CarpetRegrid2 supports both RotatingSymmetry90 and cell-centring. I do not know whether this case was tested, but even if so, there may not be a test case which means it may have broken at some point.
I don't know why "assert (all (iaxis0 % istr == 0))" fails. This may be because cell-centring is not supported, or because this routine makes an assumption about how Carpet lays out grid cells with respect to the finest grid, or because of an error in the parameter file.
Could you open a ticket, and attach the parameter file as well as some debug output, i.e. the values of reflevel, bb, bnd, axis, iaxis0, hh.baseextent(0,rl)?
-erik
On Mon, Mar 18, 2013 at 5:03 PM, Roland Haas <roland.haas@physics.gatech.edu
wrote:
Hello all,
I am trying to run a cell centered run with Carpet and RotatingSymmetry90.
Using the attached parfile (cropped from the real one) I get an error
cactus_bns_all:
/mnt/data/rhaas/postdoc/gr/Zelmani/arrangements/Carpet/CarpetRegrid2/src/property.cc:382: ibset CarpetRegrid2::rotsym90::symmetrised_regions(const gh&, const dh&, const CarpetRegrid2::level_boundary&, const std::vector<bboxset<int, 3>
&, int): Assertion `all (iaxis0 % istr == 0)' failed.
Rank 0 with PID 12717 received signal 6
Right now I can just run without the symmetry (slower of course...). Before I eventually try to debug this: is Regrid2 supposed to support 90 degree symmetry when using cell centering or was it only ever tested with vertex centering or without 90 degree symmetry?
Yours, Roland
-- My email is as private as my paper mail. I therefore support encrypting and signing email messages. Get my PGP key from http://keys.gnupg.net.
Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users