Silicon photonics convolution accelerator based on coherent chips with sub-1 pJ/MAC power consumption
Photonics Research · 01.02.2025 · 13 (2) · 497
DOI: 10.1364/prj.536939 ↗Kurz gefasst
Fachartikel in der Zeitschrift Photonics Research, online erschienen am 31.01.2025. Verfasst von Ying Zhu und 14 weiteren Autor:innen. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 60 Literaturangaben. Gefördert von: National Natural Science Foundation of China.
Abstract (Originalsprache)
Artificial intelligence (AI), owing to its substantial computing demands, necessitates computing hardware that offers both high speed and high power efficiency. A silicon photonic integrated circuit shows promise as a hardware solution due to its attributes, including high power efficiency, low latency, large bandwidth, and complementary metal–oxide–semiconductor (CMOS) compatibility. Here, we propose a silicon photonic convolution accelerator (SiPh-CA) and experimentally realize a prototype with sub-integrated coherent transmit–receive optical sub-assemblies (sub-IC-TROSAs). The prototype, compared to a previous IC-TROSA-based convolution accelerator, achieves almost the same performances of 1.024 TOPS/channel and 96.22% inference accuracy when it processes neural networks for image recognition, using half the numbers of the modulators and the drivers with which over 1/3 chip footprint and 37.01% power consumption are reduced. By incorporating a broadcasting scheme based on splitters and combiners, the approach can efficiently process multiple convolutions in parallel, achieving several tera operations per second. This scalability feature allows the SiPh-CA to process complex AI and high-performance computing tasks.
Förderung
- National Natural Science Foundation of China
Zitieren
APA
Zhu, Y., Xu, L., Hua, X., Liu, K., Liu, Y., Luo, M., Liu, J., Dang, Z., Liu, Y., Liu, M., Zhang, H., Chen, D., Wang, L., Xiao, X., & Yu, S. (2025). Silicon photonics convolution accelerator based on coherent chips with sub-1 pJ/MAC power consumption. Photonics Research, 13(2), 497. https://doi.org/10.1364/prj.536939
Vancouver
Zhu Y, Xu L, Hua X, Liu K, Liu Y, Luo M, et al. Silicon photonics convolution accelerator based on coherent chips with sub-1 pJ/MAC power consumption. Photonics Research. 2025;13(2):497. doi:10.1364/prj.536939
BibTeX
@article{Zhu2025,
author = {Zhu, Ying and Xu, Lu and Hua, Xin and Liu, Kailai and Liu, Yifan and Luo, Ming and Liu, Jia and Dang, Ziyue and Liu, Ye and Liu, Min and Zhang, Hongguang and Chen, Daigao and Wang, Lei and Xiao, Xi and Yu, Shaohua},
title = {Silicon photonics convolution accelerator based on coherent chips with sub-1 pJ/MAC power consumption},
journal = {Photonics Research},
year = {2025},
volume = {13},
number = {2},
pages = {497},
doi = {10.1364/prj.536939}
}
EUVIMED