Efficient analysis of a physica problem: different parallel computational approaches for a dynamical integrator evolution

A. GAUDIANI; A. SOBA; M. F. CARUSELA

CACIC

FAECE- UNCI Net

Año: 2013 p. 275 - 285

1330-1012

A great challenge for scientists is to execute their computational applications efficiently. Nowadays, parallel programming has become a fundamental key to achieve this goal. High-performance computing provides a solution to exploit parallel architectures in order to get optimal performance. Both parallel programming model and the system architecture will maximize the benefits if both together are suitable to the inherent parallelism of the problem. We compare three parallel algorithm performance when applied to the study of the heat transport phenomenon in a low dimensional system. We qualitatively analyze the obtained performance data based on the own characteristics of multicore architecture, shared memory and graphical multiprocesors related to programing models provided by MPI, OpenMP and Cuda. We conclude that Shared Memory and Data parallelism are the most suitable programing models to achieve a better performance of our model. We obtained an improvement near 5X, quite good for a program whose efficiency is strongly degraded by an integration process that essentially must be carried out in a serial way due to the dependence of the data.