A Novel FeFET-Based Multibit Content Addressable Memory Through Thermometer Encoding for Manhattan Distance Metric with High Area- and Energy-Efficiency
In this work, a novel ferroelectric FET (FeFET)-based multibit content addressable memory (MCAM) through modified thermometer encoding (TE), for Manhattan distance (MD) metric is proposed for the first time. By utilizing the multilevel storage of FeFET and the modified TE scheme, the arbitrary bit-length MD metric can be implemented with ultra-low hardware cost and power consumption. Moreover, an enhanced FeFET MCAM cell, which is more suitable for the proposed modified TE scheme, is further introduced to improve the MD metric precision and reliability when expanding the bit length of MD. Based on the proposed design, the MD metric is demonstrated with 4.5×/34.8× area-efficiency and 66.7×/7.9× energy-efficiency improvements compared with conventional CMOS-based digital/CAM design, showing its great potential for MD-based applications.
Authors
- Qianqian Huang (ORCID: https://orcid.org/0000-0002-3714-8581)
- Weikai Xu (ORCID: https://orcid.org/0000-0003-0176-0469)
- Ru Huang (ORCID: https://orcid.org/0000-0002-8146-4821)
- Zeyu Zhang (ORCID: https://orcid.org/0000-0002-0464-8467)
Institutions
- Peking University (CN)
Publication Details
- Published
- 2025-03-09
- DOI
- https://doi.org/10.1109/edtm61175.2025.11040605
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00