A Novel Non-Volatile Inverter-based CiM: Continuous Sign Weight Transition and Low Power on-Chip Training
Dong Zhang,Yuye Kang,Gan Liu,Zuopu Zhou,Kaizhen Han,Chen Sun,Leming Jiao,Xiaolin Wang,Yue Chen,Qiwen Kong,Zijie Zheng,Long Liu,Xiao Gong
DOI: https://doi.org/10.48550/arXiv.2209.09083
2022-09-19
Abstract:In this work, we report a novel design, one-transistor-one-inverter (1T1I), to satisfy high speed and low power on-chip training requirements. By leveraging doped HfO2 with ferroelectricity, a non-volatile inverter is successfully demonstrated, enabling desired continuous weight transition between negative and positive via the programmable threshold voltage (VTH) of ferroelectric field-effect transistors (FeFETs). Compared with commonly used designs with the similar function, 1T1I uniquely achieves pure on-chip-based weight transition at an optimized working current without relying on assistance from off-chip calculation units for signed-weight comparison, facilitating high-speed training at low power consumption. Further improvements in linearity and training speed can be obtained via a two-transistor-one-inverter (2T1I) design. Overall, focusing on energy and time efficiencies, this work provides a valuable design strategy for future FeFET-based computing-in-memory (CiM).
Mesoscale and Nanoscale Physics