arm_compute::CLGatherExKernel Class Reference

Class to define an interface for the gather kernel. More...

#include <CLGatherExKernel.h>

Collaboration diagram for arm_compute::CLGatherExKernel:

Public Member Functions
	CLGatherExKernel ()
	Construct CLGatherExKernel object.
	CLGatherExKernel (const CLGatherExKernel &)=delete
	Prevent instances of this class from being copied (As this class contains pointers).
CLGatherExKernel &	operator= (const CLGatherExKernel &)=delete
	Prevent instances of this class from being copied (As this class contains pointers).
	CLGatherExKernel (CLGatherExKernel &&)=default
	Construct CLGatherExKernel object by using default move constructor.
CLGatherExKernel &	operator= (CLGatherExKernel &&)=default
	Move assignment operator.
void	configure (const ICLTensor *input, const ICLTensor *indices, ICLTensor *output, int axis=0)
	Initialise the kernel's input, output and border mode.
void	run (const Window &window, cl::CommandQueue &queue) override
	Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue.

Static Public Member Functions
static Status	validate (const ITensorInfo *input, const ITensorInfo *indices, const ITensorInfo *output, int axis=0)
	Static function to check if given info will lead to a valid configuration of CLGatherExKernel.

Detailed Description

Class to define an interface for the gather kernel.

Definition at line 59 of file CLGatherExKernel.h.

Constructor & Destructor Documentation

CLGatherExKernel() [1/3]

CLGatherExKernel::CLGatherExKernel

(

)

Construct CLGatherExKernel object.

Definition at line 106 of file CLGatherExKernel.cpp.

  : _input(nullptr), _indices(nullptr), _output(nullptr), _axis(0)
{
}

CLGatherExKernel() [2/3]

arm_compute::CLGatherExKernel::CLGatherExKernel ( const CLGatherExKernel &  )

delete

Prevent instances of this class from being copied (As this class contains pointers).

CLGatherExKernel() [3/3]

arm_compute::CLGatherExKernel::CLGatherExKernel ( CLGatherExKernel &&  )

default

Construct CLGatherExKernel object by using default move constructor.

Parameters

[in]

CLGatherExKernel

object to move

Member Function Documentation

configure()

void CLGatherExKernel::configure	(	const ICLTensor *	input,
		const ICLTensor *	indices,
		ICLTensor *	output,
		int	axis = 0
	)

Initialise the kernel's input, output and border mode.

Parameters

[in]	input	An input tensor. Data types supported: U8/QASYMM8/S32/F32.
[in]	indices	Indices tensor. Data types supported: S32.
[out]	output	The output tensor, Data types supported: same as input1.
[in]	axis	(Optional) The axis in input to gather indices from. Negative values wrap around. Defaults to 0

Returns

N/A

Definition at line 111 of file CLGatherExKernel.cpp.

{
  ARM_COMPUTE_ERROR_ON_NULLPTR(input, output, indices);
  ARM_COMPUTE_ERROR_THROW_ON(
    validate_arguments(input->info(), indices->info(), output->info(), axis));
 
  // Configure kernel window
  auto win_config =
    validate_and_configure_window(input->info(), indices->info(), output->info(), axis);
  ARM_COMPUTE_ERROR_THROW_ON(win_config.first);
 
  _input = input;
  _output = output;
  _indices = indices;
  _axis = wrap_around(axis, static_cast<int>(input->info()->num_dimensions()));
 
  // Set build options
  CLBuildOptions build_opts;
  build_opts.add_option("-DDATA_TYPE=" + get_cl_type_from_data_type(input->info()->data_type()));
  build_opts.add_option("-DOUTPUT_DIM_Z=" +
                        support::cpp11::to_string(output->info()->dimension(2)));
  build_opts.add_option("-DINPUT_DIM_Z=" + support::cpp11::to_string(input->info()->dimension(2)));
  build_opts.add_option("-DAXIS=" + support::cpp11::to_string(_axis));
  build_opts.add_option("-DINDICES_DIM=" +
                        support::cpp11::to_string(indices->info()->num_dimensions()));
 
  // Create kernel
  _kernel = static_cast<cl::Kernel>(
    CLKernelLibraryEx::get().create_kernel("gather_ex", build_opts.options()));
  ICLKernel::configure_internal(win_config.second);
}

References arm_compute::CLKernelLibraryEx::create_kernel(), and arm_compute::CLKernelLibraryEx::get().

operator=() [1/2]

CLGatherExKernel & arm_compute::CLGatherExKernel::operator= ( CLGatherExKernel &&  )

default

Move assignment operator.

Parameters

[in]

CLGatherExKernel

object to move

References validate().

operator=() [2/2]

CLGatherExKernel & arm_compute::CLGatherExKernel::operator= ( const CLGatherExKernel &  )

delete

Prevent instances of this class from being copied (As this class contains pointers).

run()

void CLGatherExKernel::run ( const Window &  window, cl::CommandQueue &  queue  )

override

Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue.

Note

The queue is not flushed by this method, and therefore the kernel will not have been executed by the time this method returns.

Parameters

[in]	window	Region on which to execute the kernel. (Must be a valid region of the window returned by window()).
[in,out]	queue	Command queue on which to enqueue the kernel.

Returns

N/A

N/A

Definition at line 155 of file CLGatherExKernel.cpp.

{
  ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
  ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(IKernel::window(), window);
 
  Window window_collapsed = window.collapse_if_possible(ICLKernel::window(), Window::DimZ, 4);
  unsigned int idx = 0;
  add_4D_tensor_argument(idx, _input, window_collapsed);
  add_3D_tensor_argument(idx, _indices, window_collapsed);
  add_4D_tensor_argument(idx, _output, window_collapsed);
  enqueue(queue, *this, window_collapsed, lws_hint());
}

Referenced by package.infer.session::inference().

validate()

Status CLGatherExKernel::validate ( const ITensorInfo *  input, const ITensorInfo *  indices, const ITensorInfo *  output, int  axis = 0  )

static

Static function to check if given info will lead to a valid configuration of CLGatherExKernel.

Parameters

[in]	input	An input tensor. Data types supported: U8/QASYMM8/S32/F32.
[in]	indices	Indices tensor. Data types supported: S32.
[out]	output	The output tensor, Data types supported: same as input1.
[in]	axis	(Optional) The axis in input to gather indices from. Negative values wrap around. Defaults to 0

Returns

a status

Definition at line 144 of file CLGatherExKernel.cpp.

{
  ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, indices, output, axis));
  ARM_COMPUTE_RETURN_ON_ERROR(validate_and_configure_window(input->clone().get(),
                                                            indices->clone().get(),
                                                            output->clone().get(), axis)
                                .first);
  return Status{};
}

Referenced by operator=(), and arm_compute::CLGatherEx::validate().

---

The documentation for this class was generated from the following files:

• compute/ARMComputeEx/arm_compute/core/CL/kernels/CLGatherExKernel.h
• compute/ARMComputeEx/src/core/CL/kernels/CLGatherExKernel.cpp