| 74ed9c26 | 19-Nov-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
updated more files to use PetscMallocN() instead of single malloc with ugly operations
Hg-commit: 06b9e8a2056c1ddd773610bbb1de775ec9eddee0 |
| 6363de48 | 12-Nov-2009 |
Jed Brown <jed@59A2.org> |
MatSetBlockSize for SBAIJ
Hg-commit: 51786ee77a329efe5929ae68b197479011aa0015 |
| 71fd2e92 | 11-Nov-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
added check on PetscValidScalarPointer() to all the MatSetValues_() routines to check for non-double aligned
Hg-commit: 1f2e574d0828af632142de782d665f4dded95c8f |
| e37c4368 | 06-Nov-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
commit after merge
Hg-commit: dab6e8ec0fe71ad4fe0f4bf2ee374ce51b18f40d |
| 65460251 | 06-Nov-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
changed PETSC_USE_SINGLE etc to PETSC_USE_SCALAR_SINGLE to make clearer that this variable is related to defining the scalar type
Hg-commit: 091ca03e9ad7a91d59048671631a058ed1b9044d |
| 471cc821 | 04-Nov-2009 |
Hong Zhang <hzhang@mcs.anl.gov> |
enable implementation of MatConvert_MPIAIJ_MPISBAIJ(), but
not implement yet.
Hg-commit: 065c7b7b3610a870ce9c1e9da3227b798622944a |
| 547795f9 | 04-Nov-2009 |
Hong Zhang <hzhang@mcs.anl.gov> |
add support MatMult() for Hermitian matrix in mpisbaij format
add MatMultHermitianTranspose() and MatMultHermitianTransposeAdd()
Hg-commit: bd3ee2c07764767e1715e99fb8df83bd35b221a3 |
| eeffb40d | 03-Nov-2009 |
Hong Zhang <hzhang@mcs.anl.gov> |
add MatMult_SeqSBAIJ_1_Hermitian() and MatHermitianTranspose()
Hg-commit: 284b8e9149e7268648f12ec58fc3a6cb75e5468c |
| 41f059ae | 29-Oct-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
changed MatRelax() to MatSOR() to match MatSORType and PCSOR
Hg-commit: e68531df6999d872cd12e8f3ab1ea488c58078d8 |
| 26283091 | 29-Oct-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
changed PetscMap to PetscLayout
Hg-commit: b8f4632bd00d1563dcf8108fd4199169f8fd320b |
| 7a2fc3fe | 29-Oct-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
changed PetscMap to be pointer to _p_PetscMap
changed PetscMapInitialize() to PetscMapCreate() and also have it malloc the space
Hg-commit: 2d1edc265bc1b52b0bfe730b4711ac37a11744fd |
| ec1065ed | 05-Oct-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
commit after merge
Hg-commit: e4ceadb2454090be7d46b3718e609756791804d0 |
| 887ee2ca | 01-Sep-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
fixed Eisentat preconditioner for AIJ and SBAIJ to work for all omega, instead of just omega== 1
Hg-commit: d40d325f2fcbde49d908cb600d0aae362304a789 |
| a2b30743 | 20-Aug-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
added support for nonzero initial guess for Eisentat preconditioner; what UNIC needs and I forgot to put in before.
Hg-commit: c320a1da02e44812ec1e106a6b5c452116794b2e |
| 0af126e1 | 13-Aug-2009 |
Hong Zhang <hzhang@mcs.anl.gov> |
merge from release
Hg-commit: a918e316167282373d7d69800bd29ad770224d32 |
| 20f1ed55 | 13-Aug-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
slight more optimization for MatRelax_MPISBAIJ()
Hg-commit: 58e382d37ba04053e96feadcda48b8fc0ca33253 |
| effcda25 | 13-Aug-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
minor optimizations to MatRelax_MPISBAIJ()
removed global length checks on some vector operations so that one of the two vectors can be sequential
Hg-commit: 2a706c91dfbb7e840e89f07b1e42608aec372742 |
| fa22f6d0 | 12-Aug-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
commpleted parallel eisenstat for SBAIJ matrices
also fixed ksp ex10.c to support file without right hand side; use argument -rhs 0 to indicate this
Hg-commit: 74e1c15ff1669cfc4b83d5bc2c8bde8825a913
commpleted parallel eisenstat for SBAIJ matrices
also fixed ksp ex10.c to support file without right hand side; use argument -rhs 0 to indicate this
Hg-commit: 74e1c15ff1669cfc4b83d5bc2c8bde8825a913ba
show more ...
|
| 17fa8088 | 12-Aug-2009 |
Hong Zhang <hzhang@mcs.anl.gov> |
Bugfix: MatLoad_MPISBAIJ() should not require ignore_lower_triangular part
Hg-commit: c10b0cd399ba00c9d5c2871fd218e1690fc2acc0 |
| ae3dafa0 | 17-Jul-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
pull bug fix for The situation is the following: I'm calling MatSetValues with the flag ADD_VALUES and with matrix option MAT_IGNORE_ZERO_ENTRIES. I'm inserting multiple values at once, multiple colu
pull bug fix for The situation is the following: I'm calling MatSetValues with the flag ADD_VALUES and with matrix option MAT_IGNORE_ZERO_ENTRIES. I'm inserting multiple values at once, multiple columns and rows, so I provide a rank-2 matrix of values. As I'm calling this from fortran I'm also using MAT_COLUMN_ORIENTED. Now for provided rows that are not owned by the process, it jumps to mpiaij.c:394 (line numbers as in petsc-dev). On line 399, it checks for zero entries, but only checks the very first entry of the (non-owned) row. If however other entries of that same row are nonzero, the entire row is still dropped. Note that this is independent of row_oriented/column_oriented as line 396 does exactly the same.
If I don't set the option MAT_IGNORE_ZERO_ENTRIES the problem disappears. In that case however we would either have to preallocate substantially more nonzeros, or complicate the matrix assembly in our code by taking out the zero entries ourselves and call MatSetValues for each entry seperately.
Hg-commit: f09f0ab443e582d40fb2b646b2461bcc61f6e436
show more ...
|
| b400d20c | 17-Jul-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
From: s.kramer@imperial.ac.uk
Subject: non-local values being dropped in MatSetValues
Date: July 17, 2009 7:44:09 AM CDT
To: petsc-users@mcs.anl.gov
Reply-To: petsc-users@mcs.anl.gov
The si
From: s.kramer@imperial.ac.uk
Subject: non-local values being dropped in MatSetValues
Date: July 17, 2009 7:44:09 AM CDT
To: petsc-users@mcs.anl.gov
Reply-To: petsc-users@mcs.anl.gov
The situation is the following: I'm calling MatSetValues with the flag ADD_VALUES and with matrix option MAT_IGNORE_ZERO_ENTRIES. I'm inserting multiple values at once, multiple columns and rows, so I provide a rank-2 matrix of values. As I'm calling this from fortran I'm also using MAT_COLUMN_ORIENTED. Now for provided rows that are not owned by the process, it jumps to mpiaij.c:394 (line numbers as in petsc-dev). On line 399, it checks for zero entries, but only checks the very first entry of the (non-owned) row. If however other entries of that same row are nonzero, the entire row is still dropped. Note that this is independent of row_oriented/column_oriented as line 396 does exactly the same.
If I don't set the option MAT_IGNORE_ZERO_ENTRIES the problem disappears. In that case however we would either have to preallocate substantially more nonzeros, or complicate the matrix assembly in our code by taking out the zero entries ourselves and call MatSetValues for each entry seperately.
Hg-commit: d28deff5e1b8c4e99cf2d919d60e6073ac99448f
show more ...
|
| b75cf396 | 05-Jul-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
set d in MatRelax_SeqSBAIJ that it no longer needed
Hg-commit: c1444f48fd77cb2635ffc8849c7a111911162cb7 |
| a9817697 | 05-Jun-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
changed MAT_KEEP_ZEROED_ROWS to MAT_KEEP_NONZERO_PATTERN
Hg-commit: dd0eb2438b4ecde1a3517ea573f0c60a0160fd54 |
| dc2c7911 | 04-May-2009 |
Satish Balay <balay@mcs.anl.gov> |
merge from release
Hg-commit: 2b7a7d71963e7893197575adcffbd9e29a0afb26 |
| 526dfc15 | 01-May-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
petsc-maint19,395 if MatMPIXAIJSetPreallocation() was called twice, the inner matrices were created a second time leading to memory bleeding.
Fixed by allowing preallocation to be called twice and
petsc-maint19,395 if MatMPIXAIJSetPreallocation() was called twice, the inner matrices were created a second time leading to memory bleeding.
Fixed by allowing preallocation to be called twice and not recreating inner matrices after the first time.
Hg-commit: ecd2cb882bda291142224dcebc9b00d3f6cecc0e
show more ...
|