Running large-scale air pollution models on message passing machines

The protection of our environment is one of the most important problems in the modern society. Reliable and robust control strategies for keeping the pollution caused by harmful chemical compounds under certain safe levels have to be developed and used in a routine way. Large mathematical models, in which all physical and chemical processes are adequately described, can successfully be used to solve this task. However, the use of large mathematical models in which all physical and chemical processes are adequately described leads, after the application of appropriate discretization and splitting procedures, to the treatment of huge computational tasks: in a typical simulation one has to perform several hundred runs, in each of these runs one has to carry out several thousand time-steps and at each time-step one has to solve numerically systems of ODE's containing up to O(10(6)) equations. Therefore, it is difficult to treat numerically such large mathematical models even when modern computers are available. Runs of an important module of a large-scale air pollution model on parallel message passing machines will be discussed in this paper. Numerical results will be presented. Concluding remarks explaining how the performance could be further improved will be given.