Kernel which transposes the elements of a matrix. More...

#include <CpuTransposeKernel.h>

Collaboration diagram for CpuTransposeKernel:

Public Member Functions
	CpuTransposeKernel ()=default

	ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE (CpuTransposeKernel)

void	configure (const ITensorInfo src, ITensorInfo dst)
	Configure kernel for a given list of arguments. More...

void	run_op (ITensorPack &tensors, const Window &window, const ThreadInfo &info) override
	Execute the kernel on the passed window. More...

const char *	name () const override
	Name of the kernel. More...

Public Member Functions inherited from ICPPKernel
virtual	~ICPPKernel ()=default
	Default destructor. More...

virtual void	run (const Window &window, const ThreadInfo &info)
	Execute the kernel on the passed window. More...

virtual void	run_nd (const Window &window, const ThreadInfo &info, const Window &thread_locator)
	legacy compatibility layer for implemantions which do not support thread_locator In these cases we simply narrow the interface down the legacy version More...

virtual size_t	get_mws (const CPUInfo &platform, size_t thread_count) const
	Return minimum workload size of the relevant kernel. More...

Public Member Functions inherited from IKernel
	IKernel ()
	Constructor. More...

virtual	~IKernel ()=default
	Destructor. More...

virtual bool	is_parallelisable () const
	Indicates whether or not the kernel is parallelisable. More...

virtual BorderSize	border_size () const
	The size of the border for that kernel. More...

const Window &	window () const
	The maximum window the kernel can be executed on. More...

bool	is_window_configured () const
	Function to check if the embedded window of this kernel has been configured. More...

Static Public Member Functions
static Status	validate (const ITensorInfo src, const ITensorInfo dst)
	Static function to check if given info will lead to a valid configuration. More...

Static Public Member Functions inherited from ICpuKernel< CpuTransposeKernel >
static const auto *	get_implementation (const SelectorType &selector, KernelSelectionType selection_type=KernelSelectionType::Supported)
	Micro-kernel selector. More...

Additional Inherited Members
Static Public Attributes inherited from ICPPKernel
static constexpr size_t	default_mws = 1

Detailed Description

Kernel which transposes the elements of a matrix.

Definition at line 37 of file CpuTransposeKernel.h.

Constructor & Destructor Documentation

◆ CpuTransposeKernel()

CpuTransposeKernel ( )

default

Member Function Documentation

◆ ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE()

ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE ( CpuTransposeKernel )

◆ configure()

void configure	(	const ITensorInfo *	src,
		ITensorInfo *	dst
	)

Configure kernel for a given list of arguments.

Parameters

[in]	src	Srouce tensor to permute. Data types supported: All
[out]	dst	Destination tensor. Data types supported: Same as `src`

Definition at line 732 of file CpuTransposeKernel.cpp.

 {
     ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst);
  
     // Destination auto inizialitation if not yet initialized
     const TensorShape dst_shape = misc::shape_calculator::compute_transposed_shape(*src);
     auto_init_if_empty(*dst, src->clone()->set_tensor_shape(dst_shape));
  
     // Explicitly set the tensor shape to preserve dimensions
     dst->set_tensor_shape(dst_shape);
  
     // Perform validation step
     ARM_COMPUTE_ERROR_THROW_ON(validate(src, dst));
  
     // Note: This kernel performs 16 elements per iteration.
     // However, since we use a left-over for loop on both dimensions (X and Y), we cannot have any read or write out of memory
     // For this reason num_elems_processed_per_iteration_x is set to 1
     const unsigned int num_elems_processed_per_iteration_x = 1;
     const unsigned int num_elems_processed_per_iteration_y = num_elems_processed(src->element_size());
  
     // Configure kernel window
     Window win =
         calculate_max_window(*src, Steps(num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y));
  
     // The CpuTranspose doesn't need padding so update_window_and_padding() can be skipped
     Coordinates coord;
     coord.set_num_dimensions(dst->num_dimensions());
     dst->set_valid_region(ValidRegion(coord, dst->tensor_shape()));
  
     ICpuKernel::configure(win);
 }

References ARM_COMPUTE_ERROR_ON_NULLPTR, ARM_COMPUTE_ERROR_THROW_ON, arm_compute::auto_init_if_empty(), arm_compute::calculate_max_window(), arm_compute::misc::shape_calculator::compute_transposed_shape(), arm_compute::test::validation::dst, arm_compute::test::validation::dst_shape, Dimensions< T >::set_num_dimensions(), arm_compute::test::validation::src, and CpuTransposeKernel::validate().

◆ name()

const char * name ( ) const

overridevirtual

Name of the kernel.

Returns: Kernel name

Implements ICPPKernel.

Definition at line 813 of file CpuTransposeKernel.cpp.

 {
     return "CpuTransposeKernel";
 }

◆ run_op()

void run_op	(	ITensorPack &	tensors,
		const Window &	window,
		const ThreadInfo &	info
	)

overridevirtual

Execute the kernel on the passed window.

Warning: If is_parallelisable() returns false then the passed window must be equal to window()

Note: The window has to be a region within the window returned by the window() method; The width of the window has to be a multiple of num_elems_processed_per_iteration().

Parameters

[in]	tensors	A vector containing the tensors to operate on.
[in]	window	Region on which to execute the kernel. (Must be a region of the window returned by window())
[in]	info	Info about executing thread and CPU.

Reimplemented from ICPPKernel.

Definition at line 787 of file CpuTransposeKernel.cpp.

 {
     ARM_COMPUTE_UNUSED(info);
     ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
     ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(ICpuKernel::window(), window);
  
     const auto src = tensors.get_const_tensor(TensorType::ACL_SRC);
     auto       dst = tensors.get_tensor(TensorType::ACL_DST);
  
     switch (src->info()->element_size())
     {
         case 1:
             transpose_8bit_elements(src, dst, window);
             break;
         case 2:
             transpose_16bit_elements(src, dst, window);
             break;
         case 4:
             transpose_32bit_elements(src, dst, window);
             break;
         default:
             ARM_COMPUTE_ERROR("Element size not supported");
             break;
     }
 }

References arm_compute::ACL_DST, arm_compute::ACL_SRC, ARM_COMPUTE_ERROR, ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW, ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL, ARM_COMPUTE_UNUSED, arm_compute::test::validation::dst, ITensorPack::get_const_tensor(), ITensorPack::get_tensor(), arm_compute::test::validation::info, arm_compute::test::validation::src, and IKernel::window().

◆ validate()

Status validate	(	const ITensorInfo *	src,
		const ITensorInfo *	dst
	)

static

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

Returns: a status

Definition at line 764 of file CpuTransposeKernel.cpp.

 {
     ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(src);
     //Note: ARM_COMPUTE_RETURN_ERROR_ON_CPU_F16_UNSUPPORTED(input) is not needed here as this kernel doesn't use CPU FP16 instructions.
     ARM_COMPUTE_RETURN_ERROR_ON(src->data_type() == DataType::UNKNOWN);
  
     // Error if input is not 8 bit, 16bit or 32bit
     ARM_COMPUTE_RETURN_ERROR_ON_MSG(src->element_size() != 1 && src->element_size() != 2 && src->element_size() != 4,
                                     "Element size not supported");
  
     // Validate configured destination
     if (dst->total_size() != 0)
     {
         const TensorShape dst_shape = misc::shape_calculator::compute_transposed_shape(*src);
  
         ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(dst->tensor_shape(), dst_shape);
         ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO(src, dst);
         ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(src, dst);
     }
  
     return Status{};
 }

References ARM_COMPUTE_RETURN_ERROR_ON, ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES, ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS, ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO, ARM_COMPUTE_RETURN_ERROR_ON_MSG, ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR, arm_compute::misc::shape_calculator::compute_transposed_shape(), arm_compute::test::validation::dst, arm_compute::test::validation::dst_shape, arm_compute::test::validation::src, and arm_compute::UNKNOWN.

Referenced by CpuTransposeKernel::configure(), CpuTranspose::validate(), CpuMatMul::validate(), and CpuFullyConnected::validate().

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

src/cpu/kernels/CpuTransposeKernel.h
src/cpu/kernels/CpuTransposeKernel.cpp

Public Member Functions

Static Public Member Functions

Additional Inherited Members