Modeling the performance of parallel applications using model selection techniques

摘要

Nowadays, parallel architectures are changing so fast that there is a need for scalable and efficient tools tornanalyze and predict the performance of parallel applications. Analytical models are proved to be a usefulrnapproximation for characterizing parallel algorithms, but developing accurate analytical models is a hardrnissue, and, in general, they provide coarse performance predictions due to their intrinsic lack of accuracy. Inrnthis paper, we describe in detail the Tools for Instrumentation and Analysis (TIA) framework, an easy-touserntool that automatically obtains accurate performance models by means of analytical expressions. Thisrnframework automatizes most of its internal tasks, reducing opportunities for human error, and it only requiresrnthe user to focus on the metrics and execution parameters that might influence the performance, those thatrnshould be considered in the modeling process. Its main advantage over other tools is that TIA uses modelrnselection techniques that allow the automation of the modeling process. As a case of study, the use of TIA tornobtain analytical models of different implementations of the broadcast collective communication in a clusterrnof multicores is shown. The results obtained by TIA are evaluated and compared with theoretical approachesrnbased on the LogGP model.