L. A. Drummond, J. D. Farrara*, C. R. Mechoso, J. A. Spahr
Department of Atmospheric Sciences
University of California, Los Angeles
Los Angles, California, 90095-1565
Y. Chao, D. S. Katz, J. Z. Lou, P. Wang
Jet Propulsion Laboratory
California Institute of Technology
Pasadena, California, 91109
We are developing an Earth System Model (ESM) to be used in research aimed to better understand the interactions between the components of the Earth System and to eventually predict their variations. Currently, our ESM includes models of the atmosphere, oceans and the important chemical tracers therein.
This paper reports our efforts to optimize the parallel implementations of two of the four major ESM components: the atmospheric and oceanic models. So far, the work has focused on improving the load balancing and single node performance of the code in the CRAY T3D. As a result, the atmospheric and oceanic model components running side-by-side, can achieve a performance level of slightly more than 10 GFLOPS on 512 processors of that machine. We also briefly outline our plans for realizing performance levels of 100 Gflops by the end of the century.