CLReductionOperation Class Reference

Perform reduction operation. More...

#include <CLReductionOperation.h>

Collaboration diagram for CLReductionOperation:

Public Member Functions
	CLReductionOperation (std::shared_ptr< IMemoryManager > memory_manager=nullptr)
	Default Constructor. More...
	~CLReductionOperation ()
	Default Destructor. More...
	CLReductionOperation (const CLReductionOperation &)=delete
	Prevent instances of this class from being copied (As this class contains pointers) More...
	CLReductionOperation (CLReductionOperation &&)=default
	Default move constructor. More...
CLReductionOperation &	operator= (const CLReductionOperation &)=delete
	Prevent instances of this class from being copied (As this class contains pointers) More...
CLReductionOperation &	operator= (CLReductionOperation &&)=default
	Default move assignment operator. More...
void	configure (ICLTensor *input, ICLTensor *output, unsigned int axis, ReductionOperation op, bool keep_dims=true)
	Set the input and output tensors. More...
void	configure (const CLCompileContext &compile_context, ICLTensor *input, ICLTensor *output, unsigned int axis, ReductionOperation op, bool keep_dims=true)
	Set the input and output tensors. More...
void	run () override
	Run the kernels contained in the function. More...
Public Member Functions inherited from IFunction
virtual	~IFunction ()=default
	Destructor. More...
virtual void	prepare ()
	Prepare the function for executing. More...

Static Public Member Functions
static Status	validate (const ITensorInfo *input, const ITensorInfo *output, unsigned int axis, ReductionOperation op, bool keep_dims=true)
	Static function to check if given info will lead to a valid configuration of CLReductionOperation. More...

Detailed Description

Perform reduction operation.

Definition at line 44 of file CLReductionOperation.h.

Constructor & Destructor Documentation

CLReductionOperation() [1/3]

CLReductionOperation

(

std::shared_ptr< IMemoryManager >

memory_manager = nullptr

)

Default Constructor.

Parameters

[in]

memory_manager

(Optional) Memory manager.

Definition at line 41 of file CLReductionOperation.cpp.

References CLReductionOperation::~CLReductionOperation().

    : _memory_group(std::move(memory_manager)), _unreshaped_output(), _reduction_kernel(), _reshape(), _reduction_axis(), _is_reshape_required(false)
{
}

~CLReductionOperation()

~CLReductionOperation ( )

default

Default Destructor.

Referenced by CLReductionOperation::CLReductionOperation().

CLReductionOperation() [2/3]

CLReductionOperation ( const CLReductionOperation &  )

delete

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

CLReductionOperation() [3/3]

CLReductionOperation ( CLReductionOperation &&  )

default

Default move constructor.

Member Function Documentation

configure() [1/2]

void configure	(	ICLTensor *	input,
		ICLTensor *	output,
		unsigned int	axis,
		ReductionOperation	op,
		bool	keep_dims = true
	)

Set the input and output tensors.

Valid data layouts:

• All

Valid data type configurations:

src	dst
QASYMM8	QASYMM8
QASYMM8_SIGNED	QASYMM8_SIGNED
F16	F16
F32	F32
S32	S32

Parameters

[in]	input	Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32/S32.
[out]	output	Destination tensor. Data types and data layouts supported: Same as input.
[in]	axis	Axis along which to reduce. Supported reduction axis : 0, 1, 2, 3
[in]	op	Reduction operation to perform. Operations supported: MEAN_SUM, PROD, SUM_SQUARE, SUM, MIN, MAX
[in]	keep_dims	(Optional) Whether to keep the reduced dimension after the operation. Defaults to true.

Definition at line 106 of file CLReductionOperation.cpp.

References CLKernelLibrary::get().

Referenced by CLL2NormalizeLayer::configure(), and CLFFTConvolutionLayer::configure().

{
    configure(CLKernelLibrary::get().get_compile_context(), input, output, axis, op, keep_dims);
}

configure() [2/2]

void configure	(	const CLCompileContext &	compile_context,
		ICLTensor *	input,
		ICLTensor *	output,
		unsigned int	axis,
		ReductionOperation	op,
		bool	keep_dims = true
	)

Set the input and output tensors.

Parameters

[in]	compile_context	The compile context to be used.
[in]	input	Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32/S32.
[out]	output	Destination tensor. Data types and data layouts supported: Same as input.
[in]	axis	Axis along which to reduce. Supported reduction axis : 0, 1, 2, 3
[in]	op	Reduction operation to perform. Operations supported: MEAN_SUM, PROD, SUM_SQUARE, SUM, MIN, MAX
[in]	keep_dims	(Optional) Whether to keep the reduced dimension after the operation. Defaults to true.

Definition at line 111 of file CLReductionOperation.cpp.

References CLTensorAllocator::allocate(), CLTensor::allocator(), ARM_COMPUTE_ERROR_ON_NULLPTR, ARM_COMPUTE_LOG_PARAMS, arm_compute::auto_init_if_empty(), ICloneable< T >::clone(), arm_compute::misc::shape_calculator::compute_reduced_shape(), CLReshapeLayer::configure(), ITensorInfo::data_type(), ITensor::info(), MemoryGroup::manage(), arm_compute::test::validation::output_shape, and ITensorInfo::tensor_shape().

{
    ARM_COMPUTE_ERROR_ON_NULLPTR(input, output);
    ARM_COMPUTE_LOG_PARAMS(input, output, axis, op, keep_dims);
    _reduction_axis      = axis;
    _is_reshape_required = !keep_dims;

    auto *output_internal = configure_intermediate_result_vector(input, output);

    if(_is_reshape_required)
    {
        const TensorShape output_shape     = arm_compute::misc::shape_calculator::compute_reduced_shape(input->info()->tensor_shape(), axis, false);
        const auto        output_data_type = input->info()->data_type();
        auto_init_if_empty(*output->info(), input->info()->clone()->set_tensor_shape(output_shape).set_data_type(output_data_type).reset_padding().set_is_resizable(true));

        _memory_group.manage(&_unreshaped_output);
    }

    _reduction_kernel = std::make_unique<CLReductionOperationKernel>();
    _reduction_kernel->configure(compile_context, input, output_internal, axis, op);

    if(_is_reshape_required)
    {
        _reshape.configure(compile_context, &_unreshaped_output, output);
        _unreshaped_output.allocator()->allocate();
    }
}

operator=() [1/2]

CLReductionOperation& operator= ( const CLReductionOperation &  )

delete

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

operator=() [2/2]

CLReductionOperation& operator= ( CLReductionOperation &&  )

default

Default move assignment operator.

run()

void run ( )

overridevirtual

Run the kernels contained in the function.

For CPU kernels:

• Multi-threading is used for the kernels which are parallelisable.
• By default std::thread::hardware_concurrency() threads are used.

Note

CPPScheduler::set_num_threads() can be used to manually set the number of threads

For OpenCL kernels:

• All the kernels are enqueued on the queue associated with CLScheduler.
• The queue is then flushed.

Note

The function will not block until the kernels are executed. It is the user's responsibility to wait.

Will call prepare() on first run if hasn't been done

Implements IFunction.

Definition at line 139 of file CLReductionOperation.cpp.

References CLScheduler::enqueue(), CLScheduler::get(), and CLReshapeLayer::run().

Referenced by CLL2NormalizeLayer::run(), and CLFFTConvolutionLayer::run().

{
    MemoryGroupResourceScope scope_mg(_memory_group);

    CLScheduler::get().enqueue(*_reduction_kernel, false);

    if(_is_reshape_required)
    {
        _reshape.run();
    }
}

validate()

Status validate ( const ITensorInfo *  input, const ITensorInfo *  output, unsigned int  axis, ReductionOperation  op, bool  keep_dims = true  )

static

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

Parameters

[in]	input	Source tensor info. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32/S32.
[in]	output	Destination tensor info. Data types and data layouts supported: Same as input.
[in]	axis	Axis along which to reduce. Supported reduction axis : 0, 1, 2, 3
[in]	op	Reduction operation to perform. Operations supported: MEAN_SUM, PROD, SUM_SQUARE, SUM, MIN, MAX
[in]	keep_dims	(Optional) Whether to keep the reduced dimension after the operation. Defaults to true.

Returns

a status

Definition at line 48 of file CLReductionOperation.cpp.

References CLTensor::allocator(), ARM_COMPUTE_ERROR_ON_NULLPTR, ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES, ARM_COMPUTE_RETURN_ERROR_ON_MSG, ARM_COMPUTE_RETURN_ON_ERROR, ICloneable< T >::clone(), arm_compute::misc::shape_calculator::compute_reduced_shape(), arm_compute::test::validation::data_type, ITensorInfo::data_type(), ITensor::info(), ITensorAllocator::init(), arm_compute::test::validation::input, arm_compute::test::validation::input_shape, ITensorInfo::num_channels(), Dimensions< size_t >::num_max_dimensions, arm_compute::test::validation::qinfo, ITensorInfo::quantization_info(), TensorInfo::set_data_type(), ITensorInfo::set_num_channels(), ITensorInfo::set_quantization_info(), ITensorInfo::set_tensor_shape(), arm_compute::test::validation::shape, ITensorInfo::tensor_shape(), ITensorInfo::total_size(), CLReshapeLayer::validate(), and CLReductionOperationKernel::validate().

Referenced by arm_compute::test::validation::DATA_TEST_CASE(), and CLL2NormalizeLayer::validate().

{
    ARM_COMPUTE_ERROR_ON_NULLPTR(input, output);
    ARM_COMPUTE_RETURN_ERROR_ON_MSG(axis >= TensorShape::num_max_dimensions, "Reduction axis greater than max number of dimensions");
    ARM_COMPUTE_RETURN_ERROR_ON_MSG(axis > 3, "Unsupported reduction axis");

    const bool is_reshape_required = !keep_dims;

    if(is_reshape_required && output->total_size() != 0)
    {
        const TensorInfo expected_output_shape = output->clone()->set_tensor_shape(arm_compute::misc::shape_calculator::compute_reduced_shape(input->tensor_shape(), axis, keep_dims));
        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(&expected_output_shape, output);
    }

    auto *output_internal = output;

    TensorInfo output_before_reshape;
    const auto input_shape        = input->tensor_shape();
    const auto input_num_channles = input->num_channels();
    const auto input_qinfo        = input->quantization_info();
    const auto output_data_type   = output->data_type();

    auto initialize_tensorinfo = [](TensorInfo & ti, TensorShape shape, DataType data_type, int num_channels, QuantizationInfo qinfo)
    {
        ti.set_data_type(data_type).set_tensor_shape(shape).set_num_channels(num_channels).set_quantization_info(qinfo);
    };

    if(is_reshape_required)
    {
        auto shape_before_reshape = input_shape;
        shape_before_reshape.set(axis, 1);
        initialize_tensorinfo(output_before_reshape, shape_before_reshape, output_data_type, input_num_channles, input_qinfo);
        output_internal = &output_before_reshape;
    }

    ARM_COMPUTE_RETURN_ON_ERROR(CLReductionOperationKernel::validate(input, output_internal, axis, op));

    if(is_reshape_required)
    {
        ARM_COMPUTE_RETURN_ON_ERROR(CLReshapeLayer::validate(output_internal, output));
    }

    return Status{};
}

---

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

• arm_compute/runtime/CL/functions/CLReductionOperation.h
• src/runtime/CL/functions/CLReductionOperation.cpp