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 language||English (US)|
|Title of host publication||Parallel Computation - 4th International ACPC Conference Including Special Tracks on Parallel Numerics (ParNum 1999) and Parallel Computing in Image Processing, Video Processing, and Multimedia, Proceedings|
|Editors||Peter Zinterhof, Marian Vajteršic, Andreas Uhl|
|Number of pages||15|
|ISBN (Print)||3540656413, 9783540656418|
|State||Published - 1999|
|Event||4th International ACPC Conference on Parallel Computation, ACPC 1999 - Salzburg, Austria|
Duration: Feb 16 1999 → Feb 18 1999
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Other||4th International ACPC Conference on Parallel Computation, ACPC 1999|
|Period||2/16/99 → 2/18/99|
Bibliographical notePublisher Copyright:
© Springer-Verlag Berlin Heidelberg 1999.