5:15 PM - 6:30 PM
[SVC27-P04] Accelerating Two-Dimensional Lava Flow Simulations with OpenACC
Keywords:Lava Flow, Numerical Simulation, GPGPU, OpenACC
The performance of Graphics Processing Unit, or GPU, has been much grown in these few decades. While GPU has been originally developed for drawing 3D graphics for video games on computers, for example, at a high speed, GPU can be used for calculating double-precision floating-point values under certain conditions. That enhances General-Purpose Computing on GPUs, or GPGPU, in the field of high-performance computing. In particular, GPGPU research activities are rapidly expanding since the standardization of OpenACC has been initiated in 2011. Here, OpenACC is a directive-based programming model designed to provide a simple and effective approach.
Mount Fuji Research Institute has launched a research project entitled “Improvement of Numerical Simulation Techniques for Enhancing the Validity of Hazard Maps for Volcanic Eruptions” in 2020 and installed a computer with Titan V, which is a GPU developed by nVidia with a theoretical peak performance of 13.8 TFlops for single-precision floating-point calculations. We will present preliminary results of numerical simulations of lava flows that are performed with a numerical model developed by Ishihara (1990) and accelerated with OpenACC.
Mount Fuji Research Institute has launched a research project entitled “Improvement of Numerical Simulation Techniques for Enhancing the Validity of Hazard Maps for Volcanic Eruptions” in 2020 and installed a computer with Titan V, which is a GPU developed by nVidia with a theoretical peak performance of 13.8 TFlops for single-precision floating-point calculations. We will present preliminary results of numerical simulations of lava flows that are performed with a numerical model developed by Ishihara (1990) and accelerated with OpenACC.