Subregion vs norm for setting the sample geometry

[marijanbeg]

Is it then the same for running simulations in ubermag to define the geometry either region and using the saturation magnetisation, or the so called region is the total simulation space and the geometry is defined via the saturation magnetisation?

[marijanbeg]

These are three different things: (i) the region defines the cuboid in space that you want to simulate. (ii) the mesh is a spatial discretisation of this region into tiny simulation cells. (iii) to define the magnetisation on this mesh, you need to provide a vector field to Ubermag. You can do this by giving a direction and Ms value as Python functions, or by defining a spatially varying field using subregions in the mesh and then passing dictionaries to define the value and norm.