Wavelet Encoding and Multi-GPU Programming
Department of Mathematics, Massachusetts Institute of Technology, Cambridge, MA
Massachusetts Institute of Technology, 2013
@article{kessler2013wavelet,
title={Wavelet Encoding and Multi-GPU Programming},
author={Kessler, Andre},
year={2013}
}
We investigate compression of large-volume spatial data using the wavelet transform, computed massively in parallel on NVIDIA graphics processing units (GPUs). In particular, Haar basis wavelets are used to achieve compression ratios of [100x] or more. Computation is done over a set of computing nodes consisting of multiple nodes and multiple GPUs per node. Significantly more data than can be stored on-board the individual GPUs is streamed on and successfully compressed. After the compression, the data is ready to be analyzed or manipulated by other tools, after which the changed data or extracted features will be decompressed and stored.
January 3, 2014 by hgpu
