If you are using the classic direct solver for your simulation, optimization of the element numbering is critical in order to obtain the best performance from the Ansys Polyflow solver (for additional information on the classic direct solver, see Classic Direct Solver). This optimization is performed automatically by Ansys Polyflow before it begins a calculation. It is also possible to optimize the mesh in some mesh generators (I-deas and PATRAN), or in Ansys Polydata, and have Ansys Polyflow use that optimized mesh instead of performing the optimization itself.

For large 3D meshes, decomposition of the mesh into sub-parts is another method used by Ansys Polyflow to improve the solver performance. Typically, on a 3D mesh, you can expect the CPU time and memory usage to be reduced by a factor of 2 to 10, depending on the mesh size and topology.

Ansys Polyflow takes care of mesh optimization and decomposition for you to improve the solver performance when you are using the classic direct solver, but you also have the option to perform them yourself, either in your mesh generator or in Ansys Polydata.