Title | Can Emerging Computing Paradigms Help Enhancing Reliability Towards the End of Technology Roadmap? |
Authors | Wang, Runsheng Zhang, Zuodong Zhang, Yawen Hu, Yixuan Sun, Yanan Qian, Weikang Huang, Ru |
Affiliation | Peking Univ, Inst Microelect, Beijing, Peoples R China Shanghai Jiao Tong Univ, Dept Micronano Elect, Shanghai, Peoples R China Shanghai Jiao Tong Univ, UM SJTU Joint Inst & MoE Key Lab Artificial Intel, Shanghai, Peoples R China |
Issue Date | 2021 |
Publisher | 2021 IEEE INTERNATIONAL RELIABILITY PHYSICS SYMPOSIUM (IRPS) |
Abstract | With CMOS technology shrinking into nanoscale, the circuit design margin has become extremely tight due to the severer transistor aging and process variations. To relieve the circuit reliability problems, many design optimization methods have been proposed. In essence, all these methods trade off area/power /performance for reliability. In this paper, we present a new perspective to enhance design reliability: using emerging computing paradigms. As the preliminary attempts, three reliability-enhanced design flows based on approximate computing and/or stochastic computing are demonstrated. The results show that some emerging computing paradigms are inherently robust, or can trade off computing accuracy for reliability, which provides the designers with much more flexibility. It also indicates that emerging computing paradigms are very promising for circuit design with ultimately scaled CMOS and beyond CMOS devices. |
URI | http://hdl.handle.net/20.500.11897/618968 |
ISBN | 978-1-7281-6893-7 |
ISSN | 1541-7026 |
DOI | 10.1109/IRPS46558.2021.9405167 |
Indexed | CPCI-S(ISTP) |
Appears in Collections: | 信息科学技术学院 |