Implementation and performance evaluation of a parallel ocean model

摘要

Ocean models have been used to study global phenomena such as the El Nino Southern Oscillation. The effects of El Nino are felt worldwide and are of importance to all segments of society. High resolution ocean models are being developed to run more sophisticated simulations. The demand for computing power is expected to increase drastically with the use of high resolution models. Running extended simulations with high resolution models on current vector supercomputers will require computing power which is not available or prohibitively expensive. We have developed a parallel implementation suitable for running high resolution simulations. We studied performance issues such as the communication overhead and cache hit rate. Our experiments with grid partitioning and inter--processor communication techniques enabled us to obtain an efficient parallel implementation. Three grid partitioning techniques are evaluated. Two schemes to gather and scatter global grid data among processors are compared. A cache model to analyze performance is described.