ºÙºÙÊÓƵ

The U.S. Department of Energy has awarded two grants, worth $1.6 million and $1.2 million per year for five years, to projects in advanced computing led by researchers at ºÙºÙÊÓƵ. Another four grants, from a total of 30 funded through the department's (Scientific Discovery through Advanced Computing) program, include ºÙºÙÊÓƵ scientists as team members. Altogether, grants awarded to or including ºÙºÙÊÓƵ total about $12 million per year, divided among all participating institutions.

"ºÙºÙÊÓƵ has a huge footprint in this program. In computational science, this campus is really on the map," said Zhaojun Bai, professor and chair of computer science at ºÙºÙÊÓƵ. Bai is also a co-investigator on two of the projects. The success reflects the interdisciplinary nature of the campus, he said.

Projects included in the program, funded with up to $300 million over the next five years, are aimed at understanding fundamental physics, studying global climate change, designing new materials and developing new energy sources. Some are intended to develop technology for "petascale computing," using computers that work in the range of a thousand million million operations per second.

The range of grants awarded to ºÙºÙÊÓƵ researchers includes all the main areas of the SciDAC program, from basic science discovery to developing computational tools, Bai said. That reflects the environment at ºÙºÙÊÓƵ, where computational science is becoming an important tool for researchers across many departments. It also represents rapid progress since the first round of SciDAC grants, awarded in 2001, in which ºÙºÙÊÓƵ was only lightly represented. These grants are also highly competitive, he said.

The two ºÙºÙÊÓƵ-led grants were awarded to Kwan-Liu Ma, professor in the Department of Computer Science and to Giulia Galli, professor in the Department of Chemistry.

Ma's group will receive up to $1.6 million per year to establish an . The institute will develop tools to handle and study very large amounts of data created by supercomputers. It will also work to educate scientists about these tools, through conferences, summer schools and outreach programs. Co-investigators at ºÙºÙÊÓƵ are Nelson Max, professor of computer science, and John Owens, assistant professor of electrical and computer engineering. Other participating institutions are the Argonne and Sandia National Laboratories, The Ohio State University, the University of Tennessee at Knoxville, and the University of Virginia.

Galli's project, "," will study ways to simulate the behavior of atoms using the fundamental laws of quantum mechanics and working up to simulate materials and chemical reactions. The methods developed will eventually help researchers to better understand how materials behave under different conditions, and to develop new kinds of materials. The grant is for $1.2 million per year over five years. Partners include: Bai; Francois Gygi, professor of applied science; Warren Pickett, professor of physics at ºÙºÙÊÓƵ; and colleagues at Lawrence Livermore National Laboratory, MIT, Stanford University, UC Santa Barbara and the University of Illinois at Urbana-Champaign.

The other projects involving ºÙºÙÊÓƵ faculty are: "Modeling Materials at the Petascale," led by the University of Cincinnati, including Bai, Sergey Savrasov, associate professor of physics and Richard Scalettar, professor of physics; "Seeing the Unseeable," led by Lawrence Berkeley National Laboratory, including Ken Joy, professor of computer science and co-director of IDAV, and Bernd Hamann, professor of computer science and associate vice chancellor in the Office of Research; "Advanced Science via Applied Mathematics," led by the Lawrence Berkeley National Laboratory and including Gregory Miller, professor of applied science; and "Getting the Science out of the Data," led by the Lawrence Berkeley National Laboratory and including Bertram Ludaescher, associate professor in the Department of Computer Science and at the Genome Center.

The grants were announced Sept. 7 by the U.S. Department of Energy. The program is intended to help researchers at universities and the national laboratories to create software and infrastructure to use the next generation of supercomputers.

Scientists are beginning to think about how to use "petascale" computers when they become available, Bai said. Last year, the National Science Foundation issued a call for proposal to develop petascale computing. Currently, the world's fastest computer is BlueGene/L, installed at the Lawrence Livermore National Laboratory, which has been clocked at 280 trillion operations per second. The computer code that achieved that record, called , was written by Gygi while he was a researcher at the Livermore lab. A petascale computer would run about 10 times faster than BlueGene/L.