[Users] BHNS simulation issues - Hamiltonian constraint violations and general problems
Samuel Tootle
tootle at itp.uni-frankfurt.de
Mon Aug 23 14:13:43 CDT 2021
Hi Konrad,
All of the hdf5 2d outputs (e.g. admbase::lapse) include the grid data
as well at each iteration. For these quick things, I tend to use Visit
from LLNL to plot the lapse and the mesh overtop so I can see the
position of the grid as the system evolves. You could likely do the
same plot with something like KuiBit, but I haven't messed with plotting
the grid with that yet.
Also, in the output you just sent, it looks like you are using your
grhydro par file, however, in the grhydro par you sent before, quite a
lot of the puncture tracker settings are commented out including track
puncture.
Finally, the warnings of "the refined region #" is inactive for a
specific surface is also likely a contributor. I'm sure Erik et al are
more knowledgeable on resolving this, but usually this was an issue I
had based on my refinement level and spherical surface setup.
Cheers,
Samuel
On 8/23/21 7:54 PM, Konrad Topolski wrote:
> Hi
>
> I cannot set up the tracker correctly in either of these parameter files.
> The "pt_loc_" timeseries all display the (0,0,0) position at all times.
> The 'NaNMask' data is probably corrupted by simulation shutting down,
> as I cannot open it.
>
> What are the "mesh grid variables" exactly that I should use in this
> diagnostics and in which thorns should I look for them?
>
> Here's part of the log (common for both simulations in the
> PunctureTracker part):
> mesh grid variables" that
> [...]
> INFO (GRHydro): Setting up the atmosphere mask: all points are
> not_atmosphere
> INFO (PunctureTracker): Tracking punctures...
> INFO (PunctureTracker): Puncture #0 is at (17.3774,-0.0975486,0)
> INFO (PunctureTracker): Shift at puncture #0 is at (0,0,0)
> INFO (AHFinderDirect): proc 0: searching for horizon 1/1
> INFO (PunctureTracker): Setting spherical surface 0 centroid from
> puncture #0 to (17.3774,-0.0975486,0)
> 1 0.022 | 6.0400998 | -nan -nan |
> -nan -nan | 2516 3089 | 0 0
> INFO (CarpetTracker): Setting position of refined region #1 from
> surface #0 to (17.3774,-0.0975486,0)
> INFO (CarpetTracker): Refined region #2 (depending on surface #1) is
> inactive
> INFO (PunctureTracker): Tracking punctures...
> INFO (PunctureTracker): Puncture #0 is at (17.3774,-0.0975486,0)
> INFO (PunctureTracker): Shift at puncture #0 is at (-nan,-nan,-nan)
> ERROR from host nid00681 process 0
> while executing schedule bin CCTK_EVOL, routine
> PunctureTracker::PunctureTracker_Track
> in thorn PunctureTracker, file
> /lustre/tetyda/home/topolski/Cactus/arrangements/EinsteinAnalysis/PunctureTracker/src/puncture_tracker.cc:204:
> -> Shift at puncture #0 is (-nan,-nan,-nan). This likely indicates
> an error in the simulation.
>
>
> I'm also sending the initial data to confirm the initial guess for the
> puncture.
>
> Best
>
>
> niedz., 22 sie 2021 o 08:59 Samuel Tootle <tootle at itp.uni-frankfurt.de
> <mailto:tootle at itp.uni-frankfurt.de>> napisał(a):
>
> Hi Konrad,
>
> The matter looks reasonable in all the iterations until the end,
> so that's good. Have you looked at the mesh grid to see if both
> objects are being tracked properly? You asked about this being a
> possibility in your first note, but didn't mention if you looked
> at the grid or not.
>
> Cheers,
> Samuel
>
> *From: *Konrad Topolski <k.topolski2 at student.uw.edu.pl
> <mailto:k.topolski2 at student.uw.edu.pl>>
> *To: *Samuel Tootle <tootle at itp.uni-frankfurt.de
> <mailto:tootle at itp.uni-frankfurt.de>>; Einstein Toolkit Users
> <users at einsteintoolkit.org <mailto:users at einsteintoolkit.org>>
> *Date: *Aug 21, 2021 23:20:30
> *Subject: *Re: [Users] BHNS simulation issues - Hamiltonian
> constraint violations and general problems
>
> Hi Sam
>
> I am sending the pictures of the density in a few iterations.
> Something indeed goes *very *bad (and possibly in the wrong
> direction pretty soon from the start).
>
> But before a chunk of the NS turns into atmosphere it doesn't
> look like it would end up like that.
>
> This is rho.maximum:
>
> 0 0 0.00181508688452629
> 16 0.4 0.00181562065714772
> 32 0.8 0.00181716349263237
> 48 1.2 0.00181971188439579
> 64 1.6 0.00182322915996771
> 80 2 0.00182773078378139
>
> And this is rho.average:
>
> 0 0 1.23447090415481e-07
> 16 0.4 1.2346439769246e-07
> 32 0.8 1.23490260581956e-07
> 48 1.2 1.23521771090518e-07
> 64 1.6 1.23554773904041e-07
> 80 2 1.23584518348538e-07
>
>
> sob., 21 sie 2021 o 23:03 Samuel Tootle
> <tootle at itp.uni-frankfurt.de
> <mailto:tootle at itp.uni-frankfurt.de>> napisał(a):
>
> Hi Konrad,
>
> Have you checked the import of the ID from the hydro point
> of view (i.e. plot rho in for the first iterations). If
> you see drastic changes in the NS density in the initial
> iterations, it usually means the EOS manager isn't
> configured correctly which routinely results in the NS
> "vanishing" thereby destroying the solution and the
> evolution.
>
> Cheers,
> Samuel
>
> *From: *Konrad Topolski <k.topolski2 at student.uw.edu.pl
> <mailto:k.topolski2 at student.uw.edu.pl>>
> *To: *Einstein Toolkit Users <users at einsteintoolkit.org
> <mailto:users at einsteintoolkit.org>>
> *Date: *Aug 21, 2021 22:30:57
> *Subject: *[Users] BHNS simulation issues - Hamiltonian
> constraint violations and general problems
>
> Hi all
>
> I am trying to run a BH-NS simulation using at least
> one of some (and sometimes more) unofficial thors.
> That would be Kadath for the initial data and possibly
> THC for hydro evolution.
> I'm trying to run the simulation either with GRHydro
> or with THC.
>
> In both cases I believe that the initial data gets
> imported correctly - I have looked at some simple
> visuals which I'm attaching to this mail.
>
> However, as the simulation proceeds, huge hamiltonian
> constraint violations occur and e.x. the lapse gets
> driven to 0 globally. This can also be seen.
>
> How can I help this? Is it because the puncture is not
> tracked correctly? Trying to set up a PunctureTracker
> at the location of the black hole yields a failure
> very soon, as the shift at puncture is said to be
> (nan, nan, nan).
>
> I'm sending the .par files, outputs (.out), errors
> (.err) and visualization (.png).
>
> Best regards
> Konrad
> _______________________________________________
> Users mailing list
> Users at einsteintoolkit.org
> <mailto:Users at einsteintoolkit.org>
> http://lists.einsteintoolkit.org/mailman/listinfo/users
> <http://lists.einsteintoolkit.org/mailman/listinfo/users>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.einsteintoolkit.org/pipermail/users/attachments/20210823/2cfc08ed/attachment-0001.html
More information about the Users
mailing list