Device implementation and arithmetic circuit design of ternary logic

• 主要作者: Zhao, Guangchao
• 其他作者: Tay Beng Kang
• 格式: Thesis-Doctor of Philosophy
• 語言: English
• 出版: Nanyang Technological University 2024
• 主題: Engineering; Ternary logic computing; Multiple-valued logic; Logic devices based on 2D semiconductors; Ternary arithmetic circuits
• 在線閱讀: https://hdl.handle.net/10356/181052
• 機構: Nanyang Technological University

Tremendous amounts of data are being generated and should be processed during the massive employment of artificial intelligence and edge/cloud computing, necessitating computers with higher capabilities. However, as Moore's Law approaches its limit, enhancing computing capability through transistor miniaturization is becoming less feasible. New strategies for “more-than-Moore” are highly preferred to drive future technological advancements. Multiple-valued logic, especially ternary logic with its superior data density compared to traditional binary logic, offers a viable route to augment computing capability while reducing chip interconnections and dynamic power consumption. Despite the growing interest in ternary logic within academia and industry, its practical application remains limited due to the increased transistor count and connections required by earlier logic gate implementations. In this thesis, we present efficient implementation methods of basic ternary logic gates based on emerging nano-electronics including two-dimensional material-based transistors. The fabricated ternary devices show high performance in DC gain and low power consumption. Additionally, compact models of the devices have been established and calibrated, facilitating the circuit design and simulation in Electronic Design Automation (EDA) tools. At the circuit level, this thesis introduces novel ternary arithmetic circuits, including adders and multipliers, that are optimized for low power consumption and minimal delay. Finally, we showcase the practical benefits of ternary logic systems in data processing through the implementation of carry-free arithmetic circuits, underscoring their potential as a transformative computing paradigm.