arm_compute::CLCastBoolKernel Class Reference

Class for the kernel converting boolean type. More...

#include <CLCastBoolKernel.h>

Collaboration diagram for arm_compute::CLCastBoolKernel:

Public Member Functions
void	configure (const ICLTensor *input, ICLTensor *output)
	Initialise the kernel's input and output.

Static Public Member Functions
static Status	validate (const ITensorInfo *input, const ITensorInfo *output)

Detailed Description

Class for the kernel converting boolean type.

Definition at line 59 of file CLCastBoolKernel.h.

Member Function Documentation

configure()

void arm_compute::CLCastBoolKernel::configure	(	const ICLTensor *	input,
		ICLTensor *	output
	)

Initialise the kernel's input and output.

Parameters

[in]	input	Input tensor. Data types supported: U8
[in]	output	Output tensor. Data types supported: U8/QASYMM8/S16/S32/F16/F32.

Returns

N/A

Definition at line 85 of file CLCastBoolKernel.cpp.

{
  ARM_COMPUTE_ERROR_ON_NULLPTR(input, output);
 
  // Auto initialize output shape if not initialized (We can only auto-configure the shape, datatype
  // must be given)
  set_shape_if_empty(*output->info(), input->info()->tensor_shape());
 
  ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input->info(), output->info()));
 
  // Get number of elements to process per iterations
  constexpr unsigned int num_elems_processed_per_iteration = 16;
 
  // Set build options
  CLBuildOptions build_opts;
  build_opts.add_option("-DVEC_SIZE=" +
                        support::cpp11::to_string(num_elems_processed_per_iteration));
  build_opts.add_option("-DDATA_TYPE_IN=" + get_cl_type_from_data_type(input->info()->data_type()));
  build_opts.add_option("-DDATA_TYPE_OUT=" +
                        get_cl_type_from_data_type(output->info()->data_type()));
 
  // Create kernel
  const std::string kernel_name = "cast_bool";
  _kernel = static_cast<cl::Kernel>(
    CLKernelLibraryEx::get().create_kernel(kernel_name, build_opts.options()));
 
  // Configure kernel
  ICLSimple2DKernel::configure(input, output, num_elems_processed_per_iteration);
 
  // Collapse window
  const Window &full_window = window();
  Window collapsed_window = full_window.collapse_if_possible(full_window, Window::DimZ);
  ICLKernel::configure_internal(collapsed_window);
 
  // Set config_id for enabling LWS tuning
  _config_id = kernel_name;
  _config_id += "_";
  _config_id += lower_string(string_from_data_type(output->info()->data_type()));
  _config_id += "_";
  _config_id += support::cpp11::to_string(output->info()->dimension(0));
  _config_id += "_";
  _config_id += support::cpp11::to_string(output->info()->dimension(1));
}

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

validate()

Status arm_compute::CLCastBoolKernel::validate ( const ITensorInfo *  input, const ITensorInfo *  output  )

static

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

Parameters

[in]	input	Source tensor info. Data types supported: U8.
[in]	output	Destination tensor info. Data type supported: U8/S8/U16/S16/U32/S32/F16/F32.

Returns

a status

Definition at line 129 of file CLCastBoolKernel.cpp.

{
  ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, output));
 
  return Status{};
}

---

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

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