Spin-Hall-Effect-Based Stochastic Number Generator for Parallel Stochastic Computing

Jiaxi Hu; Bingzhe Li; Cong Ma; David Lilja; Steven J. Koester

doi:10.1109/TED.2019.2920401

Spin-Hall-Effect-Based Stochastic Number Generator for Parallel Stochastic Computing

Jiaxi Hu, Bingzhe Li, Cong Ma, David Lilja, Steven J. Koester

Electrical and Computer Engineering

Research output: Contribution to journal › Article › peer-review

8 Scopus citations

Abstract

Stochastic computing (SC) is a promising technology that can be used for low-cost hardware designs. However, SC suffers from its long latency. Although parallel processing can efficiently shorten the latency, duplicated stochastic number generators (SNGs) are necessary, which cause substantial hardware overhead. This paper proposes a scalable SNG based on the spin-Hall-effect (SHE), which is capable of generating multiple independent stochastic streams simultaneously. The design takes advantages of the efficient charge-to-spin conversion from the Spin-Hall material and the intrinsic stochasticity of nanomagnets. Compared to previous spintronic SNGs, the SHE-SNG can reduce the area by 1.6×-7.8× and the power by 4.9×-13× while increasing the degree of parallelism from 1 to 16. Compared to CMOS-based SNGs, the proposed SNG obtained 24×-120× and 53× reduction in terms of area and power, respectively. Finally, three benchmarks were implemented, and the results indicate that SC implementations with the proposed SHE-SNG can achieve 1.2×-29× reduction of hardware resources compared to implementations with previous CMOS-and spintronic-based designs while scaling the degree of parallelism from 1 to 64.

Original language	English (US)
Article number	8741170
Pages (from-to)	3620-3627
Number of pages	8
Journal	IEEE Transactions on Electron Devices
Volume	66
Issue number	8
DOIs	https://doi.org/10.1109/TED.2019.2920401
State	Published - Aug 2019

Bibliographical note

Funding Information:
Manuscript received April 26, 2019; revised May 26, 2019; accepted May 28, 2019. Date of publication June 19, 2019; date of current version July 23, 2019. This work was supported in part by the National Science Foundation under Grant CCF-1408123. The review of this paper was arranged by Editor A. M. P. Anantram. (Jiaxi Hu and Bingzhe Li contributed equally to this work.) (Corresponding author: Bingzhe Li.) The authors are with the Department of Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN 55445 USA (e-mail: huxxx499@umn.edu; lixx1743@umn.edu; maxxx376@umn.edu; lilja@umn.edu; skoester@umn.edu).

Publisher Copyright:
© 1963-2012 IEEE.

Keywords

Parallel processing
spintronic
stochastic computing (SC)
stochastic number generator (SNG)

Publisher link

10.1109/TED.2019.2920401

Find It

Find It at U of M Libraries

Cite this

@article{5d4bcda91d47421ba016a2eb3fa3f2ee,

title = "Spin-Hall-Effect-Based Stochastic Number Generator for Parallel Stochastic Computing",

abstract = "Stochastic computing (SC) is a promising technology that can be used for low-cost hardware designs. However, SC suffers from its long latency. Although parallel processing can efficiently shorten the latency, duplicated stochastic number generators (SNGs) are necessary, which cause substantial hardware overhead. This paper proposes a scalable SNG based on the spin-Hall-effect (SHE), which is capable of generating multiple independent stochastic streams simultaneously. The design takes advantages of the efficient charge-to-spin conversion from the Spin-Hall material and the intrinsic stochasticity of nanomagnets. Compared to previous spintronic SNGs, the SHE-SNG can reduce the area by 1.6×-7.8× and the power by 4.9×-13× while increasing the degree of parallelism from 1 to 16. Compared to CMOS-based SNGs, the proposed SNG obtained 24×-120× and 53× reduction in terms of area and power, respectively. Finally, three benchmarks were implemented, and the results indicate that SC implementations with the proposed SHE-SNG can achieve 1.2×-29× reduction of hardware resources compared to implementations with previous CMOS-and spintronic-based designs while scaling the degree of parallelism from 1 to 64.",

keywords = "Parallel processing, spintronic, stochastic computing (SC), stochastic number generator (SNG)",

author = "Jiaxi Hu and Bingzhe Li and Cong Ma and David Lilja and Koester, {Steven J.}",

note = "Funding Information: Manuscript received April 26, 2019; revised May 26, 2019; accepted May 28, 2019. Date of publication June 19, 2019; date of current version July 23, 2019. This work was supported in part by the National Science Foundation under Grant CCF-1408123. The review of this paper was arranged by Editor A. M. P. Anantram. (Jiaxi Hu and Bingzhe Li contributed equally to this work.) (Corresponding author: Bingzhe Li.) The authors are with the Department of Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN 55445 USA (e-mail: huxxx499@umn.edu; lixx1743@umn.edu; maxxx376@umn.edu; lilja@umn.edu; skoester@umn.edu). Publisher Copyright: {\textcopyright} 1963-2012 IEEE.",

year = "2019",

month = aug,

doi = "10.1109/TED.2019.2920401",

language = "English (US)",

volume = "66",

pages = "3620--3627",

journal = "IEEE Transactions on Electron Devices",

issn = "0018-9383",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "8",

}

TY - JOUR

T1 - Spin-Hall-Effect-Based Stochastic Number Generator for Parallel Stochastic Computing

AU - Hu, Jiaxi

AU - Li, Bingzhe

AU - Ma, Cong

AU - Lilja, David

AU - Koester, Steven J.

N1 - Funding Information: Manuscript received April 26, 2019; revised May 26, 2019; accepted May 28, 2019. Date of publication June 19, 2019; date of current version July 23, 2019. This work was supported in part by the National Science Foundation under Grant CCF-1408123. The review of this paper was arranged by Editor A. M. P. Anantram. (Jiaxi Hu and Bingzhe Li contributed equally to this work.) (Corresponding author: Bingzhe Li.) The authors are with the Department of Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN 55445 USA (e-mail: huxxx499@umn.edu; lixx1743@umn.edu; maxxx376@umn.edu; lilja@umn.edu; skoester@umn.edu). Publisher Copyright: © 1963-2012 IEEE.

PY - 2019/8

Y1 - 2019/8

N2 - Stochastic computing (SC) is a promising technology that can be used for low-cost hardware designs. However, SC suffers from its long latency. Although parallel processing can efficiently shorten the latency, duplicated stochastic number generators (SNGs) are necessary, which cause substantial hardware overhead. This paper proposes a scalable SNG based on the spin-Hall-effect (SHE), which is capable of generating multiple independent stochastic streams simultaneously. The design takes advantages of the efficient charge-to-spin conversion from the Spin-Hall material and the intrinsic stochasticity of nanomagnets. Compared to previous spintronic SNGs, the SHE-SNG can reduce the area by 1.6×-7.8× and the power by 4.9×-13× while increasing the degree of parallelism from 1 to 16. Compared to CMOS-based SNGs, the proposed SNG obtained 24×-120× and 53× reduction in terms of area and power, respectively. Finally, three benchmarks were implemented, and the results indicate that SC implementations with the proposed SHE-SNG can achieve 1.2×-29× reduction of hardware resources compared to implementations with previous CMOS-and spintronic-based designs while scaling the degree of parallelism from 1 to 64.

AB - Stochastic computing (SC) is a promising technology that can be used for low-cost hardware designs. However, SC suffers from its long latency. Although parallel processing can efficiently shorten the latency, duplicated stochastic number generators (SNGs) are necessary, which cause substantial hardware overhead. This paper proposes a scalable SNG based on the spin-Hall-effect (SHE), which is capable of generating multiple independent stochastic streams simultaneously. The design takes advantages of the efficient charge-to-spin conversion from the Spin-Hall material and the intrinsic stochasticity of nanomagnets. Compared to previous spintronic SNGs, the SHE-SNG can reduce the area by 1.6×-7.8× and the power by 4.9×-13× while increasing the degree of parallelism from 1 to 16. Compared to CMOS-based SNGs, the proposed SNG obtained 24×-120× and 53× reduction in terms of area and power, respectively. Finally, three benchmarks were implemented, and the results indicate that SC implementations with the proposed SHE-SNG can achieve 1.2×-29× reduction of hardware resources compared to implementations with previous CMOS-and spintronic-based designs while scaling the degree of parallelism from 1 to 64.

KW - Parallel processing

KW - spintronic

KW - stochastic computing (SC)

KW - stochastic number generator (SNG)

UR - http://www.scopus.com/inward/record.url?scp=85069914187&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85069914187&partnerID=8YFLogxK

U2 - 10.1109/TED.2019.2920401

DO - 10.1109/TED.2019.2920401

M3 - Article

AN - SCOPUS:85069914187

SN - 0018-9383

VL - 66

SP - 3620

EP - 3627

JO - IEEE Transactions on Electron Devices

JF - IEEE Transactions on Electron Devices

IS - 8

M1 - 8741170

ER -

Spin-Hall-Effect-Based Stochastic Number Generator for Parallel Stochastic Computing

Abstract

Bibliographical note

Keywords

Publisher link

Find It

Other files and links

Fingerprint

Cite this