On Sun, 21 Apr 2013 22:24:24 +0200 Mikael Mortensen <[email protected]> wrote: > > > Den Apr 20, 2013 kl. 11:44 PM skrev Jan Blechta: > > > On Sat, 20 Apr 2013 13:24:11 +0200 > > Jan Blechta <[email protected]> wrote: > >> On Fri, 19 Apr 2013 16:23:50 +0200 > >> Mikael Mortensen <[email protected]> wrote: > >>> > >>> > >>> Den Apr 19, 2013 kl. 4:04 PM skrev Jan Blechta: > >>> > >>>> Hello Mikael, > >>>> which version of FEniCS should be used with CBC.PDESys trunk. I > >>>> tried RANS model demo turbulent_apbl.py and it crashes with > >>>> DOLFIN 1.0 (reorder_dofs_serial parameter does not exist), 1.1 > >>>> (I don't know), 1.2 (some SWIG issue), master (dofmap issue). > >>> > >>> It should work with trunk, but I have not tested the turbulence > >>> models in a few weeks. Do you need it or are you just curious? A > >>> quick check and I'm running the turbulent channel with no problem > >>> using trunk. > >> > >> By trunk do you mean DOLFIN master from bitbucket? All 3 demos in > >> cfd/ransmodels/demo fail on master with message: > >> AttributeError: 'GenericDofMap' object has no attribute 'dofs' > > > > Could be useful > > https://answers.launchpad.net/dolfin/+question/223595 for fixing it. > > Yes, I mean bitbucket and master. There seems to be a problem with > setting turbulent wall functions (some specific turbulence boundary > conditions) and I cannot find the error easily. The problems have > something to do with the recent restructuring of the dofmaps. I > thought I had it fixed already, but apparently not. The turbulent > channel case still works, but the two others are broken. Could you > please confirm that the turbulent channel case works? I will try to > fix the others as soon s possible.
turbulent_apbl.py: IOError: [Errno 2] No such file or directory: '../../data/channel_Laminar_395.0.ius' turbulent_channel.py: passes OK turbulent_diffusor.py: diverges Jan > > Mikael > > > > >> > >> There's one more bug: 'sizet' instead of 'size_t'. Just grep for > >> 'sizet' > >> > >> I have working code but for some sort of parameters turbulence > >> occurs and I need to resolve it with least effort. So I'm looking > >> into your framework if it can be merged into my code. > >> > >> Jan > >> > >>> > >>>> > >>>> Also there is a bug in this file that > >>>> problem.prm['turbulence_model'] = 'OriginalV2F' should be set > >>>> before initializatin of NS solver otherwise proper ius file is > >>>> not found. > >>> > >>> Ok, thanks. > >>> > >>>> > >>>> Should these go to launchpad? Does project continue there? > >>> > >>> I haven't really decided yet. For now it does. > >>> > >>>> > >>>> BTW is now [email protected] appropriate list for FEniCS > >>>> apps? > >>> > >>> Well, it's the only one apart from the launchpad lists, right? I > >>> think there should be more specialized lists, but this has not > >>> really been decided on yet afaik. > >>> > >>> Mikael > >>> > >>> > >>> > >>>> > >>>> Jan > >>> > >>> > >> > >> _______________________________________________ > >> fenics mailing list > >> [email protected] > >> http://fenicsproject.org/mailman/listinfo/fenics > > > > _______________________________________________ > > fenics mailing list > > [email protected] > > http://fenicsproject.org/mailman/listinfo/fenics _______________________________________________ fenics mailing list [email protected] http://fenicsproject.org/mailman/listinfo/fenics
