Cryogenic Logic-In-Memory Bit-Serial Addition With Majority Gates Based On The Quantum Anomalous Hall Effect

Abstract

As CMOS technology approaches its physical limits, new alternatives are being investigated to improve computational performance and energy efficiency. Cryogenic computing offers a viable solution, enhancing computing speeds without the need for further scaling. However, the issue of the "memory wall" continues to challenge traditional von-Neumann architectures, even under cryogenic conditions. To address this, logic-in-memory Computing architectures, which enable computation within the memory unit, provide benefits by minimizing data transfer between the memory and processing units. This reduces energy consumption and cooling costs at low temperatures. In this study, we introduce CryoLiMC, a cryogenic logic-in-memory Computing framework using a non-volatile memory system based on the quantum anomalous Hall effect (QAHE). This memory system enhances energy efficiency, resilience to process variations, and scalability. With Moore's Law slowing down, alternative technologies such as memristive devices and resistance-switching memories show promise in overcoming the memory wall by enabling computation within memory arrays. Additionally, the potential of majority logic as a Boolean logic primitive for in-memory computing is explored, particularly in the context of memristive systems. The paper also examines the use of majority gates in quantum cellular automata (QCA), offering methods to simplify Boolean functions and minimize hardware requirements in QCA designs.