A GPU-efficient local ADER finite-difference method for compressible-flow DNS

Abstract

Algorithm design tailored to graphics processing units (GPUs) can substantially improve the efficiency of computational fluid dynamics (CFD) simulations. Efforts to adapt CFD solvers to GPUs often focus on spatial discretization, while time advancement commonly relies on multistage Runge--Kutta (RK) methods. However, these RK methods require repeated residual evaluations and full-domain operations, increasing the cost of compressible-flow direct numerical simulation (DNS). To reduce this overhead, we develop a local arbitrary high-order derivatives (ADER) finite-difference method that couples a compact ten-point quadratic predictor to a conservative central corrector through operator matching. With two quadrature nodes, one Picard correction, and selective dissipation, smooth convective tests recover third-order convergence under fixed-CFL refinement and near-seventh-order spatial behavior when background spatial error dominates. Temporal states remain kernel-local, and halo communication overlaps interior computation. Tests of smooth advection, periodic turbulence, and wall-bounded flows assess accuracy and flow statistics. On a 302³-node viscous Taylor--Green benchmark without heat conduction, throughput reaches 230.87 million node updates per second on one V100 GPU in double precision, 1.944 times that of spatially matched three-stage Runge--Kutta integration. Weak-scaling efficiency reaches 99.44\% from 4 to 32 V100 GPUs at 1.812 billion nodes.

Ternary-Valued Finite-Difference Time-Domain Method: Asymptotic Agreement with the Yee Scheme Through Noise-Shaped Quantisation
Ivan S. Maksymov
arXiv30 Sep 2026PreprintCC BY 4.0Full textApplied physics & instrumentation Computational physicsarXiv
Fourier Finite-Difference method and Proper Orthogonal Decomposition for quick spectral study of 2D periodic photonic structures
Mondher Besbes, Yassine Ferhi, Henri Benisty +1
Journal of the Optical Society of America. A Optics, Image Science, and Vision11 Aug 2026Peer-reviewedCC BY 4.0Full textOptics & photonicsHAL
Light pulse propagation through a dielectric slab. I. Retarded integral equations and their local finite-difference solution
Giancarlo Reali
arXiv23 Oct 2026PreprintCC BY-NC-ND 4.0Full textOptics & photonicsarXiv
High-order immersed boundary method for subsonic aeroacoustics and compressible Navier-Stokes equationsHigh-order immersed boundary method for subsonic aeroacoustics and compressible Navier-Stokes equations
Abdoulaye Ouattara, David Marx, Christian Prax +3
arXiv21 Sep 2026PreprintCC BY-NC-ND 4.0Full textFluid dynamicsarXiv
DeHNSSo-v2: A compressible, curvilinear, and efficient extension of the Delft Harmonic Navier-Stokes Solver
Pietro Carlo Boldini, Yifu Chen, Marios Kotsonis
arXiv20 Oct 2026PreprintCC BY 4.0Full textMathematical physics Computational physicsarXiv

Cite

BibTeX RIS

BibTeX

@misc{zhang2026gpuefficient,
  title = {{A GPU-efficient local ADER finite-difference method for compressible-flow DNS}},
  author = {Hong Zhang and Xian Wang and Xinliang Li and Yibing Chen and Xing Ji},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.34252},
  eprint = {2610.34252},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.34252},
}

RIS

TY  - UNPB
TI  - A GPU-efficient local ADER finite-difference method for compressible-flow DNS
AU  - Hong Zhang
AU  - Xian Wang
AU  - Xinliang Li
AU  - Yibing Chen
AU  - Xing Ji
PY  - 2026
DA  - 2026/10/25
PB  - arXiv
DO  - 10.48550/arxiv.2610.34252
UR  - https://arxiv.org/abs/2610.34252
AB  - Algorithm design tailored to graphics processing units (GPUs) can substantially improve the efficiency of computational fluid dynamics (CFD) simulations. Efforts to adapt CFD solvers to GPUs often focus on spatial discretization, while time advancement commonly relies on multistage Runge--Kutta (RK) methods. However, these RK methods require repeated residual evaluations and full-domain operations, increasing the cost of compressible-flow direct numerical simulation (DNS). To reduce this overhead, we develop a local arbitrary high-order derivatives (ADER) finite-difference method that couples a compact ten-point quadratic predictor to a conservative central corrector through operator matching. With two quadrature nodes, one Picard correction, and selective dissipation, smooth convective tests recover third-order convergence under fixed-CFL refinement and near-seventh-order spatial behavior when background spatial error dominates. Temporal states remain kernel-local, and halo communication overlaps interior computation. Tests of smooth advection, periodic turbulence, and wall-bounded flows assess accuracy and flow statistics. On a 302^3-node viscous Taylor--Green benchmark without heat conduction, throughput reaches 230.87 million node updates per second on one V100 GPU in double precision, 1.944 times that of spatially matched three-stage Runge--Kutta integration. Weak-scaling efficiency reaches 99.44\% from 4 to 32 V100 GPUs at 1.812 billion nodes.
ER  -

Abstract

Related

Cite