摘要
本文提出了一种基于DSP(D igital Signal Processing)核及其数据链路特征的优化方法,并验证了该优化方法可以使软件效率提高5~15倍.高效快速的算法程序,不但使图像的实时传输变为可能,还可以节约硬件成本,使系统变得简单,提高系统的稳定性.为解决软件效率问题,本文主要研究了去相关、内联函数、短整型数据用整型处理和软件流水技术,并把这些优化技术应用在基于C6416平台的实时图像压缩系统中,在输入数据率82M byte/s^110M byte/s、输出码流2.9M byte/s的条件下,使该系统达到了实时处理的能力.
This paper studied the optimizing technology based on the character of the DSP （ Digital Signal Processing） and its data routeway, and proved that these optimizing technologies could improve the software＇ s efficiency from five times to fifteen times. High efficient programmer can make the system real-time, and it can reduce the cost of hardware, make the system simple and improve its reliability. To improve the software efficiency ,this paper studied the removing relativity,intrinsic, short data processed as int data and software pipeline technologies, and these technologies are applied in the real-time image compression system based on C6416 ,with the inputting data rate 82M byte/s - 110M byte/s and the outputting data rate 2.9M byte/s, the system achieved the goal of real-time processing.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2006年第9期1558-1561,共4页
Acta Electronica Sinica
关键词
软件流水
去相关
内联函数
实时处理
software pipeline
remove relativity
intrinsic
real-time processing
