| #
dd905efc
|
| 21-Nov-2011 |
Barry Smith <bsmith@mcs.anl.gov> |
converted ex8.c to use SNESSetDM() instead of DMMGSetSNES(); example didn't work before (lots of valgrind errors and still doesn't work, same valgrind errors).
Hg-commit: 4ed09d944203cb936fae669d0cb
converted ex8.c to use SNESSetDM() instead of DMMGSetSNES(); example didn't work before (lots of valgrind errors and still doesn't work, same valgrind errors).
Hg-commit: 4ed09d944203cb936fae669d0cbeb23cf761e204
show more ...
|
| #
d7465e05
|
| 16-Nov-2011 |
Matthew G Knepley <knepley@gmail.com> |
Merge
Hg-commit: b62c100e29b153a87bb3f834166c8f7035f06ea4
|
| #
7f15d9d9
|
| 15-Nov-2011 |
Barry Smith <bsmith@mcs.anl.gov> |
commit after merge
Hg-commit: 623f52c9b1ca5412d9a0d84c7827c8daba914fd2
|
| #
53565b12
|
| 15-Nov-2011 |
Hong Zhang <hzhang@mcs.anl.gov> |
replace container Mat_PtAP with a special struct inside Mat_SeqAIJ;
use three subroutines for different implementations of MatPtAPNumeric_SeqAIJ_SeqAIJ()
for trade of between speed and memory
Hg-com
replace container Mat_PtAP with a special struct inside Mat_SeqAIJ;
use three subroutines for different implementations of MatPtAPNumeric_SeqAIJ_SeqAIJ()
for trade of between speed and memory
Hg-commit: e1dcb1718da62b668ab4e1d5fdea181947a15dc1
show more ...
|
| #
8cdbd757
|
| 15-Nov-2011 |
Hong Zhang <hzhang@mcs.anl.gov> |
set C->ops->matmult in symbolic MatMatMult to skip dispatch in the case of MAT_REUSE_MATRIX
Hg-commit: dde2f4286cc0274d1a97fbab8c2e7bd50c541a10
|
| #
9ce9f596
|
| 15-Nov-2011 |
Barry Smith <bsmith@mcs.anl.gov> |
commit after merge
Hg-commit: a898da55d72cb1d30e30ffc431e7663fefe9e720
|
| #
f541dd11
|
| 14-Nov-2011 |
Hong Zhang <hzhang@mcs.anl.gov> |
rm unused ierr
Hg-commit: 9456d6549b96fc236cda19578decc74a0fb55333
|
| #
e727c939
|
| 14-Nov-2011 |
Jed Brown <jed@59A2.org> |
Renamed DMGetInterpolation(), DMGetInjection(), and DMGetColoring() to DMCreateInterpolation(), etc for semantic consistency
Hg-commit: a45bcab2b329db85bdff728f3e039f094e2e00b9
|
| #
950540a4
|
| 14-Nov-2011 |
Jed Brown <jed@59A2.org> |
Rename DMGetMatrix() to DMCreateMatrix()
Hg-commit: 045e13fbce4683d1c6e7208d70318a9e4dde5c0b
|
| #
7ef18aee
|
| 11-Nov-2011 |
Barry Smith <bsmith@mcs.anl.gov> |
commit after merge
Hg-commit: 48f2ef238de19667a56ba30c53370a936b489ecc
|
| #
a4252ee0
|
| 11-Nov-2011 |
Mark F. Adams <mark.adams@columbia.edu> |
same as last commmit.
Hg-commit: 865470b29b31e08bd93fec3bd48defc4e7cba401
|
| #
874288d9
|
| 11-Nov-2011 |
Matthew G Knepley <knepley@gmail.com> |
Now a MatNullSpace attached to a Mat is automatically used in a KSP
- No longer remove null space component when executing MatMult()
- Removed printing in MatNullSpaceTest() without -mat_null_space_t
Now a MatNullSpace attached to a Mat is automatically used in a KSP
- No longer remove null space component when executing MatMult()
- Removed printing in MatNullSpaceTest() without -mat_null_space_test_view
- Added MatGetNullSpace()
Hg-commit: eb4fe8090a4ae24f4e1cb8c7e010ba1cfe573bf7
show more ...
|
| #
b6a51c36
|
| 07-Nov-2011 |
Peter Brune <brune@mcs.anl.gov> |
commit after merge
Hg-commit: 4afddb78e4d16b99952e891a10244a92ea0015ab
|
| #
b8063d65
|
| 04-May-2011 |
Barry Smith <bsmith@mcs.anl.gov> |
XXXDestroy() doesn't need custom Fortran interface so remove a bunch of them
Hg-commit: d47d53fcd75ee46a52fff678d2fde2e429a55e02
|
| #
71dad5bb
|
| 01-May-2011 |
Barry Smith <bsmith@mcs.anl.gov> |
make sbaij matrix ignore lower triangular values set into matrix by default instead of generating an error.
Hg-commit: e7acce0c00581d51a4c92022b307d38d0caa49ac
|
| #
828413b8
|
| 30-Apr-2011 |
Barry Smith <bsmith@mcs.anl.gov> |
updates to docs on Hermitian vs non-Hermitian symmetry
Hg-commit: fbe127677dc8c0ac83300394584bd88702ba1f0c
|
| #
11bd1e4d
|
| 28-Apr-2011 |
Lisandro Dalcin <dalcinl@gmail.com> |
API Change:[Mat] Remove iscopy,reuse from MatGetDiagonlBlock()
* Now it should return a borrowed reference.
* MATDENSE was updated to cache the the diagonal block
Hg-commit: dbddbea75f68dca2821e78
API Change:[Mat] Remove iscopy,reuse from MatGetDiagonlBlock()
* Now it should return a borrowed reference.
* MATDENSE was updated to cache the the diagonal block
Hg-commit: dbddbea75f68dca2821e7818ebd56b2b79bbbaf9
show more ...
|
| #
bf0cc555
|
| 27-Apr-2011 |
Lisandro Dalcin <dalcinl@gmail.com> |
MatDestroy(): cleanup the MatPreallocated() hack, fix all matrix types
Hg-commit: 8b12afa1931080f94d5efc78eba78580324579e8
|
| #
ee8a2423
|
| 27-Apr-2011 |
Lisandro Dalcin <dalcinl@gmail.com> |
MatIsTranspose(): error if any matrix does not support the test
Hg-commit: 3bdc748ba073dfa2a9dd7120451bdac24647f540
|
| #
7d50072e
|
| 24-Apr-2011 |
Jed Brown <jed@59A2.org> |
Better fieldsplit handling for Stokes using MatNest and -pc_fieldsplit_real_diagonal
User provides a MatNest J = [A B; C 0] with a truly empty block, then
also provides Jp = [Ap 0; 0 D] where Dp is
Better fieldsplit handling for Stokes using MatNest and -pc_fieldsplit_real_diagonal
User provides a MatNest J = [A B; C 0] with a truly empty block, then
also provides Jp = [Ap 0; 0 D] where Dp is an approximation of the Schur
complement (e.g. scaled mass matrix). We want to use the real diagonal
to form the kspschur, then the diagonal block will be used for
preconditioning.
Hg-commit: 1433b73803a93eea43be2ff844c1beba570a75a4
show more ...
|
| #
3bd96682
|
| 21-Apr-2011 |
Barry Smith <bsmith@mcs.anl.gov> |
commit after merge, some manual merges
Hg-commit: a60db41ebcbbb90ab0d02727666a69b6d3ee88a0
|
| #
51a6ce93
|
| 20-Apr-2011 |
Barry Smith <bsmith@mcs.anl.gov> |
commit after merge
Hg-commit: 7bfe1376f1432461d3e26c90cb5fbc8c082e5072
|
| #
0ce45336
|
| 20-Apr-2011 |
Hong Zhang <hzhang@mcs.anl.gov> |
copy Mat optins in MatConvert() instead of in MatConvert_xxx()
Add test for MatConverting_xxx_mpisbaij
Hg-commit: 6d8d6942c2dc7091bdac9b2eb9c5320f3965b209
|
| #
6bf464f9
|
| 18-Apr-2011 |
Barry Smith <bsmith@mcs.anl.gov> |
converted rest of source code to take pointers into destructor reoutines. make test workds
Hg-commit: 78c641225c0516099423eeea3cc66e68fd2f62f3
|
| #
9b08a598
|
| 15-Apr-2011 |
Barry Smith <bsmith@mcs.anl.gov> |
commit after terrible merge
Hg-commit: 6ee623422368621da3dca005325b05a2fcd6d6bb
|