]>
https://gitweb.dealii.org/ - dealii.git/log
Matthias Maier [Sat, 26 May 2018 03:08:49 +0000 (22:08 -0500)]
doc: add incompatibilities news item for create_(boundary_;)mass_matrix
Matthias Maier [Sat, 26 May 2018 03:04:25 +0000 (22:04 -0500)]
doc: add incompatibilities news item for AffineConstraints/ConstraintMatrix
Matthias Maier [Sat, 26 May 2018 06:59:12 +0000 (01:59 -0500)]
Workaround: Temporarily define ConstraintMatrix in this header
Matthias Maier [Sat, 26 May 2018 05:39:43 +0000 (00:39 -0500)]
Bugfix: Add explicit constraint_matrix.h includes to example steps
Previously they had been included by chance. Due to the refactoring all
of these includes have to be done explicitly (at least for a short
transition period.)
Daniel Arndt [Mon, 28 May 2018 20:00:05 +0000 (15:00 -0500)]
Tests: Fix the remaining tests using complex scalars
Daniel Arndt [Mon, 28 May 2018 16:22:42 +0000 (18:22 +0200)]
Tests: Fix nearly all the failing PETSc complex tests
Matthias Maier [Mon, 28 May 2018 03:51:25 +0000 (22:51 -0500)]
Tests: Fix matrix_free/step-37
Matthias Maier [Mon, 28 May 2018 03:01:59 +0000 (22:01 -0500)]
Tests: Do not include "constraint_matrix.templates.h"
Daniel Arndt [Sun, 27 May 2018 16:59:13 +0000 (18:59 +0200)]
Bugfix: Avoid warnings regarding ABI change
Matthias Maier [Sat, 26 May 2018 02:28:54 +0000 (21:28 -0500)]
RCM: Rename some stray leftovers in source
Matthias Maier [Mon, 28 May 2018 00:44:37 +0000 (19:44 -0500)]
MatrixFree: Allow for non-matching number types in constraint handling
In order to facilitate an AffineConstraints<float> in the matrix free
framework:
- templatify internal ConstraintValues::insert_entries to convert from
compatible number2 types to Number types (aka exclusively 'double').
Matthias Maier [Mon, 28 May 2018 00:25:06 +0000 (19:25 -0500)]
MatrixFree: Templatify reinit functions
Unfortunately, we have a bit of a clashing ideology here:
- The MatrixFree container is motivated by the old ConstraintMatrix
object. Meaning it expects constraints to be expressed in double.
- AffineConstraints<number>, however, should be matched with the
underlying number type. And the rest of the library now assumes that
this is the case (in particular project_matrix_free and variants...)
In order to facilitate both approaches, we have to slab another "number"
template aorund the reinit functions. That's a bit unfortunate but
cannot be avoided. :-/
Matthias Maier [Sun, 27 May 2018 23:12:39 +0000 (18:12 -0500)]
RCM: Update matrix_free, part 3: matrix_free and dof_info
Matthias Maier [Fri, 25 May 2018 22:31:35 +0000 (17:31 -0500)]
RCM: Update matrix_free, part 2: cuda_matrix_free
Matthias Maier [Fri, 25 May 2018 22:26:53 +0000 (17:26 -0500)]
RCM: Update matrix_free, part 1: update comments (no code changes)
Matthias Maier [Sat, 26 May 2018 00:13:38 +0000 (19:13 -0500)]
RCM: Update multigrid
Matthias Maier [Sat, 26 May 2018 02:24:38 +0000 (21:24 -0500)]
RCM: Update mesh_worker
Matthias Maier [Sat, 26 May 2018 01:21:29 +0000 (20:21 -0500)]
RCM: Update numerics, part 3: vector_tools
Matthias Maier [Sat, 26 May 2018 01:02:43 +0000 (20:02 -0500)]
RCM: Update numerics, part 2: update comments (no code changes)
Matthias Maier [Sat, 26 May 2018 00:52:21 +0000 (19:52 -0500)]
RCM: Update numerics, part 1: matrix_tools, matrix_creator
- Incompatible change: Change the signature of all mass matrix creator
functions to matching number types for matrices and right-hand side.
Matthias Maier [Fri, 25 May 2018 22:16:13 +0000 (17:16 -0500)]
RCM: Update lac, part 2: constrained_linear_operator
Matthias Maier [Fri, 25 May 2018 22:09:14 +0000 (17:09 -0500)]
RCM: Update lac, part 1: update comments (no code changes)
Matthias Maier [Fri, 25 May 2018 22:05:49 +0000 (17:05 -0500)]
RCM: Update non_matching: coupling
Matthias Maier [Fri, 25 May 2018 21:11:51 +0000 (16:11 -0500)]
RCM: Update fe: fe, fe_tools
Matthias Maier [Mon, 28 May 2018 19:58:42 +0000 (14:58 -0500)]
Bugfix: Use correct number type in dof_tools_constraints.(cc|inst.in)
DoFTools::make_hanging_node_constraints always uses double internally,
but is also instantiated for std::complex<double>
Matthias Maier [Sat, 26 May 2018 16:22:02 +0000 (11:22 -0500)]
Bugfix: Add missing template default parameters to dof_tools.h
Matthias Maier [Fri, 25 May 2018 18:35:17 +0000 (13:35 -0500)]
RCM: Update dofs, part 2: dof_tools
TODO: Complexify make_periodicity_constraints
Matthias Maier [Fri, 25 May 2018 18:09:26 +0000 (13:09 -0500)]
RCM: Update dofs, part 1: dof_accessor
Matthias Maier [Fri, 25 May 2018 18:00:16 +0000 (13:00 -0500)]
RCM: Update base
Matthias Maier [Fri, 25 May 2018 17:58:37 +0000 (12:58 -0500)]
RCM: Remove ConstraintMatrix, RIP 1998 - 2018
Matthias Maier [Mon, 28 May 2018 02:58:44 +0000 (21:58 -0500)]
Workaround: Add compatibility instantiations
David Wells [Mon, 28 May 2018 02:33:28 +0000 (22:33 -0400)]
Bugfix: Use std::abs instead of std::fabs.
This has a correct overload for std::complex.
Daniel Arndt [Sun, 27 May 2018 15:30:35 +0000 (17:30 +0200)]
Bugfix: Add missing instantiations
Matthias Maier [Thu, 24 May 2018 20:56:34 +0000 (15:56 -0500)]
lac: add number template parameter to affine_constraints.cc and refactor
- Add support for std::complex in internal data structures
- Refactor special instantiations into .inst.in file.
- Workaround: add two static_cast<number>(...) to resolve an ambiguity
when a containerized std::complex is multiplied with another
std::complex.
- replace standard comparator for sorting a vector of std::pair with a
lexicographic ordering where the second component is compared with
std::abs.
- templatify and simplify internal data structures
- simplify internal scratch data (only supports one uniform number type
now)
Matthias Maier [Thu, 24 May 2018 20:41:21 +0000 (15:41 -0500)]
remove static "default_empty_table"
Matthias Maier [Thu, 24 May 2018 20:28:30 +0000 (15:28 -0500)]
lac: add number template parameter to affine_constraints.templates.h
Matthias Maier [Thu, 24 May 2018 20:15:08 +0000 (15:15 -0500)]
lac: add number template parameter to affine_constraints.h
Matthias Maier [Thu, 24 May 2018 17:09:53 +0000 (12:09 -0500)]
lac: Add AffineConstraints class
For now this is a verbatim copy of the ConstraintMatrix class. The idea
is to templatify this class and after that switch the code base to it.
Matthias Maier [Thu, 24 May 2018 19:27:49 +0000 (14:27 -0500)]
lac: remove scratch_data_initializer from constraint_matrix.cc
Matthias Maier [Sat, 26 May 2018 02:00:51 +0000 (21:00 -0500)]
lac/solver_cg.h: Preparatory changes for complex number support
Matthias Maier [Sat, 26 May 2018 04:09:29 +0000 (23:09 -0500)]
Bugfix: Make sparse_matrix_ez and chunk_sparse_matrix fit for std::complex
Daniel Arndt [Sat, 26 May 2018 12:18:14 +0000 (14:18 +0200)]
Bugfix: Fix lapack stev call in TridiagonalMatrix<number>::compute_eigenvalues()
Matthias Maier [Sat, 26 May 2018 03:43:01 +0000 (22:43 -0500)]
Bugfix: Instantiate two classes for std::complex
- instantiate BlockSparseMatrix for std::complex
- instantiate TridiagonalMatrix for std::complex
Daniel Arndt [Wed, 6 Jun 2018 15:07:41 +0000 (17:07 +0200)]
Merge pull request #6719 from masterleinad/clang_tidy_1
Fix instantiations of SymmetricTensor
Matthias Maier [Wed, 6 Jun 2018 14:15:36 +0000 (09:15 -0500)]
Merge pull request #6710 from masterleinad/deprecate_parallel_distributed_vector
Deprecate parallel::distributed::Vector
Matthias Maier [Wed, 6 Jun 2018 14:13:53 +0000 (09:13 -0500)]
Merge pull request #6721 from masterleinad/clang_tidy_2
Simplify GridOut::Gnuplot
Wolfgang Bangerth [Wed, 6 Jun 2018 14:06:52 +0000 (08:06 -0600)]
Merge pull request #6720 from masterleinad/clang_tidy_3
Use static constexpr in VectorizedArray
Daniel Arndt [Wed, 6 Jun 2018 00:57:47 +0000 (02:57 +0200)]
Fix instantiations of SymmetricTensor
Martin Kronbichler [Wed, 6 Jun 2018 10:20:12 +0000 (12:20 +0200)]
Merge pull request #6717 from masterleinad/clang_tidy_6
Fix clang-tidy findings for muParser
Daniel Arndt [Tue, 5 Jun 2018 22:25:16 +0000 (00:25 +0200)]
Simplify GridOut::Gnuplot
Daniel Arndt [Tue, 5 Jun 2018 22:05:10 +0000 (00:05 +0200)]
Use static constexpr in VectorizedArray
Daniel Arndt [Tue, 5 Jun 2018 21:12:11 +0000 (23:12 +0200)]
Fix muparser
Matthias Maier [Wed, 6 Jun 2018 05:39:50 +0000 (00:39 -0500)]
Merge pull request #6716 from tjhei/install_license_md
install LICENSE.md not LICENSE
Timo Heister [Wed, 6 Jun 2018 05:30:30 +0000 (13:30 +0800)]
install LICENSE.md not LICENSE
fixes #6706
Denis Davydov [Wed, 6 Jun 2018 03:20:34 +0000 (05:20 +0200)]
Merge pull request #6712 from masterleinad/fix_cuda_9_2
Fix compile error with CUDA 9.2
Matthias Maier [Wed, 6 Jun 2018 01:56:54 +0000 (20:56 -0500)]
Merge pull request #6713 from bangerth/indent
Provide a more readable indentation.
Wolfgang Bangerth [Tue, 5 Jun 2018 23:47:37 +0000 (17:47 -0600)]
Provide a more readable indentation.
Daniel Arndt [Tue, 5 Jun 2018 22:08:43 +0000 (00:08 +0200)]
Fix compile error with CUDA 9.2
Wolfgang Bangerth [Tue, 5 Jun 2018 15:23:23 +0000 (09:23 -0600)]
Merge pull request #6707 from tamiko/safeguard
CMake: add a safeguard for accidential /source source directory
Wolfgang Bangerth [Tue, 5 Jun 2018 15:21:22 +0000 (09:21 -0600)]
Merge pull request #6706 from tamiko/update_legalese
Update LICENSE file
Daniel Arndt [Mon, 4 Jun 2018 16:38:05 +0000 (18:38 +0200)]
Deprecate parallel::distributed::Vector
Matthias Maier [Tue, 5 Jun 2018 02:21:55 +0000 (21:21 -0500)]
Make request stronger
Matthias Maier [Tue, 5 Jun 2018 02:15:36 +0000 (21:15 -0500)]
CMake: fix path in description
Matthias Maier [Mon, 4 Jun 2018 03:20:49 +0000 (22:20 -0500)]
CMake: make safeguard even more clear.
Matthias Maier [Sun, 3 Jun 2018 20:32:46 +0000 (15:32 -0500)]
CMake: add a safeguard for accidential /source source directory
This happens remarkably often. Let's improve the error message:
```
Wrong source directory specified.
You pointed cmake to the subdirectory "source" instead of the base
directory. Make sure that the path in the cmake invocation does not end in
source. If you have created a "build" subdirectory, the cmake invocation
should simply be of the form:
$ cmake [...options...] /..
(path not ending in ".../source")
```
Matthias Maier [Mon, 4 Jun 2018 23:02:22 +0000 (18:02 -0500)]
Merge pull request #6700 from masterleinad/adjust_cxx_check_flag
Modify pushed flags for CLANGAUTODEBUG_BUG
Daniel Arndt [Mon, 4 Jun 2018 22:17:04 +0000 (00:17 +0200)]
Strip the offending flags instead
Matthias Maier [Mon, 4 Jun 2018 16:56:33 +0000 (11:56 -0500)]
Merge pull request #6709 from masterleinad/fix_warnings_in_tests
Remove some more references to astyle
Daniel Arndt [Mon, 4 Jun 2018 11:44:30 +0000 (13:44 +0200)]
Remove some more references to astyle
Matthias Maier [Mon, 4 Jun 2018 03:04:13 +0000 (22:04 -0500)]
Wolfgang's suggestion - slightly modified
Matthias Maier [Sat, 2 Jun 2018 21:45:18 +0000 (16:45 -0500)]
Implement J-P's suggestions
Matthias Maier [Sat, 2 Jun 2018 17:38:36 +0000 (12:38 -0500)]
Update LICENSE file
Martin Kronbichler [Fri, 1 Jun 2018 14:24:17 +0000 (16:24 +0200)]
Merge pull request #6702 from Zonni/fix-sundials-documentation
Fix sundials documentation
Antoni Vidal [Thu, 31 May 2018 17:17:05 +0000 (19:17 +0200)]
Bug fixed if output_step function is not defined. Now example in the documentation works.
Antoni Vidal [Thu, 31 May 2018 16:51:50 +0000 (18:51 +0200)]
Fixed sundials/ida.h example of use.
Wolfgang Bangerth [Thu, 31 May 2018 12:54:59 +0000 (06:54 -0600)]
Merge pull request #6699 from masterleinad/fix_warnings_in_tests
Fix warnings in tests
Daniel Arndt [Wed, 30 May 2018 20:52:29 +0000 (22:52 +0200)]
Alternative solution to mpi/collective_02_dealii_vector
Daniel Arndt [Wed, 30 May 2018 18:57:21 +0000 (20:57 +0200)]
Fix warnings in tests, gcc-8
Daniel Arndt [Thu, 31 May 2018 06:41:07 +0000 (08:41 +0200)]
Fix warnings in tests, clang-6
Daniel Arndt [Thu, 31 May 2018 06:40:58 +0000 (08:40 +0200)]
Modify pushed flags for CLANG_AUTODEBUGBUG
David Wells [Wed, 30 May 2018 21:31:09 +0000 (17:31 -0400)]
Merge pull request #6670 from bangerth/step-6
Next step in the step-6 introduction.
Daniel Arndt [Wed, 30 May 2018 17:32:14 +0000 (19:32 +0200)]
Merge pull request #6697 from masterleinad/avoid_sleep
Avoid using unistd.h and sleep
Daniel Arndt [Wed, 30 May 2018 11:25:21 +0000 (13:25 +0200)]
Avoid using unistd.h and sleep
Daniel Arndt [Wed, 30 May 2018 09:34:55 +0000 (11:34 +0200)]
Merge pull request #6692 from davydden/feature/timer_output_width
extend TimerOutput to have dynamic Section width
Denis Davydov [Tue, 29 May 2018 22:50:01 +0000 (00:50 +0200)]
use initlog()
Denis Davydov [Tue, 29 May 2018 13:21:09 +0000 (15:21 +0200)]
extend TimerOutput to have dynamic Section width
Matthias Maier [Tue, 29 May 2018 05:32:22 +0000 (00:32 -0500)]
Merge pull request #6689 from masterleinad/fix_multigrid_transfer_04b
Fix multigrid/transfer_04b
Daniel Arndt [Mon, 28 May 2018 21:13:15 +0000 (23:13 +0200)]
Fix multigrid/transfer_04b
Wolfgang Bangerth [Mon, 28 May 2018 15:57:57 +0000 (10:57 -0500)]
Merge pull request #6674 from drwells/add-merge-tria-tolerance
Add a tolerance for merge_triangulations.
David Wells [Mon, 28 May 2018 14:26:02 +0000 (10:26 -0400)]
Merge pull request #6680 from bangerth/step-3
A few updates to the introduction of step-3.
Wolfgang Bangerth [Sun, 27 May 2018 15:31:33 +0000 (23:31 +0800)]
A few updates to the introduction of step-3.
David Wells [Thu, 24 May 2018 18:36:03 +0000 (14:36 -0400)]
Add a tolerance for merge_triangulations.
The default setting is not robust with respect to roundoff errors (or
other minor numerical errors).
David Wells [Mon, 28 May 2018 12:09:42 +0000 (08:09 -0400)]
Merge pull request #6677 from bangerth/step-2
Update some of the introduction of step-2.
David Wells [Mon, 28 May 2018 10:36:38 +0000 (06:36 -0400)]
Merge pull request #6685 from bangerth/doc-update
Document arguments in 'contrib/utilities/indent'.
Wolfgang Bangerth [Mon, 28 May 2018 07:31:54 +0000 (15:31 +0800)]
Document arguments in 'contrib/utilities/indent'.
Wolfgang Bangerth [Mon, 28 May 2018 07:24:31 +0000 (15:24 +0800)]
Uniformly use capital letters at the top of comments.
Wolfgang Bangerth [Sun, 27 May 2018 03:41:21 +0000 (11:41 +0800)]
Address David Wells' comments.
Wolfgang Bangerth [Sun, 27 May 2018 05:55:48 +0000 (13:55 +0800)]
Update some of the introduction of step-2.
Daniel Arndt [Sun, 27 May 2018 21:38:36 +0000 (23:38 +0200)]
Merge pull request #6682 from masterleinad/fix_complex_petsc
Fix undefined reference for PETSc complex scalar type
Daniel Arndt [Sun, 27 May 2018 18:57:02 +0000 (20:57 +0200)]
Fix undefined reference for PETSc complex scalar type
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.