GitLab

- Adding updates to uniform grid cell locator
  - adding OpenMP test, updating copyrights
- Adding rectilinear grid cell locator
  - adding unit tests for serial, tbb, OpenMP, and cuda
- Updating CMakeLists to honor the alphabetical ordering

Now it's in accordance with vtkm testing name convension as
UnitTests_vtkm_worklet_testing.

To get the average, we (of course) divide the sum by the amount of
values, which is returned from valuesIn.GetNumberOfComponents(). To do
this, we need to cast the number of components (returned as a
vtkm::IdComponent) to a FieldType. This is a little more complex than it
first seems because FieldType might be a Vec type. If you just try a
static_cast<FieldType>(), it will use the constructor to FieldType which
might be a Vec constructor expecting the type of the component. This in
turn could cause a warning because the vtkm::IdComponent is implicitly
converted to the Vec's component type.

Get around this problem by first casting to the component type of the
field and then constructing a field value from that.

VTK-m has been updated to replace old per device worklet testing executables with a device
dependent shared library so that it's able to accept a device adapter
at runtime.
Meanwhile, it updates the testing infrastructure APIs. vtkm::cont::testing::Run
function would call ForceDevice when needed and if users need the device
adapter info at runtime, RunOnDevice function would pass the adapter into the functor.

Optional Parser is bumped from 1.3 to 1.7.

This MR bumps the version from 1.3 to 1.7 to absorb MSC compiler
fixes and const-correctness.

If multiple threads call a function with a static function
simultaneously for the first time, C++ does not guarantee that the
static member will be constructed correctly. Make sure that it happens
correctly.

Previously, vtkm::exec::arg::BoundaryState only provided methods that
said whether or not the neighborhood extened past the boundary of a
mesh. That is fine for a 3x3x3 neighborhood, which can only extend over
the boundary by one. However, that is problematic for larger
neighborhoods where you may need to know how far neighborhood extends
over the boundary.

This changes allows you to query how far the neighborhood extends within
the constrains of the boundary.

Creating all the presets as vtkm::cont::ColorTable objects and passing
them back was problematic. It caused state to be shared and caused
issues when deallocating after the device deallocation methods were
finalized. Instead, make a simple struct and build new color tables on
the fly.

The ParaView project went through a rigourous selection process
for a short list of color tables. Let's replicate that for our
presets.

By making RuntimeDeviceInformation class template independent, vtkm is
able to detect
device info at runtime with a runtime specified deviceId. In the past
it's impossible
because the CRTP pattern does not allow function overloading(compiler
would complain
that DeviceAdapterRuntimeDetector does not have Exists() function
defined).

The field infrastructure has a "trait" that specifies the valid types
for the default active scalar field. Most filters have at most one input
field. The majority of the remaining filters use the same types for
inputs. There are some examples, however, where some auxiliary fields
use different types. This change adds a second template parameter to
FieldTraits to customize the valid types of the input scalar field to
different fields.

These changes are in anticipation of the removal of list specifier tags
on worklet ControlSignature arguments.

added `vtkm::cont::GetGlobalRuntimeDeviceTracker().ForceDevice(VTKM_DEFAULT_DEVICE_ADAPTER_TAG());`

I also cleaned up the unit test to no longer pass the device

std::containers only activate fast emplace operations if the
T types move operators are noexcept

The ParticleAdvection code would always cause cuda 9's ptxas phase
to crash, so we disable optimization to work around this issue

Previously, UnitTestProbe used a struct templated on the device adapter,
and that necessitated the use of the template keyword when using an
ambiguous type. These recent changes remove that template argument and
thus can cause a warning about using the template keyword in a non-
ambiguous type.

In some functors within Algorithm.h, the functor did not have a
default constructor that initialized its fields. This could in turn
lead to a warning about using it unitialized.

Now that the dispatcher does its own TryExecute, filters do not need to
do that. This change requires all worklets called by filters to be able
to execute without knowing the device a priori.

ContourTreeAugumneted paths became too long and caused msvc
to fail to construct pdb paths

The CUDA compiler is giving us grief while compiling
UnitTestParticleAdvection.cxx even though the code is correct. Disable
it for now until we figure out what can be done with the compiler.