Memristor-Based Logic Circuits: Concept and Prospect

Abstract

As computational demands from artificial intelligence increase, traditional von Neumann architectures face memory wall bottlenecks while CMOS technology approaches physical limitations defined by Moore’s law and Dennard scaling. This entry examines memristor-based logic circuits as potential solutions to overcome these computing constraints. We systematically analyse various memristor-based logic implementation approaches by reviewing and comparing stateful methods, including Material Implication (IMPLY) and Memristor-Aided Logic (MAGIC), as well as non-stateful techniques such as Memristor-Ratioed Logic (MRL) and Scouting Logic (SL). Each approach demonstrates unique capabilities for implementing fundamental logic operations, with progressive improvements in efficiency documented across multiple research studies. We discuss that while memristor-based logic shows promise, significant limitations exist: crossbar arrays can only integrate specific gates (NOT, NOR), MAGIC architecture struggles with multilevel cascade and large fan-out circuits, IMPLY requires numerous operational steps, and MRL faces challenges integrating CMOS-based inverters into crossbars.