Abstract:This paper presents a parallel implementation of the Discrete Unified Gas Kinetic Scheme (DUGKS) on the GPU system using the CUDA Fortran and CUDA C++ programming languages. Firstly, we conducted an extensive revision of our original CPU-based code, resulting in a threefold decrease in memory usage. This new implementation is also paired with a novel approach to compute cell face flux using trilinear interpolation. It is shown analytically that the interpolation-based approach to flux calculation is more accurate compared to the one used in the original DUGKS. The initial simulation results using this new approach suggest that trilinear interpolation can reduce numerical errors on a coarse mesh. For example, in the case of the decaying Taylor-Green vortex flow at a 128 3 mesh resolution, the relative numerical error in the energy dissipation rate at t⁎=2 , using the spectral simulation result as the benchmark, is approximately 30% lower than that of the original implementation. The improved GPU DUGKS method is applied to laminar and turbulent flows in periodic and wall-bounded boundary configurations. A performance comparison of the GPU implementation is also presented and compared to the previous CPU implementation. A maximum speedup of 7.64x was achieved on a desktop-level GPU compared to a 32-core CPU. The strong scaling test, conducted on an eight-GPU node, demonstrated the efficient utilization of available multiple GPU resources by the code. Program summary Program Title: DUGKS-GPU CPC Library link to program files: https://doi.org/10.17632/yykv5s9g2n.1 Developer's repository link: https://github.com/kairzhan/DUGKS-GPU Code Ocean capsule: https://codeocean.com/capsule/3b1e4f74-cdd3-4781-8923-8514d5923dfb/ Licensing provisions: GNU General Public License 3 Programming language: CUDA C++, CUDA Fortran Supplementary material: Nature of problem: DUGKS-GPU is an advanced numerical code designed to accelerate the simulation of turbulent fluid flows through the application of the kinetic method known as the Discrete Unified Gas Kinetic Scheme (DUGKS). This computational tool comprises a CUDA Fortran version, tailored for a single GPU, as well as a CUDA C++ version developed for multi-GPU platforms. Solution method: The code employs a second-order finite-volume discretization of the discrete velocity Boltzmann equation with the BGK collision model. It can be used to simulate small to medium-scale problems on modern multi-GPU platforms with CUDA architecture.

Fast GPU-Powered and Auto-Differentiable Forward Modeling of IFU Data Cubes

cuFFS: A GPU-accelerated code for Fast Faraday Rotation Measure Synthesis

Collision-streams: fast GPU-based collision detection for deformable models.

Capivara: A Spectral-based Segmentation Method for IFU Data Cubes

rupMC: a ray-unit parallel marching cubes algorithm on CPU/GPU heterogeneous architectures

IDEFIX: a versatile performance-portable Godunov code for astrophysical flows

XCube: Large-Scale 3D Generative Modeling using Sparse Voxel Hierarchies

Fast forward modelling of galaxy spatial and statistical distributions

Multi-GPU RI-HF Energies and Analytic Gradients $-$ Towards High Throughput Ab Initio Molecular Dynamics

Fast interactive web-based data visualizer of panoramic spectroscopic surveys

DUGKS-GPU: An efficient parallel GPU code for 3D turbulent flow simulations using Discrete Unified Gas Kinetic Scheme

Multi-GPU RI-HF Energies and Analytic Gradients─Toward High-Throughput Ab Initio Molecular Dynamics

GaussianCube: A Structured and Explicit Radiance Representation for 3D Generative Modeling

More Bang For Your Buck(et): Fast and Space-efficient Hardware-accelerated Coarse-granular Indexing on GPUs

GPU-RRTMG_SW: Accelerating a Shortwave Radiative Transfer Scheme on GPU

J-UBIK: The JAX-accelerated Universal Bayesian Imaging Kit

Fitting Galaxies on GPUs

nvblox: GPU-Accelerated Incremental Signed Distance Field Mapping

Graphics-processing-unit-accelerated ice flow solver for unstructured meshes using the Shallow-Shelf Approximation (FastIceFlo v1.0.1)

CuQ-RTM: A CUDA-based Code Package for Stable and Efficient Q-compensated Reverse Time Migration