[Users] WeylScal4 failing testsuites with ET_2023_11

[Cupp, Samuel D.]

This should be a bug that only affects debug mode. Looking through the code, I think I've tracked back the issue. The RDWR attribute for functions (set in CCTKi_CreateRDWRData) is only defined if presync_mode is set to something other than "off" (the default). As such, this section of code should be bypassed for that parameter setting. As a quick fix, you could change line 324 of repos/flesh/src/main/RDWR.cc

from

  if(cctkGH == nullptr))

to

  if(cctkGH == nullptr || CCTK_Equals(presync_mode, "off"))

Let me know if that fixes the issue.

   Samuel Cupp
   Postdoctoral Researcher
   Department of Physics
   University of Idaho
________________________________
From: Bernard Kelly <physicsbeany at gmail.com>
Sent: Thursday, February 29, 2024 12:29 PM
To: Cupp, Samuel D. <scupp1 at my.apsu.edu>
Cc: Einstein Toolkit Users <Users at einsteintoolkit.org>
Subject: Re: [Users] WeylScal4 failing testsuites with ET_2023_11

Hi Samuel. Thanks for the quick response.

For both code releases (ET_2023_05 and ET_2023_11), I have DEBUG=yes in my config.

Now that you've pointed it out, I *could* recompile from scratch with DEBUG=no ... but should I have to? And is this just covering up an actual bug (as the name implies)?

Bernard

On Thu, Feb 29, 2024 at 3:23 PM Cupp, Samuel D. <scupp1 at my.apsu.edu<mailto:scupp1 at my.apsu.edu>> wrote:
Hi Bernard,
   Is your parfile setting the parameter presync_mode? Alternatively, did you compile in debug mode? I ask this because that function has the following code

  static bool presync_only = CCTK_Equals(presync_mode, "presync-only");
#ifndef CCTK_DEBUG
  if(!presync_only)
    return true;
#endif

Unless I'm reading this section incorrectly, the code shouldn't reach that assertion unless its in debug mode or the mode is "presync-only".

   Samuel Cupp
   Postdoctoral Researcher
   Department of Physics
   University of Idaho
________________________________
From: Users <users-bounces at einsteintoolkit.org<mailto:users-bounces at einsteintoolkit.org>> on behalf of Bernard Kelly <physicsbeany at gmail.com<mailto:physicsbeany at gmail.com>>
Sent: Thursday, February 29, 2024 11:58 AM
To: Einstein Toolkit Users <Users at einsteintoolkit.org<mailto:Users at einsteintoolkit.org>>
Subject: [Users] WeylScal4 failing testsuites with ET_2023_11

Hi. I just tried running a parameter file with WeylScal4 using the ET_2023_11 release, and it fails during the pre-initial stage, with screen output like the following:

-------------
NFO (CarpetIOHDF5): I/O Method 'IOHDF5_1D' registered: 1D AMR output of grid variables to HDF5 files
INFO (CarpetIOHDF5): I/O Method 'IOHDF5_2D' registered: 2D AMR output of grid variables to HDF5 files
INFO (CarpetIOHDF5): Periodic 2D AMR output requested for:
   ML_ADMCONSTRAINTS::H
INFO (CarpetIOHDF5): I/O Method 'IOHDF5_3D' registered: 3D AMR output of grid variables to HDF5 files
INFO (CarpetIOScalar): Periodic scalar output requested for:
   ML_BSSN::H
   ML_BSSN::M1
   ML_BSSN::M2
   ML_BSSN::M3
cactus_sim_ET_2023_11_WS4_ORIG: /nobackupp19/bjkelly1/codes/Cactus_ET_2023_11/configs/sim/build/Cactus/main/RDWR.cc:336: int cctki_RDWR::CCTK_HasAccess(const _cGH *, int): Assertion `current_function->RDWR' failed.
Rank 3 with PID 73750 received signal 6
Writing backtrace to Kerr_LES_StandardSpin_VACUUM_dx032_WS4_ORIG/backtrace.3.txt
-------------

I tried to run the WeylScal4 testsuites too (Teukolsky etc), with the same "Assertion `current_function->RDWR' failed" message.

Just to check, I went back an ET release (ET_2023_05), and did the same test, and the tests ran & passed.

So did something happen between releases to change the behavior of WeylScal4? And would it affect other Kranc-generated thorns too?

Thanks, Bernard

--
------------------------------------------------------------------
Bernard Kelly -- CRESST Assistant Research Scientist, NASA/GSFC
Gravitational Astrophysics Laboratory --- Code 663

Phone: +1 (301) 286-7243 *** Fax: +1 (301) 286-2226
Web: http://science.gsfc.nasa.gov/sed/bio/bernard.j.kelly
ORCID: orcid.org/0000-0002-3326-4454<http://orcid.org/0000-0002-3326-4454>
------------------------------------------------------------------


--
------------------------------------------------------------------
Bernard Kelly -- CRESST Assistant Research Scientist, NASA/GSFC
Gravitational Astrophysics Laboratory --- Code 663

Phone: +1 (301) 286-7243 *** Fax: +1 (301) 286-2226
Web: http://science.gsfc.nasa.gov/sed/bio/bernard.j.kelly
ORCID: orcid.org/0000-0002-3326-4454<http://orcid.org/0000-0002-3326-4454>
------------------------------------------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.einsteintoolkit.org/pipermail/users/attachments/20240229/91eff5ff/attachment-0001.htm>