| 831a100d | 08-May-2012 |
Stefano Zampini <stefano.zampini@gmail.com> |
Added slot for subdomain scaling factor to pcis.h and changed creation of pcis->D in pcis.c accordingly.
Added API for setting the scaling factor.
Hg-commit: 73ed1a76f4247024fca5216f2d34881084d97e75 |
| d3ee2243 | 08-May-2012 |
Stefano Zampini <stefano.zampini@gmail.com> |
Added explicit call to MatSetBlockSize in MatCreateIS. Removed MatGetVecs_IS no longer needed.
API changes documented.
Hg-commit: 93e67397a06693ce50a3a7dee111d0d37c5f68fe |
| 4e4c7dbe | 07-May-2012 |
Stefano Zampini <stefano.zampini@gmail.com> |
Added block size to MatCreateIS and MatGetVecs_IS function. MATIS code modified accounting for changes.
Hg-commit: a942f19a37c94cb8bc154e61416dbd65cd16b4ab |
| 09d011a0 | 07-May-2012 |
Dmitry Karpeev <karpeev@mcs.anl.gov> |
Small bug fixes to ASM and GASM.
Hg-commit: d96368c4e2da6e6dccd36431051802c12e599c2c |
| 643535ae | 07-May-2012 |
Dmitry Karpeev <karpeev@mcs.anl.gov> |
Changed the printing of ASM subdomains to prevent output interleaving, but potentially making it less scalable.
Hg-commit: f6b8a51669231d79fe217abf5197884ee9c6f073 |
| 06b43e7e | 07-May-2012 |
Dmitry Karpeev <karpeev@mcs.anl.gov> |
Reorganized GASM API and improved PCView_GASM
Hg-commit: 244afa5021e891eb860fb8a64e4f7071e380677e |
| 08ea439d | 12-Jun-2012 |
Mark F. Adams <mark.adams@columbia.edu> |
added support for matload_block_size in MPIAIJ.
Hg-commit: 2af3aa094c797e8844e3e8850d0dd040109f336c |
| 5c0d651d | 09-Jun-2012 |
Jed Brown <jed@59A2.org> |
fix fieldsplit indexing bug on reset
Hg-commit: 769fda0d533de7c6e032237e398671fef0a09e90 |
| 7287d2a3 | 07-Jun-2012 |
Dmitry Karpeev <karpeev@mcs.anl.gov> |
FieldSplit should forward DMs and prefixes to Schur solvers, use correct tab levels on their PCs, and handle reset correctly.
Hg-commit: a36eb42b26eeec4cdd64c76a892badd1985fede2 |
| a35b7b57 | 06-Jun-2012 |
Dmitry Karpeev <karpeev@mcs.anl.gov> |
PC(G)ASM should gracefully handle the case of empty outer subdomain list, even when inner subdomains are present.
Hg-commit: 57647567ccf204fe4096bde8e981160b5bd0a7e9 |
| 0f188ba9 | 05-May-2012 |
Jed Brown <jed@59A2.org> |
Doc fix for -pc_fieldsplit_schur_precondition <self,user,diag>, repair some formatting
Hg-commit: 29b351f18c2eb014490c1350a75b2d55b7f22a4c |
| ed1f0337 | 05-May-2012 |
Matthew G Knepley <knepley@gmail.com> |
FIELDSPLIT now looks for near null spaces in the IS defining the split
DMMesh constructFieldSplit() now takes near null spaces and sets then in the IS
Hg-commit: 2ad289ac99e0db0a970c4d51fcd91b9580e7
FIELDSPLIT now looks for near null spaces in the IS defining the split
DMMesh constructFieldSplit() now takes near null spaces and sets then in the IS
Hg-commit: 2ad289ac99e0db0a970c4d51fcd91b9580e7b934
show more ...
|
| 2b837212 | 04-May-2012 |
Dmitry Karpeev <karpeev@mcs.anl.gov> |
Bug fix.
Hg-commit: 2384b51d81cfde4b15e880f83ed2f537808d8b1c |
| 69344418 | 04-May-2012 |
Mark F. Adams <mark.adams@columbia.edu> |
cleaned up PCSetCoordinates for GAMG-AGG.
Replaced GAMG tests ex54-56 test solutions.
Hg-commit: fb94665f494d979e8a3bbaece57e906613241315 |
| 50456592 | 03-May-2012 |
Barry Smith <bsmith@mcs.anl.gov> |
got openmp threadcomm to build shared library on apple
Hg-commit: d994d152eebc0d18819ba525d7354d58f617fcb1 |
| eec35531 | 02-May-2012 |
Matthew G Knepley <knepley@gmail.com> |
Make a new logging stage for PCApply_MG if -pc_mg_log is given
Hg-commit: f18fee1824cc7f3036264d3ed0554874d7b6622d |
| 9d1b15c3 | 02-May-2012 |
Mark F. Adams <mark.adams@columbia.edu> |
removed old code for column block sizes in GAMG.
Hg-commit: 4ca60e8beaf3c684480d127e797f44d7b491ceb3 |
| a9db87a2 | 01-May-2012 |
Matthew G Knepley <knepley@gmail.com> |
Added an optional stage for PCApply_MG since I cannot understand the MatMult() numbers without it
Hg-commit: ba95c49b8ac6efdea07e579e6a9f3ed9e26ca556 |
| a2f3521d | 01-May-2012 |
Mark F. Adams <mark.adams@columbia.edu> |
Added block size inheritence in Mat methods.
Dissable a block IS optimization in VecScatterCreate that was wrong with blocks from Mat mult methods.
Added some debug code for a user.
Refactored GAM
Added block size inheritence in Mat methods.
Dissable a block IS optimization in VecScatterCreate that was wrong with blocks from Mat mult methods.
Added some debug code for a user.
Refactored GAMG for KKT solves.
N.B., there are a lot of changes here, sorry to dump this just before a release ...
Hg-commit: a85625d104df12863d7d1735f9fdbe9b281ef7d2
show more ...
|
| 9583d628 | 28-Apr-2012 |
Barry Smith <bsmith@mcs.anl.gov> |
norm is a real number not a complex
Hg-commit: 1d36d068558494b2cba8ae03cc58e8ed8299fa27 |
| bdddcaaa | 27-Apr-2012 |
Matthew G Knepley <knepley@gmail.com> |
Fixed FieldSplit to detect a constant null space in the blocks, now ex62 works fully
- Normalized null space vector
- Pushed down S^{-1} tolerance for ex62
Hg-commit: 7d9b5a448dfeae0624c1566357e15f6
Fixed FieldSplit to detect a constant null space in the blocks, now ex62 works fully
- Normalized null space vector
- Pushed down S^{-1} tolerance for ex62
Hg-commit: 7d9b5a448dfeae0624c1566357e15f674f896fd6
show more ...
|
| 01bc414f | 26-Apr-2012 |
Dmitry Karpeev <karpeev@mcs.anl.gov> |
Initial design of DMAKKT: algegraic coarsening of KKT systems based on GAMG.
Hg-commit: 7a6b4668cbe739ac780e3e1d115967b4ccbdcf05 |
| 251f4c67 | 27-Apr-2012 |
Dmitry Karpeev <karpeev@mcs.anl.gov> |
PetscTypeCompare -> PetscObjectTypeCompare
Hg-commit: 3fc8f665421ddb862bf4d1ba95505378246340be |
| ae33e76c | 23-Apr-2012 |
Satish Balay <balay@mcs.anl.gov> |
OBJSC is used by both .c and .cu build targets [fix by using OBJSCU for cuda objects]
Hg-commit: 66e850cc6cac3abe40ce8e1a2faabe7560da6fe9 |
| a95133b1 | 17-Apr-2012 |
Barry Smith <bsmith@mcs.anl.gov> |
added use of BLASDotu() where appropriate, fixes for complex where original author didn't handle it. Fixed wrong use of bblas in maxpy for threads
Hg-commit: 9cf34ccedfe2cc536bab398b0abcbaa633e5a615 |