]>
https://gitweb.dealii.org/ - dealii.git/log
Martin Kronbichler [Wed, 29 Jul 2015 18:36:05 +0000 (20:36 +0200)]
Merge pull request #1216 from tjhei/fix_avoid_epetra_map
fix: avoid Epetra_Map in Trilinos interface
Timo Heister [Wed, 29 Jul 2015 17:02:01 +0000 (13:02 -0400)]
fix: avoid Epetra_Map in Trilinos interface
use DomainMap instead of ColMap.
Wolfgang Bangerth [Wed, 29 Jul 2015 15:37:34 +0000 (10:37 -0500)]
Merge pull request #1205 from tjhei/stl_bugs
Stl bugs
Wolfgang Bangerth [Wed, 29 Jul 2015 15:30:04 +0000 (10:30 -0500)]
Merge pull request #1214 from kronbichler/master
Fix compilation warning: &DoFCellAccessor::get_fe() == 0
Martin Kronbichler [Wed, 29 Jul 2015 14:53:31 +0000 (16:53 +0200)]
Fix compilation warning: &DoFCellAccessor::get_fe() == 0
reference cannot be bound to dereferenced null pointer in well-defined
C++ code; comparison may be assumed to always evaluate to true
Wolfgang Bangerth [Tue, 28 Jul 2015 19:31:11 +0000 (14:31 -0500)]
Merge pull request #1213 from tjhei/avoid_trilinos_epetra_map
avoid using Trilinos col/row_partitioner
Timo Heister [Tue, 28 Jul 2015 18:49:40 +0000 (14:49 -0400)]
avoid using Trilinos col/row_partitioner
Matthias Maier [Tue, 28 Jul 2015 18:45:48 +0000 (13:45 -0500)]
Merge pull request #1200 from tjhei/set_thread_limit
rework set_thread_limit()
Timo Heister [Tue, 28 Jul 2015 18:26:00 +0000 (14:26 -0400)]
fix stl errors and rework MPI recv/send
this addresses the two TODOs in the code
Timo Heister [Tue, 28 Jul 2015 17:53:44 +0000 (13:53 -0400)]
more stl fixes
Martin Kronbichler [Tue, 28 Jul 2015 17:39:59 +0000 (19:39 +0200)]
Merge pull request #1212 from bangerth/update-step-10
Minor updates to step-10.
Timo Heister [Tue, 28 Jul 2015 17:28:23 +0000 (13:28 -0400)]
early out compute_point_to_point_communication_pattern()
Fixes another stl invalid access
Wolfgang Bangerth [Tue, 28 Jul 2015 17:00:46 +0000 (12:00 -0500)]
Stylistic updates.
Wolfgang Bangerth [Tue, 28 Jul 2015 17:00:30 +0000 (12:00 -0500)]
Minor updates to step-10.
In particular, reference step-53.
Timo Heister [Mon, 27 Jul 2015 15:18:15 +0000 (11:18 -0400)]
update tests
Timo Heister [Sun, 26 Jul 2015 20:54:16 +0000 (16:54 -0400)]
rework set_thread_limit()
- allow calling set_thread_limit() more than once by reinitializing the
TBB task_scheduler
- always call set_thread_limit() using static initialization before any
user code (and thus TBB usage) happens
- as a result DEAL_II_NUM_THREADS will now always be respected even if
users never call set_thread_limit() themselves
- test suite: always enforce a maximum of 5 threads (instead of only if
not configured with MPI)
- add test for changing the thread limit
Wolfgang Bangerth [Tue, 28 Jul 2015 14:28:43 +0000 (09:28 -0500)]
Merge pull request #1211 from tjhei/fix_tests_for_intel15
tests: increase precision to fix on intel
Timo Heister [Tue, 28 Jul 2015 14:06:05 +0000 (10:06 -0400)]
tests: increase precision to fix on intel
Timo Heister [Mon, 27 Jul 2015 21:10:15 +0000 (17:10 -0400)]
fix out of bounds access
Martin Kronbichler [Tue, 28 Jul 2015 07:14:06 +0000 (09:14 +0200)]
Merge pull request #1208 from tjhei/fix_index_set_trilinos_constr
make IndexSet Epetra_Map constr explicit
Timo Heister [Tue, 28 Jul 2015 03:04:20 +0000 (23:04 -0400)]
make IndexSet Epetra_Map constr explicit
The new constructor IndexSet(const Epetra_Map &map) was not marked explicit
and caused ambiguous type conversions.
Wolfgang Bangerth [Mon, 27 Jul 2015 21:43:15 +0000 (16:43 -0500)]
Merge pull request #1206 from tjhei/remove_muelu_warnings
suppress muelu warnings
Timo Heister [Mon, 27 Jul 2015 21:40:59 +0000 (17:40 -0400)]
suppress muelu warnings
Disable extra diagnostic Wunused-but-set-variable when including
trilinos to avoid muelu warning.
Timo Heister [Mon, 27 Jul 2015 15:36:16 +0000 (11:36 -0400)]
avoid dereferencing empty std::vectors
Martin Kronbichler [Mon, 27 Jul 2015 14:39:01 +0000 (16:39 +0200)]
Merge pull request #1204 from bangerth/revise-mapping-fill_fe_values-interface
Revise mapping fill fe values interface
Wolfgang Bangerth [Sun, 26 Jul 2015 23:52:21 +0000 (18:52 -0500)]
Also update the changelog for the mapping classes.
Wolfgang Bangerth [Sun, 26 Jul 2015 23:22:44 +0000 (18:22 -0500)]
Revise the Mapping::fill_fe_*_values() interfaces.
These functions previously got a whole lot of arguments for the output arrays.
Consolidate them by simply passing the structure in which they actually lie.
While there, also do:
- Sort these arguments in such a way that they make a bit more sense.
- Provide a thorough documentation of these functions.
Wolfgang Bangerth [Sun, 26 Jul 2015 20:21:09 +0000 (15:21 -0500)]
Fix a typo in a previous patch.
Martin Kronbichler [Mon, 27 Jul 2015 06:43:31 +0000 (08:43 +0200)]
Merge pull request #1203 from bangerth/fix-pr-1183
Fix #1183.
Wolfgang Bangerth [Mon, 27 Jul 2015 03:16:27 +0000 (22:16 -0500)]
Fix #1183.
In #1183, I accidentally removed a class qualification on a call. This leads
to some 20 tests failing. This patch restores the issue.
Wolfgang Bangerth [Sun, 26 Jul 2015 22:19:39 +0000 (17:19 -0500)]
Merge pull request #1202 from tjhei/fix_changelog
move changelog entries that went into 8.3
Timo Heister [Sun, 26 Jul 2015 21:48:18 +0000 (17:48 -0400)]
move changelog entries that went into 8.3
Timo Heister [Sun, 26 Jul 2015 21:20:06 +0000 (17:20 -0400)]
Merge pull request #1199 from luca-heltai/update-travis
Update Travis setup.
Luca Heltai [Sun, 26 Jul 2015 18:59:05 +0000 (20:59 +0200)]
Updated travis config.
Luca Heltai [Sun, 26 Jul 2015 18:23:54 +0000 (20:23 +0200)]
Merge pull request #1182 from bangerth/make-mapping-internal-data-const
Make mapping internal data const
Wolfgang Bangerth [Sun, 26 Jul 2015 17:36:56 +0000 (12:36 -0500)]
Update documentation in a few other places as well.
Wolfgang Bangerth [Sun, 26 Jul 2015 17:32:14 +0000 (12:32 -0500)]
Document the various Mapping*::InternalData classes.
Wolfgang Bangerth [Sun, 26 Jul 2015 17:23:26 +0000 (12:23 -0500)]
Document Mapping::InternalDataBase better.
Wolfgang Bangerth [Sun, 26 Jul 2015 16:55:54 +0000 (11:55 -0500)]
Update documentation of MappingType.
Wolfgang Bangerth [Sun, 26 Jul 2015 16:03:29 +0000 (11:03 -0500)]
Adapt the recent changed to MappingFEField to be compatible with what we do here.
Wolfgang Bangerth [Thu, 23 Jul 2015 22:10:57 +0000 (17:10 -0500)]
Move a member variable to derived classes.
The variable isn't used in the base class, and not all derived classes
use it either.
Wolfgang Bangerth [Wed, 22 Jul 2015 21:49:36 +0000 (16:49 -0500)]
Make the internal data object of the mapping classes const.
Logically, they are input arguments to the fill_fe_values() functions, they are only
initialized in get_data() etc. Thus, make them const in the fill_fe_values functions.
There are places where mapping classes want to move data across from fill_fe_values
in terms of temporary arrays to the transform() function that may be called right
afterwards from FiniteElement::fill_fe_values(). In those cases, make these members
mutable.
Wolfgang Bangerth [Thu, 23 Jul 2015 16:19:47 +0000 (11:19 -0500)]
Let second_derivative() return a const reference, rather than a value.
This function is used in places where we want to get pointers into an array. Make sure that
is possible for both the const and the non-const version of the function. (The latter
already did this; the former didn't and that wasn't needed, but I need it for a
follow-up patch.)
Martin Kronbichler [Sun, 26 Jul 2015 12:14:14 +0000 (14:14 +0200)]
Merge pull request #1191 from luca-heltai/no-warnings-on-reinit
Deprecate Epetra_Map constructors in Trilinos::SparseMatrix and add getters for IndexSets
Martin Kronbichler [Sun, 26 Jul 2015 10:45:11 +0000 (12:45 +0200)]
Merge pull request #1197 from luca-heltai/fix-mapping-fe-field
Fixed #941.
Luca Heltai [Sun, 26 Jul 2015 10:40:20 +0000 (12:40 +0200)]
Made get_data return InternalData.
Martin Kronbichler [Sun, 26 Jul 2015 10:23:07 +0000 (12:23 +0200)]
Merge pull request #1195 from luca-heltai/fix-fe-field
Fixed bug in VectorTools::get_position_vector.
Luca Heltai [Sun, 26 Jul 2015 10:04:43 +0000 (12:04 +0200)]
Moved dof values and indices inside InternalData. Closes #941.
Luca Heltai [Sat, 25 Jul 2015 10:27:06 +0000 (12:27 +0200)]
Added entry in changes.h. Closes #1029.
Luca Heltai [Fri, 24 Jul 2015 18:23:19 +0000 (20:23 +0200)]
Added get_mpi_communicator() to Trilinos Matrix.
Luca Heltai [Fri, 24 Jul 2015 17:43:17 +0000 (19:43 +0200)]
Added IndexSet constructor with Epetra_Map
Luca Heltai [Sat, 25 Jul 2015 16:55:21 +0000 (18:55 +0200)]
Fixed documentation and a few const.
Luca Heltai [Sat, 25 Jul 2015 16:00:06 +0000 (18:00 +0200)]
Fixed bug in MappingFEField. Closes #1138.
Luca Heltai [Sat, 25 Jul 2015 15:22:14 +0000 (17:22 +0200)]
Merge pull request #1194 from luca-heltai/fix-qgauss-log
Added Assert in QGaussLogR.
Luca Heltai [Sat, 25 Jul 2015 15:13:53 +0000 (17:13 +0200)]
Fixed indentation.
Luca Heltai [Sat, 25 Jul 2015 13:45:30 +0000 (15:45 +0200)]
Added Assert in QGaussLog and fixed test. Closes #1058.
Wolfgang Bangerth [Fri, 24 Jul 2015 22:18:22 +0000 (17:18 -0500)]
Merge pull request #1193 from drwells/german-to-english-comment
Translate a comment from German to English.
David Wells [Fri, 24 Jul 2015 20:18:57 +0000 (16:18 -0400)]
Translate a comment from German to English.
Timo Heister [Fri, 24 Jul 2015 18:47:48 +0000 (14:47 -0400)]
Merge pull request #1192 from bangerth/fix-oversight
Remove an unnecessary #include file.
Wolfgang Bangerth [Fri, 24 Jul 2015 18:16:54 +0000 (13:16 -0500)]
Remove an unnecessary #include file.
Wolfgang Bangerth [Fri, 24 Jul 2015 18:03:51 +0000 (13:03 -0500)]
Merge pull request #1190 from tjhei/fix_changelog
move changelog entries that got included in 8.3
Timo Heister [Fri, 24 Jul 2015 16:59:54 +0000 (12:59 -0400)]
move changelog entries that got included in 8.3
Timo Heister [Fri, 24 Jul 2015 16:49:55 +0000 (12:49 -0400)]
Merge pull request #1180 from bangerth/fix-refinement
Make sure that computed refinement thresholds are no larger than the max indicator
Martin Kronbichler [Fri, 24 Jul 2015 16:30:02 +0000 (18:30 +0200)]
Merge pull request #1188 from bangerth/fix-comments
A couple of forgotten things.
Wolfgang Bangerth [Fri, 24 Jul 2015 15:59:12 +0000 (10:59 -0500)]
Re-break comments to the usual length.
Wolfgang Bangerth [Fri, 24 Jul 2015 15:58:35 +0000 (10:58 -0500)]
Update the numbers in a comment.
Wolfgang Bangerth [Fri, 24 Jul 2015 15:58:25 +0000 (10:58 -0500)]
Add a changelog entry.
Martin Kronbichler [Fri, 24 Jul 2015 12:23:45 +0000 (14:23 +0200)]
Merge pull request #1183 from bangerth/use-covariant-return-types
Simplify code.
Timo Heister [Fri, 24 Jul 2015 12:20:29 +0000 (08:20 -0400)]
Merge pull request #1184 from bangerth/cosmetic-changes
Cosmetic changes.
Bruno Turcksin [Thu, 23 Jul 2015 22:58:51 +0000 (17:58 -0500)]
Merge pull request #1181 from bangerth/remove-unused-variable
Remove an unused variable.
Wolfgang Bangerth [Thu, 23 Jul 2015 22:21:27 +0000 (17:21 -0500)]
Cosmetic changes.
There's a show on TV as I write on cosmetic surgery. I thought I'd do some of
my own in a couple of places I found over the past few days.
Wolfgang Bangerth [Thu, 23 Jul 2015 22:07:07 +0000 (17:07 -0500)]
Simplify code.
There are a couple of places where we call a get_data() function
that returns a pointer to a base class when we know that the actual
data type is a pointer to derived class. We need to cast right back
to the derived class.
This can be avoided by using covariant return types where the get_data()
function returns a pointer to derived right away.
Wolfgang Bangerth [Thu, 23 Jul 2015 21:29:34 +0000 (16:29 -0500)]
Remove an unused variable.
This must be a left-over from some earlier development.
Wolfgang Bangerth [Thu, 23 Jul 2015 16:00:51 +0000 (11:00 -0500)]
Make sure that compute refinement thresholds are no larger than the max indicator.
This could happen because we search the threshold in an interval larger than the
actual indicators. If we end up with a threshold that is larger than the largest
indicator, we need to cut things back.
In the process of debugging, I also found a few oddities that have to do
with the fact that we do a lot of arithmetic in double precision, but often
pass in criteria as vectors of float. Fix some of those as well.
Matthias Maier [Thu, 23 Jul 2015 13:42:06 +0000 (08:42 -0500)]
Merge pull request #1178 from tjhei/fix_test
Fix tests
Matthias Maier [Thu, 23 Jul 2015 13:41:35 +0000 (08:41 -0500)]
Merge pull request #1179 from tjhei/fix_cmake_bug
Fix cmake configure issue introduced by #1148
Timo Heister [Thu, 23 Jul 2015 13:27:41 +0000 (09:27 -0400)]
Fix cmake configure issue introduced by #1148
add missing RESET_CMAKE_REQUIRED() introduced by #1148 and commit
e02e450 .
Timo Heister [Thu, 23 Jul 2015 13:01:26 +0000 (09:01 -0400)]
make tests/fe/cell_similariy_11 work without MPI
Avoid exception "Calling set_thread_limit() more than once is not
supported!".
Timo Heister [Thu, 23 Jul 2015 13:00:33 +0000 (09:00 -0400)]
Test: adjust iteration counts
Matthias Maier [Wed, 22 Jul 2015 21:07:08 +0000 (16:07 -0500)]
Merge pull request #1167 from tjhei/test_fixes_muelu
fix muelu test output
Matthias Maier [Wed, 22 Jul 2015 21:06:13 +0000 (16:06 -0500)]
Merge pull request #1176 from tjhei/dynamic_sp_fix
fix invalid operation in DynamicSparsityPattern::iterator
Timo Heister [Wed, 22 Jul 2015 19:42:31 +0000 (15:42 -0400)]
fix invalid operation in DynamicSparsityPattern::iterator
We used to compare two std::vector iterators that might belong to
different std::vectors. This is obviously illegal.
This addresses issue #1175, see
https://github.com/dealii/dealii/issues/1175
Add a test that failed (with stl debug library only).
Martin Kronbichler [Wed, 22 Jul 2015 16:51:24 +0000 (18:51 +0200)]
Merge pull request #1173 from bangerth/more-CellSimilarity-issues
Make the CellSimilarity argument const also in Mapping.
Wolfgang Bangerth [Tue, 21 Jul 2015 14:28:51 +0000 (09:28 -0500)]
Make the CellSimilarity argument const also in Mapping.
Logically, these arguments are intended to be input arguments, so make them
const. Also, don't make them a reference as they are simply integral values.
That said, the original intention of having a reference was so that
mappings can indicate situations where cell similarity has to be
disregarded. Consequently, let the fill_fe_values() functions return a
cell similarity value that will then simply be assigned to the same
variable again.
Wolfgang Bangerth [Wed, 22 Jul 2015 12:40:26 +0000 (07:40 -0500)]
Merge pull request #1174 from kronbichler/master
Test case for cell similarity
Martin Kronbichler [Wed, 22 Jul 2015 08:17:14 +0000 (10:17 +0200)]
Test case for cell similarity
Timo Heister [Tue, 21 Jul 2015 15:46:13 +0000 (11:46 -0400)]
Merge pull request #1171 from kronbichler/master
Fix #1168, make compile with boost bind.
Martin Kronbichler [Tue, 21 Jul 2015 15:08:59 +0000 (17:08 +0200)]
Make compile with boost bind.
Timo Heister [Tue, 21 Jul 2015 15:44:21 +0000 (11:44 -0400)]
Merge pull request #1169 from bangerth/replace-random_shuffle
Replace the use of std::random_shuffle.
Wolfgang Bangerth [Tue, 21 Jul 2015 12:27:26 +0000 (07:27 -0500)]
Replace the use of std::random_shuffle.
It turns out that this function has been deprecated for C++17, see
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2014/n3924.pdf
As a side effect, the function has also gotten more predictable.
Wolfgang Bangerth [Mon, 20 Jul 2015 23:12:02 +0000 (01:12 +0200)]
Merge pull request #1122 from tjhei/iterate_index_sets
new IndexSet iterators
Timo Heister [Mon, 20 Jul 2015 21:01:27 +0000 (17:01 -0400)]
fix muelu test output
Matthias Maier [Mon, 20 Jul 2015 21:21:03 +0000 (16:21 -0500)]
Merge pull request #1166 from bangerth/clarify-fill_fe_values-arguments
Clarify fill fe values arguments
Matthias Maier [Mon, 20 Jul 2015 21:18:01 +0000 (16:18 -0500)]
Merge pull request #1164 from kronbichler/master
Reduce number of failing tests with FMA
Wolfgang Bangerth [Mon, 20 Jul 2015 18:36:59 +0000 (13:36 -0500)]
Add a changelog entry.
Wolfgang Bangerth [Mon, 20 Jul 2015 18:24:24 +0000 (13:24 -0500)]
Pass CellSimilarity arguments by value rather than by reference.
These are integral valued arguments for which references are not necessary. They are
also all input arguments.
Wolfgang Bangerth [Sun, 19 Jul 2015 14:58:34 +0000 (09:58 -0500)]
Mark two more of the FiniteElement::fill_fe_values arguments as const.
These two arguments were supposed to be input arguments, but weren't marked as const. Do
this here, and document the respective functions better.
Wolfgang Bangerth [Mon, 20 Jul 2015 18:22:51 +0000 (13:22 -0500)]
Update the code in a way that makes it clearer.
Specifically, don't modify an input argument of the function but instead use the
old value and just modify the conditions in which it is used. This makes it clearer
when certain actions have to be performed.
Wolfgang Bangerth [Mon, 20 Jul 2015 17:32:32 +0000 (12:32 -0500)]
Rewrite the FE_DGPNonparametric::fill_fe_*_values() functions.
Specifically, make sure that the FiniteElement::InternalDataBase object they receive is
really used specifically only for read-only purposes. This requires that we allocate some
memory in the fill_fe_*_values() functions for scratch arrays, but on the upside, we can
get rid of the FE_DGPNonparametric::InternalData class.
Wolfgang Bangerth [Mon, 20 Jul 2015 17:42:39 +0000 (12:42 -0500)]
Add some documentation.
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.