Feature request for Parity Symmetry thorn

Issue #1266 closed
Yosef Zlochower created an issue

I started looking at runs with parity symmetry, that is the only symmetry is the simultaneous reflection about x, y, and z. I think I was able to "implement" this symmetry by changing just a handful of lines in Rotating180 and I'm a little less sure about my changes to carpetregrid2. I was wondering, would there be any interest in adding this symmetry to ET?

Keyword:

Comments (13)

  1. Yosef Zlochower reporter
    • removed comment

    The files par_rot_diff is the diff between the new ParitySymmetry thorn and RotatingSymmetry180. carpet_diff is the diff between the standard version of carpetRegrid2 and the version consistent with the symmetry.

  2. Roland Haas
    • removed comment

    I am not sure I understand what the difference between the parity symmetry you describe and thorn reflectionsymmetry with reflect_x = reflect_y = reflect_z = yes is.

  3. anonymous
    • removed comment

    Replying to [comment:2 rhaas]:

    I am not sure I understand what the difference between the parity symmetry you describe and thorn reflectionsymmetry with reflect_x = reflect_y = reflect_z = yes is.

    The system is not x-reflection, y-reflection, or z-reflection symmetric. Rather, the only symmetry is (x, y, z) -> (-x, -y, -z). I.e., this is a bitant symmetry where we use (x,y,-z) -> (-x, -y, +z) to fill in the z<0 ghostzones.

  4. Yosef Zlochower reporter
    • removed comment

    I placed the paritysymmetry thorn in incoming. The patch to carpetregir2 will be needed for AMR runs.

    Yosef

  5. Roland Haas
    • removed comment

    carpet_diff applies and does not break the tests. However it's comments are all referring to rotatingsymmetry180. Once these are fixed to refer to parity symmetry, that patch can be applied.

    The new thorn ParitySymmetry lacks documentation and test cases. It cannot be part of the ET in this form unfortunately. Otherwise having such a symmetry would be welcome.

  6. Yosef Zlochower reporter
    • removed comment

    Here's an updated patch for CarpetRegrid2. The testsuite for ParitySymmetry is in Incoming.

  7. Roland Haas
    • removed comment

    carpet_diff2 is fine to apply (will do so later today unless there are objections).

  8. Roland Haas
    • changed status to open
    • removed comment

    On 2013-11-12 10:34, Yosef Zlochower wrote:> I was wondering about this patch. I don't see it

    in the latest development version. Can it go in after the new release (I'm assuming you don't want to add a new feature before the new release)? Indeed. Too late to apply before the release. Can be applied afterwards, only useful with Yosef's Parity thorn as well (the thorn is in incoming I believe).

  9. Roland Haas
    • removed comment

    Just checked. The thorn is in incoming, but not in a shape to include in the ET. See comment:6 it also lists eg Erik in the README file as an author (since it is a copy of RotatingSymmetry180), most likely Yosef should be added to properly attribute credit (and blame).

  10. Log in to comment