Unsteady CFD computations using vertex‐centered finite volumes for unstructured grids on Graphics Processing Units

Asouti, Varvara; Trompoukis, Xenofon; Kampolis, Ioannis C.; Giannakoglou, Kyriakos C.

doi:10.1002/fld.2352

Cited by 75 publications

(48 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The authors, later in [9] also describe a semi-automatic script based method of porting a large CFD code written in Fortran/OpenMP to NVIDIA CUDA. Similarly [7] reports the GPU performance of a Navier-Stokes solver for steady and unsteady turbulent flows on unstructured/hybrid grids. The computations were carried out on NVIDIA's GeForce GTX 285 graphics cards (in double precision arithmetic) and speed-ups up to 46× (vs a single core of two Quad Core Intel Xeon CPUs at 2.00 GHz) are reported.…”

Section: Related Workmentioning

confidence: 99%

Designing OP2 for GPU architectures

Giles

Mudalige

Spencer

et al. 2013

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

OP2 is an "active" library framework for the solution of unstructured mesh applications. It aims to decouple the specification of a scientific application from its parallel implementation to achieve code longevity and near-optimal performance through re-targeting the back-end to different multi-core/many-core hardware. This paper presents the design of the current OP2 library for generating efficient code targeting contemporary GPU platforms. In this we focus on some of the software architecture design choices and low-level optimizations to maximize performance on NVIDIA's Fermi architecture GPUs. The performance impact of these design choices is quantified on two NVIDIA GPUs (GTX560Ti, Tesla C2070) using the end-to-end performance of an industrial representative CFD application developed using the OP2 API. Results show that for each system, a number of key configuration parameters need to be set carefully in order to gain good performance. Utilizing a recently developed auto-tuning framework, we explore the effect of these parameters, their limitations and insights into optimizations for improved performance.

show abstract

Section: Related Workmentioning

confidence: 99%

Designing OP2 for GPU architectures

Giles

Mudalige

Spencer

et al. 2013

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

show abstract

“…Even in unsteady calculations [31][32][33] on GPU device, either no mesh deformation or dynamic overset grid technique is studied. Furthermore, the developed GPU-based solvers are implemented by CUDA programming model in which the core functions of the source code require the programmers to redesign in CUDA-kernel subroutines.…”

Section: Introductionmentioning

confidence: 99%

Parallelization of an Unsteady ALE Solver with Deforming Mesh Using OpenACC

Yuan

et al. 2017

Scientific Programming

View full text Add to dashboard Cite

This paper presents a parallel, GPU-based, deforming mesh-enabled unsteady numerical solver for solving moving body problems by using OpenACC. Both the 2D and 3D parallel algorithms based on spring-like deforming mesh methods are proposed and then implemented through OpenACC programming model. Furthermore, these algorithms are coupled with an unstructured mesh based, implicit time scheme integrated numerical solver, which makes the full GPU version of the solver capable of handling unsteady calculations with deforming mesh. Experiments results show that the proposed parallel deforming mesh algorithm can achieve over 2.5x speedup on K20 GPU card in comparison with 20 OpenMP threads on Intel E5-2658 V2 CPU cores. And both 2D and 3D cases are conducted to validate the efficiency, correctness, and accuracy of the present solver.

show abstract

“…He pointed out that several researchers have developed/ported CFD software to GPUs and founded significant speedups (10-50 times, depending on algorithm, approach and implementation) over a singlecore CPU. [10][11][12] Although computing unified device architecture (CUDA) reduces the difficulty of GPU general-purpose computing, porting existing CPU codes to run on the GPU requires the user to write kernels that execute on multiple cores, which hinders the use of researchers. In order to achieve semi-automatic or fully automatic from CPU to GPU, Corrigan and Lohner 13 and Chandar et al 14 developed a semiautomatic technique and CU+ +, respectively.…”

Section: Introductionmentioning

confidence: 99%

Massively parallel least squares finite element method with graphic processing unit

Zhong

Dai

et al. 2017

Advances in Mechanical Engineering

View full text Add to dashboard Cite

For the reason of enormous computational expense, although the least squares finite element method has the advantages of high accuracy, robustness and strong versatility, the application of it is limited in computational fluid dynamics. The problems solved in this article include the rewriting of branching statements and the kernel function, variable distribution and data transfer between graphic processing units, and library functions rewriting. To the best knowledge of the authors, this article is the first time to develop the parallel computing codes for single and multiple graphic processing units based on the least squares finite element method. The computational results of single and multiple graphic processing units are verified by lid-driven cavity flow. Compared with a single central processing unit on the condition of 120 3 grids, the acceleration ratios of single and dual graphic processing units are up to 70.5 times and 95.2 times, respectively, which is much higher than the previous value of 7.7. With the increase in the grid number, the acceleration ratio of single and multiple graphic processing units is expected to increase, which can greatly enhance the computational efficiency of the least squares finite element method. Therefore, it is possible to solve the massive turbulence computing by the least squares finite element method with higher efficiency.

show abstract

Unsteady CFD computations using vertex‐centered finite volumes for unstructured grids on Graphics Processing Units

Cited by 75 publications

References 22 publications

Designing OP2 for GPU architectures

Designing OP2 for GPU architectures

Parallelization of an Unsteady ALE Solver with Deforming Mesh Using OpenACC

Massively parallel least squares finite element method with graphic processing unit

Contact Info

Product

Resources

About