PCBDDC: solved memory leaks and block size issue with change of basis matrix

PCBDDC: leak when indices has to be created

PCIS: added flag to PCIS to control for reusage of local submatrices

when using change of basis with PCBDDC, local submatrices should be destroyed because the nonzero pattern is not the same of the

PCIS: added flag to PCIS to control for reusage of local submatrices

when using change of basis with PCBDDC, local submatrices should be destroyed because the nonzero pattern is not the same of the original one

show more ...

PCBDDC: bug fix for mirrored nodes

-still remains to understand why the last local node is not mapped back

PCBDDC: improved support for coarse problem redistribution in case we are on the final level

Merge branch 'barry/refactor-petscsubcomm'

PCBDDC: forgot to add bddc.c in the previous commit

PCBDDC: allow for nmin to be 0 at command line

- fix redistribution for coarse problem if requested at coarser levels

PCBDDC: multilevel fixes to SubSchurs

PCBDDC; add flag to invert or not Stildas

small fixes

PCBDDC: fixes bugs

- dirichlet dofs should be computed since the PCBDDCGraph code allows to specify nonn-consistent sets among neighbours
- add dirichlet dofs to the list of simpli scalied dofs in d

PCBDDC: fixes bugs

- dirichlet dofs should be computed since the PCBDDCGraph code allows to specify nonn-consistent sets among neighbours
- add dirichlet dofs to the list of simpli scalied dofs in deluxe
- fix number of constraints computed during adaptive selection

show more ...

bddc: c89 fixes. [triggered by b7d87ab19e1e95173f67f3da509587a15b2e6ae1]

minor cleanup of PetscSubcomm code and documentation

PCBDDC; prepare for upcoming VecLockPush/Pop

removed the PETSC_HAVE_PCBDDC nonsense. You are either part of the PETSc repository or not!
Now it is always built and tested

squash -Wmaybe-uninitialized

PCBDDC: multilevel fixes

PCBDDC: solved warning from gcc for uninitialized variables

PCBDDC: Use MatISGetMPIXAIJ for global change of variables

PCBDDC; optimized PCBDDCSetUpCorrection

now used MatMatSolve whenever possible
extensive usage of BLAS whenever possible

MATIS: move preallocation during MatISGetMPIXAIJ into a separate routine

Schur subsets can take advantage

PCBDDC; bugfix

PCBDDC: added initial implementation for adaptive selection of constraints

positive definite case covered only
untested for complex

generalized eigenvalue problem is inv(sum(inv(S_F))) = lambda inv

PCBDDC: added initial implementation for adaptive selection of constraints

positive definite case covered only
untested for complex

generalized eigenvalue problem is inv(sum(inv(S_F))) = lambda inv(sum(inv(St_F)))
this is the right one for 2D problems and for faces in 3D; it is an heuristic for edges in 3D

tested with H(div) and H(curl) (3D) and works like a charm

cholesky support for matrix inversion should still be added

show more ...

PCBDDC: add slot for vertices in sub_schurs data structure.

This will also be helpful for adaptive selection

PCBDDC; added slot for threshold when adaptively computing constraints