ORTHOGONAL REDUCTION OF DENSE MATRICES TO BIDIAGONAL FORM ON COMPUTERS WITH DISTRIBUTED-MEMORY ARCHITECTURES

Authors
KUZNETSOV SV
Citation
Sv. Kuznetsov, ORTHOGONAL REDUCTION OF DENSE MATRICES TO BIDIAGONAL FORM ON COMPUTERS WITH DISTRIBUTED-MEMORY ARCHITECTURES, Parallel computing, 24(2), 1998, pp. 305-313
Citations number
10
Categorie Soggetti
Computer Science Theory & Methods","Computer Science Theory & Methods
Journal title
Parallel computing → ACNP
ISSN journal
01678191
Volume
24
Issue
2
Year of publication
1998
Pages
305 - 313
Database
ISI
SICI code
0167-8191(1998)24:2<305:ORODMT>2.0.ZU;2-V
Abstract
In this paper, we describe a parallel implementation for the blocked r eduction of dense matrices to bidiagonal form by Householder transform ations, The method is based on the idea of using un-normed reflector v ector. Results from experiments on the Intel Paragon are given. Compar ison between the ScaLapack bidiagonalization and the method proposed i s provided. (C) 1998 Elsevier Science B.V. All rights reserved.