李路路, 何春, 宗竹林, 章凌宇. 一种基于Dadda树的乘法器设计[J]. 微电子学与计算机, 2011, 28(5): 176-179.
引用本文: 李路路, 何春, 宗竹林, 章凌宇. 一种基于Dadda树的乘法器设计[J]. 微电子学与计算机, 2011, 28(5): 176-179.
LI Lu-lu, HE Chun, ZONG Zhu-lin, ZHANG Ling-yu. A Kind of Multiplier Design Based on Dadda Tree[J]. Microelectronics & Computer, 2011, 28(5): 176-179.
Citation: LI Lu-lu, HE Chun, ZONG Zhu-lin, ZHANG Ling-yu. A Kind of Multiplier Design Based on Dadda Tree[J]. Microelectronics & Computer, 2011, 28(5): 176-179.

一种基于Dadda树的乘法器设计

A Kind of Multiplier Design Based on Dadda Tree

  • 摘要: 在基带信号处理芯片中, 面积和速度是两个关键的指标.文中在改进的booth算法基础上, 采用了Dadda树压缩算法, 通过对压缩器基本单元的改进, 同时对符号位和尾部零填充进行优化设计;不仅保持了Wallace树结构的并行计算优势, 而且面积上也得到了很大的改善;同时相对于Wallace树结构的规则结构也更利于版图设计.压缩结果采用了多层CLA块技术, 使得乘法器的速度得到进一步的提高.在0.13μm的SMIC八层金属CMOS工艺下, DC (Design Compiler) 综合结果表明, 芯片面积为20633.59μm2, 最大延迟仅为3.00ns.

     

    Abstract: In the baseband signal processing chip, the area and speed are two key indicators.In this paper, we present a design which is based on the modified booth algorithm, using the Dadda tree compression algorithm with the improvement of the basic unit, while using the method of the sign bit and zero-filling of the tail to optimize the design, not only to maintain the Wallace tree structure of the advantages of parallel computing, and area also been greatly improved.Comparing to the Wallace tree structure, the rules of the structure is more conducive to layout.With compression results using a multi-CLA block technology, the multiplier′s speed is further improved.Under the SMIC 0.13μm CMOS process with the eight-layer metal, the synthesis result of this design by DC (Design Compiler) showed that the chip area is 20633.59μm2, and the maximum delay is only 3.00ns.

     

/

返回文章
返回