Design space exploration of magnetic tunnel junction based stochastic computing in deep learning

You Wang; Yue Zhang; Youguang Zhang; Weisheng Zhao; Hao Cai; Lirida Naviner

doi:10.1145/3194554.3194619

Design space exploration of magnetic tunnel junction based stochastic computing in deep learning

You Wang
, Yue Zhang
, Youguang Zhang
, Weisheng Zhao
, Hao Cai
, Lirida Naviner

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

2 Scopus citations

Abstract

Magnetic tunnel junction (MTJ) is considered as a promising memory candidate in the more than Moore era because of high power efficiency, fast access speed, nearly infinite endurance and easy 3D integration. The nondeterministic switching behavior has been profited to exploit new directions for computing methods, such as stochastic computing. In this paper, the application of stochastic switching behavior in stochastic computing is explored for deep neural network (DNN). Stochastic computing method features low logic complexity, low energy consumption and fine-grained parallelism, boosting the performance of DNN system by combining MTJ. As a key block of stochastic computing, MTJ based true random number generator design is presented in details. The functionality has been validated by combining the hardware design and post-processing in software. Simulation results are demonstrated visibly by handwritten digits recognition test to show the accuracy. Furthermore, the performance is investigated in terms of accuracy, energy consumption and memory occupation to find more efficient techniques.

Original language	English
Title of host publication	GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI
Publisher	Association for Computing Machinery
Pages	403-408
Number of pages	6
ISBN (Electronic)	9781450357241
DOIs	https://doi.org/10.1145/3194554.3194619
State	Published - 30 May 2018
Event	28th Great Lakes Symposium on VLSI, GLSVLSI 2018 - Chicago, United States Duration: 23 May 2018 → 25 May 2018

Publication series

Name	Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI

Conference

Conference	28th Great Lakes Symposium on VLSI, GLSVLSI 2018
Country/Territory	United States
City	Chicago
Period	23/05/18 → 25/05/18

UN SDGs

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

SDG 7 Affordable and Clean Energy

Keywords

Deep neural network
Magnetic tunnel junction
Stochastic computing
True random number generator

Access to Document

10.1145/3194554.3194619

Cite this

Wang, Y., Zhang, Y., Zhang, Y., Zhao, W., Cai, H., & Naviner, L. (2018). Design space exploration of magnetic tunnel junction based stochastic computing in deep learning. In GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI (pp. 403-408). (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI). Association for Computing Machinery . https://doi.org/10.1145/3194554.3194619

@inproceedings{533050f8a52d4dc9ad2cc3c31a034624,

title = "Design space exploration of magnetic tunnel junction based stochastic computing in deep learning",

abstract = "Magnetic tunnel junction (MTJ) is considered as a promising memory candidate in the more than Moore era because of high power efficiency, fast access speed, nearly infinite endurance and easy 3D integration. The nondeterministic switching behavior has been profited to exploit new directions for computing methods, such as stochastic computing. In this paper, the application of stochastic switching behavior in stochastic computing is explored for deep neural network (DNN). Stochastic computing method features low logic complexity, low energy consumption and fine-grained parallelism, boosting the performance of DNN system by combining MTJ. As a key block of stochastic computing, MTJ based true random number generator design is presented in details. The functionality has been validated by combining the hardware design and post-processing in software. Simulation results are demonstrated visibly by handwritten digits recognition test to show the accuracy. Furthermore, the performance is investigated in terms of accuracy, energy consumption and memory occupation to find more efficient techniques.",

keywords = "Deep neural network, Magnetic tunnel junction, Stochastic computing, True random number generator",

author = "You Wang and Yue Zhang and Youguang Zhang and Weisheng Zhao and Hao Cai and Lirida Naviner",

note = "Publisher Copyright: {\textcopyright} 2018 Association for Computing Machinery.; 28th Great Lakes Symposium on VLSI, GLSVLSI 2018 ; Conference date: 23-05-2018 Through 25-05-2018",

year = "2018",

month = may,

day = "30",

doi = "10.1145/3194554.3194619",

language = "英语",

series = "Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI",

publisher = "Association for Computing Machinery ",

pages = "403--408",

booktitle = "GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI",

address = "美国",

}

Wang, Y, Zhang, Y , Zhang, Y , Zhao, W, Cai, H & Naviner, L 2018, Design space exploration of magnetic tunnel junction based stochastic computing in deep learning. in GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI. Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI, Association for Computing Machinery , pp. 403-408, 28th Great Lakes Symposium on VLSI, GLSVLSI 2018, Chicago, United States, 23/05/18. https://doi.org/10.1145/3194554.3194619

Design space exploration of magnetic tunnel junction based stochastic computing in deep learning. / Wang, You; Zhang, Yue ; Zhang, Youguang et al.
GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI. Association for Computing Machinery , 2018. p. 403-408 (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI).

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

TY - GEN

T1 - Design space exploration of magnetic tunnel junction based stochastic computing in deep learning

AU - Wang, You

AU - Zhang, Yue

AU - Zhang, Youguang

AU - Zhao, Weisheng

AU - Cai, Hao

AU - Naviner, Lirida

PY - 2018/5/30

Y1 - 2018/5/30

N2 - Magnetic tunnel junction (MTJ) is considered as a promising memory candidate in the more than Moore era because of high power efficiency, fast access speed, nearly infinite endurance and easy 3D integration. The nondeterministic switching behavior has been profited to exploit new directions for computing methods, such as stochastic computing. In this paper, the application of stochastic switching behavior in stochastic computing is explored for deep neural network (DNN). Stochastic computing method features low logic complexity, low energy consumption and fine-grained parallelism, boosting the performance of DNN system by combining MTJ. As a key block of stochastic computing, MTJ based true random number generator design is presented in details. The functionality has been validated by combining the hardware design and post-processing in software. Simulation results are demonstrated visibly by handwritten digits recognition test to show the accuracy. Furthermore, the performance is investigated in terms of accuracy, energy consumption and memory occupation to find more efficient techniques.

AB - Magnetic tunnel junction (MTJ) is considered as a promising memory candidate in the more than Moore era because of high power efficiency, fast access speed, nearly infinite endurance and easy 3D integration. The nondeterministic switching behavior has been profited to exploit new directions for computing methods, such as stochastic computing. In this paper, the application of stochastic switching behavior in stochastic computing is explored for deep neural network (DNN). Stochastic computing method features low logic complexity, low energy consumption and fine-grained parallelism, boosting the performance of DNN system by combining MTJ. As a key block of stochastic computing, MTJ based true random number generator design is presented in details. The functionality has been validated by combining the hardware design and post-processing in software. Simulation results are demonstrated visibly by handwritten digits recognition test to show the accuracy. Furthermore, the performance is investigated in terms of accuracy, energy consumption and memory occupation to find more efficient techniques.

KW - Deep neural network

KW - Magnetic tunnel junction

KW - Stochastic computing

KW - True random number generator

UR - https://www.scopus.com/pages/publications/85049443338

U2 - 10.1145/3194554.3194619

DO - 10.1145/3194554.3194619

M3 - 会议稿件

AN - SCOPUS:85049443338

T3 - Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI

SP - 403

EP - 408

BT - GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI

PB - Association for Computing Machinery

T2 - 28th Great Lakes Symposium on VLSI, GLSVLSI 2018

Y2 - 23 May 2018 through 25 May 2018

ER -

Wang Y, Zhang Y , Zhang Y , Zhao W, Cai H, Naviner L. Design space exploration of magnetic tunnel junction based stochastic computing in deep learning. In GLSVLSI 2018 - Proceedings of the 2018 Great Lakes Symposium on VLSI. Association for Computing Machinery . 2018. p. 403-408. (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI). doi: 10.1145/3194554.3194619

Design space exploration of magnetic tunnel junction based stochastic computing in deep learning

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this