A Linux cluster of personal computers for the numerical simulation of natural airflows in greenhouses using a lattice model

Abstract

Some numerical models need a considerable computational effort to run their simulations. This fact is translated into long execution times delaying the decisions that have to be taken by the modeller in order to obtain the best phenomena description. This situation is also frequent in agronomical modelling, particularly when the subject of study is the fluid flow pattern as happens when simulating the natural airflows in a greenhouse. The knowledge of the effects of different ventilator configurations allows the crop manager to improve the greenhouse's natural ventilation conditions. In this work, a Linux cluster of ten personal computers (PC) is proposed with the aim of reducing the execution time of the lattice model simulations by means of parallel computing. This model has been used for describing fluid flow patterns in the presence of solid obstacles since the 1990s. The lattice model structure makes them suitable for a straightforward parallel computing implementation. As Jiménez-Hornero et al. (2005) show, this model is suitable for simulating the two-dimensional natural airflow in the vertical cross-section of a tropical crop protection structure described by Montero et al. (2001). The performance metrics clearly show the benefit of using the proposed Linux PC cluster in terms of execution time reduction and speedup with respect to the sequential running in a single PC.