par Magolu, Monga Made
Référence Future generations computer systems, 11, 2, page (167-173)
Publication Publié, 1995
Article révisé par les pairs
Résumé : In this paper we are concerned with the implementation on distributed memory MIMD computers of parallel block preconditionings for solving large and sparse symmetric positive definite linear systems. For this purpose, a recently proposed ordering strategy is used. Experimental tests performed on a transputer-based multiprocessor computer are reported, displaying that good performance can be achieved in exchange for some optimization effort and on to the assumption that the number of processors is smaller by orders of magnitude than the size of the problem to solve, which agrees with observations made elsewhere. © 1995.