Software-based selective validation techniques for Robust CGRAs against soft errors

Yohan Ko, Jihoon Kang, Jongwon Lee, Yongjoo Kim, Joonhyun Kim, Hwisoo So, Kyoungwoo Lee, Yunheung Paek

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Coarse-Grained Reconfigurable Architectures (CGRAs) are drawing significant attention since they promise both performances with parallelism and flexibility with reconfiguration. Soft errors (or transient faults) are becoming a serious design concern in embedded systems including CGRAs since the soft error rate is increasing exponentially as technology is scaling. A recently proposed software-based technique with TMR (Triple Modular Redundancy) implemented on CGRAs incurs extreme overheads in terms of runtime and energy consumption mainly due to expensive voting mechanisms for the outputs from the triplication ofevery operation. In this article, we propose selective validation mechanisms for efficient modular redundancy techniques in the datapaths on CGRAs. Our techniques selectively validate the results at synchronous operations rather than every operation in order to reduce the expensive performance overhead from the validation mechanism. We also present an optimization technique to further improve the runtime and the energy consumption by minimizing synchronous operations where a validating mechanism needs to be applied. Our experimental results demonstrate that our selective validation-based TMR technique with our optimization on CGRAs can improve the runtime by 41.0% and the energy consumption by 26.2% on average over benchmarks as compared to the recently proposed software-based TMR technique with the full validation.

Original languageEnglish
Article number20
JournalACM Transactions on Embedded Computing Systems
Volume15
Issue number1
DOIs
Publication statusPublished - 2016 Feb 1

Fingerprint

Reconfigurable architectures
Redundancy
Energy utilization
Embedded systems

All Science Journal Classification (ASJC) codes

  • Software
  • Hardware and Architecture

Cite this

Ko, Yohan ; Kang, Jihoon ; Lee, Jongwon ; Kim, Yongjoo ; Kim, Joonhyun ; So, Hwisoo ; Lee, Kyoungwoo ; Paek, Yunheung. / Software-based selective validation techniques for Robust CGRAs against soft errors. In: ACM Transactions on Embedded Computing Systems. 2016 ; Vol. 15, No. 1.
@article{00f0d9af101e4850885bf1d9d423d22b,
title = "Software-based selective validation techniques for Robust CGRAs against soft errors",
abstract = "Coarse-Grained Reconfigurable Architectures (CGRAs) are drawing significant attention since they promise both performances with parallelism and flexibility with reconfiguration. Soft errors (or transient faults) are becoming a serious design concern in embedded systems including CGRAs since the soft error rate is increasing exponentially as technology is scaling. A recently proposed software-based technique with TMR (Triple Modular Redundancy) implemented on CGRAs incurs extreme overheads in terms of runtime and energy consumption mainly due to expensive voting mechanisms for the outputs from the triplication ofevery operation. In this article, we propose selective validation mechanisms for efficient modular redundancy techniques in the datapaths on CGRAs. Our techniques selectively validate the results at synchronous operations rather than every operation in order to reduce the expensive performance overhead from the validation mechanism. We also present an optimization technique to further improve the runtime and the energy consumption by minimizing synchronous operations where a validating mechanism needs to be applied. Our experimental results demonstrate that our selective validation-based TMR technique with our optimization on CGRAs can improve the runtime by 41.0{\%} and the energy consumption by 26.2{\%} on average over benchmarks as compared to the recently proposed software-based TMR technique with the full validation.",
author = "Yohan Ko and Jihoon Kang and Jongwon Lee and Yongjoo Kim and Joonhyun Kim and Hwisoo So and Kyoungwoo Lee and Yunheung Paek",
year = "2016",
month = "2",
day = "1",
doi = "10.1145/2843943",
language = "English",
volume = "15",
journal = "Transactions on Embedded Computing Systems",
issn = "1539-9087",
publisher = "Association for Computing Machinery (ACM)",
number = "1",

}

Software-based selective validation techniques for Robust CGRAs against soft errors. / Ko, Yohan; Kang, Jihoon; Lee, Jongwon; Kim, Yongjoo; Kim, Joonhyun; So, Hwisoo; Lee, Kyoungwoo; Paek, Yunheung.

In: ACM Transactions on Embedded Computing Systems, Vol. 15, No. 1, 20, 01.02.2016.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Software-based selective validation techniques for Robust CGRAs against soft errors

AU - Ko, Yohan

AU - Kang, Jihoon

AU - Lee, Jongwon

AU - Kim, Yongjoo

AU - Kim, Joonhyun

AU - So, Hwisoo

AU - Lee, Kyoungwoo

AU - Paek, Yunheung

PY - 2016/2/1

Y1 - 2016/2/1

N2 - Coarse-Grained Reconfigurable Architectures (CGRAs) are drawing significant attention since they promise both performances with parallelism and flexibility with reconfiguration. Soft errors (or transient faults) are becoming a serious design concern in embedded systems including CGRAs since the soft error rate is increasing exponentially as technology is scaling. A recently proposed software-based technique with TMR (Triple Modular Redundancy) implemented on CGRAs incurs extreme overheads in terms of runtime and energy consumption mainly due to expensive voting mechanisms for the outputs from the triplication ofevery operation. In this article, we propose selective validation mechanisms for efficient modular redundancy techniques in the datapaths on CGRAs. Our techniques selectively validate the results at synchronous operations rather than every operation in order to reduce the expensive performance overhead from the validation mechanism. We also present an optimization technique to further improve the runtime and the energy consumption by minimizing synchronous operations where a validating mechanism needs to be applied. Our experimental results demonstrate that our selective validation-based TMR technique with our optimization on CGRAs can improve the runtime by 41.0% and the energy consumption by 26.2% on average over benchmarks as compared to the recently proposed software-based TMR technique with the full validation.

AB - Coarse-Grained Reconfigurable Architectures (CGRAs) are drawing significant attention since they promise both performances with parallelism and flexibility with reconfiguration. Soft errors (or transient faults) are becoming a serious design concern in embedded systems including CGRAs since the soft error rate is increasing exponentially as technology is scaling. A recently proposed software-based technique with TMR (Triple Modular Redundancy) implemented on CGRAs incurs extreme overheads in terms of runtime and energy consumption mainly due to expensive voting mechanisms for the outputs from the triplication ofevery operation. In this article, we propose selective validation mechanisms for efficient modular redundancy techniques in the datapaths on CGRAs. Our techniques selectively validate the results at synchronous operations rather than every operation in order to reduce the expensive performance overhead from the validation mechanism. We also present an optimization technique to further improve the runtime and the energy consumption by minimizing synchronous operations where a validating mechanism needs to be applied. Our experimental results demonstrate that our selective validation-based TMR technique with our optimization on CGRAs can improve the runtime by 41.0% and the energy consumption by 26.2% on average over benchmarks as compared to the recently proposed software-based TMR technique with the full validation.

UR - http://www.scopus.com/inward/record.url?scp=84964893862&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84964893862&partnerID=8YFLogxK

U2 - 10.1145/2843943

DO - 10.1145/2843943

M3 - Article

VL - 15

JO - Transactions on Embedded Computing Systems

JF - Transactions on Embedded Computing Systems

SN - 1539-9087

IS - 1

M1 - 20

ER -