A fast solver for convection diffusion equations based on nested dissection with incomplete elimination

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

We present an approach for the efficient parallel solution of convection diffusion equations. Based on iterative nested dissection techniques [1] we extended these existing iterative algorithms to a solver based on nested dissection with incomplete elimination of the unknowns. Our elimination strategy is derived from physical properties of the convection diffusion equation, but is independent of the actual discretized operator. The resulting algorithm has a memory requirement that grows linearly with the number of unknowns. This also holds for the computational cost of the setup of the nested dissection structure and the individual relaxation cycles. We present numerical examples that indicate that the number of iterations needed to solve a convection diffusion equation grows only slightly with the number of unknowns, but is widely independent of the type and strength of the convection field.

Original languageEnglish
Title of host publicationEuro-Par 2000 Parallel Processing - 6th International Euro-Par Conference, Proceedings
EditorsArndt Bode, Thomas Ludwig, Wolfgang Karl, Roland Wismüller
PublisherSpringer Verlag
Pages795-805
Number of pages11
ISBN (Electronic)9783540679561
DOIs
StatePublished - 2000
Event6th International European Conference on Parallel Computing, Euro-Par 2000 - Munich, Germany
Duration: 29 Aug 20001 Sep 2000

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1900
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference6th International European Conference on Parallel Computing, Euro-Par 2000
Country/TerritoryGermany
CityMunich
Period29/08/001/09/00

Fingerprint

Dive into the research topics of 'A fast solver for convection diffusion equations based on nested dissection with incomplete elimination'. Together they form a unique fingerprint.

Cite this