Toggle Main Menu Toggle Search

Open Access padlockePrints

Parallel Algorithms for Linear Algebra on a Shared Memory Multiprocessor

Lookup NU author(s): Dr Kenneth Wright

Downloads

Full text for this publication is not currently held within this repository. Alternative links are provided below where available.


Abstract

This paper describes a variety of parallel algorithms for linear algebra problems developed using shared memory Encore Multimax Multiprocessors. Algorithms using dynamic task allocation are compared with ones which do not. Problems considered include QR and LU decomposition, orthogonal reduction of General Matrices to upper Hessenberg form and symmetric matrices to tridiagonal form. The experimental results to be presented show that dynamic task allocation can be very effective on this machine, and that very high effciency is obtainable with careful construction of the parallel algorithms even for relatively small matrices.


Publication metadata

Author(s): Kaya D, Wright K

Editor(s): Bainov, D. and Covachev, V.

Publication type: Conference Proceedings (inc. Abstract)

Publication status: Published

Conference Name: 3rd International Colloquium on Numerical Analysis

Year of Conference: 1995

Pages: 209-218

Publisher: VSP, Utrecht

Library holdings: Search Newcastle University Library for this item

ISBN: 9067641936


Share