The Earth Simulator: Designing the Future

Tetsuya Sato

Director of The Earth Simulator Center

Japan Marine Science Technology Center

(JAMSTEC)

The global environment that surrounds us, such as climate change and diastrophism, etc. is a pattern woven with cross-wires ranging from a scale of global to microscopic through generations of rain and ice. Even if, the fundamental laws of these elementary processes have been understood respectively, we do not have technique to accurately draw the global woven pattern. The computer simulation is the most promising means of understanding and clarification that we have. This simulation has been recognized as one of the means to reproduce a specific phenomenon or to clarify its mechanism. However, in real performance the Earth Simulator far exceeds, by 100 times, any existing simulator. Even if it is thought of as a simple extension of current simulation technique, it provides a very epoch-making, quantitative revolution. Moreover, the Earth Simulator is so powerful that it provides the possibility of not only a quantitative revolution but also qualitative conversion of the simulation.

Development of the Earth Simulator began in April, 1997 and was completed at the end of February in 2002. The activity of the Earth Simulator Center started in March, 2002, and the breathtaking performance in an atmospheric circulation simulation (AFES) of 26.58 Teraflops (65% of the peak performance) was shown in April, 2002. These results generated a great ripple of excitement within research communities throughout the world.

In this seminar, I will talk about some examples of the recent results and plans for the future.