Global Memory Access Modelling for Efficient Implementation of the Lattice Boltzmann Method on Graphics Processing Units

被引：0

作者：

Obrecht, Christian ^{[1
]}

Kuznik, Frederic ^{[1
]}

Tourancheau, Bernard ^{[2
]}

Roux, Jean-Jacques ^{[1
]}

机构：

[1] Univ Lyon, INSA Lyon, CNRS,Ctr Thermique Lyon, UMR 5008, Bat Sadi Carnot,9 Rue Phys, F-69621 Villeurbanne, France

[2] UCB Lyon 1, UMR 5668,ENS Lyon, INRIA, Lab Informat Parallelisme,CNRS, Lyon 63964, France

来源：

HIGH PERFORMANCE COMPUTING FOR COMPUTATIONAL SCIENCE - VECPAR 2010 | 2011年 / 6449卷

关键词：

GPU computing; CUDA; lattice Boltzmann method; CFD;

D O I：

暂无

中图分类号：

TP31 [计算机软件];

学科分类号：

081202 ; 0835 ;

摘要：

In this work, we investigate the global memory access mechanism on recent GPUs. For the purpose of this study, we created specific benchmark programs, which allowed us to explore the scheduling of global memory transactions. Thus, we formulate a model capable of estimating the execution time for a large class of applications. Our main goal is to facilitate optimisation of regular data-parallel applications on GPUs. As an example, we finally describe our CUDA implementations of LBM flow solvers on which our model was able to estimate performance with less than 5% relative error.

引用

页码：151 / +

页数：2

共 50 条

[41] Local strong form meshless method on multiple Graphics Processing Units
Kosec, G.
Zinterhof, P.
CMES-COMPUTER MODELING IN ENGINEERING & SCIENCES, 2013, 91 (05): : 377 - 396
[42] An efficient implementation of nodal discontinuous Galerkin lattice Boltzmann method and validation for direct numerical simulation of turbulent flows
Xia, Boyang
Li, Jun
COMPUTERS & MATHEMATICS WITH APPLICATIONS, 2022, 117 : 284 - 298
[43] LATTICE BOLTZMANN SIMULATIONS OF CAVITY FLOWS ON GRAPHIC PROCESSING UNIT WITH MEMORY MANAGEMENT
Hong, P. Y.
Huang, L. M.
Chang, C. Y.
Lin, C. A.
JOURNAL OF MECHANICS, 2017, 33 (06) : 863 - 871
[44] Modelling of solidification process of aluminium foams using lattice Boltzmann method
Diop, M.
Hao, H.
Dong, H. -W.
Zhang, X. -G.
Yao, S.
Jin, J. -Z.
INTERNATIONAL JOURNAL OF CAST METALS RESEARCH, 2011, 24 (3-4) : 158 - 162
[45] A note on modelling non-rectangular boundaries by the Lattice Boltzmann Method
Asian, E.
Nahavandi, A.
Taymaz, I.
Benim, A. C.
PROGRESS IN COMPUTATIONAL FLUID DYNAMICS, 2012, 12 (06): : 433 - 438
[46] Implementation of the Lattice Boltzmann Method on Heterogeneous Hardware and Platforms using OpenCL
Tekic, Predrag M.
Radjenovic, Jelena B.
Rackovic, Milos
ADVANCES IN ELECTRICAL AND COMPUTER ENGINEERING, 2012, 12 (01) : 51 - 56
[47] A highly-efficient locally encoded boundary scheme for lattice Boltzmann method on GPU
Zhang, Zehua
Peng, Cheng
Li, Chengxiang
Zhang, Hua
Xian, Tao
Wang, Lian-Ping
COMPUTER PHYSICS COMMUNICATIONS, 2024, 298
[48] Sailfish: A flexible multi-GPU implementation of the lattice Boltzmann method
Januszewski, M.
Kostur, M.
COMPUTER PHYSICS COMMUNICATIONS, 2014, 185 (09) : 2350 - 2368
[49] Application of Lattice Boltzmann Method for fluid flow modelling of FSLDR domain
Bhatt, Tirthraj
Perumal, D. Arumuga
MATERIALS TODAY-PROCEEDINGS, 2020, 22 : 2066 - 2073
[50] Simulation of fluid-structure interaction in a microchannel using the lattice Boltzmann method and size-dependent beam element on a graphics processing unit
Esfahanian, Vahid
Dehdashti, Esmaeil
Dehrouye-Semnani, Amir Mehdi
CHINESE PHYSICS B, 2014, 23 (08)

← 1 2 3 4 5 →