Performance analysis on scientific computing and cloud computing environments

Abstract

The search for better execution times was one of the motivations for the emergence of High Performance Computing (HPC) whose importance has obtained significative preponderance in the domain of scientific computing, either in applications for an academic or industrial environment. Cloud Computing proposes the integration of various technological models for the provision of a hardware infrastructure, development platforms and applications as services based on-demand. However, with the emergence of cloud computing, a new challenge arose which was to allow the execution of HPC applications in this new environment. This paper aims to demonstrate the feasibility of using a local cluster built on a private cloud for HPC experiments, and compare with applications running in a real environment for HPC.