Low cost hybrid spin-CMOS compressor for stochastic neural networks

Bingzhe Li; Jiaxi Hu; M. Hassan Najafi; Steven Koester; David J. Lilja

doi:10.1145/3299874.3317989

Low cost hybrid spin-CMOS compressor for stochastic neural networks

Bingzhe Li, Jiaxi Hu, M. Hassan Najafi, Steven Koester, David J. Lilja

Electrical and Computer Engineering

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

2 Scopus citations

Abstract

With expansion of neural network (NN) applications lowering their hardware implementation cost becomes an urgent task especially in back-end applications where the power-supply is limited. Stochastic computing (SC) is a promising solution to realize low-cost hardware designs. Implementation of matrix multiplication has been a bottleneck in previous stochastic neural networks (SC-NNs). In this paper, we introduce spintronic components into the design of SC-NNs. A novel spin-CMOS matrix multiplier is proposed in which the stochastic multiplications are performed by CMOS AND gates while the sum of products is implemented by spintronic compressor gates. The experimental results indicate that compared to the conventional binary implementations the proposed hybrid spin-CMOS architecture can achieve over 125x, 4.5x and 43x; reduction in terms of power, energy and area consumptions, respectively. Moreover, compared to previous CMOS-based SC-NNs, our design saves the power by 3.1x-7.3x, reduces energy consumption by 3.1x-7.3x and decreases area by 1.4x-7.6x while maintaining similar recognition rates.

Original language	English (US)
Title of host publication	GLSVLSI 2019 - Proceedings of the 2019 Great Lakes Symposium on VLSI
Publisher	Association for Computing Machinery
Pages	141-146
Number of pages	6
ISBN (Electronic)	9781450362528
DOIs	https://doi.org/10.1145/3299874.3317989
State	Published - May 13 2019
Event	29th Great Lakes Symposium on VLSI, GLSVLSI 2019 - Tysons Corner, United States Duration: May 9 2019 → May 11 2019

Publication series

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

Conference

Conference	29th Great Lakes Symposium on VLSI, GLSVLSI 2019
Country/Territory	United States
City	Tysons Corner
Period	5/9/19 → 5/11/19

Bibliographical note

Funding Information:
This work was supported in part by National Science Foundation grant no. CCF-1408123 and Seagate Technology.

Publisher Copyright:
© 2019 ACM.

Keywords

Compressor
Neural network
Spintronic
Stochastic computing

Access

10.1145/3299874.3317989

OpenUrl availability

Full text

Cite this

Li, B., Hu, J., Hassan Najafi, M., Koester, S., & Lilja, D. J. (2019). Low cost hybrid spin-CMOS compressor for stochastic neural networks. In GLSVLSI 2019 - Proceedings of the 2019 Great Lakes Symposium on VLSI (pp. 141-146). Article 3317989 (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI). Association for Computing Machinery. https://doi.org/10.1145/3299874.3317989

Low cost hybrid spin-CMOS compressor for stochastic neural networks. / Li, Bingzhe; Hu, Jiaxi; Hassan Najafi, M. et al.
GLSVLSI 2019 - Proceedings of the 2019 Great Lakes Symposium on VLSI. Association for Computing Machinery, 2019. p. 141-146 3317989 (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI).

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

Li, B, Hu, J, Hassan Najafi, M, Koester, S & Lilja, DJ 2019, Low cost hybrid spin-CMOS compressor for stochastic neural networks. in GLSVLSI 2019 - Proceedings of the 2019 Great Lakes Symposium on VLSI., 3317989, Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI, Association for Computing Machinery, pp. 141-146, 29th Great Lakes Symposium on VLSI, GLSVLSI 2019, Tysons Corner, United States, 5/9/19. https://doi.org/10.1145/3299874.3317989

@inproceedings{47c2b5129a1d4686a78902661d0545f3,

title = "Low cost hybrid spin-CMOS compressor for stochastic neural networks",

abstract = "With expansion of neural network (NN) applications lowering their hardware implementation cost becomes an urgent task especially in back-end applications where the power-supply is limited. Stochastic computing (SC) is a promising solution to realize low-cost hardware designs. Implementation of matrix multiplication has been a bottleneck in previous stochastic neural networks (SC-NNs). In this paper, we introduce spintronic components into the design of SC-NNs. A novel spin-CMOS matrix multiplier is proposed in which the stochastic multiplications are performed by CMOS AND gates while the sum of products is implemented by spintronic compressor gates. The experimental results indicate that compared to the conventional binary implementations the proposed hybrid spin-CMOS architecture can achieve over 125x, 4.5x and 43x; reduction in terms of power, energy and area consumptions, respectively. Moreover, compared to previous CMOS-based SC-NNs, our design saves the power by 3.1x-7.3x, reduces energy consumption by 3.1x-7.3x and decreases area by 1.4x-7.6x while maintaining similar recognition rates.",

keywords = "Compressor, Neural network, Spintronic, Stochastic computing",

author = "Bingzhe Li and Jiaxi Hu and {Hassan Najafi}, M. and Steven Koester and Lilja, {David J.}",

note = "Funding Information: This work was supported in part by National Science Foundation grant no. CCF-1408123 and Seagate Technology. Publisher Copyright: {\textcopyright} 2019 ACM.; 29th Great Lakes Symposium on VLSI, GLSVLSI 2019 ; Conference date: 09-05-2019 Through 11-05-2019",

year = "2019",

month = may,

day = "13",

doi = "10.1145/3299874.3317989",

language = "English (US)",

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

publisher = "Association for Computing Machinery",

pages = "141--146",

booktitle = "GLSVLSI 2019 - Proceedings of the 2019 Great Lakes Symposium on VLSI",

}

TY - GEN

T1 - Low cost hybrid spin-CMOS compressor for stochastic neural networks

AU - Li, Bingzhe

AU - Hu, Jiaxi

AU - Hassan Najafi, M.

AU - Koester, Steven

AU - Lilja, David J.

PY - 2019/5/13

Y1 - 2019/5/13

N2 - With expansion of neural network (NN) applications lowering their hardware implementation cost becomes an urgent task especially in back-end applications where the power-supply is limited. Stochastic computing (SC) is a promising solution to realize low-cost hardware designs. Implementation of matrix multiplication has been a bottleneck in previous stochastic neural networks (SC-NNs). In this paper, we introduce spintronic components into the design of SC-NNs. A novel spin-CMOS matrix multiplier is proposed in which the stochastic multiplications are performed by CMOS AND gates while the sum of products is implemented by spintronic compressor gates. The experimental results indicate that compared to the conventional binary implementations the proposed hybrid spin-CMOS architecture can achieve over 125x, 4.5x and 43x; reduction in terms of power, energy and area consumptions, respectively. Moreover, compared to previous CMOS-based SC-NNs, our design saves the power by 3.1x-7.3x, reduces energy consumption by 3.1x-7.3x and decreases area by 1.4x-7.6x while maintaining similar recognition rates.

AB - With expansion of neural network (NN) applications lowering their hardware implementation cost becomes an urgent task especially in back-end applications where the power-supply is limited. Stochastic computing (SC) is a promising solution to realize low-cost hardware designs. Implementation of matrix multiplication has been a bottleneck in previous stochastic neural networks (SC-NNs). In this paper, we introduce spintronic components into the design of SC-NNs. A novel spin-CMOS matrix multiplier is proposed in which the stochastic multiplications are performed by CMOS AND gates while the sum of products is implemented by spintronic compressor gates. The experimental results indicate that compared to the conventional binary implementations the proposed hybrid spin-CMOS architecture can achieve over 125x, 4.5x and 43x; reduction in terms of power, energy and area consumptions, respectively. Moreover, compared to previous CMOS-based SC-NNs, our design saves the power by 3.1x-7.3x, reduces energy consumption by 3.1x-7.3x and decreases area by 1.4x-7.6x while maintaining similar recognition rates.

KW - Compressor

KW - Neural network

KW - Spintronic

KW - Stochastic computing

UR - http://www.scopus.com/inward/record.url?scp=85083278888&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85083278888&partnerID=8YFLogxK

U2 - 10.1145/3299874.3317989

DO - 10.1145/3299874.3317989

M3 - Conference contribution

AN - SCOPUS:85083278888

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

SP - 141

EP - 146

BT - GLSVLSI 2019 - Proceedings of the 2019 Great Lakes Symposium on VLSI

PB - Association for Computing Machinery

T2 - 29th Great Lakes Symposium on VLSI, GLSVLSI 2019

Y2 - 9 May 2019 through 11 May 2019

ER -

Low cost hybrid spin-CMOS compressor for stochastic neural networks

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access

OpenUrl availability

Other files and links

Fingerprint

Cite this