Implementation of High Performance Computing (HPC) Using Computer Clusters

Abstract

The high performance computing systems or supercomputers nowadays are highly demanded especially for applications that requires high computation power in data processing such as problems involving quantum mechanical physics, financial analysis, weather forecasting, climate research, molecular modeling (computing the structures and properties of chemical compounds, biological macromolecules, polymers, and crystals), physical simulations (such as simulation of airplanes in wind tunnels and research into nuclear fusion), signal and image processing, cryptanalysis, and so on. But these kinds of machines is very costly and can’t be available for every institute or firm, hence, this work is aiming to design and develop a low cost high performance computing system using computer cluster and applying the concepts of parallel computing. The intended cluster can be built using commodity of the shelf PCs and network hardware which will result in low cost supercomputer that can be implemented simply in any institute or university. In this research a deep study was done on parallel programming considerations and requirements as a foundation for applications development on the Beowulf cluster that was designed. Special emphasize was given to Message Passing Interface (MPI) which is a programming paradigm used widely on parallel computers, especially Scalable Parallel Computers with distributed memory, and on clusters of computers. During this study in order to verify the usage of parallel computing a numerical integration example was developed and the studied concepts were applied on it.