- Dec 09, 2015
-
-
3b89e771 Drop support for compilers without `long long` 36bb09a7 Drop support for compilers with `__int64` as the only 64-bit int Acked-by:
Kitware Robot <kwrobot@kitware.com> Acked-by:
Berk Geveci <berk.geveci@kitware.com> Merge-request: !981
-
David E DeMarle authored
modernize depth sorting code modernize depth sorting code The following improvements were made to vtkDepthSortPolyData: * transfrom GetCell to GetCellPoints. Building the cells one by one is expensive and we only need points to determine the depth. We can also then access the points in place. * transform qsort to std::sort. Comparisons in std::sort get inlined so std::sort is much faster. I also reduced the overhead of swaps by using a functor so that only cell id is swapped. * use templates to deal with point types instead of going through virtual GetValue API that converts them to double. * restructure depth computations so that they can be vectorized by the compiler. * I added a Cxx test to improve test coverage of the various sorting modes supported by this class. The following improvements were made to vtkPolyData and vtkCellTypes: * move the non-virtual impementations of GetCellPoints to the header to improve compiler optimization. * add non-virtual GetCell method and also make it inline * fix an issue in cell types where the MaxId is not set correctly when you pass in pre-built arrays In tests the improved depth sort is ~ 3 x faster when sorting by first point and cell bounds, and ~ 2 x faster when sorting by parametric center. The tests made use of a ~ 8.1M cell iso-surface computed from a cosmology simulation, and gcc 4.9.2 compiled with -Wall -Wextra -DNDEBUG -Ofast -march=native -mavx -ffast-math  code | point | bounds | param -----|-------|--------|------ old | 2.58 sec | 3.05 sec | 3.19 sec new | 0.82 sec | 1.02 sec | 1.65 sec speed up | 3.14 x | 2.98 x | 1.93 x See merge request !844
-
David E DeMarle authored
Fix vtkCaptionActor2D not scaling text The text in vtkCaptionActor2D should scale to fill the rectangle defined by its positions. The scaling works as expected only if the actor's CaptionTextProperty is modified; using the default CaptionTextProperty shows the bug. This commit makes the text scale even when using the default caption text property. Fixes #15412 See merge request !802
-
a1cc14f4 Remove lingering debugging code Acked-by:
-
- Dec 08, 2015
-
-
Ken Martin authored
Forgot to remove some debug code I was using to test something
-
David E DeMarle authored
fix valgrind issue with the test itself @ben.boeckel @jpouderoux Will fix https://gitlab.kitware.com/demarle/vtk/merge_requests/new?merge_request[source_branch]=fix-delauney2d-test-memleak See merge request !982
-
f7bcf8f3 EncodeString: Never build a launcher for this executable Acked-by:
Ben Boeckel <ben.boeckel@kitware.com> Merge-request: !983
-
Brad King authored
The dependency on `vtksys` dropped by commit 708c3a5c (EncodeString: Compile independently from other VTK modules, 2015-12-04), was still needed to support `VTK_BUILD_FORWARDING_EXECUTABLES`. However, we do not actually need to build a launcher for `vtkEncodeString` because it does not link to anything.
-
Brad King authored
VTK no longer supports any compilers that do not provide this type. Therefore all code conditional on VTK_TYPE_USE_LONG_LONG can be made unconditional. Leave the macro defined to tell dependent projects that APIs using the type are available in case they still support versions of VTK that make it conditional.
-
Burlen Loring authored
The following improvements were made to vtkDepthSortPolyData: * transfrom GetCell to GetCellPoints. Building the cells one by one is expensive and we only need points to determine the depth. We can also then access the points in place. * transform qsort to std::sort. Comparisons in std::sort get inlined so std::sort is much faster. I also reduced the overhead of swaps by using a functor so that only cell id is swapped. * use templates to deal with point types instead of going through virtual GetValue API that converts them to double. * restructure depth computations so that they can be vectorized by the compiler. * I added a Cxx test to improve test coverage of the various sorting modes supported by this class. The following improvements were made to vtkPolyData and vtkCellTypes: * move the non-virtual impementations of GetCellPoints to the header to improve compiler optimization. * add non-virtual GetCell method and also make it inline * fix an issue in cell types where the MaxId is not set correctly when you pass in pre-built arrays In tests the improved depth sort is ~ 3 x faster when sorting by first point and cell bounds, and ~ 2 x faster when sorting by parametric center. The tests made use of a ~ 8.1M cell iso-surface computed from a cosmology simulation, and gcc 4.9.2 compiled with -Wall -Wextra -DNDEBUG -Ofast -march=native -mavx -ffast-math The patch includes the changes requested in the code review: transform malloc/free to new[]/delete[], and transform memset to std::fill_n.
-
Dave DeMarle authored
-
9a712588 When pwf range is much smaller than array range, vtk crashes. Acked-by:
Ken Martin <ken.martin@kitware.com> Merge-request: !972
-
David E DeMarle authored
Optimizing variant out of contingency See merge request !537
-
816a0f60 ENH: Unit test for MergeFilter Acked-by:
Cory Quammen <cory.quammen@kitware.com>
Acked-by: -
Ben Boeckel authored* gl/dgobbi/libpng-1.0.65: png: update to for/vtk Add the update.sh for third-party libpng. Remove old vtkpng, replace with upstream 1.0.65. Reviewed-by:Ben Boeckel <mathstuf@gmail.com> Merge-request: !977
-
902356a9 Fix "same expression in both branches of ternary operator". Acked-by:
David E DeMarle <dave.demarle@kitware.com> Merge-request: !980
-
Brad King authored
VTK no longer supports any compilers that have `__int64` but not a `long` or `long long` that is 64-bit. Therefore all code that is conditional on VTK_TYPE_USE___INT64 is never used and can be dropped.
-
5f18d211 Reset helper to use piece extents in serial ExtractVOI impls. Acked-by:
Andrew Bauer <andy.bauer@kitware.com> Reviewed-by:
Utkarsh Ayachit <utkarsh.ayachit@kitware.com> Merge-request: !976
-
7bbbba26 ParseOGLExt: Compile independently from other VTK modules 708c3a5c EncodeString: Compile independently from other VTK modules Acked-by:
-
Utkarsh Ayachit authored
Fixes issue caught by cppcheck. Updating the code based on the comment which suggests the intention of the developer.
-
- Dec 07, 2015
-
-
David Gobbi authored
-
Code extracted from: https://gitlab.kitware.com/third-party/png.git at commit 63adbb10997c492ec17eaab6f80734b8f988605a. -
David Gobbi authored
-
-
c9f7a5e0 vtkStreamer: deprecate the class hierarchy Acked-by:
-
8c3c7940 Revert "ThirdParty: always use a subtree option for merging" Acked-by:
David Gobbi <david.gobbi@gmail.com> Merge-request: !979
-
Ben Boeckel authored
The simpler code only works if the target directory already exists in the target location. Instead, go back to the old merge/read-tree dance which always works for initializing the directory. This reverts commit 127573c9.
-
Scott Wittenburg authoredIn ParaView, we have to supply the ui with pwf instances for the opacity and size mapping. The default pwf range is [0,0] to [1,1]. If you choose to size by an array with a range much larger than this, ParaView crashes. You can get around it by first open the pwf editor and adjusting the x range to match an array you would like to work with. This fix could cause the mapper to be too slow, so I would like to solicit other, more efficient approaches.
-
1e5594fa Change GeoJSON reader to parse LineString features into vtkPolyLine instances Acked-by:
-
Brad King authored
Do not depend on headers from vtkCommonCore.
-
Brad King authored
Do not depend on headers from vtkCommonCore or vtksys.
-
6495a540 Define target before setting target_compile_features Acked-by:
-
David C. Lonie authored
-
7e8454b9 vtkFiltersStatisticsGnuR: Simplify availability of uintptr_t Acked-by:
-
55f16689 Fix a bug with vtkDelaunay2D. ac7a28f9 Fix a bug with vtkDelaunay2D. Acked-by:
-
Joachim Pouderoux authored
In a previous patch, a mechanism was added to check edges once constrained edges were applied. However, in some cases, this check can break some constrained edges. This patch make sure to not perform this edge check for triangle edges which contains a constrained edge.
-
- Dec 05, 2015
-
-
200a974f Minor variable redec warning Acked-by:
-
Bill Lorensen authored
MergeFilter test coverage was less than 50%. This patch provides a unit test that covers 100% of the code. NOTE: While creating the unit test, a bug was found in vtkMergeFilter. Also the patch adds a number of warnings to alert users when the merge data is not compatible with the input geometry. BUG: MergeFilter was not copying added arrays properly. The test for adding an array was incorrect for cell data. -
Ken Martin authored
Fix edge flag being set and declared twice
-
- Dec 04, 2015
-
-
a4275ab0 Documentation: add docs for making a release Acked-by:
-
