首页 | 本学科首页   官方微博 | 高级检索  
     

FT-SIMD:一种高性能乘法器的设计
引用本文:李国强,陈书明,万江华,杨惠.FT-SIMD:一种高性能乘法器的设计[J].计算机工程与科学,2012,34(1):53-57.
作者姓名:李国强  陈书明  万江华  杨惠
作者单位:国防科学技术大学计算机学院,湖南长沙,410073
基金项目:核高基重大专项(2009ZX01034001006)
摘    要:为了提高多媒体数据的处理能力,高性能DSP普遍引入了SIMD技术。作为DSP重要组成部分的乘法器也必须具备这一功能。本文对SIMD乘法器的实现进行深入研究,提出了一种新的SIMD乘法器体系结构,采用两个16×8乘法器,通过对其操作数和结果进行符号扩展和拼接等处理,简单而高效地实现了16位FT-SIMD乘法器。同时,本体系结构可以扩展为32位和64位的SIMD乘法器。

关 键 词:SIMD  乘法器  5-2压缩  4-2压缩  Booth编码
收稿时间:2011-05-15
修稿时间:2011-08-18

FT-SIMD :Design of a High-Performance Multiplier
LI Guo-qiang , CHEN Shu-ming , WAN Jiang-hua , YANG Hui.FT-SIMD :Design of a High-Performance Multiplier[J].Computer Engineering & Science,2012,34(1):53-57.
Authors:LI Guo-qiang  CHEN Shu-ming  WAN Jiang-hua  YANG Hui
Affiliation:(School of Computer Science,National University of Defense Technology,Changsha 410073,China)
Abstract:In order to enhance the multimedia data processing abilities, high performance DSPs have generally introduced the SIMD technology. As an important component in DSPs, the multiplier has to support the SIMD function. After an in-depth study on the SIMD multiplier’s implementation, this work proposes a simple and highly effective 16 bit FT-SIMD multiplier, which is composed of two 16×8 bits multipliers, with the help of signed-expansion and splicing operations on the operands and results, the 16-bit SIMD multiply is effectively supported. What’s more, our multiplier can be easily expanded to support the 32-bit and 64-bit SIMD multiply operations.
Keywords:SIMD  multiplier  5-2 compression  4-2 compression  Booth coding
本文献已被 CNKI 万方数据 等数据库收录!
点击此处可从《计算机工程与科学》浏览原始摘要信息
点击此处可从《计算机工程与科学》下载全文
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