BLOCK-JACOBI SVD ALGORITHMS FOR DISTRIBUTED MEMORY SYSTEMS II: MESHES∗ (Q3146522)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 1799926
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | BLOCK-JACOBI SVD ALGORITHMS FOR DISTRIBUTED MEMORY SYSTEMS II: MESHES∗ |
scientific article; zbMATH DE number 1799926 |
Statements
BLOCK-JACOBI SVD ALGORITHMS FOR DISTRIBUTED MEMORY SYSTEMS II: MESHES∗ (English)
0 references
12 September 2002
0 references
hypercube
0 references
parallelization
0 references
Jacobi algorithm
0 references
singular value decomposition
0 references
orderings
0 references
complexity
0 references
numerical examples
0 references
0.9378231167793274
0 references
0.8777188062667847
0 references
0.8613801002502441
0 references
This paper deals with the parallelization of the two-sided Jacobi algorithm for computation of Singular Value Decomposition (SVD) on a computer with p processors, which are organized into a two-dimensonal mesh configuration. This work is a continuation of part I [ibid. 13, No. 3, 265--287 (1999; Zbl 0926.65044)] which describes a parallelization approach by columns and efficient ordering strategies for the hypercube and ring topologies. The parallelization is based on slicing the matrices by rows and columns. Orderings developed for rings and hypercubes are used and a proper assignment of submatrices to processors is given which enables their efficient parallel execution. A complexity comparison to the column-based algorithm is given. Parallel computational experiments are presented and discussed for two test matrices.
0 references