Investigating Power Reduction for NoC-based Spiking Neural Network Platforms using Channel Encoding

Neil McDonnell, Snaider Carrillo, Jim Harkin, Liam McDaid

Research output: Contribution to journalArticle

Abstract

Recent focus has been placed on exploring the possibility to switch from parallel to serial data links between NoC routers in order to improve signal integrity in the communication channel. However, moving streams of data between the parallel path of the internal router and external serial-channel links between them consumes additional power. One challenge is encoding the data and minimise the switching activity of data in the serial links in order to reduce the additional power dissipation; while under real-time and minimal hardware constraints. Consequently, proposed is a novel low area/power decision circuit for NoC channel encoding which identifies in real-time packets for encoding and extends the existing SILENT encoders/decoders to further minimise power consumption and demonstrates the power performance savings of the decision circuitand modified (en)decoders using example test traffic with the EMBRACE NoC router, a mixed signal spiking neural network (SNNs) embedded platform.
LanguageEnglish
Pages1-16
JournalInternational Journal of Adaptive, Resilient, and Autonomic Systems
Volume4
Issue number3
DOIs
Publication statusPublished - Dec 2012

Fingerprint

Routers
Neural networks
Energy dissipation
Electric power utilization
Switches
Hardware
Networks (circuits)
Network-on-chip

Cite this

@article{3960624088a445ef8e544ccfe10a0617,
title = "Investigating Power Reduction for NoC-based Spiking Neural Network Platforms using Channel Encoding",
abstract = "Recent focus has been placed on exploring the possibility to switch from parallel to serial data links between NoC routers in order to improve signal integrity in the communication channel. However, moving streams of data between the parallel path of the internal router and external serial-channel links between them consumes additional power. One challenge is encoding the data and minimise the switching activity of data in the serial links in order to reduce the additional power dissipation; while under real-time and minimal hardware constraints. Consequently, proposed is a novel low area/power decision circuit for NoC channel encoding which identifies in real-time packets for encoding and extends the existing SILENT encoders/decoders to further minimise power consumption and demonstrates the power performance savings of the decision circuitand modified (en)decoders using example test traffic with the EMBRACE NoC router, a mixed signal spiking neural network (SNNs) embedded platform.",
author = "Neil McDonnell and Snaider Carrillo and Jim Harkin and Liam McDaid",
year = "2012",
month = "12",
doi = "10.4018/jaras.2012100101",
language = "English",
volume = "4",
pages = "1--16",
journal = "International Journal of Adaptive, Resilient and Autonomic Systems",
issn = "1947-9220",
number = "3",

}

TY - JOUR

T1 - Investigating Power Reduction for NoC-based Spiking Neural Network Platforms using Channel Encoding

AU - McDonnell, Neil

AU - Carrillo, Snaider

AU - Harkin, Jim

AU - McDaid, Liam

PY - 2012/12

Y1 - 2012/12

N2 - Recent focus has been placed on exploring the possibility to switch from parallel to serial data links between NoC routers in order to improve signal integrity in the communication channel. However, moving streams of data between the parallel path of the internal router and external serial-channel links between them consumes additional power. One challenge is encoding the data and minimise the switching activity of data in the serial links in order to reduce the additional power dissipation; while under real-time and minimal hardware constraints. Consequently, proposed is a novel low area/power decision circuit for NoC channel encoding which identifies in real-time packets for encoding and extends the existing SILENT encoders/decoders to further minimise power consumption and demonstrates the power performance savings of the decision circuitand modified (en)decoders using example test traffic with the EMBRACE NoC router, a mixed signal spiking neural network (SNNs) embedded platform.

AB - Recent focus has been placed on exploring the possibility to switch from parallel to serial data links between NoC routers in order to improve signal integrity in the communication channel. However, moving streams of data between the parallel path of the internal router and external serial-channel links between them consumes additional power. One challenge is encoding the data and minimise the switching activity of data in the serial links in order to reduce the additional power dissipation; while under real-time and minimal hardware constraints. Consequently, proposed is a novel low area/power decision circuit for NoC channel encoding which identifies in real-time packets for encoding and extends the existing SILENT encoders/decoders to further minimise power consumption and demonstrates the power performance savings of the decision circuitand modified (en)decoders using example test traffic with the EMBRACE NoC router, a mixed signal spiking neural network (SNNs) embedded platform.

U2 - 10.4018/jaras.2012100101

DO - 10.4018/jaras.2012100101

M3 - Article

VL - 4

SP - 1

EP - 16

JO - International Journal of Adaptive, Resilient and Autonomic Systems

T2 - International Journal of Adaptive, Resilient and Autonomic Systems

JF - International Journal of Adaptive, Resilient and Autonomic Systems

SN - 1947-9220

IS - 3

ER -