skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Workshop report on large-scale matrix diagonalization methods in chemistry theory institute

Abstract

The Large-Scale Matrix Diagonalization Methods in Chemistry theory institute brought together 41 computational chemists and numerical analysts. The goal was to understand the needs of the computational chemistry community in problems that utilize matrix diagonalization techniques. This was accomplished by reviewing the current state of the art and looking toward future directions in matrix diagonalization techniques. This institute occurred about 20 years after a related meeting of similar size. During those 20 years the Davidson method continued to dominate the problem of finding a few extremal eigenvalues for many computational chemistry problems. Work on non-diagonally dominant and non-Hermitian problems as well as parallel computing has also brought new methods to bear. The changes and similarities in problems and methods over the past two decades offered an interesting viewpoint for the success in this area. One important area covered by the talks was overviews of the source and nature of the chemistry problems. The numerical analysts were uniformly grateful for the efforts to convey a better understanding of the problems and issues faced in computational chemistry. An important outcome was an understanding of the wide range of eigenproblems encountered in computational chemistry. The workshop covered problems involving self- consistent-field (SCF), configurationmore » interaction (CI), intramolecular vibrational relaxation (IVR), and scattering problems. In atomic structure calculations using the Hartree-Fock method (SCF), the symmetric matrices can range from order hundreds to thousands. These matrices often include large clusters of eigenvalues which can be as much as 25% of the spectrum. However, if Cl methods are also used, the matrix size can be between 10{sup 4} and 10{sup 9} where only one or a few extremal eigenvalues and eigenvectors are needed. Working with very large matrices has lead to the development of« less

Authors:
; ;  [1]
  1. eds.
Publication Date:
Research Org.:
Argonne National Lab., IL (United States)
Sponsoring Org.:
USDOE Office of Energy Research, Washington, DC (United States)
OSTI Identifier:
465681
Report Number(s):
ANL/MCS-TM-219; CONF-9605268-Absts.
ON: DE97004423
DOE Contract Number:  
W-31109-ENG-38
Resource Type:
Conference
Resource Relation:
Conference: Large-scale matrix diagonalization methods in chemistry theory institute, Argonne, IL (United States), 20-22 May 1996; Other Information: PBD: Oct 1996
Country of Publication:
United States
Language:
English
Subject:
40 CHEMISTRY; 99 MATHEMATICS, COMPUTERS, INFORMATION SCIENCE, MANAGEMENT, LAW, MISCELLANEOUS; 66 PHYSICS; CHEMISTRY; NUMERICAL SOLUTION; MATRIX ELEMENTS; MATRICES; MEETINGS; ITERATIVE METHODS; EIGENVALUES; MOLECULES; EIGENSTATES; PARALLEL PROCESSING; QUANTUM MECHANICS; VIBRATIONAL STATES

Citation Formats

Bischof, C.H., Shepard, R.L., and Huss-Lederman, S. Workshop report on large-scale matrix diagonalization methods in chemistry theory institute. United States: N. p., 1996. Web.
Bischof, C.H., Shepard, R.L., & Huss-Lederman, S. Workshop report on large-scale matrix diagonalization methods in chemistry theory institute. United States.
Bischof, C.H., Shepard, R.L., and Huss-Lederman, S. Tue . "Workshop report on large-scale matrix diagonalization methods in chemistry theory institute". United States. doi:. https://www.osti.gov/servlets/purl/465681.
@article{osti_465681,
title = {Workshop report on large-scale matrix diagonalization methods in chemistry theory institute},
author = {Bischof, C.H. and Shepard, R.L. and Huss-Lederman, S.},
abstractNote = {The Large-Scale Matrix Diagonalization Methods in Chemistry theory institute brought together 41 computational chemists and numerical analysts. The goal was to understand the needs of the computational chemistry community in problems that utilize matrix diagonalization techniques. This was accomplished by reviewing the current state of the art and looking toward future directions in matrix diagonalization techniques. This institute occurred about 20 years after a related meeting of similar size. During those 20 years the Davidson method continued to dominate the problem of finding a few extremal eigenvalues for many computational chemistry problems. Work on non-diagonally dominant and non-Hermitian problems as well as parallel computing has also brought new methods to bear. The changes and similarities in problems and methods over the past two decades offered an interesting viewpoint for the success in this area. One important area covered by the talks was overviews of the source and nature of the chemistry problems. The numerical analysts were uniformly grateful for the efforts to convey a better understanding of the problems and issues faced in computational chemistry. An important outcome was an understanding of the wide range of eigenproblems encountered in computational chemistry. The workshop covered problems involving self- consistent-field (SCF), configuration interaction (CI), intramolecular vibrational relaxation (IVR), and scattering problems. In atomic structure calculations using the Hartree-Fock method (SCF), the symmetric matrices can range from order hundreds to thousands. These matrices often include large clusters of eigenvalues which can be as much as 25% of the spectrum. However, if Cl methods are also used, the matrix size can be between 10{sup 4} and 10{sup 9} where only one or a few extremal eigenvalues and eigenvectors are needed. Working with very large matrices has lead to the development of},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Tue Oct 01 00:00:00 EDT 1996},
month = {Tue Oct 01 00:00:00 EDT 1996}
}

Conference:
Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share: