首页 | 本学科首页   官方微博 | 高级检索  
     检索      

一种基于稀疏矩阵的多核并行扰码方法
引用本文:马万治,王俊,赵宏志,唐友善.一种基于稀疏矩阵的多核并行扰码方法[J].现代电子技术,2012,35(1):68-70,74.
作者姓名:马万治  王俊  赵宏志  唐友善
作者单位:电子科学技术大学通信抗干扰技术国家级重点实验室,四川成都,611731
基金项目:国家自然科学基金资助项目(60832007,60901018,60902027,61001087,U1035002/L05); 国家重大科技专项(2009ZX03003-008-01,2010ZX03003-002-01); 中央高校基本科研业务费专项资金资助项目(ZYGX2009J008)
摘    要:针对多核环境中高速无线信号的加扰、解扰,提出了一种基于稀疏矩阵的多核并行扰码方法。首先对输入信号进行串/并转换,并将各路信号分别送入对应的处理器核;考虑基于稀疏矩阵的并行扰码生成器,在单个处理器核内,将其生成的伪随机码与输入信号进行模二加运算,得到单路信号的扰码输出;最后将多路并行的扰码输出变换为串行输出。运算量分析结果表明,采用IEEE 802.11n中的扰码生成多项式,与普通矩阵乘法实现的多核并行扰码方法相比,基于稀疏矩阵的多核并行扰码方法,其运算量降低了一个数量级。

关 键 词:稀疏矩阵  多核  并行扰码  运算量

A parallel scrambling method based on sparse matrix in multicore processors
MA Wan-zhi , WANG Jun , ZHAO Hong-zhi , TANG You-xi.A parallel scrambling method based on sparse matrix in multicore processors[J].Modern Electronic Technique,2012,35(1):68-70,74.
Authors:MA Wan-zhi  WANG Jun  ZHAO Hong-zhi  TANG You-xi
Institution:(National Key Lab.of Science and Technology on Communications,UESTC,Chengdu 611731,China)
Abstract:Considering scrambling and descrambling of high-speed wireless signal in multicore processors,a parallel scrambling method based on sparse matrix in multicore processors is proposed.Firstly,the input signal was passed through deserializer,and fed into specific processor.Then,in the specific processor,the sum modulo two of the input signal and the PN(pseudo-random) code generated by the Parallel Scrambler based on sparse matrix,was acted as the scrambling output of each input signal.At last,all parallel scrambling outputs were passed through serializer,and converted into serial output.By analysis of the calculated amount,using generating polynomial of scrambler in IEEE 802.11n,comparing with parallel scrambling method based on common matrix multiplication in multicore processors,the calculated amount of the parallel scrambling method based on sparse matrix in multicore processors decreased one order of magnitude.
Keywords:sparse matrix  multicore processors  parallel scrambling  calculated amount
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号