Performance Evaluation of Parallel Message Passing and Thread Programming Model on Multicore Architectures

TL;DR

This paper compares the performance of MPI-2 and OpenMP on multicore shared memory systems using a custom matrix multiplication benchmark to evaluate speed and throughput differences.

Contribution

It provides a comparative analysis of MPI-2 and OpenMP performance on multicore architectures using a novel small-scale benchmark.

Findings

MPI-2 and OpenMP show different scalability behaviors.

OpenMP outperforms MPI-2 in certain shared memory scenarios.

The custom benchmark highlights specific strengths of each model.

Abstract

The current trend of multicore architectures on shared memory systems underscores the need of parallelism. While there are some programming model to express parallelism, thread programming model has become a standard to support these system such as OpenMP, and POSIX threads. MPI (Message Passing Interface) which remains the dominant model used in high-performance computing today faces this challenge. Previous version of MPI which is MPI-1 has no shared memory concept, and Current MPI version 2 which is MPI-2 has a limited support for shared memory systems. In this research, MPI-2 version of MPI will be compared with OpenMP to see how well does MPI perform on multicore / SMP (Symmetric Multiprocessor) machines. Comparison between OpenMP for thread programming model and MPI for message passing programming model will be conducted on multicore shared memory machine architectures to see who has a better performance in terms of speed and throughput. Application used to assess the scalability of the evaluated parallel programming solutions is matrix multiplication with customizable matrix dimension. Many research done on a large scale parallel computing which using high scale benchmark such as NSA Parallel Benchmark (NPB) for their testing standarization [1]. This research will be conducted on a small scale parallel computing that emphasize more on the performance evaluation between MPI and OpenMPI parallel programming model using self created benchmark.