Versioned distributed arrays for resilience in scientific applications: Global View Resilience

A. Chien; P. Balaji; P. Beckman; N. Dun; A. Fang; H. Fujita; K. Iskra; Z. Rubenstein; Z. Zheng; R. Schreiber; J. Hammond; J. Dinan; I. Laguna; D. Richards; A. Dubey; B. Van Straalen; M. Hoemmen; M. Heroux; K. Teranishi; A. Siegel

doi:10.1016/j.procs.2015.05.187

Versioned distributed arrays for resilience in scientific applications: Global View Resilience

A. Chien
, P. Balaji
, P. Beckman
, N. Dun
, A. Fang
, H. Fujita
, K. Iskra
, Z. Rubenstein
, Z. Zheng
, R. Schreiber
, J. Hammond
, J. Dinan
, I. Laguna
, D. Richards
, A. Dubey
, B. Van Straalen
, M. Hoemmen
, M. Heroux
, K. Teranishi
, A. Siegel

Programming Systems

Research output: Contribution to journal › Conference article › peer-review

24 Scopus citations

Abstract

Exascale studies project reliability challenges for future high-performance computing (HPC) systems. We propose the Global View Resilience (GVR) system, a library that enables applications to add resilience in a portable, application-controlled fashion using versioned distributed arrays. We describe GVR's interfaces to distributed arrays, versioning, and cross-layer error recovery. Using several large applications (OpenMC, the preconditioned conjugate gradient solver PCG, ddcMD, and Chombo), we evaluate the programmer effort to add resilience. The required changes are small (<2% LOC), localized, and machine-independent, requiring no software architecture changes. We also measure the overhead of adding GVR versioning and show that generally overheads <2% are achieved. We conclude that GVR's interfaces and implementation are flexible and portable and create a gentle-slope path to tolerate growing error rates in future systems.

Original language	English
Pages (from-to)	29-38
Number of pages	10
Journal	Procedia Computer Science
Volume	51
Issue number	1
DOIs	https://doi.org/10.1016/j.procs.2015.05.187
State	Published - 2015
Event	International Conference on Computational Science, ICCS 2002 - Amsterdam, Netherlands Duration: Apr 21 2002 → Apr 24 2002

Funding

This work was supported by the Office of Advanced Scientific Computing Research, Office of Science, U.S. Department of Energy, under Award DE-SC0008603 and Contract DE-AC02-06CH11357. This work was completed in part with resources provided by: the University of Chicago Research Computing Center, , the resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231, and resources of the Argonne Leadership Computing Facility, which is a DOE Office of Science User Facility supported under Contract DE-AC02-06CH11357.

Keywords

Application-based fault tolerance
Exascale
Fault tolerance
Resilience
Scalable computing

Access to Document

10.1016/j.procs.2015.05.187

Cite this

Chien, A., Balaji, P., Beckman, P., Dun, N., Fang, A., Fujita, H., Iskra, K., Rubenstein, Z., Zheng, Z., Schreiber, R., Hammond, J., Dinan, J., Laguna, I., Richards, D., Dubey, A., Van Straalen, B., Hoemmen, M., Heroux, M., Teranishi, K., & Siegel, A. (2015). Versioned distributed arrays for resilience in scientific applications: Global View Resilience. Procedia Computer Science, 51(1), 29-38. https://doi.org/10.1016/j.procs.2015.05.187

@article{20ae276fba0f483fa09313f0a1e760c2,

title = "Versioned distributed arrays for resilience in scientific applications: Global View Resilience",

abstract = "Exascale studies project reliability challenges for future high-performance computing (HPC) systems. We propose the Global View Resilience (GVR) system, a library that enables applications to add resilience in a portable, application-controlled fashion using versioned distributed arrays. We describe GVR's interfaces to distributed arrays, versioning, and cross-layer error recovery. Using several large applications (OpenMC, the preconditioned conjugate gradient solver PCG, ddcMD, and Chombo), we evaluate the programmer effort to add resilience. The required changes are small (<2\% LOC), localized, and machine-independent, requiring no software architecture changes. We also measure the overhead of adding GVR versioning and show that generally overheads <2\% are achieved. We conclude that GVR's interfaces and implementation are flexible and portable and create a gentle-slope path to tolerate growing error rates in future systems.",

keywords = "Application-based fault tolerance, Exascale, Fault tolerance, Resilience, Scalable computing",

author = "A. Chien and P. Balaji and P. Beckman and N. Dun and A. Fang and H. Fujita and K. Iskra and Z. Rubenstein and Z. Zheng and R. Schreiber and J. Hammond and J. Dinan and I. Laguna and D. Richards and A. Dubey and \{Van Straalen\}, B. and M. Hoemmen and M. Heroux and K. Teranishi and A. Siegel",

note = "Publisher Copyright: {\textcopyright} The Authors. Published by Elsevier B.V.; International Conference on Computational Science, ICCS 2002 ; Conference date: 21-04-2002 Through 24-04-2002",

year = "2015",

doi = "10.1016/j.procs.2015.05.187",

language = "English",

volume = "51",

pages = "29--38",

journal = "Procedia Computer Science",

issn = "1877-0509",

publisher = "Elsevier",

number = "1",

}

Chien, A, Balaji, P, Beckman, P, Dun, N, Fang, A, Fujita, H, Iskra, K, Rubenstein, Z, Zheng, Z, Schreiber, R, Hammond, J, Dinan, J, Laguna, I, Richards, D, Dubey, A, Van Straalen, B, Hoemmen, M, Heroux, M, Teranishi, K & Siegel, A 2015, 'Versioned distributed arrays for resilience in scientific applications: Global View Resilience', Procedia Computer Science, vol. 51, no. 1, pp. 29-38. https://doi.org/10.1016/j.procs.2015.05.187

TY - JOUR

T1 - Versioned distributed arrays for resilience in scientific applications

T2 - International Conference on Computational Science, ICCS 2002

AU - Chien, A.

AU - Balaji, P.

AU - Beckman, P.

AU - Dun, N.

AU - Fang, A.

AU - Fujita, H.

AU - Iskra, K.

AU - Rubenstein, Z.

AU - Zheng, Z.

AU - Schreiber, R.

AU - Hammond, J.

AU - Dinan, J.

AU - Laguna, I.

AU - Richards, D.

AU - Dubey, A.

AU - Van Straalen, B.

AU - Hoemmen, M.

AU - Heroux, M.

AU - Teranishi, K.

AU - Siegel, A.

PY - 2015

Y1 - 2015

N2 - Exascale studies project reliability challenges for future high-performance computing (HPC) systems. We propose the Global View Resilience (GVR) system, a library that enables applications to add resilience in a portable, application-controlled fashion using versioned distributed arrays. We describe GVR's interfaces to distributed arrays, versioning, and cross-layer error recovery. Using several large applications (OpenMC, the preconditioned conjugate gradient solver PCG, ddcMD, and Chombo), we evaluate the programmer effort to add resilience. The required changes are small (<2% LOC), localized, and machine-independent, requiring no software architecture changes. We also measure the overhead of adding GVR versioning and show that generally overheads <2% are achieved. We conclude that GVR's interfaces and implementation are flexible and portable and create a gentle-slope path to tolerate growing error rates in future systems.

AB - Exascale studies project reliability challenges for future high-performance computing (HPC) systems. We propose the Global View Resilience (GVR) system, a library that enables applications to add resilience in a portable, application-controlled fashion using versioned distributed arrays. We describe GVR's interfaces to distributed arrays, versioning, and cross-layer error recovery. Using several large applications (OpenMC, the preconditioned conjugate gradient solver PCG, ddcMD, and Chombo), we evaluate the programmer effort to add resilience. The required changes are small (<2% LOC), localized, and machine-independent, requiring no software architecture changes. We also measure the overhead of adding GVR versioning and show that generally overheads <2% are achieved. We conclude that GVR's interfaces and implementation are flexible and portable and create a gentle-slope path to tolerate growing error rates in future systems.

KW - Application-based fault tolerance

KW - Exascale

KW - Fault tolerance

KW - Resilience

KW - Scalable computing

UR - https://www.scopus.com/pages/publications/84939214452

U2 - 10.1016/j.procs.2015.05.187

DO - 10.1016/j.procs.2015.05.187

M3 - Conference article

AN - SCOPUS:84939214452

SN - 1877-0509

VL - 51

SP - 29

EP - 38

JO - Procedia Computer Science

JF - Procedia Computer Science

IS - 1

Y2 - 21 April 2002 through 24 April 2002

ER -

Versioned distributed arrays for resilience in scientific applications: Global View Resilience

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this