Multi-thread implementations of the lattice Boltzmann method on non-uniform grids for CPUs and GPUs

Schönherr, Martin; Kucher, Kostyantyn; Geier, Martin; Stiebler, Maik; Freudiger, Sören; Krafczyk, Manfred

doi:10.1016/j.camwa.2011.04.012

Cited by 105 publications

(74 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Henceforth, we consider to use a Structure of Array (SoA) approach instead of using AoS. This approach has proven to be a very efficient memory mapping on CUDA-based implementations [5], [3], [16], [15], [20]. Despite that this approach is more amenable to vectorized instructions, we find a large space of memory, as big as the size of the fluid domain, among the different latticevelocities.…”

Section: Lbm Implementatonmentioning

confidence: 99%

“…Several refinement techniques have been implemented for LBM-based solvers, such as adaptive mesh refinement (AMR) [19], multi-grid [15], and multidomain [14]. Each of these techniques exhibit its own advantages and disadvantages.…”

Section: Introductionmentioning

confidence: 99%

“…A very recent work that covers a subject closely related with the present contribution is [15], where a new and efficient 2D implementation of LBM method for muti-grid flows is presented. Here, we focus on a different approach based on multidomain coupled with LBM introduced in the work presented by D. Lagrava et al [14].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Multi-domain Grid Refinement for Lattice-Boltzmann Simulations on Heterogeneous Platforms

Valero-Lara

Jansson

2015

2015 IEEE 18th International Conference on Computational Science and Engineering

View full text Add to dashboard Cite

Abstract-The main contribution of the present work consists of several parallel approaches for grid refinement based on a multi-domain decomposition for lattice-Boltzmann simulations. The proposed method for discretizing the fluid incorporates different regular Cartesian grids with no homogeneous spatial domains, which are in need to be communicated each other.Three different parallel approaches are proposed, homogeneous Multicore, homogeneous GPU, and heterogeneous Multicore-GPU. Although, the homogeneous implementations exhibit satisfactory results, the heterogeneous approach achieves up to 30% extra efficiency, in terms of Millions of Fluid Lattice Updates per Second (MFLUPS), by overlapping some of the steps on both architectures, Multicore and GPU.

show abstract

Section: Lbm Implementatonmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multi-domain Grid Refinement for Lattice-Boltzmann Simulations on Heterogeneous Platforms

Valero-Lara

Jansson

2015

2015 IEEE 18th International Conference on Computational Science and Engineering

View full text Add to dashboard Cite

show abstract

“…The computational multi-domain algorithm is summarized in Appendix C. It is worth noting that the fine grid scales that can not be resolved by the coarse grid are removed using a filtering operator. For more details, as well as alternative algorithms the reader should refer to [45,46,47,48,49]. The overlapping between a fine and a coarse grid is illustrated in Fig.…”

Section: Multi-domain Algorithm and Domain Boundary Conditionsmentioning

confidence: 99%

PROTEUS: A coupled iterative force-correction immersed-boundary cascaded lattice Boltzmann solver for moving and deformable boundary applications

Falagkaris

Ingram

Markakis

et al. 2018

Computers & Mathematics with Applications

View full text Add to dashboard Cite

Many realistic fluid flow problems are characterised by high Reynolds numbers and complex moving or deformable geometries. In our previous study, we presented a novel coupling between an iterative force-correction immersed boundary and a multi-domain cascaded lattice Boltzmann method, Falagkaris et al., and investigated flows around rigid bodies at Reynolds numbers up to 10 5 . Here, we extend its application to flows around moving and deformable bodies with prescribed motions. Emphasis is given on the influence of the internal mass on the computation of the aerodynamic forces including deforming boundary applications where the rigid body approximation is no longer valid. Both the rigid body and the internal Lagrangian points approximations are examined. The resulting solver has been applied to viscous flows around an in-line oscillating cylinder, a pitching foil, a plunging SD7003 airfoil and a plunging and flapping NACA-0014 airfoil. Good agreement with experimental results and other numerical schemes has been obtained. It is shown that the internal Lagrangian points approximation accurately captures the internal mass effects in linear and angular motions, as well as in deforming motions, at Reynolds numbers up to 4 · 10 4 . In all cases, the aerodynamic loads are significantly affected by the internal fluid forces.

show abstract

“…Other implementations [7,9,32,34] have used the traditional implementation sketched in Algorithm 2, which performs the collision operation before the streaming. It is known as the "push" scheme [28] since it loads the distribution function from the current lattice point and then it "pushes" (scatters) the updated values to its neighbours.…”

mentioning

confidence: 99%

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

Valero-Lara

Igual

Prieto-Matías

et al. 2015

Journal of Computational Science

View full text Add to dashboard Cite

a b s t r a c tWe 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 hostaccelerator 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.

show abstract

Multi-thread implementations of the lattice Boltzmann method on non-uniform grids for CPUs and GPUs

Cited by 105 publications

References 19 publications

Multi-domain Grid Refinement for Lattice-Boltzmann Simulations on Heterogeneous Platforms

Multi-domain Grid Refinement for Lattice-Boltzmann Simulations on Heterogeneous Platforms

PROTEUS: A coupled iterative force-correction immersed-boundary cascaded lattice Boltzmann solver for moving and deformable boundary applications

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

Contact Info

Product

Resources

About