Design and performance characterization of electronic structure calculations on massively parallel supercomputers

A case study of GPAW on the Blue Gene/P architecture

N. A. Romero, Christian Glinsvad, Ask Hjorth Larsen, J. Enkovaara, S. Shende, V. A. Morozov, Jens Jørgen Mortensen

Research output: Contribution to journalJournal articleResearchpeer-review

Abstract

Density function theory (DFT) is the most widely employed electronic structure method because of its favorable scaling with system size and accuracy for a broad range of molecular and condensed-phase systems. The advent of massively parallel supercomputers has enhanced the scientific community's ability to study larger system sizes. Ground-state DFT calculations on∼103 valence electrons using traditional O(N3) algorithms can be routinely performed on present-day supercomputers. The performance characteristics of these massively parallel DFT codes on>104 computer cores are not well understood. The GPAW code was ported an optimized for the Blue Gene/P architecture. We present our algorithmic parallelization strategy and interpret the results for a number of benchmark test cases.
Original languageEnglish
JournalConcurrency and Computation: Practice & Experience
Volume27
Issue number1
Pages (from-to)69-93
ISSN1532-0626
DOIs
Publication statusPublished - 2013

Keywords

  • Blue gene
  • DFT
  • Electronic structure
  • GPAW
  • High-performance computing
  • Massive parallelization

Cite this

@article{ef8ec5ab70c94238bd03c16fe07ea6c0,
title = "Design and performance characterization of electronic structure calculations on massively parallel supercomputers: A case study of GPAW on the Blue Gene/P architecture",
abstract = "Density function theory (DFT) is the most widely employed electronic structure method because of its favorable scaling with system size and accuracy for a broad range of molecular and condensed-phase systems. The advent of massively parallel supercomputers has enhanced the scientific community's ability to study larger system sizes. Ground-state DFT calculations on∼103 valence electrons using traditional O(N3) algorithms can be routinely performed on present-day supercomputers. The performance characteristics of these massively parallel DFT codes on>104 computer cores are not well understood. The GPAW code was ported an optimized for the Blue Gene/P architecture. We present our algorithmic parallelization strategy and interpret the results for a number of benchmark test cases.",
keywords = "Blue gene, DFT, Electronic structure, GPAW, High-performance computing, Massive parallelization",
author = "Romero, {N. A.} and Christian Glinsvad and Larsen, {Ask Hjorth} and J. Enkovaara and S. Shende and Morozov, {V. A.} and Mortensen, {Jens J{\o}rgen}",
year = "2013",
doi = "10.1002/cpe.3199",
language = "English",
volume = "27",
pages = "69--93",
journal = "Concurrency and Computation: Practice & Experience",
issn = "1532-0626",
publisher = "John Wiley & Sons Ltd",
number = "1",

}

Design and performance characterization of electronic structure calculations on massively parallel supercomputers : A case study of GPAW on the Blue Gene/P architecture. / Romero, N. A.; Glinsvad, Christian; Larsen, Ask Hjorth; Enkovaara, J.; Shende, S.; Morozov, V. A.; Mortensen, Jens Jørgen.

In: Concurrency and Computation: Practice & Experience, Vol. 27, No. 1, 2013, p. 69-93.

Research output: Contribution to journalJournal articleResearchpeer-review

TY - JOUR

T1 - Design and performance characterization of electronic structure calculations on massively parallel supercomputers

T2 - A case study of GPAW on the Blue Gene/P architecture

AU - Romero, N. A.

AU - Glinsvad, Christian

AU - Larsen, Ask Hjorth

AU - Enkovaara, J.

AU - Shende, S.

AU - Morozov, V. A.

AU - Mortensen, Jens Jørgen

PY - 2013

Y1 - 2013

N2 - Density function theory (DFT) is the most widely employed electronic structure method because of its favorable scaling with system size and accuracy for a broad range of molecular and condensed-phase systems. The advent of massively parallel supercomputers has enhanced the scientific community's ability to study larger system sizes. Ground-state DFT calculations on∼103 valence electrons using traditional O(N3) algorithms can be routinely performed on present-day supercomputers. The performance characteristics of these massively parallel DFT codes on>104 computer cores are not well understood. The GPAW code was ported an optimized for the Blue Gene/P architecture. We present our algorithmic parallelization strategy and interpret the results for a number of benchmark test cases.

AB - Density function theory (DFT) is the most widely employed electronic structure method because of its favorable scaling with system size and accuracy for a broad range of molecular and condensed-phase systems. The advent of massively parallel supercomputers has enhanced the scientific community's ability to study larger system sizes. Ground-state DFT calculations on∼103 valence electrons using traditional O(N3) algorithms can be routinely performed on present-day supercomputers. The performance characteristics of these massively parallel DFT codes on>104 computer cores are not well understood. The GPAW code was ported an optimized for the Blue Gene/P architecture. We present our algorithmic parallelization strategy and interpret the results for a number of benchmark test cases.

KW - Blue gene

KW - DFT

KW - Electronic structure

KW - GPAW

KW - High-performance computing

KW - Massive parallelization

U2 - 10.1002/cpe.3199

DO - 10.1002/cpe.3199

M3 - Journal article

VL - 27

SP - 69

EP - 93

JO - Concurrency and Computation: Practice & Experience

JF - Concurrency and Computation: Practice & Experience

SN - 1532-0626

IS - 1

ER -