An Overview of Time-Based Computing with Stochastic Constructs

M. Hassan Najafi; Shiva Jamali-Zavareh; David J. Lilja; Marc D. Riedel; Kia Bazargan; Ramesh Harjani

doi:10.1109/MM.2017.4241345

An Overview of Time-Based Computing with Stochastic Constructs

M. Hassan Najafi, Shiva Jamali-Zavareh, David J. Lilja, Marc D. Riedel, Kia Bazargan, Ramesh Harjani

Electrical and Computer Engineering

Research output: Contribution to journal › Review article › peer-review

2 Scopus citations

Abstract

Computing on time-based data is a recent evolution of research in stochastic computing. As with stochastic computing, complex functions can be computed with remarkably low area cost. Unlike stochastic computing, the latency and energy efficiency are very favorable compared to computations on conventional binary radix. In this article, the authors review and evaluate the design and implementation of arithmetic operations on time-encoded signals, with a particular focus on low power. The advantages, challenges, and potential applications are discussed.

Original language	English (US)
Article number	8119711
Pages (from-to)	62-71
Number of pages	10
Journal	IEEE Micro
Volume	37
Issue number	6
DOIs	https://doi.org/10.1109/MM.2017.4241345
State	Published - Nov 1 2017

Keywords

energy-efficient computing
mixed-signal design
pulse width modulation
stochastic computing (SC)
time-based computing
time-encoded values
ultra-low-power processing

UN SDGs

This output contributes to the following Sustainable Development Goal(s)

Access

10.1109/MM.2017.4241345

Fingerprint Dive into the research topics of 'An Overview of Time-Based Computing with Stochastic Constructs'. Together they form a unique fingerprint.

Cite this

@article{db92810ee755403394d0cd0dde97055a,

title = "An Overview of Time-Based Computing with Stochastic Constructs",

abstract = "Computing on time-based data is a recent evolution of research in stochastic computing. As with stochastic computing, complex functions can be computed with remarkably low area cost. Unlike stochastic computing, the latency and energy efficiency are very favorable compared to computations on conventional binary radix. In this article, the authors review and evaluate the design and implementation of arithmetic operations on time-encoded signals, with a particular focus on low power. The advantages, challenges, and potential applications are discussed.",

keywords = "energy-efficient computing, mixed-signal design, pulse width modulation, stochastic computing (SC), time-based computing, time-encoded values, ultra-low-power processing",

author = "Najafi, {M. Hassan} and Shiva Jamali-Zavareh and Lilja, {David J.} and Riedel, {Marc D.} and Kia Bazargan and Ramesh Harjani",

year = "2017",

month = nov,

day = "1",

doi = "10.1109/MM.2017.4241345",

language = "English (US)",

volume = "37",

pages = "62--71",

journal = "IEEE Micro",

issn = "0272-1732",

publisher = "IEEE Computer Society",

number = "6",

}

TY - JOUR

T1 - An Overview of Time-Based Computing with Stochastic Constructs

AU - Najafi, M. Hassan

AU - Jamali-Zavareh, Shiva

AU - Lilja, David J.

AU - Riedel, Marc D.

AU - Bazargan, Kia

AU - Harjani, Ramesh

PY - 2017/11/1

Y1 - 2017/11/1

N2 - Computing on time-based data is a recent evolution of research in stochastic computing. As with stochastic computing, complex functions can be computed with remarkably low area cost. Unlike stochastic computing, the latency and energy efficiency are very favorable compared to computations on conventional binary radix. In this article, the authors review and evaluate the design and implementation of arithmetic operations on time-encoded signals, with a particular focus on low power. The advantages, challenges, and potential applications are discussed.

AB - Computing on time-based data is a recent evolution of research in stochastic computing. As with stochastic computing, complex functions can be computed with remarkably low area cost. Unlike stochastic computing, the latency and energy efficiency are very favorable compared to computations on conventional binary radix. In this article, the authors review and evaluate the design and implementation of arithmetic operations on time-encoded signals, with a particular focus on low power. The advantages, challenges, and potential applications are discussed.

KW - energy-efficient computing

KW - mixed-signal design

KW - pulse width modulation

KW - stochastic computing (SC)

KW - time-based computing

KW - time-encoded values

KW - ultra-low-power processing

UR - http://www.scopus.com/inward/record.url?scp=85038072183&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85038072183&partnerID=8YFLogxK

U2 - 10.1109/MM.2017.4241345

DO - 10.1109/MM.2017.4241345

M3 - Review article

AN - SCOPUS:85038072183

VL - 37

SP - 62

EP - 71

JO - IEEE Micro

JF - IEEE Micro

SN - 0272-1732

IS - 6

M1 - 8119711

ER -