fixed some function defs to not generate Fortran stubs and
put the stubs in the correct library

Hg-commit: 4cfaf012d56ebe0f6c82c5fbcd4f2b7f03226189

Added a new option
- MAT_UNUSED_NONZERO_LOCATION_ERR indicates when unused space has been allocated

Hg-commit: d10d5d18ac9b373ad66e42b56d03063833c579c1

added MatScale() and MatMultTranspose() for MatCreateNormal()

Hg-commit: b9b626ca65b765dd89971c89c38ddf39cb7ed6e6

slightly more work on wirebasket interpolation for DA

Hg-commit: 204f4784aa40185738e1762998751adb77ac467c

added MatScale() MatDiagonalScale() to MATCOMPOSITE per petsc-maint 18500
warning: not tested.

Hg-commit: 2c125e6c8461666c920507c51bfebca66ccd9d7c

removed -I$PETSC_DIR from make system and changed the corresponding
#include "src/dsdsdsds" to "../src/sdsdsds" in the source code

Hg-commit: 6e3c4cc4b4bfcb366bae9b71afcb5009c1f0429d

From: dalcinl@gmail.com
Subject: about viewer formats
Date: November 3, 2008 7:50:10 AM CST
To: petsc-dev@mcs.anl.gov
Reply-To: petsc-dev@mcs.anl.gov
Barry, I saw you commit removing PETSC_V

From: dalcinl@gmail.com
Subject: about viewer formats
Date: November 3, 2008 7:50:10 AM CST
To: petsc-dev@mcs.anl.gov
Reply-To: petsc-dev@mcs.anl.gov
Barry, I saw you commit removing PETSC_VIEWER_BINARY_NATIVE. This
makes me think that PETSC_VIEWER_ASCII_DEFAULT and
PETSC_VIEWER_BINARY_DEFAULT (which is NOT used in sources) are a bit
redundat, Why not to remove both and add a new PETSC_VIEWER_DEFAULT?
Does it make sense?

Hg-commit: 9957a5666f6211584952873f183a5f14ec823bbe

show more ...

try to save PetscMalloc calls in MatSetValuesBlocked[Local] fallback

Hg-commit: 2e4c27ae85e48a03904def93b05fc36c8d6fcff8

added MatCompositeSetType() MatComposite can now do multiply in addition to add

Hg-commit: 4c5e06c87351afd0630cd1b00305dc0ca4344355

* Using MUMPS, MatGetInfo(...,&info), info.nz_allocated is given for the
initial matrix (not the factorized one). Currently I use :
- Mat_MUMPS *lu=(Mat_MUMPS*)mat->spptr;
- lu->id.INFOG(20)
to get

* Using MUMPS, MatGetInfo(...,&info), info.nz_allocated is given for the
initial matrix (not the factorized one). Currently I use :
- Mat_MUMPS *lu=(Mat_MUMPS*)mat->spptr;
- lu->id.INFOG(20)
to get the same information. It's a dirty solution...

Added MatFactorGetSolverPackage(), also properly printed with -mat_view_info and -ksp_view

Hg-commit: d45534d976bbd57e4a4d4c4173150bb1403ccc76

show more ...

Added MatGetRowMinAbs()

Hg-commit: 641ef60822bb7bdb01f8731c75ca4e57cbd2fc5b

pass MatFactorInfo as const into factorization routines, they shouldn't be changing it.

Hg-commit: 06e09652bd7d2f02616c7256594ec565736aabbf

added profiling for VecView/Load and MatView/Load

Hg-commit: 5b55a0869ec22b6e2043299c07c23f6c9364fd0d

ISGetIndices() returns const array, it is not something that should be changed

Hg-commit: 279c07e55dc4d24fe01a982d22cefc6e74f39257

Added arg to MatNullSpaceTest()

Hg-commit: 897ade46c5d4e4a1b7f08209c1ab7c2120348c6d

commit after merge; lots of GU conflicts needed to be resolved, why? I don't know, most looked
like they could be done automatically

Hg-commit: 2b292236ee5585a41c97ddcafd3f409331a1dd72

1) The first argument to MatXXXFactorSymbolic/Numeric() is now the factored matrix
2) The assigment of the MatXXXFactorNumeric ops pointer is now done in the symbolic factorization
the assigment o

1) The first argument to MatXXXFactorSymbolic/Numeric() is now the factored matrix
2) The assigment of the MatXXXFactorNumeric ops pointer is now done in the symbolic factorization
the assigment of the MatSolve pointer is now done in the numeric factorization
(This is because only then do you know the right pointer that should be used)

Hg-commit: efc00afbc7c5103f2f56fd96963db2c4cf4e1ed4

show more ...

rm redundant declaration

Hg-commit: e53c043cd8a9ac51bd86ecf8a39492c976009307

moved assignment of LUFactorNumeric and CholeskyFactorNumeric into appropriate symbolic factorization (previously it was set
in the function table or in the getfactor
moved assignment of solve() i

moved assignment of LUFactorNumeric and CholeskyFactorNumeric into appropriate symbolic factorization (previously it was set
in the function table or in the getfactor
moved assignment of solve() into appropriate numeric factorization, previously it was in function table or get factor or symbolic factor

Hg-commit: 83ba8e21d43889fe945455ea7c1fb95368f1a281

show more ...

fixed some memory bleeding
changed PetscMap in Vec and Mat to be an allocated shared
structure with referencing counting. Shared by Vecs obtained with
VecDuplicate() and MatGetVecs()

Hg-commit

fixed some memory bleeding
changed PetscMap in Vec and Mat to be an allocated shared
structure with referencing counting. Shared by Vecs obtained with
VecDuplicate() and MatGetVecs()

Hg-commit: 9cebf4ba1b4632fc12cb8afbc01d420ee612ad72

show more ...

removed -mf_fd_coloring_jacobian_period and -dmmg_jacobian_period and function call friends because -snes_lag_jacobian does all that is needed
modified SNESComputeJacobian() to call the MatAssemblyBe

removed -mf_fd_coloring_jacobian_period and -dmmg_jacobian_period and function call friends because -snes_lag_jacobian does all that is needed
modified SNESComputeJacobian() to call the MatAssemblyBegin/End() on matrix-free A everytime even if lag is not set to 1
this is what people would expect

Hg-commit: 296a77abbd7661b7fbc32fd263fb1abe69a4a9a5

show more ...

basic Schur complement preconditioner is now working (no preconditioner on the Schur solve yet)

Hg-commit: b92a24fa5a99cb9f0f1ebe423a22c2cab0402a12

fix typo, routine name is MatGetOwnershipRangesColumn()

Hg-commit: e351a1f54f0ab6f83cd2b30d0c90ea740ec257e6

MatGetOwnershipRangesXXX() calls preallocate if needed.

Hg-commit: c88e91d8d146b2ffeebd8466b196c6d7209b51ef

added MatGetOwnershipRangesColumn()

Hg-commit: 03eea7bc8c561f690126477bfcba7b1d4b9a8a1d