Toward global and grid computing for large scale linear algebra problems
Laurent Choy, Serge Petiton
IEEE International Conference on Cluster
Computing (Cluster 2005) Boston, Massachusetts, USA, September 27 -
30, 2005
Abstract
On the one hand, although platforms of grid and global computing have been developed independently, they can be considered as complementary tools. On the other hand, linear algebra problems have seldom been studied for global computing although many applications use them.
In this paper, we focus on the eigenproblem with the bisection algorithm. We fit this method on a platform of grid computing, OmniRPC, and on a platform of global computing, XtremWeb, which are deployed on two different geographic sites. The combination of many software and many sites allows to do a wide range of tests and, then, to analyze their results.