]>
https://gitweb.dealii.org/ - dealii.git/log
Matthias Maier [Mon, 19 Apr 2021 20:43:20 +0000 (15:43 -0500)]
free allocated MPI resources before we call MPI_Finalize
Matthias Maier [Mon, 19 Apr 2021 20:26:19 +0000 (15:26 -0500)]
force initialization of static struct
Matthias Maier [Sun, 18 Apr 2021 21:11:33 +0000 (16:11 -0500)]
Utilities::MPI::mpi_sum_avg: create MPI_Datatype and MPI_Op once
Instead of repeatedly creating and freeing an MPI_Datatype and MPI_Op
for every invocation of Utilities::MPI::mpi_sum_avg let us define these
objects statically and create them only once.
This works around a memory leak issue with OpenMPI 4.1.0 that will leak
memory in the subsequent MPI_Allreduce call.
Upstream bug report: https://github.com/open-mpi/ompi/issues/8827
Matthias Maier [Mon, 19 Apr 2021 18:17:41 +0000 (13:17 -0500)]
add a changes entry
Matthias Maier [Mon, 19 Apr 2021 20:48:51 +0000 (15:48 -0500)]
add a test
Matthias Maier [Mon, 19 Apr 2021 18:13:50 +0000 (13:13 -0500)]
MPI_InitFinalize: add an at_mpi_init and at_mpi_finalize signal handler
Daniel Arndt [Mon, 19 Apr 2021 16:46:27 +0000 (12:46 -0400)]
Merge pull request #12055 from bangerth/grammar
Fix grammar.
Wolfgang Bangerth [Mon, 19 Apr 2021 16:15:45 +0000 (10:15 -0600)]
Fix grammar.
Daniel Arndt [Mon, 19 Apr 2021 15:22:39 +0000 (11:22 -0400)]
Merge pull request #12052 from peterrum/vector_tools_evaluate_at_points_02_fix
Make test remote_point_evaluation/vector_tools_evaluate_at_points_02 more robust
David Wells [Mon, 19 Apr 2021 15:15:55 +0000 (11:15 -0400)]
Merge pull request #12051 from peterrum/fix_compute_diagonal_tests
Fix matrix_free/compute_diagonal tests in release mode
David Wells [Mon, 19 Apr 2021 15:13:45 +0000 (11:13 -0400)]
Merge pull request #12053 from peterrum/ecl_02_fix
Fix test matrix_free/ecl_02
Peter Munch [Sun, 18 Apr 2021 10:54:59 +0000 (12:54 +0200)]
Fix test matrix_free/ecl_02
Peter Munch [Sun, 18 Apr 2021 07:46:00 +0000 (09:46 +0200)]
Make test remote_point_evaluation/vector_tools_evaluate_at_points_02 more robust
Peter Munch [Sun, 18 Apr 2021 07:29:36 +0000 (09:29 +0200)]
Fix matrix_free/compute_diagonal tests in release mode
David Wells [Sat, 17 Apr 2021 13:24:19 +0000 (09:24 -0400)]
Merge pull request #12031 from bangerth/kinsol
Update the KINSOL callbacks to what KINSOL current provides.
Wolfgang Bangerth [Fri, 16 Apr 2021 22:06:26 +0000 (16:06 -0600)]
Add tests for the new KINSOL interface.
Wolfgang Bangerth [Fri, 16 Apr 2021 22:01:00 +0000 (16:01 -0600)]
Merge pull request #12049 from drwells/vtk-compatibility-tria
Remove remaining create_triangulation_compatibility calls.
Wolfgang Bangerth [Fri, 16 Apr 2021 21:58:58 +0000 (15:58 -0600)]
Merge pull request #12048 from drwells/gmsh-compatibility-tria
Get rid of create_triangulation_compatibility in the GMSH reader.
David Wells [Fri, 16 Apr 2021 17:06:43 +0000 (13:06 -0400)]
Deprecate create_triangulation_compatibility.
David Wells [Fri, 16 Apr 2021 15:41:15 +0000 (11:41 -0400)]
Remove the rest of the create_triangulation_compatibility calls.
David Wells [Fri, 16 Apr 2021 17:03:17 +0000 (13:03 -0400)]
Make the vertex swapping algorithm work with the new-style ordering.
This is necessary for the tecplot reader, which relies on this behavior.
David Wells [Fri, 16 Apr 2021 14:59:54 +0000 (10:59 -0400)]
Get rid of create_triangulation_compatibility in the GMSH reader.
David Wells [Fri, 16 Apr 2021 13:51:58 +0000 (09:51 -0400)]
Merge pull request #12027 from bangerth/cmake
Improve error messages during configuration.
Daniel Arndt [Fri, 16 Apr 2021 04:11:57 +0000 (00:11 -0400)]
Merge pull request #12026 from bangerth/cmake-3
Use a better filename for SUNDIALS finding.
Daniel Arndt [Fri, 16 Apr 2021 02:54:16 +0000 (22:54 -0400)]
Merge pull request #12030 from bangerth/kinsol-versions
Support different KINSOL versions.
Wolfgang Bangerth [Fri, 16 Apr 2021 00:07:18 +0000 (18:07 -0600)]
Merge pull request #12045 from drwells/dbmesh-xda-compatibility-tria
Remove two more create_triangulation_compatibility calls.
Wolfgang Bangerth [Fri, 16 Apr 2021 00:06:21 +0000 (18:06 -0600)]
Merge pull request #12046 from drwells/grid-generator-compatibility-tria
Remove calls to create_triangulation_compatibility in GridGenerator.
Wolfgang Bangerth [Thu, 8 Apr 2021 17:40:37 +0000 (11:40 -0600)]
Improve error messages during configuration.
Wolfgang Bangerth [Thu, 8 Apr 2021 17:42:11 +0000 (11:42 -0600)]
Use a better filename for SUNDIALS finding.
Wolfgang Bangerth [Thu, 8 Apr 2021 21:53:16 +0000 (15:53 -0600)]
Support different KINSOL versions.
David Wells [Thu, 15 Apr 2021 14:27:07 +0000 (10:27 -0400)]
Merge pull request #12044 from tamiko/fix_configure
CMake: FindKOKKOS: Work around configuration failure
David Wells [Thu, 15 Apr 2021 14:25:45 +0000 (10:25 -0400)]
Remove calls to create_triangulation_compatibility in GridGenerator.
David Wells [Thu, 15 Apr 2021 13:18:39 +0000 (09:18 -0400)]
Merge the two read_xda functions.
David Wells [Thu, 15 Apr 2021 13:12:54 +0000 (09:12 -0400)]
Get rid of create_triangulation_compatibility in the XDA reader.
David Wells [Wed, 14 Apr 2021 04:03:20 +0000 (00:03 -0400)]
Get rid of create_triangulation_compatibility in the DB Mesh reader.
Daniel Arndt [Thu, 15 Apr 2021 12:42:15 +0000 (08:42 -0400)]
Merge pull request #12043 from marcfehling/tut-list
Remove hp::DoFHandler from tutorial list.
Matthias Maier [Thu, 15 Apr 2021 04:48:56 +0000 (23:48 -0500)]
CMake: FindKOKKOS: Work around configuration failure
Configuration did fail in the following situation
- Kokkos installed via Trilinos (MPI enabled)
- deal.II configured without MPI
In this situation FindTRILINOS.cmake bails out because of an MPI
configuration mismatch, so the manual search for Kokkos in
FindKOKKOS.cmake is triggered. The latter one fails because Kokkos
installed via Trilinos did export a valid project configuration only
containing the Kokkos::kokkos target and not the Kokkos::kokkoscore
target.
Work around this issue by only configuring if Kokkos::kokkoscore has
indeed been imported.
Marc Fehling [Thu, 15 Apr 2021 03:15:01 +0000 (21:15 -0600)]
Remove hp::DoFHandler from tutorial list.
Bruno Turcksin [Thu, 15 Apr 2021 01:16:29 +0000 (21:16 -0400)]
Merge pull request #12042 from drwells/ucd-compatibility-tria
Get rid of create_triangulation_compatibility in the UCD reader.
Bruno Turcksin [Wed, 14 Apr 2021 16:03:57 +0000 (12:03 -0400)]
Merge pull request #12039 from drwells/unv-compatibility-tria
Get rid of create_triangulation_compatibility in the UNV reader.
David Wells [Wed, 14 Apr 2021 03:39:29 +0000 (23:39 -0400)]
Get rid of create_triangulation_compatibility in the UCD reader.
To keep reading input consistently this also requires that we slightly modify
the cell inversion algorithm. This is backwards-compatible since the option to
use the new style ordering was added in
7f9c1023b61 .
David Wells [Tue, 13 Apr 2021 14:22:31 +0000 (10:22 -0400)]
Get rid of create_triangulation_compatibility in the UNV reader.
This case is interesting since the current code only works by coincidence. Our
test files have vertices numbered like
2 ---- 3
| |
| |
1 ---- 0
which we read in in the old-style (UCD) numbering, which is like mirroring
across the y-axis. This causes all the cells to have a negative measure (but they
aren't pinched). GridReordering then fixes this by swapping vertices 1 and 3 of
every single cell - i.e., adding another -1 factor to the measure, which now
causes all cells to have positive measure. Something similar happens in 3D.
As far as I can tell there is no widely accepted node numbering for UNV like
there is for everything else so I am going by what Salome generates. Since we
now use their numbering directly (instead of something that worked by
coincidence) the output files are slightly different but all cells do end up
looking correct (nothing is twisted when we visualize with VTK).
Bruno Turcksin [Wed, 14 Apr 2021 02:10:47 +0000 (22:10 -0400)]
Merge pull request #12041 from marcfehling/cell-weights-empty
Bug in weighting_callback for uninitialized DoFHandlers.
Bruno Turcksin [Wed, 14 Apr 2021 02:07:43 +0000 (22:07 -0400)]
Merge pull request #12040 from peterrum/process_dof_indices_fix_set_periodicity_constraints
Fix process_dof_indices (for artificial faces in DoFTools::set_periodicity_constraints)
Marc Fehling [Tue, 13 Apr 2021 21:38:59 +0000 (15:38 -0600)]
Bug in weighting_callback for uninitialized DoFHandlers.
Peter Munch [Tue, 13 Apr 2021 19:58:57 +0000 (21:58 +0200)]
Fix process_dof_indices (for artificial faces in DoFTools::set_periodicity_constraints)
David Wells [Tue, 13 Apr 2021 14:56:13 +0000 (10:56 -0400)]
Merge pull request #12038 from bangerth/61
Clarify an issue in the introduction of step-61.
David Wells [Tue, 13 Apr 2021 14:43:36 +0000 (10:43 -0400)]
Merge pull request #12032 from peterrum/dof_handler_artificial_dofs
Do not allocate memory for DoFs on artificial cells
Wolfgang Bangerth [Mon, 12 Apr 2021 17:24:33 +0000 (11:24 -0600)]
Clarify an issue in the introduction of step-61.
Peter Munch [Fri, 9 Apr 2021 15:24:19 +0000 (17:24 +0200)]
Do not allocate memory for DoFs on artificial cells
Timo Heister [Sat, 10 Apr 2021 17:14:19 +0000 (13:14 -0400)]
Merge pull request #12036 from bangerth/uniformity
Make tooltips more uniform in style.
Martin Kronbichler [Sat, 10 Apr 2021 13:49:55 +0000 (15:49 +0200)]
Merge pull request #12033 from peterrum/find_vector_in_mf
Compare pointers in find_vector_in_mf
Wolfgang Bangerth [Fri, 9 Apr 2021 22:13:58 +0000 (16:13 -0600)]
Make tooltips more uniform in style.
Peter Munch [Fri, 9 Apr 2021 19:45:51 +0000 (21:45 +0200)]
Compare pointers in find_vector_in_mf
Bruno Turcksin [Fri, 9 Apr 2021 02:54:39 +0000 (22:54 -0400)]
Merge pull request #12028 from bangerth/cmake-2
Make cmake messages easier to read.
Wolfgang Bangerth [Thu, 8 Apr 2021 19:30:13 +0000 (13:30 -0600)]
Update the KINSOL callbacks to what KINSOL current provides.
Wolfgang Bangerth [Thu, 8 Apr 2021 17:41:21 +0000 (11:41 -0600)]
Make cmake error messages easier to read.
Marc Fehling [Wed, 7 Apr 2021 05:23:58 +0000 (07:23 +0200)]
Merge pull request #12020 from bangerth/p2
Fix a typo.
Bruno Turcksin [Wed, 7 Apr 2021 02:16:51 +0000 (22:16 -0400)]
Merge pull request #12019 from bangerth/15-5
Break generation of graphical data into its own function in step-15.
Bruno Turcksin [Wed, 7 Apr 2021 02:14:49 +0000 (22:14 -0400)]
Merge pull request #12021 from bangerth/p1
Wolfgang Bangerth [Tue, 6 Apr 2021 21:47:09 +0000 (15:47 -0600)]
Avoid unnecessary include files.
Wolfgang Bangerth [Tue, 6 Apr 2021 21:46:45 +0000 (15:46 -0600)]
Add missing include files.
Wolfgang Bangerth [Tue, 6 Apr 2021 21:47:34 +0000 (15:47 -0600)]
Fix a typo.
Wolfgang Bangerth [Wed, 31 Mar 2021 15:56:25 +0000 (09:56 -0600)]
Break generation of graphical data into its own function in step-15.
David Wells [Tue, 6 Apr 2021 18:44:31 +0000 (14:44 -0400)]
Merge pull request #11869 from dangars/hdf5_bool
Add bool type to HDF5
Daniel Arndt [Tue, 6 Apr 2021 17:55:34 +0000 (13:55 -0400)]
Merge pull request #12018 from bangerth/table-3
Wolfgang Bangerth [Mon, 5 Apr 2021 22:32:38 +0000 (16:32 -0600)]
Rename more local variables. Add to the documentation.
Wolfgang Bangerth [Mon, 5 Apr 2021 21:28:38 +0000 (15:28 -0600)]
Slightly simplify the AlignedVector::reserve() function.
Wolfgang Bangerth [Mon, 5 Apr 2021 21:26:59 +0000 (15:26 -0600)]
Rename variables in AlignedVector::reserve().
Wolfgang Bangerth [Mon, 5 Apr 2021 21:24:50 +0000 (15:24 -0600)]
Simplify the logic of AlignedVector::reserve().
Daniel Arndt [Tue, 6 Apr 2021 00:31:47 +0000 (20:31 -0400)]
Merge pull request #12017 from bangerth/table-2
Rename member variables.
Wolfgang Bangerth [Mon, 5 Apr 2021 17:14:22 +0000 (11:14 -0600)]
Rename member variables.
David Wells [Mon, 5 Apr 2021 19:11:57 +0000 (15:11 -0400)]
Merge pull request #12009 from bangerth/table
Use std::unique_ptr to manage memory in AlignedVector.
Daniel Arndt [Mon, 5 Apr 2021 18:32:47 +0000 (14:32 -0400)]
Merge pull request #11965 from peterrum/compute_nonzero_components_assert
FETools::Compositing::compute_nonzero_components(): add assert
Marc Fehling [Mon, 5 Apr 2021 17:49:12 +0000 (19:49 +0200)]
Merge pull request #12010 from tjhei/fewer_prepare_refine_calls
p::d:Triangulation: fewer prepare_coarsening_and_refinement calls
Daniel Arndt [Mon, 5 Apr 2021 17:45:07 +0000 (13:45 -0400)]
Merge pull request #11995 from bangerth/p2
Wolfgang Bangerth [Mon, 5 Apr 2021 17:19:26 +0000 (11:19 -0600)]
Add a test.
Wolfgang Bangerth [Fri, 2 Apr 2021 21:21:33 +0000 (15:21 -0600)]
Use std::unique_ptr to manage memory in AlignedVector.
Wolfgang Bangerth [Wed, 31 Mar 2021 03:32:39 +0000 (21:32 -0600)]
Simplify step-15 with the new function.
Wolfgang Bangerth [Wed, 31 Mar 2021 03:39:03 +0000 (21:39 -0600)]
Simplify step-11 somewhat.
Wolfgang Bangerth [Wed, 31 Mar 2021 03:30:38 +0000 (21:30 -0600)]
Adjust tests.
Wolfgang Bangerth [Wed, 31 Mar 2021 03:25:57 +0000 (21:25 -0600)]
Add a changelog entry.
Wolfgang Bangerth [Wed, 31 Mar 2021 14:51:09 +0000 (08:51 -0600)]
Add a default argument.
Wolfgang Bangerth [Wed, 31 Mar 2021 03:25:49 +0000 (21:25 -0600)]
Deprecate a DoFTools function that returns by argument, rather than directly.
We want to move away from these kinds of functions, so deprecate it and replace it by
a function that does the right thing instead.
Wolfgang Bangerth [Wed, 31 Mar 2021 03:18:53 +0000 (21:18 -0600)]
Add a changelog entry.
Wolfgang Bangerth [Wed, 31 Mar 2021 03:18:42 +0000 (21:18 -0600)]
Deprecate a function in DoFTools.
While there also update the documentation of its replacement a bit.
Peter Munch [Thu, 25 Mar 2021 22:02:05 +0000 (23:02 +0100)]
FETools::Compositing::compute_nonzero_components(): add assert
Timo Heister [Sun, 4 Apr 2021 15:20:20 +0000 (11:20 -0400)]
add tests
Timo Heister [Mon, 28 Sep 2020 19:14:52 +0000 (15:14 -0400)]
reduce number of prepare_coarsening_and_refinement() calls
I noticed that we call prepare_coarsening_and_refinement() many times
when using a p::d::Triangulation. Many of these are certainly
unnecessary:
1. distributed tria for mesh_reconstruction_after_repartitioning. This
is used to remove all refinement and smoothing will be called inside the
execute...() call below. No need to call our more expensive version.
2. serial execute_coarsening_and_refinement() should not call the
derived version, which in turn calls the serial version.
For the test tests/distributed_grids/refine_periodic_01:
Before: 154 sequential, 154 parallel
After: 126 sequential, 53 parallel
fixup
Jean-Paul Pelteret [Sun, 4 Apr 2021 19:59:04 +0000 (21:59 +0200)]
Merge pull request #11990 from vachan-potluri/subdivided_hyper_cube-doc
Fix mistake in parameter description
Jean-Paul Pelteret [Sun, 4 Apr 2021 19:56:22 +0000 (21:56 +0200)]
Merge pull request #12005 from bangerth/kinsol
Some KINSOL documentation updates.
Jean-Paul Pelteret [Sun, 4 Apr 2021 19:42:21 +0000 (21:42 +0200)]
Merge pull request #12006 from bangerth/whitespace
Unify endlines at the end of files.
Jean-Paul Pelteret [Sun, 4 Apr 2021 19:40:28 +0000 (21:40 +0200)]
Merge pull request #12011 from tjhei/readme-typo
small readme.html fixes
Timo Heister [Sun, 4 Apr 2021 16:10:27 +0000 (12:10 -0400)]
small readme.html fixes
Daniel Garcia-Sanchez [Sun, 4 Apr 2021 10:18:55 +0000 (12:18 +0200)]
Add changelog
Daniel Garcia-Sanchez [Sun, 4 Apr 2021 10:14:04 +0000 (12:14 +0200)]
Add HDF5 documentation for datatypes
Daniel Garcia-Sanchez [Sat, 3 Apr 2021 19:15:01 +0000 (21:15 +0200)]
Write test hdf5_07
David Wells [Sat, 3 Apr 2021 02:03:04 +0000 (22:03 -0400)]
Merge pull request #12004 from bangerth/15-4
Use an easier-to-read do...while loop.
Wolfgang Bangerth [Fri, 2 Apr 2021 17:51:13 +0000 (11:51 -0600)]
Unify endlines at the end of files.
Wolfgang Bangerth [Fri, 2 Apr 2021 17:12:53 +0000 (11:12 -0600)]
Merge pull request #11999 from masterleinad/avoid_polymorphic_use_arborx_access_traits
Avoid usage of the ArborX access traits in polymorphic contexts
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.