Compute Library
 21.02
CLArithmeticAddition Class Reference

Basic function to run opencl::kernels::ClSaturatedArithmeticKernel for addition. More...

#include <CLElementwiseOperations.h>

Collaboration diagram for CLArithmeticAddition:
[legend]

Public Member Functions

 CLArithmeticAddition ()
 Default Constructor. More...
 
 ~CLArithmeticAddition ()
 Default Destructor. More...
 
 CLArithmeticAddition (const CLArithmeticAddition &)=delete
 Prevent instances of this class from being copied (As this class contains pointers) More...
 
 CLArithmeticAddition (CLArithmeticAddition &&)
 Default move constructor. More...
 
CLArithmeticAdditionoperator= (const CLArithmeticAddition &)=delete
 Prevent instances of this class from being copied (As this class contains pointers) More...
 
CLArithmeticAdditionoperator= (CLArithmeticAddition &&)
 Default move assignment operator. More...
 
void configure (ICLTensor *input1, ICLTensor *input2, ICLTensor *output, ConvertPolicy policy, const ActivationLayerInfo &act_info=ActivationLayerInfo())
 Initialise the kernel's inputs, output and conversion policy. More...
 
void configure (const CLCompileContext &compile_context, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, ConvertPolicy policy, const ActivationLayerInfo &act_info=ActivationLayerInfo())
 Initialise the kernel's inputs, output and conversion policy. 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 *input1, const ITensorInfo *input2, const ITensorInfo *output, ConvertPolicy policy, const ActivationLayerInfo &act_info=ActivationLayerInfo())
 Static function to check if given info will lead to a valid configuration of opencl::kernels::ClSaturatedArithmeticKernel for addition. More...
 

Detailed Description

Basic function to run opencl::kernels::ClSaturatedArithmeticKernel for addition.

Note
The tensor data type for the inputs must be U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/S32/F16/F32.
The function performs an arithmetic addition between two tensors.

Definition at line 41 of file CLElementwiseOperations.h.

Constructor & Destructor Documentation

◆ CLArithmeticAddition() [1/3]

Default Constructor.

Definition at line 45 of file CLElementwiseOperations.cpp.

References CLArithmeticAddition::operator=(), and CLArithmeticAddition::~CLArithmeticAddition().

46  : _impl(std::make_unique<Impl>())
47 {
48 }

◆ ~CLArithmeticAddition()

~CLArithmeticAddition ( )
default

Default Destructor.

Referenced by CLArithmeticAddition::CLArithmeticAddition().

◆ CLArithmeticAddition() [2/3]

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

◆ CLArithmeticAddition() [3/3]

Default move constructor.

Member Function Documentation

◆ configure() [1/2]

void configure ( ICLTensor input1,
ICLTensor input2,
ICLTensor output,
ConvertPolicy  policy,
const ActivationLayerInfo act_info = ActivationLayerInfo() 
)

Initialise the kernel's inputs, output and conversion policy.

Valid configurations (Input1,Input2) -> Output :

  • (U8,U8) -> U8
  • (U8,U8) -> S16
  • (S16,U8) -> S16
  • (U8,S16) -> S16
  • (S16,S16) -> S16
  • (S32,S32) -> S32
  • (F16,F16) -> F16
  • (F32,F32) -> F32
  • (QASYMM8,QASYMM8) -> QASYMM8
  • (QASYMM8_SIGNED,QASYMM8_SIGNED) -> QASYMM8_SIGNED
  • (QSYMM16,QSYMM16) -> QSYMM16
Parameters
[in,out]input1First tensor input. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/S32/F16/F32. The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
[in,out]input2Second tensor input. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/S32/F16/F32. The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
[out]outputOutput tensor. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/S32/F16/F32.
[in]policyPolicy to use to handle overflow.
[in]act_info(Optional) Activation layer information in case of a fused activation.

Definition at line 53 of file CLElementwiseOperations.cpp.

References CLKernelLibrary::get().

Referenced by CLRNNLayer::configure(), CLFFTConvolutionLayer::configure(), CLLSTMLayerQuantized::configure(), CLLSTMLayer::configure(), and CLQLSTMLayer::configure().

54 {
55  configure(CLKernelLibrary::get().get_compile_context(), input1, input2, output, policy, act_info);
56 }
static CLKernelLibrary & get()
Access the KernelLibrary singleton.
void configure(ICLTensor *input1, ICLTensor *input2, ICLTensor *output, ConvertPolicy policy, const ActivationLayerInfo &act_info=ActivationLayerInfo())
Initialise the kernel&#39;s inputs, output and conversion policy.

◆ configure() [2/2]

void configure ( const CLCompileContext compile_context,
const ICLTensor input1,
const ICLTensor input2,
ICLTensor output,
ConvertPolicy  policy,
const ActivationLayerInfo act_info = ActivationLayerInfo() 
)

Initialise the kernel's inputs, output and conversion policy.

Valid configurations (Input1,Input2) -> Output :

  • (U8,U8) -> U8
  • (U8,U8) -> S16
  • (S16,U8) -> S16
  • (U8,S16) -> S16
  • (S16,S16) -> S16
  • (S32,S32) -> S32
  • (F16,F16) -> F16
  • (F32,F32) -> F32
  • (QASYMM8,QASYMM8) -> QASYMM8
  • (QASYMM8_SIGNED,QASYMM8_SIGNED) -> QASYMM8_SIGNED
  • (QSYMM16,QSYMM16) -> QSYMM16
Parameters
[in]compile_contextThe compile context to be used.
[in,out]input1First tensor input. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/S32/F16/F32. The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
[in,out]input2Second tensor input. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/S32/F16/F32. The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
[out]outputOutput tensor. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/S32/F16/F32.
[in]policyPolicy to use to handle overflow.
[in]act_info(Optional) Activation layer information in case of a fused activation.

Definition at line 58 of file CLElementwiseOperations.cpp.

References ITensor::info().

60 {
61  _impl->src_0 = input1;
62  _impl->src_1 = input2;
63  _impl->dst = output;
64  _impl->op = std::make_unique<opencl::ClAdd>();
65  _impl->op->configure(compile_context, input1->info(), input2->info(), output->info(), policy, act_info);
66 }

◆ operator=() [1/2]

CLArithmeticAddition& operator= ( const CLArithmeticAddition )
delete

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

Referenced by CLArithmeticAddition::CLArithmeticAddition().

◆ operator=() [2/2]

CLArithmeticAddition & operator= ( CLArithmeticAddition &&  )
default

Default move assignment operator.

◆ run()

void run ( )
overridevirtual

Run the kernels contained in the function.

For Neon 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 73 of file CLElementwiseOperations.cpp.

References arm_compute::ACL_DST, arm_compute::ACL_SRC_0, arm_compute::ACL_SRC_1, ITensorPack::add_tensor(), and arm_compute::test::validation::dst.

Referenced by CLQLSTMLayer::prepare(), CLRNNLayer::run(), CLFFTConvolutionLayer::run(), CLLSTMLayerQuantized::run(), CLLSTMLayer::run(), and CLQLSTMLayer::run().

74 {
75  ITensorPack pack;
76  pack.add_tensor(TensorType::ACL_SRC_0, _impl->src_0);
77  pack.add_tensor(TensorType::ACL_SRC_1, _impl->src_1);
78  pack.add_tensor(TensorType::ACL_DST, _impl->dst);
79 
80  _impl->op->run(pack);
81 }

◆ validate()

Status validate ( const ITensorInfo input1,
const ITensorInfo input2,
const ITensorInfo output,
ConvertPolicy  policy,
const ActivationLayerInfo act_info = ActivationLayerInfo() 
)
static

Static function to check if given info will lead to a valid configuration of opencl::kernels::ClSaturatedArithmeticKernel for addition.

Valid configurations (Input1,Input2) -> Output :

  • (U8,U8) -> U8
  • (U8,U8) -> S16
  • (S16,U8) -> S16
  • (U8,S16) -> S16
  • (S16,S16) -> S16
  • (S32,S32) -> S32
  • (F16,F16) -> F16
  • (F32,F32) -> F32
  • (QASYMM8,QASYMM8) -> QASYMM8
  • (QASYMM8_SIGNED,QASYMM8_SIGNED) -> QASYMM8_SIGNED
  • (QSYMM16,QSYMM16) -> QSYMM16
Parameters
[in]input1First tensor input info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/S32/F16/F32.
[in]input2Second tensor input info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/S32/F16/F32.
[in]outputOutput tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/S32/F16/F32.
[in]policyPolicy to use to handle overflow.
[in]act_info(Optional) Activation layer information in case of a fused activation.
Returns
a status

Definition at line 68 of file CLElementwiseOperations.cpp.

References ClAdd::validate().

Referenced by arm_compute::test::validation::DATA_TEST_CASE(), arm_compute::test::validation::TEST_CASE(), CLRNNLayer::validate(), CLLSTMLayerQuantized::validate(), CLLSTMLayer::validate(), and CLQLSTMLayer::validate().

69 {
70  return opencl::ClAdd::validate(input1, input2, output, policy, act_info);
71 }
static Status validate(const ITensorInfo *src1, const ITensorInfo *src2, const ITensorInfo *dst, ConvertPolicy policy, const ActivationLayerInfo &act_info=ActivationLayerInfo())
Static function to check if given info will lead to a valid configuration of ClAdd.
Definition: ClAdd.cpp:41

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