A5020592501- Gamma-ray bursts and supernovae
- Solidification and crystal growth phenomena
- SAS software applications and methods
- Aluminum Alloy Microstructure Properties
- Astro and Planetary Science
- Astronomical Observations and Instrumentation
- Parallel Computing and Optimization Techniques
- Distributed and Parallel Computing Systems
- Meteorological Phenomena and Simulations
- Metallurgy and Material Forming
- Advanced Data Storage Technologies
- Air Quality Monitoring and Forecasting
- Astronomy and Astrophysical Research
- Metallurgical Processes and Thermodynamics
- Astrophysics and Cosmic Phenomena
- Nuclear Physics and Applications
- Lattice Boltzmann Simulation Studies
- Particle Detector Development and Performance
- Pulsars and Gravitational Waves Research
- nanoparticles nucleation surface interactions
- Computer Graphics and Visualization Techniques
- Precipitation Measurement and Analysis
- Wind and Air Flow Studies
- Urban Heat Island Mitigation
- Fluid Dynamics and Thin Films
The University of Tokyo
2017-2024
Ion Technology Center (Japan)
2017-2024
Tokyo Institute of Technology
2007-2016
Computing Center
2016
Institute of Experimental Physics of the Slovak Academy of Sciences
2014
Hokkaido University
2014
KTH Royal Institute of Technology
2009
Stockholm University
2009
Kanazawa University
2008
The mechanical properties of metal materials largely depend on their intrinsic internal microstructures. To develop engineering with the expected properties, predicting patterns in solidified metals would be indispensable. phase-field simulation is most powerful method known to simulate micro-scale dendritic growth during solidification a binary alloy. evaluate realistic description solidification, however, requires computing large number complex nonlinear terms over fine-grained grid. Due...
Regional weather forecasting demands fast simulation over fine-grained grids, resulting in extremely memory- bottlenecked computation, a difficult problem on conventional supercomputers. Early work accelerating mainstream code WRF using GPUs with their high memory performance, however, resulted only minor speedup due to partial GPU porting of the huge code. Our full CUDA high- resolution prediction model ASUCA is first such one we know date; next-generation, production developed by Japan...
Abstract Grain growth, a competitive growth of crystal grains accompanied by curvature-driven boundary migration, is one the most fundamental phenomena in context metallurgy and other scientific disciplines. However, true picture grain still controversial, even for simplest (or ‘ideal’) case. This problem can be addressed only large-scale numerical simulation. Here, we analyze ideal via ultra-large-scale phase-field simulations on supercomputer elucidating corresponding authentic statistical...
Primary arm arrays formed during the directional solidification of a single-crystal binary alloy were investigated by performing large-scale phase-field simulations using GPU supercomputer TSUBAME2.5 at Tokyo Institute Technology. The primary array and spacing Voronoi decomposition Delaunay triangulation, respectively. It was concluded that hexagonal dominant for both dendrite cell structures penta–hepta defects, which are typical defects in patterns, formed. arms continuously moved such...
Recent developments in machine learning have enabled accurate predictions of the dynamics slow structural relaxation glass-forming systems. However, existing models for these tasks are mostly designed such that they learn a single dynamic quantity and relate it to features glassy liquids. In this study, we propose graph neural network model, “BOnd TArgeting Network,” learns relative motion between neighboring pairs particles, addition self-motion particles. By relating two different...
A GPU code has been developed for a phase-field lattice Boltzmann (PFLB) method, which can simulate the dendritic growth with motion of solids in dilute binary alloy melt. The accelerated PFLB method implemented using CUDA C. equiaxed shear flow and settling condition have simulated by code. It confirmed that simulations were efficiently introducing computation. characteristic dendrite morphologies depend on melt could also be simulations.
Large-scale phase-field studies of three-dimensional (3D) dendrite competitive growth at the converging grain boundary (GB) a bicrystal binary alloy were carried out using GPU-rich supercomputer TSUBAME 2.5 Tokyo Institute Technology. First, series thin-sample simulations performed to investigate effects thickness, unfavorably oriented (UO) inclination angle, and arrangement on an unusual overgrowth phenomenon whereby favorably (FO) is overgrown by UO grain. It was concluded that easily...
The Large Area Telescope (LAT) instrument on the Fermi mission will reveal rich spectral and temporal gamma-ray burst (GRB) phenomena in >100 MeV band. synergy with Fermi's Gamma-ray Burst Monitor detectors link these observations to those well explored 10–1000 keV range; addition of band resolve theoretical uncertainties about emission both prompt afterglow phases. Trigger algorithms be applied LAT data onboard spacecraft ground. sensitivity triggers differ because available computing...
Abstract We present ensemble-based large-eddy simulations based on a lattice Boltzmann method for realistic urban area. A plume-dispersion model enables real-time simulation over several kilometres by applying local mesh-refinement method. assess problems in the complex environment of Oklahoma City 16 July using mesoscale velocity boundary conditions produced Weather Research and Forecasting model, as well building structures plant-canopy introduced into model. Ensemble calculations are...
Numerical weather prediction is one of the major applications in high performance computing and demands fast high-precision simulation over fine-grained grids. While utilizing hundreds CPUs certainly most common way to get for large scale simulations, we have another solution use GPUs as massively parallel platform. In order drastically shorten runtime a code, rewrite its huge entire code GPU from scratch CUDA. The ASUCA resolution meso-scale atmosphere model that being developed by Japan...
Phase-field method has emerged as the most powerful numerical scheme to simulate dendrite growth. However, phase-field simulations of growth performed so far are limited two-dimension or single in three-dimension because large computational cost involved. To express actual solidification microstructures, multiple dendrites with different preferred directions should be computed at same time. In this study, order enable large-scale simulations, we developed a code using graphics processing...
The weather prediction code demands large computational performance to achieve fast and high-resolution simulations. Skillful programming techniques are required for obtaining good parallel efficiency on GPU supercomputers. Our framework-based ASUCA has achieved scalability with hiding complicated implementation optimizations distributed GPUs, contributing increasing the maintainability, is a next-generation high resolution meso-scale atmospheric model being developed by Japan Meteorological...
Stencil-based applications such as CFD have succeeded in obtaining high performance on GPU supercomputers. The problem sizes of these are limited by the device memory capacity, which is typically smaller than host memory. On supercomputers, a locality improvement technique using temporal blocking method with swapping between and enables large computation beyond capacity. However, because loop management data movement across memories increase programming difficulty, applying this methodology...
Development of MITSuME is reported. Two 50-cm optical telescopes have been built at Akeno in Yamanashi prefecture and Okayama Astrophysical Observatory (OAO) prefecture. Three CCD cameras for simultaneous g'RcIc photometry are to be mounted on each focal plane, covering a wide FOV about 30" x 30". The limiting magnitude V fainter than 18. In addition these two telescopes, 91-cm IR telescope with 1 deg field view being OAO, which performs YJHK bands. These robotic can start the observation...
Abstract Using a pulse-fit method, we investigated the spectral lags between traditional gamma-ray band (50–400 keV) and X-ray (6–25 for 8 GRBs with known redshifts (GRB 010921, GRB 020124, 020127, 021211, 030528, 040924, 041006, 050408), detected WXM FREGATE instruments aboard HETE-2 satellite. We found several relations individual pulses lag other observables, such as luminosity, pulse duration, peak energy, $E_{\rm peak}$. The obtained results are consistent those BATSE, indicating that...
The multiphase field (MPF) method is recognized as a powerful numerical to simulate microstructural evolutions, such solidification, grain growth, recrystallization, and phase transformation, in various materials. However, because we need solve the time evolution equations for multiple-order parameters derived from total Gibbs free energy, MPF simulations are very computationally expensive. In this paper, use graphics processing unit (GPU) accelerate two-dimensional simulation of...
A high-productivity framework for multi-GPU and multi-CPU computation of stencil applications is proposed. Our implemented in C++ CUDA languages. It automatically translates user-written functions that update a grid point generates both GPU CPU codes. The programmers write user code just the language, can execute translated on either multiple multicore CPUs or GPUs with optimization. be executed auto-tuning mechanism overlapping method to hide communication cost by computation. also OpenMP....