Skip to content


    Skip to content

    FROM THE DIRECTOR | Contents | Next

    Toward New Scientific Horizons with Terascale Computing

    BY
    Sid Karin
    NPACI and SDSC Director

    A t the recent dedication ceremony for NPACI's newest IBM RS/6000 SP supercomputer, named Blue Horizon, speakers representing NPACI, IBM, the NSF and Congress, the campus administration, and the scientific community, discussed the opportunities such a machine brings to the U.S. academic community. Capable of 1 trillion operations per second, Blue Horizon is the most powerful computer available to the U.S. academic community. The name "Blue Horizon" was chosen both to recognize the partnership with IBM that put the machine at SDSC and to suggest the new horizons toward which today's computational scientists continue to strive. Of course, science, just like a ship, never reaches the horizon, but constantly pushes it forward.

    OPPORTUNITY

    LEADERSHIP

    RESPONSIBILITY

    OPPORTUNITY

    The articles in this issue of enVision feature some of the first scientific computations to take advantage of Blue Horizon's unique computing capability. From fundamental properties of matter to the evolution of the universe, Blue Horizon not only can solve large problems more quickly, but also can tackle problems of a size that simply could not fit within the physical constraints of earlier generations of supercomputers.

    The astrophysics research of the team led by Lars Hernquist at Harvard University is a perfect example (see page 4). In 1995, Hernquist's team used weeks on SDSC's then-most powerful computer, a Cray C90, to simulate a pair of colliding galaxies comprised of 250,000 particles--at the time, the largest such simulation to date. Fast forward to 2000. In March, Hernquist's team simulated a collision between two galaxies comprised of more than 100 million particles--400 times the size of the 1995 model.

    The additional particles will provide not just finer detail, but lift the simulation past a mathematical threshold for a type of error inherent in models with fewer particles. Thus, these simulations represent new science, not just bigger computations. The same holds true for the other researchers featured in this magazine, and similar opportunities await scientists in other fields.

    Top| Contents | Next

    LEADERSHIP

    By bringing these opportunities to the scientific community, Blue Horizon also firmly solidifies the leadership of the National Science Foundation, the PACI program, and NPACI in academic computing. The scientific advances made possible by such technological innovation are driving the booming U.S. economy.

    NPACI's leadership, in partnership with IBM, has catapulted academic high-performance computing into the ranks of the Top 10 most powerful computers in the world. And the NSF's new Terascale Computing competition will soon push academic computing into the upper reaches of the Top 10. Future budget increases should be used to secure an even stronger position for the NSF as a world leader in high-performance computing.

    Within NPACI, we have extended our leadership role beyond operating a large computer system such as Blue Horizon. For example, NPACI's Strategic Applications Collaborations (SAC) teams are working with leading computational scientists to port and optimize their codes so that these scientists can quickly begin making new discoveries.

    Thanks to the leadership of the SAC teams, new science computations, using a thousand or more processors, have been conducted on Blue Horizon since January, less than a month after the machine's acceptance. And the machine will soon become available for production runs through the efforts of both SDSC's Scientific Computing and Advanced Systems groups.

    Top| Contents | Next

    RESPONSIBILITY

    NPACI's leadership role in teraflops-scale computing brings additional responsibilities. To ensure that Blue Horizon contributes to significant new science, the National Resource Allocation Committee will ensure that allocations are preferentially awarded to projects that take advantage of its unique capability to push forward the scientific horizon of their discipline.

    On a more practical note, NPACI and SDSC are responsible for keeping Blue Horizon in production--a specialized task for which few other organizations are qualified. NPACI's SAC teams will also continue to collaborate with scientists to move codes to Blue Horizon and jump-start new science. Further, these SAC projects are chosen so that not just the selected scientist benefits, but the entire field, either through scientific results or the availability of programs to other researchers.

    In addition, the results of basic research, from SAC projects and allocated users, benefit all disciplines, as NSF Director Rita Colwell has said. Advances in physics benefit chemistry, which affects biology, which influences environmental science. In keeping with the overall NSF mission and NPACI responsibility, Blue Horizon extends the capability for advancing the horizons of basic understanding in all disciplines. *

    Top| Contents | Next