Free-surface flow simulations require high-resolution grids to capture phenomena at the interface and huge computational time. This research aims to develop a simulation method to realize large-scale free-surface flow simulations using the lattice Boltzmann method and multiple GPUs. By introducing the adaptive mesh refinement (AMR) method which adapts high-resolution grids to the free-surfaces to the lattice Boltzmann method, the number of lattice points can be greatly reduced. In the calculation of the AMR method, the spatial distribution of computational load changes with time, therefore the number of lattice points assigned to each GPU is kept equal by dynamic domain partitioning using a space-filling curve. We demonstrate large-scale free-surface slow simulations such as a dam breaking problem using multiple GPUs on the TSUBAME3.0 supercomputer at the Tokyo Institute of Technology.
各種検索
サポート
ヘルプ