CD var/Includes CD:A comparative study of sequential and parallel sorting algorithms

Includes references(81-89 p.)

'ABSTRACT Steady growth in technology has conveyed significant change in the designs of processors which has now made it possible for manufacturers to design multiple processors on a single chip. Presently, Supercomputers comprises group of interconnected nodes that are designed to work together to solve complex and advanced computational problems. One of the most popular programming model used in optimizing sequential codes is the Message Passing Interface, and it does this by parallelizing the codes on multiprocessor architectures. In this study, parallel sorting algorithms are compared to their sequential counterparts, using C++ programming language. The parallel sorting algorithms are implemented using MPI library on a distributed architecture based on windows platform. From the experiments carried out it can be seen that the parallel merge sort algorithm has higher performance than the parallel quick sort algorithm. For 4000 size of array, quick sort has speedup 1.6 in parallel with two PCs ,whereas merge sort has speedup 2.42 compared to their sequential algorithms. Key words: Parallel Computing, Message Passing Interface, Sorting Algorithms.'