Thermal Heating in ReRAM Crossbar Arrays: Challenges and Solutions

Abstract

The high speed, scalability, and parallelism offered by ReRAM crossbar arrays foster the development of ReRAM-based next-generation AI accelerators. At the same time, the sensitivity of ReRAM to temperature variations decreases $\text{R}_{ON}/\text{R}_{OFF}$ ratio and negatively affects the achieved accuracy and reliability of the hardware. Various works on temperature-aware optimization and remapping in ReRAM crossbar arrays reported up to 58% improvement in accuracy and $2.39\times $ ReRAM lifetime enhancement. This paper classifies the challenges caused by thermal heat, starting from constraints in ReRAM cells’ dimensions and characteristics to their placement in the architecture. In addition, it reviews the available solutions designed to mitigate the impact of these challenges, including emerging temperature-resilient Deep Neural Network (DNN) training methods. Our work also provides a summary of the techniques and their advantages and limitations.