moved the const out of XXType definitions and into the Set and Get Methods

Hg-commit: d4d9a237045a151847ef3121b00df73d952d0999

replace c++ style comments with c style

Hg-commit: 409221fc4b0988054f4fa7c997094d26ce9bdb50

initial conversion of Mat derived solver types like Mumps etc to use MatGetFactor() instead
most code does not yet run but all code should be using the new format.

Hg-commit: ac83415153993ee58dc4

initial conversion of Mat derived solver types like Mumps etc to use MatGetFactor() instead
most code does not yet run but all code should be using the new format.

Hg-commit: ac83415153993ee58dc4194c173418bfbd73bc73

show more ...

commit after merge

Hg-commit: 0abb3c3b0f4a77d6d1d68b240a2ff94021eabd3f

removed no longer needed converts of Fortran to C MPI_Comm (since MPI 1.1 has a standard for this)

Hg-commit: 263aca0fcfe7efcaed30a4abb46e648de3bb24ab

Merge

Hg-commit: fa3e6714ee07a74a50771a91df7da838e348b8a6

Fix for bug in binary viewer for MPIDENSE matrix when output in parallel with size > 2
- We were not resetting the pointer to memory to the base

Hg-commit: b87e4e992f1e70f247f49df4cc12dd55c1682df6

error message when binary viewing mpidense matrix but have not set PETSC_VIEWER_BINARY_NATIVE
docs for MatShellSetOperation() say must return 0 on return

Hg-commit: 8699883c0d648218c56a47cda79ff5898

error message when binary viewing mpidense matrix but have not set PETSC_VIEWER_BINARY_NATIVE
docs for MatShellSetOperation() say must return 0 on return

Hg-commit: 8699883c0d648218c56a47cda79ff5898d508418

show more ...

completed rough tutorial

Hg-commit: 4863b347d5336a5880eb54981a37d446e1ce18a9

doc fixes

Hg-commit: bfa2f409adde725d41f6e90dc43db1923443b052

initialize idx in the interface MatGetRowMaxAbs() instead of MatGetRowMaxAbs_xxx()

Hg-commit: 97c7a21b689e8f747918fb86b44687c6e389e8b9

stopped work on MatMatMult_MPIDense_MPIDense because of aparent bugs in PLAPACK.

Hg-commit: c809f2c48c997f9790c9f01b439406d85cc28d6a

fixed bugs where proper wrapping of #if defined(PETSC_HAVE_PLAPACK) was not done

Hg-commit: 5026de2f7da910abf735b42e337331d39f1c84e7

again I was checking *matout vs A when reuse was MAT_INITIAL_MATRIX so *matout
was not initialized and valgrind chocked

Hg-commit: a294f51766ee1fb1e137fa155a98da00d66ab152

fixed illegal access to *matout in MatTranspose_ routines when MAT_INITIAL_MATRIX
had to remove check that one use A, MAT_INITIAL_MATRIX, &A because cannot check &A since may not be set

Hg-commit:

fixed illegal access to *matout in MatTranspose_ routines when MAT_INITIAL_MATRIX
had to remove check that one use A, MAT_INITIAL_MATRIX, &A because cannot check &A since may not be set

Hg-commit: bdfca46607ae0c6b051c05f93f67da04e9fc248d

show more ...

more work on matmatmult_mpidense_mpidense, still broken

Hg-commit: 4f6e126896f77d306a6ad982267d65ff4ac63717

fixed MatMatMultTransposeNumeric_SeqDense_SeqDense() to use correct calling
sequence of dgemm() when 'T' is the first argument.

Hg-commit: a95fc948f53f6aafba0a33632da3e103e2359ed8

moved copy from PETSc mpidense to plapack array now handled in seperate routine
moved creation of plapack array into its own routine

Hg-commit: 854828b0ebe7da46975c3953775a8adccb6431e5

further updates to pull plapack setup out of mpidense code

Hg-commit: 38c7bf68622f2e77d3dd5b6ca952c988e8916f78

removed comm_2d and nprows/npcols from lu object in MPIDense

Hg-commit: 2714c409d8d43554e1f1514a242cda4f00155c84

moved PLA_Init() and Finalize() out of individual mpidense code into MatInitializePackage() and in PetscFinalize()

Hg-commit: ea67190658453d8ff8814eb61be9de7b9706d883

added error checking to all PLA_ codes
added _isnan and _finite support for Microsoft compilers for PetscIsInfOrNan()

Hg-commit: 8b8cd2980f212d95aec23ec7a89ea1b9d33d6dd1

moved creation of plapack matrix into symbolic code from numeric code

Hg-commit: 9e056e25d3fc729aa7dde686668ab8f87440047c

reuse plapack matrix for repeated factorizations

Hg-commit: c45c3f360aa355b2fd3c3216dcbb28d8485264e9

compile fixes

Hg-commit: 34fc387cabc19db41ecb0237fe8ff938cc572822