An evaluation of User-Level Failure Mitigation support in MPI

As the scale of computing platforms becomes increasingly extreme, the requirements for application fault tolerance are increasing as well. Techniques to address this problem by improving the resilience of algorithms have been developed, but they currently receive no support from the programming mode...

Full description

Saved in:
Bibliographic Details
Published inComputing Vol. 95; no. 12; pp. 1171 - 1184
Main Authors Bland, Wesley, Bouteiller, Aurelien, Herault, Thomas, Hursey, Joshua, Bosilca, George, Dongarra, Jack J.
Format Journal Article
LanguageEnglish
Published Vienna Springer Vienna 01.12.2013
Springer Nature B.V
Subjects
Algorithms
Artificial Intelligence
Computer Appl. in Administrative Data Processing
Computer Communication Networks
Computer programming
Computer Science
Failure
Fault tolerance
Information Systems Applications (incl.Internet)
Libraries
Semantics
Software
Software Engineering
Studies
Users
Fault tolerance
User-level fault mitigation
68M15
68M14
MPI
Online AccessGet full text

Cover

More Information
Summary:As the scale of computing platforms becomes increasingly extreme, the requirements for application fault tolerance are increasing as well. Techniques to address this problem by improving the resilience of algorithms have been developed, but they currently receive no support from the programming model, and without such support, they are bound to fail. This paper discusses the failure-free overhead and recovery impact of the user-level failure mitigation proposal presented in the MPI Forum. Experiments demonstrate that fault-aware MPI has little or no impact on performance for a range of applications, and produces satisfactory recovery times when there are failures.
ISSN:0010-485X
1436-5057
DOI:10.1007/s00607-013-0331-3