Simulating a P system based efficient solution to SAT by using GPUs
Research Group on Natural Computing, Dpt. of Computer Science and Artificial Intelligence, University of Seville, Avda. Reina Mercedes s/n, 41012 Sevilla, Spain
Journal of Logic and Algebraic Programming, Volume 79, Issue 6, August 2010, Pages 317-325 (02 April 2010)
@article{cecilia2010simulating,
title={Simulating a P system based efficient solution to SAT by using GPUs},
author={Cecilia, J.M. and Garc{‘i}a, J.M. and Guerrero, G.D. and Mart{‘i}nez-del-Amor, M.A. and P{‘e}rez-Hurtado, I. and P{‘e}rez-Jim{‘e}nez, M.J.},
journal={Journal of Logic and Algebraic Programming},
volume={79},
number={6},
pages={317–325},
issn={1567-8326},
year={2010},
publisher={Elsevier}
}
P systems are inherently parallel and non-deterministic theoretical computing devices defined inside the field of Membrane Computing. Many P system simulators have been presented in this area, but they are inefficient since they can not handle the parallelism of these devices. Nowadays, we are witnessing the consolidation of the GPUs as a parallel framework to compute general purpose applications. In this paper, we analyse GPUs as an alternative parallel architecture to improve the performance in the simulation of P systems, and we illustrate it by using the case study of a family of P systems that provides an efficient and uniform solution to the SAT problem. Firstly, we develop a simulator that fully simulates the computation of the P system, demonstrating that GPUs are well suited to simulate them. Then, we adapt this simulator to the GPU architecture idiosyncrasies, improving the performance of the previous simulator.
November 20, 2010 by hgpu
