Analysis and simulation of graphs applied to learning with parallel programming in HPC

Abstract

Large-scale graph analysis or also called network analysis of networks is supported by different algorithms, among the most relevant are PageRank (Web page ranking), Betweenness centrality (centrality in a graph) and Community Detection, these by of their complexity and the large amount of data that process diverse applications, increasingly need to use computational resources such as processor, memory and storage, for these reasons, it is necessary to apply high performance computing or HPC (High Performance Computing) but it would not be useful to apply HPC without having designed these algorithms in parallel programming, in this part there have been many studies on its application and methodologies to do it. The purpose of this work is to create a framework that allows computer science students to abstract a computer system based on the parallel programming paradigm, which implies that students to get acquainted with the resolution of algorithmic problems in a more natural way and away from the typical sequential thinking., The development of a graph analysis design pattern oriented to parallel programming in HPC, complemented with the design of didactic learning techniques in the network such as laboratories and/or simulators are key in the development of this framework.