Massively Parallel Implementation of the Penn State/ncar Mesoscale Model

1. INTRODUCTION Parallel computing promises signiicant improvements in both the raw speed and cost performance of mesoscale atmospheric models. On distributed-memory massively parallel computers available today, the performance of a mesoscale model will exceed that of conventional supercom-puters; on the teraaops machines expected within the next ve years, performance will increase by several orders of magnitude. As a result, scientists will be able to consider larger problems, more complex model processes, and ner resolutions. In this paper, we report on a project at Argonne National Laboratory that will allow scientists to take advantage of parallel computing technology. This Massively Parallel Mesoscale Model (MPMM) will be functionally equivalent to the Penn State/NCAR Mesoscale Model (MM). In a prototype study, we produced a parallel version of MM4 using a static (compile-time) coarse-grained \patch" decomposition. This code achieves one-third the performance of a one-processor CRAY Y-MP on twelve Intel i860 microprocessors. The current version of MPMM is based on MM5 and uses a more ne-grained approach , decomposing the grid as nely as the mesh itself allows so that each horizontal grid cell is a parallel process. This will allow the code to utilize many hundreds of processors. A high-level language for expressing parallel programs is used to implement communication streams between the processes in a way that permits dynamic remap-ping to the physical processors of a particular parallel computer. This facilitates load balancing, grid nesting, and coupling with graphical systems and other models. In the rest of this paper, we rst introduce the Penn State/NCAR model and discuss issues that must be addressed in a massively parallel implementation. We then discuss the prototype parallel MM4 and the subsequent ne-grained paral-lelization of MM5 to produce MPMM. 1.1 Penn State/NCAR Model The Penn State/NCAR Mesoscale Model is a limited-area primitive equation model, designed to simulate meso-alpha scale (200{2000 km) and meso-beta scale (20{200 km) atmospheric circulation systems (Anthes, 1987; Grell et al., 1992). It has been developed over a period of twenty years, rst at the Pennsylvania State University and more recently also at the National Center for Atmospheric Research. The time integration uses a fourth-order nite diier-ence scheme (Brown and Campana, 1978) with optional split-explicit time stepping. Model physics is highly conngurable across a variety of convec-tion, clouds, planetary boundary layer, radiative transfer, and other standard and optional param-eterizations. The model may be run nonhydro-statically for high-resolution simulations (Dudhia, 1992). The model provides …