8:30 AM-9:30 AM
Room: Texas Ballroom B
Chair: Michael T. Heath, University of Illinois, Urbana
Numerical simulation plays an increasingly important role in many scientific and engineering applications, including global climate modeling and computational fluid dynamics. In the case of the nation's stockpile stewardship program, it plays a central role, as evidenced by the U.S. Department of Energy's Accelerated Strategic Computing Initiative (ASCI). This ambitious program is opening the door to terascale scientific simulation, and in so doing, redefining the state of the art in supercomputing. ASCI already has fielded three 1TF machines and will deploy the world's first 100TF machine in 2004. These massively parallel computers will have upwards of 100,000 processors and there are myriad mathematical and computer science issues that must be addressed to ensure that these machines can be employed effectively to explore complex physical phenomena.
The speaker will discuss the challenges facing ASCI, with emphasis on scalability (of platforms, algorithms, and implementation). He will discuss the need for scalable numerical algorithms, such as multilevel methods for the solution of PDEs, and will present some recent results in this area. He also will motivate the urgent need for scalable data exploration methodologies, such as multiresolution data analysis and pattern recognition techniques, that are needed to help scientists glean insight from their petabytes of simulation output. The need to develop the mathematical foundations for simulation science also will be mentioned. The presentation will include several recent "first of their kind" simulation results drawn from the defense, environmental, and life sciences that demonstrate the utility---and difficulties---of terascale scientific simulation.
Steven F. Ashby
Center for Applied Scientific Computing, Lawrence Livermore National Laboratory