[Carpet] Prolongation order

[Erik Schnetter]

On Apr 3, 2008, at 10:28:02, Ian Hinder wrote:
> But from your description above, I would have thought the number of
> ghost zones required would be 2 for the 5th order stencil, but
> (prolongation_order_space+1)/2 gives 3?

A 5th order stencil (where "5" is counted according to Carpet  
conventions) requires 6 points, hence 3 the stencil radius is 3.

> Am I correct in deducing that Carpet does not fetch points from remote
> processors to interpolate them, but instead the interpolation  
> happens on
> the processor which owns the coarse grid, and the interpolated values
> are then returned to the processor which owns the fine grid, hence the
> need for ghost zones?


Yes, this is so.  The reason is that it is much cheaper to communicate  
interpolation results than interpolation stencils, hence interpolation  
always happens on the sending processor.

-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>   http://www.cct.lsu.edu/~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: PGP.sig
Type: application/pgp-signature
Size: 194 bytes
Desc: This is a digitally signed message part
Url : /archives/developers/attachments/20080403/18c12851/attachment.pgp