# DistributedGeneratedMeshGenerator

<!—!syntax description /MeshGenerators/DistributedGeneratedMeshGenerator—>

## Description

Similar to GeneratedMeshGenerator - builds lines, rectangles and rectangular prisms. It differs though in the way the mesh is constructed in parallel. While GeneratedMeshGenerator creates a full copy of the mesh on every processor - DistributedGeneratedMeshGenerator only creates the elements / nodes each processor is assigned. This makes it _much_ faster in parallel and much leaner in memory.

note

Be aware: DistributedGeneratedMeshGenerator turns on parallel_type = distributed. So make sure that everything you're using in your problem works with that before using this!

DistributedGeneratedMeshGenerator works by first creating a "dual graph" of the element connectivity - before ever building an elements. It then uses METIS to partition that graph - assigning elements to processors. Then, each processor can read the partition map and build only the elements that need to be on that processor. Final steps include adding in "ghosted" elements and making sure that boundary conditions are right.