Deep neural network (DNN) accelerators with improved energy and delay are desirable for meeting the requirements of hardware targeted for IoT and edge computing systems. Convolutional neural networks (CoNNs) belong to one of the most popular types of DNN architectures. is paper presents the design and evaluation of an accelerator for CoNNs. e system-level architecture is based on mixed-signal, cellular neural networks (CeNNs). Speciically, we present (i) the implementation of diierent layers, including convolution, ReLU, and pooling, in a CoNN using CeNN, (ii) modiied CoNN structures with CeNN-friendly layers to reduce computational overheads typically associated with a CoNN, (iii) a mixed-signal CeNN architecture that performs CoNN computations in the analog and mixed signal domain, and (iv) design space exploration that identiies what CeNN-based algorithm and architectural features fare best compared to existing algorithms and architectures when evaluated over common datasets-MNIST and CIFAR-10. Notably, the proposed approach can lead to 8.7× improvements in energy-delay product (EDP) per digit classiication for the MNIST dataset at iso-accuracy when compared with the state-of-the-art DNN engine, while our approach could ooer 4.3× improvements in EDP when compared to other network implementations for the CIFAR-10 dataset.
translated by 谷歌翻译