Parallel Unrestricted MP2 Analytic Gradients Using the Distributed Data Interface

A scalable distributed-data parallel analytic gradient algorithm for unrestricted second-order Møller−Plesset perturbation theory is presented. Features of the implementation using the Distributed Data Interface are discussed in detail. Benchmark timing calculations on a parallel cluster system are presented for a variety of gold cluster molecules. The speedups, parallel efficiencies, and percentage parallelism for these calculations are reported. Disciplines Chemistry Comments Reprinted (adapted) with permission from Journal of Physical Chemistry A 108 (2004): 3103, doi:10.1021/ jp031142t. Copyright 2004 American Chemical Society. This article is available at Iowa State University Digital Repository: http://lib.dr.iastate.edu/chem_pubs/448 Parallel Unrestricted MP2 Analytic Gradients Using the Distributed Data Interface† Christine M. Aikens and Mark S. Gordon* Department of Chemistry, Iowa State UniVersity, Ames, Iowa 50011 ReceiVed: October 8, 2003; In Final Form: December 23, 2003 A scalable distributed-data parallel analytic gradient algorithm for unrestricted second-order Møller-Plesset perturbation theory is presented. Features of the implementation using the Distributed Data Interface are discussed in detail. Benchmark timing calculations on a parallel cluster system are presented for a variety of gold cluster molecules. The speedups, parallel efficiencies, and percentage parallelism for these calculations are reported.