Figuring Out Water
June / July 2010
Volume 8 Number 3
Water may be one of the most precious substances on Earth, but scientists know less than you might think about how it behaves at the molecular level. For example, how do billions of tiny aerosol particles interact when they form clouds? What molecular changes occur within large quantities of water, and how might certain polymers clean up water when it’s contaminated?
Recently, a group of Ames Laboratory researchers received the Department of Energy’s prestigious award known as the Innovative and Novel Computational Impact on Theory and Experiment, or INCITE, for their proposal to study these and other fundamental questions about the nature and behavior of water.
The Ames Lab group includes Theresa Windus, associate scientist at Ames and a professor of chemistry at Iowa State University; Mark Gordon, director of the applied mathematics and computational sciences program at Ames and Frances M. Craig, chair in chemistry at ISU; Monica Lamm, Ames associate scientist and ISU assistant professor of chemical and biological engineering; and Michael Schmidt, an Ames associate chemist.
Unlike most research grants, the INCITE award doesn’t directly involve money. Instead, in another example of how DOE research centers collaborate, the Ames team was granted 8 million processor hours on an IBM Blue Gene supercomputer—one of the world’s fastest—located at Argonne National Laboratory in Illinois. To understand just how much processing power that is, think of a city with 8 million people, in which everyone uses their personal computers to work on the same problem for one hour.
In fact, the award received by the team is a fraction of the 1.6 billion hours of supercomputing time granted by the DOE to 69 research teams located throughout the country as part of the highly competitive INCITE program.
Coveted time on the Blue Gene supercomputer will enable the team to accomplish in weeks or months what would have taken years or decades to complete in the past. To perform their work, the group will rely on a key tool: a suite of supercomputer-enabled software programs known as the General Atomic and Molecular Electronic Structure System, or GAMESS. Mark Gordon is the lead developer of GAMESS code and has devoted much of his career to its refinement.
The group’s work will generate terabytes worth of data, a common result with research involving supercomputers. “The more large-scale computing you do, the more you generate large amounts of data. So the issue arises, how do you best manage that data?”says Gordon.
The data will prove invaluable to the researchers and other scientists probing the dynamics of water. Indeed, the questions the group hopes their data will address touch on vital areas such as climate change and scarce resources. Take cloud formation, for example. As Theresa Windus observes, “The composition of the aerosols determines if a cloud is made up of a lot of small particles or large particles, which in turn impacts the size of clouds, their longevity and the probability they will produce rain.”
Similarly, Monica Lamm, who will study ways of purifying water, hopes her work will tackle an equally pressing global problem. “Clean water is going to be a huge issue as the world’s water resources become increasingly stretched in the years ahead,” she says.
It’s ironic, perhaps, that one of the planet’s most advanced computers will be used to study one of earth’s simplest substances. Yet for the scientists involved, the thrill of discovery remains the same. “The opportunity to have your science run on these machines is very exciting,” Lamm says, adding, “and the fact that we don’t know what the results will tell us is also exciting.”
With more than 163,000 processors and a peak performance of 557 teraflops (trillion floating point operations per second), the Blue Gene is an important collaborative tool for scientists working at DOE research facilities, nationwide.
Mark Ingebretsen is a communications specialist at Ames Laboratory. This article originally appeared in Inquiry, which highlights Ames research.
Copyright © 2012 | Innovation America