EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Memristor‐Based In‐Memory Computing System‐on‐Chip with Efficient Depthwise Convolution

Wenhao Song, Fuxi Cai, Sangsu Park, Zhaoyan Li, Seonghun Lee, Mingyi Rao, Woohyung Choi, Seungin Shin, Yan Zhu, Sujin Chae, Myoungsub Kim, Mingche Wu und 10 weitere

Advanced Intelligent Systems · 2026-09 · 8 (9)

DOI: 10.1002/aisy.202501225 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Advanced Intelligent Systems, online erschienen am 04.06.2026. Verfasst von Wenhao Song und 21 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 20 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In‐memory computing (IMC) using memristor crossbar arrays offers a promising approach to accelerating neural networks by performing analog computation directly within memory, thereby minimizing data movement and enhancing energy efficiency. While prior work has focused primarily on dense neural operations, depthwise convolution (DWC), a key operation in lightweight models like MobileNets, poses unique challenges due to its limited cross‐channel data reuse and irregular access patterns. This work proposes a novel IMC architecture tailored for efficient DWC acceleration using memristors. Our design features a zig–zag selection‐line topology that significantly improves memory utilization compared to conventional 1T1R crossbars while preserving dense operations’ throughput and energy efficiency. We fabricate a system‐on‐chip (SoC) implementing the proposed architecture and experimentally validate its performance. Fabricated in a 65 nm complementary metal–oxide–semiconductor process, the SoC delivers a measured energy efficiency of 21.3 TOPS/W at 100 MHz and lifts the per‐array weight utilization of depthwise layers to ∼100%. End‐to‐end deployment of a tailored MobileNetV1 on the visual‐wake‐words dataset reaches an inference accuracy of 80.36%, on par with the same software model quantized to 4‐bit precision. This first hardware demonstration of depthwise‐separable convolution on a memristor‐based IMC SoC establishes the readiness of memristor‐based IMC for next‐generation edge‐artificial intelligence accelerators.

Zitieren

APA

Song, W., Cai, F., Park, S., Li, Z., Lee, S., Rao, M., Choi, W., Shin, S., Zhu, Y., Chae, S., Kim, M., Wu, M., Choi, H., Dao, Y., Yi, J., Cho, Y., Cha, S. Y., Ge, N., Xia, Q., …, & Yang, J. J. (2026). A Memristor‐Based In‐Memory Computing System‐on‐Chip with Efficient Depthwise Convolution. Advanced Intelligent Systems, 8(9). https://doi.org/10.1002/aisy.202501225

Vancouver

Song W, Cai F, Park S, Li Z, Lee S, Rao M, et al. A Memristor‐Based In‐Memory Computing System‐on‐Chip with Efficient Depthwise Convolution. Advanced Intelligent Systems. 2026;8(9). doi:10.1002/aisy.202501225

BibTeX
@article{Song2026,
  author = {Song, Wenhao and Cai, Fuxi and Park, Sangsu and Li, Zhaoyan and Lee, Seonghun and Rao, Mingyi and Choi, Woohyung and Shin, Seungin and Zhu, Yan and Chae, Sujin and Kim, Myoungsub and Wu, Mingche and Choi, Hyejung and Dao, Yuan and Yi, Jaeyun and Cho, Youngman and Cha, Seon Yong and Ge, Ning and Xia, Qiangfei and Kim, Soo Gil and Hu, Miao and Yang, J. Joshua},
  title = {A Memristor‐Based In‐Memory Computing System‐on‐Chip with Efficient Depthwise Convolution},
  journal = {Advanced Intelligent Systems},
  year = {2026},
  volume = {8},
  number = {9},
  doi = {10.1002/aisy.202501225}
}
Kostenlos starten →