GPU-accelerated "green" technology for fast analysis of nano-composites

Presented to the 10th Annual Symposium on Graduate Research and Scholarly Projects (GRASP) held at the Heskett Center, Wichita State University, April 25, 2014.Research completed at Department of Electrical Engineering and Computer Science, College of EngineeringIn this work, we investigate an energy-efficient simulation technique using NVIDIA compute unified device architecture (CUDA) to fast analysis of nanomaterial properties. In this approach, multithreaded parallel programming is used to solve the concurrent/parallel segments of the problem. The sequential portion of the program is executed on a multicore central processing unit (CPU), and the parallel portion of the program is executed on a graphics processing unit (GPU). Experimental results from finite difference method (FDM)-based on Laplace's equation demonstrate up to 257x speedup and 97% energy savings over a parallel MATLAB implementation while solving a 4Kx4K problem with reasonable accuracy.Graduate School, Academic Affairs, University Librarie