Applied Mathematics and Computation, cilt.156, sa.1, ss.189-209, 2004 (SCI-Expanded)
The article presents an experimental evaluation of the effect of deflation on accuracy. A number of test matrices were used to test Cuppen's algorithm since the amount of deflation in this method depends on the test matrix. The motivation for a non-recursive version was to compare this with the recursive version and to investigate the relationship between the recursive and non-recursive implementations of Cuppen's divide-and-conquer method. We have concentrated in particular on algorithms written in the C++ programming language. C++ is an object-oriented programming language, which can provide various types of matrix classes. The use of C++ implies some storage organisation for array elements. The primitive arrays provided in C++ use storage by rows. As far as we are aware there is little previous work on parallelising these algorithms using C++. Consequently in this work we have considered implementation of the serial using the object-oriented programming language C++. © 2003 Elsevier Inc. All rights reserved.