Several revisions ago, the ability to use virtual methods in the
execution environment was deprecated. Completely remove this
functionality for the VTK-m 2.0 release.
Deprecate `VirtualObjectHandle` and all other classes that are used to
implement objects with virtual methods in the execution environment.
Additionally, the code is updated so that if the
`VTKm_NO_DEPRECATED_VIRTUAL` flag is set none of the code is compiled at
all. This opens us up to opportunities that do not work with virtual
methods such as backends that do not support virtual methods and dynamic
libraries for CUDA.
The basic type operators in `Types.h` (i.e. `vtkm::Add`,
`vtkm::Subtract`, `vtkm::Multiply` and `vtkm::Divide`) required the same
type for both arguments. This caused problems when used with `Reduce`
and the initial value type did not match exactly.
Use some tricks from `BinaryOperators.h` to be flexible about using
different types.
The newer version of `ArrayHandle` no longer supports different types of
portals for different devices. Thus, the `ReadPortalType` and
`WritePortalType` are sufficient for all types of portals across all
devices.
This significantly simplifies supporting execution objects on devices,
and thus this change also includes many changes to various execution
objects to remove their dependence on the device adapter tag.
The actual code for AtomicArrayExecutionObject does not need to be
specialized by the device. The functionality is implemented by calling
the vtkm::Atomic* methods, which are properly implemented on each
device.
What was previously declared as `ArrayHandleNewStyle` is now just the
implementation of `ArrayHandle`. The old implementation of `ArrayHandle`
has been moved to `ArrayHandleDeprecated`, and `ArrayHandle`s still
using this implementation must declare `VTKM_ARRAY_HANDLE_DEPRECATED` to
use it.
The old atomic compare and swap operations (`vtkm::AtomicCompareAndSwap`
and `vtkm::exec::AtomicArrayExecutionObject::CompareAndSwap`) had an
order of arguments that was confusing. The order of the arguments was
shared pointer (or index), desired value, expected value. Most people
probably assume expected value comes before desired value. And this
order conflicts with the order in the `std` methods, GCC atomics, and
Kokkos.
Change the interface of atomic operations to be patterned off the
`std::atomic_compare_exchange` and `std::atomic<T>::compare_exchange`
methods. First, these methods have a more intuitive order of parameters
(shared pointer, expected, desired). Second, rather than take a value
for the expected and return the actual old value, they take a pointer to
the expected value (or reference in `AtomicArrayExecutionObject`) and
modify this value in the case that it does not match the actual value.
This makes it harder to mix up the expected and desired parameters.
Also, because the methods return a bool indicating whether the value was
changed, there is an additional benefit that compare-exchange loops are
implemented easier.
For example, consider you want to apply the function `MyOp` on a
`sharedValue` atomically. With the old interface, you would have to do
something like this.
```cpp
T oldValue;
T newValue;
do
{
oldValue = *sharedValue;
newValue = MyOp(oldValue);
} while (vtkm::AtomicCompareAndSwap(sharedValue, newValue, oldValue) != oldValue);
```
With the new interface, this is simplfied to this.
```cpp
T oldValue = *sharedValue;
while (!vtkm::AtomicCompareExchange(sharedValue, &oldValue, MyOp(oldValue));
```
We have made several improvements to adding data into an `ArrayHandle`.
## Moving data from an `std::vector`
For numerous reasons, it is convenient to define data in a `std::vector`
and then wrap that into an `ArrayHandle`. It is often the case that an
`std::vector` is filled and then becomes unused once it is converted to an
`ArrayHandle`. In this case, what we really want is to pass the data off to
the `ArrayHandle` so that the `ArrayHandle` is now managing the data and
not the `std::vector`.
C++11 has a mechanism to do this: move semantics. You can now pass
variables to functions as an "rvalue" (right-hand value). When something is
passed as an rvalue, it can pull state out of that variable and move it
somewhere else. `std::vector` implements this movement so that an rvalue
can be moved to another `std::vector` without actually copying the data.
`make_ArrayHandle` now also takes advantage of this feature to move rvalue
`std::vector`s.
There is a special form of `make_ArrayHandle` named `make_ArrayHandleMove`
that takes an rvalue. There is also a special overload of
`make_ArrayHandle` itself that handles an rvalue `vector`. (However, using
the explicit move version is better if you want to make sure the data is
actually moved.)
## Make `ArrayHandle` from initalizer list
A common use case for using `std::vector` (particularly in our unit tests)
is to quickly add an initalizer list into an `ArrayHandle`. Now you can
by simply passing an initializer list to `make_ArrayHandle`.
## Deprecated `make_ArrayHandle` with default shallow copy
For historical reasons, passing an `std::vector` or a pointer to
`make_ArrayHandle` does a shallow copy (i.e. `CopyFlag` defaults to `Off`).
Although more efficient, this mode is inherintly unsafe, and making it the
default is asking for trouble.
To combat this, calling `make_ArrayHandle` without a copy flag is
deprecated. In this way, if you wish to do the faster but more unsafe
creation of an `ArrayHandle` you should explicitly express that.
This requried quite a few changes through the VTK-m source (particularly in
the tests).
## Similar changes to `Field`
`vtkm::cont::Field` has a `make_Field` helper function that is similar to
`make_ArrayHandle`. It also features the ability to create fields from
`std::vector`s and C arrays. It also likewise had the same unsafe behavior
by default of not copying from the source of the arrays.
That behavior has similarly been depreciated. You now have to specify a
copy flag.
The ability to construct a `Field` from an initializer list of values has
also been added.
Often when a user gives memory to an `ArrayHandle`, she wants data to be
written into the memory given to be used elsewhere. Previously, the
`Buffer` objects would delete the given buffer as soon as a write buffer
was created elsewhere. That was a problem if a user wants VTK-m to write
results right into a given buffer.
Instead, when a user provides memory, "pin" that memory so that the
`ArrayHandle` never deletes it.
The buffer class encapsulates the movement of raw C arrays between
host and devices.
The `Buffer` class itself is not associated with any device. Instead,
`Buffer` is used in conjunction with a new templated class named
`DeviceAdapterMemoryManager` that can allocate data on a given
device and transfer data as necessary. `DeviceAdapterMemoryManager`
will eventually replace the more complicated device adapter classes
that manage data on a device.
The code in `DeviceAdapterMemoryManager` is actually enclosed in
virtual methods. This allows us to limit the number of classes that
need to be compiled for a device. Rather, the implementation of
`DeviceAdapterMemoryManager` is compiled once with whatever compiler
is necessary, and then the `RuntimeDeviceInformation` is used to
get the correct object instance.
If you gave ReduceByKey a fancy output array that decorated another
array, you could get a runtime error for using an invalid array (if the
device adapter used the generic algorithm). The problem was that
ReduceByKey creates a temporary array, and that array was given the same
storage as the output array. That might not be valid for fancy arrays,
so instead use the default storage for the temporary array.
If you gave ScanInclusiveByKey a fancy output array that decorated
another array, you would get a runtime error for using an invalid array.
The problem was that ScanInclusiveByKey creates a temporary output array
and then copies the result to the actual output array. The problem was
that the temporary output array was given the same storage as the output
array, which won't work if the output array is fancy. Instead, make the
storage for the temporary array default.
Previously, when a ReadPortal or a WritePortal was returned from an
ArrayHandle, it had wrapped in it a Token that was attached to the
ArrayHandle. This Token would prevent other reads and writes from the
ArrayHandle.
This added safety in the form of making sure that the ArrayPortal was
always valid. Unfortunately, it also made deadlocks very easy. They
happened when an ArrayPortal did not leave scope immediately after use
(which is not all that uncommon).
Now, the ArrayPortal no longer locks up the ArrayHandle. Instead, when
an access happens on the ArrayPortal, it checks to make sure that
nothing has happened to the data being accessed. If it has, a fatal
error is reported to the log.
To get a portal to access ArrayHandle values in the control
environment, you now use the ReadPortal and WritePortal methods.
The portals returned are wrapped in an ArrayPortalToken object
so that the data between the portal and the ArrayHandle are
guaranteed to be consistent.
It is questionable whether there is a point to having a token object
when transfering a virtual object to a device (since there is a handle
object that is managing it anyway). Back out of passing the token all
the way down unless there is an actual need for that.
Marked the old versions of PrepareFor* that do not use tokens as
deprecated and moved all of the code to use the new versions that
require a token. This makes the scope of the execution object more
explicit so that it will be kept while in use and can potentially be
reclaimed afterward.
This behaves just like `ScanExclusive`, but rather than returning the
total sum, it is appended to the end of the output array.
This is in preparation for the CellSetExplicit refactoring described in
issue #408.
- Use AtomicInterface to implement device-specific atomic operations.
- Remove DeviceAdapterAtomicArrayImplementations.
- Extend supported atomic types to include unsigned 32/64-bit ints.
- Add a static_assert to check that AtomicArray type is supported.
- Add documentation for AtomicArrayExecutionObject, including a CAS
example.
- Add a `T Get(idx)` method to AtomicArrayExecutionObject that does
an atomic load, and update existing CAS usage to use this instead
of `Add(idx, 0)`.
As the RuntimeDeviceTracker is a per thread construct we now make
it explicit that you can only get a reference to the per-thread
version and can't copy it.
The consistent API for control to execution memory transfers is
the ArrayHandle class. Previously the tests would verify memory
transfer by calling the ArrayManagerExecution class directly. This
is problematic as the class isn't used by ArrayHandle<T, StorageBasic>.
Fixes#277
DeviceAdapterError existed to make sure that the default device adapter
template was being handled properly. Since the default device adapter doesn't
exist, and nothing is templated over it we can now remove DeviceAdapterError.
661fb64de AtomicInterfaceControl functions are marked with VTKM_SUPPRESS_EXEC_WARNINGS
0c70f9b9a Add BitFieldIn/Out/InOut worklet signature tags.
a66510e81 Add ArrayHandleBitField, a boolean-valued AH backed by a BitField.
56cc5c3d3 Add support for BitFields.
d01b97382 Allow VTKM_SUPPRESS_EXEC_WARNINGS to be used inside macros.
2f2ca9370 Add bit operations FindFirstSetBit and CountSetBits to Math.h.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !1629
BitFields are:
- Stored in memory using a contiguous buffer of bits.
- Accessible via portals, a la ArrayHandle.
- Portals operate on individual bits or words.
- Operations may be atomic for safe use from concurrent kernels.
The new BitFieldToUnorderedSet device algorithm produces an ArrayHandle
containing the indices of all set bits, in no particular order.
The new AtomicInterface classes provide an abstraction into bitwise
atomic operations across control and execution environments and are used
to implement the BitPortals.
When reducing an input type that differs from the output type
you need to write a custom binary operator that also implements
how to do the unary transformation.
The timer class now is asynchronous and device independent. it's using an
similiar API as vtkOpenGLRenderTimer with Start(), Stop(), Reset(), Ready(),
and GetElapsedTime() function. For convenience and backward compability, Each
Start() function call will call Reset() internally and each GetElapsedTime()
function call will call Stop() function if it hasn't been called yet for keeping
backward compatibility purpose.
Bascially it can be used in two modes:
* Create a Timer without any device info. vtkm::cont::Timer time;
* It would enable timers for all enabled devices on the machine. Users can get a
specific elapsed time by passing a device id into the GetElapsedtime function.
If no device is provided, it would pick the maximum of all timer results - the
logic behind this decision is that if cuda is disabled, openmp, serial and tbb
roughly give the same results; if cuda is enabled it's safe to return the
maximum elapsed time since users are more interested in the device execution
time rather than the kernal launch time. The Ready function can be handy here
to query the status of the timer.
* Create a Timer with a device id. vtkm::cont::Timer time((vtkm::cont::DeviceAdapterTagCuda()));
* It works as the old timer that times for a specific device id.
`vtkm::cont::testing` now initializes with logging enabled and support
for device being passed on the command line, `vtkm::testing` only
enables logging.
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 OpenMP Device Reduction algorithm previously used a std::vector<T>
to store the reduction results of each thread. This caused problems
when T=bool as the types became a proxy type which isn't usable
with vtkm BinaryOperators.
Additionally by fixing this issue in the FunctorsOpenMP we
can remove a workaround in FunctorsGeneral that caused
compile failures when using complex BinaryOperators
such as MinAndMax.