Optimization Of JPEG Encoding Based On Embedded Multi-core Processor

With the rapid development of embedded technology and multimedia technology, the use of single-core processor devices already can not meet the growing wealth of multimedia information processing, performance of processor directly affects the user experience, embedded processor develop from single-core to dual-core, from dual-core to quad-core development even to the eight-core, quantity of processor core rises exponentially. Nowadays, the appearance of multi-core processors caused a huge challenge to the device’s battery. Various chip manufacturers continue to optimize the performance of multi-core processors and power problems are one of the hottest current embedded technology development. Therefore, under the conditions of the same resources, optimization of embedded multi-core processors is a research direction that is the much-needed and has significant practical value.In this study, we propose two optimization methods, ARM NEON instruction set and multi-core processors, aimed to accelerate the processing speed of the processor, reducing processing time and improve the efficiency of the processor, and reduce processor power consumption and radiating. NEON instruction set is an extension instruction set based on the new ARM v7 architecture, NEON instruction set is based on hardware support for NEON technology, the optimization use its characteristics to accelerate JPEG encoding and decoding by NEON technology, and conducted a number of experiments to validate NEON technology for JPEG codec has a better optimization of acceleration effect. Optimization of multi-core processors is based on program which have been optimized by NEON technology, it is a deeper optimization, as used in this experiment development board is a quad-core processor, it can used up to four threads in parallel processing tasks. In the process of optimization, analyze each part of JPEG encoding carefully and the feasibility of parallel processing, develop appropriate parallel processing solutions to optimize the program based on multi-core processor, and verify its performance by experiment, record the processing time which used for comparative analysis to the NEON technology in the treatment of time, analyze each core processor activity with professional software. Experimental results show that, compared with JPEG codec, optimization based on NEON technology has better results, the processing time is greatly shorten, also optimization based on multi-core processor has better acceleration effect than optimization based on NEON technology.