Fast Simulation of Large-Scale Floods Based on GPU Parallel Computing

Abstract

Abstract:Computing speed is a significant issue of large-scale flood simulations for real-time response to disaster prevention and mitigation. Even today, most of the large-scale flood simulations are generally run on supercomputers due to the massive amounts of data and computations necessary. In this work, a two-dimensional shallow water model based on an unstructured Godunov-type finite volume scheme was proposed for flood simulation. To realize a fast simulation of large-scale floods on a personal computer, a Graphics Processing Unit (GPU)-based, high-performance computing method using the OpenACC application was adopted to parallelize the shallow water model. An unstructured data management method was presented to control the data transportation between the GPU and CPU (Central Processing Unit) with minimum overhead, and then both computation and data were offloaded from the CPU to the GPU, which exploited the computational capability of the GPU as much as possible. The parallel model was validated using various benchmarks and real-world case studies. The results demonstrate that speed-ups of up to one order of magnitude can be achieved in comparison with the serial model. The proposed parallel model provides a fast and reliable tool with which to quickly assess flood hazards in large-scale areas and, thus, has a bright application prospect for dynamic inundation risk identification and disaster assessment.

SPH Modeling of Water-Related Natural Hazards

“…Several examples of solvers based on shallow water equations on GPU are in [98,[122][123][124][125][126]. These allow for large-scale flood simulations to be performed using only one GPU.…”

“…That new version of IBER solved 24 h of an extreme flash flood in less than 10 min. The work in [126] presented a shallow water model based on the finite volume approach implemented for GPU with the OpenACC language [128]. OpenACC allows GPU parallelization to be implemented automatically using compiler directives like OpenMP for CPUs.…”

Parallelizing Multiple Flow Accumulation Algorithm using CUDA and OpenACC

“…The experiments show a significant improvement in computational efficiency that opens up the possibility of using their solution for real-time forecasting of flood events as well. Liu et al [13] proposed parallel OpenACC implementation of the two-dimensional shallow water model for flood simulation. The results of their experiments demonstrated achievements of up to one order of magnitude in speedup in comparison with the serial solution.…”

Effects of Bridge Piers on Flood Hazards: A Case Study on the Jialing River in China

“…In recent years, many one-and two-dimensional (1D, 2D) numerical models have been developed and applied to the investigation of river and urban floods [4][5][6][7][8][9][10][11][12][13][14]. For the 1D numerical model, Tomasz et al [12] carried out an assessment of the impact of bridges on flood flow of the Warta River near Wronki in Poland using the HEC-RAS software package.…”

“…In recent years, with the development of the LiDAR and Graphics processing units (GPUs) techniques, 2D numerical models have been extensively investigated and applied to flood inundation [20,21]. Very accurate, inexpensive DEM data has been obtained easily by LiDAR, and the GPU technique significantly improved the computational efficiency of the numerical model and maked available very accurate numerical methods [10,22]. Meesuk et al [23] demonstrated the accuracy of LiDAR data for a numerical flood model.…”