The invention discloses a
parallel computing method for a distributed hydrodynamic model of a large-scale
watershed system. The
parallel computing method includes performing a
parallel algorithm to a distributed numerical model of the large-scale
watershed, with regular nets as basic computing elements and partitioned zones as compute nodes. The problem of limited 2G
memory array of a win32 program in a large-scale
watershed simulation is solved by diving the watershed zone and performing a distributed cluster computing, and the large-scale watershed
simulation is fine; based on a distributed
message passing model, data are exchanged by establishing adjacent partitioned overlap zones, thus error correction of
boundary data of the partitioned node is achieved, and consistency and accuracy of the whole watershed
simulation are realized; according to a parallel general protocol integrating OPENMP and MPI (
message passing interface), computing parallelization of and in the partitioned nodes is achieved, and efficient fine simulation of the large-scale water shed is realized. By the use of the
parallel computing method, the numerical simulation of the hydrodynamic force of the large-scale watershed
system is available with low cost and high efficiency and speed.