Browse by author
Lookup NU author(s): Professor Martin Downie, Professor Grant Hearn, Professor Chris Phillips
Full text for this publication is not currently held within this repository. Alternative links are provided below where available.
Systolic algorithms often present an attractive parallel programming paradigm. However, the unavailability of specialised hardware for efficient implementation means that such algorithms are often dismissed as being of theoretical interest only. In this paper we report on experience with implementing a systolic algorithm for matrix factorisation and present a modified version that is expected to lead to acceptable performance on a distributed memory multicomputer. The origin of the problem that generates the full complex matrix in the first place is in the field of hydrodynamics.
Author(s): Seo S-G, Downie MJ, Hearn GE, Phillips C
Editor(s): Palma, J.M.L.M., Dongarra, J., Hernandez, V
Publication type: Conference Proceedings (inc. Abstract)
Publication status: Published
Conference Name: 3rd International Conference on Vector and Parallel Processing (VECPAR)
Year of Conference: 1999
Pages: 355-364
ISSN: 0302-9743 (Print) 1611-3349 (Online)
Publisher: Springer-Verlag
URL: http://dx.doi.org/10.1007/10703040_27
DOI: 10.1007/10703040_27
Library holdings: Search Newcastle University Library for this item
Series Title: Lecture Notes in Computer Science
ISBN: 9783540662280
