Cubic indices are used to map the spatial information to the grid points.
A Hilbert space-filling curve is used to map the spatial information to
the grid points.
Number of levels in the grid hierarchy used for multigrid. Positive numbers indicate an absolute number of levels, negative numbers are subtracted from the maximum number of levels possible.
Calculate the optimal number of levels for the grid.
Use a trilinear interpolation. This is implemented similar to restriction and prolongation
operations in multigrid methods. This ensures that both directions are adjoint to each other.
Use the nearest neighbor for the regridding. This is faster than the linear interpolation.
Rescale the regridded quantities. Not using a rescaling can lead to bad results if the ratio of the grid spacings is large.
Do not rescale the regridded quantities.
Scale the regridded quantities by the 2-norm of the quantity on the overlapping
region of the grid.
The spacing between the points in the mesh. This controls the quality of the discretization: smaller spacing gives more precise results but increased computational cost.
When using curvilinear coordinates, this is a canonical spacing that will be changed locally by the transformation. In periodic directions, your spacing may be slightly different than what you request here, since the box size must be an integer multiple of the spacing.
The default value is defined by the image resolution if BoxShape = box_image. Othewise here is no default otherwise.
It is possible to have a different spacing in each one of the Cartesian directions if we define Spacing as block of the form
spacing_x | spacing_y | spacing_z