Detail výsledku
Parallelisation of the 3D Fast Fourier Transform Using the Hybrid OpenMP/MPI Decomposition
Jaroš Jiří, prof. Ing., Ph.D., UPSY (FIT)
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.
Fast Fourier Transform, Decompostion, Hybrid Paralellisation, OpenMP, MPI.
@inproceedings{BUT111848,
author="Vojtěch {Nikl} and Jiří {Jaroš}",
title="Parallelisation of the 3D Fast Fourier Transform Using the Hybrid OpenMP/MPI Decomposition",
booktitle="Mathematical and Engineering Methods in Computer Science",
year="2014",
series="Lecture Notes in Computer Science",
volume="8934",
pages="100--112",
publisher="Springer International Publishing",
address="Heidelberg",
doi="10.1007/978-3-319-14896-0\{_}9",
isbn="978-3-319-14895-3",
url="https://www.fit.vut.cz/research/publication/10706/"
}Centrum excelence IT4Innovations, MŠMT, Operační program Výzkum a vývoj pro inovace, ED1.1.00/02.0070, zahájení: 2011-01-01, ukončení: 2015-12-31, ukončen
Simulace šíření ultrazvukových vln v lidském mozku, OVZ, Ostatní veřejné zdroje financování nepatřící do institucionálních zdrojů ČR, SoMoPro-II-2013-IG, zahájení: 2014-01-01, ukončení: 2016-12-31, ukončen