Wolfgang Bangerth [Thu, 6 Nov 2014 04:29:26 +0000 (22:29 -0600)]
Remove a broken link.
Wolfgang Bangerth [Wed, 5 Nov 2014 19:12:13 +0000 (13:12 -0600)]
Merge pull request #220 from Rombur/trilinos_11_2
Update the documentation about requiring at least Trilinos 11.2
Bruno Turcksin [Tue, 4 Nov 2014 16:07:02 +0000 (10:07 -0600)]
Update the documentation about requiring at least Trilinos 11.2
Bruno Turcksin [Wed, 5 Nov 2014 16:15:31 +0000 (10:15 -0600)]
Merge pull request #222 from kronbichler/matrix_free_1d
Further actions to make MatrixFree/FEEvaluation work in 1D
Timo Heister [Wed, 5 Nov 2014 15:41:11 +0000 (10:41 -0500)]
Merge pull request #221 from kronbichler/master
Fix a compiler warning with gcc-4.9.
gcc-4.9 seems to be able to see that accessing R[1] in 1D is out of bound
for spacedim == 1 but it can omit the warning if accessed inside a switch
statement. Fix this warning (even though we won't use SphericalManifold
for spacedim == 1).
Martin Kronbichler [Wed, 5 Nov 2014 13:02:16 +0000 (14:02 +0100)]
Further actions to make MatrixFree/FEEvaluation work in 1D
- Added test for matrix-vector product
- Made compile with clang.
Martin Kronbichler [Wed, 5 Nov 2014 09:05:17 +0000 (10:05 +0100)]
Fix a compiler warning with gcc-4.9
gcc-4.9 seems to be able to see that accessing p[1] in 1D is out of bound
for spacedim == 1 but it can omit the warning if accessed inside a switch
statement. Fix this warning (even though we won't use SphericalManifold
for spacedim == 1).
Martin Kronbichler [Wed, 5 Nov 2014 09:04:11 +0000 (10:04 +0100)]
Merge pull request #213 from kronbichler/master
Combined add and inner product for vectors: add_and_dot
Martin Kronbichler [Wed, 5 Nov 2014 09:01:09 +0000 (10:01 +0100)]
Merge branch 'master' of https://github.com/dealii/dealii
Martin Kronbichler [Wed, 5 Nov 2014 08:51:56 +0000 (09:51 +0100)]
Merge pull request #219 from rudraa/add-missing-partial-specialization
Resolve two conflicting partial specializations by adding a third, most ...
Shiva Rudraraju [Tue, 4 Nov 2014 15:42:47 +0000 (10:42 -0500)]
Resolve two conflicting partial specializations by adding a third, most specialized case.
Martin Kronbichler [Tue, 4 Nov 2014 09:47:34 +0000 (10:47 +0100)]
Merge pull request #218 from tamiko/trilinos_version_requirement
Increase version requirement for Trilinos to 11.2
Matthias Maier [Tue, 4 Nov 2014 08:54:43 +0000 (09:54 +0100)]
Increase version requirement for Trilinos to 11.2
Martin Kronbichler [Fri, 31 Oct 2014 19:01:43 +0000 (13:01 -0600)]
Merge pull request #217 from bangerth/introduce-dealii-version-macro
Introduce a macro DEAL_II_VERSION_GTE.
Wolfgang Bangerth [Fri, 31 Oct 2014 18:48:20 +0000 (13:48 -0500)]
Introduce a macro DEAL_II_VERSION_GTE.
Martin Kronbichler [Sun, 26 Oct 2014 08:25:43 +0000 (09:25 +0100)]
Combined add and inner product for vectors: add_and_dot
- Implemented a new method add_and_dot for all vector classes (deal.II's vectors,
block vectors, PETSc, Trilinos vectors)
- Specialized function for deal.II vector that does add_and_dot in one sweep over
the data by a new AddAndDot struct that is passed to the accumulate function. This
also extends the other structs for inner products and norms by one (unused) argument.
- New tests for this feature.
- Added new do_vectorized operation to all accumulation operations that uses the
operations of VectorizedArray<double/float> for the sums. This helps to reduce the
operations and thus increase performance when memory transfer is not the limit.
This functionality is called in the regular part of the accumulate function which
now sits in its own method 'accumulate_regular' (length divisible by 32). This
functionality can unfortunately not be realized by the OPENMP_SIMD pragma recently
introduced because that leads to non-reproducible results when the vector memory
starts at different memory locations. VectorizedArray is only supported on some
platforms and thus this functionality may not be available everywhere. However, good
code will be generated unconditionally (except on old machines from around 2004-2008
where unaligned reads also from aligned memory are slow, but those are not relevant
any more).
- To implemented vectorized pow operations, I also added a std::pow function for
VectorizedArray.
- The memory for parallel::distributed::Vector is now allocated to 64 byte boundaries
just as is the memory of Vector<Number>.
Wolfgang Bangerth [Thu, 30 Oct 2014 22:32:38 +0000 (17:32 -0500)]
Merge pull request #204 from bangerth/use-signals-in-solvers
Use signals in solvers
Wolfgang Bangerth [Sat, 25 Oct 2014 02:37:52 +0000 (21:37 -0500)]
Document in classes derived from Solver that they use the signal/slot mechanism.
Wolfgang Bangerth [Sat, 25 Oct 2014 02:31:40 +0000 (21:31 -0500)]
Add a mechanism to attach slots to the signal that now controls convergence.
Document this with an example that illustrates how we can observe
the iteration in process.
Wolfgang Bangerth [Thu, 16 Oct 2014 15:27:11 +0000 (10:27 -0500)]
Introduce signals into the Solver base class.
This patch uses signals to determine whether a current iterate
satisfies the convergence criterion. Currently, the only slot that
is connected to a signal is the SolverControl::check() function,
but users will be able to connect different slots there as well,
closely monitoring the convergence process.
Adjust the testcases.
Timo Heister [Thu, 30 Oct 2014 18:08:00 +0000 (14:08 -0400)]
Merge pull request #215 from bangerth/cleanups-1
Cleanups and doc updates
Luca Heltai [Thu, 30 Oct 2014 16:42:42 +0000 (17:42 +0100)]
Merge pull request #216 from tamiko/add_make_tuple
Also expose make_tuple in std_cxx11
Matthias Maier [Thu, 30 Oct 2014 16:37:29 +0000 (17:37 +0100)]
Also expose make_tuple in std_cxx11
Our C++11 compatibility headers did not expose (std|boost)::make_tuple, fix
this.
Wolfgang Bangerth [Thu, 30 Oct 2014 16:06:46 +0000 (11:06 -0500)]
Fix a typo.
Wolfgang Bangerth [Thu, 30 Oct 2014 15:22:09 +0000 (10:22 -0500)]
Leave a note and a TODO.
Wolfgang Bangerth [Thu, 30 Oct 2014 15:16:04 +0000 (10:16 -0500)]
Hide a variable in the only scope where it is used.
Also put braces around a longish code block.
Wolfgang Bangerth [Thu, 30 Oct 2014 15:14:56 +0000 (10:14 -0500)]
Make code slightly easier to read.
Wolfgang Bangerth [Mon, 27 Oct 2014 15:59:35 +0000 (10:59 -0500)]
Merge pull request #205 from davydden/master
Added set_matrix_mode to the SLEPcWrappers::TransformationBase class.
Denis Davydov [Mon, 27 Oct 2014 13:02:19 +0000 (14:02 +0100)]
made argument of set_matrix_mode const.
Added a reference to SLEPc with possible values.
Timo Heister [Mon, 27 Oct 2014 12:26:53 +0000 (08:26 -0400)]
Merge pull request #211 from bangerth/remove-broken-link
Remove a link that is no longer live.
Wolfgang Bangerth [Sat, 25 Oct 2014 04:08:01 +0000 (23:08 -0500)]
Remove a link that is no longer live.
Wolfgang Bangerth [Sat, 25 Oct 2014 02:41:49 +0000 (21:41 -0500)]
Merge pull request #210 from tjhei/documentation_fixes
Documentation fixes
Timo Heister [Fri, 24 Oct 2014 13:01:28 +0000 (09:01 -0400)]
wrap comments
Timo Heister [Fri, 24 Oct 2014 12:55:24 +0000 (08:55 -0400)]
spell check, minor edits
Timo Heister [Fri, 24 Oct 2014 12:48:02 +0000 (08:48 -0400)]
document project_boundary_values_div_conforming better
Timo Heister [Thu, 23 Oct 2014 13:53:15 +0000 (09:53 -0400)]
Merge pull request #209 from kronbichler/master
Disable nonlocal graph in Trilinos sparsity pattern (revisit in 11.13)
Martin Kronbichler [Thu, 23 Oct 2014 11:39:58 +0000 (13:39 +0200)]
Disable nonlocal graph in Trilinos sparsity pattern because Trilinos has a bug.
Timo Heister [Thu, 23 Oct 2014 11:05:03 +0000 (07:05 -0400)]
Merge pull request #206 from bangerth/remove-faulty-test
The test verifies that it doesn't matter whether we first call
DoFHandler::distribute_dofs() and then DoFHandler::distribute_mg_dofs(),
or the other way around. But we have recently changed the behavior
to require exactly this order, so part of the test is no longer of interest.
Wolfgang Bangerth [Wed, 22 Oct 2014 21:33:57 +0000 (16:33 -0500)]
Remove part of a test.
The test verifies that it doesn't matter whether we first call
DoFHandler::distribute_dofs() and then DoFHandler::distribute_mg_dofs(),
or the other way around. But we have recently changed the behavior
to require exactly this order, so this part of the test is no longer of interest.
Timo Heister [Wed, 22 Oct 2014 20:30:45 +0000 (16:30 -0400)]
Merge pull request #207 from bangerth/fix-test
We have recently allowed that one can set a vector to zero, even if it has
ghost elements. But this test verified that we throw an exception in this
case. Remove this one test.
Timo Heister [Wed, 22 Oct 2014 20:20:34 +0000 (16:20 -0400)]
Merge pull request #208 from bangerth/augment-trilinos-detection
Also ensure that we have ML as part of Trilinos.
Wolfgang Bangerth [Wed, 22 Oct 2014 18:57:46 +0000 (13:57 -0500)]
Also ensure that we have ML as part of Trilinos.
Wolfgang Bangerth [Wed, 22 Oct 2014 18:51:58 +0000 (13:51 -0500)]
Fix a test.
We have recently allowed that one can set a vector to zero, even if it has
ghost elements. But this test verified that we throw an exception in this
case. Remove this one test.
Denis Davydov [Tue, 21 Oct 2014 17:57:28 +0000 (19:57 +0200)]
Added set_matrix_mode to the SLEPcWrappers::TransformationBase class.
Guido Kanschat [Fri, 17 Oct 2014 08:54:07 +0000 (10:54 +0200)]
Merge pull request #202 from kronbichler/master
Improved AVX-512 vectorization code path. The original version of the AVX-512 vectorization (DEAL_II_VECTORIZATION_LEVEL = 3) was not properly tested. When I did some tests with gcc, it turned out that I had to fix a few issues. Now it compiles fine and the generated assembler code looks good, but there is no hardware out to test it yet.
Martin Kronbichler [Fri, 17 Oct 2014 08:21:56 +0000 (10:21 +0200)]
Improved AVX-512 vectorization code path
The original version of the AVX-512 vectorization (DEAL_II_VECTORIZATION_LEVEL = 3) was not properly tested. When I did some tests with gcc, it turned out that I had to fix a few issues. Now it compiles fine and the generated assembler code looks good, but there is no hardware out to test it yet.
I also improved some code comments.
Martin Kronbichler [Fri, 17 Oct 2014 05:09:36 +0000 (07:09 +0200)]
Merge pull request #201 from bangerth/doc-update-5
Update an error message.
Wolfgang Bangerth [Fri, 17 Oct 2014 02:37:42 +0000 (21:37 -0500)]
Update a comment.
Martin Kronbichler [Thu, 16 Oct 2014 06:20:49 +0000 (08:20 +0200)]
Merge pull request #200 from bangerth/doc-updates-4
Update documentation.
Wolfgang Bangerth [Wed, 15 Oct 2014 19:39:28 +0000 (14:39 -0500)]
More updates.
Wolfgang Bangerth [Wed, 15 Oct 2014 19:26:45 +0000 (14:26 -0500)]
Update documentation.
The text describing the Solver base class must easily be 15 years old.
I can see 3 Germans writing it, accent, language problems, and all.
Timo Heister [Wed, 15 Oct 2014 13:29:00 +0000 (09:29 -0400)]
Merge pull request #199 from bangerth/doc-update-3
Minor doc update for the new manifold_id field.
Wolfgang Bangerth [Wed, 15 Oct 2014 13:17:32 +0000 (08:17 -0500)]
Minor doc update.
We had forgotten to update the documentation for the new manifold_id field.
Also a few other touches of TLC.
Timo Heister [Tue, 14 Oct 2014 21:11:11 +0000 (17:11 -0400)]
Merge pull request #198 from tjhei/gla_broken_ghost_vec
PETSc and Trilinos vectors with ghost entries can now be reset to zero using = 0.0
Timo Heister [Tue, 14 Oct 2014 21:07:29 +0000 (17:07 -0400)]
Allow v=0.0 for ghosted PETSc and Trilinos vectors
Timo Heister [Tue, 14 Oct 2014 20:19:34 +0000 (16:19 -0400)]
Allow v=0.0 for ghosted PETSc and Trilinos vectors
Timo Heister [Tue, 14 Oct 2014 19:53:10 +0000 (15:53 -0400)]
add failing test for ghosted vectors
Wolfgang Bangerth [Tue, 14 Oct 2014 14:59:32 +0000 (09:59 -0500)]
Merge pull request #196 from bangerth/step-52-updates
Step 52 updates
Wolfgang Bangerth [Mon, 13 Oct 2014 23:59:45 +0000 (18:59 -0500)]
Give it a once over.
No functional changes, just updates to documentation and style.
Wolfgang Bangerth [Mon, 13 Oct 2014 23:59:24 +0000 (18:59 -0500)]
Also remove output files.
Wolfgang Bangerth [Mon, 13 Oct 2014 23:59:16 +0000 (18:59 -0500)]
Minor documentation updates.
Wolfgang Bangerth [Fri, 10 Oct 2014 19:11:12 +0000 (14:11 -0500)]
Merge pull request #194 from tjhei/remove_face_it_assert
Remove Assert in CellAccessor::face()
Timo Heister [Fri, 10 Oct 2014 14:44:25 +0000 (10:44 -0400)]
Remove Assert in CellAccessor::face()
The Assert checked that DoFs are distributed before allowing the use of
::face(), which is overly restrictive.
Also see https://groups.google.com/d/topic/dealii/G3Xp-Uv72_M/discussion
Wolfgang Bangerth [Fri, 10 Oct 2014 14:54:42 +0000 (09:54 -0500)]
Merge pull request #191 from tjhei/renumber_dofs_asserts
Check in DoFHandler::renumber_dofs() that DoFs are distributed
Timo Heister [Thu, 9 Oct 2014 14:17:38 +0000 (10:17 -0400)]
Check in DoFHandler::renumber_dofs() that DoFs are distributed
- This adds several Asserts to check that DoFs are distributed
before calling renumber_dofs() for active or level DoFs.
- Note that we now also require that you call distribute_dofs()
before distribute_mg_dofs().
- Updated documentation.
Martin Kronbichler [Fri, 10 Oct 2014 12:13:17 +0000 (14:13 +0200)]
Merge pull request #193 from bangerth/omit-explicit-instantiations
Remove explicit instantiations.
Wolfgang Bangerth [Fri, 10 Oct 2014 12:03:13 +0000 (07:03 -0500)]
Remove explicit instantiations.
Since these functions were moved to the header file, we no longer
need to explicitly instantiate them since every .cc file that uses
them will generate an instantiation where necessary.
Wolfgang Bangerth [Fri, 10 Oct 2014 11:51:39 +0000 (06:51 -0500)]
Merge pull request #192 from vijaysm/vijaysm/tria_objects_release_fix
Making next_free_pair_object and next_free_single_object as inline functions
Vijay Mahadevan [Thu, 9 Oct 2014 16:47:34 +0000 (11:47 -0500)]
Making the next_free_pair_object and next_free_single_object as inline functions.
Also contains a patch from Wolfgang with changes to the instantiation file.
Background: In release mode, GNU-4.8.x (OSX) seem to optimize away the calls when
specializing the template functions and then symbol definitions seem to be different
between debug and optimized modes. This causes undefined linkage for release mode.
References:
[1] https://groups.google.com/forum/#!topic/dealii/c2qjAsuJ7Mc
[2] https://code.google.com/p/dealii/issues/detail?id=137
[3] https://groups.google.com/forum/#!msg/dealii/6RXRR-iac0Y/Tjzwg1rFRMQJ
Bruno Turcksin [Thu, 9 Oct 2014 14:21:08 +0000 (09:21 -0500)]
Merge pull request #190 from tjhei/fix_typo
fix typo in comment
Timo Heister [Thu, 9 Oct 2014 13:34:11 +0000 (09:34 -0400)]
fix typo in comment
Bruno Turcksin [Wed, 8 Oct 2014 14:48:01 +0000 (09:48 -0500)]
Merge pull request #187 from agrayver/petsc_mat_alloc
Preallocation of PETScWrappers::MPI::SparseMatrix with off-diagonal block nonzeros.
Alexander Grayver [Wed, 8 Oct 2014 12:40:45 +0000 (14:40 +0200)]
Merge branch 'petsc_mat_alloc' of https://github.com/agrayver/dealii into petsc_mat_alloc
Alexander Grayver [Fri, 3 Oct 2014 17:57:50 +0000 (19:57 +0200)]
Enable PETScWrappers::MPI::SparseMatrix interface to account for off-diagonal elements during preallocation.
Matthias Maier [Wed, 8 Oct 2014 11:44:52 +0000 (13:44 +0200)]
Merge pull request #188 from andre-massing/massing/fix-mumps-detection-on-ubuntu-14.04
Added additional names for scalapack and blacs related libs on ubuntu 14...
Alexander Grayver [Wed, 8 Oct 2014 09:40:21 +0000 (11:40 +0200)]
Indent files properly.
Andre Massing [Fri, 3 Oct 2014 21:58:25 +0000 (23:58 +0200)]
Added additional names for scalapack and blacs related libs on ubuntu 14.04
Alexander Grayver [Fri, 3 Oct 2014 17:57:50 +0000 (19:57 +0200)]
Enable PETScWrappers::MPI::SparseMatrix interface to account for off-diagonal elements during preallocation.
Wolfgang Bangerth [Wed, 1 Oct 2014 14:23:48 +0000 (09:23 -0500)]
Merge pull request #185 from tamiko/fix_cxx_test
Fix cxx test
Matthias Maier [Wed, 1 Oct 2014 13:10:45 +0000 (15:10 +0200)]
CMake: Bugfix: Also sanitize (cpp) macros in enable_if_links
Matthias Maier [Wed, 1 Oct 2014 13:08:18 +0000 (15:08 +0200)]
CMake: Remove test DEAL_II_HAVE_CXX11_ICCBOOSTSIGNALSBUG_OK
We cannot sanely test for this regression prior to boost setup. And after
boost setup it is incredibly ugly to do so because DEAL_II_WITH_STDCXX11 is
already processed.
Wolfgang Bangerth [Tue, 30 Sep 2014 23:32:14 +0000 (18:32 -0500)]
Merge pull request #153 from bangerth/query-dof-associations
Allow querying whether a DoF is located on a vertex, line, etc.
Wolfgang Bangerth [Wed, 17 Sep 2014 14:42:44 +0000 (09:42 -0500)]
Allow querying whether a DoF is located on a vertex, line, etc.
Wolfgang Bangerth [Tue, 30 Sep 2014 19:58:30 +0000 (14:58 -0500)]
Merge pull request #184 from tamiko/fix_cxx_test
CMake: Only use internal boost header (if available) for regression check
Matthias Maier [Tue, 30 Sep 2014 18:46:11 +0000 (20:46 +0200)]
CMake: Only use internal boost header (if available) for regression check
Fix a regression where a test for an icc regression in
check_01_cxx_features.cmake incorrectly used a generic <boost/signals2/...>
include. Such an include file doesn't have to be available at this early
point in the configuration run. User internal boost headers (if available)
instead - otherwise just test nothing.
Wolfgang Bangerth [Mon, 29 Sep 2014 18:59:48 +0000 (13:59 -0500)]
Merge pull request #177 from bangerth/document-ghosted-vectors
Provide better documentation on ghosted vectors.
Wolfgang Bangerth [Fri, 26 Sep 2014 23:39:29 +0000 (18:39 -0500)]
Provide better documentation on ghosted vectors.
This includes a discussion of what ghosted vectors are and what their
semantics are.
Martin Kronbichler [Mon, 29 Sep 2014 16:13:51 +0000 (18:13 +0200)]
Merge pull request #183 from bangerth/harden-mpi-reductions
Harden mpi reductions
Wolfgang Bangerth [Mon, 29 Sep 2014 15:24:58 +0000 (10:24 -0500)]
Add a bunch of missing tests for Utilities::MPI::max.
Wolfgang Bangerth [Mon, 29 Sep 2014 15:12:57 +0000 (10:12 -0500)]
MPI_Allreduce is apparently unable to accept input and output arguments at the same location.
Fix this by determining when &input=&output and in that case pass MPI_IN_PLACE.
Matthias Maier [Mon, 29 Sep 2014 14:51:55 +0000 (16:51 +0200)]
Merge pull request #181 from bangerth/add-fusion-to-boost
Add fusion to boost
Wolfgang Bangerth [Mon, 29 Sep 2014 14:42:36 +0000 (09:42 -0500)]
Merge pull request #182 from kronbichler/master
Circumvent a bug in Trilinos 11.10
Martin Kronbichler [Mon, 29 Sep 2014 14:35:54 +0000 (16:35 +0200)]
Circumvent a bug in Trilinos 11.10
If we avoid calling GlobalAssemble when copying from an Epetra_FECrsMatrix that uses a
non-local matrix (enabled for Trilinos 11.9 and greater), we do not run into some bad
code in Trilinos causing a segfault. At this point of the code, it is enough if we only
call FillComplete because there should be no communication after all.
Wolfgang Bangerth [Mon, 29 Sep 2014 11:38:20 +0000 (06:38 -0500)]
Add a test to avoid this problem from happening again in the future.
Wolfgang Bangerth [Mon, 29 Sep 2014 11:28:12 +0000 (06:28 -0500)]
Add BOOST.fusion as apparently we need that in boost/math/special_functions/erf.hpp.
Toby D. Young [Sun, 28 Sep 2014 22:03:34 +0000 (00:03 +0200)]
Merge pull request #180 from bangerth/comment-on-spurious-eigenvalues
Comment on spurious eigenvalues
Wolfgang Bangerth [Sun, 28 Sep 2014 21:00:54 +0000 (16:00 -0500)]
Add a changelog entry.
Wolfgang Bangerth [Sun, 28 Sep 2014 18:38:50 +0000 (13:38 -0500)]
Extend discussion about spurious eigenvalues.
Wolfgang Bangerth [Sun, 28 Sep 2014 18:38:02 +0000 (13:38 -0500)]
Rename a couple of variables and extend discussion.
Wolfgang Bangerth [Sun, 28 Sep 2014 18:05:49 +0000 (13:05 -0500)]
Reindent and fix the issues discussed in the original pull request.
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.