Accelerating VASP electronic structure calculations using graphic processing units

Handle URI:
http://hdl.handle.net/10754/597443
Title:
Accelerating VASP electronic structure calculations using graphic processing units
Authors:
Hacene, Mohamed; Anciaux-Sedrakian, Ani; Rozanska, Xavier; Klahr, Diego; Guignon, Thomas; Fleurat-Lessard, Paul
Abstract:
We present a way to improve the performance of the electronic structure Vienna Ab initio Simulation Package (VASP) program. We show that high-performance computers equipped with graphics processing units (GPUs) as accelerators may reduce drastically the computation time when offloading these sections to the graphic chips. The procedure consists of (i) profiling the performance of the code to isolate the time-consuming parts, (ii) rewriting these so that the algorithms become better-suited for the chosen graphic accelerator, and (iii) optimizing memory traffic between the host computer and the GPU accelerator. We chose to accelerate VASP with NVIDIA GPU using CUDA. We compare the GPU and original versions of VASP by evaluating the Davidson and RMM-DIIS algorithms on chemical systems of up to 1100 atoms. In these tests, the total time is reduced by a factor between 3 and 8 when running on n (CPU core + GPU) compared to n CPU cores only, without any accuracy loss. © 2012 Wiley Periodicals, Inc.
Citation:
Hacene M, Anciaux-Sedrakian A, Rozanska X, Klahr D, Guignon T, et al. (2012) Accelerating VASP electronic structure calculations using graphic processing units. Journal of Computational Chemistry 33: 2581–2589. Available: http://dx.doi.org/10.1002/jcc.23096.
Publisher:
Wiley-Blackwell
Journal:
Journal of Computational Chemistry
KAUST Grant Number:
UK-C0017
Issue Date:
20-Aug-2012
DOI:
10.1002/jcc.23096
PubMed ID:
22903247
Type:
Article
ISSN:
0192-8651
Sponsors:
Contract/grant sponsor: King Abdullah University of Science and Technology (KAUST, Award No. UK-C0017).
Appears in Collections:
Publications Acknowledging KAUST Support

Full metadata record

DC FieldValue Language
dc.contributor.authorHacene, Mohameden
dc.contributor.authorAnciaux-Sedrakian, Anien
dc.contributor.authorRozanska, Xavieren
dc.contributor.authorKlahr, Diegoen
dc.contributor.authorGuignon, Thomasen
dc.contributor.authorFleurat-Lessard, Paulen
dc.date.accessioned2016-02-25T12:33:21Zen
dc.date.available2016-02-25T12:33:21Zen
dc.date.issued2012-08-20en
dc.identifier.citationHacene M, Anciaux-Sedrakian A, Rozanska X, Klahr D, Guignon T, et al. (2012) Accelerating VASP electronic structure calculations using graphic processing units. Journal of Computational Chemistry 33: 2581–2589. Available: http://dx.doi.org/10.1002/jcc.23096.en
dc.identifier.issn0192-8651en
dc.identifier.pmid22903247en
dc.identifier.doi10.1002/jcc.23096en
dc.identifier.urihttp://hdl.handle.net/10754/597443en
dc.description.abstractWe present a way to improve the performance of the electronic structure Vienna Ab initio Simulation Package (VASP) program. We show that high-performance computers equipped with graphics processing units (GPUs) as accelerators may reduce drastically the computation time when offloading these sections to the graphic chips. The procedure consists of (i) profiling the performance of the code to isolate the time-consuming parts, (ii) rewriting these so that the algorithms become better-suited for the chosen graphic accelerator, and (iii) optimizing memory traffic between the host computer and the GPU accelerator. We chose to accelerate VASP with NVIDIA GPU using CUDA. We compare the GPU and original versions of VASP by evaluating the Davidson and RMM-DIIS algorithms on chemical systems of up to 1100 atoms. In these tests, the total time is reduced by a factor between 3 and 8 when running on n (CPU core + GPU) compared to n CPU cores only, without any accuracy loss. © 2012 Wiley Periodicals, Inc.en
dc.description.sponsorshipContract/grant sponsor: King Abdullah University of Science and Technology (KAUST, Award No. UK-C0017).en
dc.publisherWiley-Blackwellen
dc.subjectaccelerated computingen
dc.subjectFortran scientific computingen
dc.subjectgraphics processing units plane-wavesen
dc.subjecthybrid computingen
dc.subjectVASPen
dc.titleAccelerating VASP electronic structure calculations using graphic processing unitsen
dc.typeArticleen
dc.identifier.journalJournal of Computational Chemistryen
dc.contributor.institutionIFP Energies nouvelles, Rueil-Malmaison, Franceen
dc.contributor.institutionUniversite de Lyon, Lyon, Franceen
dc.contributor.institutionMaterials Design, Montrouge, Franceen
dc.contributor.institutionTotal, Courbevoie, Franceen
kaust.grant.numberUK-C0017en
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.