A COMPARISON OF DATA-PARALLEL AND MESSAGE-PASSING VERSIONS OF THE MIAMI ISOPYCNIC COORDINATE OCEAN MODEL (MICOM)

A two-pronged effort to convert a recently developed ocean circulation model written in Fortran-77 for execution on massively parallel compu ters is described. A data-parallel version was developed for the CM-5 manufactured by Thinking Machines, Inc., while a message-passing versi on was developed for both the Gray T3D and the Silicon Graphics ONYX w orkstation. Since the time differentiation scheme in the ocean model i s fully explicit and does not require solution of elliptic partial dif ferential equations, adequate machine utilization has been achieved wi thout major changes to the original algorithms. We developed a partiti oning strategy for the message passing version that significantly redu ces memory requirements and increases model speed. On a per-node basis (a T3D node is one Alpha processor, a CM-5 node is one Spare chip and four vector units), the T3D and CM-5 are found to execute our ''large '' model version consisting of 511 x 511 horizontal mesh points at rou ghly the same speed.