On Wed, Feb 27, 2013 at 12:05 PM, Jed Brown <jedbrown@mcs.anl.gov> wrote:

On Wed, Feb 27, 2013 at 2:09 AM, Dmitry Karpeev <karpeev@mcs.anl.gov> wrote:

I made some line comments. Did you test this with petsc-3.2 or earlier?
I tested with petsc-dev, petsc-3.3 and petsc-3.2. 
 DMlibMeshCreateXXXDecomposition() is "broken" (produces no decompositions at the moment),
but the API compatibility is otherwise restored (relative to petsc-dev).  I will work on finishing DMlibMeshCreateXXXDecomposition() work in the coming week or so.  Note that there appears to be a regression on the petsc-dev side with respect to the VI solve, but I haven't had a chance to chase it down, yet.
The symptom is this: miscellaneous/ex7 with -snes_type virs works fine with petsc-3.3 but doen't appear to work with -snes_type vinewtonrsls and petsc-dev, while the libMesh code is the same except for the DMlibMesh plumbing.  I'll try to take a look at it when I have a free minute.

If there was a regression in PETSc, we need to have a better test.