Evolving the message passing programming model via a fault-tolerant, object-oriented transport layer

Jeremiah J. Wilke; Hemanth Kolla; Keita Teranishi; David S. Hollman; Janine C. Bennett; Nicole Slattengren

doi:10.1145/2751504.2751513

Evolving the message passing programming model via a fault-tolerant, object-oriented transport layer

Jeremiah J. Wilke
, Hemanth Kolla
, Keita Teranishi
, David S. Hollman
, Janine C. Bennett
, Nicole Slattengren

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

2 Scopus citations

Abstract

In this position paper, we argue for improved fault-tolerance of an MPI code by introducing lightweight virtualization into the MPI interface. In particular, we outline key-value store semantics for MPI send/recv calls, thereby creating a far more expressive programming model. The general message passing semantics and imperative style of MPI application codes would remain essentially unchanged. However, the additional expressiblity of the programming model 1) enables the underlying transport layer to handle faulttolerance more transparently to the application developer, and 2) provides an evolutionary code path towards more declarative asynchronous programming models. The core contribution of this paper is an initial implementation of the DHARMA transport layer that provides the new, required functionality to support the MPI key-value store model.

Original language	English
Title of host publication	FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015
Publisher	Association for Computing Machinery, Inc
Pages	41-46
Number of pages	6
ISBN (Electronic)	9781450335690
DOIs	https://doi.org/10.1145/2751504.2751513
State	Published - Jun 15 2015
Externally published	Yes
Event	5th Workshop on Fault Tolerance for HPC at eXtreme Scale, FTXS 2015 - Portland, United States Duration: Jun 15 2015 → …

Publication series

Name	FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015

Conference

Conference	5th Workshop on Fault Tolerance for HPC at eXtreme Scale, FTXS 2015
Country/Territory	United States
City	Portland
Period	06/15/15 → …

Funding

The authors would like to thank Craig Ulmer, Gary Templet, and Abhinav Vishnu for useful discussions. This work was supported by the U.S. Department of Energy (DOE) Na- tional Nuclear Security Administration (NNSA) Advanced Simulation and Computing (ASC) program. Sandia Na- tional Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned sub- sidiary of Lockheed Martin Corporation, for the U.S. Department of Energy's National Nuclear Security Administration under contract DE-AC04-94AL85000

Access to Document

10.1145/2751504.2751513

Cite this

Wilke, J. J., Kolla, H., Teranishi, K., Hollman, D. S., Bennett, J. C., & Slattengren, N. (2015). Evolving the message passing programming model via a fault-tolerant, object-oriented transport layer. In FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015 (pp. 41-46). (FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015). Association for Computing Machinery, Inc. https://doi.org/10.1145/2751504.2751513

Wilke, Jeremiah J. ; Kolla, Hemanth ; Teranishi, Keita et al. / Evolving the message passing programming model via a fault-tolerant, object-oriented transport layer. FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015. Association for Computing Machinery, Inc, 2015. pp. 41-46 (FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015).

@inproceedings{fa8ffbfc833849a69adebe71fb7b6bb7,

title = "Evolving the message passing programming model via a fault-tolerant, object-oriented transport layer",

abstract = "In this position paper, we argue for improved fault-tolerance of an MPI code by introducing lightweight virtualization into the MPI interface. In particular, we outline key-value store semantics for MPI send/recv calls, thereby creating a far more expressive programming model. The general message passing semantics and imperative style of MPI application codes would remain essentially unchanged. However, the additional expressiblity of the programming model 1) enables the underlying transport layer to handle faulttolerance more transparently to the application developer, and 2) provides an evolutionary code path towards more declarative asynchronous programming models. The core contribution of this paper is an initial implementation of the DHARMA transport layer that provides the new, required functionality to support the MPI key-value store model.",

author = "Wilke, \{Jeremiah J.\} and Hemanth Kolla and Keita Teranishi and Hollman, \{David S.\} and Bennett, \{Janine C.\} and Nicole Slattengren",

note = "Publisher Copyright: {\textcopyright} 2015 ACM.; 5th Workshop on Fault Tolerance for HPC at eXtreme Scale, FTXS 2015 ; Conference date: 15-06-2015",

year = "2015",

month = jun,

day = "15",

doi = "10.1145/2751504.2751513",

language = "English",

series = "FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015",

publisher = "Association for Computing Machinery, Inc",

pages = "41--46",

booktitle = "FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015",

}

Wilke, JJ, Kolla, H, Teranishi, K, Hollman, DS, Bennett, JC & Slattengren, N 2015, Evolving the message passing programming model via a fault-tolerant, object-oriented transport layer. in FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015. FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015, Association for Computing Machinery, Inc, pp. 41-46, 5th Workshop on Fault Tolerance for HPC at eXtreme Scale, FTXS 2015, Portland, United States, 06/15/15. https://doi.org/10.1145/2751504.2751513

Evolving the message passing programming model via a fault-tolerant, object-oriented transport layer. / Wilke, Jeremiah J.; Kolla, Hemanth; Teranishi, Keita et al.
FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015. Association for Computing Machinery, Inc, 2015. p. 41-46 (FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015).

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

TY - GEN

T1 - Evolving the message passing programming model via a fault-tolerant, object-oriented transport layer

AU - Wilke, Jeremiah J.

AU - Kolla, Hemanth

AU - Teranishi, Keita

AU - Hollman, David S.

AU - Bennett, Janine C.

AU - Slattengren, Nicole

PY - 2015/6/15

Y1 - 2015/6/15

N2 - In this position paper, we argue for improved fault-tolerance of an MPI code by introducing lightweight virtualization into the MPI interface. In particular, we outline key-value store semantics for MPI send/recv calls, thereby creating a far more expressive programming model. The general message passing semantics and imperative style of MPI application codes would remain essentially unchanged. However, the additional expressiblity of the programming model 1) enables the underlying transport layer to handle faulttolerance more transparently to the application developer, and 2) provides an evolutionary code path towards more declarative asynchronous programming models. The core contribution of this paper is an initial implementation of the DHARMA transport layer that provides the new, required functionality to support the MPI key-value store model.

AB - In this position paper, we argue for improved fault-tolerance of an MPI code by introducing lightweight virtualization into the MPI interface. In particular, we outline key-value store semantics for MPI send/recv calls, thereby creating a far more expressive programming model. The general message passing semantics and imperative style of MPI application codes would remain essentially unchanged. However, the additional expressiblity of the programming model 1) enables the underlying transport layer to handle faulttolerance more transparently to the application developer, and 2) provides an evolutionary code path towards more declarative asynchronous programming models. The core contribution of this paper is an initial implementation of the DHARMA transport layer that provides the new, required functionality to support the MPI key-value store model.

UR - https://www.scopus.com/pages/publications/84979742781

U2 - 10.1145/2751504.2751513

DO - 10.1145/2751504.2751513

M3 - Conference contribution

AN - SCOPUS:84979742781

T3 - FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015

SP - 41

EP - 46

BT - FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015

PB - Association for Computing Machinery, Inc

T2 - 5th Workshop on Fault Tolerance for HPC at eXtreme Scale, FTXS 2015

Y2 - 15 June 2015

ER -

Wilke JJ, Kolla H, Teranishi K, Hollman DS, Bennett JC, Slattengren N. Evolving the message passing programming model via a fault-tolerant, object-oriented transport layer. In FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015. Association for Computing Machinery, Inc. 2015. p. 41-46. (FTXS 2015 - Proceedings of the 2015 Workshop on Fault Tolerance for HPC at eXtreme Scale, Part of HPDC 2015). doi: 10.1145/2751504.2751513

Evolving the message passing programming model via a fault-tolerant, object-oriented transport layer

Abstract

Publication series

Conference

Funding

Access to Document

Other files and links

Fingerprint

Cite this