Adjusting process count on demand for petascale global optimization

Masha Sosonkina; Layne T. Watson; Nicholas R. Radcliffe; Rafael T. Haftka; Michael W. Trosset

doi:10.1016/j.parco.2012.11.001

Adjusting process count on demand for petascale global optimization

Masha Sosonkina, Layne T. Watson, Nicholas R. Radcliffe, Rafael T. Haftka, Michael W. Trosset

Research output: Contribution to journal › Article › peer-review

Abstract

There are many challenges that need to be met before efficient and reliable computation at the petascale is possible. Many scientific and engineering codes running at the petascale are likely to be memory intensive, which makes thrashing a serious problem for many petascale applications. One way to overcome this challenge is to use a dynamic number of processes, so that the total amount of memory available for the computation can be increased on demand. This paper describes modifications made to the massively parallel global optimization code pVTdirect in order to allow for a dynamic number of processes. In particular, the modified version of the code monitors memory use and spawns new processes if the amount of available memory is determined to be insufficient. The primary design challenges are discussed, and performance results are presented and analyzed.

Original language	English (US)
Pages (from-to)	21-35
Number of pages	15
Journal	Parallel Computing
Volume	39
Issue number	1
DOIs	https://doi.org/10.1016/j.parco.2012.11.001
State	Published - Jan 2013
Externally published	Yes

Bibliographical note

KAUST Repository Item: Exported on 2020-10-01
Acknowledgements: The authors thank the National Energy Research Scientific Computing Center (NERSC) for use of the Carver cluster, and Aron Ahmadia at the King Abdullah University of Science and Technology (KAUST) for use of the Shaheen and Neser clusters. The authors are thankful to the anonymous reviewers for their insights that helped improve the paper.
This publication acknowledges KAUST support, but has no KAUST affiliated authors.

Access to Document

10.1016/j.parco.2012.11.001

Cite this

@article{dc8e9ae5a79d4553a69a4a08bc5b1265,

title = "Adjusting process count on demand for petascale global optimization",

abstract = "There are many challenges that need to be met before efficient and reliable computation at the petascale is possible. Many scientific and engineering codes running at the petascale are likely to be memory intensive, which makes thrashing a serious problem for many petascale applications. One way to overcome this challenge is to use a dynamic number of processes, so that the total amount of memory available for the computation can be increased on demand. This paper describes modifications made to the massively parallel global optimization code pVTdirect in order to allow for a dynamic number of processes. In particular, the modified version of the code monitors memory use and spawns new processes if the amount of available memory is determined to be insufficient. The primary design challenges are discussed, and performance results are presented and analyzed.",

author = "Masha Sosonkina and Watson, {Layne T.} and Radcliffe, {Nicholas R.} and Haftka, {Rafael T.} and Trosset, {Michael W.}",

note = "KAUST Repository Item: Exported on 2020-10-01 Acknowledgements: The authors thank the National Energy Research Scientific Computing Center (NERSC) for use of the Carver cluster, and Aron Ahmadia at the King Abdullah University of Science and Technology (KAUST) for use of the Shaheen and Neser clusters. The authors are thankful to the anonymous reviewers for their insights that helped improve the paper. This publication acknowledges KAUST support, but has no KAUST affiliated authors.",

year = "2013",

month = jan,

doi = "10.1016/j.parco.2012.11.001",

language = "English (US)",

volume = "39",

pages = "21--35",

journal = "Parallel Computing",

issn = "0167-8191",

publisher = "Elsevier B.V.",

number = "1",

}

TY - JOUR

T1 - Adjusting process count on demand for petascale global optimization

AU - Sosonkina, Masha

AU - Watson, Layne T.

AU - Radcliffe, Nicholas R.

AU - Haftka, Rafael T.

AU - Trosset, Michael W.

N1 - KAUST Repository Item: Exported on 2020-10-01 Acknowledgements: The authors thank the National Energy Research Scientific Computing Center (NERSC) for use of the Carver cluster, and Aron Ahmadia at the King Abdullah University of Science and Technology (KAUST) for use of the Shaheen and Neser clusters. The authors are thankful to the anonymous reviewers for their insights that helped improve the paper. This publication acknowledges KAUST support, but has no KAUST affiliated authors.

PY - 2013/1

Y1 - 2013/1

N2 - There are many challenges that need to be met before efficient and reliable computation at the petascale is possible. Many scientific and engineering codes running at the petascale are likely to be memory intensive, which makes thrashing a serious problem for many petascale applications. One way to overcome this challenge is to use a dynamic number of processes, so that the total amount of memory available for the computation can be increased on demand. This paper describes modifications made to the massively parallel global optimization code pVTdirect in order to allow for a dynamic number of processes. In particular, the modified version of the code monitors memory use and spawns new processes if the amount of available memory is determined to be insufficient. The primary design challenges are discussed, and performance results are presented and analyzed.

AB - There are many challenges that need to be met before efficient and reliable computation at the petascale is possible. Many scientific and engineering codes running at the petascale are likely to be memory intensive, which makes thrashing a serious problem for many petascale applications. One way to overcome this challenge is to use a dynamic number of processes, so that the total amount of memory available for the computation can be increased on demand. This paper describes modifications made to the massively parallel global optimization code pVTdirect in order to allow for a dynamic number of processes. In particular, the modified version of the code monitors memory use and spawns new processes if the amount of available memory is determined to be insufficient. The primary design challenges are discussed, and performance results are presented and analyzed.

UR - http://hdl.handle.net/10754/597469

UR - https://linkinghub.elsevier.com/retrieve/pii/S0167819112000890

U2 - 10.1016/j.parco.2012.11.001

DO - 10.1016/j.parco.2012.11.001

M3 - Article

SN - 0167-8191

VL - 39

SP - 21

EP - 35

JO - Parallel Computing

JF - Parallel Computing

IS - 1

ER -

Adjusting process count on demand for petascale global optimization

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this