WeylScal4 teukolskyID test is failing

Issue #893 closed
Ian Hinder created an issue

The test WeylScal4/teukolskyID started failing on 13-May-2012. See http://damiana2.aei.mpg.de/~ianhin/testsuites/einsteintoolkit/, specifically http://damiana2.aei.mpg.de/~ianhin/testsuites/einsteintoolkit/einsteintoolkit/2012-05-13-010001/WeylScal4_teukolskyID.1.diffs

One of the following commits is probably responsible:

simfactory Correction (eschnett, Sun, 13 May 2012 03:20:57 +0000) (details)

simfactory Support Titan, the Perimeter HPC system (eschnett, Sat, 12 May 2012 22:32:04 +0000) (details)

arrangements/EinsteinAnalysis/AHFinderDirect Support CCTK_REAL4 (eschnett, Sat, 12 May 2012 00:40:26 +0000) (details)

flesh Clean up CCTK_GFINDEX definitions (eschnett, Sat, 12 May 2012 00:41:49 +0000) (details)

repos/carpet CarpetLib: disable debug output (Roland Haas, Wed, 9 May 2012 16:56:06 -0700) (details)

arrangements/CactusNumerical/Periodic Add parameter to poison periodic boundaries (eschnett, Fri, 11 May 2012 23:20:57 +0000) (details)

arrangements/EinsteinBase/ADMBase Allow different boundary conditions in ADMBase (eschnett, Fri, 11 May 2012 23:15:39 +0000) (details)

arrangements/EinsteinEvolve/GRHydro Support real16 (and real4) in GRHydro (eschnett, Fri, 11 May 2012 22:46:36 +0000) (details)

arrangements/EinsteinEvolve/GRHydro GRHydro: correct reconstruction of velocities in ePPM (rhaas, Fri, 11 May 2012 22:17:18 +0000) (details)

repos/carpet Remove all #warning statements (Erik Schnetter, Fri, 11 May 2012 13:43:32 -0400) (details)

CarpetLib: Correct determining overlap zones (Erik Schnetter, Fri, 11 May 2012 12:55:38 -0400) (details)

CarpetLib: Correct determining buffer and overlap zones (Erik Schnetter, Fri, 11 May 2012 12:52:59 -0400) (details)

LoopControl: Explain why downcast is safe (Erik Schnetter, Fri, 11 May 2012 12:52:09 -0400) (details)

arrangements/CactusNumerical/SphericalSurface Use cast to avoid warning. Comment out unused variable. (eschnett, Fri, 11 May 2012 16:34:56 +0000) (details)

arrangements/CactusPUGH/PUGHReduce Remove unused variables (eschnett, Fri, 11 May 2012 16:50:05 +0000) (details)

arrangements/EinsteinInitialData/NoExcision Remove #warning (eschnett, Fri, 11 May 2012 16:50:34 +0000) (details)

flesh Declare string array as "const char*" (eschnett, Fri, 11 May 2012 16:50:51 +0000) (details)

arrangements/CactusBase/IOUtil Remove unused variable (eschnett, Fri, 11 May 2012 16:31:42 +0000) (details)

Keyword: WeylScal4
Keyword: testsuite

Comments (6)

  1. Roland Haas
    • removed comment

    strong contenders are the ADMBase change. The test does not use GRHydro or PUGH or NoExcision or AHFinderDirect.

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

    It seems like this is caused by ADMBase r66 not taking the boundary width into account anymore and instead always setting the width to 1. The attached patch to ADMBase restores the old width and makes the test pass again. This assumes of course that the change in ADMBase was not intentional.

  3. Log in to comment