AL-Rafidain Journal of Computer Sciences and Mathematics

Abstract

The aim of the project is to develop parallel approaches for Gaussian Elimination Methods that are used in linear programming  to solve linear module  systems.
            Most of these models are time-consuming when executed and processed in the sequential microprocessor computers. During the project, we try to decrease this time and increase the efficiency of the algorithm for the Gaussian Elimination Method, through developing parallel methods appropriate to be executed on MIMD type computers.
            In this paper, three algorithms were suggested for paralleling a developed algorithm of Gaussian Elimination Method and a comparison was made between the three algorithms and the original.
            As we have been able to accelerate the three parallel methods and the speedup was one of the following:
          Speedup =  ,  no. of processor is (50)
            In general, the practical results and the suggested programs for these new algorithms proved to be better in performance than their analogues that are executed in computers of sequential processor in view of the two elements of execution time and algorithm time.