An Energy-Efficient Deep Convolutional Neural Network Accelerator Featuring Condi

Abstract

With its algorithmic success in many machine learning tasks and applications, deep convolutional neural net- works (DCNNs) have been implemented with custom hardware in a number of prior works. However, such works have not exploited conditional/approximate computing to the utmost toward elim- inating redundant computations of CNNs. This article presents a DCNN accelerator featuring a n ovel conditional computing scheme that synergistically combines precision cascading (PC) with zero skipping (ZS