Adjusting process count on demand for petascale global optimization

Handle URI:
http://hdl.handle.net/10754/597469
Title:
Adjusting process count on demand for petascale global optimization
Authors:
Sosonkina, Masha; Watson, Layne T.; Radcliffe, Nicholas R.; Haftka, Rafael T.; Trosset, Michael W.
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.
Citation:
Sosonkina M, Watson LT, Radcliffe NR, Haftka RT, Trosset MW (2013) Adjusting process count on demand for petascale global optimization. Parallel Computing 39: 21–35. Available: http://dx.doi.org/10.1016/j.parco.2012.11.001.
Publisher:
Elsevier BV
Journal:
Parallel Computing
Issue Date:
Jan-2013
DOI:
10.1016/j.parco.2012.11.001
Type:
Article
ISSN:
0167-8191
Sponsors:
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.
Appears in Collections:
Publications Acknowledging KAUST Support

Full metadata record

DC FieldValue Language
dc.contributor.authorSosonkina, Mashaen
dc.contributor.authorWatson, Layne T.en
dc.contributor.authorRadcliffe, Nicholas R.en
dc.contributor.authorHaftka, Rafael T.en
dc.contributor.authorTrosset, Michael W.en
dc.date.accessioned2016-02-25T12:40:20Zen
dc.date.available2016-02-25T12:40:20Zen
dc.date.issued2013-01en
dc.identifier.citationSosonkina M, Watson LT, Radcliffe NR, Haftka RT, Trosset MW (2013) Adjusting process count on demand for petascale global optimization. Parallel Computing 39: 21–35. Available: http://dx.doi.org/10.1016/j.parco.2012.11.001.en
dc.identifier.issn0167-8191en
dc.identifier.doi10.1016/j.parco.2012.11.001en
dc.identifier.urihttp://hdl.handle.net/10754/597469en
dc.description.abstractThere 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.en
dc.description.sponsorshipThe 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.en
dc.publisherElsevier BVen
dc.titleAdjusting process count on demand for petascale global optimizationen
dc.typeArticleen
dc.identifier.journalParallel Computingen
dc.contributor.institutionDepartment of Modeling, Simulation and Visualization Engineering, Old Dominion University, Norfolk, VA, USAen
dc.contributor.institutionU.S. DOE Ames Laboratory, Iowa State University, Ames, IA, USAen
dc.contributor.institutionDepartment of Computer Science, Virginia Polytechnic Institute and State University, Blacksburg, VA, USAen
dc.contributor.institutionDepartment of Mathematics, Virginia Polytechnic Institute and State University, Blacksburg, VA, USAen
dc.contributor.institutionDepartment of Mechanical and Aerospace Engineering, University of Florida, Gainesville, FL, USAen
dc.contributor.institutionDepartment of Statistics, Indiana University, Bloomington, IN, USAen
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.