]>
https://gitweb.dealii.org/ - dealii.git/log
Wolfgang Bangerth [Mon, 2 Apr 2018 16:17:00 +0000 (10:17 -0600)]
Update thread_management.h
Wolfgang Bangerth [Sun, 1 Apr 2018 17:00:52 +0000 (11:00 -0600)]
Update description of Thread/Task::return_value().
Denis Davydov [Sun, 1 Apr 2018 08:30:30 +0000 (10:30 +0200)]
Merge pull request #6132 from drwells/check-fe-system-dynamic-cast
Explicitly check the result of a dynamic_cast.
Denis Davydov [Sun, 1 Apr 2018 08:29:47 +0000 (10:29 +0200)]
Merge pull request #6101 from davydden/doc/fix_latex_lapack_scalapack
fix LaTeX in Lapack and ScaLapack
David Wells [Sat, 31 Mar 2018 20:33:54 +0000 (16:33 -0400)]
Explicitly check the result of a dynamic_cast.
Daniel Arndt [Sat, 31 Mar 2018 22:29:32 +0000 (00:29 +0200)]
Merge pull request #6131 from drwells/move-dtor-code-inside-try
Move some destructor code inside a try block.
Bruno Turcksin [Sat, 31 Mar 2018 20:21:57 +0000 (16:21 -0400)]
Merge pull request #6114 from bangerth/movable
Enable writing Threads::Thread<T> for types T that are only movable.
David Wells [Sat, 31 Mar 2018 20:21:18 +0000 (16:21 -0400)]
Merge pull request #6130 from tamiko/fix_a_test_14
petsc/solver_02: Make range more forgiving
David Wells [Sat, 31 Mar 2018 20:13:59 +0000 (16:13 -0400)]
Move some destructor code inside a try block.
This was caught by coverity.
Matthias Maier [Sat, 31 Mar 2018 16:30:38 +0000 (11:30 -0500)]
petsc/solver_02: Make range more forgiving
Depending on the PETSc version the solver takes roughly between 1069 and
1141 cycles. Just check for a genereous number of cycles (between 1050
and 1150).
Daniel Arndt [Sat, 31 Mar 2018 06:50:23 +0000 (13:50 +0700)]
Merge pull request #6110 from drwells/avoid-slepc-deprecated
Avoid using a deprecated SLEPc function.
David Wells [Wed, 28 Mar 2018 21:00:45 +0000 (17:00 -0400)]
Avoid using a deprecated SLEPc function.
EPSComputeResidualNorm and EPSComputeRelativeError were both deprecated
in the 3.6 release in favor of EPSComputeError.
Matthias Maier [Fri, 30 Mar 2018 20:06:06 +0000 (15:06 -0500)]
Merge pull request #6126 from masterleinad/fix_libc++
Avoid using functionality removed in libc++ for C++17
Matthias Maier [Fri, 30 Mar 2018 19:48:56 +0000 (14:48 -0500)]
Merge pull request #6120 from tamiko/fix_a_test_12
Tests: Slightly change stopping criterion in petsc/solver_02
Daniel Arndt [Fri, 30 Mar 2018 17:03:17 +0000 (00:03 +0700)]
Merge pull request #6122 from tamiko/fix_a_test_13
Tests: update petsc_complex/solver_real_02
Daniel Arndt [Fri, 30 Mar 2018 16:10:47 +0000 (18:10 +0200)]
Fix remainig occurrences of std::bind1st and std::bind2nd in bundled boost
Daniel Arndt [Fri, 30 Mar 2018 10:23:57 +0000 (12:23 +0200)]
Detect std::auto_ptr support
Daniel Arndt [Fri, 30 Mar 2018 10:54:48 +0000 (12:54 +0200)]
Avoid auto_ptr in bundled boost
Daniel Arndt [Fri, 30 Mar 2018 10:54:37 +0000 (12:54 +0200)]
Avoid auto_ptr in bundled muParser
Jean-Paul Pelteret [Fri, 30 Mar 2018 07:51:43 +0000 (09:51 +0200)]
Merge pull request #6123 from bangerth/fix-is_contiguous
Make sure is_contiguous() compares addresses, not values, of array objects.
Wolfgang Bangerth [Thu, 29 Mar 2018 21:42:29 +0000 (15:42 -0600)]
Add test.
Wolfgang Bangerth [Thu, 29 Mar 2018 21:33:19 +0000 (15:33 -0600)]
Make sure is_contiguous() compares addresses, not values, of array objects.
Matthias Maier [Thu, 29 Mar 2018 21:07:02 +0000 (16:07 -0500)]
Tests: update petsc_complex/solver_real_02
Make this test a bit more stable to avoid a floating point exception,
see pull request #6120
Set number of allowed iterations to the same value that is use in
petsc/solver_02
Matthias Maier [Thu, 29 Mar 2018 20:16:40 +0000 (15:16 -0500)]
Tests: Slightly change stopping criterion in petsc/solver_02
The Chebyshev smoother for PETSc is quite a tricky beast. Maybe it would
be best to simply not use it. Anyway, it turns out that simply changing
the stopping criterion for the inner smoother a bit let's this test
pass. [1]
Let's just give up and let the test pass...
Fixes #5320
[1] Under certain circumstances the test fails with a floating point
exception (division by zero) with newer PETSc 3.8.X.
Daniel Arndt [Thu, 29 Mar 2018 20:07:19 +0000 (03:07 +0700)]
Merge pull request #6111 from Rombur/cuda_warnings
Fix warnings from AssertCuda and AssertCusparse
Wolfgang Bangerth [Thu, 29 Mar 2018 16:47:40 +0000 (10:47 -0600)]
Merge pull request #6112 from masterleinad/fix_filtered_iterator_ptr
Let Predicate*::clone return a std::unique_ptr
Daniel Arndt [Thu, 29 Mar 2018 07:49:16 +0000 (14:49 +0700)]
Merge pull request #6116 from bangerth/step-12
Remove unnecessary semi-colons in step-12.
Daniel Arndt [Thu, 29 Mar 2018 06:49:03 +0000 (13:49 +0700)]
Merge pull request #6117 from bangerth/step-12-2
Adapt step-12 some further.
Wolfgang Bangerth [Thu, 29 Mar 2018 03:40:49 +0000 (21:40 -0600)]
Adapt step-12 some further.
In #5988, I added a function beta(...) that computes the advection direction
in step-12, but I forgot to update one other place where we computed this
direction as well. Use the new function in there as well.
Wolfgang Bangerth [Thu, 29 Mar 2018 03:31:52 +0000 (21:31 -0600)]
Remove unnecessary semi-colons in step-12.
Wolfgang Bangerth [Thu, 29 Mar 2018 01:31:07 +0000 (19:31 -0600)]
Add a changelog entry.
Wolfgang Bangerth [Thu, 29 Mar 2018 01:30:45 +0000 (19:30 -0600)]
Add tests.
Wolfgang Bangerth [Thu, 29 Mar 2018 01:30:28 +0000 (19:30 -0600)]
Enable writing Threads::Thread<T> for types T that are only movable.
Previously, it was not possible to write Threads::new_thread(&foo) or Threads::new_task(&foo)
for functions that return an object whose type is not copyable, but is movable. That's
because in a number of places in the internal machinery of the Threads::Thread and
Threads::Task classes, we didn't account for such types.
This patch fixes this.
Daniel Arndt [Wed, 28 Mar 2018 16:37:13 +0000 (18:37 +0200)]
Let Predicate*::clone return a std::unique_ptr
Bruno Turcksin [Wed, 28 Mar 2018 20:24:04 +0000 (16:24 -0400)]
Fix warnings from AssertCuda and AssertCusparse
David Wells [Wed, 28 Mar 2018 20:22:14 +0000 (16:22 -0400)]
Merge pull request #6105 from tamiko/fix_gla_tests
Tests: Update gla/* test outputs
David Wells [Wed, 28 Mar 2018 20:19:44 +0000 (16:19 -0400)]
Merge pull request #6087 from bangerth/fix-petsc
Fix iterating over entries of parallel PETSc matrices.
David Wells [Wed, 28 Mar 2018 20:17:40 +0000 (16:17 -0400)]
Merge pull request #6109 from masterleinad/use_reference
Avoid copies clang-tidy complains about
Martin Kronbichler [Wed, 28 Mar 2018 12:52:55 +0000 (14:52 +0200)]
Merge pull request #6107 from davydden/feature/p_vector_compress_split
la::p::Vector: split compress() and update_ghosts() calls into chunks
Bruno Turcksin [Wed, 28 Mar 2018 12:45:34 +0000 (08:45 -0400)]
Merge pull request #5989 from drwells/update-tbb
Update tbb
Denis Davydov [Wed, 28 Mar 2018 09:00:23 +0000 (11:00 +0200)]
extend description
Daniel Arndt [Wed, 28 Mar 2018 08:40:33 +0000 (10:40 +0200)]
Avoid copies clang-tidy complains about
Denis Davydov [Tue, 27 Mar 2018 11:03:05 +0000 (13:03 +0200)]
la::p::Vector: split compress() and update_ghosts() calls into chunks
Bruno Turcksin [Wed, 28 Mar 2018 02:16:22 +0000 (22:16 -0400)]
Merge pull request #6080 from BenBrands/Inverse
ScaLAPACKMatrix::invert for non-symmetric matrices
Wolfgang Bangerth [Tue, 27 Mar 2018 21:21:14 +0000 (15:21 -0600)]
Add a changelog entry.
Wolfgang Bangerth [Wed, 21 Mar 2018 00:45:41 +0000 (18:45 -0600)]
Add a test.
Daniel Arndt [Tue, 27 Mar 2018 18:50:53 +0000 (01:50 +0700)]
Merge pull request #6097 from bangerth/timer-output
Use TimerOutput::Scope in step-32.
Wolfgang Bangerth [Wed, 21 Mar 2018 00:45:24 +0000 (18:45 -0600)]
Fix iterating over entries of parallel PETSc matrices.
Previously, it was not possible to iterate over the local range because
one would have to call
matrix.end(row)
where 'row' is the last locally owned row, and that triggered an
assertion because this end iterator is also the begin iterator of
the next row -- which is not locally owned any more.
Fix this.
David Wells [Tue, 27 Mar 2018 16:32:49 +0000 (12:32 -0400)]
Merge pull request #6108 from bangerth/fix-step-35
Fix odd use of comma operator.
Wolfgang Bangerth [Tue, 27 Mar 2018 16:17:45 +0000 (10:17 -0600)]
Fix odd use of comma operator.
step-35 uses an odd way of concatenating two 'void' statements, using
operator comma where a semicolon was intended. Fix this.
This was pointed out to me by my student Chen Shen.
Denis Davydov [Tue, 27 Mar 2018 04:42:08 +0000 (06:42 +0200)]
Merge pull request #6104 from masterleinad/fix_minmax_cell_diameter
Let GridTools::minimal/maximal_cell_diameter return global values
Matthias Maier [Mon, 26 Mar 2018 20:50:19 +0000 (15:50 -0500)]
Tests: Add an output variant for gla/block_mat_03
Matthias Maier [Mon, 26 Mar 2018 20:49:33 +0000 (15:49 -0500)]
Tests: rename output variant
All of these output variants are for complex number, thus call the file
"output.complex-output" instead of "output.2".
Matthias Maier [Mon, 26 Mar 2018 20:47:16 +0000 (15:47 -0500)]
Tests: Remove unnecessary with_trilinos=true guard
This subdirectory is only every parsed if Trilinos, PETSc and MPI
support is enabled.
Daniel Arndt [Mon, 26 Mar 2018 19:39:04 +0000 (21:39 +0200)]
Let GridTools::minimal/maximal_cell_diameter return global values
Bruno Turcksin [Mon, 26 Mar 2018 18:44:16 +0000 (14:44 -0400)]
Merge pull request #6103 from tamiko/update_test_script
CMake: Do not require a special name for TRACK="Continuous"
Wolfgang Bangerth [Fri, 23 Mar 2018 22:27:00 +0000 (16:27 -0600)]
Use TimerOutput::Scope in step-32.
This program still used the old, manual enter_section/leave_section()
functions. We know these days how to do this better.
Martin Kronbichler [Mon, 26 Mar 2018 15:56:06 +0000 (17:56 +0200)]
Merge pull request #6102 from davydden/bugfix/parall_vector_compress
la::p::Vector::compress(insert) now correctly zeroes ghosts
Matthias Maier [Mon, 26 Mar 2018 13:03:42 +0000 (08:03 -0500)]
CMake: Do not require a special name for TRACK=continuous
Denis Davydov [Mon, 26 Mar 2018 13:32:20 +0000 (15:32 +0200)]
bugfix: la::p::Vector::compress(insert) now correctly zeroes ghosts
Denis Davydov [Mon, 26 Mar 2018 07:17:37 +0000 (09:17 +0200)]
Merge pull request #6088 from tamiko/preprocess_prm_in
CMake: Preprocess testsuite parameter files
Martin Kronbichler [Mon, 26 Mar 2018 07:13:50 +0000 (09:13 +0200)]
Merge pull request #6100 from masterleinad/fix_minmax_cell_diameter
Fix GridTools::minimal/maximal_cell_diameter for p::d::Triangulation
Martin Kronbichler [Mon, 26 Mar 2018 06:58:40 +0000 (08:58 +0200)]
Merge pull request #6099 from davydden/feature/improve_partitioner
improve MPI::Partitioner for the case of empty owned DoFs with with non-empty ghosts
Denis Davydov [Sun, 25 Mar 2018 08:47:45 +0000 (10:47 +0200)]
doc: fix LaTeX in Lapack and ScaLapack
Denis Davydov [Sun, 25 Mar 2018 12:06:06 +0000 (14:06 +0200)]
Merge pull request #6095 from davydden/doc/partitioner
give example of Utilities::MPI::Partitioner
Denis Davydov [Fri, 23 Mar 2018 20:26:04 +0000 (21:26 +0100)]
doc: give example of Utilities::MPI::Partitioner
Matthias Maier [Sun, 25 Mar 2018 07:18:34 +0000 (02:18 -0500)]
Merge pull request #6090 from luca-heltai/fix-occ-normal-to-mesh
Fix bug in NormalToMeshProjection
Daniel Arndt [Sat, 24 Mar 2018 17:30:33 +0000 (00:30 +0700)]
Fix GridTools::minimal/maximal_cell_dimaeter for p::d::Triangulation
Denis Davydov [Sat, 24 Mar 2018 19:26:25 +0000 (20:26 +0100)]
improve MPI::Partitioner to support empty ArrayView for owned DoFs
Denis Davydov [Sat, 24 Mar 2018 10:28:50 +0000 (11:28 +0100)]
Merge pull request #6098 from tamiko/add_rpath
CMake: Add rpath to example steps
Matthias Maier [Sat, 24 Mar 2018 01:05:21 +0000 (20:05 -0500)]
CMake: Add rpath to example steps
Quoting from the documentation [1]
"""
CMAKE_INSTALL_RPATH_USE_LINK_PATH is an interesting and very useful
option. When building a target with RPATH, CMake determines the RPATH
by using the directories of all libraries to which this target links.
Some of these libraries may be located in the same build tree, e.g.
libbar.so, these directories are also added to the RPATH. If this
option is enabled, all these directories except those which are also
in the build tree will be added to the install RPATH automatically.
The only directories which may then still be missing from the RPATH
are the directories where the libraries from the same project (i.e.
libbar.so) are installed to. If the install directory for the
libraries is not one of the systems default library directories, you
have to add this directory yourself to the install RPATH by setting
CMAKE_INSTALL_RPATH accordingly.
"""
Well, let's simply add the missing rpath to the install location for our
example binaries then. We guard this with
CMAKE_INSTALL_RPATH_USE_LINK_PATH in case augment the INSTALL_RPATH
so that CMake doesn't populate any rpath in case the user doesn't want
that.
[1] https://cmake.org/Wiki/CMake_RPATH_handling
Fixes #6096
Daniel Arndt [Fri, 23 Mar 2018 17:01:58 +0000 (00:01 +0700)]
Merge pull request #6081 from bangerth/fix-copy_from
Fix quadratic behavior of SparsityPattern::copy_from().
Matthias Maier [Fri, 23 Mar 2018 16:41:12 +0000 (11:41 -0500)]
Merge pull request #6094 from masterleinad/fix_cone_04
Fix cone_04
David Wells [Fri, 23 Mar 2018 14:29:15 +0000 (10:29 -0400)]
Merge pull request #6092 from tamiko/fix_a_test_10
Tests: Provide output variants for clang-6 and libc++
Daniel Arndt [Fri, 23 Mar 2018 11:13:05 +0000 (12:13 +0100)]
Fix cone_04
Daniel Arndt [Fri, 23 Mar 2018 03:37:43 +0000 (10:37 +0700)]
Merge pull request #5959 from drwells/transpose-table-iterator
Transpose table iterator
Daniel Arndt [Fri, 23 Mar 2018 02:14:15 +0000 (09:14 +0700)]
Merge pull request #6091 from tamiko/fix_check
CMake: Also propagate compiler and linker flags to the boost check
Daniel Arndt [Fri, 23 Mar 2018 00:19:31 +0000 (07:19 +0700)]
Merge pull request #6093 from tamiko/fix_a_test_11
Tests: Do not print memory consumptions in fe/fe_move_0? tests
Matthias Maier [Thu, 22 Mar 2018 20:26:00 +0000 (15:26 -0500)]
Tests: Do not print memory consumptions in fe/fe_move_0? tests
The memory consumption is simply not stable with respect to architecture
and standard library.
The test remains useful - the actual contents of the moved elements is
still printed and compared in a unique (and stable) manner.
Further, remove output variants for clang9_darwin - I checked that the
output is now the same.
Matthias Maier [Thu, 22 Mar 2018 20:22:12 +0000 (15:22 -0500)]
Tests: Provide output variant for clang-6 and libc++
The renumbering in dofs/dof_renumbering_02 changes slightly when libc++
is used. Let's add an output variant.
David Wells [Sun, 18 Mar 2018 22:41:09 +0000 (18:41 -0400)]
Implement a random-access iterator for TransposeTable.
Matthias Maier [Thu, 22 Mar 2018 18:26:24 +0000 (13:26 -0500)]
CMake: Also propagate compiler and linker flags to the boost check
David Wells [Thu, 1 Mar 2018 15:03:54 +0000 (10:03 -0500)]
Add a new header for implementing random-access iterators.
Matthias Maier [Wed, 21 Mar 2018 21:52:22 +0000 (16:52 -0500)]
doc: Update documentation
Wolfgang Bangerth [Wed, 21 Mar 2018 20:09:26 +0000 (14:09 -0600)]
Add another test.
Wolfgang Bangerth [Wed, 21 Mar 2018 20:09:12 +0000 (14:09 -0600)]
Ensure we get no segmentation faults for empty matrices.
Wolfgang Bangerth [Wed, 21 Mar 2018 20:08:55 +0000 (14:08 -0600)]
Relax a couple of assertions for empty matrices.
Luca Heltai [Wed, 21 Mar 2018 17:09:11 +0000 (18:09 +0100)]
Fix bug in normal to mesh projection due to Boundary -> Manifold
Daniel Arndt [Wed, 21 Mar 2018 16:56:40 +0000 (23:56 +0700)]
Merge pull request #6071 from tamiko/fix_a_test_4
tests: sundials/harmonic_oscillator_04 increase precision
Matthias Maier [Wed, 21 Mar 2018 15:28:25 +0000 (10:28 -0500)]
add a news/changes entry
Matthias Maier [Wed, 21 Mar 2018 15:20:03 +0000 (10:20 -0500)]
CMakes: Preprocess testsuite parameter files
Add a third test variant consisting of a .prm.in file:
test.output
test.prm.in
The test.prm.in file is preprocessed by CONFIGURE_FEATURE to a test.prm
file substituting all @VARIABLE@ string with the corresponding CMake
variable.
Matthias Maier [Wed, 21 Mar 2018 15:08:41 +0000 (10:08 -0500)]
Tests: Add a test that checks the *.prm.in mechanism
This small test checks whether the testsuite correctly picks up an
configures a test consisting of
parameter_file.prm.in
parameter_file.output
Matthias Maier [Wed, 21 Mar 2018 15:06:05 +0000 (10:06 -0500)]
Tests: Add a test that checks the *.prm mechanism
This small test checks whether the testsuite correctly picks up a test
consisting of
parameter_file.prm
parameter_file.output
Benjamin Brands [Tue, 20 Mar 2018 22:34:54 +0000 (23:34 +0100)]
ScaLAPACKMatrix::invert() for non-symmetric matrices
Martin Kronbichler [Wed, 21 Mar 2018 10:48:57 +0000 (11:48 +0100)]
Merge pull request #6078 from tamiko/fix_a_test_7
tests: lac/*vector_add_and_dot*
Martin Kronbichler [Wed, 21 Mar 2018 10:48:30 +0000 (11:48 +0100)]
Merge pull request #6079 from tamiko/fix_a_test_8
VectorTools: Increase max iterations in project_matrix_free
Martin Kronbichler [Wed, 21 Mar 2018 10:47:28 +0000 (11:47 +0100)]
Merge pull request #6083 from bangerth/fix-step-35
Do not make a variable in step-35 static.
Wolfgang Bangerth [Tue, 20 Mar 2018 23:06:57 +0000 (17:06 -0600)]
Add a changelog entry.
Wolfgang Bangerth [Tue, 20 Mar 2018 23:11:51 +0000 (17:11 -0600)]
Do not make a variable in step-35 static.
Static variables are only initialized once at the beginning of the run. Here,
we have a vector that is marked as 'static' but this really only works because
we never do mesh refinement in this program and consequently (i) the originally
set size continues to be correct, and (ii) the code below sets every vector
element, rather than add to it.
Matthias Maier [Tue, 20 Mar 2018 23:10:54 +0000 (18:10 -0500)]
Merge pull request #6082 from bangerth/fix-typo
Fix a grammar error.
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.