李志刚, 陈辉, 刘鹏, 武继刚. 基于择多-非-图的忆阻加法器设计[J]. 微电子学与计算机, 2022, 39(5): 104-110. DOI: 10.19304/J.ISSN1000-7180.2021.1130
引用本文: 李志刚, 陈辉, 刘鹏, 武继刚. 基于择多-非-图的忆阻加法器设计[J]. 微电子学与计算机, 2022, 39(5): 104-110. DOI: 10.19304/J.ISSN1000-7180.2021.1130
LI Zhigang, CHEN Hui, LIU Peng, WU Jigang. Memristive adder design based on majority-inverter-graph[J]. Microelectronics & Computer, 2022, 39(5): 104-110. DOI: 10.19304/J.ISSN1000-7180.2021.1130
Citation: LI Zhigang, CHEN Hui, LIU Peng, WU Jigang. Memristive adder design based on majority-inverter-graph[J]. Microelectronics & Computer, 2022, 39(5): 104-110. DOI: 10.19304/J.ISSN1000-7180.2021.1130

基于择多-非-图的忆阻加法器设计

Memristive adder design based on majority-inverter-graph

  • 摘要: 针对基于忆阻器的数字逻辑设计中存在延时开销和面积开销较高的问题,提出了一种双择多-非-图(Double Majority-Inverter-Graph,DMIG)逻辑.该逻辑能在一个时钟周期内同步实现2个择多-非-图逻辑.通过初始化DMIG结构中的忆阻器为不同的逻辑状态,并在DMIG结构的两端施加不同的电压,在一步内并行实现了两种不同的基础逻辑门.此外,利用逻辑综合方法优化全加器的布尔逻辑表达式,进一步设计了基于DMIG的一位全加器,并针对延时和面积开销两个性能分别提出了两种不同的优化方法。在Spice仿真环境下利用VTEAM模型仿真验证了DMIG和全加器,电路的延时开销和面积开销大幅度降低.延时优化加法器设计利用4个忆阻器通过4个步骤来实现,而面积优化加法器设计利用3个忆阻器通过5个步骤来实现。与现有的逻辑设计相比,所提出的设计减少了运算步骤及所需忆阻器数量,有效地优化了忆阻加法器性能.

     

    Abstract: Aiming at the problems of high latency and area overhead in the digital logic design based on memristor, A Double Majority-Inverter-Graph Logic (DMIG) for the memristive logic designisproposed. The proposed logic realizes 2 MIG logic synchronously in one clock cycle. Two different basic logic gates are implemented synchronously by initializing the memristor to different logic states and applying different voltages to both ends of the DMIG. In addition, this paper optimizes the Boolean logic expression of full adder by the logic synthesis method and designs a one-bit full adder based on DMIG. Two different optimization methods are proposed to reduce the performance of delay and area overhead.The proposed DMIG and one-bit full adder are simulated and verified in spice simulation environment by utilizing VTEAM, and the delay and area overhead of the circuit are greatly reduced. The delay optimized adder is realized within 4 memristors and 4 steps, while the area optimized adder is realized within 3 memristors and 5 steps. Compared to the existing logic design work, the proposed design reduces the calculation steps and the number of required memristors, and effectively optimizes the performance of the memristive adder.

     

/

返回文章
返回