]>
https://gitweb.dealii.org/ - dealii.git/log
David Wells [Tue, 14 Aug 2018 22:27:42 +0000 (18:27 -0400)]
Merge pull request #7060 from bangerth/brace-initializers
Use brace initializers for default arguments.
Wolfgang Bangerth [Tue, 14 Aug 2018 22:17:04 +0000 (16:17 -0600)]
Merge pull request #7051 from masterleinad/deprecate_mgtools_extract_non_interface_dofs
Deprecate MGTools::extract_non_interface_dofs
Daniel Arndt [Tue, 14 Aug 2018 20:59:38 +0000 (22:59 +0200)]
Merge pull request #7057 from Rombur/cuda_misc
CUDA preprocessor guard
Wolfgang Bangerth [Tue, 14 Aug 2018 18:11:54 +0000 (12:11 -0600)]
Merge pull request #7049 from drwells/add-manifold-extrusion
Add manifold extrusion
Wolfgang Bangerth [Tue, 14 Aug 2018 15:22:30 +0000 (09:22 -0600)]
Indent correctly.
Wolfgang Bangerth [Tue, 14 Aug 2018 15:21:43 +0000 (09:21 -0600)]
Use brace initializers for default arguments.
Wolfgang Bangerth [Tue, 14 Aug 2018 12:00:04 +0000 (06:00 -0600)]
Merge pull request #7058 from dangars/fix_bug_active_fe_transfer
Fix bug in template instantiations of active_fe_indices_transfer.cc
Daniel Garcia-Sanchez [Tue, 14 Aug 2018 10:07:25 +0000 (12:07 +0200)]
Fix bug in template instantiations of active_fe_indices_transfer.cc
Wolfgang Bangerth [Mon, 13 Aug 2018 23:01:36 +0000 (17:01 -0600)]
Merge pull request #6961 from marcfehling/active-fe-transfer
Add class 'ActiveFEIndicesTransfer' for hp::DoFHandler on p::d::Triangulation.
Bruno Turcksin [Mon, 13 Aug 2018 14:41:30 +0000 (14:41 +0000)]
Better guard CUDA code
Use __CUDACC__ guard around CUDA to make sure that the CUDA code is only sent to
nvcc and not the host compiler when CUDA support is enable.
Marc Fehling [Thu, 19 Jul 2018 20:22:40 +0000 (14:22 -0600)]
Add class 'ActiveFEIndicesTransfer' for hp::DoFHandler on p::d::Triangulation.
David Wells [Sat, 11 Aug 2018 20:48:42 +0000 (16:48 -0400)]
Remove internal::plate_with_a_hole.
David Wells [Sat, 11 Aug 2018 20:44:44 +0000 (16:44 -0400)]
Remove the material_id technique from plate_with_a_hole.
David Wells [Sat, 4 Aug 2018 02:48:21 +0000 (22:48 -0400)]
Add an option to copy manifold ids in extrude_triangulation.
Bruno Turcksin [Thu, 2 Aug 2018 14:40:39 +0000 (14:40 +0000)]
Fix a typo
Martin Kronbichler [Mon, 13 Aug 2018 11:52:46 +0000 (13:52 +0200)]
Merge pull request #6277 from niveshd/fe_enriched_overlapping_domains
Fe enriched overlapping domains
Denis Davydov [Mon, 13 Aug 2018 11:41:56 +0000 (13:41 +0200)]
Merge pull request #7055 from masterleinad/alternative_output_unsubscribe_subscriptor
Add alternative (clang) output for unsubscribe_subscriptor
Daniel Arndt [Mon, 13 Aug 2018 10:42:22 +0000 (12:42 +0200)]
Add alternative (clang) output for unsubscribe_subscriptor
Jean-Paul Pelteret [Mon, 13 Aug 2018 10:24:59 +0000 (12:24 +0200)]
Merge pull request #7054 from masterleinad/fix_changelog_entry
Fix typo in changelog entry
Daniel Arndt [Mon, 13 Aug 2018 10:17:41 +0000 (12:17 +0200)]
Fix changelog entry
Timo Heister [Sun, 12 Aug 2018 15:14:32 +0000 (17:14 +0200)]
Merge pull request #7047 from masterleinad/subscriptor_decrease_counters
Decrement counters in Subscriptor only upon success
Daniel Arndt [Sun, 12 Aug 2018 10:44:38 +0000 (12:44 +0200)]
Deprecate MGTools::extract_non_interface_dofs
David Wells [Sun, 29 Jul 2018 20:31:03 +0000 (16:31 -0400)]
Clean up GridGenerator::extrude_triangulation.
1. Rename 's' to subcell_data.
2. Use a range for loop over the cells
3. Use more descriptive names for loop indices (e.g., vertex_n instead of i)
4. Copy a point directly instead of componentwise
5. Use the range-for loops for cells and faces.
6. Rewrite the first paragraph of the documentation to be clearer.
Daniel Arndt [Sat, 11 Aug 2018 19:35:26 +0000 (21:35 +0200)]
Decrease counters in Subscriptor only upon success
David Wells [Sat, 11 Aug 2018 15:03:19 +0000 (11:03 -0400)]
Merge pull request #7045 from Rombur/cuda_cmake
Fix build system when using CMake 3.12
Martin Kronbichler [Sat, 11 Aug 2018 10:08:07 +0000 (12:08 +0200)]
Merge pull request #7044 from BenBrands/remove_bug_ScaLAPACK_save_load
ScaLAPACKMatrix: remove small bug in parallel save/load
Martin Kronbichler [Sat, 11 Aug 2018 10:07:36 +0000 (12:07 +0200)]
Merge pull request #7036 from Rombur/cuda_constraints
Cuda constraints for matrix-free
Bruno Turcksin [Sat, 11 Aug 2018 03:26:11 +0000 (23:26 -0400)]
Fix build system when using CMake 3.12
Bruno Turcksin [Mon, 6 Aug 2018 13:54:32 +0000 (09:54 -0400)]
Add changelog entry
Bruno Turcksin [Mon, 6 Aug 2018 13:42:14 +0000 (09:42 -0400)]
Add tests
Bruno Turcksin [Mon, 6 Aug 2018 13:42:05 +0000 (09:42 -0400)]
Add support for hanging nodes
Benjamin Brands [Fri, 10 Aug 2018 05:33:08 +0000 (07:33 +0200)]
add changelog
Benjamin Brands [Thu, 9 Aug 2018 10:59:31 +0000 (12:59 +0200)]
remove inconsistency/bug in cases of serial save/ parallel load and vice versa
David Wells [Wed, 8 Aug 2018 21:07:48 +0000 (17:07 -0400)]
Merge pull request #7013 from Rombur/cuda_kernels
Move cuda kernels to their own file and namespace
Bruno Turcksin [Fri, 29 Jun 2018 19:06:56 +0000 (19:06 +0000)]
Move cuda kernels to their own file and namespace.
Move kernels from cuda_vector to their own file, remove reference to Vector, and
add a couple of new kernels.
Bruno Turcksin [Wed, 8 Aug 2018 14:42:06 +0000 (14:42 +0000)]
Fix typo and wrong formula in the documentation
Daniel Arndt [Wed, 8 Aug 2018 14:47:50 +0000 (16:47 +0200)]
Merge pull request #7043 from tjhei/doc-fix-tagfiles
doc: fix formatting of pre block
Timo Heister [Wed, 8 Aug 2018 13:47:51 +0000 (15:47 +0200)]
doc: fix formatting of pre block
Daniel Arndt [Wed, 8 Aug 2018 12:51:33 +0000 (14:51 +0200)]
Merge pull request #7031 from drwells/add-active_face_iterators
Add Triangulation::active_face_iterators().
Daniel Arndt [Tue, 7 Aug 2018 22:24:08 +0000 (00:24 +0200)]
Merge pull request #7027 from masterleinad/check-inconsistent-manifold-ids
Check for consistent manifold ids in 2D when creating a Triangulation
Daniel Arndt [Tue, 7 Aug 2018 20:44:02 +0000 (22:44 +0200)]
Changelog entry
Daniel Arndt [Tue, 7 Aug 2018 20:41:56 +0000 (22:41 +0200)]
Update a comment in GridGenerator::extract_boundary_mesh
Martin Kronbichler [Tue, 7 Aug 2018 14:56:18 +0000 (16:56 +0200)]
Merge pull request #7030 from drwells/add-initlog-to-grid-tests
Add initlog to grid tests
Martin Kronbichler [Tue, 7 Aug 2018 14:18:44 +0000 (16:18 +0200)]
Merge pull request #7034 from kronbichler/mf_transform_q_points_to_basis
Add transformation function between bases in matrix-free
David Wells [Tue, 7 Aug 2018 13:59:55 +0000 (09:59 -0400)]
Merge pull request #7038 from drwells/fix-c++17-special-math-functions
Make C++17 support independent of the special math functions.
Martin Kronbichler [Tue, 7 Aug 2018 11:23:59 +0000 (13:23 +0200)]
Remove old logfile from tests.
Martin Kronbichler [Mon, 6 Aug 2018 15:52:35 +0000 (17:52 +0200)]
Changelog.
Martin Kronbichler [Mon, 6 Aug 2018 15:52:22 +0000 (17:52 +0200)]
Test case.
Martin Kronbichler [Mon, 6 Aug 2018 15:52:12 +0000 (17:52 +0200)]
Add transform_from_q_points_to_basis in inverse mass matrix.
Daniel Arndt [Tue, 7 Aug 2018 08:32:23 +0000 (10:32 +0200)]
Fix preprocessor directive
Daniel Arndt [Tue, 7 Aug 2018 07:20:11 +0000 (09:20 +0200)]
Add comments to GridIn::read_ucd
David Wells [Tue, 7 Aug 2018 03:02:10 +0000 (23:02 -0400)]
Make C++17 support independent of the special math functions.
Clang does not yet have these, but supports a large subset of C++17. Since these
are a library feature its easy enough to just fall back to boost.
David Wells [Mon, 6 Aug 2018 19:30:19 +0000 (15:30 -0400)]
Merge pull request #7035 from dangars/camelcase_namespace
Use CamelCase in namespaces
Daniel Arndt [Mon, 6 Aug 2018 18:11:55 +0000 (20:11 +0200)]
Simplify check for boundary ids
Daniel Garcia-Sanchez [Mon, 6 Aug 2018 18:11:32 +0000 (20:11 +0200)]
Use CamelCase in namespaces
Bruno Turcksin [Mon, 6 Aug 2018 13:54:19 +0000 (09:54 -0400)]
Do not instantiate matrix-free if dim=1
David Wells [Sun, 5 Aug 2018 23:40:06 +0000 (19:40 -0400)]
Use initlog() in grid/ tests.
David Wells [Sun, 5 Aug 2018 23:37:42 +0000 (19:37 -0400)]
Add another argument to LogStream::attach.
Some tests that do not use initlog() have output that is formatted in a
different way than the one set up by the default flags.
David Wells [Mon, 6 Aug 2018 00:50:12 +0000 (20:50 -0400)]
Add Triangulation::active_face_iterators().
This function is like Triangulation::active_cell_iterators(), but for faces.
Martin Kronbichler [Sun, 5 Aug 2018 20:07:10 +0000 (22:07 +0200)]
Merge pull request #7028 from masterleinad/cell_categorization_sse2
Reduce output of cell_categorization_02
Denis Davydov [Sun, 5 Aug 2018 19:27:21 +0000 (21:27 +0200)]
Merge pull request #6241 from BenBrands/copy_functions
ScaLAPACKMatrix: copy functions
Daniel Arndt [Sun, 5 Aug 2018 19:05:11 +0000 (21:05 +0200)]
Merge pull request #6904 from GrahamBenHarper/fe_br
Implement The Bernardi-Raugel Finite Element
Daniel Arndt [Sat, 4 Aug 2018 23:10:42 +0000 (01:10 +0200)]
Reduce output of cell_categorization_02
Martin Kronbichler [Sun, 5 Aug 2018 11:05:30 +0000 (13:05 +0200)]
Merge pull request #7029 from masterleinad/remove_unused_la_d_v_import
Remove unused variables in LA::d::Vector::import
Daniel Arndt [Sun, 5 Aug 2018 09:28:34 +0000 (11:28 +0200)]
Unify error messages in Triangulation::create_triangulation
Daniel Arndt [Sun, 5 Aug 2018 09:01:10 +0000 (11:01 +0200)]
Remove unused variables in LA::d::Vector::import
Daniel Arndt [Sat, 4 Aug 2018 16:41:29 +0000 (18:41 +0200)]
Check for consistent manifold ids in 2D when creating a Triangulation
Daniel Arndt [Sat, 4 Aug 2018 16:51:51 +0000 (18:51 +0200)]
Replace invalid_manifold_id by flat_manifold_id
Jean-Paul Pelteret [Sat, 4 Aug 2018 15:22:58 +0000 (17:22 +0200)]
Merge pull request #7026 from bangerth/fix-coding-conventions
Fix formatting of our coding conventions.
Denis Davydov [Sat, 4 Aug 2018 04:01:20 +0000 (06:01 +0200)]
Merge pull request #7025 from bangerth/add-reference
Make an argument of GridGenerator::merge_triangulations() a reference.
Daniel Arndt [Fri, 3 Aug 2018 23:23:35 +0000 (01:23 +0200)]
More improvements to coding_conventions.h
Daniel Arndt [Fri, 3 Aug 2018 22:23:51 +0000 (00:23 +0200)]
Merge pull request #7002 from marcfehling/vardata-particle
Applies variable size transfer to 'ParticleHandler'.
Wolfgang Bangerth [Fri, 3 Aug 2018 22:10:12 +0000 (16:10 -0600)]
Fix formatting of our coding conventions.
This file used to be a stand-alone HTML file but is now run through
doxygen. We need to use different markup for that.
Marc Fehling [Fri, 6 Jul 2018 21:11:33 +0000 (15:11 -0600)]
Applies variable size transfer to 'ParticleHandler'.
Wolfgang Bangerth [Fri, 3 Aug 2018 16:35:43 +0000 (10:35 -0600)]
Make an argument a reference.
Denis Davydov [Fri, 3 Aug 2018 15:59:04 +0000 (17:59 +0200)]
Merge pull request #7012 from masterleinad/merge_multiple
Allow merging multiple Triangulation objects at once
Martin Kronbichler [Fri, 3 Aug 2018 11:13:07 +0000 (13:13 +0200)]
Merge pull request #7024 from masterleinad/test_higher_order_vtu
Provide test for higher order VTU output
Daniel Arndt [Fri, 3 Aug 2018 08:32:58 +0000 (10:32 +0200)]
Fix warning regarding unused parameter
Daniel Arndt [Fri, 3 Aug 2018 08:32:29 +0000 (10:32 +0200)]
Provide test for higher order VTU output
Daniel Arndt [Fri, 3 Aug 2018 07:11:40 +0000 (09:11 +0200)]
Merge pull request #6900 from drwells/update-step-9
Update step 9
Martin Kronbichler [Fri, 3 Aug 2018 07:07:40 +0000 (09:07 +0200)]
Merge pull request #7022 from Rombur/import
Fix bug in import in LinearAlgebra::distributed::Vector
Jean-Paul Pelteret [Fri, 3 Aug 2018 06:36:57 +0000 (08:36 +0200)]
Merge pull request #7015 from agrayver/high_order_vtu
Add support for high-order VTU output
David Wells [Sun, 15 Jul 2018 16:37:23 +0000 (12:37 -0400)]
step-9: Modify the refinement cycle and linear solver.
1. The different refinement strategy seems to work better with the higher degree
finite element.
2. Bicgstab has trouble converging (and does not converge in the max norm) for
finer grids. GMRES is more consistent for this problem.
3. I added some extra console output to show the residual (which does not always
converge with Bicgstab) and the number of DoFs.
David Wells [Thu, 12 Jul 2018 13:19:08 +0000 (09:19 -0400)]
step-9: Label terms in the weak form.
David Wells [Sun, 8 Jul 2018 19:17:53 +0000 (15:17 -0400)]
step-9: Redo the graphical output.
David Wells [Sun, 8 Jul 2018 18:22:40 +0000 (14:22 -0400)]
step-9: Declare the class after the data.
This fixes compilation.
David Wells [Sun, 8 Jul 2018 17:59:25 +0000 (13:59 -0400)]
step-9: Use a higher degree finite element.
This improves the resolution of the solution without requiring millions of
DoFs. I also added more output patches.
David Wells [Sun, 8 Jul 2018 17:39:38 +0000 (13:39 -0400)]
step-9: Overhaul the handling of Workstream objects.
1. Use AffineConstraints::distribute_local_to_global: This was already
introduced in step-6.
2. Move some objects into AssemblyScratchData instead of recreating them on the
stack during function calls.
3. Move some objects into EstimateScratchData (same reasons as above).
David Wells [Sun, 8 Jul 2018 17:31:14 +0000 (13:31 -0400)]
step-9: Make the implementation match the documentation.
There was a disconnect between what was described in the introduction and the
actual class implementation.
David Wells [Sun, 8 Jul 2018 03:14:57 +0000 (23:14 -0400)]
step-9: do some simple modernizations.
1. Use AffineConstraints<double>, not ConstraintMatrix.
2. Use lambdas instead of std::functions.
3. Use auto in a few places instead of very long type names.
4. Use range-based for loops.
David Wells [Sun, 15 Jul 2018 16:45:34 +0000 (12:45 -0400)]
step-9: Improve the documentation and typography.
1. Remove an inaccurate remark on SUPG: The gradient of test functions scales
like 1/h, so scaling it by h makes the term O(1), not O(h).
2. General improvements: Reword some sentences and fix some typos.
3. Remark on the choice of diffusion coefficient.
4. Notate floating point numbers as, e.g., '0.'.
5. Clarify text on TensorFunction. AFAIK this class never returned anything but
Tensors so this documentation was incorrect.
6. Use face_n instead of face or face_no to indicate the face number.
7. Move the Tensor projected_gradient closer to where it is actually used and
remove a remark about allocating it.
8. Mark an unchanged Tensor (gradient) as const.
9. Prefer p.norm() to std::sqrt(p.norm_square()).
10. Add 'using namespace dealii;' to improve the doxygen result. Since we strip
all preceding dealii namespace qualifiers this prevents doxygen from writing
::MultithreadInfo::set_thread_limit();
which looks odd.
11. Clarify that Assert aborts (it does not terminate).
David Wells [Sun, 8 Jul 2018 02:16:47 +0000 (22:16 -0400)]
step-9: Remove some unneeded code.
1. The default constructor destroys the DoFHandler correctly already, so we can
get rid of the explicit declaration.
2. Remove explicit value_list overrides: Modern compilers are good enough at
devirtualizing that this optimization is no longer necessary.
3. Remove some empty constructors.
Daniel Arndt [Thu, 2 Aug 2018 21:10:42 +0000 (23:10 +0200)]
Merge pull request #7017 from davydden/bugfix/cell_categorization_mg
fix cell_vectorization_category with MG
Bruno Turcksin [Thu, 2 Aug 2018 15:50:30 +0000 (11:50 -0400)]
Add changelog entry
Bruno Turcksin [Wed, 1 Aug 2018 18:09:51 +0000 (18:09 +0000)]
Fix a bug in distributed::Vector::import
The import function would fail if the ReadWriteVector would only contains
ghosted values
Denis Davydov [Wed, 1 Aug 2018 22:17:48 +0000 (00:17 +0200)]
minor fix of matrix-free cell categorization on MG levels
Daniel Arndt [Thu, 2 Aug 2018 15:18:10 +0000 (17:18 +0200)]
Merge pull request #7019 from masterleinad/common_type_comparison
Use common type in AssertIndexRange for the comparison
Martin Kronbichler [Thu, 2 Aug 2018 12:04:12 +0000 (14:04 +0200)]
Merge pull request #7016 from masterleinad/fix_restrictions_warnings_tests
Fix restrictions and warnings for some tests
Martin Kronbichler [Thu, 2 Aug 2018 11:49:13 +0000 (13:49 +0200)]
Merge pull request #7018 from masterleinad/higher_order_cell_docu
Fix doxygen documentation for DataOutBase::VtkFlags::write_higher_order_cells
Daniel Arndt [Thu, 2 Aug 2018 09:04:43 +0000 (11:04 +0200)]
Use common type in AssertIndexRange for the comparison
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.