CUDA Accelerated Entropy Constrained Vector Quantization and Multiple K-Means
NVIDIA UK Ltd., London, UK
The Fourth International Conference on Advanced Communications and Computation (INFOCOMP), 2014
@inproceedings{ashley2014cuda,
title={CUDA Accelerated Entropy Constrained Vector Quantization and Multiple K-Means},
author={Ashley, John and Braverman, Amy},
booktitle={INFOCOMP 2014, The Fourth International Conference on Advanced Communications and Computation},
pages={30–34},
year={2014}
}
Multi-trial sampled K-means performance and scalability is studied as a stepping stone towards a Graphical Processing Unit implementation of Entropy Constrained Vector Quantization for interactive data compression. Basic parallelization strategies and data layout impacts are explored with K-means. The K-means implementation is extended to Entropy Constrained Vector Quantization, and additional tuning specific to the anticipated use case is performed. The results obtained are sufficiently promising that this will in the next phase be applied to the interactive exploration and visualization of very large satellite datasets.
August 1, 2014 by hgpu
