Performance Analysis and Optimization of Hermite Methods on NVIDIA GPUs Using CUDA
The University of New Mexico, Albuquerque, New Mexico
The University of New Mexico, 2014
@phdthesis{dye2014performance,
title={Performance Analysis and Optimization of Hermite Methods on NVIDIA GPUs Using CUDA},
author={Dye, Evan T},
year={2014}
}
In this thesis we present the first, to our knowledge, implementation and performance analysis of Hermite methods on GPU accelerated systems. We give analytic background for Hermite methods; give implementations of the Hermite methods on traditional CPU systems as well as on GPUs; give the reader background on basic CUDA programming for GPUs; discuss performance characteristics of GPUs; we give recommended design choices for GPU implementations of Hermite methods; and present and discuss examples which illustrate the effect these design choices have on performance. Lastly, we present areas of future research that may yield increased performance for Hermite methods on GPUs.
February 2, 2015 by hgpu
