[Carpet] Enabling/disabling storage and time levels

[Erik Schnetter]

On Wednesday 13 July 2005 10:39, Ian Hawke wrote:
> I believe that it should abort if it does the wrong thing. The
> question is what is "wrong".
>
> There are two reasons why you wouldn't want to do standard
> prolongation. One is that the GF is time independent (coordinates,
> static conformal factor) so it is either inaccurate or a waste of
> time to prolongate. The other is that it doesn't make sense for some
> reason (a mask stored as a real number). There you need copying or
> some other prescription. I would say that these are a sufficiently
> small set of the total number of GFs that they should be the special
> case; that is, I don't like option (3).

There are two other reasons why people don't want / use prolongation.  
One is that all analysis thorns that were written for PUGH use only a 
single time level, and some allocate storage only during analysis, so 
these routines don't work with Carpet out of the box.  Some may say 
that these routines should either work out of the box and do something 
"sensible".

The other reason is that people are not willing to spend the memory for 
having three time levels for an analysis quantity such as the 
constraints.  Compared to what you spend for the evolved variables, and 
if you have many analysis variables, this can be a real factor.  This 
is in fact the current driving point: More analysis variables, but 
using less memory.

-erik

-- 
Erik Schnetter <schnetter at aei.mpg.de>   http://www.aei.mpg.de/~eschnett/

My email is as private as my paper mail.  I therefore support encrypting
and signing email messages.  Get my PGP key from www.keyserver.net.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : /archives/developers/attachments/20050713/d2fc04a6/attachment.pgp