Skip to main content Skip to main navigation

Publikation

Devices, Synthesis, Circuits, and Verification for Resistive RAM-Based Digital Computing-in-Memory

Chandan Jha; Ankit Bende; Xingyue Qian; Simranjeet Singh; Weikang Qian; Vikas Rana; Farhad Merchant; Rolf Drechsler
In: Proceedings of the International Conference on Intelligent Computing and Systems at the Edge (ICEdge). International Conference on Intelligent Computing and Systems at the Edge (ICEdge-2025), December 18-20, Bangalore, India, 12/2025.

Zusammenfassung

Electronic Design Automation (EDA) has enabled the development of multi-billion-transistor designs in modern chips. Hence, similar methodologies are required to make digital computing using emerging technologies feasible. While for transistors, the development has happened over nearly eight decades, for emerging technologies, parallel efforts are being taken to enable their rapid feasibility. Some methodologies have been developed from scratch, while several are taken from existing technologies and tailored for emerging technologies. This paper attempts to highlight the recent developments in the area of using Resistive Random Access Memory (RRAM) for performing digital Computing-in-Memory (CiM). We have divided the paper into four sections and follow a bottom-up approach. First, we discuss the RRAM devices and their properties that make them feasible for digital computing. Second, we discuss the method to perform the mapping of an arbitrary design onto the RRAM crossbar. Third, we discuss the automated methods to generate the netlists of the mapping on the RRAM crossbars. Last, we discuss the automated formal verification strategies that are being employed to ensure the correctness of the transformation process at each of the synthesis and the netlist generation stages.