Title :
GPU-based spatially divided predictive partitioned vector quantization for gifts ultraspectral data compression
Author :
Wei, Shih-Chieh ; Huang, Bormin
Author_Institution :
Dept. of Inf. Manage., Tamkang Univ., Tamsui, Taiwan
Abstract :
Predictive partitioned vector quantization (PPVQ) has been proven to be an effective lossless compression scheme for ultraspectral sounder data. In previous work, we have identified the two most time-consuming stages of PPVQ for implementation on GPU. By using 4 GPUs and a spectral division design in sharing the workload, we showed a 42x speedup on NASA´s Geostationary Imaging Fourier Transform Spectrometer (GIFTS) dataset compared to its original single-threaded CPU code. In this paper, an alternative spatial division design is developed to run on 4 GPUs. The experiment on the GIFTS dataset shows that a 72x speedup can be further achieved by this new design of the GPU-based PPVQ compression scheme.
Keywords :
Fourier transform spectra; computer graphic equipment; coprocessors; multiprocessing systems; vector quantisation; GIFTS ultraspectral data compression; GPU-based spatially divided predictive partitioned vector quantization; NASA; geostationary imaging Fourier transform spectrometer; lossless compression scheme; single-threaded CPU code; spectral division design; Graphics processing unit; Instruction sets; Kernel; Pipelines; Training; Vector quantization; Vectors; GIFTS sounder data; Graphic processor unit; data compression;
Conference_Titel :
Geoscience and Remote Sensing Symposium (IGARSS), 2011 IEEE International
Conference_Location :
Vancouver, BC
Print_ISBN :
978-1-4577-1003-2
DOI :
10.1109/IGARSS.2011.6048932
