TBB 2020 introduced a new class called `task_group`. TBB 2021 removed
the old class `task` as its functionality was replaced by `task_group`.
Our parallel radix sort for TBB was implemented using `task`s, so change
it to use `task_group` (which actually cleans up the code a bit).
Years ago we discovered a problem with TBB's parallel sort, which we
patch in our local repo and submitted a change to TBB, which has been
accepted.
The code to decide whether to use our parallel_sort patch does not work
with the latest versions of TBB because it requires including a header
that changed names to get the TBB version.
We no longer support any TBB version with this bug, so just remove the
patch from VTK-m.
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.
This class was used indirectly by the old `ArrayHandle`, through
`ArrayHandleTransfer`, to move data to and from a device. This
functionality has been replaced in the new `ArrayHandle`s through the
`Buffer` class (which can be compiled into libraries rather than make
every translation unit compile their own template).
This commit removes `ArrayManagerExecution` and all the implementations
that the device adapters were required to make. None of this code was in
any use anymore.
When `DeviceAdapterAlgorithm::ScheduleTask` was called directly (i.e.
not through `Schedule`), nothing was added to the log. Adding
`VTKM_LOG_SCOPE` to these methods so that all scheduling is added to the
performance log.
The new name reflects better what the underlying algorithm does. It also
helps prevent confusion about what types of data the locator is good
for. The old name suggested it only worked for structured grids, which
is not the case.
As we remove more and more virtual methods from VTK-m, I expect several
users will be interested in completely removing them from the build for
several reasons.
1. They may be compiling for hardware that does not support virtual
methods.
2. They may need to compile for CUDA but need shared libraries.
3. It should go a bit faster.
To enable this, a CMake option named `VTKm_NO_DEPRECATED_VIRTUAL` is
added. It defaults to `OFF`. But when it is `ON`, none of the code that
both uses virtuals and is deprecated will be built.
Currently, only `ArrayHandleVirtual` is deprecated, so the rest of the
virtual classes will still be built. As we move forward, more will be
removed until all virtual method functionality is removed.
Now that we have atomic free functions (e.g. `vtkm::AtomicAdd()`), we no
longer need special implementations for control and each execution
device. (Well, technically we do have special implementations for each,
but they are handled with compiler directives in the free functions.)
Convert the old atomic interface classes (`AtomicInterfaceControl` and
`AtomicInterfaceExecution`) to use the new atomic free functions. This
will allow us to test the new atomic functions everywhere that atomics
are used in VTK-m.
Once verified, we can deprecate the old atomic interface classes.
Some of the code in the base `vtkm` namespace is device specific. For
example, the functions in `Math.h` are customized for specific devices.
Thus, we want this code to be specially compiled and run on these
devices.
Previously, we made a header file and then added separate tests to each
device package. That was created before we had ways of running on any
device. Now, it is much easier to compile the test a single time for all
devices and use the `ALL_BACKENDS` feature of `vtkm_unit_tests` CMake
function to automatically create the test for all devices.
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.
The `ArrayHandleStreaming` class stems from an old research project
experimenting with bringing data from an `ArrayHandle` in parts and
overlapping device transfer and execution. It works, but only in very
limited contexts. Thus, it is not actually used today. Plus, the feature
requires global indexing to be permutated throughout the worklet
dispatching classes of VTK-m for no further reason.
Because it is not really used, there are other more promising approaches
on the horizon, and it makes further scheduling improvements difficult,
we are removing this functionality.
1f1688483 Initial infrastructure to allow WorkletMapField to have 3D scheduling
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Kenneth Moreland <kmorel@sandia.gov>
Merge-request: !1938
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.
The old version of ExecutionObject (that only takes a device) is still
supported, but you will get a deprecated warning if that is what is
defined.
Supporing this also included sending vtkm::cont::Token through the
vtkm::cont::arg::Transport mechanism, which was a change that propogated
through a lot of code.
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.
It is very easy to cause ODR violations with DeviceAdapterTagCuda.
If you include that header from a C++ file and a CUDA file inside
the same program we an ODR violation. The reasons is that the C++
versions will say the tag is invalid, and the CUDA will say the
tag is valid.
The solution to this is that any compilation unit that includes
DeviceAdapterTagCuda from a version of VTK-m that has CUDA enabled
must be invoked by the cuda compiler.
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.
6797c6e33 Specify return type for GetTimerImpl
25f3432b1 Increase the conditions on which Timer is tested
4d9ce2488 Synchronize CUDA timer when stopping it
85265a9c8 Add const correctness to Timer
465508993 Allow resetting Timer with a new device
dd4a93952 Enable initializing Timer with a DeviceAdapterId
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Haocheng LIU <haocheng.liu@kitware.com>
Merge-request: !1562
