[1]魏武杨靓.图像处理中数据复用及存储层次设计的研究[J].计算机技术与发展,2012,(12):43-46.
　WEI Wu,YANG Liang.Data Reuse and Memory Hierarchy Design in Image Processing[J].,2012,(12):43-46.
点击复制

图像处理中数据复用及存储层次设计的研究()

分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

Keywords:: data-reuse level; block sizes; memory hierarchy ; assembly language

摘要:: 图像处理过程中数据传输量大，通过数据复用可以减少数据传输量提高执行效率。同时图像分块的大小在很大程度上决定了存储器大小，而存储器大小对系统的功耗影响巨大。文中通过对分块图像平滑算法的研究，提出了四种数据复用层次。再以TSIOI为仿真平台，在不同的分块大小下编写汇编代码模拟数据流在不同存储器间的流动，然后得出其执行效率、存储器大小和数据复用层次、块大小等的关系。为硬件设计中存储器体系结构的优化设计提供了参考

Abstract:: There was large amount data transfer in image processing. It could reduce data transmission quantity and improve the efficiency by using data reuse. And block size largely determined the memory size which had a great influence on system consumption. The four data reuse levels were presented based on research of the block image smoothing algorithm and various block methods in this paper. Finally simulations were done on the platform of TS101 in assembly language and different block sizes. Then the relationship between efficiency and memory size was obtained as well as the relationship between data reuse level and block size. It provided a reference design for the optimization of the memory architecture in hardware design

备注/Memo:: 中国航天集团课题预言项目（417010204）魏武（1986-），男，硕士研究生，主要研究方向为计算机系统结构、数字电路设计；杨靓，研究员，硕士生导师，主要研究方向为嵌入式高速计算机系统结构、ASIC没计、数字图像处理

更新日期/Last Update: 1900-01-01