PARALLELIZATION AND DISTRIBUTION OF A COUPLED ATMOSPHERE OCEAN GENERAL-CIRCULATION MODEL

The distribution of a climate model across homogeneous and heterogeneo us computer environments with nodes that can reside at geographically different locations is investigated. This scientific application consi sts of an atmospheric general circulation model (AGCM) coupled to an o ceanic general circulation model (OGCM). Three levels of code decompos ition are considered to achieve a high degree of parallelism and to ma sk communication with computation. First, the domains of both the grid point AGCM and OGCM are divided into subdomains for which calculations are carried out concurrently (domain decomposition). Second, the mode l is decomposed based on the diversity of tasks performed by its major components (task decomposition). Three such components are identified : (a) AGCM/physics, which computes the effects on the grid-scale flow of subgrid-scale processes such as convection and turbulent mixing; (b ) AGCM/dynamics, which computes the evolution of the flow governed by the primitive equations; and (c) the OGCM. Task decomposition allows t he AGCM/dynamics and OGCM calculations to be carried out concurrently. Last, computation and communication are organized in such a way that the exchange of data between different tasks is carried out in subdoma ins of the model domain (I/O decomposition). In a dedicated computer n etwork environment, the wall-clock time required by the resulting dist ributed application is reduced to that for the AGCM /physics, with the other two components and interprocess communications running in paral lel. The network bandwidth requirements for the distributed applicatio n are analyzed. It is assumed that the wall-clock time required to run the AGCM/physics for the model atmosphere in a dedicated computer env ironment is fixed at a value corresponding to high network efficiency. The analysis shows that, for computer environments based on nodes equ ivalent to the Intel Touchstone Delta, a bandwidth approaching that of the Gigabit Network is required for an efficient operation of the dis tributed application with model resolution double that used in current studies of the climate system if output is visualized in real time. I t is argued that distribution of a climate model based on domain, task , and I/O decomposition has the potential for significant and eventual ly superlinear speedup in model execution, which will facilitate perfo rmance of the long integrations required by climate studies.