]>
https://gitweb.dealii.org/ - dealii.git/log
Martin Kronbichler [Fri, 13 Jan 2017 17:47:45 +0000 (18:47 +0100)]
Cleanup step-37 regarding shared_ptr usage.
Martin Kronbichler [Fri, 13 Jan 2017 17:47:18 +0000 (18:47 +0100)]
Disable warning about deprecated functionality.
Martin Kronbichler [Fri, 13 Jan 2017 17:47:02 +0000 (18:47 +0100)]
Add additional access method to the underlying shared ptr
Bruno Turcksin [Fri, 13 Jan 2017 13:44:47 +0000 (08:44 -0500)]
Merge pull request #3785 from tamiko/cuda_fixes_3
CMake: Various fixes for the cuda configuration
Matthias Maier [Fri, 13 Jan 2017 11:13:47 +0000 (05:13 -0600)]
CMake: Remove obsolete file
Daniel Arndt [Fri, 13 Jan 2017 10:14:41 +0000 (11:14 +0100)]
Merge pull request #3763 from elddle/Fix-get_ria_vector
Fix a get_ria_vector for FE_BDM
Matthias Maier [Fri, 13 Jan 2017 09:35:49 +0000 (03:35 -0600)]
Merge pull request #3761 from jppelteret/symm_tensor_contract_3
Add contract3() function with SymmetricTensor arguments.
Karl Ljungkvist [Fri, 13 Jan 2017 08:14:02 +0000 (09:14 +0100)]
remove forgotten dot
Matthias Maier [Thu, 12 Jan 2017 17:03:13 +0000 (11:03 -0600)]
add a changes file
Matthias Maier [Thu, 12 Jan 2017 17:00:30 +0000 (11:00 -0600)]
CMake: Make sure the cuda wrapper only uses -std=c++11
Matthias Maier [Thu, 12 Jan 2017 16:54:39 +0000 (10:54 -0600)]
CMake: Properly propagate all compiler flags to the cuda wrapper
Matthias Maier [Thu, 12 Jan 2017 16:34:42 +0000 (10:34 -0600)]
CMake: Rename all object targets to obj_<foo>_(debug|release)
This change is necessary to please the cuda wrapper nvcc that insists on
exporting the target name as preprocessor definition. Unfortunately, a
period "." is an invalid character and we end up with warnings of the
form
"missing whitespace after the macro name"
Fix this by renaming all object targets to only contain underscores.
Matthias Maier [Thu, 12 Jan 2017 16:24:07 +0000 (10:24 -0600)]
CMake: Avoid bogus preprocessor definitions
The macro CUDA_WARP_SRCS expects the target to have a set of
preprocessor definitions that does not contain an empty element.
Matthias Maier [Thu, 12 Jan 2017 15:46:50 +0000 (09:46 -0600)]
CMake: Fix a typo
Martin Kronbichler [Fri, 13 Jan 2017 08:01:34 +0000 (09:01 +0100)]
Merge pull request #3764 from Rombur/shared_ptr
Use shared_ptr to store MatrixFree data in MatrixFreeOperators::Base
Daniel Arndt [Thu, 12 Jan 2017 23:12:16 +0000 (00:12 +0100)]
Merge pull request #3778 from tamiko/add_a_boost_check
CMake: Define BOOST_NO_CXX11_HDR_UNORDERED_MAP for bundled boost if <gcc-4.8
Jean-Paul Pelteret [Thu, 12 Jan 2017 22:50:05 +0000 (23:50 +0100)]
Merge pull request #3777 from jppelteret/step-18-update_results
Update results for step-18
Jean-Paul Pelteret [Thu, 12 Jan 2017 13:06:26 +0000 (14:06 +0100)]
Update results for step-18
fixes #3757
Matthias Maier [Thu, 12 Jan 2017 22:20:22 +0000 (16:20 -0600)]
CMake: Refactor macros and augment comment
Matthias Maier [Thu, 12 Jan 2017 13:26:27 +0000 (07:26 -0600)]
CMake: Define BOOST_NO_CXX11_HDR_UNORDERED_MAP for bundled boost if <gcc-4.8
Closes #3624.
Wolfgang Bangerth [Thu, 12 Jan 2017 21:47:55 +0000 (14:47 -0700)]
Merge pull request #3779 from tamiko/cuda_fixes_1
CMake: let *_cuda targets depend on *.inst targets
elddle [Thu, 12 Jan 2017 16:46:08 +0000 (11:46 -0500)]
Fix get_ria_vector for BDM
David Wells [Thu, 12 Jan 2017 16:45:30 +0000 (11:45 -0500)]
Merge pull request #3783 from tjhei/fix_compressed_fe_index_is_active
fix hp fe_index_is_active()
Wolfgang Bangerth [Thu, 12 Jan 2017 16:40:25 +0000 (09:40 -0700)]
Merge pull request #3780 from drwells/doxygen-link-fixes
Doxygen link fixes
Wolfgang Bangerth [Thu, 12 Jan 2017 16:10:16 +0000 (09:10 -0700)]
Merge pull request #3781 from drwells/analogous-grammar-fix
Write 'is analogous to', not 'is analogy to'.
David Wells [Thu, 12 Jan 2017 15:08:20 +0000 (10:08 -0500)]
Fix some dead UNV links.
Timo Heister [Thu, 12 Jan 2017 15:50:43 +0000 (10:50 -0500)]
fix hp fe_index_is_active()
David Wells [Thu, 12 Jan 2017 15:23:53 +0000 (10:23 -0500)]
Write 'is analogous to', not 'is analogy to'.
Jean-Paul Pelteret [Tue, 10 Jan 2017 16:21:00 +0000 (17:21 +0100)]
Documentation update
Change ".." to "\ldots" in tensor and symmetric tensor documentation
Jean-Paul Pelteret [Tue, 10 Jan 2017 15:10:26 +0000 (16:10 +0100)]
Add contract3() function with SymmetricTensor arguments.
fixes #2691
Timo Heister [Thu, 12 Jan 2017 15:12:13 +0000 (10:12 -0500)]
Merge pull request #3776 from tamiko/fix_spaces_testsuite
CMake: Simplify run_test.sh and avoid array handling magic
David Wells [Thu, 12 Jan 2017 15:01:58 +0000 (10:01 -0500)]
Fix doxygen formatting of URLs with hyphens.
The script wrapcomments.py was previously allowed to split lines after hyphens
which results in broken links.
Matthias Maier [Thu, 12 Jan 2017 15:06:48 +0000 (09:06 -0600)]
CMake: let *_cuda targets depend on *.inst targets
Bruno Turcksin [Tue, 10 Jan 2017 23:08:04 +0000 (18:08 -0500)]
Use shared_ptr to store MatrixFree data in MatrixFreeOperators::Base
Matthias Maier [Thu, 12 Jan 2017 11:47:55 +0000 (05:47 -0600)]
CMake: Update DEAL_II_ADD_TEST macro
Matthias Maier [Thu, 12 Jan 2017 11:41:43 +0000 (05:41 -0600)]
CMake: Simplify run_test.sh and avoid array handling magic
Matthias Maier [Thu, 12 Jan 2017 11:43:12 +0000 (05:43 -0600)]
Merge pull request #3775 from kronbichler/fix_test_1
Fix test output
Martin Kronbichler [Thu, 12 Jan 2017 07:56:29 +0000 (08:56 +0100)]
Fix test output
Martin Kronbichler [Thu, 12 Jan 2017 05:58:36 +0000 (06:58 +0100)]
Merge pull request #3773 from kronbichler/mg_transfer_mf_legendre
Allow for FE_DGQLegendre in MGTransferMatrixFree
Martin Kronbichler [Thu, 12 Jan 2017 05:48:50 +0000 (06:48 +0100)]
Merge pull request #3771 from bangerth/fix-face-domination
Fix face domination between FE_Q and FE_DGQ.
Wolfgang Bangerth [Wed, 11 Jan 2017 20:19:50 +0000 (13:19 -0700)]
Add changelog entry.
Wolfgang Bangerth [Wed, 11 Jan 2017 20:19:47 +0000 (13:19 -0700)]
Add test.
Wolfgang Bangerth [Wed, 11 Jan 2017 20:18:32 +0000 (13:18 -0700)]
Fix face domination between FE_Q and FE_DGQ.
Right now, one got an answer if one asked the FE_DGQ whether or not
it dominates the FE_Q, but not the other way around. Make this symmetric.
While there, also include FE_DGP in the test.
Daniel Arndt [Thu, 12 Jan 2017 00:40:43 +0000 (01:40 +0100)]
Merge pull request #3770 from kronbichler/introduce_fe_dgq_legendre
Introduce FE_DGQLegendre and FE_DGQHermite
David Wells [Thu, 12 Jan 2017 00:09:45 +0000 (19:09 -0500)]
Merge pull request #3772 from bangerth/fix-type-mismatch
Fix missing 'spacedim' template argument.
Daniel Arndt [Wed, 11 Jan 2017 22:15:06 +0000 (23:15 +0100)]
Merge pull request #3768 from masterleinad/test_fe_mf
Implement function to check for MatrixFree support
Martin Kronbichler [Wed, 11 Jan 2017 22:04:15 +0000 (23:04 +0100)]
Merge pull request #3769 from kronbichler/fix_parallel_vector_comm
Change communication pattern in parallel vector.
Martin Kronbichler [Wed, 11 Jan 2017 21:58:42 +0000 (22:58 +0100)]
Allow for FE_DGQLegendre in MGTransferMatrixFree
Martin Kronbichler [Wed, 11 Jan 2017 18:14:45 +0000 (19:14 +0100)]
List incompatible change
Martin Kronbichler [Wed, 11 Jan 2017 18:14:34 +0000 (19:14 +0100)]
Add another test
Martin Kronbichler [Wed, 11 Jan 2017 17:49:02 +0000 (18:49 +0100)]
Introduce FE_DGQLegendre and FE_DGQHermite
Wolfgang Bangerth [Wed, 11 Jan 2017 20:27:08 +0000 (13:27 -0700)]
Add changelog entry.
Wolfgang Bangerth [Wed, 11 Jan 2017 20:26:56 +0000 (13:26 -0700)]
Fix missing 'spacedim' template argument.
Bruno Turcksin [Wed, 11 Jan 2017 19:55:17 +0000 (14:55 -0500)]
Merge pull request #3755 from jppelteret/fix_step-18_rotation_matrix
Fix rotation matrix definition in step-18
Daniel Arndt [Wed, 11 Jan 2017 17:40:21 +0000 (18:40 +0100)]
Merge pull request #3767 from kronbichler/chebyshev_common_vector_interface
Use common vector interface for Chebyshev smoother
Daniel Arndt [Mon, 9 Jan 2017 13:44:33 +0000 (14:44 +0100)]
Implement function to check for MatrixFree support
Martin Kronbichler [Wed, 11 Jan 2017 17:29:33 +0000 (18:29 +0100)]
Change communication pattern in parallel vector.
Jean-Paul Pelteret [Mon, 9 Jan 2017 18:15:08 +0000 (19:15 +0100)]
Update step-18 to use rotation matrix defined in physics module.
Jean-Paul Pelteret [Mon, 9 Jan 2017 17:06:18 +0000 (18:06 +0100)]
Added functions to compute rotation matrices in 2d and 3d.
Martin Kronbichler [Wed, 11 Jan 2017 14:58:28 +0000 (15:58 +0100)]
Use common vector interface for Chebyshev smoother
David Wells [Wed, 11 Jan 2017 13:48:54 +0000 (08:48 -0500)]
Merge pull request #3754 from bangerth/fix-abf-polynomials-pointer
Do not use a plain pointer in PolynomialsABF.
Wolfgang Bangerth [Tue, 10 Jan 2017 02:47:48 +0000 (19:47 -0700)]
Add changelog entry.
Wolfgang Bangerth [Mon, 9 Jan 2017 15:52:15 +0000 (08:52 -0700)]
Do not use a plain pointer in PolynomialsABF.
These objects are being copied by the FE_Poly* classes, so plain pointers without
dedicated copy constructors and operators are likely going to lead to memory
corruption. It's not clear to me how this ever worked, but it's easy to fix.
Bruno Turcksin [Tue, 10 Jan 2017 18:52:42 +0000 (13:52 -0500)]
Merge pull request #3762 from drwells/make-member-const
Make a member variable const.
David Wells [Tue, 10 Jan 2017 16:10:38 +0000 (11:10 -0500)]
Make a member variable const.
This is copied from an input parameter and never changed so it may as
well be marked as const.
Wolfgang Bangerth [Tue, 10 Jan 2017 13:36:23 +0000 (06:36 -0700)]
Merge pull request #3758 from Rombur/reinit
Do not use deprecated reinit function.
Martin Kronbichler [Tue, 10 Jan 2017 09:45:45 +0000 (10:45 +0100)]
Merge pull request #3760 from masterleinad/fix_filename
Rename misnamed file
Daniel Arndt [Tue, 10 Jan 2017 09:43:17 +0000 (10:43 +0100)]
Rename misnamed file
Bruno Turcksin [Tue, 10 Jan 2017 02:21:15 +0000 (21:21 -0500)]
Do not use deprecated reinit function.
Bruno Turcksin [Tue, 10 Jan 2017 02:20:20 +0000 (21:20 -0500)]
Merge pull request #3752 from bangerth/fix-abf-in-3d
Fix PolynomialsABF::compute_n_pols() in 3d.
Daniel Arndt [Mon, 9 Jan 2017 20:42:31 +0000 (21:42 +0100)]
Merge pull request #3749 from jppelteret/fix_tests_01
Redo output for physics/step-44* tests
Daniel Arndt [Mon, 9 Jan 2017 20:41:36 +0000 (21:41 +0100)]
Merge pull request #3735 from guidokanschat/tangential_laplace
tangential jumps for Laplacian
Daniel Arndt [Mon, 9 Jan 2017 20:36:36 +0000 (21:36 +0100)]
Merge pull request #3753 from bangerth/make-var-const
Make a member variable 'const'.
Guido Kanschat [Mon, 9 Jan 2017 06:15:35 +0000 (07:15 +0100)]
be a bit more explicit
Martin Kronbichler [Mon, 9 Jan 2017 16:53:12 +0000 (17:53 +0100)]
Merge pull request #3691 from drwells/constraint-petsc-vector-fast-path
Add a faster code path for PETSc vectors.
Wolfgang Bangerth [Mon, 9 Jan 2017 15:16:54 +0000 (08:16 -0700)]
Make a member variable 'const'.
We can do so because it is initialized in the constructor.
Wolfgang Bangerth [Mon, 9 Jan 2017 15:10:28 +0000 (08:10 -0700)]
Add changelog entry.
David Wells [Wed, 21 Dec 2016 17:28:48 +0000 (12:28 -0500)]
Bulk write vector values in distribute_local_to_global.
This patch lowers the total number of calls to VecSetValues in
ConstraintMatrix::distribute_local_to_global by writing new vector values to a
temporary array before sending them off to PETSc. There are no notable
performance improvements for the other vector classes (though this certainly
should not hurt).
I used step-40 to check that the performance improved. Here is the (filtered)
output with this patch:
[drwells@archway step-40]$ make run | grep assembly
| assembly | 1 | 0.00966s | 8.7% |
| assembly | 1 | 0.0122s | 6.4% |
| assembly | 1 | 0.0217s | 5.9% |
| assembly | 1 | 0.0431s | 5.6% |
| assembly | 1 | 0.0848s | 5.6% |
| assembly | 1 | 0.173s | 5.6% |
| assembly | 1 | 0.351s | 5.5% |
| assembly | 1 | 0.729s | 5.5% |
and on master:
[drwells@archway step-40]$ make run | grep assembly
| assembly | 1 | 0.00672s | 8.9% |
| assembly | 1 | 0.0133s | 7.2% |
| assembly | 1 | 0.0258s | 6.9% |
| assembly | 1 | 0.0513s | 6.8% |
| assembly | 1 | 0.101s | 6.6% |
| assembly | 1 | 0.203s | 6.5% |
| assembly | 1 | 0.414s | 6.4% |
| assembly | 1 | 0.93s | 6.9% |
so it appears that the whole assembly process takes about 10% less time.
David Wells [Wed, 28 Dec 2016 17:39:22 +0000 (12:39 -0500)]
Add an 'add' method to the native parallel vectors.
This is necessary for an upcoming patch that uses the method inside
ConstraintMatrix::distribute_local_to_global.
Martin Kronbichler [Mon, 9 Jan 2017 11:22:52 +0000 (12:22 +0100)]
Merge pull request #3750 from masterleinad/matrix_free_comm
Automatically set MPI_Comm in MatrixFree
Daniel Arndt [Sun, 8 Jan 2017 21:56:29 +0000 (22:56 +0100)]
Automatically set MPI_Comm in MatrixFree
Martin Kronbichler [Mon, 9 Jan 2017 07:27:39 +0000 (08:27 +0100)]
Merge pull request #3740 from davydden/doc/debug_Og
add a note on -Og and -O0 debug flags
Martin Kronbichler [Mon, 9 Jan 2017 07:21:51 +0000 (08:21 +0100)]
Merge pull request #3748 from Rombur/vector_operations
Encapsulate functions in VectorOperations namespace
Martin Kronbichler [Mon, 9 Jan 2017 07:21:20 +0000 (08:21 +0100)]
Merge pull request #3745 from tamiko/fix_spaces_testsuite
Bugfix: Fix up a previous attempt to quote correctly
Wolfgang Bangerth [Mon, 9 Jan 2017 04:28:25 +0000 (21:28 -0700)]
Fix the computation of the number of polynomials the ABF element has.
Matthias Maier [Sun, 8 Jan 2017 23:10:25 +0000 (17:10 -0600)]
Bugfix: Quote in a posix compliant manner
Matthias Maier [Sun, 8 Jan 2017 16:51:12 +0000 (10:51 -0600)]
Bugfix: Fix up previous attempt to quote correctly
In reference to #3742
Jean-Paul Pelteret [Sun, 8 Jan 2017 21:36:37 +0000 (22:36 +0100)]
Redo output for physics/step-44* tests
Instead of checking the convergence rates for each Newton iteration
against some benchmark result that is machine specific, we now monitor
the solution at a fixed point. This should remain unchanged if the
problem converges correctly for each timestep. The nonlinear solver
output is now send to the file "stdout".
Fixes #3747
Bruno Turcksin [Sun, 8 Jan 2017 20:26:44 +0000 (15:26 -0500)]
Encapsulate functions from vector_operations_internal in VectorOperations namespace.
Bruno Turcksin [Sun, 8 Jan 2017 18:48:15 +0000 (13:48 -0500)]
Merge pull request #3739 from bangerth/fix-hp-indices-compression
Fix and enable compression of hp DoF indices.
Wolfgang Bangerth [Sun, 8 Jan 2017 18:15:02 +0000 (11:15 -0700)]
Merge pull request #3746 from drwells/add-higher-mapping-test
Add a test checking ChartManifold/MappingQ convergence rates.
David Wells [Sun, 8 Jan 2017 17:49:21 +0000 (12:49 -0500)]
Add a test checking ChartManifold/MappingQ convergence rates.
Wolfgang Bangerth [Sun, 8 Jan 2017 15:47:22 +0000 (08:47 -0700)]
Merge pull request #3738 from drwells/more-precise-cell-similarity
More precise cell similarity
David Wells [Sun, 8 Jan 2017 01:26:59 +0000 (20:26 -0500)]
Do not reuse CellSimilarity for higher order MappingQGeneric maps.
This commit fixes bug where two cells, both alike in vertices (but varying in
curvature) would sometimes be treated as translations of each-other by
MappingQGeneric. This lead to an artificial limit of second order convergence
since Jacobian weights (and other values) were not correctly updated.
Wolfgang Bangerth [Fri, 6 Jan 2017 19:07:49 +0000 (12:07 -0700)]
Add a changelog entry.
Wolfgang Bangerth [Fri, 6 Jan 2017 15:14:03 +0000 (08:14 -0700)]
Fix and enable compression of hp DoF indices.
A previous patch, a long time ago, implemented a scheme to compress the data structure
that stores the DoF indices in the hp case. For reasons no longer clear, the functions
that do the compression and decompression just returned immediately, without actually
doing anything. Maybe not surprisingly, the algorithm -- not executed after all -- did
not work.
This patch fixes the algorithm at various locations and enables the compression.
Wolfgang Bangerth [Sun, 8 Jan 2017 00:33:11 +0000 (17:33 -0700)]
Merge pull request #3742 from tamiko/fix_spaces_testsuite
Bugfix: Quote paths in run_test.sh
Wolfgang Bangerth [Sat, 7 Jan 2017 23:25:22 +0000 (16:25 -0700)]
Merge pull request #3744 from drwells/add-parallel-block-vector-file
Add another output file for a test.
David Wells [Sat, 7 Jan 2017 04:36:38 +0000 (23:36 -0500)]
Add more output files for a test.
This test checks memory consumption so it is reasonable to have a few of these.
Guido Kanschat [Thu, 5 Jan 2017 17:07:23 +0000 (18:07 +0100)]
tangential jumps for Laplacian
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.