[ET Trac] [Einstein Toolkit] #481: Tests should be converted to use new symmetry thorns

Einstein Toolkit trac-noreply at einsteintoolkit.org
Wed Jul 27 18:57:16 CDT 2011


#481: Tests should be converted to use new symmetry thorns
-----------------------------------+----------------------------------------
 Reporter:  hinder                 |       Owner:     
     Type:  defect                 |      Status:  new
 Priority:  major                  |   Milestone:     
Component:  EinsteinToolkit thorn  |     Version:     
 Keywords:                         |  
-----------------------------------+----------------------------------------
 It was decided
 (http://cactuscode.org/pipermail/developers/2010-September/006087.html)
 that the built-in CartGrid3D symmetries are deprecated and will not work
 with the Mercurial version of Carpet.  Any test cases that use
 Grid::domain != "full" should be modified to use ReflectionSymmetry
 instead.

 In general, this should be possible without regenerating the test output.
 However, for the case of AHFinderDirect, the horizon finder actually uses
 a different internal multipatch grid structure if it detects that a
 CartGrid3D symmetry is being used.  This means that the horizon shapes and
 quantities are different at the level of numerical error with
 ReflectionSymmetry.  I propose to convert the tests to use
 ReflectionSymmetry and regenerate the test data (the grid function phi is
 the same with the new symmetry thorn, it's only the horizons which are
 different).  The failing tests are related to checkpoint and recovery, and
 are not designed to test the symmetry mechanism.

-- 
Ticket URL: <https://trac.einsteintoolkit.org/ticket/481>
Einstein Toolkit <http://einsteintoolkit.org>
The Einstein Toolkit


More information about the Trac mailing list