[Carpet] Error when using moving grids

[Yosef Zlochower]

Erik Schnetter wrote:
> On Aug 23, 2006, at 15:44:39, Yosef Zlochower wrote:
> 
>>> I take back my statement that there should be no time interpolation.
>>> In order to fill the past time levels during regridding, Carpet  has  to
>>> interpolate in time.  With prolongation_order_time=0, Carpet uses   only
>>> 1 time level to do so, namely the current time level -- and that   means
>>> that this "interpolation" has to occur at just the time of that  time
>>> level.
>>
>>
>> I don't understand why Carpet would need to interpolate in time.
>> Data on current and past time levels can  be obtained by spatial
>> interpolation only.
> 
> 
> I call in "time interpolation" because this is a generic expression.   
> In your case, the interpolation happens to be exact.
> 
> If you specify an nth order time interpolation, then Carpet uses the  
> most current n+1 time levels to interpolate.  This could be changed  -- 
> Carpet could instead choose some other set of time levels for its  
> interpolations -- but it is unambiguous.
> 
> In your case, Carpet needs to fill in the past time level with data  
> only from the current time level.  That is not an interpolation any  
> more, it is an extrapolation, which is why Carpet complains.
> 
>>> I think you should set the order of time interpolation to a value
>>> corresponding to the number of time levels which you use.  I  assume  
>>> you
>>> use 2 time levels, hence using linear interpolation would be   adequate.
>>> Try prolongation_order_time=1.
>>>
>>> -erik
>>>
>>
>> I get a similar error with 1st order prolongation
>>
>> WARNING level 0 in thorn CarpetLib processor 0 host master
>>    (line 1680 of
>> /home/yosef/NCactus/Cactus_New/configs/carpetmol/build/CarpetLib/ 
>> data.cc):
>>    -> Internal error: extrapolation in time.  time=-1  times=[1,0]
> 
> 
> Do you get this error with the exact parameter file that you sent  
> earlier, except that you changed the prolongation order in time?

No, that parfile takes several hours to crash. I modified the parfile to
remove all but two levels of refinement, and used the carpetRegrid 
thorn with a moving refinement region. The error went away when I set
the prolongation order to 2. I actually had been using 3 time levels
(although only 2 were needed). When I modified the code to use
2 timelevels I stopped seeing the error.

Yosef
> 
> -erik
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> developers mailing list
> developers at lists.carpetcode.org
> http://lists.carpetcode.org/listinfo/developers