Modeling of Heat Diffusion Through Isotropic Media Using Graphical Processing Units
Centre for NDE, Indian Institute of Technology – Madras, Chennai 600036, T.N., India
National Seminar & Exhibition on Non-Destructive Evaluation (NDE 2011), 2011
@article{nahas2011modeling,
title={MODELING OF HEAT DIFFUSION THROUGH ISOTROPIC MEDIA USING GRAPHICAL PROCESSING UNITS},
author={Nahas, CA and Krishnamurthy, CV and Rajagopal, P. and Balasubramaniam, K.},
year={2011}
}
With accurate simulation of ever-more complex phenomena desired, numerical models are getting increasingly sophisticated and thus take considerable amount of time to run on normal computers. Parallel computing has emerged as an important paradigm in response, allowing engineers to run programs faster. In recent years graphics processing units (GPU) are among the massively parallel devices widely available to the commodity market. General purpose computing on these devices became mainstream after the introduction of NVIDA ‘compute unified device architecture’ (CUDA). In this paper, we develop CUDA implementations of the finite difference time domain (FDTD) scheme for two- and three-dimensional heat diffusion through homogenous media, a phenomenon of much interest to the NDE community. Simulations are run on the commodity GPU GeForce 9800GT which has 112 CUDA cores. Results show a strong improvement in the speed of computation as compared to that of a serial implementation on a CPU. We also discuss factors that can improve accuracy and speed of GPU based computation.
August 5, 2012 by hgpu
