[Carpet] Carpet Status Report (2006-05-13)

[Erik Schnetter]

Carpet Status Report

The last month saw a flurry of development.  I hope that the changes  
did not destabilise the distribution.

The Coarse grid spacing is now defined again in global mode when  
there is only a single patch.  With multiple patches, different grids  
can have different grid spacings, so there is no unique coarse grid  
spacing in global mode in that case.

Steve White updated many parameter files and test cases.

Christian Ott introduced a schedule group "PreRegrid", which is  
executed just before Carpet changes the grid hierarchy.  Routines  
which decide whether, when, and how to regrid should be scheduled  
there instead of in the poststep bin.

The parameter Carpet::enable_all_storage = no should now work.  It  
still has many rough edges, especially for I/O methods which want to  
calculate reductions, which is not possible unless a variable has  
storage all the time.  This parameter may save a substantial amount  
of memory.  Note: MoL's scratch variables could also be dynamically  
enabled and disabled to save more memory.

Carpet's Bugzilla is being used again.  People report problems, and  
people solve them.  (I solve them, but maybe this will change in the  
future...)

There was a discussion about which grid points should be affected by  
restriction.  Carpet switched back to the old behaviour, and a new  
parameter CarpetLib::omit_prolongation_points_when_restricting can be  
used to experiment with Carpet's behaviour.

CarpetIOHDF5 now supports reading the grid structure from checkpoint  
files.  Yay!

Carpet now checks for proper nesting of the refinement hierarchy.   
Improperly nested hierarchies lead to warnings (and probably bad  
convergence), but Carpet does not abort.  Check for these warnings.

There is a new thorn CarpetRegrid2 which makes setting up a grid  
hierarchy much easier.  Instead of defining the extent of every  
region, one specifies a centre, and then an extent for each  
refinement level.  The hierarchy is automatically made consistent  
with itself and with the boundary conditions.

-erik

-- 
Erik Schnetter <schnetter at cct.lsu.edu>

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: 186 bytes
Desc: This is a digitally signed message part
Url : /archives/developers/attachments/20060513/74d9450d/attachment.pgp