Phase-aware optimization in approximate computing

Subrata Mitra, Manish K. Gupta, Sasa Misailovic, Saurabh Bagchi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper shows that many applications exhibit execution-phase-specific sensitivity towards approximation of the internal subcomputations. Therefore, approximation in certain phases can be more beneficial than others. Further, this paper presents Opprox, a novel system for application's execution-phase-aware approximation. For a user provided error budget and target input parameters, Opprox identifies different program phases and searches for profitable approximation settings for each phase of the application execution. Our evaluation with five benchmarks and four existing transformations show that our phase-aware optimization on average does 14% less work for a 5% error tolerance bound and 42% less work for a 20% tolerance bound.

Original languageEnglish (US)
Title of host publicationCGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization
EditorsVijay Janapa Reddi, Aaron Smith, Lingjia Tang
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages185-196
Number of pages12
ISBN (Electronic)9781509049318
DOIs
StatePublished - Feb 23 2017
Event2017 International Symposium on Code Generation and Optimization, CGO 2017 - Austin, United States
Duration: Feb 4 2017Feb 8 2017

Publication series

NameCGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization

Other

Other2017 International Symposium on Code Generation and Optimization, CGO 2017
CountryUnited States
CityAustin
Period2/4/172/8/17

Fingerprint

Optimization
Computing
Approximation
Tolerance
Benchmark
Internal
Target
Evaluation

Keywords

  • Approximate Computing
  • Execution Phases

ASJC Scopus subject areas

  • Software
  • Control and Optimization

Cite this

Mitra, S., Gupta, M. K., Misailovic, S., & Bagchi, S. (2017). Phase-aware optimization in approximate computing. In V. J. Reddi, A. Smith, & L. Tang (Eds.), CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization (pp. 185-196). [7863739] (CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CGO.2017.7863739

Phase-aware optimization in approximate computing. / Mitra, Subrata; Gupta, Manish K.; Misailovic, Sasa; Bagchi, Saurabh.

CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization. ed. / Vijay Janapa Reddi; Aaron Smith; Lingjia Tang. Institute of Electrical and Electronics Engineers Inc., 2017. p. 185-196 7863739 (CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Mitra, S, Gupta, MK, Misailovic, S & Bagchi, S 2017, Phase-aware optimization in approximate computing. in VJ Reddi, A Smith & L Tang (eds), CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization., 7863739, CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization, Institute of Electrical and Electronics Engineers Inc., pp. 185-196, 2017 International Symposium on Code Generation and Optimization, CGO 2017, Austin, United States, 2/4/17. https://doi.org/10.1109/CGO.2017.7863739
Mitra S, Gupta MK, Misailovic S, Bagchi S. Phase-aware optimization in approximate computing. In Reddi VJ, Smith A, Tang L, editors, CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization. Institute of Electrical and Electronics Engineers Inc. 2017. p. 185-196. 7863739. (CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization). https://doi.org/10.1109/CGO.2017.7863739
Mitra, Subrata ; Gupta, Manish K. ; Misailovic, Sasa ; Bagchi, Saurabh. / Phase-aware optimization in approximate computing. CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization. editor / Vijay Janapa Reddi ; Aaron Smith ; Lingjia Tang. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 185-196 (CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization).
@inproceedings{e5ae7b028f8d4c1aa68acfbcde88dbdd,
title = "Phase-aware optimization in approximate computing",
abstract = "This paper shows that many applications exhibit execution-phase-specific sensitivity towards approximation of the internal subcomputations. Therefore, approximation in certain phases can be more beneficial than others. Further, this paper presents Opprox, a novel system for application's execution-phase-aware approximation. For a user provided error budget and target input parameters, Opprox identifies different program phases and searches for profitable approximation settings for each phase of the application execution. Our evaluation with five benchmarks and four existing transformations show that our phase-aware optimization on average does 14{\%} less work for a 5{\%} error tolerance bound and 42{\%} less work for a 20{\%} tolerance bound.",
keywords = "Approximate Computing, Execution Phases",
author = "Subrata Mitra and Gupta, {Manish K.} and Sasa Misailovic and Saurabh Bagchi",
year = "2017",
month = "2",
day = "23",
doi = "10.1109/CGO.2017.7863739",
language = "English (US)",
series = "CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "185--196",
editor = "Reddi, {Vijay Janapa} and Aaron Smith and Lingjia Tang",
booktitle = "CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization",
address = "United States",

}

TY - GEN

T1 - Phase-aware optimization in approximate computing

AU - Mitra, Subrata

AU - Gupta, Manish K.

AU - Misailovic, Sasa

AU - Bagchi, Saurabh

PY - 2017/2/23

Y1 - 2017/2/23

N2 - This paper shows that many applications exhibit execution-phase-specific sensitivity towards approximation of the internal subcomputations. Therefore, approximation in certain phases can be more beneficial than others. Further, this paper presents Opprox, a novel system for application's execution-phase-aware approximation. For a user provided error budget and target input parameters, Opprox identifies different program phases and searches for profitable approximation settings for each phase of the application execution. Our evaluation with five benchmarks and four existing transformations show that our phase-aware optimization on average does 14% less work for a 5% error tolerance bound and 42% less work for a 20% tolerance bound.

AB - This paper shows that many applications exhibit execution-phase-specific sensitivity towards approximation of the internal subcomputations. Therefore, approximation in certain phases can be more beneficial than others. Further, this paper presents Opprox, a novel system for application's execution-phase-aware approximation. For a user provided error budget and target input parameters, Opprox identifies different program phases and searches for profitable approximation settings for each phase of the application execution. Our evaluation with five benchmarks and four existing transformations show that our phase-aware optimization on average does 14% less work for a 5% error tolerance bound and 42% less work for a 20% tolerance bound.

KW - Approximate Computing

KW - Execution Phases

UR - http://www.scopus.com/inward/record.url?scp=85016065847&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85016065847&partnerID=8YFLogxK

U2 - 10.1109/CGO.2017.7863739

DO - 10.1109/CGO.2017.7863739

M3 - Conference contribution

AN - SCOPUS:85016065847

T3 - CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization

SP - 185

EP - 196

BT - CGO 2017 - Proceedings of the 2017 International Symposium on Code Generation and Optimization

A2 - Reddi, Vijay Janapa

A2 - Smith, Aaron

A2 - Tang, Lingjia

PB - Institute of Electrical and Electronics Engineers Inc.

ER -