Add functions for deallocating aijcusparse data structures.

bugfix for MatGetRedundantMatrix() with user provided subcomm

Merge branch 'dmeiser/pullrequest-fix-cusp-bjacobi2' (pull request #178)

* dmeiser/pullrequest-fix-cusp-bjacobi2:
small bug fixes and typos,
test cases that exercise the bjacobi preconditioner w

Merge branch 'dmeiser/pullrequest-fix-cusp-bjacobi2' (pull request #178)

* dmeiser/pullrequest-fix-cusp-bjacobi2:
small bug fixes and typos,
test cases that exercise the bjacobi preconditioner with cusp and cusparse solvers,
removal of cudaMemcpyAsync (and the cudaHostRegister/cudaHostUnregister),
and replacement of VecGetArray/VecSetArray calls in PCApply_BJacobi_Singleblock
with a new vector method VecGetLocalVector

Conflicts:
include/petsc-private/vecimpl.h
src/vec/vec/impls/seq/seqcusp/veccusp.cu
src/vec/vec/utils/vpscat.h

show more ...

rm psubcomm from Mat_Redundant

Use current value for default in PetscOptionsEnum().

In options handling for matrices of type:
seqcusp
seqcusparse
mpicusp
mpicusparse

Whitespace.

update use of PetscOptionsBool() to not rely on default setting automatically

Merge branch 'maint'

cleanup of use of PetscOptionsInt and friends including use of flag final argument

Silent valgrind warning complaining about uninitialized memory

reduce dependancy on psubcomm for MatGetRedundantMatrix_MPIAIJ()

rename MatCreateMPIAIJConcatenateSeqAIJ() to MatCreateMPIMatConcatenateSeqMat()

mv struct 'Mat_Redundant' to '_p_Mat’ for simplify codes

add MatGetRedundantMatrix_MPIBAIJ() -- almost same as MatGetRedundantMatrix_MPIAIJ()

add MatCreateMPIBAIJConcatenateSeqBAIJ() to be used by MatGetRedundant for sbaij format; bs>1 is not tested yet

Merge branch 'barry/docs-matblocksize'

Merge branch 'maint'

Conflicts:
config/BuildSystem/config/packages/Chaco.py

rm warnings for uninitialized *_oth

MatGetVecs() and KSPGetVecs() replaced with MatCreateVecs() and KSPCreateVecs()
since they are creators not getters

fix build warning as reported from the nightly tests

make name of MatMatMultNumeric_MPIAIJ_MPIAIJ_xxx consistent with MatMatMultSymbolic

same fix for MatMatMult_xxx_nonscalable()

when np=1, A_loc = A

bugfix for MatMatMut_MPIAIJ_MPIAIJ when np=1 by pass parts of the code that for np>1
See email from James Balasalle.

add MatConvert_MPIAIJ_Elemental()