As data-centric applications escalate in complexity and volume, the inherent latency and energy inefficiency of von Neumann architectures pose significant challenges to next-generation computing systems. In-memory computing, leveraging emerging memristive devices, has emerged as a transformative paradigm to overcome these bottlenecks by enabling logic operations within the memory fabric. Here, we present a reconfigurable, fully hardware-realized logic-in-memory computing platform based on complementary resistive-switching (CRS) memristor crossbar arrays that physically implements the complete set of 16 two-input Boolean logic functions─and further enables a half-adder circuit─within an ultraefficient three-step operation cycle. Our system employs a four-variable logic model mapped onto Ag/GO:Au NPs/Al/GO:Au NPs/ITO CRS devices, exploiting their robust, repeatable resistive switching to achieve scalable, reconfigurable logic-in-memory functionality. Built on a graphene-oxide-based composite material system modified by Au nanoparticles, this work not only demonstrates a compact, low-power alternative to conventional CMOS logic but also delineates a viable pathway for future high-density, neuromorphic hardware systems with embedded logic capabilities.
Wang et al. (Tue,) studied this question.