Abstract

Parallel processing is necessary in many computer science domains. Graphics cards, even non-professional ones, are designed to perform very fast parallel processing tasks. This paper will provide an overview of recent mobile NVIDIA Graphics Processing Unit architecture, the benefits for using a GPU in parallel programming and will investigate the effect of operation grouping on the parallel processing performance of image pixels. The GPU performance will be compared to that of a multi core CPU designed for use in mobile platforms, in a pixel targeting image-processing algorithm.