There are various reasons why you might want to execute something but
not have a specific device to execute on. To mange this, add a general
function that will try a list of devices in order and attempt to run on
them in order.
It is more common to use degrees when specifying a transform (thanks to
the classic OpenGL interface). Also, camera specifies the field of view in
degrees, which made rotations inconsistent. This change unifies all that.
Affine transformations of homogeneous coordinates using 4x4 matrices are
quite common in visualization. Create a new math header file in the base
vtkm namespace that has common functions for such coordinates.
Much of this implementation was taken from the rendering matrix helpers.
First, be more explicit when we mean a range of values in a field or a
spacial bounds. Use the Range and Bounds structs in Field and
CoordinateSystem to make all of this more clear (and reduce a bit of
code as well).
After the most recent changes, I noticed that the matrix unit test
started failing for some optimized compiles. I'm not sure if it was
these changes or others. What I think happened is that it was a check
for a Matrix operation that should be invalid. It checked the valid flag
without checking the data (which, of course, is invalid). However, I
think the optimizer saw that the data generated was never used so
removed that part of the computation so the invalid flag was never set.
Add a condition that uses the result even though it should never be
called to hopefully force the compiler to compute it.
Fix issues with UnitTestVectorAnalysis for MSVC.
We have been having a problem with one of the MSVC dashboards failing
the UnitTestVectorAnalysis test. The test just dies in the middle with
no indication of what problematic thing was run.
After playing with this for quite a while, I found that it could by
triggered exclusively in the Lerp test. I further found that if I
switched the order of the test and check Lerp the test worked. This is
strange behavior and leads me to believe one of the following is going
on:
1. There is an error such as an invalid memory access happening in
the VTK-m code that is sometimes corrupting the stack.
2. Somewhere there is an expression that has undefined behavior. Usually
it works OK, but some optimization sequence causes it to fail.
3. There is a bug in one of the compiler's optimizations.
It concerns me that I cannot identify exactly where the problem lies.
I've looked very hard at the vtkm::Vec and vtkm::Lerp code to try to find
possible problems, but I have not been able to find anything.
See merge request !399
We have been having a problem with one of the MSVC dashboards failing
the UnitTestVectorAnalysis test. The test just dies in the middle with
no indication of what problematic thing was run.
After playing with this for quite a while, I found that it could by
triggered exclusively in the Lerp test. I further found that if I
switched the order of the test and check Lerp the test worked. This is
strange behavior and leads me to believe one of the following is going
on:
1. There is an error such as an invalid memory access happening in
the VTK-m code that is sometimes corrupting the stack.
2. Somewhere there is an expression that has undefined behavior. Usually
it works OK, but some optimization sequence causes it to fail.
3. There is a bug in one of the compiler's optimizations.
It concerns me that I cannot identify exactly where the problem lies.
I've looked very hard at the vtkm::Vec and vtkm::Lerp code to try to find
possible problems, but I have not been able to find anything.
All the other math functions are in the vtkm package. This one was in
vtkm::exec because it uses a callback method. This can be problematic on
CUDA the the declaration of NewtonsMethod does not match the callback
method. However, we now have a VTKM_SUPPRESS_EXEC_WARNINGS macro that
allows a VTKM_EXEC_CONT_EXPORT function (like NewtonsMethod) to call
either a VTKM_EXEC_EXPORT or VTKM_CONT_EXPORT without a warning.
We have decided that we do not need the concept of an Extent in our data
model. Instead, we simply use dimensions, which can be represented in a
vtkm::Id3.
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.
There was an error in the test_equal comparison that would return true
when the second value was 0 (or close to 0) and the first value was not.
This was a bug I introduced with commit
b270438fea8a9fe4200322bfe6344ab2075c4d9a. I clearly misinterpreted how a
conditional worked.
This class implicitly stores the point coordinates for a rectilinear
cell (such as a voxel) with just the location of the lower left point
and the spacing in each dimension. In addition to saving space, this
class should allow cell-specific functions to specialize for faster
processing.
The PrintSummary for CoordinateSystem went in an infinite loop. It was
supposed to call PrintSummary of its superclass (Field), but instead it
called itself.
The PrintSummary for Field only worked for fields of type vtkm::Float32.
To make it work for all array types, I added a PrintSummary method to
DynamicArrayHandle, and Field calls that without trying to cast to a
static type.
Variable topology fields
Changes to fetching in topology maps that lets you properly deal with cases where you do not know how many values are being fetched at compile time. For example, explicit cell sets can have any number of cell shapes that have different numbers of nodes.
This change should resolve issue #26.
See merge request !128
The PGI compiler appearently saw the condition (var != var) and optimized
it to always be false. However, in this particular case var was holding
a NaN value that we were testing to make sure it does not equal itself.
Get around the problem by comparing the variable to the result of a
function call.
This class holds a Vec and exposes some number of components. The class
is used when you need a Vec of a size that is not known at compile time
but that a maximum length of reasonable size is known.
Some changes to the Vec class and VecTraits in anticipation of creating
Vec-like objects. The following changes are made:
* Add GetNumberOfComponents to Vec, which returns NUM_COMPONENTS.
* Likewise, all VecTraits have a GetNumberOfComponents method.
* The ToVec method in VecTraits is changed to CopyInto so that it can be
used when the length of the Vec-like is not known. CopyInto is also
added to Vec.
* VecTraits has a typedef named IsSizeStatic which is set to
VecTraitsTagSizeStatic when the number of components is known at compile
time and VecTraitsTagSizeVariable when the number of components is not
known until runtime.
C and C++ has a funny feature where operations on small integers (char
and short) actually promote the result to a 32 bit integer. Most often
in our code the result is pushed back to the same type, and picky compilers
can then give a warning about an implicit type conversion (that we
inevitably don't care about). Here are a lot of changes to suppress
the warnings.
On one of my compile platforms, GCC was giving conversion warnings from
any boost include that was not wrapped in pragmas to disable conversion
warnings. To make things easier and more robust, I created a pair of
macros, VTKM_BOOST_PRE_INCLUDE and VTKM_BOOST_POST_INCLUDE, that should
be wrapped around any #include of a boost header file.
The test_equal method compares the ratio of the two values to decide if
they are close enough. Although there is a previous check to make sure
that neither value is too close to zero, the MSVC sometimes gives a
warning because it cannot trace the flow of the check. Add another
conditional (that will never actually be executed) to check a second time
that we never divide by 0.
Unlike most compilers, MSVC will give conversion warnings when
implicitly converting a 32-bit int to a 32-bit float because the
mantissa can potentially drop some of the lower order bits of large
ints. Use static casts to (hopefully) remove the warnings.
Some compilers, particularly icc, do strange things to literals during
optimization that makes them slightly different. This change was originally
to fix a testing issue with the icc build. It is not actually fixing the
problem, but I am leaving in the change because it is good practice.
Robert Maynard pointed out that the unary operator- I added to Vec could
lead to undesirable behavior for vectors of unsigned integer types. This
changed makes the definition of operator- conditional on the component
type being either a signed integer or a float type.
I also added some more actual testing of the new operator.
Previously, the templated Vec classes had overloaded multiply operators
to allow it to be multiplied by any basic type (float, double, int,
long, char, etc.). This was there to make it easier to multiply a vector
by a scalar without having to jump through introspection hoops for the
component types. However, using them almost always resulted in a
conversion warning on some type of compiler, so I think it is easier
just to remove them.
Also fix some issues that caused the compile to fail when trying to
run some of the math functions on a CUDA device. In particular, CUDA
is picky about using a global const on a device when the const type is
not one of the basic C types.