Computational & Technology Resources
an online resource for computational,
engineering & technology publications
Civil-Comp Proceedings
ISSN 1759-3433
CCP: 52
ADVANCES IN COMPUTATIONAL MECHANICS WITH HIGH PERFORMANCE COMPUTING
Edited by: B.H.V. Topping
Paper III.3

Direct Domain Decomposition Method on a Distributed Memory Architecture

J.F. Lassignardie, Y. Escaig and G. Touzot

INSA de Rouen, St Etienne du Rouvray, France

Full Bibliographic Reference for this paper
J.F. Lassignardie, Y. Escaig, G. Touzot, "Direct Domain Decomposition Method on a Distributed Memory Architecture", in B.H.V. Topping, (Editor), "Advances in Computational Mechanics with High Performance Computing", Civil-Comp Press, Edinburgh, UK, pp 65-72, 1998. doi:10.4203/ccp.52.3.3
Abstract
This article presents different issues of the implementation of a domain decomposition method on distributed memory systems. The proposed domain decomposition method uses direct solvers for the local problems and for the interface problem. This method, which is sequentially as efficient as other direct solvers (like sparse solvers) shows an inherent parallelism during the resolution of the local problems. Besides, the interface problem, of which the matrix is stored in a skyline format, can be solved in parallel. The parallel performances of the method, implemented on a Cray T3E, are analysed and compared with a algebraic parallelisation of a finite element problem. The method appears to be efficient and better than a traditionnal parallelisation of a finite element problem and also offers a better scalability if the size of the problem increases.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £56 +P&P)