TY - JOUR
T1 - HEIF
T2 - Highly Efficient Stochastic Computing-Based Inference Framework for Deep Neural Networks
AU - Li, Zhe
AU - Li, Ji
AU - Ren, Ao
AU - Cai, Ruizhe
AU - Ding, Caiwen
AU - Qian, Xuehai
AU - Draper, Jeffrey
AU - Yuan, Bo
AU - Tang, Jian
AU - Qiu, Qinru
AU - Wang, Yanzhi
N1 - Publisher Copyright:
© 1982-2012 IEEE.
PY - 2019/8
Y1 - 2019/8
N2 - Deep convolutional neural networks (DCNNs) are one of the most promising deep learning techniques and have been recognized as the dominant approach for almost all recognition and detection tasks. The computation of DCNNs is memory intensive due to large feature maps and neuron connections, and the performance highly depends on the capability of hardware resources. With the recent trend of wearable devices and Internet of Things, it becomes desirable to integrate the DCNNs onto embedded and portable devices that require low power and energy consumptions and small hardware footprints. Recently stochastic computing (SC)-DCNN demonstrated that SC as a low-cost substitute to binary-based computing radically simplifies the hardware implementation of arithmetic units and has the potential to satisfy the stringent power requirements in embedded devices. In SC, many arithmetic operations that are resource-consuming in binary designs can be implemented with very simple hardware logic, alleviating the extensive computational complexity. It offers a colossal design space for integration and optimization due to its reduced area and soft error resiliency. In this paper, we present HEIF, a highly efficient SC-based inference framework of the large-scale DCNNs, with broad applications including (but not limited to) LeNet-5 and AlexNet, that achieves high energy efficiency and low area/hardware cost. Compared to SC-DCNN, HEIF features: 1) the first (to the best of our knowledge) SC-based rectified linear unit activation function to catch up with the recent advances in software models and mitigate degradation in application-level accuracy; 2) the redesigned approximate parallel counter and optimized stochastic multiplication using transmission gates and inverse mirror adders; and 3) the new optimization of weight storage using clustering. Most importantly, to achieve maximum energy efficiency while maintaining acceptable accuracy, HEIF considers holistic optimizations on cascade connection of function blocks in DCNN, pipelining technique, and bit-stream length reduction. Experimental results show that in large-scale applications HEIF outperforms previous SC-DCNN by the throughput of 4.1 ×, by area efficiency of up to 6.5 ×, and achieves up to 5.6 × energy improvement.
AB - Deep convolutional neural networks (DCNNs) are one of the most promising deep learning techniques and have been recognized as the dominant approach for almost all recognition and detection tasks. The computation of DCNNs is memory intensive due to large feature maps and neuron connections, and the performance highly depends on the capability of hardware resources. With the recent trend of wearable devices and Internet of Things, it becomes desirable to integrate the DCNNs onto embedded and portable devices that require low power and energy consumptions and small hardware footprints. Recently stochastic computing (SC)-DCNN demonstrated that SC as a low-cost substitute to binary-based computing radically simplifies the hardware implementation of arithmetic units and has the potential to satisfy the stringent power requirements in embedded devices. In SC, many arithmetic operations that are resource-consuming in binary designs can be implemented with very simple hardware logic, alleviating the extensive computational complexity. It offers a colossal design space for integration and optimization due to its reduced area and soft error resiliency. In this paper, we present HEIF, a highly efficient SC-based inference framework of the large-scale DCNNs, with broad applications including (but not limited to) LeNet-5 and AlexNet, that achieves high energy efficiency and low area/hardware cost. Compared to SC-DCNN, HEIF features: 1) the first (to the best of our knowledge) SC-based rectified linear unit activation function to catch up with the recent advances in software models and mitigate degradation in application-level accuracy; 2) the redesigned approximate parallel counter and optimized stochastic multiplication using transmission gates and inverse mirror adders; and 3) the new optimization of weight storage using clustering. Most importantly, to achieve maximum energy efficiency while maintaining acceptable accuracy, HEIF considers holistic optimizations on cascade connection of function blocks in DCNN, pipelining technique, and bit-stream length reduction. Experimental results show that in large-scale applications HEIF outperforms previous SC-DCNN by the throughput of 4.1 ×, by area efficiency of up to 6.5 ×, and achieves up to 5.6 × energy improvement.
KW - ASIC
KW - convolutional neural network
KW - deep learning
KW - energy-efficient
KW - optimization
KW - stochastic computing (SC)
UR - http://www.scopus.com/inward/record.url?scp=85049489419&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85049489419&partnerID=8YFLogxK
U2 - 10.1109/TCAD.2018.2852752
DO - 10.1109/TCAD.2018.2852752
M3 - Article
AN - SCOPUS:85049489419
SN - 0278-0070
VL - 38
SP - 1543
EP - 1556
JO - IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
JF - IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
IS - 8
M1 - 8403283
ER -