Investigating performance variations of an optimized GPU-ported granulometry algorithm
GIPSA-lab, UMR5216 CNRS/INPG/UJF/U.Stendhal, F-38402 GRENOBLE CEDEX, France
Conference on Design and Architectures for Signal and Image Processing (DASIP), 2012
@inproceedings{boulos2012investigating,
title={Investigating performance variations of an optimized GPU-ported granulometry algorithm},
author={Boulos, Vincent and Fristot, Vincent and Houzet, Dominique and Salvo, Luc and Lhuissier, Pierre},
booktitle={Design and Architectures for Signal and Image Processing (DASIP), 2012 Conference on},
pages={1–6},
year={2012},
organization={IEEE}
}
In this article, we present an optimized GPU implementation of a granulometry algorithm which is used a lot in the study of material domain. The main contribution to this algorithm is the binarization of the input data which increases throughput while reducing data allocated memory space. Also, the optimized GPU implementation brings an order of magnitude speedup compared to a CPU multi-threaded implementation. Furthermore, we investigate the reasons why GPU performance drop for different input data dimensions. Three main factors are exposed: under-exploited threads, threadblocks and streaming multiprocessors. This study should help the reader understand the tight relation that exists between the CUDA programming paradigm and the gpu architecture as well as some main bottlenecks.
February 22, 2013 by hgpu
