Energy efficient in-memory machine learning for data intensive image-processing by non-volatile domain-wall memory

Hao Yu; Yuhao Wang; Shuai Chen; Wei Fei; Chuliang Weng; Junfeng Zhao; Zhulin Wei

doi:10.1109/ASPDAC.2014.6742888

Energy efficient in-memory machine learning for data intensive image-processing by non-volatile domain-wall memory

Hao Yu^*
, Yuhao Wang
, Shuai Chen
, Wei Fei
, Chuliang Weng
, Junfeng Zhao
, Zhulin Wei

^*Corresponding author for this work

Nanyang Technological University
Huawei Shannon Laboratory

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

28 Scopus citations

Abstract

Image processing in conventional logic-memory I/O-integrated systems will incur significant communication congestion at memory I/Os for excessive big image data at exa-scale. This paper explores an in-memory machine learning on neural network architecture by utilizing the newly introduced domain-wall nanowire, called DW-NN. We show that all operations involved in machine learning on neural network can be mapped to a logic-in-memory architecture by non-volatile domain-wall nanowire. Domain-wall nanowire based logic is customized for in machine learning within image data storage. As such, both neural network training and processing can be performed locally within the memory. The experimental results show that system throughput in DW-NN is improved by 11.6x and the energy efficiency is improved by 92x when compared to conventional image processing system.

Original language	English
Title of host publication	2014 19th Asia and South Pacific Design Automation Conference, ASP-DAC 2014 - Proceedings
Pages	191-196
Number of pages	6
DOIs	https://doi.org/10.1109/ASPDAC.2014.6742888
State	Published - 2014
Externally published	Yes
Event	2014 19th Asia and South Pacific Design Automation Conference, ASP-DAC 2014 - Suntec, Singapore Duration: 20 Jan 2014 → 23 Jan 2014

Publication series

Name	Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC

Conference

Conference	2014 19th Asia and South Pacific Design Automation Conference, ASP-DAC 2014
Country/Territory	Singapore
City	Suntec
Period	20/01/14 → 23/01/14

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 7 Affordable and Clean Energy

Access to Document

10.1109/ASPDAC.2014.6742888

Cite this

Yu, H., Wang, Y., Chen, S., Fei, W., Weng, C., Zhao, J., & Wei, Z. (2014). Energy efficient in-memory machine learning for data intensive image-processing by non-volatile domain-wall memory. In 2014 19th Asia and South Pacific Design Automation Conference, ASP-DAC 2014 - Proceedings (pp. 191-196). Article 6742888 (Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC). https://doi.org/10.1109/ASPDAC.2014.6742888

@inproceedings{6dc93cfbaffa4bda818319ac3245725a,

title = "Energy efficient in-memory machine learning for data intensive image-processing by non-volatile domain-wall memory",

abstract = "Image processing in conventional logic-memory I/O-integrated systems will incur significant communication congestion at memory I/Os for excessive big image data at exa-scale. This paper explores an in-memory machine learning on neural network architecture by utilizing the newly introduced domain-wall nanowire, called DW-NN. We show that all operations involved in machine learning on neural network can be mapped to a logic-in-memory architecture by non-volatile domain-wall nanowire. Domain-wall nanowire based logic is customized for in machine learning within image data storage. As such, both neural network training and processing can be performed locally within the memory. The experimental results show that system throughput in DW-NN is improved by 11.6x and the energy efficiency is improved by 92x when compared to conventional image processing system.",

author = "Hao Yu and Yuhao Wang and Shuai Chen and Wei Fei and Chuliang Weng and Junfeng Zhao and Zhulin Wei",

year = "2014",

doi = "10.1109/ASPDAC.2014.6742888",

language = "英语",

isbn = "9781479928163",

series = "Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC",

pages = "191--196",

booktitle = "2014 19th Asia and South Pacific Design Automation Conference, ASP-DAC 2014 - Proceedings",

note = "2014 19th Asia and South Pacific Design Automation Conference, ASP-DAC 2014 ; Conference date: 20-01-2014 Through 23-01-2014",

}

Yu, H, Wang, Y, Chen, S, Fei, W, Weng, C, Zhao, J & Wei, Z 2014, Energy efficient in-memory machine learning for data intensive image-processing by non-volatile domain-wall memory. in 2014 19th Asia and South Pacific Design Automation Conference, ASP-DAC 2014 - Proceedings., 6742888, Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC, pp. 191-196, 2014 19th Asia and South Pacific Design Automation Conference, ASP-DAC 2014, Suntec, Singapore, 20/01/14. https://doi.org/10.1109/ASPDAC.2014.6742888

Energy efficient in-memory machine learning for data intensive image-processing by non-volatile domain-wall memory. / Yu, Hao; Wang, Yuhao; Chen, Shuai et al.
2014 19th Asia and South Pacific Design Automation Conference, ASP-DAC 2014 - Proceedings. 2014. p. 191-196 6742888 (Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Energy efficient in-memory machine learning for data intensive image-processing by non-volatile domain-wall memory

AU - Yu, Hao

AU - Wang, Yuhao

AU - Chen, Shuai

AU - Fei, Wei

AU - Weng, Chuliang

AU - Zhao, Junfeng

AU - Wei, Zhulin

PY - 2014

Y1 - 2014

N2 - Image processing in conventional logic-memory I/O-integrated systems will incur significant communication congestion at memory I/Os for excessive big image data at exa-scale. This paper explores an in-memory machine learning on neural network architecture by utilizing the newly introduced domain-wall nanowire, called DW-NN. We show that all operations involved in machine learning on neural network can be mapped to a logic-in-memory architecture by non-volatile domain-wall nanowire. Domain-wall nanowire based logic is customized for in machine learning within image data storage. As such, both neural network training and processing can be performed locally within the memory. The experimental results show that system throughput in DW-NN is improved by 11.6x and the energy efficiency is improved by 92x when compared to conventional image processing system.

AB - Image processing in conventional logic-memory I/O-integrated systems will incur significant communication congestion at memory I/Os for excessive big image data at exa-scale. This paper explores an in-memory machine learning on neural network architecture by utilizing the newly introduced domain-wall nanowire, called DW-NN. We show that all operations involved in machine learning on neural network can be mapped to a logic-in-memory architecture by non-volatile domain-wall nanowire. Domain-wall nanowire based logic is customized for in machine learning within image data storage. As such, both neural network training and processing can be performed locally within the memory. The experimental results show that system throughput in DW-NN is improved by 11.6x and the energy efficiency is improved by 92x when compared to conventional image processing system.

UR - https://www.scopus.com/pages/publications/84897875079

U2 - 10.1109/ASPDAC.2014.6742888

DO - 10.1109/ASPDAC.2014.6742888

M3 - 会议稿件

AN - SCOPUS:84897875079

SN - 9781479928163

T3 - Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC

SP - 191

EP - 196

BT - 2014 19th Asia and South Pacific Design Automation Conference, ASP-DAC 2014 - Proceedings

T2 - 2014 19th Asia and South Pacific Design Automation Conference, ASP-DAC 2014

Y2 - 20 January 2014 through 23 January 2014

ER -

Yu H, Wang Y, Chen S, Fei W, Weng C, Zhao J et al. Energy efficient in-memory machine learning for data intensive image-processing by non-volatile domain-wall memory. In 2014 19th Asia and South Pacific Design Automation Conference, ASP-DAC 2014 - Proceedings. 2014. p. 191-196. 6742888. (Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC). doi: 10.1109/ASPDAC.2014.6742888

Energy efficient in-memory machine learning for data intensive image-processing by non-volatile domain-wall memory

Abstract

Publication series

Conference

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this