Ryo Onishi

Taketo Tominaga, Ryo Onishi

Multiphase turbulent flow phenomena are observed not only in industrial devices but also in environmental flows, and direct numerical simulation (DNS) plays a key role in their investigation. Many numerical models have been developed; nevertheless, few models are highly optimized for GPU platforms, which represent the current mainstream in high-performance computing (HPC). In this study, we developed LCS.jl (Lagrangian Cloud Simulator in Julia), a single-source and multi-platform multiphase turbulence simulation model implemented in Julia language and KernelAbstractions.jl. Validation results confirmed that the present fluid and particle statistics agree well with those obtained in prior studies. A GPU-native particle communication algorithm based on prefix-scan reduced the particle communication cost from approximately 78% (CPU-delegated) to 10% of total execution time. LCS.jl achieved computational performance equivalent to the Fortran implementation in many-processes computations. For GPUs, strong scaling efficiency was maintained above 85% (up to 256 GPUs) and weak scaling efficiency above 90% (up to 216 GPUs) on TSUBAME4.0 (a GPU supercomputer at the Institute of Science Tokyo). LCS.jl achieved a maximum speedup of 18.0x on GPUs over CPUs. A trial heterogeneous execution achieved a 72% reduction in execution time compared to the CPU-only configuration even in configurations where the GPU was not the primary compute device. These results demonstrate that LCS.jl is a multiphase turbulence simulation platform that achieves both portability and scalability across a variety of computational resource configurations.

Yuki Yasuda, Ryo Onishi

This study employs a neural network that represents the solution to a Schrödinger bridge problem to perform super-resolution of 2-m temperature in an urban area. Schrödinger bridges generally describe transformations between two data distributions based on diffusion processes. We use a specific Schrödinger-bridge model (SM) that directly transforms low-resolution data into high-resolution data, unlike denoising diffusion probabilistic models (simply, diffusion models; DMs) that generate high-resolution data from Gaussian noise. Low-resolution and high-resolution data were obtained from separate numerical simulations with a physics-based model under common initial and boundary conditions. Compared with a DM, the SM attains comparable accuracy at one-fifth the computational cost, requiring 50 neural-network evaluations per datum for the DM and only 10 for the SM. Furthermore, high-resolution samples generated by the SM exhibit larger variance, implying superior uncertainty quantification relative to the DM. Owing to the reduced computational cost of the SM, our results suggest the feasibility of real-time ensemble micrometeorological prediction using SM-based super-resolution.

Yuki Yasuda, Ryo Onishi

This paper investigates the super-resolution (SR) of velocity fields in two-dimensional fluids from the viewpoint of rotational equivariance. SR refers to techniques that estimate high-resolution images from those in low resolution and has lately been applied in fluid mechanics. The rotational equivariance of SR models is defined as the property in which the super-resolved velocity field is rotated according to a rotation of the input, which leads to the inference covariant to the orientation of fluid systems. Generally, the covariance in physics is related to symmetries. To clarify a relationship to symmetries, the rotational consistency of datasets for SR is newly introduced as the invariance of pairs of low- and high-resolution velocity fields with respect to rotation. This consistency is sufficient and necessary for SR models to acquire rotational equivariance from large datasets with supervised learning. Such a large dataset is not required when rotational equivariance is imposed on SR models through weight sharing of convolution kernels as prior knowledge. Even if a fluid system has rotational symmetry, this symmetry may not carry over to a velocity dataset, which is not rotationally consistent. This inconsistency can occur when the rotation does not commute with the generation of low-resolution velocity fields. These theoretical suggestions are supported by the results from numerical experiments, where two existing convolutional neural networks (CNNs) are converted into rotationally equivariant CNNs and the inferences of the four CNNs are compared after the supervised training.