Research and Analysis of Parallel Performance with MPI Odd-Even Sorting Algorithm on Super Cheap Computing Cluster

Abstract

This paper is applied an odd even transposition sorting algorithm with Raspberry Pi cluster to break down a computational problem into several parallel tasks of a distributed system. According to the test results, the algorithm can apply the distributed memory of multiple nodes to sort the large dataset with the medium performance for parallel speed computing to solve the real world complex parallel and distributed computing system. In this case, it used the MPI to establish the communication and synchronization between the processors. The experiment is tested upon ten nodes Raspberry Pi (one Raspberry Pi has four cores) with forty cores computer cluster. It also compared upon the sorting method with one to forty processes on Raspberry Pi for various dataset. This work can benefit to know the HPC using super cheap cluster computer and to support the computer science students in applying the parallel and distributed computing using MPI.