Adding PeriodicCarpet to the Einstein Toolkit

Create issue
Issue #1139 closed
Eloisa Bentivegna created an issue

Old thorn Periodic does not behave correctly with AMR patches that overlap the periodic boundaries (see ticket #694). A significantly different mechanism to handle has been implemented, by Erik, in LSUThorns/PeriodicCarpet. This thorn should be added to the Einstein Toolkit, and slowly replace Periodic.

Before introducing this thorn to the ET, we need at least one testcase with AMR.


Comments (9)

  1. Roland Haas
    • removed comment

    Really replace? Will it work with PUGH (required at least for test cases if not for much else anymore).

  2. Erik Schnetter
    • removed comment

    This thorn is specific to Carpet, as the name indicates. PUGH provides built-in periodic boundaries and doesn't need an external thorn.

    I believe the thorn is ready to be included except for the lack of documentation.

  3. Eloisa Bentivegna reporter
    • removed comment

    And the lack of a test with AMR that proves that the thorn succeeds where Periodic fails. I'm working on this now.

  4. Eloisa Bentivegna reporter
    • removed comment

    I've now committed an AMR test and some documentation. Please take a look and see if anything can be added or improved.

  5. Frank Löffler
    • removed comment

    After a quick glance - Please mention in "param.ccl: peridic" that this enables all directions (in particular that you don't need this if you set (some of) periodic_[xyz]). Alternatively, the parameter could be named periodic_xyz. - Please document that none of these parameters can be used to turn periodic boundary conditions off. In particular, 'periodic = yes; periodic_z = no' wouldn't do what you might expect.

  6. Log in to comment