]>
https://gitweb.dealii.org/ - dealii.git/log
Denis Davydov [Wed, 7 Feb 2018 13:24:37 +0000 (14:24 +0100)]
extend VectorTools::project() to take fe_component
Denis Davydov [Tue, 23 Jan 2018 12:03:40 +0000 (13:03 +0100)]
Merge pull request #5780 from masterleinad/link_error_geometryinfo0
Avoid link errors in GeometryInfo<0>
Daniel Arndt [Tue, 23 Jan 2018 09:20:31 +0000 (10:20 +0100)]
Avoid link errors in GeometryInfo<0>
Wolfgang Bangerth [Mon, 22 Jan 2018 21:42:09 +0000 (14:42 -0700)]
Merge pull request #5777 from masterleinad/changelog_typo
Fix typo in changelog entry
Daniel Arndt [Mon, 22 Jan 2018 21:39:15 +0000 (22:39 +0100)]
Fix typo in changelog entry
Denis Davydov [Mon, 22 Jan 2018 18:53:24 +0000 (19:53 +0100)]
Merge pull request #5776 from bangerth/fix-step-42
step-42: Move a statement to where it belongs.
Wolfgang Bangerth [Mon, 22 Jan 2018 18:37:28 +0000 (11:37 -0700)]
Move a statement to where it belongs.
Daniel Arndt [Mon, 22 Jan 2018 17:56:42 +0000 (18:56 +0100)]
Merge pull request #5728 from masterleinad/clang-tidy-new
Be explicit about virtual functions called in con-/destructors
Wolfgang Bangerth [Mon, 22 Jan 2018 17:46:35 +0000 (10:46 -0700)]
Merge pull request #5775 from Rombur/fix_test
Require metis for a test to run
Jean-Paul Pelteret [Mon, 22 Jan 2018 14:26:01 +0000 (15:26 +0100)]
Merge pull request #5766 from bangerth/doc-update-2
Address what 'sequential' means in the context of DoFHandlers.
Jean-Paul Pelteret [Mon, 22 Jan 2018 14:13:57 +0000 (15:13 +0100)]
Merge pull request #5754 from Rombur/cuda_documentation
Make it clear that our CUDA wrappers only support float and double
Bruno Turcksin [Mon, 22 Jan 2018 14:11:08 +0000 (09:11 -0500)]
Merge pull request #5774 from davydden/bugfix_makeunique
fix make_unique
Bruno Turcksin [Mon, 22 Jan 2018 14:00:55 +0000 (09:00 -0500)]
Require metis for the test to run
Denis Davydov [Mon, 22 Jan 2018 12:48:23 +0000 (13:48 +0100)]
fix make_unique
Martin Kronbichler [Mon, 22 Jan 2018 08:15:05 +0000 (09:15 +0100)]
Merge pull request #5761 from drwells/improve-2d-transfinite-guess
Improve 2d transfinite guess
Denis Davydov [Mon, 22 Jan 2018 07:59:29 +0000 (08:59 +0100)]
Merge pull request #5734 from BenBrands/extend_scalapack_more
Extend scalapack more
Jean-Paul Pelteret [Sun, 21 Jan 2018 07:14:23 +0000 (08:14 +0100)]
Merge pull request #5772 from drwells/use-std-abs
Use std::abs instead of abs.
David Wells [Sat, 20 Jan 2018 21:52:42 +0000 (16:52 -0500)]
Use std::abs instead of abs.
On some compilers (notably GCC 4.8.4, in some configurations) 'abs' refers to
'int abs(int)', and gives the wrong answer for floating point arguments.
David Wells [Sat, 20 Jan 2018 20:48:41 +0000 (15:48 -0500)]
Merge pull request #5765 from bangerth/doc-update
Update a comment for PETScWrappers::PreconditionBlockJacobi.
David Wells [Fri, 19 Jan 2018 20:23:22 +0000 (15:23 -0500)]
Differentiate between quads in 2D and 3D.
In 3D, we may use eight surrounding points either as the eight vertices
of a hexahedron or as eight surrounding points on the surface of a
hexahedron. This patch adds some logic to tell the two cases apart based
on some angle calculations.
David Wells [Thu, 18 Jan 2018 21:43:59 +0000 (16:43 -0500)]
Guard a check with an if statement.
David Wells [Thu, 18 Jan 2018 21:42:00 +0000 (16:42 -0500)]
Make the TFI chart point calculation more robust.
If the guess computed by the shortcut is not accurate enough then we
should try to use the affine approximation instead.
David Wells [Thu, 18 Jan 2018 20:16:34 +0000 (15:16 -0500)]
Add a test for the 2D TFI change.
Denis Davydov [Sat, 20 Jan 2018 16:21:35 +0000 (17:21 +0100)]
Merge pull request #5756 from masterleinad/hdf5_quick
Provide HDF5 quicktest
Daniel Arndt [Thu, 18 Jan 2018 17:10:45 +0000 (18:10 +0100)]
Provide HDF5 quicktest
Jean-Paul Pelteret [Sat, 20 Jan 2018 09:34:22 +0000 (10:34 +0100)]
Merge pull request #5768 from bangerth/doc-update-4
Update class doc of p::s::Triangulation.
Jean-Paul Pelteret [Sat, 20 Jan 2018 09:24:19 +0000 (10:24 +0100)]
Merge pull request #5767 from bangerth/doc-update-3
Address what 'sequential' means in the context of communicators.
Wolfgang Bangerth [Sat, 20 Jan 2018 01:11:00 +0000 (18:11 -0700)]
Update class doc of p::s::Triangulation.
Wolfgang Bangerth [Sat, 20 Jan 2018 01:05:03 +0000 (18:05 -0700)]
Provide a link into the glossary.
Wolfgang Bangerth [Sat, 20 Jan 2018 01:04:32 +0000 (18:04 -0700)]
Address what 'sequential' means in the context of communicators.
Wolfgang Bangerth [Sat, 20 Jan 2018 00:59:55 +0000 (17:59 -0700)]
Address what 'sequential' means in the context of DoFHandlers.
Wolfgang Bangerth [Sat, 20 Jan 2018 00:37:19 +0000 (17:37 -0700)]
Update a comment for PETScWrappers::PreconditionBlockJacobi.
Benjamin Brands [Wed, 17 Jan 2018 12:09:03 +0000 (13:09 +0100)]
reviewer corrections
Jean-Paul Pelteret [Fri, 19 Jan 2018 17:45:36 +0000 (18:45 +0100)]
Merge pull request #5759 from bangerth/fix
Fix DoFTools::locally_owned_dofs_per_subdomain() for empty processors.
David Wells [Thu, 18 Jan 2018 18:37:43 +0000 (13:37 -0500)]
Implement a 2D-specific chart point estimate.
The current algorithm calculates a reasonable guess for the chart point
in 3D. Using this 3D algorithm in 2D results in relatively bad guesses:
for i = 5:
affine approx chart point: 0.5 3.64977e-14
pulled back chart point: 1 0.5
for i = 6:
affine approx chart point: 0.5 0.5
pulled back chart point: 0.5 8.678e-17
for i = 7:
affine approx chart point: 0.5 1
pulled back chart point: 0.5 1
This patch adds a 2D-specific algorithm that simply averages the two
adjacent vertices (e.g., chart point 5 is assumed to be an average of
chart points 1 and 3).
This issue was reported by Juan Carlos Araujo Cabarcas, who found a test
case where the affine guess was sufficiently inaccurate that
TransfiniteInterpolationManifold::pull_back failed to find the right
point. This change fixes that test case.
Martin Kronbichler [Fri, 19 Jan 2018 13:38:04 +0000 (14:38 +0100)]
Merge pull request #5760 from bangerth/add-tests
Add tests for complex-valued add_and_dot()
Bruno Turcksin [Fri, 19 Jan 2018 13:34:03 +0000 (08:34 -0500)]
Merge pull request #5750 from bangerth/fix-cuda-vector
Fix CUDAVector::add_and_dot() for complex-valued vectors.
Martin Kronbichler [Fri, 19 Jan 2018 08:18:23 +0000 (09:18 +0100)]
Merge pull request #5755 from drwells/transfinite-message
Print out more info when TFI encounters errors.
Wolfgang Bangerth [Thu, 18 Jan 2018 22:28:39 +0000 (15:28 -0700)]
Add changelog entry.
Wolfgang Bangerth [Thu, 18 Jan 2018 22:28:31 +0000 (15:28 -0700)]
Add test.
Wolfgang Bangerth [Thu, 18 Jan 2018 22:28:12 +0000 (15:28 -0700)]
Fix DoFTools::locally_owned_dofs_per_subdomain() for empty processors.
Wolfgang Bangerth [Thu, 18 Jan 2018 04:06:59 +0000 (21:06 -0700)]
Fix CUDAVector::add_and_dot() for complex-valued vectors.
Wolfgang Bangerth [Thu, 18 Jan 2018 13:54:28 +0000 (06:54 -0700)]
Mark NumberTraits::conjugate() as DEAL_II_CUDA_HOST_DEV.
Daniel Arndt [Thu, 18 Jan 2018 23:15:26 +0000 (00:15 +0100)]
Merge pull request #5757 from bangerth/add-static-asserts
Add static_asserts for complex-valued cases.
Wolfgang Bangerth [Thu, 18 Jan 2018 22:50:36 +0000 (15:50 -0700)]
Merge pull request #5742 from drwells/fe-system-dereference-check
Rearrange a pointer check to satisfy coverity.
Wolfgang Bangerth [Thu, 18 Jan 2018 22:45:56 +0000 (15:45 -0700)]
Add a test for LinearAlgebra::distributed::Vector::add_and_dot().
Wolfgang Bangerth [Thu, 18 Jan 2018 22:39:34 +0000 (15:39 -0700)]
Add a test for LinearAlgebra::Vector::add_and_dot().
Wolfgang Bangerth [Thu, 18 Jan 2018 22:33:55 +0000 (15:33 -0700)]
Add a test for Vector::add_and_dot().
Wolfgang Bangerth [Thu, 18 Jan 2018 22:33:04 +0000 (15:33 -0700)]
Add small comments.
Jean-Paul Pelteret [Thu, 18 Jan 2018 19:24:49 +0000 (20:24 +0100)]
Merge pull request #5758 from dangars/typo-doc-DataPostprocessorTensor
Typo in the documentation of DataPostprocessorTensor
Daniel Garcia-Sanchez [Thu, 18 Jan 2018 18:47:48 +0000 (19:47 +0100)]
Typo in the documentation of DataPostprocessorTensor
Wolfgang Bangerth [Thu, 18 Jan 2018 18:18:45 +0000 (11:18 -0700)]
Add static_asserts for complex-valued cases.
Wolfgang Bangerth [Thu, 18 Jan 2018 17:13:41 +0000 (10:13 -0700)]
Merge pull request #5748 from masterleinad/inline_geometry_info
Inline GeometryInfo members
David Wells [Thu, 18 Jan 2018 17:02:22 +0000 (12:02 -0500)]
Print out more info when TFI encounters errors.
the upper loop bound, c, cannot be larger than the maximum valid index
into the relevant array: therefore we should use '<=' instead of '<' in
a loop bounded above by c.
Bruno Turcksin [Thu, 18 Jan 2018 14:20:32 +0000 (09:20 -0500)]
Make it clear that our CUDA wrappers only support float and double
Daniel Arndt [Wed, 17 Jan 2018 15:20:03 +0000 (16:20 +0100)]
Inline GeometryInfo
Martin Kronbichler [Thu, 18 Jan 2018 08:25:28 +0000 (09:25 +0100)]
Merge pull request #5735 from Rombur/cmake_cusparse
Add cuSPARSE support in the build system
Martin Kronbichler [Thu, 18 Jan 2018 08:24:53 +0000 (09:24 +0100)]
Merge pull request #5737 from Rombur/cuda_warning
Fix warnings: do not use Assert in CUDA kernel
Martin Kronbichler [Thu, 18 Jan 2018 08:24:23 +0000 (09:24 +0100)]
Merge pull request #5736 from Rombur/cuh
Change .cuh extension to h
Martin Kronbichler [Thu, 18 Jan 2018 08:23:31 +0000 (09:23 +0100)]
Merge pull request #5740 from masterleinad/internal_alignedvector_const
Make parameters and variables for the internal::AlignedVector classes const
Martin Kronbichler [Thu, 18 Jan 2018 08:22:13 +0000 (09:22 +0100)]
Merge pull request #5749 from bangerth/doc-update
Update documentation of *::Vector::add_and_dot for complex vectors.
Martin Kronbichler [Thu, 18 Jan 2018 08:15:40 +0000 (09:15 +0100)]
Merge pull request #5745 from masterleinad/initialize_constexpr_arrays
Initialize constexpr arrays in GeometryInfo
Wolfgang Bangerth [Thu, 18 Jan 2018 04:04:04 +0000 (21:04 -0700)]
Wrap comments.
Wolfgang Bangerth [Thu, 18 Jan 2018 04:03:47 +0000 (21:03 -0700)]
Update documentation of *::Vector::add_and_dot for complex vectors.
Daniel Arndt [Wed, 17 Jan 2018 23:52:24 +0000 (00:52 +0100)]
Merge pull request #5744 from masterleinad/fix_coverity
Fix issues reported by Coverity
Daniel Arndt [Wed, 17 Jan 2018 15:30:12 +0000 (16:30 +0100)]
Initialize constexpr arrays in GeometryInfo
Daniel Arndt [Wed, 17 Jan 2018 13:00:09 +0000 (14:00 +0100)]
Fix issues reported by Coverity
David Wells [Tue, 16 Jan 2018 19:11:19 +0000 (14:11 -0500)]
Rearrange a pointer check to satisfy coverity.
Coverity (the static analysis program) raises an issue with the way we
check the result of a dynamic_cast. This commit rearranges things so
that we cannot enter the relevant block unless dynamic_cast succeeds,
regardless of the assertion.
Jean-Paul Pelteret [Wed, 17 Jan 2018 15:35:26 +0000 (16:35 +0100)]
Merge pull request #5743 from bangerth/doc-update
Clarify documentation of (hp::)?DoFHandler::distribute_dofs().
Wolfgang Bangerth [Tue, 16 Jan 2018 03:49:08 +0000 (20:49 -0700)]
Clarify documentation of (hp::)?DoFHandler::distribute_dofs().
Daniel Arndt [Tue, 16 Jan 2018 16:50:59 +0000 (17:50 +0100)]
Make parameters and variables for the internal::AlignedVector classes const
Denis Davydov [Tue, 16 Jan 2018 16:31:12 +0000 (17:31 +0100)]
Merge pull request #5731 from masterleinad/fix_rt_bubbles_tests
Increase precision in FE_RT_Bubbles tests
Jean-Paul Pelteret [Tue, 16 Jan 2018 06:38:00 +0000 (07:38 +0100)]
Merge pull request #5733 from niveshd/ExtendTestFE_Enriched
Minor change in FE Enriched element
Jean-Paul Pelteret [Tue, 16 Jan 2018 06:37:16 +0000 (07:37 +0100)]
Merge pull request #5738 from bangerth/doc-update
Fix a grammar error.
Wolfgang Bangerth [Tue, 16 Jan 2018 01:34:57 +0000 (18:34 -0700)]
Fix a grammar error.
Bruno Turcksin [Tue, 16 Jan 2018 00:23:03 +0000 (19:23 -0500)]
Fix warnings: do not use Assert in CUDA kernel
Bruno Turcksin [Mon, 15 Jan 2018 23:04:50 +0000 (18:04 -0500)]
Change .cuh extension to .h
Bruno Turcksin [Mon, 15 Jan 2018 23:00:19 +0000 (18:00 -0500)]
Blacklist header files containing CUDA code from all-headers tests
Header files that contain CUDA codes cannot be passed to the host compiler.
Thus, we blacklist them in all-headers tests.
Daniel Arndt [Sat, 13 Jan 2018 22:10:01 +0000 (23:10 +0100)]
Increase precision in FE_RT_Bubbles tests and filter small values out
Bruno Turcksin [Mon, 15 Jan 2018 20:54:21 +0000 (15:54 -0500)]
Add cuSPARSE support in the build system
Daniel Arndt [Mon, 15 Jan 2018 17:44:19 +0000 (18:44 +0100)]
Merge pull request #5690 from tamiko/fix_mpi_check
CMake: remove mpi two-pass search
Benjamin Brands [Mon, 15 Jan 2018 16:59:28 +0000 (17:59 +0100)]
fix return type of pslange_
Benjamin Brands [Mon, 15 Jan 2018 13:29:07 +0000 (14:29 +0100)]
added least squares routine and tests
Benjamin Brands [Mon, 15 Jan 2018 16:52:35 +0000 (17:52 +0100)]
added SVD computation and tests
Benjamin Brands [Wed, 10 Jan 2018 10:35:07 +0000 (11:35 +0100)]
changed eigenvalue/eigenvector computation
nivesh [Mon, 15 Jan 2018 15:28:40 +0000 (16:28 +0100)]
Bug fix in FE_Enriched element. Avoid evaluating quadrature points if no dofs are assigned
Daniel Arndt [Mon, 15 Jan 2018 08:47:40 +0000 (09:47 +0100)]
Merge pull request #5719 from bangerth/doc-update
Update some comments around renumbering DoFs.
Daniel Arndt [Sun, 14 Jan 2018 22:59:01 +0000 (23:59 +0100)]
Add comments
Wolfgang Bangerth [Sun, 14 Jan 2018 15:49:57 +0000 (08:49 -0700)]
Merge pull request #5730 from masterleinad/clang-tidy-new__3
Make some variables const (references)
Jean-Paul Pelteret [Sun, 14 Jan 2018 08:02:11 +0000 (09:02 +0100)]
Merge pull request #5729 from masterleinad/clang-tidy-new_2
Use nullptr instead of 0 or NULL
Daniel Arndt [Sat, 13 Jan 2018 18:39:08 +0000 (19:39 +0100)]
Make some variables const (references)
Daniel Arndt [Sat, 13 Jan 2018 21:31:46 +0000 (22:31 +0100)]
Merge pull request #5704 from BenBrands/extend_scalapack
add template wrappers for ScaLAPACK functions
Daniel Arndt [Sat, 13 Jan 2018 13:53:04 +0000 (14:53 +0100)]
Be explicit about virtual functions called in constructors and destructors
Daniel Arndt [Sat, 13 Jan 2018 16:37:31 +0000 (17:37 +0100)]
Use nullptr
Daniel Arndt [Sat, 13 Jan 2018 01:51:38 +0000 (02:51 +0100)]
Merge pull request #5722 from eldarkh/add-fe-rt-bubbles
New Finite Element RT+Bubbles (PR #4540 part II)
Matthias Maier [Fri, 12 Jan 2018 21:47:39 +0000 (15:47 -0600)]
CMake: make the MPI compatibility check less strict
When configuring with CMake 3.10 and providing compiler wrappers,
CC=mpicc CXX=mpicxx cmake ...
the new FindMPI.cmake library does not export MPI_CXX_LIBRARIES and
MPI_C_LIBRARIES any more. So, make the check less strict.
Matthias Maier [Fri, 12 Jan 2018 21:31:14 +0000 (15:31 -0600)]
CMake: Support a FindMPI.cmake of CMake version 3.10 and later
This change should also be backward compatible to older CMake versions.
(Fingers crossed.)
Matthias Maier [Mon, 1 Jan 2018 19:41:12 +0000 (13:41 -0600)]
add changes entry
Matthias Maier [Mon, 1 Jan 2018 19:33:47 +0000 (13:33 -0600)]
CMake: Disable MPI autodetection
Matthias Maier [Mon, 1 Jan 2018 19:30:05 +0000 (13:30 -0600)]
CMake: remove mpi two-pass search
As discussed in https://github.com/dealii/dealii/issues/5510, we must
not assume anything about the internal workings of CMake's FindMPI.cmake
module. In particular, it is not sufficient to set CMAKE_<LANG>_COMPILER
to an MPI-compiler wrapper in order for FindMPI.cmake to actually detect
MPI. So, remove this logic in our FindMPI.cmake module.
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.