Adapting data processing methods to modern GPU architecture
Lodz University of Technology, Institute of Applied Computer Science, ul. Stefanowskiego 18/22, 90-924 Lodz, Poland
XV International PhD Workshop OWD, 2013
@article{szczepaniak2013adapting,
title={Adapting data processing methods to modern GPU architecture},
author={Szczepaniak, Bartlomiej},
year={2013}
}
Wavelet transform have a wide area of application in many scientific areas, for example signal processing, image compression [6] or data mining [4] [5]. Present requirements demand preforming large amount of calculations in the minimum time. For that reason the goal of this paper is to present an approach that will fulfill mentioned requirements, by adapting algorithms for signal processing to GPU demand. The article analyzes potential benefits of usage of the implementation of lattice structure on Graphics Processing Unit (GPU) and compares obtained results with the implementation on Central Processing Unit (CPU).
September 29, 2013 by hgpu
