Memristors based Computation and Synthesis

Overview

• Memristor is a fundamental circuit element with non-volatile properties that are being explored for beyond-CMOS computing.
• Memristors can perform basic logic operations and are used in applications like Dot Product Engines and Convolutional Neural Networks.
• This paper presents a new behavioral model of memristors and uses it to build a 32-bit ripple carry adder.
• The paper compares the area, power, and time delay of the 32-bit Ripple Carry Adder using memristors against 45nm CMOS technology.

Plain English Explanation

The paper discusses memristors, which are a type of electronic component that was identified as the fourth fundamental circuit element in 1971. Memristors have some unique properties, like being able to "remember" their state even when the power is turned off. This makes them interesting for use in beyond-CMOS computing - computers that go beyond the traditional silicon-based transistors.

Researchers have found ways to use memristors to perform basic logic operations, like AND, OR, and XOR. They've also used memristors in more complex applications, like Dot Product Engines and Convolutional Neural Networks.

In this paper, the authors present a new way of modeling how memristors behave. They then use this model to build a 32-bit ripple carry adder - a basic component for doing math in computers. The paper compares this memristor-based adder to one made using traditional 45nm silicon transistors, looking at things like how much space it takes up, how much power it uses, and how fast it is.

Technical Explanation

The paper introduces a new behavioral model for memristors, which are described as the fourth fundamental circuit element after resistors, capacitors, and inductors. Memristors are of interest due to their non-volatile and spintronic properties, making them a viable solution for beyond-CMOS computing.

The authors leverage this new memristor model to design and implement a 32-bit ripple carry adder. Ripple carry adders are a basic building block for arithmetic operations in digital systems. The paper provides a detailed description of the memristor-based adder architecture and its operation.

Through extensive simulations, the authors compare the area, power, and time delay of the 32-bit memristor-based ripple carry adder against an equivalent design implemented using 45nm CMOS technology. The results show that the memristor-based adder offers significant improvements in terms of area and power consumption, although it may have a slightly higher time delay.

Critical Analysis

The paper provides a comprehensive analysis of using memristors to implement a ripple carry adder, a fundamental building block for digital arithmetic. The authors' introduction of a new memristor behavioral model is a valuable contribution, as accurate modeling is crucial for designing and simulating memristor-based circuits.

While the results demonstrate the potential advantages of memristor-based adders over CMOS counterparts, the paper does not address some important considerations. For example, it does not discuss the challenges around memristor variability and reliability, which can be significant hurdles for practical implementation. Additionally, the paper focuses solely on the ripple carry adder and does not explore the use of memristors in more complex arithmetic units or other digital logic circuits.

Further research is needed to fully assess the viability of memristor-based computing, including investigating techniques to mitigate variability, improve reliability, and expand the range of applications beyond basic arithmetic operations.

Conclusion

This paper presents a new behavioral model for memristors and leverages it to design and evaluate a 32-bit ripple carry adder. The results show that memristor-based adders can offer significant improvements in area and power consumption compared to traditional CMOS implementations, although they may have slightly higher time delays.

While the findings are promising, the paper does not address some key challenges, such as memristor variability and reliability, that will need to be overcome for widespread adoption of memristor-based computing. Additional research is necessary to further explore the potential of memristors in more complex digital logic and arithmetic circuits.