CRC查询表及其并行矩阵生成方法  

CRC Lookup-table and its Implementation by Parallel Matrix

作　　者：梁海华[1] 盘丽娜[1] 赵秀兰[1] 李克清[1] 

机构地区：[1]常熟理工学院计算机科学与工程学院,常熟215500

出　　处：《计算机科学》2012年第B06期154-158,共5页Computer Science

基　　金：苏州市2010年科技计划项目(SYND201002);常熟理工学院青年教师科研启动项目(KYZ2010136Z)资助

摘　　要：循环冗余校验码(CRC)被广泛应用于通信领域。直接按位计算CRC校验值的方法难以满足高速链路要求,查询表方法与并行处理在很大程度上可以突破处理速度上的瓶颈。对查询表与并行处理矩阵之间的关系进行探讨,给出任意阶次生成多项式及任意处理位宽的查询表并行矩阵生成方法,并衍生出块处理方法。对表查询方法、并行矩阵查询方法及块处理方法的性能进行了比较分析,结果表明,并行处理位宽w增加,处理时间减少,并行矩阵方法在存储空间上表现更优;块处理通过减小校验序列长度,运算速度显著提高。Cyclic redundancy check（CRC） has already been used in the fields of communication widely.However,straightforward CRC implementation based on the bit cannot meet the requirements of high-speed link.By the lookup-table or parallel algorithm,the bottle-neck of speed can be resolved on a large scale.The relation between lookup-table and parallel matrix was investigated,from which lookup-table responding to polynomial of any order and any bit width processing can be given and deriving procedure of blocks.Comparison analysis on the performance of lookup-table,parallel matrix and blocks shows that less time is consumed while increasing width,and parallel matrix has better performance in requirement of memory space,and by reducing length of checking sequence,computational speed of blocks is increased significantly.

关 键 词：循环冗余校验码 位计算 查询表 并行处理矩阵 并行处理位宽 块处理 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]