[Users] Issue with AHFinderDirect thorn

[Erik Schnetter]

Hector

This error is reported by the elliptic solver that AHFinderDirect is
using. It is usually caused either by a horizon surface that has
become very distorted or is too close to a puncture when the horizon
cannot be found, or when the ADM variables contain nans.

If this is the first time you are looking for a particular horizon,
then it might be that your initial guess is off. If you have been
tracking this horizon for some time, then it is likely that your
simulation has gone bad and has produced nans.

Since your setup was working with the previous version of the toolkit,
you might be able to run the same simulation with both versions and
look for the first differences, either in horizon location/shape or in
the ADM variables. This might help find the cause for this behaviour.

-erik

On Thu, Jun 2, 2022 at 9:20 AM Hector L Iglesias <higlesia at utexas.edu> wrote:
>
> Hello,
>
> I am running a BBH simulation with the latest ETK release (the Johnson release) using some local thorns, and I'm encountering this particular error:
>
> WARNING level 0 from host chaac.cns.utexas.edu process 0
>   while executing schedule bin CCTK_POSTSTEP, routine AHFinderDirect::AHFinderDirect_find_horizons
>   in thorn AHFinderDirect, file /home/hli75/Cactus/configs/etabeta/build/AHFinderDirect/jtutil/error_exit.cc:78:
>   -> ***** row_sparse_Jacobian__UMFPACK::solve_linear_system():
>         error return status=1 from umfpack_numeric() routine
>
>
> WIth a previous version of ETK, it is working fine. Is there something I can do to correct this? I've attached the .out and .par files for the simulation below.
>
> Thank you!
>
> Hector Iglesias
> _______________________________________________
> Users mailing list
> Users at einsteintoolkit.org
> http://lists.einsteintoolkit.org/mailman/listinfo/users



-- 
Erik Schnetter <schnetter at gmail.com>
http://www.perimeterinstitute.ca/personal/eschnetter/