]>
https://gitweb.dealii.org/ - dealii.git/log
Luca Heltai [Sat, 6 Jun 2015 12:51:36 +0000 (14:51 +0200)]
Merge pull request #992 from ESeNonFossiIo/null_operator
Null operator
ESeNonFossiIo [Wed, 3 Jun 2015 14:09:55 +0000 (16:09 +0200)]
is_null_operator
ESeNonFossiIo [Wed, 3 Jun 2015 14:00:21 +0000 (16:00 +0200)]
null_operator for rectangular and square matrices
Wolfgang Bangerth [Wed, 3 Jun 2015 15:57:01 +0000 (08:57 -0700)]
Merge pull request #985 from drwells/remove-unused-private-members
Silence warnings about unused private members.
Bruno Turcksin [Wed, 3 Jun 2015 12:37:54 +0000 (07:37 -0500)]
Merge pull request #987 from angelrca/vtools_interp_fenot
VectorTools::interpolate changed to allow FE_Nothing
Luca Heltai [Wed, 3 Jun 2015 10:17:18 +0000 (12:17 +0200)]
Merge pull request #990 from ESeNonFossiIo/upper_triangular_block_bug
a bug fixed
angelrca [Wed, 3 Jun 2015 09:56:56 +0000 (11:56 +0200)]
changes.h, blocks moved
angelrca [Wed, 3 Jun 2015 09:52:38 +0000 (11:52 +0200)]
interpolate_nothing_03.cc date changed
angelrca [Wed, 3 Jun 2015 09:29:22 +0000 (11:29 +0200)]
test case interpolate_nothing_03 added
angelrca [Wed, 3 Jun 2015 08:43:36 +0000 (10:43 +0200)]
doc/news/changes.h line added
ESeNonFossiIo [Wed, 3 Jun 2015 08:32:36 +0000 (10:32 +0200)]
a test for upper triangular matrix with non trivial block matrices
angelrca [Wed, 3 Jun 2015 07:49:47 +0000 (09:49 +0200)]
indentation corrected
David Wells [Tue, 2 Jun 2015 21:22:22 +0000 (17:22 -0400)]
Silence a warning about an unused private member.
The GmvFlags struct is not used in the current implementation, but it
may be used in the future.
David Wells [Tue, 2 Jun 2015 21:19:43 +0000 (17:19 -0400)]
Remove 'dummy' members and allow empty structs.
Empty structs are implementation-dependent in C, but are acceptable in
C++. Therefore unused private members are not needed (and they cause
clang to generate warnings).
ESeNonFossiIo [Tue, 2 Jun 2015 21:04:19 +0000 (23:04 +0200)]
a bug fixed
Luca Heltai [Tue, 2 Jun 2015 20:30:36 +0000 (22:30 +0200)]
Merge pull request #988 from tjhei/intel_simd
intel and openmp-simd
Luca Heltai [Tue, 2 Jun 2015 20:29:42 +0000 (22:29 +0200)]
Merge pull request #989 from tjhei/muparser_v2_2_4
update muparser to 2.2.4
Timo Heister [Tue, 2 Jun 2015 18:52:32 +0000 (14:52 -0400)]
update muparser to 2.2.4
Timo Heister [Tue, 2 Jun 2015 17:36:41 +0000 (13:36 -0400)]
intel and openmp-simd
- Intel 13 does not support -openmp-simd and generates a bogus warning
(confusing it with -o). Do not try to check for it.
- Intel 15 renamed the flag to -qopenmp-simd (see
https://software.intel.com/en-us/node/522965)
angelrca [Tue, 2 Jun 2015 16:50:43 +0000 (18:50 +0200)]
vector_tools::interpolation identation corrected
angelrca [Tue, 2 Jun 2015 15:34:01 +0000 (17:34 +0200)]
Merge branch 'vtools_interp_fenot' of https://github.com/angelrca/dealii into vtools_interp_fenot
angelrca [Tue, 2 Jun 2015 14:32:24 +0000 (16:32 +0200)]
VectorTools::interpolate changed
angelrca [Tue, 2 Jun 2015 14:32:24 +0000 (16:32 +0200)]
VectorTools::interpolate changed
Luca Heltai [Tue, 2 Jun 2015 12:54:29 +0000 (14:54 +0200)]
Merge pull request #986 from ESeNonFossiIo/null_operator
null operator
ESeNonFossiIo [Tue, 2 Jun 2015 10:39:38 +0000 (12:39 +0200)]
null operator
Wolfgang Bangerth [Mon, 1 Jun 2015 12:23:35 +0000 (07:23 -0500)]
Merge pull request #983 from drwells/memory-usage-strlen-header
Fix two issues with C string memory consumption.
David Wells [Mon, 1 Jun 2015 12:02:49 +0000 (08:02 -0400)]
Fix two issues with C string memory consumption.
These were pointed out by @msteigemann.
Including cstring to obtain strlen is not necessary with some GCC
releases, but may be needed on other platforms.
I rewrote the call to strlen so that it does not rely on the fact that
sizeof(char) == 1 on most platforms.
Wolfgang Bangerth [Mon, 1 Jun 2015 03:03:47 +0000 (22:03 -0500)]
Merge pull request #982 from drwells/refactor-data-out-flags-base-class
Refactor data out flags base class
David Wells [Sun, 31 May 2015 22:52:04 +0000 (18:52 -0400)]
Add an '@ingroup output' label to an output flag.
David Wells [Sun, 31 May 2015 22:41:08 +0000 (18:41 -0400)]
Consistently label constructors.
David Wells [Sun, 31 May 2015 22:39:56 +0000 (18:39 -0400)]
Delete extra 'public:' labels.
Things in structs are public by default.
David Wells [Sun, 31 May 2015 22:36:58 +0000 (18:36 -0400)]
Remove default constructor implementations.
David Wells [Sun, 31 May 2015 22:33:30 +0000 (18:33 -0400)]
Switch to just spaces, instead of tabs and spaces.
David Wells [Sun, 31 May 2015 22:28:45 +0000 (18:28 -0400)]
Replace all set_flags functions with a template.
This sacrifices some type safety in optimized mode to decrease
redundancy.
David Wells [Sun, 31 May 2015 22:19:04 +0000 (18:19 -0400)]
Remove empty declare and parse parameter methods.
The default implementation (also an empty block) is covered by the base
class.
David Wells [Sun, 31 May 2015 22:11:39 +0000 (18:11 -0400)]
Improve the TecplotFlags memory estimate.
David Wells [Sun, 31 May 2015 21:58:24 +0000 (17:58 -0400)]
Remove extra memory_consumption methods.
The base class (introduced by the previous commit) implements the
behavior common to all of these methods.
David Wells [Sun, 31 May 2015 21:50:18 +0000 (17:50 -0400)]
Add a base class for all output flag classes.
The base class uses "the curiously recurring template pattern" technique
so that they can all share a default implementation of memory
consumption (i.e., something akin to "return sizeof(*this)" works for
child classes unless overridden).
Wolfgang Bangerth [Sun, 31 May 2015 02:10:44 +0000 (21:10 -0500)]
Merge pull request #981 from tjhei/inst_mismatch
fix instantiation of subdivided_hyper_rectangle
Timo Heister [Sun, 31 May 2015 01:41:04 +0000 (21:41 -0400)]
fix instantiation of subdivided_hyper_rectangle
intel compiler complains about a missing instantiation because of a missing const
Wolfgang Bangerth [Sun, 31 May 2015 01:23:45 +0000 (20:23 -0500)]
Merge pull request #980 from tjhei/fix_deprecated_statement
fix placement of deprecation keyword
Timo Heister [Sun, 31 May 2015 01:11:59 +0000 (21:11 -0400)]
fix placement of deprecation keyword
Guido Kanschat [Fri, 29 May 2015 22:16:20 +0000 (00:16 +0200)]
Merge pull request #978 from davydden/local_integrators2
exploit symmetries in Local integrators of mass and laplace matrices
avoid copying the matrix entries and thus side effects
closes issue #976
Denis Davydov [Fri, 29 May 2015 20:24:02 +0000 (22:24 +0200)]
accumulate sums over n_components
Denis Davydov [Fri, 29 May 2015 18:59:10 +0000 (20:59 +0200)]
exploit symmetries in Local integrators of mass and laplace matrices without assignment operation
Wolfgang Bangerth [Fri, 29 May 2015 17:52:16 +0000 (10:52 -0700)]
Merge pull request #975 from guidokanschat/documentation_relaxation_block
Add some documentation on the special structure of RelaxationBlock::A…
Guido Kanschat [Fri, 29 May 2015 15:52:51 +0000 (17:52 +0200)]
Add some documentation on the special structure of RelaxationBlock::AdditionalData
Martin Kronbichler [Fri, 29 May 2015 06:35:16 +0000 (08:35 +0200)]
Merge pull request #972 from bangerth/fix-test-4
Fix a test where we may inadvertently divide by zero.
Guido Kanschat [Fri, 29 May 2015 05:46:35 +0000 (07:46 +0200)]
Merge pull request #973 from bangerth/fix-condition
Use a standard method to find whether a number is normal.
Wolfgang Bangerth [Thu, 28 May 2015 20:37:37 +0000 (15:37 -0500)]
Use a standard method to find whether a number is normal.
In the current context, this slightly changes the behavior in that the new
condition also catches infinities, rather than just NaNs as before, but the
downstream code will still yield the same answers.
Wolfgang Bangerth [Thu, 28 May 2015 20:26:21 +0000 (15:26 -0500)]
Fix a test where we may inadvertently divide by zero.
The test is also just overly convoluted.
Luca Heltai [Thu, 28 May 2015 17:41:23 +0000 (19:41 +0200)]
Merge pull request #971 from bangerth/fix-test-3
Fix test.
Wolfgang Bangerth [Thu, 28 May 2015 16:35:15 +0000 (11:35 -0500)]
Fix test.
The test uses acosh(0.5), but acosh is only defined for arguments >=1.
Wolfgang Bangerth [Thu, 28 May 2015 16:01:03 +0000 (12:01 -0400)]
Merge pull request #970 from guidokanschat/fix_newton_threshold
fix glitch with assemble_threshold, avoid division by zero
Luca Heltai [Thu, 28 May 2015 15:13:21 +0000 (17:13 +0200)]
Merge pull request #969 from tamiko/rename_two_more_header_guards
Rename two more header guards
Matthias Maier [Thu, 28 May 2015 14:00:32 +0000 (16:00 +0200)]
Merge pull request #759 from guidokanschat/fix_arpack_vectors
Fix Arpack solver eigenvectors
Guido Kanschat [Thu, 28 May 2015 12:58:36 +0000 (14:58 +0200)]
fix glitch with assemble_threshold, avoid division by zero
Matthias Maier [Thu, 28 May 2015 12:04:20 +0000 (14:04 +0200)]
Rename two more header guards
Wolfgang Bangerth [Thu, 28 May 2015 06:24:35 +0000 (02:24 -0400)]
Merge pull request #967 from guidokanschat/fix_clear_in_relaxation_block
RelaxationBlock::clear() did not clear additional_data; fixed now
Guido Kanschat [Thu, 28 May 2015 05:50:05 +0000 (07:50 +0200)]
Merge pull request #966 from bangerth/fix-test-2
Fix a test output.
Guido Kanschat [Thu, 28 May 2015 05:29:31 +0000 (07:29 +0200)]
RelaxationBlock::clear() did not clear additional_data; fixed now
Wolfgang Bangerth [Thu, 28 May 2015 04:08:22 +0000 (23:08 -0500)]
Fix a test output.
Luca Heltai [Thu, 28 May 2015 00:01:36 +0000 (02:01 +0200)]
Merge pull request #964 from bangerth/fix-test
Fix a test.
Luca Heltai [Wed, 27 May 2015 23:58:03 +0000 (01:58 +0200)]
Merge pull request #961 from tamiko/refactor_block_linear_operator
Refactoring and a bug fix for LinearOperator
Wolfgang Bangerth [Wed, 27 May 2015 23:49:15 +0000 (18:49 -0500)]
Fix a test.
The test calls acosh(0.5), but acosh is really only defined for arguments >=1.
Matthias Maier [Wed, 27 May 2015 20:11:55 +0000 (22:11 +0200)]
Write a single-line statement in two lines
Matthias Maier [Wed, 27 May 2015 20:07:23 +0000 (22:07 +0200)]
Reword a description for a workaround
Matthias Maier [Wed, 27 May 2015 12:24:42 +0000 (14:24 +0200)]
Refactoring III: Reintroduce block operations and tests
This reintroduces all block operations in a new header
"block_linear_operator.h". Tests renamed accordingly.
Matthias Maier [Wed, 27 May 2015 12:23:37 +0000 (14:23 +0200)]
Refactoring II: Reintroduce PackagedOperation and related tests
This reintroduces PackagedOperation and related tests under a new header
file "packaged_operation.h". Tests renamed accordingly.
Matthias Maier [Wed, 27 May 2015 12:21:47 +0000 (14:21 +0200)]
Refactoring I: Remove PackagedOperation and block stuff
Remove all PackagedOperation related code and code for block structures
from linear_operator.h, remove corresponding tests and rename remaining
tests.
Matthias Maier [Wed, 27 May 2015 10:53:58 +0000 (12:53 +0200)]
Workaround for a regression in <=gcc-4.7 with partial template default values
Wolfgang Bangerth [Wed, 27 May 2015 15:54:45 +0000 (11:54 -0400)]
Merge pull request #962 from tamiko/rename_header_guards
Rename header guards from __deal2__ to dealii__
Matthias Maier [Wed, 27 May 2015 12:37:28 +0000 (14:37 +0200)]
Rename header guards from __deal2__ to dealii__
Matthias Maier [Wed, 27 May 2015 10:17:21 +0000 (12:17 +0200)]
Merge pull request #960 from ESeNonFossiIo/linear_operator_fix
add template
ESeNonFossiIo [Wed, 27 May 2015 09:58:52 +0000 (11:58 +0200)]
add template
Wolfgang Bangerth [Wed, 27 May 2015 06:48:57 +0000 (02:48 -0400)]
Merge pull request #957 from QiaoLei-88/fixDynamicSparsityPatternIterators
fix undefined behavior DynamicSparsityPatternIterators::Accessor::operator <
Martin Kronbichler [Wed, 27 May 2015 06:28:45 +0000 (08:28 +0200)]
Merge pull request #959 from bangerth/silence-warnings
Silence warnings about unused arguments.
Wolfgang Bangerth [Wed, 27 May 2015 02:14:37 +0000 (21:14 -0500)]
Silence warnings about unused arguments.
Also explain why we don't need the argument.
Lei Qiao [Sat, 23 May 2015 22:47:57 +0000 (17:47 -0500)]
fix undefined behavior DynamicSparsityPatternIterators::Accessor::operator <
Luca Heltai [Sun, 24 May 2015 12:21:04 +0000 (14:21 +0200)]
Merge pull request #950 from ESeNonFossiIo/block_lin_op_inv
Block lin op inv
ESeNonFossiIo [Thu, 21 May 2015 07:27:24 +0000 (09:27 +0200)]
documentation
indentation and an assert
layout
a small error
documentation
small bugs + rebase
small change in the comment
notation
a small comment
renaming the function
a -> block_matrix
ESeNonFossiIo [Wed, 20 May 2015 16:52:01 +0000 (18:52 +0200)]
preconditioner_from_diagonal_inverse
preconditioner
inverse check added
ESeNonFossiIo [Wed, 20 May 2015 13:36:44 +0000 (15:36 +0200)]
upper triangular block matrix
upper triangular block matrix tests
comments
crossed tests
small typos
ESeNonFossiIo [Wed, 20 May 2015 08:18:44 +0000 (10:18 +0200)]
lower triangular block matrix
lower triangular matrix tests
Wolfgang Bangerth [Sat, 23 May 2015 05:12:03 +0000 (01:12 -0400)]
Merge pull request #947 from msteigemann/branch_petsc_vector
Add a clear function for PETSc serial and parallel vectors
msteigemann [Thu, 21 May 2015 05:53:24 +0000 (07:53 +0200)]
Add a clear function for PETSc serial and parallel vectors
Matthias Maier [Fri, 22 May 2015 23:58:00 +0000 (01:58 +0200)]
Merge pull request #953 from bangerth/check-astyle-version
Check that we have the right astyle version.
Wolfgang Bangerth [Fri, 22 May 2015 03:58:50 +0000 (22:58 -0500)]
Check that we have the right astyle version.
Wolfgang Bangerth [Thu, 21 May 2015 17:56:49 +0000 (13:56 -0400)]
Merge pull request #942 from mtezzele/hp_ref_bernstein
hp refinement for bernstein
Marco Tezzele [Mon, 18 May 2015 09:19:12 +0000 (11:19 +0200)]
hp refinement for bernstein
Timo Heister [Thu, 21 May 2015 04:38:39 +0000 (00:38 -0400)]
Merge pull request #949 from bangerth/fix-petsc-gla-test
Rewrite test.
Wolfgang Bangerth [Thu, 21 May 2015 03:57:59 +0000 (22:57 -0500)]
Rewrite test.
The test uses PETScWrappers::Vector, but this class is now deprecated and indeed
produces an error now (because the respective include file has been removed
from the generic_linear_algebra.h include file.
Wolfgang Bangerth [Thu, 21 May 2015 02:58:29 +0000 (22:58 -0400)]
Merge pull request #943 from davydden/local_integrators
exploit symmetries in Local integrators of mass and laplace matrices
Martin Kronbichler [Wed, 20 May 2015 10:51:42 +0000 (12:51 +0200)]
Merge pull request #945 from drwells/modify-stokes-diffusion
Modify the default Stokes equation in step-22
Luca Heltai [Wed, 20 May 2015 07:50:29 +0000 (09:50 +0200)]
Merge pull request #944 from mtezzele/failing_mapfe
Failing test for mapping_fe_field with fe_bernstein
Luca Heltai [Tue, 19 May 2015 21:10:44 +0000 (23:10 +0200)]
Merge pull request #946 from tamiko/fix_documentation
Documentation: Update forgotten change in cmake.html about default locations
Matthias Maier [Tue, 19 May 2015 20:17:53 +0000 (22:17 +0200)]
Documentation: Update forgotten change in cmake.html about default locations
David Wells [Tue, 19 May 2015 16:05:11 +0000 (12:05 -0400)]
Explain the sparsity pattern changes in step-22.
In particular, step-22 has used three different sparsity patterns at
different points.
David Wells [Tue, 19 May 2015 16:01:39 +0000 (12:01 -0400)]
Update the step-22 benchmark as well.
David Wells [Tue, 19 May 2015 15:45:19 +0000 (11:45 -0400)]
Update the step-22 numerical results.
The solution difference norms are significantly different because of the
additional factor of two. I recovered the original differences (up to
about 10%) by removing the factor.
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.