Multiplexer Optimization for Adders in Stochastic Computing

Sercan Aygun; M. Hassan Najafi; Lida Kouhalvandi; Ece Olcay Gunes

doi:10.1145/3611315.3633256

Multiplexer Optimization for Adders in Stochastic Computing

Sercan Aygun, M. Hassan Najafi, Lida Kouhalvandi, Ece Olcay Gunes

Department of Electronics and Communication Engineering

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

1 Citation (Scopus)

Abstract

This study presents an optimization algorithm for multiplexer (MUX)-based scaled addition for stochastic computing (SC). Accumulation operation can be performed in SC using a MUX unit. Cascaded structures of 2m-to-1 MUXs are used for the accumulation of multiple terms. Optimizing these designs holds significance in cases of accumulating a large number of inputs. The depth of the cascaded MUXs varies with m, affecting the hardware cost, delay, and accuracy. The proposed algorithm performs stage-wise optimization of m. Evaluation results show a lower hardware cost and a higher accuracy compared to the standard MUX-based SC addition using 2-to-1 MUXs for SC-based neural networks.

Original language	English
Title of host publication	Proceedings of the 18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023
Publisher	Association for Computing Machinery
ISBN (Electronic)	9798400703256
DOIs	https://doi.org/10.1145/3611315.3633256
Publication status	Published - 18 Dec 2023
Event	18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023 - Dresden, Germany Duration: 18 Dec 2023 → 20 Dec 2023

Publication series

Name	ACM International Conference Proceeding Series

Conference

Conference	18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023
Country/Territory	Germany
City	Dresden
Period	18/12/23 → 20/12/23

Bibliographical note

Publisher Copyright:
© 2023 ACM.

Funding

This work was supported in part by National Science Foundation (NSF) grant #2019511, the Louisiana Board of Regents Support Fund #LEQSF(2020-23)-RD-A-26, and generous gifts from Cisco, Xilinx, and Nvidia.

Funders	Funder number
National Science Foundation	2019511
Louisiana Board of Regents	2020-23)-RD-A-26

Keywords

Multiplexer
neural networks
optimization
stochastic computing

Access to Document

10.1145/3611315.3633256

Cite this

@inproceedings{692467ef3adf4ce3b8390fbd15fd6710,

title = "Multiplexer Optimization for Adders in Stochastic Computing",

abstract = "This study presents an optimization algorithm for multiplexer (MUX)-based scaled addition for stochastic computing (SC). Accumulation operation can be performed in SC using a MUX unit. Cascaded structures of 2m-to-1 MUXs are used for the accumulation of multiple terms. Optimizing these designs holds significance in cases of accumulating a large number of inputs. The depth of the cascaded MUXs varies with m, affecting the hardware cost, delay, and accuracy. The proposed algorithm performs stage-wise optimization of m. Evaluation results show a lower hardware cost and a higher accuracy compared to the standard MUX-based SC addition using 2-to-1 MUXs for SC-based neural networks.",

keywords = "Multiplexer, neural networks, optimization, stochastic computing",

author = "Sercan Aygun and Najafi, \{M. Hassan\} and Lida Kouhalvandi and Gunes, \{Ece Olcay\}",

note = "Publisher Copyright: {\textcopyright} 2023 ACM.; 18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023 ; Conference date: 18-12-2023 Through 20-12-2023",

year = "2023",

month = dec,

day = "18",

doi = "10.1145/3611315.3633256",

language = "English",

series = "ACM International Conference Proceeding Series",

publisher = "Association for Computing Machinery",

booktitle = "Proceedings of the 18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023",

}

Aygun, S, Najafi, MH, Kouhalvandi, L & Gunes, EO 2023, Multiplexer Optimization for Adders in Stochastic Computing. in Proceedings of the 18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023., 18, ACM International Conference Proceeding Series, Association for Computing Machinery, 18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023, Dresden, Germany, 18/12/23. https://doi.org/10.1145/3611315.3633256

Multiplexer Optimization for Adders in Stochastic Computing. / Aygun, Sercan; Najafi, M. Hassan; Kouhalvandi, Lida et al.
Proceedings of the 18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023. Association for Computing Machinery, 2023. 18 (ACM International Conference Proceeding Series).

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

TY - GEN

T1 - Multiplexer Optimization for Adders in Stochastic Computing

AU - Aygun, Sercan

AU - Najafi, M. Hassan

AU - Kouhalvandi, Lida

AU - Gunes, Ece Olcay

PY - 2023/12/18

Y1 - 2023/12/18

N2 - This study presents an optimization algorithm for multiplexer (MUX)-based scaled addition for stochastic computing (SC). Accumulation operation can be performed in SC using a MUX unit. Cascaded structures of 2m-to-1 MUXs are used for the accumulation of multiple terms. Optimizing these designs holds significance in cases of accumulating a large number of inputs. The depth of the cascaded MUXs varies with m, affecting the hardware cost, delay, and accuracy. The proposed algorithm performs stage-wise optimization of m. Evaluation results show a lower hardware cost and a higher accuracy compared to the standard MUX-based SC addition using 2-to-1 MUXs for SC-based neural networks.

AB - This study presents an optimization algorithm for multiplexer (MUX)-based scaled addition for stochastic computing (SC). Accumulation operation can be performed in SC using a MUX unit. Cascaded structures of 2m-to-1 MUXs are used for the accumulation of multiple terms. Optimizing these designs holds significance in cases of accumulating a large number of inputs. The depth of the cascaded MUXs varies with m, affecting the hardware cost, delay, and accuracy. The proposed algorithm performs stage-wise optimization of m. Evaluation results show a lower hardware cost and a higher accuracy compared to the standard MUX-based SC addition using 2-to-1 MUXs for SC-based neural networks.

KW - Multiplexer

KW - neural networks

KW - optimization

KW - stochastic computing

UR - https://www.scopus.com/pages/publications/85184301041

U2 - 10.1145/3611315.3633256

DO - 10.1145/3611315.3633256

M3 - Conference contribution

AN - SCOPUS:85184301041

T3 - ACM International Conference Proceeding Series

BT - Proceedings of the 18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023

PB - Association for Computing Machinery

T2 - 18th ACM International Symposium on Nanoscale Architectures, NANOARCH 2023

Y2 - 18 December 2023 through 20 December 2023

ER -

Multiplexer Optimization for Adders in Stochastic Computing

Abstract

Publication series

Conference

Bibliographical note

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this