On 28 January 2010 11:33, Anders Logg <[email protected]> wrote:
On Thu, Jan 28, 2010 at 10:16:44AM +0100, Kristian Oelgaard wrote:On 27 January 2010 23:39, Anders Logg <[email protected]> wrote: >Some results from tabulate_tensor seemed to be off by a factor 2 (in 2D). >I tracked this down to a scaling issue where we need to scale the >weights on a facet such that the sum of the weights is 1. > >I have added this in monomialintegration.py where I multiply by 2 in >2D and by 6 in 3D. You don't need any scaling, but if we did shouldn't it be 8 in 3D? We also had a scaling factor of 0.125 in the old FFC. This is why some results are 'still' wrong for quadrature I think.You're correct. I hadn't pulled the FIAT update. Looks much better now! :-)
Good, will you revert the changes you made in FFC? Kristian
I think the scaling should have been a factor 6 which is the inverse area of the reference tetrahedron but I guess it's not important. -- Anders -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) iEYEARECAAYFAkthZ+YACgkQTuwUCDsYZdF8sgCeOq3LZBS5N8IFMf/FSRAMIQ6K tYoAn3cXmECKegM5gbZa7jcAmPdb60gq =HZzv -----END PGP SIGNATURE-----
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Mailing list: https://launchpad.net/~ffc Post to : [email protected] Unsubscribe : https://launchpad.net/~ffc More help : https://help.launchpad.net/ListHelp
