PARALLEL COMPUTATION DESIGN OF DETERMINANT AND INVERSE OF A MATRICE USING COMPUTATIONAL GEOMETRY ANALYSIS TECHNIQUE

Abstract

The aim of this research is to study the use of computational geometry analysis technique in detail to design a parallel computation to determine determinant and inverse of a matrice using Gauss-Jordan method. In comparing with dependence graph technique, it is well known that computational geometry analysis gives us an  easier way in determining determinant and inverse of a matrice, moreover for handling a computation domain with dimension of higher than 3. The type of scheduling function used in this research is of linear type. Some important facts are discovered in this research, the most important one is, this technique gives us the ease in arranging data scheduling scheme based on the observation of the computation domain. The other important fact is, it is also give us the ease to understand the scheme of the data scheduling since this technique can be collaborated with dependence graph technique for visualizing the data scheduling