EUVIMEDEUVIMED entdecken
Crossref

A 28-nm 3.32-nJ/Frame Compute-in-Memory CNN Processor With Layer Fusion for Always-on Applications

Yuanzhe Zhao, Pengyu He, Yan Zhu, Rui P. Martins, Chi-Hang Chan, Minglei Zhang

IEEE Transactions on Circuits and Systems I: Regular Papers · 2025-10 · 72 (10) · 5683-5695

DOI: 10.1109/tcsi.2025.3542590 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Circuits and Systems I: Regular Papers, 2025-10. Verfasst von Yuanzhe Zhao und 5 weiteren Autor:innen. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 36 Literaturangaben. Gefördert von: National Key Research and Development Program, Macau Science and Technology Development Fund, University of Macau Research.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This work presents an always-on CNN processor featuring compute-in-memory (CIM) and layer-fusion (LF) techniques. It demonstrates an end-to-end neural network (NN) inference while eliminating memory accesses for both the weight and inter-layer activation, thus significantly saving associated energy. The analog LF-CIM processor combines charge- and time-domain multiply-accumulate (MAC) arrays in a fusional manner, avoiding the need for the interfacial analog-to-digital and digital-to-analog converters; it also supports a highly sparsity-adaptive network and exhibits low-voltage-supplied tolerance to circuit noise and variations, which facilitates an energy-efficient and robust CNN processor, simultaneously. Furthermore, the full-precision activation originating from the LF reduces the NN parameters and operations, while the inference accuracy is improved by the presented ensemble NN. The prototype processor is fabricated in a 28-nm CMOS process, demonstrating a fully on-chip MNIST inference with 97.9% accuracy. It operates at 3,508 frames per second while consuming $11.6~\mu $ W at a 0.5-V supply; the achieved efficiency of 3.32 nJ/frame is over 50-fold than the state-of-the-art end-to-end MNIST accelerators.

Förderung

  • National Key Research and Development Program
  • Macau Science and Technology Development Fund
  • University of Macau Research

Zitieren

APA

Zhao, Y., He, P., Zhu, Y., Martins, R. P., Chan, C., & Zhang, M. (2025). A 28-nm 3.32-nJ/Frame Compute-in-Memory CNN Processor With Layer Fusion for Always-on Applications. IEEE Transactions on Circuits and Systems I: Regular Papers, 72(10), 5683-5695. https://doi.org/10.1109/tcsi.2025.3542590

Vancouver

Zhao Y, He P, Zhu Y, Martins RP, Chan C, Zhang M. A 28-nm 3.32-nJ/Frame Compute-in-Memory CNN Processor With Layer Fusion for Always-on Applications. IEEE Transactions on Circuits and Systems I: Regular Papers. 2025;72(10):5683-5695. doi:10.1109/tcsi.2025.3542590

BibTeX
@article{Zhao2025,
  author = {Zhao, Yuanzhe and He, Pengyu and Zhu, Yan and Martins, Rui P. and Chan, Chi-Hang and Zhang, Minglei},
  title = {A 28-nm 3.32-nJ/Frame Compute-in-Memory CNN Processor With Layer Fusion for Always-on Applications},
  journal = {IEEE Transactions on Circuits and Systems I: Regular Papers},
  year = {2025},
  volume = {72},
  number = {10},
  pages = {5683--5695},
  doi = {10.1109/tcsi.2025.3542590}
}
Kostenlos starten →