Design Space Exploration for ReRAM-based Architectures to Address Scaling Non-idealities
This work addresses a practical problem for hardware designers of ReRAM-based systems by providing an incremental improvement over time-consuming simulation methods.
The paper tackles the challenge of balancing scaling effects and peripheral circuit power in ReRAM-based in-memory computing architectures by proposing a design space exploration framework that identifies optimal configurations, achieving efficient high-performance design without exhaustive simulations.
ReRAM-based in-memory computing (IMC) architectures are promising candidates for energy-efficient matrix-vector multiplication. While scaling the size of ReRAM arrays allows for the amortization of power-hungry peripheral circuits like DACs and ADCs, it simultaneously introduces more parasitic along the signal path. Because of these challenges, current design methodologies often lack practical guidelines to balance these effects at early design stage, forcing designers to rely on time-consuming, iterative transistor-level simulations. In this work, we propose a comprehensive framework for design space exploration that enables the selection of optimal array size, ADC resolution, and system frequency without requiring exhaustive simulations. The framework utilizes a specialized testbench to extract parameters from a limited set of representative transistor-level simulations. These parameters are then used to accurately predict the performance of arbitrary architectures. We demonstrate the effectiveness of this framework through two realistic design cases aimed at maximizing energy efficiency (TOPs/s/W). The results show that the framework successfully identifies optimal architectural configurations under strict power and error constraints, providing an efficient path for high-performance IMC design.