Research Interest

Computer systems and architecture, heterogeneous computing, mobile computing, neural network accelerators, performance characterization, power and energy efficiency

Education
  • Ph.D. in Electrical and Electronic Engineering, Yonsei University, Seoul, South Korea (Feb. 2025 (anticipated))
  • B.S. in Electrical and Electronic Engineering, Yonsei University, Seoul, South Korea (Aug. 2018)
  • Publications
  • B. Kim, C. Park, H. Kim, K. Rao, and W. J. Song, "NPUsim: Full-System, Cycle-Accurate, Value-Aware Functional Simulations of DNN Accelerators", Under revision.
  • T.Lim, H. Kim, J. Park, B. Kim, and W. Song, "RoTA: Rotational Torus Accelerator for Wear Leveling of Neural Processing Elements", Deisgn, Automation and Test in Europe Conference, Mar. 2025, To appear
  • C. Park*, B. Kim*, S. Ryu, and W. J. Song, (*equal contribution)"NeuroSpector: Systematic Optimization of Dataflow Scheduling in Deep Neural Network Accelerators", IEEE Transactions on Parallel and Distributed Computing, vol.34, no.8, Aug. 2023, pp.2279-2294
  • B. Kim, S. Lee, A. Trivedi, and W. J. Song, "Energy-Efficient Acceleration of Neural Network Applications on Edge Devices", IEEE Access, vol. 8, Nov. 2020, pp.216259-216270.
  • B. Kim, S. Lee, C. Park, H. Kim, and W. J. Song, "The Nebula Benchmark Suite: Implications of Lightweight Neural Networks", IEEE Transactions on Computers, vol. 70, Oct. 2020, pp.1887-1900.
  • H. Kim, S. Ahn, Y. Oh, B. Kim, W. Ro, and W. J. Song, "Duplo: Lifting Redundant Memory Accesses of Deep Neural Networks for GPU Tensor Cores", The 53rd IEEE/ACM International Symposium on Microarchitecture, Oct. 2020, pp.364-379.
  • Workshops
  • B. Kim, C. Park, T. Lim, and W. J. Song, "NPUsim: Full-System, Cycle-Accurate, Functional Simulations of Deep Neural Network Accelerators", Workshop on Modeling and Simulation of Systems and Applications, Oct. 2021.
  • B. Kim, S. Lee, and W. J. Song, "Nebula: Lightweight Neural Network Benchmarks", Workshop on Modeling and Simulation of Systems and Applications, Aug. 2020.
  • Poster
  • C. Park, S. Koong, B. Kim, T. Lim, and W. J. Song, "Fornax: Lightweight Energy-Efficient DNN Accelerator Architecture for Edge Devices", IEEE/ACM Design Automation Conference, Work-in-Progress (WIP), July, 2023.
  • T. Lim, H. Kim, J. Park, B. Kim, and W. J. Song "Wear Leveling of Processing Element Arrays in Deep Neural Network Accelerators", IEEE/ACM Design Automation Conference, Work-in-Progress (WIP), July, 2023.
  • Software
  • NPUsim: A Full-System NPU Microarchitecture Simulation Framework
  • NeuroSpector: Dataflow and Mapping Optimization of Deep Neural Network Accelerators.
  • Nebula: Lightweight Neural Network Benchmarks.
  • Patents
  • W. J. Song, B. Kim, C. Park, S. Koong, and T. Lim, "Deep Neural Network Accelerator For Optimized Data Processing, And Control Method Of The Deep Neural Network Accelerator,” US18/127,875, Mar., 2023.
  • W. J. Song, C. Park, B. Kim, and S. Ryu, "Neural Network Computing Device and Control Method,” US17/883,010, Aug., 2022.
  • W. J. Song, B. Kim, and S. Lee, "Operation Device of Artificial Neural Network, Operation Method of Artificial Neural Network and Computer Program Stored in A Recording Medium to Execute The Method," 10-2021-0071874, June 2021.
  • W. J. Song, W. Ro, H. Kim, S. Ahn, Y. Oh, and B. Kim, "Operation Device of Convolutional Neural Network, Operation Method of Convolutional Neural Network and Computer Program Stored in Recording Medium to Execute The Method," 10-2021-0066176, May 2021.
  • Scholarship

    Ph.D. Fellowship
    National Research Foundation of Korea (June 2021 - May 2023)

    Contact

    50 Yonsei-ro Eng-B705, Seoul, Seodaemun-gu 03722, South Korea