Hello Martin I am patching many charts to define a manifold to cover the sphere. I use MappingManifold. I suspect at the interface of two charts, which also coincides with the mesh partition boundary, there may be some round off error.
Best praveen On Wed, Nov 23, 2016 at 7:12 PM, Martin Kronbichler < [email protected]> wrote: > Dear Praveen, > > Called compress(VectorOperation::insert), but the element received from a > remote processor, value 3.343515457745695e-15, does not match with the > value -3.343515457745695e-15 on the owner processor 0 > > > I dont understand why this error is triggered. > > The error says that two processors have set the same vector entry to > different values. On a VectorOperation::insert operation, the vector does > not know which of the two conflicting values it should choose. This should > not happen. Can you describe what kind of function you pass to the > VectorTools::interpolate function? Given that the value is relatively > small, I assume this happens because two processors decide to put the node > location of a certain degree of freedom on a distance of machine epsilon > apart. Given that they create different meshes, they will follow different > paths down to a certain vertex along which they might refine differently. > > The reason the check fails is that we measure a relative distance between > the two entries 3.343515457745695e-15 and -3.343515457745695e-15, which > is of course O(1) and fails. Since I'm hesitating to simply put an > additional absolute check (what should happen if 1e-15 is a big number and > small numbers are 1e-30?), I would try to get a global idea of the > magnitude of expected entries in the vector by computing an infinity norm. > This is a debug section only, so the cost is no problem. Any other idea? > > Best, > Martin > > -- > The deal.II project is located at http://www.dealii.org/ > For mailing list/forum options, see https://groups.google.com/d/ > forum/dealii?hl=en > --- > You received this message because you are subscribed to the Google Groups > "deal.II User Group" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
