Parallelisation of the 3D Fast Fourier Transform Using the Hybrid OpenMP/MPI Decomposition

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F14%3APU112218" target="_blank" >RIV/00216305:26230/14:PU112218 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.fit.vut.cz/research/publication/10706/" target="_blank" >https://www.fit.vut.cz/research/publication/10706/</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-319-14896-0_9" target="_blank" >10.1007/978-3-319-14896-0_9</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Parallelisation of the 3D Fast Fourier Transform Using the Hybrid OpenMP/MPI Decomposition

Popis výsledku v původním jazyce

The 3D fast Fourier transform (FFT) is the heart of many simulation methods. Although the efficient parallelisation of the FFT has been deeply studied over last few decades, many researchers only focused on either pure message passing (MPI) or shared memory (OpenMP) implementations. Unfortunately, pure MPI approaches cannot exploit the shared memory within the cluster node and the OpenMP cannot scale over multiple nodes. This paper proposes a 2D hybrid decomposition of the 3D FFT where the domain is decomposed over the first axis by means of MPI while over the second axis by means of OpenMP. The performance of the pro-posed method is thoroughly compared with the state of the art libraries (FFTW, PFFT, P3DFFT) on three supercomputer systems with up to 16k cores. The experimental results show that the hybrid implementation offers 10-20% higher performance and better scaling especially for high core counts.

Název v anglickém jazyce

Parallelisation of the 3D Fast Fourier Transform Using the Hybrid OpenMP/MPI Decomposition

Popis výsledku anglicky

The 3D fast Fourier transform (FFT) is the heart of many simulation methods. Although the efficient parallelisation of the FFT has been deeply studied over last few decades, many researchers only focused on either pure message passing (MPI) or shared memory (OpenMP) implementations. Unfortunately, pure MPI approaches cannot exploit the shared memory within the cluster node and the OpenMP cannot scale over multiple nodes. This paper proposes a 2D hybrid decomposition of the 3D FFT where the domain is decomposed over the first axis by means of MPI while over the second axis by means of OpenMP. The performance of the pro-posed method is thoroughly compared with the state of the art libraries (FFTW, PFFT, P3DFFT) on three supercomputer systems with up to 16k cores. The experimental results show that the hybrid implementation offers 10-20% higher performance and better scaling especially for high core counts.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

V - Vyzkumna aktivita podporovana z jinych verejnych zdroju

Rok uplatnění

2014

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Mathematical and Engineering Methods in Computer Science

ISBN

978-3-319-14895-3

ISSN

—

e-ISSN

—

Počet stran výsledku

13

Strana od-do

100-112

Název nakladatele

Springer International Publishing

Místo vydání

Heidelberg

Místo konání akce

Telč

Datum konání akce

17. 10. 2014

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

000357573300009