PCM: Precision-Controlled Memory System for Energy Efficient Deep Neural Network Training

Kim, Boyeal; Lee, Sang Hyun; Kim, Hyun; Nguyen, Duy-Thanh; Le, Minh-Son; Chang, Ik Joon; Kwon, Daewoong; Yoo, Jin Hyeok; Choi, Jun Won; Lee, Hyuk-Jae

doi:10.23919/date48585.2020.9116530

Cited by 10 publications

(1 citation statement)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, computational complexity, memory usage, and corresponding power consumption are almost linearly proportional to the number of network layers [7], [8]. Consequently, CNN suffers from significant computational complexity and power consumption [9]. Especially, memory usage reduction is critical for hardware implementations because a memory access consumes 100× more energy than an arithmetic operation.…”

Section: Introductionmentioning

confidence: 99%

DC-AC: Deep Correlation-Based Adaptive Compression of Feature Map Planes in Convolutional Neural Networks

Bae

Lee

Kim

2021

2021 IEEE International Symposium on Circuits and Systems (ISCAS)

Self Cite

View full text Add to dashboard Cite

Deep learning has been successfully deployed to a broad range of applications with its outstanding performance. Supporting an efficient hardware architecture is critical to making effective use of a deep learning approach with proven algorithm performance. One challenge in implementation of deep learning algorithm is to reduce memory bandwidth because a single memory access normally consumes 100× more energy than an arithmetic operation. To reduce the memory bandwidth, deep learning data could be compressed and decompressed before memory write/read operations. Especially, feature maps, which account for a significant portion of the convolutional neural network (CNN), could be compressed further by reducing the correlations between feature map planes. This paper proposes a compression method for feature maps in CNN that adaptively exploits the varying correlation between feature map planes. For every feature map plane, the proposed method searches the most similar plane among nearby planes in the same layer, and compresses the residual of the two planes instead of the plane itself. Experimental results show that the average bit length to store feature maps is reduced by 14.2% compared to the compression without correlation reduction, and the CNN accuracy does not change and additional training is also not required because the proposed method applies lossless compression.

show abstract

Section: Introductionmentioning

confidence: 99%

DC-AC: Deep Correlation-Based Adaptive Compression of Feature Map Planes in Convolutional Neural Networks

Bae

Lee

Kim

2021

2021 IEEE International Symposium on Circuits and Systems (ISCAS)

Self Cite

View full text Add to dashboard Cite

show abstract

Recent implications towards sustainable and energy efficient AI and big data implementations in cloud-fog systems: A newsworthy inquiry

Ikhlasse¹,

Duthil

Courboulay

et al. 2022

Journal of King Saud University - Computer and Information Scie

View full text Add to dashboard Cite

Resource constrained neural network training

Pietrołaj,

Blok

2024

Sci Rep

View full text Add to dashboard Cite

Modern applications of neural-network-based AI solutions tend to move from datacenter backends to low-power edge devices. Environmental, computational, and power constraints are inevitable consequences of such a shift. Limiting the bit count of neural network parameters proved to be a valid technique for speeding up and increasing efficiency of the inference process. Hence, it is understandable that a similar approach is gaining momentum in the field of neural network training. In the face of growing complexity of neural network architectures, reducing resources required for preparation of new models would not only improve cost efficiency but also enable a variety of new AI applications on modern personal devices. In this work, we present a deep refinement of neural network parameters limitation with the use of the asymmetric exponent method. In addition to the previous research, we study new techniques of floating-point variables limitation, representation, and rounding. Moreover, by leveraging exponent offset, we present floating-point precision adjustments without an increase in variables’ bit count. The proposed method allowed us to train LeNet, AlexNet and ResNet-18 convolutional neural networks with a custom 8-bit floating-point representation achieving minimal or no results degradation in comparison to baseline 32-bit floating-point variables.

show abstract

PCM: Precision-Controlled Memory System for Energy Efficient Deep Neural Network Training

Cited by 10 publications

References 12 publications

DC-AC: Deep Correlation-Based Adaptive Compression of Feature Map Planes in Convolutional Neural Networks

DC-AC: Deep Correlation-Based Adaptive Compression of Feature Map Planes in Convolutional Neural Networks

Recent implications towards sustainable and energy efficient AI and big data implementations in cloud-fog systems: A newsworthy inquiry

Resource constrained neural network training

Contact Info

Product

Resources

About