Hello all,
does anyone feel confident to state if / when one can skip the SYNC in POSTRESTRICT ? Assuming all computation by user thorns in the bin is purely pointwise of course.
My run has no symmetries, and (for the sake of the argument) assume that the resolution is high enough so that the buffer region of level N+1 is not affected by the grid poinst in level N that are affected by restriction from level N+1. So basically (as far as I understand this) my question boils down do: does restriction fill in values in ghost points?
Yours, Roland
Hello all,
My run has no symmetries, and (for the sake of the argument) assume that the resolution is high enough so that the buffer region of level N+1 is not affected by the grid poinst in level N that are affected by restriction from level N+1.
Uh, nevermind. This never happens. The buffer points will always be affected by what was just restricted into the coarse grid (unless one uses "additional overlap zones").
Yours, Roland
users@lists.einsteintoolkit.org
-
Roland Haas -
Roland Haas