Six research teams have begun using the first phase of the Blue Waters sustained-petascale supercomputer to study some of the most challenging problems in science and engineering, from supernovae to climate change to the molecular mechanism of HIV infection.
The Blue Waters Early Science System, which is made up of 48 Cray XE6 cabinets, represents about 15 percent of the total Blue Waters computational system and is currently the most powerful computing resource available through the National Science Foundation.
Once fully deployed, Blue Waters is expected to make arithmetic calculations at a sustained rate in excess of 1,000-trillion operations per second (a "petaflop" per second). It will enable researchers across a variety of disciplines to tackle some of the most challenging research issues in science and engineering.
"This is an exciting and important milestone in the Blue Waters project," says Irene Qualters, program director of the NSF Office of Cyberinfrastructure. "It began as an idea, and now thanks to sustained collaborative efforts by the entire project team, the vendor and the science teams, this computational tool is beginning to advance fundamental understanding in a wide range of scientific topics."
More than two dozen research teams have been awarded Petascale Computing Resource Allocations (PRAC) through a competitive NSF-led process. The PRAC awards enable these teams to work with NCSA to prepare their codes to take full advantage of Blue Waters and other extreme-scale computing systems. The teams submitted proposals outlining how they could use the Early Science System during the limited time it is available before being integrated into the full Blue Waters system.
"All of these outstanding science and engineering teams are poised to do great, boundary-expanding work. The achievements of the first set of pioneers will soon be followed by those of their colleagues when the full system becomes available later this year," says NCSA Director Thom Dunning, principal investigator for the Blue Waters project.
The Early Science System research teams and their projects are:
Homayoun Karimabadi's team, University of California-San Diego, is modeling high-temperature plasmas, including magnetic reconnection and flux transfer events to better understand the impact of the solar wind and solar flares on the Earth's atmosphere.
Brian O'Shea and his team, Michigan State University, are simulating the formation and evolution of the Milky Way's most distant ancestors, a population of small galaxies formed shortly after the Big Bang. These simulations will be more accurate and full featured than any performed before. Previous simulations have modeled volumes of 1 megaparsec; the highest resolution subvolume of O'Shea's simulation will be more than 200 times this size.
Klaus Schulten and his team at the University of Illinois at Urbana-Champaign are studying the protein capsid that encases the HIV-1 genome. The process through which this capsid disassembles, releasing its genetic material, is a critical step in HIV infection. Schulten's group will simulate a cylindrical capsid consisting of 12.5 million atoms.
Robert Sugar's team, University of California, Santa Barbara, is conducting lattice quantum chromodynamics studies, which deal with sub-atomic physics. The first goal is to examine the charmonium spectrum (the bound state of charms and anti-charms); this requires challenging simulations with very small lattice spacing and very large lattice dimensions. The team's second effort deals with exotic mesons, which are prime territory in which to hunt for gluons. Sugar's team aims to use the Blue Waters Early Science System to confirm or refute that this state exists and can be identified experimentally.
Stanford Woosley's team, University of California Observatories, is researching explosive burning in Type Ia supernovae, which are used as "standard candles" for surveying astronomically vast distances. Using the Early Science System, Woosley's team will be able to achieve unprecedented resolution at the finest level of their adaptive mesh refinement simulations.
Donald Wuebbles and his team will simulate the end of both the 20th and 21st centuries at 0.25° global resolution. These high-resolution time slices will enable his team to explore changes in the frequency and intensity of extreme events, such as tropical cyclones and mid-continental thunderstorms, that are not adequately resolved in global climate models at lower resolution. By using the Early Science System, Wuebbles' team hopes to contribute high-resolution results in time for the next assessment report of the Intergovernmental Panel on Climate Change.