Vectorization of the general Monte Carlo classical trajectory program VENUS

Journal of Computational Chemistry - Tập 12 Số 8 - Trang 1014-1024 - 1991
Xiche Hu1, William L. Hase1, Tony Pirraglia2
1Dept. of Chemistry, Wayne State University, Detroit, Michigan 48202
2High Performance Computing Solution Development, IBM, Kingston, New York 12401

Tóm tắt

AbstractThe general chemical dynamics computer program VENUS is used to perform classical trajectory simulations for large polyatomic systems, with many atoms and complicated potential energy functions. To simulate an ensemble of many trajectories requires a large amount of CPU time. Since each trajectory is independent, it is possible to parallel process a large set of trajectories instead of processing the trajectories by the conventional sequential approach. This enhances the vectorizability of the VENUS program, since the integration of Hamilton's equations of motion and the gradient evaluation, which comprise 97.8% of the CPU, can each be parallel processed. In this article, the vectorization and ensuing optimization of VENUS on the CRAY‐YMP and IBM‐3090 are presented in terms of both global strategies and technical details. A switching algorithm is designed to enhance the vector performance and to minimize the memory storage. A performance of 140 MFLOPS and a vector/scalar execution rate ratio of 10.6 are observed when this new version of VENUS is used to study the association of CH3 with the H(Ar)12 cluster on the CRAY‐YMP.

Từ khóa


Tài liệu tham khảo

Bunker D. L., 1971, Met Comput. Phys., 10, 287

Porter R. N., 1976, Dynamics of Molecular Collisions, 1

10.1007/978-1-4613-2913-8_10

10.1007/978-1-4613-2913-8_16

Hase W. L., 1980, Aspects of the Kinetics and Dynamics of Surface Reactions, 109

Goldstein H., 1950, Classical Mechanics

Levine R. D., 1987, Molecular Reaction Dynamics and Chemical Reactivity

10.1002/9780470141205

10.1021/j100410a037

10.1021/ja00287a028

10.1002/9780470142875.ch6

10.1146/annurev.pc.36.100185.002521

X.HuandW. L.Hase in preparation.

Perera L., 1990, J. Chem. Phys., 90, 7354, 10.1063/1.456215

Wilson S., 1989, Methods in Computational Chemistry, 1

Taylor P. R., 1989, Methods in Computational Chemistry, 63

Diercksen G. H. F., 1982, Methods in Computational Molecular Physics, 335

Duff I. S., 1989, Computer Physics Reports, 11, 1, 10.1016/0167-7977(89)90016-6

Jensen E. F., 1987, Supercomputer Research in Chemistry and Chemical Engineering, 10.1021/bk-1987-0353

10.1007/BF00526412

10.1002/jcc.540070215

10.1002/jcc.540110209

10.1016/0167-7977(88)90014-7

10.1016/0167-8191(88)90006-3

W. L.Hase R. J.Duchovic X.Hu K. F.Lim D.‐H.Lu K. N.Swamy S. R.Vande Linde andR. J.Wolf VENUS to be submitted toQCPE. VENUS is an enhanced version of MERCURY. W.L. Hase MERCURY QCPE 3 453 (1983).

Designing and Writing FORTRAN Program for Vector and Parallel Processing(Publication No. SC23‐033700) IBM 1986.

IBM VS FORTRAN Version 2.4 Programming Guide(Publication No. SC26‐4222‐04) IBM 1989.

IBM VS FORTRAN Version 2.4 Language and Library Reference Manual(Publication No. SC26‐4221‐04) IBM 1989.

CFT77 Reference Manual(Publication No. SR‐0018) Cray Research Inc. Mendota Heights Minnesota 1987.

UNICOS User Commands Reference Manual(Publication No. SR‐2011) Cray Research Inc. Mendota Heights Minnesota 1988.

Cray CFT Optimization Guide On Line Documents Pittsburgh Supercomputer Center.

10.1063/1.430787

10.1021/j150651a021

10.1021/ja00244a010

10.1119/1.1934101

Gear C. W., 1971, Numerical Initial Value Problems in Ordinary Differential Equations

Lapidus L., 1971, Numerical Solution of Ordinary Differential Equations

10.1002/jcc.540070108

Street W. B., 1978, Mol. Phys., 35, 539

Course materials from IBM vendor workshop at Kingston Dallas.

We plan to ultimately make this vectorized VENUS available through QCPE. Until that is done interested parties can receive copies of the program by contacting the authors.

Thông tin
Thông tin xuất bản

Journal of Computational Chemistry

Tập 12 Số 8

1014-1024

Thông tin tác giả