General Circulation Models (GCMs) are at the top of the hierarchy of numerical models that are used to study the Earth's climate. To increase the significance of predictions using GCMs requires ensembles of integrations that in turn demand large amounts of computing resources. GCMs codes are particularly difficult to optimize in view of their heterogeneity. In this paper we focus on code optimization for GCMs of the atmosphere (AGCMs), one of the major components of the climate system. In this paper, we present our efforts in optimizing the parallel UCLA AGCM code. The UCLA AGCM is a state-of-the-art finite-difference model of the global atmosphere. Our optimization efforts include the implementation of load balancing schemes, new physical parameterizations of atmospheric processes, code restructuring and use of special mathematical functions. At the beginning of this work, the overall execution time of the code was 459 seconds per simulated day in 256 nodes of a CRAY T3D. At present, the same model configuration requires 51 seconds per simulated day in 256 nodes of a CRAY T3E-900, which is approximately 9 times faster. The peak model performance is about 40 GFLOPs on 512 T3E-900 nodes. We present results in support of our conclusion that major advances in our ability to carry out longer and more detailed climate simulations depend primarily upon development of more powerful supercomputers and that code optimization, for a particular computer architecture, and development of more efficient algorithms can be nearly as important.

An analysis is presented of the primary factors influencing the performance of a parallel implementation of the UCLA atmospheric general circulation model (AGCM) on distributed-memory, massively parallel computer systems. Several modifications to the original parallel AGCM code aimed at improving its numerical efficiency, load-balance and single-node code performance are discussed. The impact of these optimization strategies on the performance on two of the state-of-the-art parallel computers, the Intel Paragon and Cray T3D, is presented and analyzed. It is found that implementation of a load-balanced FFT algorithm results in a reduction in overall execution time of approximately 45% compared to the original convolution-based algorithm. Preliminary results of the application of a load-balancing scheme for the physics part of the AGCM code suggest that additional reductions in execution time of 10–15% can be achieved. Finally, several strategies for improving the single-node performance of the code are presented, and the results obtained thus far suggest that reductions in execution time in the range of 35–45% are possible. © 1998 John Wiley & Sons, Ltd.

An analysis is presented of several factors influencing the performance of a parallel implementation of the UCLA atmospheric general circulation model(AGCM) on massively parallel computer systems. Several modifications to the parallel AGCM code aimed at improving its numerical efficiency, interprocessor communication cost, load-balance and cache efficiency are discussed. The impact of some of the optimization strategies on the performance of the AGCM code as we implemented on several state-of-the-art parallel computers, including the Intel Paragon, Cray T3D and IBM SP2, is presented and analyzed.

A methodology for distribution of a global climate model among computers connected by a wide area network is presented. The application consists of a model of the global atmosphere coupled to a model of the world ocean. It is demonstrated that a 'metacomputer' consisting of a CRAY Y-MP at the Jet Propulsion Laboratory and an Intel Delta at the California Institute of Technology connected by a highspeed (Gigabit per second) network can result in a superlinear speedup of execution of the atmospheric component of the global climate model code, despite the added overheads due to latency and communication delays.>

As part of a long range plan to develop a comprehensive climate systems modeling capability, the authors have taken the atmospheric general circulation model originally developed by Arakawa and collaborators at UCLA and have recast it in a portable, parallel form. The code uses an explicit time-advance procedure on a staggered three-dimensional Eulerian mesh. They have implemented a two-dimensional latitude/longitude domain decomposition message passing strategy. Both dynamic memory management and interprocess communication are handled with macro constructs that are preprocessed prior to compilation. The code can be moved about a variety of platforms, including massively parallel processors, workstation clusters, and vector processors, with a mere change of three parameters. Performance on the various platforms as well as issues associated with coupling different models for major components of the climate system are discussed.>

The authors investigate the distribution of a climate model across homogeneous and heterogeneous computer environments with nodes that can reside at geographically different locations. The application consists of an atmospheric general circulation model (AGCM) coupled to an oceanic general circulation model (OGCM). Three levels of code decomposition are considered to achieve a high degree of parallelism and to mask communication with computation. First, the domains of both the grid-point AGCM and OGCM are divided into sub-domains for which calculations are carried out concurrently (domain decomposition). Second, the model is decomposed based on the diversity of tasks performed by its major components (task decomposition). Last, computation and communication are organized in such a way that the exchange of data between different tasks is carried out in subdomains of the model domain (I/O decomposition). In a dedicated computer/network environment, the wall-clock time required by the resulting distributed application is reduced to that for the AGCM/Physics, with the other two model components and interprocessor communications running in parallel.>

We discuss a method for distributing a coupled atmosphere-ocean general circulation model across high-speed gigabit per second networks.Based on the special characteristics of the model code, three levels of decomposition are considered to achieve parallelism.It is argued that the distribution will enable the utilization of computing resources at geographically separated locations and speed up model execution nonlinearly through the usage of heterogeneous computing environments.1