Canaries in a coal mine: Using application-level checkpoints to detect memory failures

Patrick M. Widener; Kurt B. Ferreira; Scott Levy; Nathan Fabian

doi:10.1007/978-3-319-27308-2_54

Canaries in a coal mine: Using application-level checkpoints to detect memory failures

Patrick M. Widener, Kurt B. Ferreira, Scott Levy, Nathan Fabian

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Scopus citations

Abstract

Memory failures in future extreme scale applications are a significant concern in the high-performance computing community and have attracted much research attention. We contend in this paper that using application checkpoint data to detect memory failures has potential benefits and is preferable to examining application memory. To support this contention, we describe the application of machine learning techniques to evaluate the veracity of checkpoint data. Our preliminary results indicate that supervised decision tree machine learning approaches can effectively detect corruption in restart files, suggesting that future extreme-scale applications and systems may benefit from incorporating such approaches in order to cope with memory failues.

Original language	English
Title of host publication	Euro-Par 2015
Subtitle of host publication	Parallel Processing Workshops - Euro-Par 2015 International Workshops, Revised Selected Papers
Editors	Sascha Hunold, Josef Weidendorfer, Domingo Gimenez, Laura Ricci, Stefan Lankes, Alexandru Costan, Ana Lucia Varbanescu, Stephen L. Scott, María Engracia Gómez Requena, Vittorio Scarano, Alexandru Iosup, Michael Alexander
Publisher	Springer Verlag
Pages	669-681
Number of pages	13
ISBN (Print)	9783319273075
DOIs	https://doi.org/10.1007/978-3-319-27308-2_54
State	Published - 2015
Externally published	Yes
Event	International Workshops on Parallel Processing Workshops, Euro-Par 2015 - Vienna, Austria Duration: Aug 24 2015 → Aug 25 2015

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9523
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	International Workshops on Parallel Processing Workshops, Euro-Par 2015
Country/Territory	Austria
City	Vienna
Period	08/24/15 → 08/25/15

Funding

P. M. Widener, K. B. Ferreira, N. Fabian—Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly-owned subsidiary of Lockheed MartinCorporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.

Funders	Funder number
U.S. Department of Energy
National Nuclear Security Administration	DE-AC04-94AL85000

Access to Document

10.1007/978-3-319-27308-2_54

Cite this

Widener, P. M., Ferreira, K. B., Levy, S., & Fabian, N. (2015). Canaries in a coal mine: Using application-level checkpoints to detect memory failures. In S. Hunold, J. Weidendorfer, D. Gimenez, L. Ricci, S. Lankes, A. Costan, A. L. Varbanescu, S. L. Scott, M. E. G. Requena, V. Scarano, A. Iosup, & M. Alexander (Eds.), Euro-Par 2015: Parallel Processing Workshops - Euro-Par 2015 International Workshops, Revised Selected Papers (pp. 669-681). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9523). Springer Verlag. https://doi.org/10.1007/978-3-319-27308-2_54

Widener, Patrick M. ; Ferreira, Kurt B. ; Levy, Scott et al. / Canaries in a coal mine : Using application-level checkpoints to detect memory failures. Euro-Par 2015: Parallel Processing Workshops - Euro-Par 2015 International Workshops, Revised Selected Papers. editor / Sascha Hunold ; Josef Weidendorfer ; Domingo Gimenez ; Laura Ricci ; Stefan Lankes ; Alexandru Costan ; Ana Lucia Varbanescu ; Stephen L. Scott ; María Engracia Gómez Requena ; Vittorio Scarano ; Alexandru Iosup ; Michael Alexander. Springer Verlag, 2015. pp. 669-681 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{83894bb864a24ee9b255f2b324b004b5,

title = "Canaries in a coal mine: Using application-level checkpoints to detect memory failures",

abstract = "Memory failures in future extreme scale applications are a significant concern in the high-performance computing community and have attracted much research attention. We contend in this paper that using application checkpoint data to detect memory failures has potential benefits and is preferable to examining application memory. To support this contention, we describe the application of machine learning techniques to evaluate the veracity of checkpoint data. Our preliminary results indicate that supervised decision tree machine learning approaches can effectively detect corruption in restart files, suggesting that future extreme-scale applications and systems may benefit from incorporating such approaches in order to cope with memory failues.",

author = "Widener, {Patrick M.} and Ferreira, {Kurt B.} and Scott Levy and Nathan Fabian",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2015.; International Workshops on Parallel Processing Workshops, Euro-Par 2015 ; Conference date: 24-08-2015 Through 25-08-2015",

year = "2015",

doi = "10.1007/978-3-319-27308-2_54",

language = "English",

isbn = "9783319273075",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "669--681",

editor = "Sascha Hunold and Josef Weidendorfer and Domingo Gimenez and Laura Ricci and Stefan Lankes and Alexandru Costan and Varbanescu, {Ana Lucia} and Scott, {Stephen L.} and Requena, {Mar{\'i}a Engracia G{\'o}mez} and Vittorio Scarano and Alexandru Iosup and Michael Alexander",

booktitle = "Euro-Par 2015",

}

Widener, PM, Ferreira, KB, Levy, S & Fabian, N 2015, Canaries in a coal mine: Using application-level checkpoints to detect memory failures. in S Hunold, J Weidendorfer, D Gimenez, L Ricci, S Lankes, A Costan, AL Varbanescu, SL Scott, MEG Requena, V Scarano, A Iosup & M Alexander (eds), Euro-Par 2015: Parallel Processing Workshops - Euro-Par 2015 International Workshops, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9523, Springer Verlag, pp. 669-681, International Workshops on Parallel Processing Workshops, Euro-Par 2015, Vienna, Austria, 08/24/15. https://doi.org/10.1007/978-3-319-27308-2_54

Canaries in a coal mine: Using application-level checkpoints to detect memory failures. / Widener, Patrick M.; Ferreira, Kurt B.; Levy, Scott et al.
Euro-Par 2015: Parallel Processing Workshops - Euro-Par 2015 International Workshops, Revised Selected Papers. ed. / Sascha Hunold; Josef Weidendorfer; Domingo Gimenez; Laura Ricci; Stefan Lankes; Alexandru Costan; Ana Lucia Varbanescu; Stephen L. Scott; María Engracia Gómez Requena; Vittorio Scarano; Alexandru Iosup; Michael Alexander. Springer Verlag, 2015. p. 669-681 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9523).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Canaries in a coal mine

T2 - International Workshops on Parallel Processing Workshops, Euro-Par 2015

AU - Widener, Patrick M.

AU - Ferreira, Kurt B.

AU - Levy, Scott

AU - Fabian, Nathan

N1 - Publisher Copyright: © Springer International Publishing Switzerland 2015.

PY - 2015

Y1 - 2015

N2 - Memory failures in future extreme scale applications are a significant concern in the high-performance computing community and have attracted much research attention. We contend in this paper that using application checkpoint data to detect memory failures has potential benefits and is preferable to examining application memory. To support this contention, we describe the application of machine learning techniques to evaluate the veracity of checkpoint data. Our preliminary results indicate that supervised decision tree machine learning approaches can effectively detect corruption in restart files, suggesting that future extreme-scale applications and systems may benefit from incorporating such approaches in order to cope with memory failues.

AB - Memory failures in future extreme scale applications are a significant concern in the high-performance computing community and have attracted much research attention. We contend in this paper that using application checkpoint data to detect memory failures has potential benefits and is preferable to examining application memory. To support this contention, we describe the application of machine learning techniques to evaluate the veracity of checkpoint data. Our preliminary results indicate that supervised decision tree machine learning approaches can effectively detect corruption in restart files, suggesting that future extreme-scale applications and systems may benefit from incorporating such approaches in order to cope with memory failues.

UR - http://www.scopus.com/inward/record.url?scp=84952050378&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-27308-2_54

DO - 10.1007/978-3-319-27308-2_54

M3 - Conference contribution

AN - SCOPUS:84952050378

SN - 9783319273075

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 669

EP - 681

BT - Euro-Par 2015

A2 - Hunold, Sascha

A2 - Weidendorfer, Josef

A2 - Gimenez, Domingo

A2 - Ricci, Laura

A2 - Lankes, Stefan

A2 - Costan, Alexandru

A2 - Varbanescu, Ana Lucia

A2 - Scott, Stephen L.

A2 - Requena, María Engracia Gómez

A2 - Scarano, Vittorio

A2 - Iosup, Alexandru

A2 - Alexander, Michael

PB - Springer Verlag

Y2 - 24 August 2015 through 25 August 2015

ER -

Widener PM, Ferreira KB, Levy S, Fabian N. Canaries in a coal mine: Using application-level checkpoints to detect memory failures. In Hunold S, Weidendorfer J, Gimenez D, Ricci L, Lankes S, Costan A, Varbanescu AL, Scott SL, Requena MEG, Scarano V, Iosup A, Alexander M, editors, Euro-Par 2015: Parallel Processing Workshops - Euro-Par 2015 International Workshops, Revised Selected Papers. Springer Verlag. 2015. p. 669-681. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-27308-2_54

Canaries in a coal mine: Using application-level checkpoints to detect memory failures

Abstract

Publication series

Conference

Funding

Access to Document

Other files and links

Fingerprint

Cite this