On Fri, May 7, 2010 at 3:53 PM, Karen Lee <[email protected]> wrote: > >> Distinct, yes, but have you checked whether it is inverted? Assuming >> this is a tet, it will be inverted (and thus have negative jacobians) >> if you take one of the vertices and 'pull it through' the opposite >> face. You say the solution came from a previous libmesh run, what >> mesh format did you write in? >> >> > It is a tet... I'm not sure how I check the Jacobian (using the standard > shape function?). > > It seems strange to me that I would have to take one of the vertices and > "move it through" the opposite face. I have literally a million elements and > I doubt I can do that for every element... Is there a way I can bypass this > and make it happy?
Oh, I'm not saying you *should* move vertices, I was just trying to explain how an element can become inverted by reading in nodes in the wrong order. > I wrote in the gmv format, which I read in using the tetgen format (since > I'm used to reading in tetgen...) Would this be the potential issue? I don't see how that's possible... the tetgen reader shouldn't work with GMV files and vice-versa. -- John ------------------------------------------------------------------------------ _______________________________________________ Libmesh-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
