Comparing MPI and OpenMP implementations of the 0-1 Knapsack Problem

Abstract

This paper describes and compares two parallel implementations of the 0-1 Knapsack Problem. We have used the C++ programming language to develop two exact algorithms based on the Branch-and-Bound technique and implemented them in supercomputers and in a PC cluster. MPI has been used in the implementation of the Message Passing code and OpenMP was chosen for the Shared Memory code. Computational results on a Sunfire 6800 SMP, an Origin 3000 and a PC cluster are presented.