Deployment on GPUs of an application in computational atomic physics

P. Fortin, R. Habel, F. Jézéquel, J.-L. Lamotte, N.S. Scott

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

This paper describes the deployment on GPUs of PROP, a program of the 2DRMP suite which models electron collisions with H-like atoms and ions. Because performance on GPUs is better in single precision than in double precision, the numerical stability of the PROP program in single precision has been studied. The numerical quality of PROP results computed in single precision and their impact on the next program of the 2DRMP suite has been analyzed. Successive versions of the PROP program on GPUs have been developed in order to improve its performance. Particular attention has been paid to the optimization of data transfers and of linear algebra operations. Performance obtained on several architectures (including NVIDIA Fermi) are presented.
Original languageEnglish
Title of host publication25th IEEE International Symposium on Parallel and Distributed Processing
Subtitle of host publicationIPDPS 2011
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Pages1359-1366
Number of pages8
ISBN (Electronic)978-0-7695-4577-6
ISBN (Print)978-1-61284-425-1
DOIs
Publication statusPublished - May 2011
Event12th IEEE International Workshop on Parallel and Distributed Scientific and Engineering Computing (PDSEC-11) - Anchorage, Alaska, United States
Duration: 16 May 201120 May 2011

Conference

Conference12th IEEE International Workshop on Parallel and Distributed Scientific and Engineering Computing (PDSEC-11)
CountryUnited States
CityAnchorage, Alaska
Period16/05/201120/05/2011

Bibliographical note

ISSN: 1530-2075

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Software
  • Theoretical Computer Science

Fingerprint Dive into the research topics of 'Deployment on GPUs of an application in computational atomic physics'. Together they form a unique fingerprint.

Cite this