Avaliação do uso de paralelismo em Python para um preditor de interações atômicas de estruturas tridimencionais de proteínas

Abstract

With the technological advances achieved in the field of computing, parallelism has consequently been increasingly used. In this context, tools that calculate distance and atomic interactions are usually coded without using this device. Thus, this work aims to optimize one of these tools, Ysera, to evaluate the impact of using parallelism in your code. The optimization was carried out with the restructuring of its code so that it worked in parallel with resources available for the Python language, the multiprocessing. The performance gain evaluation was done through several software executions sequential and parallel, using input files in PDB format of different sizes. Evaluating its execution time, memory consumption, and processing. The impact that the increase in the number of processes causes on its performance was also evaluated. As a result, a performance improvement of up to more than 50% was obtained compared to the original version of the tool, reducing the computational time to obtain the desired output data.