Ternary computing using a novel spintronic multi-operator logic-in-memory architecture

This paper presents a novel multi-operator ternary logic-in-memory (LiM) design that addresses the challenges of the Von-Neumann bottleneck architecture. The proposed design leverages magnetic tunnel junctions (MTJ), which provide nonvolatile memory capabilities. Nonvolatility ensures that data is retained even when powered down, enhancing reliability and reducing energy waste during idle periods. The proposed LiM architecture can perform all fundamental ternary logic operations, such as AND/NAND, OR/NOR, and XNOR/XOR, within the memory itself, eliminating the need for separate logic units and reducing data transfer delays. The proposed innovative approach based on the well-known MTJ and FinFET technologies offers substantial improvements over existing architectures, with a 78 % reduction in delay and an 86 % reduction in power consumption. Digital image processing simulations, as real-world applications, also demonstrate the design's practical effectiveness by achieving up to 98 % reduction in energy consumption. This makes the proposed ternary LiM architecture a promising solution for future high-performance, energy-efficient computing systems.