Implementation of Parallel Algorithms on Cluster of Workstations

Most of the DNA sequence comparison techniques are based on heuristic as it is the fastest even though it does not guarantees that the optimum results will be produced. In DNA sequence comparison - a basic operation in sequencing projects most of the methods uses heuristic approach. The algorithm proposed by Needleman and Wunsch obtains the best global alignments at the expense of very high computing power and huge memory requirements. In this paper we have presented and evaluated our experiments with two different methods to run the Needleman and Wunsch algorithm on a cluster of workstations using a Distributed Shared Memory System. Our results on cluster of workstations which consist of total 24 cores presented very good speedups and shows that notable improvements can be achieved, depending on the methods used. Also, we present some theoretical remarks on how to reduce the amount of memory used.