In‐Memory Stateful Logic Computing Using Memristors: Gate, Calculation, and Application

Abstract

Combining Boolean logic and nonvolatile memory functions, memristor-based stateful logic circuits can minimize data movement during the computing process to achieve futuristic in-memory computing. This may solve the problem of the von Neumann bottleneck in the current computing architecture. Herein, the recent developments in memristor-based stateful logic computation are discussed from several perspectives, including the device, circuit, operational principles, and applications. Stateful logic gates correspond to in-memory logic gates, with the inputs and outputs having identical physical entities, such as resistance. The developments in stateful logic primitive gates reported in the past decade are summarized. The influential factors in their actual implementation for stateful logic computation are also discussed. Then, methods for allocating the logic gates into the crossbar array are explained, which are for implementing the complex computing instances in the crossbar array. Finally, several data-intensive applications achieved using the in-memory computing method are discussed for the practical application of stateful logic in future computing systems.

Error Detection and Correction Method Toward Fully Memristive Stateful Logic Design

“…Along with the control voltages, this circuit structure is called a PMR-two-2IMP gate. [13] Any complicated Boolean logic function can be realized by cascading the stateful logic gate and the false (reset) operation. To perform the IMP logic function, as shown in Figure 3b, two positive voltage pulses of V P with an amplitude lower than V set , and V Q with an amplitude larger than V set , are applied to P and Q, respectively.…”

“…Along with the control voltages, this circuit structure is called a PMR‐two‐2IMP gate. [ 13 ] Any complicated Boolean logic function can be realized by cascading the stateful logic gate and the false (reset) operation.…”

“…[1][2][3][4][5] To date, many stateful logic gates based on various circuit structures have been proposed. [6][7][8][9][10][11][12] According to the naming rule proposed by Xu et al, [13] the "Structure-N-n Function" format is used to define different types of stateful logic gates, among which, 2 parallel memristors þ a series resistor for two inputs imply logic (PMR-two-2IMP), [10] 3 parallel memristors þ a series resistor for two inputs nor logic (PMASM-two-3NOR), [11] and 2 parallel memristors þ 1 anti-serial memristor for two inputs nor logic (PMASM-two-3NOR) [12] are the classical gates that have received extensive attention in both experiments and applications. [14][15][16][17][18][19][20][21][22][23] The stateful logic operation is inherently based on the conditionally resistive switching of the resistance states under a group of fixed control voltages.…”

Cationic Interstitials: An Overlooked Ionic Defect in Memristors

“…As the memories and CPUs are separated in the current von Neumann computer system, the data have to be transferred between them through the limited bandwidth buses, which limits the time and energy efficiencies in the data processing. Such issue could be addressed by achieving an in-memory computing paradigm, for which the memristor is a suitable device because of its higher data storage density (Cheng et al, 2017;Chen et al, 2019) and excellent physical characteristics of conditional switching and physical MAC operation (Yang et al, 2017;Zhou et al, 2019;Xu et al, 2021;Li et al, 2022). Meanwhile, they can also bridge various electrical devices and be applied in energy storage, remote sensing, low-power applications, etc.…”

Cellular automata imbedded memristor-based recirculated logic in-memory computing

“…Matrix multiplication based on memristive circuits always has been used to accelerate the data-intensive computing tasks like artificial neural networks 17 . Logical operation based on memristor is usually designed as logic circuit, such as adder 18,19 or logic gate 20,21 . Part of CA's transition rules can be transformed into corresponding Boolean functions, which implies that CA can be implemented by using memristive circuits.…”