Climate Modeling and Global Change

Partnership

Web Site: http://www.cerfacs.fr

Context

The aim is the understanding and forecasting of the world's environment and climate on regional to global scales. The strategy used by the CERAFCS strongly relies upon a dual approach based on theoretical and modelling studies as well as on the development of high-level software needed to address the various coupling issues arising in climate science. This is done through coordination and participation in national and international research programs

Use case

The proposed climate modeling application consists of simulations of present climate followed by the 21st century, for a total of 150 years (one scenario). A scenario combines 1800 simulations of one month each (150×12), launched one after the other. The results from the nth monthly simu- lation are the starting point of the (n+1)th. For the whole experiment, several scenarios are performed simultaneously. A monthly simulation can be divided into a preprocessing phase, a main-processing parallel task, and a post-processing phase of analysis.

Contribution

• Analyze and Modeling the climatology application (Ocean-Atmosphere).
• Three scheduling heuristics has been designed to improve the application.
• Yield to gains of up to 12%
• Application used in heterogeneous grid composed of homogeneous clusters.

Technical Form

References

• Y. Caniou, E. Caron, G. Charrier, A. Chis, F. Desprez, and E. Maisonnave. Ocean-atmosphere modelization over the grid. In W.-C. Feng and Y. Yang, editors, The 37th International Conference on Parallel Processing (ICPP 2008), pages 206–213, Portland, Oregon. USA, Sept. 2008. IEEE.
  
• Y. Caniou, E. Caron, G. Charrier, F. Desprez, E. Maisonnave, and V. Pichon. Ocean-atmosphere application scheduling within DIET. In APDCT-08 Symposium. International Symposium on Advanced in Parallel and Distributed Computing Techniques, pages 675–680, Sydney. Australia., Dec. 2008. In conjunction with ISPA’2008, IEEE Computer Society.