]>
https://gitweb.dealii.org/ - dealii.git/log
Denis Davydov [Fri, 17 Feb 2017 15:09:14 +0000 (16:09 +0100)]
make GridTools::laplace_transform() internally solve for displacement by default
while there, adjust grid_transform_02 to produce a better
looking mesh.
Wolfgang Bangerth [Thu, 16 Feb 2017 06:32:36 +0000 (23:32 -0700)]
Merge pull request #3967 from MathiasMen/Vectorization_SymmetricTensor
Vectorization of SymmetricTensor
Denis Davydov [Wed, 15 Feb 2017 10:18:31 +0000 (11:18 +0100)]
Merge pull request #3978 from masterleinad/c++_symmetric
Use C++11 instead of boost in base/symmetric_tensor_36
Daniel Arndt [Wed, 15 Feb 2017 10:05:40 +0000 (11:05 +0100)]
Use C++11 instead of boost in base/symmetric_tensor_36
Martin Kronbichler [Wed, 15 Feb 2017 09:29:07 +0000 (10:29 +0100)]
Merge pull request #3977 from masterleinad/fix_fe_enriched_test
Fix fe/fe_enriched_step-36b for 64bit indices
Daniel Arndt [Wed, 15 Feb 2017 09:24:37 +0000 (10:24 +0100)]
Fix fe/fe_enriched_step-36b for 64bit indices
Martin Kronbichler [Wed, 15 Feb 2017 07:17:39 +0000 (08:17 +0100)]
Merge pull request #3976 from davydden/tests/disable_petsc_complex
disable tests which are not meant for complex-valued PETSc
Denis Davydov [Wed, 15 Feb 2017 07:14:44 +0000 (08:14 +0100)]
disable tests which are not meant for complex-valued PETSc
Mathias Mentler [Sun, 12 Feb 2017 10:00:15 +0000 (11:00 +0100)]
Removed use of constructor of Number in SymmetricTensor as this could
also call the non-existant constructor of VectorizedArray. Replaced
by call to static member of template struct defined in
vectorization.h and numbers.h
Bruno Turcksin [Tue, 14 Feb 2017 18:54:50 +0000 (13:54 -0500)]
Merge pull request #3974 from masterleinad/fix_linear_operator_10
Don't allow 64bit-indices in PreconditionAMGMueLu
Daniel Arndt [Tue, 14 Feb 2017 13:37:24 +0000 (14:37 +0100)]
Don't allow 64bit-indices in PreconditionAMGMueLu
Timo Heister [Mon, 13 Feb 2017 17:46:44 +0000 (12:46 -0500)]
Merge pull request #3970 from davydden/doc/fe_evaluation_dim
hack doxygen to show additional methods of FEEValution
Denis Davydov [Thu, 9 Feb 2017 16:56:09 +0000 (17:56 +0100)]
hack doxygen to show additional methods of FEEValution for partial specialization
Martin Kronbichler [Mon, 13 Feb 2017 11:02:40 +0000 (12:02 +0100)]
Merge pull request #3971 from davydden/fe_enriched_prolongation
FE_Enriched: add prolongation/restriction matrices
Denis Davydov [Mon, 13 Feb 2017 09:00:23 +0000 (10:00 +0100)]
FE_Enriched: add prolongation/restriction matrices
Daniel Arndt [Sun, 12 Feb 2017 23:27:36 +0000 (00:27 +0100)]
Merge pull request #3932 from sairajat/master
Fixes bug in Pattern::Selection
sairajat [Sun, 12 Feb 2017 21:23:23 +0000 (16:23 -0500)]
Fix indentation
sairajat [Sun, 12 Feb 2017 21:14:43 +0000 (16:14 -0500)]
Fix Tests
Denis Davydov [Sun, 12 Feb 2017 06:11:04 +0000 (07:11 +0100)]
Merge pull request #3966 from drwells/remove-petsc-vector-base-copy
Remove the PETSc VectorBase copy operator.
sairajat [Sun, 12 Feb 2017 05:48:33 +0000 (00:48 -0500)]
Fixed Typo and added comment in output of test
sairajat [Sun, 12 Feb 2017 04:40:53 +0000 (23:40 -0500)]
Added space at end
David Wells [Sat, 11 Feb 2017 22:35:14 +0000 (17:35 -0500)]
Make VectorBase::operator= private and undefined.
This operator was formerly implicitly defined, which provides, almost surely,
the wrong behavior since the underlying Vec will then be destroyed twice.
Since both inheriting classes (Vector and MPI::Vector) define their own
operator= overloads this operator is also not necessary.
David Wells [Sat, 11 Feb 2017 22:29:44 +0000 (17:29 -0500)]
Fix, and deprecate, a PETSc parallel vector copy constructor.
The former implementation of this copy constructor contained a subtle bug: since
VectorBase::operator= is implicitly defined, calling it simply copies (byte for
byte) the members of a VectorBase. This is erroneous because now two different
destructors will call VecDestroy on the underlying Vec.
The fix is simple: (nearly) preserve the current behavior by just deep copying
the vector.
Since one should never have an explicit instance of VectorBase (and there is no
reasonable way to copy a parallel vector and then enforce a possibly different
local_size) this constructor I also marked this constructor for deprecation: one
should use one of the concrete classes (Vector or MPI::Vector) in applications.
David Wells [Sat, 11 Feb 2017 21:31:43 +0000 (16:31 -0500)]
Reindent some sample code comments.
Denis Davydov [Sat, 11 Feb 2017 17:52:35 +0000 (18:52 +0100)]
Merge pull request #3963 from drwells/fix-block-vector-deprecation-warning
Fix a deprecation warning in user codes.
Timo Heister [Sat, 11 Feb 2017 16:42:03 +0000 (11:42 -0500)]
Merge pull request #3961 from drwells/fix-clang-warnings-2
Fix some clang warnings.
Timo Heister [Sat, 11 Feb 2017 16:40:49 +0000 (11:40 -0500)]
Merge pull request #3964 from drwells/remove-preceding-underscore
Rename an identifier that started with an underscore.
David Wells [Sat, 11 Feb 2017 14:44:14 +0000 (09:44 -0500)]
Rename an identifier that started with an underscore.
Identifiers starting with a single underscore followed by a capital letter are
reserved by the C standard.
David Wells [Sat, 11 Feb 2017 13:40:38 +0000 (08:40 -0500)]
Fix a deprecation warning in user codes.
Since supports_distributed_data is deprecated, instantiating a BlockVector
raises a deprecation warning in user codes. Fix this by simply using the type
trait replacement.
Denis Davydov [Sat, 11 Feb 2017 12:52:04 +0000 (13:52 +0100)]
Merge pull request #3962 from masterleinad/restrict_test
grid_out_per_processor_vtu_02 requires METIS
David Wells [Sat, 11 Feb 2017 12:26:59 +0000 (07:26 -0500)]
Remove some unused function arguments.
Daniel Arndt [Sat, 11 Feb 2017 12:30:02 +0000 (13:30 +0100)]
grid_out_per_processor_vtu_02 requires METIS
David Wells [Sat, 11 Feb 2017 12:26:11 +0000 (07:26 -0500)]
Silence some warnings in OpenCascade.
Clang complains that OpenCascade checks whether or not a certain undefined macro
evaluates to true. Get around this with diagnostic guards.
Denis Davydov [Sat, 11 Feb 2017 08:42:18 +0000 (09:42 +0100)]
Merge pull request #3960 from kronbichler/classify_tests_with_p4est
Mark tests that need p4est
Martin Kronbichler [Sat, 11 Feb 2017 08:32:41 +0000 (09:32 +0100)]
Mark tests that need p4est
Daniel Arndt [Fri, 10 Feb 2017 21:33:48 +0000 (22:33 +0100)]
Merge pull request #3959 from masterleinad/fix_distribute_flux_sparsity_pattern
Fix mpi/distribute_flux_sparsity_pattern
Daniel Arndt [Fri, 10 Feb 2017 20:39:26 +0000 (21:39 +0100)]
Fix mpi/distribute_flux_sparsity_pattern
Timo Heister [Fri, 10 Feb 2017 20:03:54 +0000 (15:03 -0500)]
Merge pull request #3953 from tjhei/test_checksum
replace call to md5sum in tests
Bruno Turcksin [Thu, 9 Feb 2017 22:33:32 +0000 (17:33 -0500)]
Merge pull request #3952 from tcclevenger/write_mesh_per_processor_as_vtu_fixed_bug
fixed bug in write_mesh_per_processor_as_vtu, added test
sairajat [Thu, 9 Feb 2017 17:39:10 +0000 (12:39 -0500)]
Added change to doc/news/minor
Wolfgang Bangerth [Thu, 9 Feb 2017 17:05:05 +0000 (10:05 -0700)]
Merge pull request #3955 from davydden/doc/submit_symmetric_gradient
fix documentation of FEEvaluationAccess<dim,dim,Number>::submit_symme…
Denis Davydov [Thu, 9 Feb 2017 17:00:53 +0000 (18:00 +0100)]
fix documentation of FEEvaluationAccess<dim,dim,Number>::submit_symmetric_gradient
Timo Heister [Thu, 9 Feb 2017 14:03:37 +0000 (09:03 -0500)]
replace call to md5sum in tests
tcclevenger [Thu, 9 Feb 2017 13:38:27 +0000 (08:38 -0500)]
fixed bug in write_mesh_per_processor_as_vtu, added test
Previous implementation did not account for non-distributed multilevel
hierarchy or active cells having different level_subdomain_id and
subdomain_id.
made requested changes
typo in changes. GridOut function, not GridTools
Denis Davydov [Thu, 9 Feb 2017 10:37:38 +0000 (11:37 +0100)]
Merge pull request #3947 from jppelteret/trilinos_lin_ops-enable_deal_II_solver
Enable deal.II solver use with Trilinos LinearOperator
Jean-Paul Pelteret [Wed, 8 Feb 2017 22:43:22 +0000 (23:43 +0100)]
Fix comment indentation
Jean-Paul Pelteret [Wed, 8 Feb 2017 08:04:22 +0000 (09:04 +0100)]
Enable deal.II solver use with Trilinos LinearOperator
By making a distinction between an `inverse_payload` created for
Trilinos solvers/preconditioners and all other solver/preconditioner
types, we allow a `TrilinosPayload` to be created when the
solver/preconditioner is incompatible with `Epetra_Multivector`. In
turn, it is now possible to use the deal.II solvers (e.g.
`dealii::SolverCG<TrilinosWrappers::MPI::Vector>`) with `Trilinos`
`LinearOperator`s as such a call no longer results in a compiler error.
Daniel Arndt [Wed, 8 Feb 2017 23:10:26 +0000 (00:10 +0100)]
Merge pull request #3931 from jppelteret/fix_tests_03
Refactor schur_complement operator
Jean-Paul Pelteret [Mon, 6 Feb 2017 14:43:21 +0000 (15:43 +0100)]
Test: Make lac/schur_complement_05 a bit more robust
Do not output residuals and convergence numbers. We now only check
whether we do the right thing by comparing the difference of the result
to a reference computation.
Denis Davydov [Wed, 8 Feb 2017 21:13:49 +0000 (22:13 +0100)]
Merge pull request #3951 from Rombur/step_18
Fix typo.
Bruno Turcksin [Wed, 8 Feb 2017 21:11:23 +0000 (16:11 -0500)]
Fix typo.
Matthias Maier [Wed, 8 Feb 2017 20:14:25 +0000 (14:14 -0600)]
Merge pull request #3950 from dealii/revert-3914-fix_tests_3
Revert "Tests: output variants for opencascade 6.9.1"
Denis Davydov [Wed, 8 Feb 2017 19:00:42 +0000 (20:00 +0100)]
Revert "Tests: output variants for opencascade 6.9.1"
Wolfgang Bangerth [Wed, 8 Feb 2017 18:28:07 +0000 (11:28 -0700)]
Merge pull request #3914 from tamiko/fix_tests_3
Tests: output variants for opencascade 6.9.1
Wolfgang Bangerth [Wed, 8 Feb 2017 16:47:28 +0000 (09:47 -0700)]
Merge pull request #3946 from davydden/tests/fe_enriched_step_36
use jacobi preconditioner in fe_enriched test eigenproblems
Martin Kronbichler [Wed, 8 Feb 2017 14:26:53 +0000 (15:26 +0100)]
Merge pull request #3940 from davydden/tests/parallel_block_vector_03
add extra output for parallel_block_vector_03
Martin Kronbichler [Wed, 8 Feb 2017 14:21:13 +0000 (15:21 +0100)]
Merge pull request #3949 from masterleinad/reduce_fe_abf_02
Simplify fe/abf_02
Daniel Arndt [Wed, 8 Feb 2017 12:54:10 +0000 (13:54 +0100)]
Simplify fe/abf_02
Denis Davydov [Tue, 7 Feb 2017 22:07:28 +0000 (23:07 +0100)]
tests: use jacobi preconditioner in fe_enriched test eigenproblems
Wolfgang Bangerth [Wed, 8 Feb 2017 05:55:03 +0000 (22:55 -0700)]
Merge pull request #3945 from drwells/clang-warnings
Fix some simple clang warnings.
David Wells [Wed, 8 Feb 2017 03:12:46 +0000 (22:12 -0500)]
Remove some unused variables.
This was caught by clang 3.9.
David Wells [Wed, 8 Feb 2017 03:12:11 +0000 (22:12 -0500)]
Check if a preprocessor variable is defined before use.
Clang complains about this with -Wundef on.
sairajat [Tue, 7 Feb 2017 22:23:03 +0000 (17:23 -0500)]
Merge branch 'master' of https://github.com/sairajat/dealii
sairajat [Tue, 7 Feb 2017 21:17:53 +0000 (16:17 -0500)]
Bug Fix in Patterns::Selection
Wolfgang Bangerth [Tue, 7 Feb 2017 22:15:31 +0000 (15:15 -0700)]
Merge pull request #3936 from davydden/tests/memory_consumption_01
tests: extra output for memory_consumption_01
Wolfgang Bangerth [Tue, 7 Feb 2017 22:14:22 +0000 (15:14 -0700)]
Merge pull request #3938 from davydden/tests/oce_step_create_write
extra output for OCE 0.18
Wolfgang Bangerth [Tue, 7 Feb 2017 22:06:10 +0000 (15:06 -0700)]
Merge pull request #3943 from davydden/tests/no_flux_constraints_extra
extra output for no_flux_constraints when md5sum is not installed
Timo Heister [Tue, 7 Feb 2017 21:41:41 +0000 (16:41 -0500)]
Merge pull request #3941 from davydden/tests/fix_slepc_solve_04
fix slepc/solve_04
sairajat [Tue, 7 Feb 2017 21:17:53 +0000 (16:17 -0500)]
Bug Fix in Patterns::Selection
sairajat [Tue, 7 Feb 2017 21:14:58 +0000 (16:14 -0500)]
Fixed bug in Patterns::Selection
Denis Davydov [Tue, 7 Feb 2017 21:11:16 +0000 (22:11 +0100)]
tests: extra output for no_flux_constraints when md5sum is not installed
Denis Davydov [Tue, 7 Feb 2017 20:57:00 +0000 (21:57 +0100)]
tests: fix slepc/solve_04
Wolfgang Bangerth [Tue, 7 Feb 2017 20:50:14 +0000 (13:50 -0700)]
Merge pull request #3937 from davydden/tests/metis_extra_output
extra output for METIS tests
Denis Davydov [Tue, 7 Feb 2017 20:27:59 +0000 (21:27 +0100)]
tests: add extra output for parallel_blocK_vector_03
Denis Davydov [Tue, 7 Feb 2017 20:08:59 +0000 (21:08 +0100)]
tests: extra output for OCE 0.18
Denis Davydov [Tue, 7 Feb 2017 20:02:05 +0000 (21:02 +0100)]
tests: extra output for METIS tests
Denis Davydov [Tue, 7 Feb 2017 19:50:36 +0000 (20:50 +0100)]
tests: extra output for memory_consumption_01
Wolfgang Bangerth [Tue, 7 Feb 2017 17:57:38 +0000 (10:57 -0700)]
Merge pull request #3934 from tjhei/warn_unref
remove unused variable names
Timo Heister [Tue, 7 Feb 2017 16:24:49 +0000 (11:24 -0500)]
remove unused variable names
MSVC complains about this...
Martin Kronbichler [Tue, 7 Feb 2017 06:30:56 +0000 (07:30 +0100)]
Merge pull request #3928 from masterleinad/fix_filtered_iterator
Fix Initialization of FilteredIterator from base class
Daniel Arndt [Mon, 6 Feb 2017 16:50:21 +0000 (17:50 +0100)]
Add test for copying FilteredOperator
Daniel Arndt [Mon, 6 Feb 2017 18:37:33 +0000 (19:37 +0100)]
Merge pull request #3918 from koecher/deprecate_supports_distributed_data_static_member
deprecates static support_distributed_data member of vector declarations
Uwe Köcher [Mon, 6 Feb 2017 17:40:48 +0000 (18:40 +0100)]
adds DEAL_II_DEPRECATED to PREDEFINED of doc/doxygen/options.dox.in
sairajat [Mon, 6 Feb 2017 17:09:35 +0000 (12:09 -0500)]
Fixes bug in Pattern::Selection
Wolfgang Bangerth [Mon, 6 Feb 2017 15:47:50 +0000 (08:47 -0700)]
Merge pull request #3930 from tjhei/intel_inline_warning
remove intel warning about always_inline
Timo Heister [Mon, 6 Feb 2017 14:11:52 +0000 (09:11 -0500)]
Merge pull request #3850 from bangerth/add-abf-test
Add a test for the ABF element in 3d.
Timo Heister [Mon, 6 Feb 2017 14:10:25 +0000 (09:10 -0500)]
remove intel warning about always_inline
disable the following incorrect warning:
```
include/deal.II/base/tensor.h(824): warning #3415: the "always_inline"
attribute is ignored on non-inline functions
inline DEAL_II_ALWAYS_INLINE
```
Jean-Paul Pelteret [Mon, 6 Feb 2017 13:29:11 +0000 (14:29 +0100)]
Refactor SchurComplement operator.
Subsequent to the implementation of Payloads for LinearOperators, it is
necessary that compound operations such as this one be expressed in
terms of overloaded operations (operator+,* etc.) instead of in a
low-level format, with the operator functions v_mult etc. expanded
manually. This is because the payload is not correctly configured in the
latter case, and the operations defined on the Payload-level would not
align with those being performed by the higher-level LinearOperator
interface. When operator+,* etc. are called, the equivalent operations
for the payload are automatically set up.
Daniel Arndt [Sun, 5 Feb 2017 16:41:23 +0000 (17:41 +0100)]
Fix Initialization of FilteredIterator from base class
Wolfgang Bangerth [Sun, 5 Feb 2017 01:39:11 +0000 (18:39 -0700)]
Merge pull request #3925 from masterleinad/update_fe_project_test
Update fe/fe_project_2d
Daniel Arndt [Sat, 4 Feb 2017 17:22:05 +0000 (18:22 +0100)]
Update fe/fe_project_2d
Wolfgang Bangerth [Sat, 4 Feb 2017 15:01:02 +0000 (08:01 -0700)]
Merge pull request #3903 from tjhei/rework_instantiations
cleanup template-arguments.in
Timo Heister [Sat, 4 Feb 2017 14:33:23 +0000 (09:33 -0500)]
Merge pull request #3924 from drwells/boost-serialization-macro-warning
Avoid using 'defined' inside a macro.
Uwe Köcher [Fri, 3 Feb 2017 11:52:30 +0000 (12:52 +0100)]
deprecates static support_distributed_data member of vector declarations
David Wells [Sat, 4 Feb 2017 02:17:49 +0000 (21:17 -0500)]
Avoid using 'define' inside a macro.
This is technically illegal and clang rightfully complains.
Patch from https://github.com/boostorg/serialization/pull/49
Timo Heister [Fri, 3 Feb 2017 22:40:42 +0000 (17:40 -0500)]
fix expand-instantiations
- remove trailing newlines in pattern values
- skip spaces and newlines before first pattern
- report errors if pattern is not found
Martin Kronbichler [Fri, 3 Feb 2017 21:25:59 +0000 (22:25 +0100)]
Merge pull request #3923 from tjhei/undo_static_assert
Revert #3861 "add new macro for static asserts"
Martin Kronbichler [Fri, 3 Feb 2017 21:25:30 +0000 (22:25 +0100)]
Merge pull request #3891 from tjhei/step-57-fixes
update step-57
Denis Davydov [Fri, 3 Feb 2017 20:27:33 +0000 (21:27 +0100)]
Merge pull request #3916 from tamiko/do_not_use_as_needed_in_check
CMake: Disable -Wl,--as-needed in DEAL_II_HAVE_USABLE_FLAGS
Matthias Maier [Fri, 3 Feb 2017 08:35:32 +0000 (02:35 -0600)]
CMake: Disable -Wl,--as-needed in DEAL_II_HAVE_USABLE_FLAGS
This works around a serious regression with ld.bfd in combination with
-Wl,--as-needed when compiling a simple
int main () { return 0; }
and linking against a *huge* list of (entirely unused) libraries.
See #3686
In the beginning the Universe was created. This has made a lot of
people very angry and has been widely regarded as a bad move.
Douglas Adams
Typeset in
Trocchi and Trocchi Bold Sans Serif.