Iterated Runge-Kutta methods on distributed memory multiprocessors

Author

Rauber, Thomas ; Rünger, Gudula

Author_Institution

Comput. Sci. Dept., Saarlandes Univ., Saarbrucken, Germany

fYear

1995

fDate

25-27 Jan 1995

Firstpage

12

Lastpage

19

Abstract

In this article, we consider the iterated Runge-Kutta (IRK) method which is an iteration method based on a predictor-corrector scheme for the solution of ordinary differential equations. The method uses embedded formulae to control the stepsize. We present different algorithms of the IRK method on distributed memory multiprocessors using appropriate communication primitives. The theoretical performance analysis and a runtime simulation allow us to value the presented algorithms. An implementation on the Intel iPSC/860 confirms the predicted runtimes

Keywords

Runge-Kutta methods; differential equations; distributed memory systems; iterative methods; parallel algorithms; Runge-Kutta methods; distributed memory multiprocessors; embedded formulae; iterated Runge-Kutta; iteration method; ordinary differential equations; performance analysis; predictor-corrector scheme; Analytical models; Communication system control; Computer science; Differential equations; Parallel algorithms; Partial differential equations; Performance analysis; Power engineering and energy; Predictive models; Runtime;

fLanguage

English

Publisher

ieee

Conference_Titel

Parallel and Distributed Processing, 1995. Proceedings. Euromicro Workshop on

Conference_Location

San Remo

Print_ISBN

0-8186-7031-2

Type

conf

DOI

10.1109/EMPDP.1995.389159

Filename

389159