English

Reuse Kernels or Activations? A Flexible Dataflow for Low-latency Spectral CNN Acceleration

Hardware Architecture 2023-10-18 v1 Signal Processing

Abstract

Spectral-domain CNNs have been shown to be more efficient than traditional spatial CNNs in terms of reducing computation complexity. However they come with a `kernel explosion' problem that, even after compression (pruning), imposes a high memory burden and off-chip bandwidth requirement for kernel access. This creates a performance gap between the potential acceleration offered by compression and actual FPGA implementation performance, especially for low-latency CNN inference. In this paper, we develop a principled approach to overcoming this performance gap and designing a low-latency, low-bandwidth, spectral sparse CNN accelerator on FPGAs. First, we analyze the bandwidth-storage tradeoff of sparse convolutional layers and locate communication bottlenecks. We then develop a dataflow for flexibly optimizing data reuse in different layers to minimize off-chip communication. Finally, we propose a novel scheduling algorithm to optimally schedule the on-chip memory access of multiple sparse kernels and minimize read conflicts. On a state-of-the-art FPGA platform, our design reduces data transfers by 42\% with DSP utilization up to 90\% and achieves inference latency of 9 ms for VGG16, compared to the baseline state-of-the-art latency of 68 ms.

Keywords

Cite

@article{arxiv.2310.10902,
  title  = {Reuse Kernels or Activations? A Flexible Dataflow for Low-latency Spectral CNN Acceleration},
  author = {Yue Niu and Rajgopal Kannan and Ajitesh Srivastava and Viktor Prasanna},
  journal= {arXiv preprint arXiv:2310.10902},
  year   = {2023}
}

Comments

11 pages, 11 figures Accepted to ACM/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA) 2020

R2 v1 2026-06-28T12:52:47.043Z