A 40-nm 1.89-pJ/SOP Scalable Convolutional Spiking Neural Network Learning Core With On-Chip Spatiotemporal Back-Propagation
Pai-Yu Tan, Cheng‐Wen Wu
Abstract
In recent years, progress in spiking neural network (SNN) research has generated growing interest in specialized SNN hardware. However, most of the hardware studies are about inference-only engines, and the training process for low-power endpoint devices remains arduous due to high numerical precision requirements. In this article, we introduce a scalable convolutional SNN learning core for energy-efficient training utilizing the spatiotemporal back-propagation (STBP) algorithm. We modify the STBP algorithm with five hardware-based enhancing methods, which minimize the hardware implementation cost without losing accuracy. We propose a unified core architecture encompassing three data-flow modes for three training phases. It also offers multicore scalability for wider and deeper models. A 40-nm prototype chip has been implemented, achieving peak training and inference efficiencies of 3 and 7.7 TOPS/W, respectively, at 90% input sparsity and an energy per synaptic operation (SOP) metric of 1.89 pJ/SOP. Based on the chip, our multicore prototype system attains a competitive accuracy of 99.1% on MNIST. We also exhibit the first on-chip training results on SVHN and CIFAR10, with an average of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$36\times $ </tex-math></inline-formula> improvement in efficiency compared with a typical commercial GPU platform.