GitLab

Using v0+w*(v1-v0) can be numerically unstaqble due to floating point
arithmatic errors.  Additionaly, the newer form also resolves type
conversion warnings.

3f7d9429 Disable vectorization on debug Intel Compiler builds.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !363

8818cfc6 Disable vectorization when building with ICC 14.0
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !362

73b90a2e Don't provide vectorization hints for operations that can be done inplace.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !360

81147365 fix a long long to int warning.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !361

The intel compiler generates internal compiler errors when you use
pragma simd and haven't defined NDEBUG.

We have found that ICC 14.0 produces bad simd code that causes
vtkm worklets to throw SIGBUS signals. This issue was resolved in
ICC 15.0

Previously we hinted to the compiler that it should vectorized operations
where the input and output are the same array. This obviously caused problems,
and these hints had to be removed.

In the future we need to first check for aliased arrays, and go from there.

85084f2c ScatterIdentity::GetVisitArray parameters are now named properly
86ecad65 ScatterIdentity::GetOutputToInputMap parameters are now named properly
40896e2b Allocate the scatter arrays to be proper length.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Kenneth Moreland <kmorel@sandia.gov>
Merge-request: !359

Now it is clear you pass in the input range and you will get the output
range

Now it is clear you pass in the input range and you will get the output
range.

The previous code would over allocate all the scatter arrays since
it was computing the output range, than using that as the input range.

e4237c3a Fix warnings found by the dashboard machines.
179b48e0 Reduce compile time for MarchingCubes by passing less info by Invoke.
8e4a47ef Update IsosurfaceUniformGrid to use the marching cubes filter.
d370155e MarchingCubes filter now generates coordinates when point merging is enabled.
c00fb53b Marching Cubes now generates vertices when merge duplicates is enabled.
f699c986 Renamed the ```Convert``` method to ```ApplyPolicy```
8e72ec8e Switch filter::threshold over to have a lower and upper bounds.
bcee8270 First draft of vtkm::filter design.
...
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !311

8683240b  vtkm::exec::FunctorBase now properly initializes ErrorMessageBuffer.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !358

06424786 Correct implicit function test failures when vectorization is enabled.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !356

ee4e490f Moved SMP atomic operations from TBB to General.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !357

When vectorization is enabled we start using 'fast' math, and therefore instead
of getting perfect 0.0 values we get values that are extremely close to zero.

f01fd68b DataSetBuilderUniform now builds when the float default is double.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !355

Even when running with the serial backend, the compiler might enable SIMD
vectorization when optimizations are turned on. When this occurs, we need
to use properly atomic Add's and CAE's.

When the number of parameters to a worklet hits 6+ like the EdgeGenerate
worklet did, the compile times when some of those parameters are dynamic array
handle is very bad.

For example when building the MarchingCubes filter for VTK integration the
EdgeGenerate worklet would take 800+ seconds to compile, but when we switch
it to have fewer arguments the compile time comes down to ~250sec.

This way we have an example of how to use the filter code.

The original name was a really bad name, so thank to Ken for suggesting
a better name.

The original implementation wasn't flexible enough to handle the requirements
that filter requires ( mainly policy support ).

This isn't needed now that vtkm::cont::Field can compute bounds.

5ddade7a Adding some basic documentation on atomics.
e5c4aa3f Fixing cuda index error
00a7f6c1 Correcting function call
249cce35 Adding type restrictions to serial atomics
43131ee0 Adding comments about CAS
40b6db7e Inserted missing ,
3b46706e Adding compare and swap and removing unsigned atomics
12632de3 Casting Uint64 to long long
...
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !336

19dc5672 Correct CUDA fallback to "fermi" when trying to do auto detection.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !353

55280637 Correct the VTKM_THIRDPARTY_PRE_INCLUDE/POST_INCLUDE for MSVC.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !354

We now properly suppress a collection of warnings from thirdparty headers on
windows.