Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Chinese Journal of Lasers, Volume. 49, Issue 12, 1219001(2022)

Advances and Challenges of Optoelectronic Intelligent Computing

Junwei Cheng1, Xueyi Jiang1, Hailong Zhou1, and Jianji Dong1,2、*
Author Affiliations
  • 1Wuhan National Laboratory for Optoelectronics, Huazhong University of Science and Technology, Wuhan 430074, Hubei, China
  • 2Optics Valley Laboratory, Wuhan 430074, Hubei, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (8)
    Equations (0)
    References (89)
    Cited By (11)
    Tools
    Paper Information
    Topics
    Figures & Tables(8)
    Categories of photonic MVM. (a) PLC-based photonic MVM; (b) MZI-based photonic MVM; (c) WDM-based photonic MVM

    Fig. 1. Categories of photonic MVM. (a) PLC-based photonic MVM; (b) MZI-based photonic MVM; (c) WDM-based photonic MVM

    Download full size
    Summary of applications in optoelectronic intelligent computing

    Fig. 2. Summary of applications in optoelectronic intelligent computing

    Download full size
    In situ training of optical neural networks through BP algorithm. (a) Chip of integrated optical neural networks[77]; (b) forward propagation of diffractive optical neural networks[78]; (c) backward propagation of diffractive optical neural networks[78]

    Fig. 3. In situ training of optical neural networks through BP algorithm. (a) Chip of integrated optical neural networks[77]; (b) forward propagation of diffractive optical neural networks[78]; (c) backward propagation of diffractive optical neural networks[78]

    Download full size
    Online training of optoelectronic intelligent computing chip through SGD algorithm. (a) Brief flow chart of SGD algorithm; (b) multifunctional on-chip polarization processor[42]; (c) photonic accelerator for Google PageRank algorithm[43]; (d) self-configuring and fully reconfigurable silicon photonic signal processor[44]

    Fig. 4. Online training of optoelectronic intelligent computing chip through SGD algorithm. (a) Brief flow chart of SGD algorithm; (b) multifunctional on-chip polarization processor[42]; (c) photonic accelerator for Google PageRank algorithm[43]; (d) self-configuring and fully reconfigurable silicon photonic signal processor[44]

    Download full size
    Three typical optoelectronic intelligent computing architectures. (a) Coherent MZI mesh[40]; (b) photonic accelerator based on time-wavelength interleaving[62]; (c) integrated photonic tensor core based on PCM[61]

    Fig. 5. Three typical optoelectronic intelligent computing architectures. (a) Coherent MZI mesh[40]; (b) photonic accelerator based on time-wavelength interleaving[62]; (c) integrated photonic tensor core based on PCM[61]

    Download full size
    Factors influencing computing capacity and energy consumption

    Fig. 6. Factors influencing computing capacity and energy consumption

    Download full size

    • Table 1. Comparison of computing capacity and energy efficiency of microelectronic chips and optoelectronic chips

      View table

      Table 1. Comparison of computing capacity and energy efficiency of microelectronic chips and optoelectronic chips

      TechnologyComputing capacity /(1012 operation/s)Energy efficiency
      Coherent MZI mesh[40]3.230 fJ per MAC
      Time-wavelength interleaving photonic convolutional accelerator[62]110.39 fJ per MAC
      Photonic WDM/PCM in-memory computing[61]4.30217 fJ per MAC
      Google TPU[79]230.43 pJ per MAC
      NVIDIA Tesla T4[80]1301.08 pJ per MAC
      HUAWEI Ascend 310[81]161 pJ per MAC
      HUAWEI Ascend 910[82]6401.09 pJ per MAC

    • Table 2. Computing capacity and energy efficiency of different optoelectronic computing architectures

      View table

      Table 2. Computing capacity and energy efficiency of different optoelectronic computing architectures

      PerformanceCoherent MZI meshPhotonic accelerator based on time-wavelength interleavingIntegrated photonic tensor core based on PCM
      Formula for computing capacitym×2×N2×10112R×1τ2×M×mo
      Computing capacity6.4×1012 operation/s11×1012 operation/s4.302×1012 operation/s
      Formula for energy efficiencyee＝cc/poee＝cc/poee＝cc/po
    Tools

    Get Citation

    Copy Citation Text

    Junwei Cheng, Xueyi Jiang, Hailong Zhou, Jianji Dong. Advances and Challenges of Optoelectronic Intelligent Computing[J]. Chinese Journal of Lasers, 2022, 49(12): 1219001

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Set citation alerts for article
    Save article for my favorites
    Paper Information

    Received: Feb. 14, 2022

    Accepted: Apr. 24, 2022

    Published Online: Jun. 13, 2022

    The Author Email: Jianji Dong (jjdong@hust.edu.cn)

    DOI:10.3788/CJL202249.1219001

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology