Non-standard parallel solution strategies for distributed sparse linear systems

Yousef Saad, Maria Sosonkina

Research output: Chapter in Book/Report/Conference proceedingConference contribution

10 Scopus citations

Abstract

A number of techniques are described for solving sparse linear systems on parallel platforms. The general approach used is a domai n-decomposition type method in which a processor is assigned a certain number of rows of the linear system to be solved. Strategies that are discussed include non-standard graph partitioners, and a forced loadbalance technique for the local iterations. A common practice when partitioning a graph is to seek to minimize the number of cut-edges and to have an equal number of equations per processor. It is shown that partitioners that take into account the values of the matrix entries may be more effective.

Original languageEnglish (US)
Title of host publicationParallel Computation - 4th International ACPC Conference Including Special Tracks on Parallel Numerics (ParNum 1999) and Parallel Computing in Image Processing, Video Processing, and Multimedia, Proceedings
EditorsPeter Zinterhof, Marian Vajteršic, Andreas Uhl
PublisherSpringer Verlag
Pages13-27
Number of pages15
ISBN (Print)3540656413, 9783540656418
DOIs
StatePublished - 1999
Event4th International ACPC Conference on Parallel Computation, ACPC 1999 - Salzburg, Austria
Duration: Feb 16 1999Feb 18 1999

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1557
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other4th International ACPC Conference on Parallel Computation, ACPC 1999
CountryAustria
CitySalzburg
Period2/16/992/18/99

Fingerprint Dive into the research topics of 'Non-standard parallel solution strategies for distributed sparse linear systems'. Together they form a unique fingerprint.

Cite this