049d0cca8 cmake: namespace vtkm export targets
d4fc84f12 CI: CUDA build jobs dont need cuda-rt
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !2939
The PerlinNoise source has a mode where if a seed is not set, it will
choose a new seed every time it is executed. It did this by using the
value 0 as an indicator to do this (and initializing the Seed to 0).
However, there was a problem with one of the benchmarks that was
specifically setting the seed to 0 and getting unexpected results.
Fix the problem by adding a separate, hidden member of the PerlinNoise
class that keeps track of whether to generate new seeds or not.
This should help catch regressions that only affect a single filter. It
also allows us to set up a partitioned test that has many small blocks.
It also fixes an issue we were seeing on the dashboard where the
contour benchmark was apparently running out of resources.
Added a name option that allows the same benchmark executable to be used
in multiple benchmark tests. This allows the benchmarks to be separated.
Also added an option to pass customized arguments to the benchmark
executable to overwrite the default values.
aa7b83bb2 Handle random seed generation better for PerlinNoise
84bc72312 Make source parameters more clear
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Li-Ta Lo <ollie@lanl.gov>
Merge-request: !2933
During the VTK-m 1.8 and 1.9 development, the filter infrastructure was
overhauled. Part of this created a completely new set of base classes. To
avoid confusion with the original filter base classes and ease transition,
the new filter base classes were named `NewFilter*`. Eventually after all
filters were transitioned, the old filter base classes were deprecated.
With the release of VTK-m 2.0, the old filter base classes are removed. The
"new" filter base classes are no longer new. Thus, they have been renamed
simply `Filter` (and `FilterField`).
Originally, most of the sources used constructor parameters to set the
various options of the source. Although convenient, it was difficult to
keep track of what each parameter meant. To make the code more clear,
source parameters are now set with accessor functions (e.g.
`SetPointDimensions`). Although this makes code more verbose, it helps
prevent mistakes and makes the changes more resilient to future changes.
19f752960 Remove Filter::CreateResult that takes a vector of CoordinateSystems
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Li-Ta Lo <ollie@lanl.gov>
Merge-request: !2934
This method is principally obsolete and should have been removed when
coordinate systems were converted into fields. I noticed it was still
there when updating the documentation.
This removal required some changes to the CleanGrid filter, but this
actually simplified the code.
a4435a6b2 Change wavelet dim to 256 and numPart to 1.
fecc703a0 Add header for vtkm::cont::PartitionedDataSet
90586d36d Change default waveletdim back to 256.
b7605689a Add multiblock benchmarks for filters.
4ca68d9c0 Merge branch 'master' of https://gitlab.kitware.com/vtk/vtk-m
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !2930
With the major revision 2.0 of VTK-m, many items previously marked as
deprecated were removed. If updating to a new version of VTK-m, it is
recommended to first update to VTK-m 1.9, which will include the deprecated
features but provide warnings (with the right compiler) that will point to
the replacement code. Once the deprecations have been fixed, updating to
2.0 should be smoother.
4ca64f0f0 Back out of DataSet::CopyPartsFromExcept method
2d30e6d45 Coordinate systems are stored as Fields
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Li-Ta Lo <ollie@lanl.gov>
Merge-request: !2920
Previously, `DataSet` managed `CoordinateSystem`s separately from `Field`s.
However, a `CoordinateSystem` is really just a `Field` with some special
attributes. Thus, coordiante systems are now just listed along with the
rest of the fields, and the coordinate systems are simply strings that
point back to the appropriate field. (This was actually the original
concept for `DataSet`, but the coordinate systems were separated from
fields for some now obsolete reasons.)
Previously, point fields compressed by ZFP were attached as point fields
on the output. However, using them as a point field would cause
problems. So, instead attache them as `WholeDataSet` fields.
Also fixed a problem where the 1D decompressor created an output of the
wrong size.