[Users] McLachlan constraint time levels

[Erik Schnetter]

Yes, that would be a good idea and would make the behaviour clearer.
However, as the number of time levels is chosen via a parameter, the
prolongation type also needs to be a parameter. Currently, Carpet
changes the prolongation type internally if there are not enough time
levels, and then outputs this warning.

-erik

On Thu, Feb 10, 2011 at 12:19 AM, Bruno C. Mundim <bcmsma at astro.rit.edu> wrote:
> Ok, that seems fair enough, but shouldn't we then tag those groups with
> "prolongation=none"? that would avoid the warning messages when looking at
> 3D output. What do you think?
>
> Thanks,
> Bruno.
>
>
> Erik Schnetter wrote:
>>
>> Bruno
>>
>> This depends on whether you want the mesh refinement boundaries of the
>> constraint variables be interpolated from the next coarser grid, or
>> remain undefined/set to zero. In principle, these boundaries are not
>> important if you are looking for 3D output, but if you want to look at
>> reductions then you will need time interpolation. However, this
>> requires in addition that the constraints be calculated at every
>> iteration, and is thus expensive both in terms of memory requirements
>> and computing time requirements. Most people are therefore fine with
>> undefined prolongation boundaries and only look at e.g. 1D output,
>> ignoring the ghost and boundary points.
>>
>> -erik
>>
>> On Wed, Feb 9, 2011 at 5:34 PM, Bruno Coutinho Mundim
>> <bcmsma at astro.rit.edu> wrote:
>>>
>>> Hi,
>>>
>>> I am receiving the messages of the following kind in my simulation with
>>> McLachlan:
>>>
>>> (line 1084 of ./Cactus/arrangements/Carpet/CarpetReduce/src/reduce.cc):
>>>  -> Grid function "ML_BSSN::H" has only 1 time levels on refinement
>>> level 0; this is not enough for time interpolation
>>>
>>> and it is indeed true that the constraint variables has 1 time level
>>> only:
>>>
>>> CTK_REAL ML_Ham type=GF timelevels=1 tags='tensortypealias="Scalar"
>>> tensorweight=0'
>>> {
>>>  H
>>> } "ML_Ham"
>>>
>>>
>>> I was wondering if we shouldn't be using 3 time levels instead.
>>>
>>> Thanks,
>>> Bruno.
>>> _______________________________________________
>>> Users mailing list
>>> Users at einsteintoolkit.org
>>> http://lists.einsteintoolkit.org/mailman/listinfo/users
>>>
>>
>>
>>
>



-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~eschnett/