Eyeriss: An Energy-Efficient Reconfigurable Accelerator for Deep Convolutional Neural Networks

📋 论文概要

该论文提出了一种名为Eyeriss的能量可重构加速器，用于高效处理深度卷积神经网络（CNN）中的卷积运算，解决了CNN计算中高能耗和低能效的问题。通过优化数据流和架构设计，Eyeriss在保持高吞吐量的同时显著降低了能耗。

💡 主要创新点

• 提出了一种新颖的行静态数据流（Row-Stationary Dataflow），通过最小化数据移动来降低能耗。
• 设计了一个可重构的加速器架构，支持不同形状和规模的CNN层，提高了灵活性。
• 实现了片上存储层次和计算单元的协同优化，进一步提升了能效。

🏷 关键词

📄 原文摘要

Nvidia, Westford, MA 1 2 Deep learning using convolutional neural networks (CNN) gives state-of-the-art accuracy on many computer vision tasks (e.g. object detection, recognition, segmentation). Convolutions account for over 90% of the processing in CNNs for both inference/testing and training, and fully convolutional networks are increasingly being used. To achieve state-of-the-art accuracy requires CNNs with not only a larger number of layers, but also millions of filters weights, and varying shapes (i.e. filter sizes, number of filters, number of channels) as shown in Fig. 14.5.1. For instance, AlexNet [1] uses 2.3 million weights (4.6MB of storage) and requires 666 million MACs per 227×227 image (13kMACs/pixel). VGG16 [2] uses 14.7 million weights (29.4MB of storage) and requires 15.3 billion MACs per 224×224 image (306kMACs/pixel). The large number of filter weights and channels results in substantial data movement, which consumes significant

👥 作者与机构

Yu-Hsin Chen1, Tushar Krishna1, Joel Emer1,2, Vivienne Sze1

Massachusetts Institute of Technology, Cambridge, MA,