Architecture-aware approximate computing

Mustafa Karakoy, Orhan Kislal, Xulong Tang, Mahmut Kandemir, Meenakshi Arunachalam

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

Abstract

Observing that many application programs from different domains can live with less-than-perfect accuracy, existing techniques try to trade off program output accuracy with performance-energy savings. While these works provide point solutions, they leave three critical questions regarding approximate computing unanswered: (i) what is the maximum potential of skipping (i.e., not performing) data accesses under a given inaccuracy bound?; (ii) can we identify the data accesses to drop randomly, or is being architecture aware critical?; and (iii) do two executions that skip the same number of data accesses always result in the same output quality (error)? This paper first provides answers to these questions using ten multithreaded workloads, and then presents a program slicing-based approach that identifies the set of data accesses to drop. Results indicate 8.8% performance improvement and 13.7% energy saving are possible when we set the error bound to 2%, and the corresponding improvements jump to 15% and 25%, respectively, when the error bound is raised to 4%.

Original languageEnglish (US)
Title of host publicationSIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems
PublisherAssociation for Computing Machinery, Inc
Pages23-24
Number of pages2
ISBN (Electronic)9781450366786
DOIs
StatePublished - Jun 20 2019
Event14th Joint Conference of International Conference on Measurement and Modeling of Computer Systems, SIGMETRICS 2019 and IFIP Performance Conference 2019, SIGMETRICS/Performance 2019 - Phoenix, United States
Duration: Jun 24 2019Jun 28 2019

Publication series

NameSIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems

Conference

Conference14th Joint Conference of International Conference on Measurement and Modeling of Computer Systems, SIGMETRICS 2019 and IFIP Performance Conference 2019, SIGMETRICS/Performance 2019
CountryUnited States
CityPhoenix
Period6/24/196/28/19

Fingerprint

Energy conservation
Application programs

All Science Journal Classification (ASJC) codes

  • Hardware and Architecture
  • Computer Networks and Communications
  • Computational Theory and Mathematics

Cite this

Karakoy, M., Kislal, O., Tang, X., Kandemir, M., & Arunachalam, M. (2019). Architecture-aware approximate computing. In SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems (pp. 23-24). (SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems). Association for Computing Machinery, Inc. https://doi.org/10.1145/3309697.3331508
Karakoy, Mustafa ; Kislal, Orhan ; Tang, Xulong ; Kandemir, Mahmut ; Arunachalam, Meenakshi. / Architecture-aware approximate computing. SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems. Association for Computing Machinery, Inc, 2019. pp. 23-24 (SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems).
@inproceedings{88be6fe35f70406699ea37f7d044f85e,
title = "Architecture-aware approximate computing",
abstract = "Observing that many application programs from different domains can live with less-than-perfect accuracy, existing techniques try to trade off program output accuracy with performance-energy savings. While these works provide point solutions, they leave three critical questions regarding approximate computing unanswered: (i) what is the maximum potential of skipping (i.e., not performing) data accesses under a given inaccuracy bound?; (ii) can we identify the data accesses to drop randomly, or is being architecture aware critical?; and (iii) do two executions that skip the same number of data accesses always result in the same output quality (error)? This paper first provides answers to these questions using ten multithreaded workloads, and then presents a program slicing-based approach that identifies the set of data accesses to drop. Results indicate 8.8{\%} performance improvement and 13.7{\%} energy saving are possible when we set the error bound to 2{\%}, and the corresponding improvements jump to 15{\%} and 25{\%}, respectively, when the error bound is raised to 4{\%}.",
author = "Mustafa Karakoy and Orhan Kislal and Xulong Tang and Mahmut Kandemir and Meenakshi Arunachalam",
year = "2019",
month = "6",
day = "20",
doi = "10.1145/3309697.3331508",
language = "English (US)",
series = "SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems",
publisher = "Association for Computing Machinery, Inc",
pages = "23--24",
booktitle = "SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems",

}

Karakoy, M, Kislal, O, Tang, X, Kandemir, M & Arunachalam, M 2019, Architecture-aware approximate computing. in SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems. SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems, Association for Computing Machinery, Inc, pp. 23-24, 14th Joint Conference of International Conference on Measurement and Modeling of Computer Systems, SIGMETRICS 2019 and IFIP Performance Conference 2019, SIGMETRICS/Performance 2019, Phoenix, United States, 6/24/19. https://doi.org/10.1145/3309697.3331508

Architecture-aware approximate computing. / Karakoy, Mustafa; Kislal, Orhan; Tang, Xulong; Kandemir, Mahmut; Arunachalam, Meenakshi.

SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems. Association for Computing Machinery, Inc, 2019. p. 23-24 (SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems).

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

TY - GEN

T1 - Architecture-aware approximate computing

AU - Karakoy, Mustafa

AU - Kislal, Orhan

AU - Tang, Xulong

AU - Kandemir, Mahmut

AU - Arunachalam, Meenakshi

PY - 2019/6/20

Y1 - 2019/6/20

N2 - Observing that many application programs from different domains can live with less-than-perfect accuracy, existing techniques try to trade off program output accuracy with performance-energy savings. While these works provide point solutions, they leave three critical questions regarding approximate computing unanswered: (i) what is the maximum potential of skipping (i.e., not performing) data accesses under a given inaccuracy bound?; (ii) can we identify the data accesses to drop randomly, or is being architecture aware critical?; and (iii) do two executions that skip the same number of data accesses always result in the same output quality (error)? This paper first provides answers to these questions using ten multithreaded workloads, and then presents a program slicing-based approach that identifies the set of data accesses to drop. Results indicate 8.8% performance improvement and 13.7% energy saving are possible when we set the error bound to 2%, and the corresponding improvements jump to 15% and 25%, respectively, when the error bound is raised to 4%.

AB - Observing that many application programs from different domains can live with less-than-perfect accuracy, existing techniques try to trade off program output accuracy with performance-energy savings. While these works provide point solutions, they leave three critical questions regarding approximate computing unanswered: (i) what is the maximum potential of skipping (i.e., not performing) data accesses under a given inaccuracy bound?; (ii) can we identify the data accesses to drop randomly, or is being architecture aware critical?; and (iii) do two executions that skip the same number of data accesses always result in the same output quality (error)? This paper first provides answers to these questions using ten multithreaded workloads, and then presents a program slicing-based approach that identifies the set of data accesses to drop. Results indicate 8.8% performance improvement and 13.7% energy saving are possible when we set the error bound to 2%, and the corresponding improvements jump to 15% and 25%, respectively, when the error bound is raised to 4%.

UR - http://www.scopus.com/inward/record.url?scp=85069210662&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85069210662&partnerID=8YFLogxK

U2 - 10.1145/3309697.3331508

DO - 10.1145/3309697.3331508

M3 - Conference contribution

AN - SCOPUS:85069210662

T3 - SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems

SP - 23

EP - 24

BT - SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems

PB - Association for Computing Machinery, Inc

ER -

Karakoy M, Kislal O, Tang X, Kandemir M, Arunachalam M. Architecture-aware approximate computing. In SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems. Association for Computing Machinery, Inc. 2019. p. 23-24. (SIGMETRICS Performance 2019 - Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems). https://doi.org/10.1145/3309697.3331508