Towards theoretical cost limit of stochastic number generators for stochastic computing

Meng Yang; Bingzhe Li; David J. Lilja; Bo Yuan; Weikang Qian

doi:10.1109/ISVLSI.2018.00037

Towards theoretical cost limit of stochastic number generators for stochastic computing

Meng Yang, Bingzhe Li, David J. Lilja, Bo Yuan, Weikang Qian

Electrical and Computer Engineering

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

16 Scopus citations

Abstract

Stochastic number generator (SNG) is one important component of stochastic computing (SC). An SNG usually consists of a random number source (RNS) and a probability conversion circuit (PCC). The SNGs occupy a large portion of the total area and power of a stochastic circuit. Thus, it is critical to lower the area and power of the SNGs. The existing methods only focused on simplifying the RNSs inside the SNGs, such as sharing the RNSs and using emerging devices. However, how to reduce the area and power of PCCs is never studied. In this work, we explore this problem and propose a solution that can effectively reduce the area and power of PCCs. We also study the theoretical limit on the cost of SNG and find that our proposed design approaches the limit. The experimental results show that our design can gain up to 2x improvement in power-delay product over the traditional SNGs.

Original language	English (US)
Title of host publication	Proceedings - 2018 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2018
Publisher	IEEE Computer Society
Pages	154-159
Number of pages	6
ISBN (Print)	9781538670996
DOIs	https://doi.org/10.1109/ISVLSI.2018.00037
State	Published - Aug 7 2018
Event	17th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2018 - Hong Kong, Hong Kong Duration: Jul 9 2018 → Jul 11 2018

Publication series

Name	Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI
Volume	2018-July
ISSN (Print)	2159-3469
ISSN (Electronic)	2159-3477

Other

Other	17th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2018
Country/Territory	Hong Kong
City	Hong Kong
Period	7/9/18 → 7/11/18

Bibliographical note

Funding Information:
This work is supported in part by National Natural Science Foundation of China (NSFC) under grant no. 61472243 and 61204042 and in part by National Science Foundation (NSF) of U.S.A. under grant no. CCF-1408123 and CCF-1438286. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSFC and NSF.

Publisher Copyright:
© 2018 IEEE.

Keywords

Probability Conversion Circuit
Stochastic number generator
Theoretical Limit

Publisher link

10.1109/ISVLSI.2018.00037

Find It

Find It at U of M Libraries

Cite this

Yang, M., Li, B., Lilja, D. J., Yuan, B., & Qian, W. (2018). Towards theoretical cost limit of stochastic number generators for stochastic computing. In Proceedings - 2018 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2018 (pp. 154-159). [8429358] (Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI; Vol. 2018-July). IEEE Computer Society. https://doi.org/10.1109/ISVLSI.2018.00037

Towards theoretical cost limit of stochastic number generators for stochastic computing. / Yang, Meng; Li, Bingzhe; Lilja, David J. et al.
Proceedings - 2018 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2018. IEEE Computer Society, 2018. p. 154-159 8429358 (Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI; Vol. 2018-July).

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

Yang, M, Li, B, Lilja, DJ, Yuan, B & Qian, W 2018, Towards theoretical cost limit of stochastic number generators for stochastic computing. in Proceedings - 2018 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2018., 8429358, Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI, vol. 2018-July, IEEE Computer Society, pp. 154-159, 17th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2018, Hong Kong, Hong Kong, 7/9/18. https://doi.org/10.1109/ISVLSI.2018.00037

@inproceedings{94d88aa96c534393af871bdea0599a6a,

title = "Towards theoretical cost limit of stochastic number generators for stochastic computing",

abstract = "Stochastic number generator (SNG) is one important component of stochastic computing (SC). An SNG usually consists of a random number source (RNS) and a probability conversion circuit (PCC). The SNGs occupy a large portion of the total area and power of a stochastic circuit. Thus, it is critical to lower the area and power of the SNGs. The existing methods only focused on simplifying the RNSs inside the SNGs, such as sharing the RNSs and using emerging devices. However, how to reduce the area and power of PCCs is never studied. In this work, we explore this problem and propose a solution that can effectively reduce the area and power of PCCs. We also study the theoretical limit on the cost of SNG and find that our proposed design approaches the limit. The experimental results show that our design can gain up to 2x improvement in power-delay product over the traditional SNGs.",

keywords = "Probability Conversion Circuit, Stochastic number generator, Theoretical Limit",

author = "Meng Yang and Bingzhe Li and Lilja, {David J.} and Bo Yuan and Weikang Qian",

note = "Funding Information: This work is supported in part by National Natural Science Foundation of China (NSFC) under grant no. 61472243 and 61204042 and in part by National Science Foundation (NSF) of U.S.A. under grant no. CCF-1408123 and CCF-1438286. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSFC and NSF. Publisher Copyright: {\textcopyright} 2018 IEEE.; 17th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2018 ; Conference date: 09-07-2018 Through 11-07-2018",

year = "2018",

month = aug,

day = "7",

doi = "10.1109/ISVLSI.2018.00037",

language = "English (US)",

isbn = "9781538670996",

series = "Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI",

publisher = "IEEE Computer Society",

pages = "154--159",

booktitle = "Proceedings - 2018 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2018",

}

TY - GEN

T1 - Towards theoretical cost limit of stochastic number generators for stochastic computing

AU - Yang, Meng

AU - Li, Bingzhe

AU - Lilja, David J.

AU - Yuan, Bo

AU - Qian, Weikang

N1 - Funding Information: This work is supported in part by National Natural Science Foundation of China (NSFC) under grant no. 61472243 and 61204042 and in part by National Science Foundation (NSF) of U.S.A. under grant no. CCF-1408123 and CCF-1438286. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSFC and NSF. Publisher Copyright: © 2018 IEEE.

PY - 2018/8/7

Y1 - 2018/8/7

N2 - Stochastic number generator (SNG) is one important component of stochastic computing (SC). An SNG usually consists of a random number source (RNS) and a probability conversion circuit (PCC). The SNGs occupy a large portion of the total area and power of a stochastic circuit. Thus, it is critical to lower the area and power of the SNGs. The existing methods only focused on simplifying the RNSs inside the SNGs, such as sharing the RNSs and using emerging devices. However, how to reduce the area and power of PCCs is never studied. In this work, we explore this problem and propose a solution that can effectively reduce the area and power of PCCs. We also study the theoretical limit on the cost of SNG and find that our proposed design approaches the limit. The experimental results show that our design can gain up to 2x improvement in power-delay product over the traditional SNGs.

AB - Stochastic number generator (SNG) is one important component of stochastic computing (SC). An SNG usually consists of a random number source (RNS) and a probability conversion circuit (PCC). The SNGs occupy a large portion of the total area and power of a stochastic circuit. Thus, it is critical to lower the area and power of the SNGs. The existing methods only focused on simplifying the RNSs inside the SNGs, such as sharing the RNSs and using emerging devices. However, how to reduce the area and power of PCCs is never studied. In this work, we explore this problem and propose a solution that can effectively reduce the area and power of PCCs. We also study the theoretical limit on the cost of SNG and find that our proposed design approaches the limit. The experimental results show that our design can gain up to 2x improvement in power-delay product over the traditional SNGs.

KW - Probability Conversion Circuit

KW - Stochastic number generator

KW - Theoretical Limit

UR - http://www.scopus.com/inward/record.url?scp=85052102840&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85052102840&partnerID=8YFLogxK

U2 - 10.1109/ISVLSI.2018.00037

DO - 10.1109/ISVLSI.2018.00037

M3 - Conference contribution

AN - SCOPUS:85052102840

SN - 9781538670996

T3 - Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI

SP - 154

EP - 159

BT - Proceedings - 2018 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2018

PB - IEEE Computer Society

T2 - 17th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2018

Y2 - 9 July 2018 through 11 July 2018

ER -

Towards theoretical cost limit of stochastic number generators for stochastic computing

Abstract

Publication series

Other

Bibliographical note

Keywords

Publisher link

Find It

Other files and links

Fingerprint

Cite this