| 6abce2d9 | 27-Jul-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
DMMGSetNullSpace() did not destroy vectors creating internally that were then passed into MatNullSpaceCreate() leading to memory leak in src/snes/examples/tutorials/ex29.c
Hg-commit: 837aa9994d4211c
DMMGSetNullSpace() did not destroy vectors creating internally that were then passed into MatNullSpaceCreate() leading to memory leak in src/snes/examples/tutorials/ex29.c
Hg-commit: 837aa9994d4211cc073c44aafca32fbcbbbdede0
show more ...
|
| f6d58c54 | 25-Jul-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
finished support for MatFDColoringCreate() etc for BAIJ matrices
Hg-commit: 61c9609b5ffa140392733ce982b1260ca27cbd92 |
| 3acb8795 | 23-Jul-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
more work on adding FD computation of Jacobian for BAIJ matrices
Hg-commit: 89075e85186765069f32cddc814471a4b6a51957 |
| 5f94a25c | 21-Jul-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
fixed cut and paste bug introduced into MatGetColoring_Natural()
added MatType argument to DAGetColoring()
Hg-commit: 9fa545833a0c398df641458d95d67e8b23418eed |
| c60f0209 | 16-Jul-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
moved all the vector fortran kernels to correct location and the block invert kernels used by BAIJ and SBAIJ
Hg-commit: e1f076dca7c745ad92f3e0051127f9b15b56c9dc |
| a0c198e4 | 01-Jun-2009 |
Satish Balay <balay@mcs.anl.gov> |
merge from release
Hg-commit: 768f8db0f66e5e62ac4fcfea107b7de09dfc5daa |
| 45732ad0 | 01-Jun-2009 |
Satish Balay <balay@mcs.anl.gov> |
use correct MANSEC in makefiles as per generatefortranstubs.py [for include/fortran/ftn-auto/*.h files]
Hg-commit: b729083a7e07dd5d7724ed19736aad998359952b |
| f91d8e95 | 01-May-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
started hypre HYPRE_StructMatrix based PETSc matrix type
hypre contains now outdated struct stencil and matrix stuff (will be removed next)
Hg-commit: 09c845b0ad214d20d8d422125e3595c17ebee30a |
| 62e02987 | 27-Apr-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
commit after merge of patch from 3.0.0 for faster PetscBinaryWrite()
Hg-commit: 272378aa41244d7473de0ea3caa121579e4031f8 |
| 6e0c8459 | 27-Apr-2009 |
Satish Balay <balay@mcs.anl.gov> |
#19373 c++ compile fixes
Hg-commit: e16911e351a95349c6228121da73e2676bb8091f |
| ca4b5a82 | 24-Apr-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
commit after pull of two patches from 3.0.0
Hg-commit: 09be8607786d079dca91b5fdb1f1e4f203cff325 |
| 16efb6df | 24-Apr-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
removed Victor's doc that DMMGSetNullSpace() did not exist for Fortran because I added the Fortran interface to petsc-3.0.0
Hg-commit: eb594657a7eee1207cc73ef1daadc22dfe4b830b |
| e157ba5d | 24-Apr-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
added Fortran version of DMMGSetNullSpace for petsc-maint 19,266
Hg-commit: 65458427c6a576df472719b2a58b7f99e0993726 |
| 5bb3c79a | 16-Apr-2009 |
Victor Eijkhout <eijkhout@tacc.utexas.edu> |
Added note: Fortran version not available
Hg-commit: c1a363508fa124a0182027f2282e3f566a0692a5 |
| 90d69ab7 | 28-Mar-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
changed PetscOptionsName() and PetscOptionsHasName() to return true if option is set, regardless of the value it is set to, include 0 and false
changed most uses of these routines to the Truth vari
changed PetscOptionsName() and PetscOptionsHasName() to return true if option is set, regardless of the value it is set to, include 0 and false
changed most uses of these routines to the Truth varient
Hg-commit: fd6366abd7eb1f3ec925faa0ecd078d5cc025eb2
show more ...
|
| dbf0e21d | 08-Feb-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
fixes to mpibaij.c found in nightly build
fixes to wrong use of CHKERRQ() throughout code
Hg-commit: 733dea1e39d18170feab68a0e7a269e3248a628a |
| d6767e86 | 30-Jan-2009 |
Satish Balay <balay@mcs.anl.gov> |
tree e0679aa08652
parent 5ebb3561e28e
author hzhang@petsc.mcs.anl.gov 1232741388 21600
committer hzhang@petsc.mcs.anl.gov 1232741388 21600
revision 13335
branch default
put back the old checks that
tree e0679aa08652
parent 5ebb3561e28e
author hzhang@petsc.mcs.anl.gov 1232741388 21600
committer hzhang@petsc.mcs.anl.gov 1232741388 21600
revision 13335
branch default
put back the old checks that was deleted in change 4f50b4403303
The checks in DMMGSetSNES() are used when pc=none,
e.g., /src/snes/examples/tutorials/ex22.c in which matformat=mffd.
Hg-commit: fa29c5396e00aa375157a3b5f801680e9602b0b4
show more ...
|
| 9b828f69 | 23-Jan-2009 |
Hong Zhang <hzhang@mcs.anl.gov> |
put back the old checks that was deleted in change 4f50b4403303
The checks in DMMGSetSNES() are used when pc=none,
e.g., /src/snes/examples/tutorials/ex22.c in which matformat=mffd.
Hg-commit: 2e04b
put back the old checks that was deleted in change 4f50b4403303
The checks in DMMGSetSNES() are used when pc=none,
e.g., /src/snes/examples/tutorials/ex22.c in which matformat=mffd.
Hg-commit: 2e04b24709cdfb14afc0b52af8ffc8138dcfd34d
show more ...
|
| cfc797a6 | 20-Jan-2009 |
Barry Smith <bsmith@mcs.anl.gov> |
DMMGSetSNES was setting the coarsest grid problem for -dmmg_grid_sequence to use LU as the coarse grid
solver, ignoring the fact that it likely doesn't exist for parallel matrices. Changed to simply
DMMGSetSNES was setting the coarsest grid problem for -dmmg_grid_sequence to use LU as the coarse grid
solver, ignoring the fact that it likely doesn't exist for parallel matrices. Changed to simply discard
mg on the coarsest grid nonlinear solver and to use redundant if parallel otherwise LU.
From: Xuefeng.Li@Colorado.EDU
Subject: [PETSC #18923] Clarification: option -dmmg_grid_sequence
Date: January 20, 2009 9:47:08 AM CST
Not perfect solution, what if mat supported parallel LU? At this point we don't even know the matrix type so
cannot check this. What we might want is have the PC try a variety of preconditioners starting with LU if
certain flag set? No infrastructure for this currently.
Hg-commit: 4f50b440330337e2324d509b1b620016deda1557
show more ...
|
| 415c4166 | 12-Dec-2008 |
Matthew Knepley <knepley@mcs.anl.gov> |
Fixed declarations
Hg-commit: deeef823b3160d312e9b42fc0b44e06fabf64a49 |
| c01a0f9a | 05-Dec-2008 |
Barry Smith <bsmith@mcs.anl.gov> |
updated mpi4py.py to use tarball for mpi4py code
unused variable in scotch
added support for DMMGSetSNESLocal() in Fortran with an example
Hg-commit: f8100c9c50da1c0788bdafde09f06e1f0882b143 |
| d6679a64 | 05-Dec-2008 |
Satish Balay <balay@mcs.anl.gov> |
do not auto-generate DMMGSetSNESLocal()
Hg-commit: 4d11ac382287b34a9d7b4f1b82a643eba85f3fff |
| c61a7d01 | 05-Dec-2008 |
Barry Smith <bsmith@mcs.anl.gov> |
removed PetscObjectPublish/AMS memory snooper from PETSc examples (someone found them
and got confused). They have not worked in what 7 years?
Hg-commit: 7f3df0204d19278a8719fbde38749c3babb959fa |
| 3f6d85c9 | 21-Nov-2008 |
Barry Smith <bsmith@mcs.anl.gov> |
updated to handle generation of fortran interface definitions if
--with-fortran-interfaces is used.
Hg-commit: 0baa789046fea4675ba0b102771a008c7623d68a |
| 7677d60d | 12-Nov-2008 |
Matthew Knepley <knepley@mcs.anl.gov> |
Fix edit
Hg-commit: b7ed6119f6f700275b02413d5b3cdac1827a1e2f |