Chemistry Computing at the Petascale

Radically advanced methods for solving the quantum electronic structure of atoms, molecules, and nanoscale systems for petascale computational chemistry

George I Fann
Oak Ridge National Laboratory

This is a tightly coupled project with the “out-of-phase” SciDAC Basic Energy Sciences Project, “Advanced Methods for Electronic Structure,” (PI: Robert J. Harrison).
This project will develop and implement advanced mathematical methods and software in a continuing collaboration with computational chemists that will lead to radical advances in the capabilities of quantum chemical methods to describe efficiently and with controlled precision the electronic structure of atoms, molecules and nano-chemical systems. Specific activities include the development of advanced mathematics and computational methods/algorithms for quantum chemistry problems in:

  • eliminating basis set error and reducing the scaling of one- and many-body methods using techniques from local correction, explicitly correlated wave functions and multiresolution analysis;
  • computing efficiently accurate energy differences for large systems using local correction methods and multiresolution Green’s function techniques; and
  • reducing the complexity and increasing the performance of software for present and anticipated high-performance computers.
Basic research into theory, algorithms, and numerical methods will be pursued in close collaboration with chemists, mathematicians, and computer scientists. Software development will emphasize novel, production-ready capabilities for petaflop computers with the massively parallel computational chemistry code NWChem and other computational chemistry packages and in the mathematical software framework Multiresolution ADaptive NumErical Scientific Simulation (MADNESS).

Advances in materials and chemistry are often critical to progress in all three mission areas. For example, chemistry research leads to the development of such advances as efficient combustion systems with reduced emissions of pollutants; new solar photoconversion processes; improved catalysts for clean and efficient production of fuels and chemicals; and better separations and analytical methods for applications in energy processes, environmental remediation, and waste management. Computational Chemistry has become a vital element in the chemist’s toolbox, an equal partner with spectroscopy and other experimental analysis tools.
To be relevant to chemical research or design, simulations must reliably meet specific criteria for both the precision of predictions, reliability, and their speed of execution. This project is expected to advance each of those areas.

Science Application: Materials Science and Chemistry

Project Title: Advanced Mathematics for Electronic Structure

Principal Investigator: George I Fann
Affiliation: Oak Ridge National Laboratory

Participating Institutions and Co-Investigators:
Oak Ridge National Laboratory - Robert Harrison

Funding Partners: U.S. Department of Energy - Office of Science, Advanced Scientific Computing Research.

Budget and Duration: Approximately $0.3 million per year for three years 1

Other SciDAC Materials Science & Chemistry efforts



1Subject to acceptable progress review and the availability of appropriated funds

 


Home  |  ASCR  |  Contact Us  |  DOE disclaimer