In this paper, we present the developments realized in the OURAGAN project
around the parallelization of a MATLAB-like tool called SCILAB. These devel
opments use high-performance numerical libraries and different approaches b
ased either on the duplication Of SCI LAB processes or on computational ser
vers. This tool, SCILAB(//), allows users to perform high-level operations
on distributed matrices in a metacomputing environment. We also present per
formance results on different architectures. (C) 2001 Elsevier Science B.V.
All rights reserved.