For release: Dec. 19, 2002
Contacts:
Mark Gordon, Applied Mathematics and Computational Sciences, gordon@ameslab.gov
Michael Schmidt, Applied Mathematics and Computational Sciences, mike@si.fi.ameslab.gov
Brett Bode, Applied Mathematics and Computational Sciences, brett@ameslab.gov
Saren
Johnston, Public Affairs, (515) 294-34740
AMES, Iowa – Researchers at the U. S. Department of Energy’s Ames Laboratory at Iowa State University are fostering and expanding a computational chemistry code that provides extensive and detailed information about how things work on the molecular scale. The General Atomic and Molecular Electronic Structure System, known as GAMESS, includes a hierarchy of quantum chemistry methods that helps solve problems relating to molecules, making possible the design of new fuels and optical materials and the development of coatings that are resistant to extreme environments.
“All of chemistry, which also means all of biology, fundamentally involves molecular processes – that is, the way in which molecules react or behave,” said Mark Gordon, director of Ames Lab’s Applied Mathematics and Computational Sciences Program and an ISU Distinguished Professor of Liberal Arts and Sciences. “Chemistry, biochemistry, biology, materials science, physics – all ultimately can be reduced to what molecules do and why they do it. GAMESS allows you to answer those questions.”
Using GAMESS, Gordon, his co-workers and the students in his research group are making major contributions to the design of new rocket fuels for the Air Force. “We’re also involved in a Grand Challenge program for the Department of Defense, using GAMESS to help design new optical materials, fuels and wear-resistant coatings,” said Gordon. “And some of my former students are now using GAMESS to study biological processes, such as protein behavior.”
The history of GAMESS goes back to 1977 when the initial version of the code was assembled, under the direction of Dr. Michel Dupuis, from several existing quantum chemistry computer programs by the staff at the National Resource for Computations in Chemistry. The original project was funded by the National Science Foundation and the Department of Energy from 1977 to 1981. However, the continued development of GAMESS from 1982 to the present is directly due to the funding provided by the Air Force Office of Scientific Research; to the nurturing environment provided by Gordon; and to the expertise demonstrated by Gordon, Ames Lab associate scientist Michael Schmidt and many students and postdocs in developing new functionalities and parallelizing the code. Making a code operate efficiently in parallel refers to organizing it so that it can take advantage of parallel computers – those with several processors that can work on different parts of a single problem simultaneously. In addition, contributions to the GAMESS program from Gordon’s many colleagues and from students who have been in his research group have added tremendously to the current robust nature of the code.
“I’ve always had at least two or three students and postdocs specifically working on how to make GAMESS more parallel,” said Gordon. A major result of this ongoing effort is that an increasing percentage of the features in GAMESS, especially the most computationally demanding ones, can be done in parallel. “Developing parallel code is something we do better than most,” he said.
In addition to excelling at writing parallel code for GAMESS, Gordon and his Ames Lab collaborators, Schmidt and Klaus Ruedenberg, who is also an ISU Distinguished Professor of Liberal Arts and Sciences, are particularly adept at creating sophisticated and complex methods for the software suite that address unusual chemical species. “When you have chemical reactions going on, especially under severe conditions, such as high temperature or high pressure, you encounter chemical species that you would not even call molecules,” Gordon explained. “These are things that don’t hang around very long, and you may never see them. They’re intermediates and may have lifetimes of only picoseconds or femtoseconds, but they may be very important in the overall reaction process. One of our goals is to develop methods to treat these very unusual species, and the more nontraditional the method, the harder it is to make parallel.”
An innovative research tool, GAMESS includes a novel graphics visualization program, MacMolPlt, written for Macintosh computers by Ames Lab associate scientist, Brett Bode. Gordon said MacMolPlt eases the task of interpreting the various complicated calculations performed by GAMESS. Ryan Olson, a graduate student in Gordon’s group, is developing WinMolPlt, a Windows version of MacMolPlt. “Ryan has half a dozen people around the world testing this program, so now you can visualize the results of GAMESS calculations if you have a Mac or a PC,” said Gordon.
A clearly unique feature of GAMESS is the effective fragment potential, or EFP, which was initially developed by Gordon and Jan Jensen, a former graduate student in Gordon’s group now at the University of Iowa, in collaboration with Dr. Walter Stevens and his group at the National Institute of Standards and Technology. Stevens is now Director of the DOE Chemical Sciences Program. Over the past decade, several graduate students and postdocs have contributed to the development of the EFP method, including current students Heather Netzloff and Ivana Adamovic and postdoc Jie Song.
“Sometimes scientists deal with systems in which there are so many atoms and electrons that quantum mechanics can’t be done, even at simple levels,” said Gordon. The EFP method is based on quantum mechanics, but it’s not quantum mechanics – it’s a simple but sophisticated model potential that represents most of the quantum mechanics effects at a very reduced computational effort. The EFP method can be combined with actual quantum mechanics in such a way that a quantum mechanical description is required only in that part of a chemical system that is undergoing a chemical change. The remainder of the system is treated with EFPs, so that the entire calculation takes orders of magnitude less computer time than a fully quantum calculation. “The EFP is a sophisticated model to predict how solvents effect chemical reactions and to predict the behavior of liquids,” said Gordon, “and it’s working very well.”
GAMESS is used at well over 5,000 sites worldwide, ranging from high schools to research universities in the United States and abroad to government laboratories to the private sector. It is distributed at no cost to users by accessing www.msg.ameslab.gov and signing a license agreement.
The GAMESS research is funded by grants from the Air Force Office of Scientific Research, by the Department of Defense and by a Department of Energy SciDAC, or Scientific Discovery through Advanced Computing, grant to Gordon, Ruedenberg, Schmidt and Jim Evans, an Ames Laboratory scientist and an ISU professor of mathematics. Ames Laboratory is operated for the DOE by Iowa State University. The Lab conducts research into various areas of national concern, including energy resources, high-speed computer design, environmental cleanup and restoration, and the synthesis and study of new materials.
###
Last revision: 12/20/02