Accelerating fluid-solid simulations (Lattice-Boltzmann & Immersed-Boundary) on heterogeneous architectures

Pedro Valero-Lara; Francisco D. Igual; Manuel Prieto-Matías; Alfredo Pinelli; Julien Favier

doi:10.1016/j.jocs.2015.07.002

Accelerating fluid-solid simulations (Lattice-Boltzmann & Immersed-Boundary) on heterogeneous architectures

Pedro Valero-Lara, Francisco D. Igual, Manuel Prieto-Matías, Alfredo Pinelli, Julien Favier

Research output: Contribution to journal › Article › peer-review

39 Scopus citations

Abstract

We propose a numerical approach based on the Lattice-Boltzmann (LBM) and Immersed Boundary (IB) methods to tackle the problem of the interaction of solids with an incompressible fluid flow, and its implementation on heterogeneous platforms based on data-parallel accelerators such as NVIDIA GPUs and the Intel Xeon Phi. We explain in detail the parallelization of these methods and describe a number of optimizations, mainly focusing on improving memory management and reducing the cost of host-accelerator communication. As previous research has consistently shown, pure LBM simulations are able to achieve good performance results on heterogeneous systems thanks to the high parallel efficiency of this method. Unfortunately, when coupling LBM and IB methods, the overheads of IB degrade the overall performance. As an alternative, we have explored different hybrid implementations that effectively hide such overheads and allow us to exploit both the multi-core and the hardware accelerator in a cooperative way, with excellent performance results.

Original language	English
Pages (from-to)	249-261
Number of pages	13
Journal	Journal of Computational Science
Volume	10
DOIs	https://doi.org/10.1016/j.jocs.2015.07.002
State	Published - Sep 1 2015
Externally published	Yes

Funding

This research is supported by the Spanish Government Research Contracts TIN2012-32180, Ingenio 2010 Consolider ESP00C-07-20811, MTM2013-40824 and SEV-2013-0323, by the EU-FET grant EUNISON 308874 and by the Basque government BERC 2014-2017 contract. We also thank the support of the computing facilities of Extremadura Research Centre for Advanced Technologies (CETA-CIEMAT) and NVIDIA GPU Research Center program for the provided resources.

Funders	Funder number
BERC
EU-FET	EUNISON 308874
Seventh Framework Programme	308874

Keywords

Computational fluid dynamics
Fluid-solid interaction
Heterogeneous computing
Immersed-Boundary method
Lattice-Boltzmann method
Parallel computing

Access to Document

10.1016/j.jocs.2015.07.002

Cite this

@article{44b1059242f4433fa86d3318c8014ddc,

title = "Accelerating fluid-solid simulations (Lattice-Boltzmann & Immersed-Boundary) on heterogeneous architectures",

abstract = "We propose a numerical approach based on the Lattice-Boltzmann (LBM) and Immersed Boundary (IB) methods to tackle the problem of the interaction of solids with an incompressible fluid flow, and its implementation on heterogeneous platforms based on data-parallel accelerators such as NVIDIA GPUs and the Intel Xeon Phi. We explain in detail the parallelization of these methods and describe a number of optimizations, mainly focusing on improving memory management and reducing the cost of host-accelerator communication. As previous research has consistently shown, pure LBM simulations are able to achieve good performance results on heterogeneous systems thanks to the high parallel efficiency of this method. Unfortunately, when coupling LBM and IB methods, the overheads of IB degrade the overall performance. As an alternative, we have explored different hybrid implementations that effectively hide such overheads and allow us to exploit both the multi-core and the hardware accelerator in a cooperative way, with excellent performance results.",

keywords = "Computational fluid dynamics, Fluid-solid interaction, Heterogeneous computing, Immersed-Boundary method, Lattice-Boltzmann method, Parallel computing",

author = "Pedro Valero-Lara and Igual, {Francisco D.} and Manuel Prieto-Mat{\'i}as and Alfredo Pinelli and Julien Favier",

note = "Publisher Copyright: {\textcopyright} 2015 Elsevier B.V.",

year = "2015",

month = sep,

day = "1",

doi = "10.1016/j.jocs.2015.07.002",

language = "English",

volume = "10",

pages = "249--261",

journal = "Journal of Computational Science",

issn = "1877-7503",

publisher = "Elsevier",

}

TY - JOUR

T1 - Accelerating fluid-solid simulations (Lattice-Boltzmann & Immersed-Boundary) on heterogeneous architectures

AU - Valero-Lara, Pedro

AU - Igual, Francisco D.

AU - Prieto-Matías, Manuel

AU - Pinelli, Alfredo

AU - Favier, Julien

PY - 2015/9/1

Y1 - 2015/9/1

N2 - We propose a numerical approach based on the Lattice-Boltzmann (LBM) and Immersed Boundary (IB) methods to tackle the problem of the interaction of solids with an incompressible fluid flow, and its implementation on heterogeneous platforms based on data-parallel accelerators such as NVIDIA GPUs and the Intel Xeon Phi. We explain in detail the parallelization of these methods and describe a number of optimizations, mainly focusing on improving memory management and reducing the cost of host-accelerator communication. As previous research has consistently shown, pure LBM simulations are able to achieve good performance results on heterogeneous systems thanks to the high parallel efficiency of this method. Unfortunately, when coupling LBM and IB methods, the overheads of IB degrade the overall performance. As an alternative, we have explored different hybrid implementations that effectively hide such overheads and allow us to exploit both the multi-core and the hardware accelerator in a cooperative way, with excellent performance results.

AB - We propose a numerical approach based on the Lattice-Boltzmann (LBM) and Immersed Boundary (IB) methods to tackle the problem of the interaction of solids with an incompressible fluid flow, and its implementation on heterogeneous platforms based on data-parallel accelerators such as NVIDIA GPUs and the Intel Xeon Phi. We explain in detail the parallelization of these methods and describe a number of optimizations, mainly focusing on improving memory management and reducing the cost of host-accelerator communication. As previous research has consistently shown, pure LBM simulations are able to achieve good performance results on heterogeneous systems thanks to the high parallel efficiency of this method. Unfortunately, when coupling LBM and IB methods, the overheads of IB degrade the overall performance. As an alternative, we have explored different hybrid implementations that effectively hide such overheads and allow us to exploit both the multi-core and the hardware accelerator in a cooperative way, with excellent performance results.

KW - Computational fluid dynamics

KW - Fluid-solid interaction

KW - Heterogeneous computing

KW - Immersed-Boundary method

KW - Lattice-Boltzmann method

KW - Parallel computing

UR - http://www.scopus.com/inward/record.url?scp=84941285531&partnerID=8YFLogxK

U2 - 10.1016/j.jocs.2015.07.002

DO - 10.1016/j.jocs.2015.07.002

M3 - Article

AN - SCOPUS:84941285531

SN - 1877-7503

VL - 10

SP - 249

EP - 261

JO - Journal of Computational Science

JF - Journal of Computational Science

ER -

Accelerating fluid-solid simulations (Lattice-Boltzmann & Immersed-Boundary) on heterogeneous architectures

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this