Nov. 16, 2006
Previous Next

2006 Gordon Bell prizes go to LLNL's BlueGene/L simulations

TAMPA BAY, Fla. — A large-scale electronic structure simulation of the heavy metal molybdenum conducted on the world’s fastest supercomputer, BlueGene/L, today earned a team led by a former and current Lawrence Livermore National Laboratory (LLNL) scientists the 2006 Gordon Bell Prize for "peak performance."

Three teams were selected from six finalists for Gordon Bell Prizes in two categories. A team from IBM’s Thomas J. Watson Research Center also was awarded a "Special Achievement" Gordon Bell Prize for work on BlueGene/L and a Japanese team earned a Gordon Bell peak performance honorable mention. The awards were announced at SC06 in Tampa, Fla., an international conference of high performance computing, networking, storage and analysis. Named for one of the founding fathers of supercomputing, the prestigious Gordon Bell Prize is awarded to innovators who advance high-performance computing.

Led by Francois Gygi, formerly of Lawrence Livermore and currently at UC Davis, LLNL’s peak performance team includes IBM’s TJ Watson Research Center, Carnegie Mellon University and the Institute of Analysis and Scientific Computing at the Vienna University of Technology, Vienna, Austria. Their entry was titled "Large-Scale Structure Calculations of High-Z Metals on the BlueGene/L Platform."

Molybdenum, a high-Z or heavy metal, is of particular interest to scientists with the National Nuclear Security Administration’s (NNSA) Stockpile Stewardship Program, the effort to ensure the safety, security and reliability of the nation’s nuclear deterrent without underground testing. NNSA scientists use simulation to better understand the effects of aging on nuclear materials. The team used Qbox, a first principles molecular dynamics (FPMD) code, to achieve simulations of unprecedented scale and detail.

Qbox demonstrates excellent parallel efficiency and peak performance. A sustained peak performance of 207.3 TFlop/s was measured on 65,536 nodes, corresponding to 56.5 percent of the theoretical full machine peak using all 128k CPUs.

"Conducting predictive simulations of heavy metals, such as molybdenum, is a very challenging problem which requires a machine of BlueGene/L’s capabilities," said Erik Draeger of Livermore’s Center for Applied Scientific Computing. "It’s a challenge because this code has to do many things very efficiently to accurately render a simulation. Heavy metals are very complex elements."

This kind of simulation is much larger than any previously feasible FPMD simulation. Such calculations open the way for accurate simulations of the properties of metals, including the calculation of melting temperatures, defect energies and defect migration processes, studying the effects of aging on the structural and electronic properties of heavy metals and materials subjected to extreme conditions. Understanding the properties of heavy metals is critical to the Stockpile Stewardship Program

Visit the SC06 Website for more information on the Gordon Bell Prizes and to read abstracts of the winning entries.

Founded in 1952, Lawrence Livermore National Laboratory has a mission to ensure national security and to apply science and technology to the important issues of our time. Lawrence Livermore National Laboratory is managed by the University of California for the U.S. Department of Energy’s National Nuclear Security Administration.