Hello all,
i have a few questions regarding the example of a collapsing NS from the recent Einstein Toolkit paper...
in figure 19, the evolution of the radius of the NS is shown, alongside with the appearance of the Apparent horizon...
the plot shows the evolution up to a time of roughly 180 (in M_sun)...my first question is: for how long in total was the simulation running? for how much longer is the evolution stable AFTER the BH has been formed? does the code eventually crash or was there only interest in showing the successful collapse and then stop the simulation manually after that point?
furthermore, in the tov_collapse_vc_l3.par file for the collapse for the paper there the following parameters for a thorn *corecollapsecontrol* the following section:
*corecollapsecontrol::rho_max_list[0] = 2.5e15* *corecollapsecontrol::rho_max_list[1] = 3.5e15* *corecollapsecontrol::check_every = 64* * * *corecollapsecontrol::rho_max_every = 16* *CoreCollapseControl::handle_PMR = yes* *CoreCollapseControl::output_control = no* *CoreCollapsecontrol::bounce_rho = 2.0e33* *corecollapsecontrol::bounce_criterion = "density"* *corecollapsecontrol::bounce_entropy = 3.0e0*
what is this *corecollpasecontrol* thorn doing? and where can it be obtained if it is necessary for successful simulations of NS to BH collapse?
best wishes,
Vassili
Hello Vassilios,
the plot shows the evolution up to a time of roughly 180 (in M_sun)...my first question is: for how long in total was the simulation running? for how much longer is the evolution stable AFTER the BH has been formed? does the code eventually crash or was there only interest in showing the successful collapse and then stop the simulation manually after that point?
I ran the NS collapse run for the paper. Mostly there was only interest in getting the collapse. The code fails shortly after that though (with NaNs inside the AH I think is what it was).
furthermore, in the tov_collapse_vc_l3.par file for the collapse for the paper there the following parameters for a thorn *corecollapsecontrol* the following section:
These have to go. I'll put up a replacement that uses public thorns (CoreCollapseControl is not yet public [though there are no objections to making it public from the author]).
what is this *corecollpasecontrol* thorn doing? and where can it be obtained if it is necessary for successful simulations of NS to BH collapse?
It does many things but we only really need functionality to steer some grid functions.
I attach an updated parameter file which uses Frank Loeffler's Trigger thorn (https://docs.einsteintoolkit.org/et-docs/Thorns_we_know_of). Please let me know if this solves your problem (I only tested that I can start it), then I can (barring protests from the other authors), update the parameter file in the repository.
Yours, Roland
Hello Roland,
thanks for the quick reply...
when i did the collapse (with less refinement levels and resolution due to RAM restrictions on my iMac) the code crashes as well after the Bh has been formed...when using AHfinderDirect, the simulation stops upon the following error:
[1mWARNING level 0 in thorn AHFinderDirect processor 0 host permeability.daa.uv.es (line 78 of /Users/vass/programs/Cactus/arrangements/EinsteinAnalysis/AHFinderDirect/src/jtutil/error_exit.cc): -> [0m ***** row_sparse_Jacobian__UMFPACK::solve_linear_system(): error return status=1 from umfpack_numeric() routine
and when running without the AHFinderdirect, the run produces meaningless results (the density maximum and minimum as well as the lapse min and max get set to 1.797693e+308 and -1.797693e+308) however, there are no NaNs caught by the NaNChecker...
thanks for pointing out the trigger thorn, is it triggering more refinement levels once the density has reached a certain threshold?
what about the general stability of simulations after a BH has been formed? is it yet possible to evolve the system for a long time after the formation of a BH, using the current tools available in the ET? i am asking this because we are interested in gravitational collapse of NS triggered by winds past the NS and it would be desirable to observe the long term evolution of the NS-wind and later BH-wind system...
best wishes,
Vassili
On Mon, Jul 23, 2012 at 7:31 PM, Roland Haas <roland.haas@physics.gatech.edu
wrote:
Hello Vassilios,
the plot shows the evolution up to a time of roughly 180 (in M_sun)...my first question is: for how long in total was the simulation running? for how much longer is the evolution stable AFTER the BH has been formed?
does
the code eventually crash or was there only interest in showing the successful collapse and then stop the simulation manually after that point?
I ran the NS collapse run for the paper. Mostly there was only interest in getting the collapse. The code fails shortly after that though (with NaNs inside the AH I think is what it was).
furthermore, in the tov_collapse_vc_l3.par file for the collapse for the paper there the following parameters for a thorn *corecollapsecontrol*
the
following section:
These have to go. I'll put up a replacement that uses public thorns (CoreCollapseControl is not yet public [though there are no objections to making it public from the author]).
what is this *corecollpasecontrol* thorn doing? and where can it be obtained if it is necessary for successful simulations of NS to BH
collapse? It does many things but we only really need functionality to steer some grid functions.
I attach an updated parameter file which uses Frank Loeffler's Trigger thorn (https://docs.einsteintoolkit.org/et-docs/Thorns_we_know_of). Please let me know if this solves your problem (I only tested that I can start it), then I can (barring protests from the other authors), update the parameter file in the repository.
Yours, Roland
-- My email is as private as my paper mail. I therefore support encrypting and signing email messages. Get my PGP key from http://keys.gnupg.net.
Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hello Vassilios,
when i did the collapse (with less refinement levels and resolution due to RAM restrictions on my iMac) the code crashes as well after the Bh has been formed...when using AHfinderDirect, the simulation stops upon the following error:
[1mWARNING level 0 in thorn AHFinderDirect processor 0 host permeability.daa.uv.es (line 78 of /Users/vass/programs/Cactus/arrangements/EinsteinAnalysis/AHFinderDirect/src/jtutil/error_exit.cc):
- -> [0m ***** row_sparse_Jacobian__UMFPACK::solve_linear_system():
error return status=1 from umfpack_numeric() routine
and when running without the AHFinderdirect, the run produces meaningless results (the density maximum and minimum as well as the lapse min and max get set to 1.797693e+308 and -1.797693e+308) however, there are no NaNs caught by the NaNChecker...
The huge min/max of lapse would point to it picking up poison somewhere. I would have expected NaNChecker (which checks only rho, gxx and kxx) to trigger soon (it only checks every 64 iterations) after. This actually sounds like a bug since there should never be poison left one the grid. I'd have to re-run it again (did you use current code or code from a previous release) to test what happens. One can get the simulation to proceed after collapse but it takes a bit of fine-tuning as far as I know.
thanks for pointing out the trigger thorn, is it triggering more refinement levels once the density has reached a certain threshold?
Yes, that was the intention. Trigger allows general settings to be made based on conditions. For each trigger there is a Trigger_Relation parameter (which defaults to ">") that is used to evaluate the condition. The ones I set up should increase the number of refined levels to 6 and 7 when the density goes above 2.5e15 and 3.5e15 respectively.
what about the general stability of simulations after a BH has been formed? is it yet possible to evolve the system for a long time after the formation of a BH, using the current tools available in the ET? i am asking this because we are interested in gravitational collapse of NS triggered by winds past the NS and it would be desirable to observe the long term evolution of the NS-wind and later BH-wind system...
One can get it to evolve afterwards as far as I know but it requires some fine tuning and possibly use of more current code.
Yours, Roland
- -- My email is as private as my paper mail. I therefore support encrypting and signing email messages. Get my PGP key from http://keys.gnupg.net.
users@lists.einsteintoolkit.org
-
Roland Haas -
Roland Haas -
Vassilios Mewes