e-Science in high energy density science research

• Published in: Fusion engineering and design Vol. 83; no. 2; pp. 525 - 529
• Main Authors: Nishihara, K., Fukuda, Y., Shimada, K., Taniguchi, M., Zhakhovskii, V., Fujioka, S., Shigemori, K., Sakane, E., Shimojo, S., Ueshima, Y., Okamoto, T., Sasaki, A., Sunahara, A., Nakajima, T.
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Amsterdam Elsevier B.V 01.04.2008; New York, NY Elsevier Science
• Subjects: Applied sciences; Controled nuclear fusion plants; e-Science; Energy; Energy. Thermal use of fuels; EUV; Exact sciences and technology; Grid computing; High energy density science; Installations for energy generation and conversion: thermal and electrical energy; Laser fusion; Laser produced plasma; Lithography; XML; High energy density science; Lithography; XML; EUV; Database; Grid computing; Laser produced plasma; e-Science; Laser fusion; Computer simulation; Control system; Molecular dynamics; Algorithm; Research and development; Nuclear fusion reactor; High energy; Laser; Computer system
• ISSN: 0920-3796; 1873-7196
• DOI: 10.1016/j.fusengdes.2007.12.010

We present three achievements related to e-Science in high energy density science research. We have developed a test module of new data base system for nation-wide users of the high power lasers at Osaka University with web technology, flexible control system and XML-database, which provide efficient, accurate and flexible R&D processing. We also constructed a EUV-GRID portal system for integrated computer simulations of laser produced plasma–extreme ultraviolet (LPP–EUV) light source development, which provides flexible framework for automatic sequencing of workflow for the integrated simulations. A dynamical domain decomposition method in molecular dynamic (MD) simulations is required to obtain a good adaptive load balancing for heterogeneous computing environments such as grid. We have developed a new algorithm of the dynamical domain decomposition. We successfully performed large scale MD-message passing interface (MPI) simulations on cluster computers connected through Super-SINET using the National Research Grid Initiative (NAREGI GRID) grid middleware and the new algorithm.