An Efficient Implementation of Entropic Lattice Boltzmann Method in a Hybrid CPU-GPU Computing Environment

2014 ◽  
pp. 136-148
Author(s):  
Yu Ye ◽  
Peng Chi ◽  
Yan Wang
Keyword(s):  
Lattice Boltzmann Method ◽  
Lattice Boltzmann ◽  
Gpu Computing ◽  
Efficient Implementation ◽  
Computing Environment ◽  
Boltzmann Method

ON VECTORIZATION FOR LATTICE BASED SIMULATIONS

2013 ◽  
Vol 24 (12) ◽  
pp. 1340011 ◽  
Author(s):  
ANIRUDDHA G. SHET ◽  
K. SIDDHARTH ◽  
SHAHAJHAN H. SORATHIYA ◽  
ANAND M. DESHPANDE ◽  
SUNIL D. SHERLEKAR ◽  
...  
Keyword(s):  
Data Structure ◽  
Lattice Boltzmann Method ◽  
Lattice Boltzmann ◽  
Performance Optimization ◽  
Memory Bandwidth ◽  
Bandwidth Utilization ◽  
Computing Environment ◽  
Optimization Framework ◽  
Boltzmann Method ◽  
Type Lattice

We present a vector-friendly blocked computing strategy for the lattice Boltzmann method (LBM). This strategy, along with a recently developed data structure, Structure of Arrays of Structures (SoAoS), is implemented for multi-relaxation type lattice Boltzmann (LB). The proposed methodology enables optimal memory bandwidth utilization in the advection step and high compute efficiency in the collision step of LB implementation. In a dense computing environment, current performance optimization framework for LBM is able to achieve high single-core efficiency.

Implementation of Multi-GPU Based Lattice Boltzmann Method for Flow Through Porous Media

2015 ◽  
Vol 7 (1) ◽  
pp. 1-12 ◽  
Author(s):  
Changsheng Huang ◽  
Baochang Shi ◽  
Nanzhong He ◽  
Zhenhua Chai
Keyword(s):  
Porous Media ◽  
Lattice Boltzmann Method ◽  
Lattice Boltzmann ◽  
Message Passing Interface ◽  
Gpu Computing ◽  
Processing Unit ◽  
Flow Through Porous Media ◽  
Communication Time ◽  
Flow Through ◽  
Boltzmann Method

AbstractThe lattice Boltzmann method (LBM) can gain a great amount of performance benefit by taking advantage of graphics processing unit (GPU) computing, and thus, the GPU, or multi-GPU based LBM can be considered as a promising and competent candidate in the study of large-scale fluid flows. However, the multi-GPU based lattice Boltzmann algorithm has not been studied extensively, especially for simulations of flow in complex geometries. In this paper, through coupling with the message passing interface (MPI) technique, we present an implementation of multi-GPU based LBM for fluid flow through porous media as well as some optimization strategies based on the data structure and layout, which can apparently reduce memory access and completely hide the communication time consumption. Then the performance of the algorithm is tested on a one-node cluster equipped with four Tesla C1060 GPU cards where up to 1732 MFLUPS is achieved for the Poiseuille flow and a nearly linear speedup with the number of GPUs is also observed.

A Novel Dynamics Lattice Boltzmann Method for Gas Flows

2010 ◽  
Author(s):  
C. T. Hsu ◽  
S. W. Chiang ◽  
K. F. Sin
Keyword(s):  
Boltzmann Equation ◽  
Coordinate System ◽  
Lattice Boltzmann Method ◽  
Lattice Boltzmann ◽  
Physical Space ◽  
Efficient Implementation ◽  
Gas Flows ◽  
Lattice Node ◽  
Quadrature Scheme ◽  
Boltzmann Method

The lattice Boltzmann method (LBM), where discrete velocities are specifically assigned to ensure that a particle leaves one lattice node always resides on another lattice node, has been developed for decades as a powerful numerical tool to solve the Boltzmann equation for gas flows. The efficient implementation of LBM requires that the discrete velocities be isotropic and that the lattice nodes be homogeneous. These requirements restrict the applications of the currently-used LBM schemes to incompressible and isothermal flows. Such restrictions defy the original physics of Boltzmann equation. Much effort has been devoted in the past decades to remove these restrictions, but of less success. In this paper, a novel dynamic lattice Boltzmann method (DLBM) that is free of the incompressible and isothermal restrictions is proposed and developed to simulate gas flows. This is achieved through a coordinate transformation featured with Galilean translation and thermal normalization. The transformation renders the normalized Maxwell equilibrium distribution with directional isotropy and spatial homogeneity for the accurate and efficient implementation of the Gaussian-Hermite quadrature. The transformed Boltzmann equation contains additional terms due to local convection and acceleration. The velocity quadrature points in the new coordinate system are fixed while the correspondent points in the physical space change from time to time and from position to position. By this dynamic quadrature nature in the physical space, we term this new scheme as the dynamic quadrature scheme. The lattice Boltzmann method (LBM) with the dynamic quadrature scheme is named as the dynamic lattice Boltzmann method (DLBM). The transformed Boltzmann equation is then solved in the new coordinate system based on the fixed quadrature points. Validations of the DLBM have been carried with several benchmark problems. Cavity flows problem are used. Excellent agreements are obtained as compared with those obtained from the conventional schemes. Up to date, the DLBM algorithm can run up to Mach number at 0.3 without suffering from numerical instability. The application of the DLBM to the Rayleigh-Bernard thermal instability problem is illustrated, where the onset of 2D vortex rolls and 3D hexagonal cells are well-predicted and are in excellent agreement with the theory. In summary, a novel dynamic lattice Boltzmann method (DLBM) has been proposed with algorithm developed for numerical simulation of gas flows. This new DLBM has been demonstrated to have removed the incompressible and isothermal restrictions encountered by the traditional LBM.

scholarly journals An efficient implementation of the lattice Boltzmann method for hybrid supercomputers

2015 ◽  
pp. 205-214
Author(s):  
Д.А. Бикулов
Keyword(s):  
Lattice Boltzmann Method ◽  
Lattice Boltzmann ◽  
Graphics Processing Units ◽  
Efficient Implementation ◽  
State University ◽  
Memory Space ◽  
Boltzmann Method ◽  
Graphics Processing ◽  
Moscow State University

Рассмотрены особенности эффективной реализации метода решеточных уравнений Больцмана (Lattice Boltzmann method, LBM) для гибридных суперкомпьютерных систем с множеством видеокарт. Описаны основные стратегии по сокращению требуемой для работы LBM памяти на графическом ускорителе. Представлены результаты измерения зависимости производительности реализованного программного модуля от числа задействованных видеокарт, полученные на суперкомпьютере Ломоносов. A number of features of an efficient implementation of the lattice Boltzmann method (LBM) for hybrid supercomputers with many graphics processing units (GPU) are discussed. The main strategies for reducing the memory space required by LBM are described. The performance dependence of the implemented solver on the number of the GPUs in use is analyzed for the Lomonosov supercomputer installed at Moscow State University.

Sign in / Sign up

Export Citation Format

Share Document