EUVIMEDEUVIMED entdecken
Crossref

Eyelet: A Cross-Mesh NoC-Based Fine-Grained Sparse CNN Accelerator for Spatio-Temporal Parallel Computing Optimization

Bowen Yao, Liansheng Liu, Yu Peng, Xiyuan Peng, Rong Xu, Heming Liu, Haijun Zhang

IEEE Transactions on Circuits and Systems I: Regular Papers · 2025-04 · 72 (4) · 1634-1647

DOI: 10.1109/tcsi.2024.3483308 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Circuits and Systems I: Regular Papers, 2025-04. Verfasst von Bowen Yao und 6 weiteren Autor:innen. Die Arbeit wurde bisher 8-mal zitiert (Crossref) und nennt 46 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Fine-grained sparse convolutional neural networks (CNNs) achieve a better trade-off between model accuracy and size than coarse-grained sparse CNNs. Due to irregular data structures and unbalanced computation loads, fine-grained sparse CNNs struggle to fully leverage the performance advantages of computation and storage on general-purpose edge hardware. However, existing custom sparse accelerators are designed from the perspective of emulating a balanced load by software or computational strategies, neglecting the exploration of the computing architecture’s adaptability and parallelism for fine-grained sparse models. To address these challenges, a cross-mesh NoC-based accelerator architecture is proposed. This architecture aligns with the irregular characteristics of fine-grained sparse CNN weights and enhances the spatio-temporal parallelism of fine-grained sparse CNNs. First, a sparse multiplier unit (SMU) array and an adder array are designed to enable parallel execution of convolution multiplication and accumulation operations. Then, element-wise unroll-based nonzero weight multiplication is mapped to the SMU array to provide more flexible spatial parallelism. A horizontal and vertical cross-mesh NoC is proposed for flexible dataflow scheduling between the SMU and adder arrays to further improve temporal parallelism. This architecture allows the multiplication and accumulation operations in convolution to be decoupled and pipelined with negligible latency. Finally, the proposed accelerator architecture is implemented on the ZU9EG platform. The experimental results show that the proposed accelerator achieves frame rates of 509.9, 249.3, 100.7, 48.4, and 168.9 frames per second (FPS) for AlexNet, VGG-16, ResNet-18, MobileNet-v2, and EfficientNet, respectively. Compared with related works, this accelerator achieves inference speed and energy efficiency improvements of $1.1\times \sim 36.1\times $ and $2.4\times \sim 13.4\times $ , respectively.

Zitieren

APA

Yao, B., Liu, L., Peng, Y., Peng, X., Xu, R., Liu, H., & Zhang, H. (2025). Eyelet: A Cross-Mesh NoC-Based Fine-Grained Sparse CNN Accelerator for Spatio-Temporal Parallel Computing Optimization. IEEE Transactions on Circuits and Systems I: Regular Papers, 72(4), 1634-1647. https://doi.org/10.1109/tcsi.2024.3483308

Vancouver

Yao B, Liu L, Peng Y, Peng X, Xu R, Liu H, et al. Eyelet: A Cross-Mesh NoC-Based Fine-Grained Sparse CNN Accelerator for Spatio-Temporal Parallel Computing Optimization. IEEE Transactions on Circuits and Systems I: Regular Papers. 2025;72(4):1634-1647. doi:10.1109/tcsi.2024.3483308

BibTeX
@article{Yao2025,
  author = {Yao, Bowen and Liu, Liansheng and Peng, Yu and Peng, Xiyuan and Xu, Rong and Liu, Heming and Zhang, Haijun},
  title = {Eyelet: A Cross-Mesh NoC-Based Fine-Grained Sparse CNN Accelerator for Spatio-Temporal Parallel Computing Optimization},
  journal = {IEEE Transactions on Circuits and Systems I: Regular Papers},
  year = {2025},
  volume = {72},
  number = {4},
  pages = {1634--1647},
  doi = {10.1109/tcsi.2024.3483308}
}
Kostenlos starten →