Vis enkel innførsel

Performance and Energy Efficiency of CUDA and OpenCL for GPU Computing using Python

dc.contributor.authorHolm, Håvard Heitlo
dc.contributor.authorBrodtkorb, André R.
dc.contributor.authorSætra, Martin Lilleeng
dc.date.accessioned2021-01-15T09:53:24Z
dc.date.accessioned2021-02-25T11:49:29Z
dc.date.available2021-01-15T09:53:24Z
dc.date.available2021-02-25T11:49:29Z
dc.date.issued2020
dc.identifier.citationHolm HH, Brodtkorb A, Sætra ML. Performance and Energy Efficiency of CUDA and OpenCL for GPU Computing using Python. Advances in Parallel Computing. 2020;36:593-604en
dc.identifier.issn0927-5452
dc.identifier.issn1879-808X
dc.identifier.urihttps://hdl.handle.net/10642/9742
dc.description.abstractIn this work, we examine the performance and energy efficiency when using Python for developing HPC codes running on the GPU. We investigate the portability of performance and energy efficiency between CUDA and OpenCL; between GPU generations; and between low-end, mid-range and high-end GPUs. Our findings show that for some combinations of GPU and GPU code, there is a significant speedup for CUDA over OpenCL, but that this does not hold in general. Our experiments show that performance in general varies more between different GPUs, than between using CUDA and OpenCL. Finally, we show that tuning for performance is a good way of tuning for energy efficiency.en
dc.description.sponsorshipThis work is supported by the Research Council of Norway through grant number 250935 (GPU Ocean). The GPU Ocean project received support from UNINETT Sigma2 - the National Infrastructure for High Performance Computing and Data Storage in Norway under project number nn9550k.en
dc.language.isoenen
dc.publisherIOS Pressen
dc.relation.ispartofseriesAdvances in Parallel Computing;Volume 36: Parallel Computing: Technology Trends
dc.rightsCreative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licenseen
dc.subjectGPU Computingen
dc.subjectCompute unified device architectureen
dc.subjectOpen compute languagesen
dc.subjectHigh performance computingen
dc.subjectShallow-water simulationsen
dc.subjectPower efficiencyen
dc.subjectOpenCLen
dc.titlePerformance and Energy Efficiency of CUDA and OpenCL for GPU Computing using Pythonen
dc.typeJournal articleen
dc.typePeer revieweden
dc.date.updated2021-01-15T09:53:23Z
dc.description.versionpublishedVersionen
dc.identifier.doihttps://doi.org/10.3233/APC200089
dc.identifier.cristin1776313
dc.source.journalAdvances in Parallel Computing
dc.relation.projectIDNorges forskningsråd: 250935
dc.relation.projectIDNorges forskningsråd: 250935 (GPU Ocean)
dc.relation.projectIDNotur/NorStore: NN9550K


Tilhørende fil(er)

Thumbnail
Filnavn:
APC-36-APC200089.pdf
Størrelse:
536.1Kb
Format:
PDF
Åpne

Denne innførselen finnes i følgende samling(er)

Vis enkel innførsel