more windows fixesHg-commit: 5c2f9a2740763d845002a2a4d856be2ee0182a65
commit after mergeHg-commit: 023d6c6af2b31621eff92a00274d614b53ecfb4b
changed sqrt() to PetscSqrtReal() to work properly with __float128changed some %G to use %14.12e and (double) to work with __float128Hg-commit: a8faa4e64896cf76f3f64094484ef5e06761417b
bugfix of MatView() for factored seqaij and seqsbaij matricesHg-commit: b2a05b2919dfa8cfa9deeaf89b795711aa2a0691
bugfix of MatGetDiagonal_SeqSBAIJ() for cholesky factored matrix (see [petsc-users] MatGetDiagonal)Hg-commit: a19846efa232a041ff3b723c0e4b68f1f36fb5b7
Use PetscBLASInt for argument to BLASHg-commit: 911e554eb253ac3f69d560f4f4cf5ddc379e8328
Remove some warnings due to -Wunused-but-set-variable (part of -Wall)Hg-commit: c822e1971e7b63bec2bc56d447818b9fc0eb6a7a
MatAXPY() was incorrect for BAIJ and SBAIJ with block size greater than 1Hg-commit: 581c78c5887744d26043b7987a4b6c48b8ede761
added matrix object to MatCheckPivot routines to allow displaying more information on failurePCEXOTIC turns off DM object in PCMG so that PCSetUp_MG() doesn't use the DM object to generate incorrect
added matrix object to MatCheckPivot routines to allow displaying more information on failurePCEXOTIC turns off DM object in PCMG so that PCSetUp_MG() doesn't use the DM object to generate incorrect work vectorsSNESSetDM,KSPSetDM and PCSetDM allow using 0 dm to clear current DM.Hg-commit: d3ce2e5ec6f3ec4502b397e3155085dddbdcd62e
show more ...
fix compile warningHg-commit: edf28cf744303287f4f6f051ac9a2cacc977018b
code cleanupHg-commit: 9991dd9df66aa85697ef434d5378ca43333e929a
bug fixesHg-commit: a2cdd4bed691425566464b7641d858f37743937f
various MatDuplicate_XXX() that previously used PetscLayoutCopy() now use PetscLayoutReference() to share the same layout.Hg-commit: fd9a32ce98293ea5f22d7487fd95e4b6f875cb6c
Fixes for insertion error options.Option should be set if "real" preallocation information wasprovided. If PETSC_DECIDE was used, then do not error. The option isinherited through a DisAssemble_*
Fixes for insertion error options.Option should be set if "real" preallocation information wasprovided. If PETSC_DECIDE was used, then do not error. The option isinherited through a DisAssemble_* operation.Hg-commit: e675ef955f926f7569d867832ae3d74f566041a1
changed docs about preallocating diagonal entries of matrix to only when planning to factorHg-commit: 62debd9818aa7a0dec4c1125f964afa7ec1bc5c4
mat_view routines intentionaly set low preallocation - so also disable MAT_NEW_NONZERO_LOCATION_ERRHg-commit: edc5daef6e31f57103f93b02b118ea3971db190f
Preallocation routines set MAT_NEW_NONZERO_ALLOCATION_ERRHg-commit: 631dbd3be1c5a3778cfbb484eb5cec7d7e4571a3
Mass revert file modes that were made executable in f6e61ee9c329Hg-commit: 116d90d0b66125cee0db3e7351047057e756baf5
changed MatMatMult to MatTranposeMatMult.Hg-commit: f6e61ee9c329774cd5343eea1ee6b7432be17c3f
Add InsertMode argument to PetscTableAdd()Hg-commit: 2cec7fc947c83dd32154ec42fd5ca5e989a3ecc5
Documentation fixes for options for specific matrix solversHg-commit: 100174d3da01faeba57dc2a3a183284bcb1b0654
Mat preallocation: check valid type and logically collective block sizeHg-commit: 14994399c05663553b5763e55847cd01074430ab
reorganized SNESVI to two solvers SNESVIRS and SNESVISS in seperate directories without being intertwined though they share some common coderemoved rsaug from being built since is brokenHg-commit
reorganized SNESVI to two solvers SNESVIRS and SNESVISS in seperate directories without being intertwined though they share some common coderemoved rsaug from being built since is brokenHg-commit: 31c8ce4101ddfbfe4fc6d005d66ef14f58ee082a
fixed a few PetscOptionsGetInt(PETSC_NULL, to use object prefixHg-commit: 18077e155b608bf1c6379b862b95c69ac5c98077
bugfix for MatIncreaseOverlap_MPISBAIJ()Hg-commit: b6f7404c321fc742b796808d1e54b77a6175d5e4
1...<<21222324252627282930>>...69