[Users] Loss of convergence with subcycling in time

[Miguel Zilhão]

hi all,

i haven’t been able to join the weekly calls lately, but i wanted to 
bring this up again.

i recall that the problem is that, unless i’m missing something, the 
electric constraint isn’t converging at all in a very simple setting of 
a static charged black hole setup. this could be a real issue, 
especially since referees often ask for convergence tests and plots (and 
thus this could affect potential publications).

i’m happy to help dig into this further, but i’m not super familiar with 
Carpet's internals. should i file a bug report, or is someone already 
looking into it?

thanks,
Miguel


On 19/11/2025 17:33, users-bounces at einsteintoolkit.org wrote:
> hi Erik, all,
> 
> thanks for your input. for this specific example the grid is not moving 
> at all, so i was expecting it to be a simpler setting than the case with 
> binary BHs...
> 
> is it then expected to see all that noise propagating out of the buffer 
> region and contaminating the whole grid in such a short amount of time 
> (plot attached)? this is a bit uncomfortable, since the convergence on 
> the electric constraint violation is completely lost... if it were a 
> matter of dropping from 4th to 2nd order (as is the case in the L2 norm 
> of the Hamiltonian constraint) i'd be fine with it, but a complete lack 
> of convergence is difficult to justify...
> 
> thanks,
> Miguel
> 
> 
> On 19/11/2025 16:23, Erik Schnetter wrote:
>> On Nov 19, 2025, at 10:26, Steven Brandt via Users 
>> <users at einsteintoolkit.org> wrote:
>>>
>>> On 11/19/2025 8:24 AM, Erik Schnetter wrote:
>>>> Miguel
>>>>
>>>> If I recall correctly, Ian Hinder studied convergence of black hole 
>>>> simulations with subcycling in time in great detail. The Einstein 
>>>> Toolkit gallery example for GW150914 contains the respective 
>>>> distilled knowledge. https://einsteintoolkit.org/gallery/bbh/index.html
>>>>
>>>> Some important details that I recall:
>>>> - You can regrid only when the fine and coarse grids are aligned
>>>> - You cannot use time interpolation at all. You need to use enough 
>>>> buffer zones for all the RK substeps for all the fine timesteps for 
>>>> each coarse time step. With 3 ghost zones and RK4 you need 21 buffer 
>>>> zones.
>>>
>>> Does no time interpolation mean no dense output? That didn't exist 
>>> when Ian did these tests, right?
>>>
>> Output doesn't affect time evolution, so it doesn't matter which way 
>> you output things. Of course, if you use second-order accurate 
>> interpolation to output a quantity you cannot expect 4th order 
>> convergence for these quantities. If you output time-interpolated 
>> values of e.g. the lapse then you should check convergence only for 
>> the fine grid values of the lapse there, not for the interpolated 
>> coarse grid values.
>>
>> -erik
>>
>> -- 
>> Erik Schnetter <schnetter at gmail.com>
>> http://www.perimeterinstitute.ca/personal/eschnetter/
>>
>>
>> _______________________________________________
>> Users mailing list
>> Users at einsteintoolkit.org
>> http://lists.einsteintoolkit.org/mailman/listinfo/users
> 
> _______________________________________________
> Users mailing list
> Users at einsteintoolkit.org
> http://lists.einsteintoolkit.org/mailman/listinfo/users