A 65-nm Energy-Efficient Interframe Data Reuse Neural Network Accelerator for Video Applications
Yixiong Yang, Yongpan Liu, Zhe Yuan, Wenyu Sun, Ruoyang Liu, Jingyu Wang, Jinshan Yue, Xiaoyu Feng, Zhuqing Yuan, Xueqing Li, Huazhong Yang
Abstract
An energy-efficient convolutional neural network (CNN) accelerator is proposed for the video application. Previous works exploited the sparsity of differential (Diff) frame activation, but the improvement is limited as many Diff-frame data is small but non-zero. Processing of irregular sparse data also leads to low hardware utilization. To solve these problems, two key innovations are proposed in this article. First, we implement a hybrid-precision inter-frame-reuse architecture which takes advantage of both low bit-width and high sparsity of Diff-frame data. This technology can accelerate 3.2 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\times $ </tex-math></inline-formula> inference speed with no accuracy loss. Second, we design a conv-pattern-aware processing array that achieves the 2.48 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\times $ </tex-math></inline-formula> –14.2 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\times $ </tex-math></inline-formula> PE utilization rate to process sparse data for different convolution kernels. The accelerator chip was implemented in 65-nm CMOS technology. To the best of our knowledge, it is the first silicon-proven CNN accelerator that supports inter-frame data reuse. Attributed to the inter-frame similarity, this video CNN accelerator reaches the minimum energy consumption of 24.7 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu \text{J}$ </tex-math></inline-formula> /frame in the MobileNet-slim model, which is 76.3% less than the baseline.