Summative report of the public competition research and development on software for computational science and engineering in the fiscal year 1997 through 2002

Office of Planning; Center for Promotion of Computational Science and Engineering

JAERI-Review 2005-038, 68 Pages, 2005/09

JAERI-Review-2005-038.pdf:3.85MB

Japan Atomic Energy Research Institute started the public competition research and development on software for computational science and engineering in 1997, and closed it in 2002. This report describes the system of the competition research and development, application situations, R & D subjects adopted, evaluation findings, outputs produced, achievements and problems, as a summative report of practice of the system for six years.

Summaries of research and development activities by using JAERI computer system in FY2003 (April 1, 2003 - March 31, 2004)

Center for Promotion of Computational Science and Engineering

JAERI-Review 2005-008, 199 Pages, 2005/03

JAERI-Review-2005-008.pdf:30.15MB

The Center for Promotion of Computational Science and Engineering (CCSE) in Japan Atomic Energy Research Institute (JAERI) installed large computer system including super-computers in order to support research and development activities in JAERI. CCSE is also operate anad manage the computer system and network system. This report presents the survey results of research and development activities by using JAERI computer system in FY2003.

Report of the evaluation by the Ad Hoc Review Committee on Computational Science and Engineering; Result evaluation in fiscal year 2000

Research Evaluation Committee

JAERI-Review 2001-025, 17 Pages, 2001/06

JAERI-Review-2001-025.pdf:1.62MB

no abstracts in English

High performance implementation of nuclear fusion simulation code on GPU cluster

Matsumoto, Kazuya; Asahi, Yuichi*; Ina, Takuya; Idomura, Yasuhiro

no journal, , 

We present the implementation and performance evaluation results of the plasma physics simulation code called GT5D on a GPU cluster. In this study, an iterative matrix solver, which is identified as a performance bottleneck in the code, is tuned on the GPU. The measured performance is compared with attainable performance calculated by the roofline model. Additionally, we show the implementation with direction communications between GPUs for utilizing many GPUs.