Parallelization in OM2p/OM3p mainly takes advantage of the first two approaches listed in “Parallel Processing”. Large meshes may be divided via domain decomposition into smaller portions that can be solved on separate processes. If multiple phase advances are requested, they may be distributed among available processes (subject to available memory) since these may be solved independently. Note that multiple modes at a given phase advance cannot be solved independently.
The OM2p/OM3p linear solver is parallelized as described in “Linear Solver”.