Toward local failure local recovery resilience model using MPI-ULFM

Keita Teranishi; Michael A. Heroux

doi:10.1145/2642769.2642774

Toward local failure local recovery resilience model using MPI-ULFM

Keita Teranishi
, Michael A. Heroux

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

51 Scopus citations

Abstract

The current system reaction to the loss of a single MPI process is to kill all the remaining processes and restart the application from the most recent checkpoint. This approach will become unfeasible for future extreme scale systems. We address this issue using an emerging resilient computing model called Local Failure Local Recovery (LFLR) that provides application developers with the ability to recover locally and continue application execution when a process is lost. We discuss the design of our software framework to enable the LFLR model using MPI-ULFM and demonstrate the resilient version of MiniFE that achieves a scalable recovery from process failures.

Original language	English
Title of host publication	Proceedings of the 21st European MPI Users' Group Meeting, EuroMPI/ASIA 2014
Publisher	Association for Computing Machinery
Pages	51-56
Number of pages	6
ISBN (Electronic)	9781450328753
DOIs	https://doi.org/10.1145/2642769.2642774
State	Published - Sep 9 2014
Externally published	Yes
Event	21st European MPI Users' Group Meeting, EuroMPI/ASIA 2014 - Kyoto, Japan Duration: Sep 9 2014 → Sep 12 2014

Publication series

Name	ACM International Conference Proceeding Series
Volume	09-12-September-2014

Conference

Conference	21st European MPI Users' Group Meeting, EuroMPI/ASIA 2014
Country/Territory	Japan
City	Kyoto
Period	09/9/14 → 09/12/14

Funding

The authors would like to thank George Bosilca, Robert Clay, Pedro Diniz and Mark Hoemmen for interesting discussions related to this work. This work was supported by the U.S. Department of Energy (DOE) National Nuclear Security Administration (NNSA) Advanced Simulation and Computing (ASC) program. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.

Keywords

Fault Tolerance
MPI
PDE solvers
Scientific Computing
User Level Fault Mitigation

Access to Document

10.1145/2642769.2642774

Cite this

@inproceedings{27a4788ca44d49cb9fa203d3366592b2,

title = "Toward local failure local recovery resilience model using MPI-ULFM",

abstract = "The current system reaction to the loss of a single MPI process is to kill all the remaining processes and restart the application from the most recent checkpoint. This approach will become unfeasible for future extreme scale systems. We address this issue using an emerging resilient computing model called Local Failure Local Recovery (LFLR) that provides application developers with the ability to recover locally and continue application execution when a process is lost. We discuss the design of our software framework to enable the LFLR model using MPI-ULFM and demonstrate the resilient version of MiniFE that achieves a scalable recovery from process failures.",

keywords = "Fault Tolerance, MPI, PDE solvers, Scientific Computing, User Level Fault Mitigation",

author = "Keita Teranishi and Heroux, \{Michael A.\}",

note = "Publisher Copyright: {\textcopyright} ACM 2014.; 21st European MPI Users' Group Meeting, EuroMPI/ASIA 2014 ; Conference date: 09-09-2014 Through 12-09-2014",

year = "2014",

month = sep,

day = "9",

doi = "10.1145/2642769.2642774",

language = "English",

series = "ACM International Conference Proceeding Series",

publisher = "Association for Computing Machinery",

pages = "51--56",

booktitle = "Proceedings of the 21st European MPI Users' Group Meeting, EuroMPI/ASIA 2014",

}

Teranishi, K & Heroux, MA 2014, Toward local failure local recovery resilience model using MPI-ULFM. in Proceedings of the 21st European MPI Users' Group Meeting, EuroMPI/ASIA 2014. ACM International Conference Proceeding Series, vol. 09-12-September-2014, Association for Computing Machinery, pp. 51-56, 21st European MPI Users' Group Meeting, EuroMPI/ASIA 2014, Kyoto, Japan, 09/9/14. https://doi.org/10.1145/2642769.2642774

Toward local failure local recovery resilience model using MPI-ULFM. / Teranishi, Keita; Heroux, Michael A.
Proceedings of the 21st European MPI Users' Group Meeting, EuroMPI/ASIA 2014. Association for Computing Machinery, 2014. p. 51-56 (ACM International Conference Proceeding Series; Vol. 09-12-September-2014).

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

TY - GEN

T1 - Toward local failure local recovery resilience model using MPI-ULFM

AU - Teranishi, Keita

AU - Heroux, Michael A.

N1 - Publisher Copyright: © ACM 2014.

PY - 2014/9/9

Y1 - 2014/9/9

N2 - The current system reaction to the loss of a single MPI process is to kill all the remaining processes and restart the application from the most recent checkpoint. This approach will become unfeasible for future extreme scale systems. We address this issue using an emerging resilient computing model called Local Failure Local Recovery (LFLR) that provides application developers with the ability to recover locally and continue application execution when a process is lost. We discuss the design of our software framework to enable the LFLR model using MPI-ULFM and demonstrate the resilient version of MiniFE that achieves a scalable recovery from process failures.

AB - The current system reaction to the loss of a single MPI process is to kill all the remaining processes and restart the application from the most recent checkpoint. This approach will become unfeasible for future extreme scale systems. We address this issue using an emerging resilient computing model called Local Failure Local Recovery (LFLR) that provides application developers with the ability to recover locally and continue application execution when a process is lost. We discuss the design of our software framework to enable the LFLR model using MPI-ULFM and demonstrate the resilient version of MiniFE that achieves a scalable recovery from process failures.

KW - Fault Tolerance

KW - MPI

KW - PDE solvers

KW - Scientific Computing

KW - User Level Fault Mitigation

UR - https://www.scopus.com/pages/publications/84958972240

U2 - 10.1145/2642769.2642774

DO - 10.1145/2642769.2642774

M3 - Conference contribution

AN - SCOPUS:84958972240

T3 - ACM International Conference Proceeding Series

SP - 51

EP - 56

BT - Proceedings of the 21st European MPI Users' Group Meeting, EuroMPI/ASIA 2014

PB - Association for Computing Machinery

T2 - 21st European MPI Users' Group Meeting, EuroMPI/ASIA 2014

Y2 - 9 September 2014 through 12 September 2014

ER -

Toward local failure local recovery resilience model using MPI-ULFM

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this