innovative, high-performance systems PSC has brought into being. Each of these systems pushed the technological envelope, and each time PSC has made them into productive research tools for the national scientific community.” “It is widely recognized that university research is at the heart of most commercial innovation, and is a key driver of economic prosperity,” said Mark A. Nordenberg, chancellor of the University of Pittsburgh. “The fact that Western Pennsylvania is home to this extraordinary national resource is a reflection of the scientific strengths that exist in this region. When major research universities such as the University of Pittsburgh and Carnegie Mellon University join forces with industry — in this case, Westinghouse and Cray Inc. — and with the National Science Foundation, other federal agencies, and the Commonwealth, the result is a powerful partnership. We are deeply grateful for the NSF grant that made it possible to acquire ‘Big Ben’ and believe that the Pittsburgh Supercomputing Center is poised to build further on its strong history as one of the world’s leading centers for supercomputing.” “The Pittsburgh Supercomputing Center should be proud of its continued leadership in providing high-performance computational resources,” said Steve Tritch, president and CEO of Westinghouse Electric Company. “Westinghouse is proud of its long-term relationship with the PSC and the collaborative effort in providing innovative solutions to matters in science and engineering.” “We’re extremely gratified to be able to introduce the Cray XT3 system for the NSF,” said PSC scientific directors Michael Levine and Ralph Roskies in a joint statement. “and to make it available as a lead resource on the TeraGrid. As with previous new systems, going back to the CRAY X-MP in 1986, we expect Big Ben to follow in the footsteps of LeMieux and the CRAY T3E before that as a major contributor to U.S. science.” The Cray XT3 architecture of Big Ben is related to but differs from the “Red Storm” system, built by Cray for classified research at Sandia National Laboratory. Big Ben houses its 2,090 processors (AMD Opteron, 2.4 GHz) in 22 cabinets or nodes, each containing up to 96 processors. The Cray SeaStar interconnect, configured as a 3-D torus, sustains 6.5 gigabytes per second bidirectional communication between processors. Each processor has one gigabyte of memory, for more than two terabytes of aggregate memory. Big Ben uses a parallel file system — the Lustre File System from ClusterFileSystems (created from a Carnegie Mellon University collaboration) — which on key applications has demonstrated very efficient performance. Big Ben is served by 245 terabytes of local disk storage and an additional 200 terabytes of global disk. The Pittsburgh Supercomputing Center is supported by several federal agencies, the Commonwealth of Pennsylvania and private industry.