基于多核的Rijndael算法的并行优化与实现  

Parallel optimization and realization of Rijndael based on multi-core

作　　者：钱晓捷[1] 师攀攀[1] 王建辉[1] 

机构地区：[1]郑州大学信息工程学院,河南郑州450001

出　　处：《计算机工程与设计》2012年第6期2197-2202,2235,共7页Computer Engineering and Design

摘　　要：为提高Rijndael算法的执行效率,从对Rijndael算法的核心部分-轮变换的分析入手,在多核平台下提出了一种基于数据流分解的方式对Rijndael算法进行并行优化。采用的主要方法是将轮变换的各构成变换对整个状态的作用分割成对状态的每个组成单元的作用,使得各构成变换可以并行进行。实验结果表明,采用基于数据流的分解方式对Rijndael算法分解可有效的提高其执行效率。To improve the execution efficiency of Rijndael,starting from the analysis of the core component of Rijndael-the round transformation,a method of adopting data flow decomposition to optimize Rijndael based on multi-core is presented.The main method used is dividing each constitute transformation＇s acting on all states into acting on each composition unit,so each constitute transformation is executed in parallel.Experiments show that the method based on data flow decomposition to decompose Rijndael improve its execution efficiency effectively.

关 键 词：RIJNDAEL算法 轮变换 多核 数据流分解 并行优化 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]