Committee:

Dr. Shimeng Yu, ECE, Chair , Advisor

Dr. Sung-Kyu Lim, ECE

Dr. Arijit Raychowdhury, ECE

Dr. Shaolan Li, ECE

Dr. Jae-Sun Seo, Arizona State

Abstract: The objective of this research is to accelerate deep neural networks (DNNs) with emerging non-volatile memories (eNVMs) based computing-in-memory (CIM) architecture. The research first focuses on the inference acceleration and proposes a resistive random access memory (RRAM) based CIM architecture. Two generations of RRAM testchips that monolithically integrate the RRAM memory array and CMOS peripheral circuits are designed and fabricated using commercial embedded RRAM process respectively. This research develops a PyTorch based framework that incorporates the device characteristics into the DNN model to evaluate the impact of the eNVM nonidealities on training/inference accuracy. Furthermore, to overcome the challenges posed by the asymmetric conductance tuning behavior of typical eNVMs, this research proposes a novel 2-transistor-1-FeFET (ferroelectric field effect transistor) based synaptic weight cell that exploits hybrid precision for in-situ training and inference, which achieves near-software classification accuracy on MNIST and CIFAR-10 dataset.