Computational performance evaluation of a limited area meteorological model by using the earth simulator

It is well known that meteorological models are computationally demanding, and they require both accurate and efficient numerical models on high-performance parallel computing. The regional version is nested in a global/larger regional model (coarse grid) by 1-way interactive nesting. The parallel computation of the nonhydrostatic atmospheric model, which has been developed at ESC, is implemented by a hybrid parallelization model: intra- and internode parallelisms are performed by microtasking and by MPI, respectively. To evaluate runtime performances, suitable compilation options and environment variables have been set. Because the program cannot be run sequentially, both speed-up and efficiency have been calculated using the elapsed time obtained run sequentially, both speed-up and efficiency have been calculated using the obtained values in all performed tests that show a speed-up varying nearby linearly and an efficiency staying always up to 70%. Varying the number of nodes on each domain such values of speed-up and efficiency get worse as increasing the number of processors, each sub-domain becomes too small. © 2007