TitleA Parallel Error Diffusion Implementation on a GPU (In Proceedings)
inProceedings of SPIE: IS&T/SPIE Electronic Imaging 2011 / Parallel Processing for Imaging Applications
Author(s) Yao Zhang, John Recker, Robert Ulichney, Giordano Beretta, Ingeborg Tastl, I-Jong Lin, John D. Owens
Keyword(s)Halftoning, Error Di ffusion, Parallel Processing, GPU Computing.
Year January 2011
LocationSan Francisco, CA
DateJanuary 23-27, 2011
Abstract In this paper, we investigate the suitability of the GPU for a parallel implementation of the pinwheel error di usion. We demonstrate a high-performance GPU implementation by efficiently parallelizing and unrolling the image processing algorithm. Our GPU implementation achieves a 10-30x speedup over a two-threaded CPU error di ffusion implementation with comparable image quality. We have conducted experiments to study the performance and quality tradeoff s for di fferences in image block sizes. We also present a performance analysis at assembly level to understand the performance bottlenecks.