Project/Area Number |
03044157
|
Research Category |
Grant-in-Aid for international Scientific Research
|
Allocation Type | Single-year Grants |
Section | Joint Research |
Research Institution | National Institute for Fusion Science |
Principal Investigator |
SATO Tetsuya National Institute for Fusion Science, 理論・シミュレーション研究センター, センター長教授 (80025395)
|
Co-Investigator(Kenkyū-buntansha) |
BRACKBILL J. ロスアラモス国立研究所, 主任研究員
DECYK V. カリフォルニア大学, ロサンジェルス校, 主任研究員
TROYON F. Plasma Physics Research Center, SFIT, Lausanne, 所長
DUM C. マックスプランク研究所, 主任研究員
ADAM J.C. Centre de Physique Theorique, Ecole Polytechnique, 計算機センター, センター長
CARRERAS B.A オークリッジ国立研究所, 主任研究員
OKAMOTO Masao National Institute for Fusion Science, 大型ヘリカル研究部, 教授 (70115541)
UCHIDA Yutaka The University of Tokyo, 理学部, 教授 (90012814)
HAYASHI Takaya National Institute for Fusion Science, 計算機センター, 助教授 (60156445)
TAKAMARU Hisanori National Institute for Fusion Science, 理論・シミュレーション研究センター, 助手 (20241234)
TODO Yasushi National Institute for Fusion Science, 理論・シミュレーション研究センター, 助手 (00249971)
HORIUCHI Ritoku National Institute for Fusion Science, 理論・シミュレーション研究センター, 助教授 (00229220)
WATANABE Kunihiko National Institute for Fusion Science, 理論・シミュレーション研究センター, 助教授 (40220876)
TANAKA Motohiko National Institute for Fusion Science, 理論・シミュレーション研究センター, 助教授 (80167501)
CARRERAS Benjamin Oak Ridge National Laboratory
DUM Christian Max Planck Institute for Physics and Astrophysics
BRACKBILL Jeremiash Los Alamos National Laboratory
DECYK Viktor University of Calfornia, Los Angeles
J U Brackbil ロスアラモス国立研究所, 主任研究員
V Decyk カリフォルニア大学ロサンジェルス校, 主任研究員
F Troyon ローザンヌプラズマ物理研究所, 所長
C Dum マックスプランク研究所, 主任研究員
J C Adam パリ総合技術大学, 計算機センター長
B A Carreras オークリッジ国立研究所, 主任研究員
上村 鉄雄 核融合科学研究所, 理論・シミュレーション・研究センター, 教授 (20023731)
MATSUDA Y. リバモア国立研究所, 主任研究員
西原 功修 大阪大学, レーザー核融合研究センター, 教授 (40107131)
|
Project Period (FY) |
1991 – 1993
|
Project Status |
Completed (Fiscal Year 1993)
|
Budget Amount *help |
¥9,500,000 (Direct Cost: ¥9,500,000)
Fiscal Year 1993: ¥3,500,000 (Direct Cost: ¥3,500,000)
Fiscal Year 1992: ¥3,000,000 (Direct Cost: ¥3,000,000)
Fiscal Year 1991: ¥3,000,000 (Direct Cost: ¥3,000,000)
|
Keywords | parallel processor / fusion plasma / simulation / processor element / network / program language / algorism / data access / 超多重並列計算機 / シミュレ-ション / 記憶方式 / 演算命令方式 / 記憶素子 |
Research Abstract |
For the last three years we have investigated the appropriateness of existing massively parallel processors in Japan, U.S.A. and Europe as a vital tool for studying the fusion plasma dynamics and conducted an international collaboration on fusion plasma simulations with making use of some of the massively parallel processors. The summary of our collaboration research is described in the following. (1) The parallel processors are classified by three architectures, i.e., a) the memory architecture, b) the network architecture of the processor element, and c) the instruction method. (2) For the magnetohydrodynamic simulation which handles near-term interactions it is found that it is not so difficult to parallelize the calculation procedure and the data access by using multiple-instruction distributed-memory machine. (3) Recently the data access speed among processor elements has been improved largely, say, by the factor of 10-40, thus this issue is becoming less important for parallel processing. Accordingly, we have reached to a conclusion that the grid-type connection system is the most appropriate network system for the magnetohydrodynamic simulation because of its simple geometry. (4) For the particle simulation all particle informations are gathered into the fluid quantities and then must be distributed to all processor elements at each time step. Thus, it is quite difficult to obtain high parallel performance under any existing machine architectures. (5) It is concluded that FORTRAN90, a program language for parallel processors, may not be appropriate for the fusion plasma simulation, because it includes many exceptional expressions and machine dependent languages.
|