PCBDDC; added slot for threshold when adaptively computing constraints

PCBDDC; clean up of sub_schurs/deluxe code

* options database are now part of sub_schurs setup
* eliminated unneeded field from sub_schurs structure
* added 2 interface functions to init and setup s

PCBDDC; clean up of sub_schurs/deluxe code

* options database are now part of sub_schurs setup
* eliminated unneeded field from sub_schurs structure
* added 2 interface functions to init and setup sub_schurs when setupping the preconditioner
* if a change of basis is requested, releveant matrices are projected
* still need to be tested when the BDDC is reused

show more ...

PCBDDC: compute sum of Stildas

PCBDDC; added Stildas to sub_schurs data structure

still don't know if they should belong to the structure or just their sums

simplify option object names

PCBDDC: added initial support for computation of Stildas

the computation is very intensive, need to find an alternative

PCBDDC: added private function to explicitly compute local schur complements

Fix signature of VecGetLocalVector and related functions.

The second argument should be 'Vec w' rather than 'Vec *w' because these
functions expect a fully constructed vector.

PCBDDC: removed old SubSchursSetUp code

PCBDDC: added support for explicit computation of Schur complements using MUMPS in sub_schurs code

PCBDDC: add function to extract sequential matrices from parallel or sequential matrices using unordered index sets

This is just to allow code reuse inside PCBDDC

added --with-threadsafety flag
other cleanup for threadsafety
removed Euclid due to bit rot

fixed squaring graph on all levels in GAMG

Merge branch 'master' of bitbucket.org:petsc/petsc into mark/gamg-crs-ksp2

PCBDDC: store principal minors of local Schur complements by columns

This shoudl facilitate the computation of adaptive constraints

PCBDDC: remove unneeeded index sets

PCBDDC: small fixes

PCBDDC: move computation of explicit schur complements from deluxe code to sub_schurs setup

PCBDDC: make deluxe->seq_mat and sum of schur complements members of sub_schurs

updated new XXXSetFromOptions_YYY() methods to take PetscOptionsObjectType

fixed hypre code to work with non-global object version of PetscOptionsBegin()

moved PetscOptionsObject from global variable to local variable in functions

manual merge of conflict

PCBDDC; remove unneeded copies

PCBDDC: use MatISGetMPIXAIJ to compute sum of schur complements