Concurrent Meshing
You can use multiple processors to speed up volume mesh generation and mesh individual parts on separate processors.
This feature is useful when meshing large multi-part assemblies of relatively small parts in per-part mode.
Instead of meshing each individual part successively in serial, the concurrent mesher meshes separate parts simultaneously on different processors that are available.
Concurrent meshing behaves in a similar way to parallel meshing, however, the distribution of parts between processors is different:
- During concurrent meshing, each processor meshes a different part.
- During parallel meshing, parts are subdivided and meshed between different processors.
All except one of the processors that are available for concurrent meshing are assigned separate parts for meshing. This processor acts as a control node and delegates work to the other processors. For this reason, there is no improvement in speed-up when using fewer than three processors for concurrent meshing.
You can use Part Priority to define the order in which parts are meshed concurrently.
Meshes that are generated using concurrent meshing are non-conformal and are of comparable quality to meshes that are generated in serial.
The following meshers support concurrent meshing:
- Surface Remesher
- Surface Wrapper
- Trimmed Cell Mesher
- Polyhedral Mesher (including the underlying tetrahedral mesher)
- Tetrahedral Mesher
- Prism Layer Mesher
- Thin Mesher
- Advancing Layer Mesher
- Directed Meshing
Currently, only the Generalized Cylinder Mesher does not support concurrent meshing.