计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (7): 104-107.

• 产品、研发、测试 • 上一篇    下一篇

对数/指数算法的研究及其VLSI实现

赵海燕 周晓方 周电   

  1. 复旦大学 复旦大学专用集成电路与系统国家重点实验室
  • 收稿日期:2006-07-04 修回日期:1900-01-01 出版日期:2007-03-01 发布日期:2007-03-01
  • 通讯作者: 赵海燕

Arithmetic Research of logarithm and anti-logarithm converters and VLSI implementation

Haiyan Zhao   

  • Received:2006-07-04 Revised:1900-01-01 Online:2007-03-01 Published:2007-03-01
  • Contact: Haiyan Zhao

摘要: 本文提出一种二进制数的指数/对数运算的线性近似的改进算法,并VLSI实现。该算法能较好的提高精度,相比于现有最新文献提出的算法,对数运算的相对误差减少了46.1%,指数运算的相对误差减少了32.2%。实现时,设计了前导1探测电路和减小误差的误差补偿电路。该算法VLSI实现简单,只需组合逻辑就能在一个时钟周期内就能得到计算结果。

关键词: 对数/指数运算 线性近似 硬件实现

Abstract: This paper presents an improved binary-to-binary logarithm and anti-logarithm line approximation including its VLSI implementation. This arithmetic can increase precision well, compared with the newest references, the logarithm converter’s percent error reduces 46.1 percent and the anti-logarithm converter’s percent error reduces 32.2 percent. A leading-one detector circuit is designed to obtain the leading-one position of an input binary word and error-correcting circuit is designed to reduce error. The implementation is simply, uses combinational logic only and calculates the approximation in a single clock cycle.

Key words: Logarithm/anti-logarithm converter Line approximation VLSI implementation