A parallel adaptive version of the block-based Gauss-Jordan algorithm

Author

Melab, N. ; Talbi, E.-G. ; Petiton, S.

Author_Institution

Lab. d´´Inf. Fondamentale de Lille, Univ. des Sci. et Tech. de Lille Flandres Artois, Villeneuve d´´Ascq, France

fYear

1999

fDate

12-16 Apr 1999

Firstpage

350

Lastpage

354

Abstract

This paper presents a parallel adaptive version of the block-based Gauss-Jordan algorithm used in numerical analysis to invert matrices. This version includes a characterization of the workload of processors and a mechanism of its adaptive folding/unfolding. The application is implemented and experimented with MARS in dedicated and non-dedicated environments. The results show that an absolute efficiency of 92% is possible on a cluster of DEC/ALPHA processors interconnected by a Gigaswitch network and an absolute efficiency of 67% can be obtained on an Ethernet network of SUN-Sparc4 workstations. Moreover the adaptability of the algorithm is experimented on a non-dedicated meta-system including both the two parks of machines

Keywords

parallel algorithms; workstation clusters; Ethernet network; Gigaswitch network; MARS; SUN-Sparc4 workstations; adaptive folding; block-based Gauss-Jordan algorithm; cluster of DEC/ALPHA processors; numerical analysis; parallel adaptive version; Fault tolerance; Gaussian processes; Ice; LAN interconnection; Mars; Network topology; Numerical analysis; Parallel programming; Programming environments; Workstations;

fLanguage

English

Publisher

ieee

Conference_Titel

Parallel Processing, 1999. 13th International and 10th Symposium on Parallel and Distributed Processing, 1999. 1999 IPPS/SPDP. Proceedings

Conference_Location

San Juan

Print_ISBN

0-7695-0143-5

Type

conf

DOI

10.1109/IPPS.1999.760499

Filename

760499