CENNA: Cost-Effective Neural Network Accelerator

Convolutional neural networks (CNNs) are widely adopted in various applications. State-of-the-art CNN models deliver excellent classification performance, but they require a large amount of computation and data exchange because they typically employ many processing layers. Among these processing layers, convolution layers, which carry out many multiplications and additions, account for a major portion of computation and memory access. Therefore, reducing the amount of computation and memory access is the key for high-performance CNNs. In this study, we propose a cost-effective neural network accelerator, named CENNA, whose hardware cost is reduced by employing a cost-centric matrix multiplication that employs both Strassen’s multiplication and a naive multiplication. Furthermore, the convolution method using the proposed matrix multiplication can minimize data movement by reusing both the feature map and the convolution kernel without any additional control logic. In terms of throughput, power consumption, and silicon area, the efficiency of CENNA is up to 88 times higher than that of conventional designs for the CNN inference.

[1]  Dajiang Zhou,et al.  Chain-NN: An energy-efficient 1D chain architecture for accelerating deep convolutional neural networks , 2017, Design, Automation & Test in Europe Conference & Exhibition (DATE), 2017.

[2]  John Sartori,et al.  Power balanced pipelines , 2012, IEEE International Symposium on High-Performance Comp Architecture.

[3]  Joel Emer,et al.  Eyeriss: a spatial architecture for energy-efficient dataflow for convolutional neural networks , 2016, CARN.

[4]  Eugenio Culurciello,et al.  An Analysis of Deep Neural Network Models for Practical Applications , 2016, ArXiv.

[5]  Kaiming He,et al.  Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[6]  Jason Cong,et al.  Minimizing Computation in Convolutional Neural Networks , 2014, ICANN.

[7]  Hyoukjun Kwon,et al.  MAERI: Enabling Flexible Dataflow Mapping over DNN Accelerators via Reconfigurable Interconnects , 2018, ASPLOS.

[8]  Marian Verhelst,et al.  An Energy-Efficient Precision-Scalable ConvNet Processor in 40-nm CMOS , 2017, IEEE Journal of Solid-State Circuits.

[9]  Pilsung Kang,et al.  Sentiment Classification with Word Attention based on Weakly Supervised Learning with a Convolutional Neural Network , 2017, ArXiv.

[10]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[11]  Shuang Wu,et al.  Training and Inference with Integers in Deep Neural Networks , 2018, ICLR.

[12]  Soheil Ghiasi,et al.  Fast and Energy-Efficient CNN Inference on IoT Devices , 2016, ArXiv.

[13]  Kevin Clark Hamilton.  Optimization of energy and throughput for pipelined VLSI interconnect , 2010 .

[14]  Luca Benini,et al.  Origami: A 803-GOp/s/W Convolutional Network Accelerator , 2015, IEEE Transactions on Circuits and Systems for Video Technology.

[15]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[16]  Rita Cucchiara,et al.  DARPA benchmark image processing on SIMD parallel machines , 1996, Proceedings of 1996 IEEE Second International Conference on Algorithms and Architectures for Parallel Processing, ICA/sup 3/PP '96.

[17]  Ameet Talwalkar,et al.  Divide-and-Conquer Matrix Factorization , 2011, NIPS.

[18]  Kiyoung Choi,et al.  SoC Architecture for Automobile Vision System , 2014 .

[19]  Christos-Savvas Bouganis,et al.  DroNet: Efficient convolutional neural network detector for real-time UAV applications , 2018, 2018 Design, Automation & Test in Europe Conference & Exhibition (DATE).

[20]  Michael Gschwind,et al.  Integrated analysis of power and performance for pipelined microprocessors , 2004, IEEE Transactions on Computers.

[21]  S. K. Nandy,et al.  Accelerating BLAS on Custom Architecture through Algorithm-Architecture Co-design , 2016, ArXiv.

[22]  Tian Guo,et al.  Cloud-Based or On-Device: An Empirical Study of Mobile Deep Inference , 2017, 2018 IEEE International Conference on Cloud Engineering (IC2E).

[23]  Andrew Lavin,et al.  Fast Algorithms for Convolutional Neural Networks , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[24]  Kilian Q. Weinberger,et al.  Densely Connected Convolutional Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[25]  Weisong Shi,et al.  PI-Edge: A Low-Power Edge Computing System for Real-Time Autonomous Driving Services , 2018, ArXiv.

[26]  Aren Jansen,et al.  CNN architectures for large-scale audio classification , 2016, 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[27]  Yu Wang,et al.  HARDWARE-FRIENDLY CONVOLUTIONAL NEURAL NETWORK WITH EVEN-NUMBER FILTER SIZE , 2016 .

[28]  Gerald Penn,et al.  Convolutional Neural Networks for Speech Recognition , 2014, IEEE/ACM Transactions on Audio, Speech, and Language Processing.

[29]  Shuang Wu,et al.  Convolution with even-sized kernels and symmetric padding , 2019, NeurIPS.

[30]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[31]  Yoshua Bengio,et al.  Gradient-based learning applied to document recognition , 1998, Proc. IEEE.