Atsuo Ozaki

Design and implementation of parallel and distributed wargame simulation system

Simulation based education and training, especially wargame simulations are being widely used in the field of defence modeling and simulation communities. In order to efficiently train students and trainees, the wargame simulations must have both high performance and high fidelity. In this paper, we discuss design and implementation issues for a tentative wargame simulation on a distribution system. This wargame simulation system is based on high-level architecture and employs optimization to achieve a high performance and high fidelity simulation system.

High Performance Road Traffic Simulation based on Dynamic Time Step Synchronization Method

We have already proposed DTSS (event aware Dynamic Time Step Synchronization method) for speeding up moving objects (MOs) simulation, e.g. aircraft, ships, vehicles and so on, and evaluated it on the HLA (High Level Architecture, IEEE1516) computer cluster environment. However, speeding up the simulation for a number of MOs on a single processor is required as well, when the size of the problem is increased. In this paper, we propose DTSS- RT (DTSS for Road Traffic simulation) that can speed up the simulation by reducing the simulation cost of each MO and the synchronization cost between MOs without degrading simulation accuracy. For evaluating DTSS-RT, we employed a road traffic sub-simulator of RoboCup rescue simulation. The results show that the performance of the simulation based on DTSS-RT is improved approximately 2 to 3 times over that of the conventional method except when the situation includes a heavy traffic jam.

Distributed Parallel Backprojection for Real-Time Stripmap SAR Imaging on GPU Clusters

Parallelization on a GPU (graphics processing unit) cluster is an effective approach to reducing the huge computation time of backprojection, which is the most accurate SAR (synthetic aperture radar) imaging algorithm for reconstructing images with no errors caused by the platform motion. To obtain accurate imagery in real-time, we developed a distributed parallel backprojection algorithm for stripmap SAR on GPU clusters, which reconstruct the image while receiving signals from the remote platform. In the case of receiving the 1.9 GiB signals from the remote storage through 1GbE, we found that 16 GPUs on the 4 nodes are 11.5 times faster than 1 GPU, and they finished the imaging 1.0s after receiving all signals.

Performance Evaluation of Adapting Many-Core Processors to the Optimization of a Shaped Reflector Antenna

In the design of reflector antennas, optimization of the reflector surface requires a long computation time. To overcome this problem, we use parallel processing based on GPUs. In this paper, we present the performance of a GPU-based implementation for optimizing a shaped reflector antenna. The experimental results show that our implementation with 1-4 GPUs is about 8-20 times faster than that with a 6-threaded CPU.

Dynamic time step synchronization method for road traffic simulation - application to RoboCup rescue simulation -

We have already proposed DTSS (event aware dynamic time step synchronization method) for speeding up moving objects (MOs) simulation, e.g. aircraft, ships, vehicles and so on, in a computer cluster environment. However, speeding up the simulation for a number of MOs on a single processor is required as well, when the problem size is enlarged. In this paper, we propose DTSS-RT (DTSS for road traffic simulation), which can speed up the simulation by reducing the simulation cost of each MO and synchronization cost between MOs without degrading the simulation accuracy. For evaluating DTSS-RT, we employed a road traffic sub-simulator of RoboCup rescue simulation. The results show that the performance of the simulation based on DTSS-RT is improved approximately 2 - 3 times over that of the conventional method except when the situation includes a heavy traffic jam.