rm 'statement not reached' -- reported by PETSc blame digest

Merge branch 'tisaac/mat-get-submatrix-unordered'

Merge branch 'barry/fix-matload-uneven-rows'

Conflicts:
src/mat/examples/tests/makefile

MatLoad_MPIXAIJ() did not work correctly with user provided decomposition of the matrix rows
also added test example for this case

Optimized MatShift() for a completely empty matrix since this is commonly used with SLEPc to produce diagonal matrices

Reported-by: Jose Roman <jroman@dsic.upv.es>

completed basic use of MatSetValuesBlocked_SeqBAIJ_Inlined() by MatSetValuesBlocked_MPIBAIJ() to
allow proper global indices being printed when error generated by new nonzero locations.

Further opti

completed basic use of MatSetValuesBlocked_SeqBAIJ_Inlined() by MatSetValuesBlocked_MPIBAIJ() to
allow proper global indices being printed when error generated by new nonzero locations.

Further optimizations of MatSetValuesBlocked_MPIBAIJ() are now possible and should be implemented in the future!

show more ...

Added inline version of MatSetValuesBlocked_SeqBAIJ() to be used by MatSetValuesBlocked_MPIBAIJ() to print
error message for nonzero block with global indices

in error messages in MatSetValuesBlocked_Seq(S)BAIJ clarify the indices are block not point

clarify in error messages for MPI Mat block set values that the printed indices are block based, not point based

When generating error new nonzero location in MatSetValues_MPIxxx() calls print the global row/column number of the offending location
Previously this printed the number inside the local sequential p

When generating error new nonzero location in MatSetValues_MPIxxx() calls print the global row/column number of the offending location
Previously this printed the number inside the local sequential portion of the matrix which is confusing for users

Note that this cannot be done trivially for MatSetValuesBlocked_MPIxxx() because the insertion in the submatrix is in a seperate function that does not know about the calling function

show more ...

removed preallocation from MatConvert_Basic because it was totally wrong and caused crashes; slow code is better than buggy code!

update petsc4py to latest version since previous version did not work with new PETSc install directory structure
MatConvert_SeqAIJ_SeqBAIJ does not work with block size > 1 so error out instead of co

update petsc4py to latest version since previous version did not work with new PETSc install directory structure
MatConvert_SeqAIJ_SeqBAIJ does not work with block size > 1 so error out instead of confusing user with incorrect matrix

show more ...

MatGetSubMatrix_SeqBAIJ(): fix bad counting

Fix MatGetSubMatrix_SeqBAIJ()

Remove ISSorted() check, correctly compute compressed indices

MatGetSubMatrix(): to allow unsorted index sets in XAIJ formats.

I think the added generality of allowing index sets to be unsorted is
worth the cost of sorting matrix rows in the created submatrice

MatGetSubMatrix(): to allow unsorted index sets in XAIJ formats.

I think the added generality of allowing index sets to be unsorted is
worth the cost of sorting matrix rows in the created submatrices.

show more ...

MatLoad_ and VecLoad_ routines that use options database information from the binary file .info files need to PetscViewerSetUp() before accessing the options database because the "lazy" setup of bina

MatLoad_ and VecLoad_ routines that use options database information from the binary file .info files need to PetscViewerSetUp() before accessing the options database because the "lazy" setup of binary viewers defered the opening of the files until binary reading was first done which was after the calls to determine matrix and vector block sizes.

show more ...

MatAssembly: move check for InsertMode consistency into MatStashScatter impls

The BTS implementation now encodes InsertMode in-band to avoid another
set of messages for MAT_SUBSET_OFF_PROC_ENTRIES.

MatAssembly: move check for InsertMode consistency into MatStashScatter impls

The BTS implementation now encodes InsertMode in-band to avoid another
set of messages for MAT_SUBSET_OFF_PROC_ENTRIES. It does the reduction
in debug mode for better error reporting.

show more ...

PetscViewerBinaryGetDescriptor() is now a collective and must be treated as such
fixed some calls to PetscViewerBinarySetUseMPIIO()
removed some test runs that take too long

MatStash BTS: add MAT_SUBSET_OFF_PROC_ENTRIES, impl with neighbor-only comm

Add MatOption to imprint a communication pattern on the first assembly
and reuse on subsequent assemblies. Messages of si

MatStash BTS: add MAT_SUBSET_OFF_PROC_ENTRIES, impl with neighbor-only comm

Add MatOption to imprint a communication pattern on the first assembly
and reuse on subsequent assemblies. Messages of size 0 are sent if no
values need to be communicated to that rank. Since there are no
collectives (even Ibarrier) and no separate headers, this should be more
efficient for large process counts. The BTS implementation must
currently be activated via the -matstash_bts option.

show more ...

Merge branch 'maint'

Conflicts:
src/mat/impls/baij/seq/baij.h
src/mat/impls/sbaij/mpi/mpisbaij.c
src/mat/impls/sbaij/seq/sbaij.h

rm white spaces

baij: move 'seqbaij' prototypes required by mpibaij.c to baij.h
This fixes link failure with VS2005 cl

libpetsc.lib(mpibaij.o) : error LNK2019: unresolved external symbol "int __cdecl MatRestoreRow_

baij: move 'seqbaij' prototypes required by mpibaij.c to baij.h
This fixes link failure with VS2005 cl

libpetsc.lib(mpibaij.o) : error LNK2019: unresolved external symbol "int __cdecl MatRestoreRow_SeqBAIJ(struct _p_Mat *,int,int *,int * * const,class std::complex<double> * * const)" (?MatRestoreRow_SeqBAIJ@@YAHPEAU_p_Mat@@HPEAHQEAPEAHQEAPEAV?$complex@N@std@@@Z) referenced in function MatCreateMPIMatConcatenateSeqMat_MPIBAIJ

show more ...

add new MatGetSubMatrix_MPISBAIJ (see issue #7)

Added VecLockPush() and VecLockPop() to make vector read only
Used VecLockPush() and VecLockPop() in a couple of key locations; PCApply(), KSPSolve(), MatMult()
Fixed many uses of VecGetArray() to Ve

Added VecLockPush() and VecLockPop() to make vector read only
Used VecLockPush() and VecLockPop() in a couple of key locations; PCApply(), KSPSolve(), MatMult()
Fixed many uses of VecGetArray() to VecGetArrayRead()
Added VecGet/RestoreArrayPair() for commonly occuring pattern
Added const to many locations driven by use of VecGetArrayRead()

show more ...

Merge branch 'maint'