Mingyuan Sun

Logic-in-memory (LIM) is an attractive approach to overcome the limitations of the conventional von Neumann architecture, especially in big data applications. Recently, many works have been focused on implementing Boolean logic or majority gate logic in memory using non-volatile storage devices. This work presents a programmable logic-in-memory (PLIM) architecture based on 22nm Fully Depleted Silicon on Insulator (FD-SOI) technology, which can map XOR, AND, and OR logic operations. Additionally, the architecture can be independently programmed in the column direction. The proposed architecture introduces a minimal hardware overhead and enables logic programming operations through simple manipulation.

In response to the escalating demand for edge encryption in the big data era, we introduce the implementation of the Memristor-based Transient Chaotic Neural Network (MTCNN) on FPGA. This innovative approach aims to elevate the security of key generation for the Advanced Encryption Standard (AES) algorithm. Through meticulous mapping and optimization of MTCNN-AES encryption algorithm on FPGA, we achieve efficient edge encryption. Notably, the FPGA-implemented MTCNN-AES features a key space 22 orders of magnitude larger than traditional AES, underscoring its substantial resilience against brute-force attacks. The proposed approach aligns with federal information processing standards, emphasizing its reliability, robustness and randomness.