Compute Library
 22.08
ClQuantizeKernel Class Reference

Interface for the quantization layer kernel. More...

#include <ClQuantizeKernel.h>

Collaboration diagram for ClQuantizeKernel:
[legend]

Public Member Functions

 ClQuantizeKernel ()
 
 ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE (ClQuantizeKernel)
 
void configure (const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst)
 Set the input, output. More...
 
void run_op (ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override
 Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue. More...
 
- Public Member Functions inherited from ICLKernel
 ICLKernel ()
 Constructor. More...
 
cl::Kernel & kernel ()
 Returns a reference to the OpenCL kernel of this object. More...
 
CLKernelType type () const
 Returns the CL kernel type. More...
 
template<typename T >
void add_1D_array_argument (unsigned int &idx, const ICLArray< T > *array, const Strides &strides, unsigned int num_dimensions, const Window &window)
 Add the passed 1D array's parameters to the object's kernel's arguments starting from the index idx. More...
 
void add_1D_tensor_argument (unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 1D tensor's parameters to the object's kernel's arguments starting from the index idx. More...
 
void add_1D_tensor_argument_if (bool cond, unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 1D tensor's parameters to the object's kernel's arguments starting from the index idx if the condition is true. More...
 
void add_2D_tensor_argument (unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 2D tensor's parameters to the object's kernel's arguments starting from the index idx. More...
 
void add_2D_tensor_argument_if (bool cond, unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 2D tensor's parameters to the object's kernel's arguments starting from the index idx if the condition is true. More...
 
void add_3D_tensor_argument (unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 3D tensor's parameters to the object's kernel's arguments starting from the index idx. More...
 
void add_4D_tensor_argument (unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 4D tensor's parameters to the object's kernel's arguments starting from the index idx. More...
 
void add_5D_tensor_argument (unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 5D tensor's parameters to the object's kernel's arguments starting from the index idx. More...
 
void add_3d_tensor_nhw_argument (unsigned int &idx, const ICLTensor *tensor)
 Add the passed NHW 3D tensor's parameters to the object's kernel's arguments by passing strides, dimensions and the offset to the first valid element in bytes. More...
 
void add_4d_tensor_nhwc_argument (unsigned int &idx, const ICLTensor *tensor)
 Add the passed NHWC 4D tensor's parameters to the object's kernel's arguments by passing strides, dimensions and the offset to the first valid element in bytes. More...
 
virtual void run (const Window &window, cl::CommandQueue &queue)
 Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue. More...
 
virtual void run_composite_op (ITensorPack &tensors, const Window &window, cl::CommandQueue &queue, const experimental::dynamic_fusion::ClExecutionDescriptor &exec_desc)
 The execution is carried out through run_op method. But the run_op method needs to be extended to include ClExecutionDescriptor as now LWS GWS tuning will be separated from the IKernel. More...
 
template<typename T >
void add_argument (unsigned int &idx, T value)
 Add the passed parameters to the object's kernel's arguments starting from the index idx. More...
 
void set_lws_hint (const cl::NDRange &lws_hint)
 Set the Local-Workgroup-Size hint. More...
 
cl::NDRange lws_hint () const
 Return the Local-Workgroup-Size hint. More...
 
void set_wbsm_hint (const cl_int &wbsm_hint)
 Set the workgroup batch size modifier hint. More...
 
cl_int wbsm_hint () const
 Return the workgroup batch size modifier hint. More...
 
const std::string & config_id () const
 Get the configuration ID. More...
 
void set_target (GPUTarget target)
 Set the targeted GPU architecture. More...
 
void set_target (cl::Device &device)
 Set the targeted GPU architecture according to the CL device. More...
 
GPUTarget get_target () const
 Get the targeted GPU architecture. More...
 
size_t get_max_workgroup_size ()
 Get the maximum workgroup size for the device the CLKernelLibrary uses. More...
 
template<unsigned int dimension_size>
void add_tensor_argument (unsigned &idx, const ICLTensor *tensor, const Window &window)
 
template<typename T , unsigned int dimension_size>
void add_array_argument (unsigned &idx, const ICLArray< T > *array, const Strides &strides, unsigned int num_dimensions, const Window &window)
 Add the passed array's parameters to the object's kernel's arguments starting from the index idx. 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 Windowwindow () 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 ICLKernel
static constexpr unsigned int num_arguments_per_3d_tensor_nhw ()
 Returns the number of arguments enqueued per NHW 3D Tensor object. More...
 
static constexpr unsigned int num_arguments_per_4d_tensor_nhwc ()
 Returns the number of arguments enqueued per NHWC 4D Tensor object. More...
 
static constexpr unsigned int num_arguments_per_1D_array ()
 Returns the number of arguments enqueued per 1D array object. More...
 
static constexpr unsigned int num_arguments_per_1D_tensor ()
 Returns the number of arguments enqueued per 1D tensor object. More...
 
static constexpr unsigned int num_arguments_per_2D_tensor ()
 Returns the number of arguments enqueued per 2D tensor object. More...
 
static constexpr unsigned int num_arguments_per_3D_tensor ()
 Returns the number of arguments enqueued per 3D tensor object. More...
 
static constexpr unsigned int num_arguments_per_4D_tensor ()
 Returns the number of arguments enqueued per 4D tensor object. More...
 
static cl::NDRange gws_from_window (const Window &window)
 Get the global work size given an execution window. More...
 

Detailed Description

Interface for the quantization layer kernel.

Note
The implementation supports only 3D input tensors.

Definition at line 41 of file ClQuantizeKernel.h.

Constructor & Destructor Documentation

◆ ClQuantizeKernel()

Definition at line 64 of file ClQuantizeKernel.cpp.

References arm_compute::ELEMENTWISE.

65 {
67 }
Elementwise CL kernel type.
Definition: CLTypes.h:85

Member Function Documentation

◆ ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE()

ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE ( ClQuantizeKernel  )

◆ configure()

void configure ( const CLCompileContext compile_context,
const ITensorInfo src,
ITensorInfo dst 
)

Set the input, output.

Parameters
[in]compile_contextThe compile context to be used.
[in]srcSource tensor info. Data types supported: QASYMM8/QASYMM8_SIGNED/F32/F16.
[out]dstDestination tensor info with the same dimensions of input. Data types supported: QASYMM8/QASYMM8_SIGNED/QASYMM16.
Note
Output auto initialization is not supported by this kernel

Definition at line 69 of file ClQuantizeKernel.cpp.

References CLBuildOptions::add_option(), CLBuildOptions::add_option_if(), ARM_COMPUTE_ERROR_ON, ARM_COMPUTE_ERROR_ON_NULLPTR, ARM_COMPUTE_ERROR_THROW_ON, arm_compute::calculate_max_window(), arm_compute::ceil_to_multiple(), arm_compute::create_kernel(), ITensorInfo::data_type(), Window::DimX, ITensorInfo::element_size(), arm_compute::float_to_string_with_full_precision(), arm_compute::get_cl_type_from_data_type(), arm_compute::quantization::get_min_max_values_from_quantized_data_type(), arm_compute::get_padding_info(), arm_compute::has_padding_changed(), arm_compute::is_data_type_float(), arm_compute::is_data_type_quantized_asymmetric(), UniformQuantizationInfo::offset, CLBuildOptions::options(), arm_compute::test::validation::qinfo, ITensorInfo::quantization_info(), UniformQuantizationInfo::scale, arm_compute::test::validation::src, ITensorInfo::tensor_shape(), arm_compute::support::cpp11::to_string(), QuantizationInfo::uniform(), arm_compute::cpu::kernels::validate_arguments(), and Dimensions< T >::x().

70 {
72 
73  auto padding_info = get_padding_info({ src, dst });
74 
76 
77  const int vec_size_x = 16 / src->element_size();
78  const int input_width_x = src->tensor_shape().x();
79  const bool multi_access_x = (input_width_x / vec_size_x > 0);
80 
81  const UniformQuantizationInfo qinfo = dst->quantization_info().uniform();
82  const DataType output_data_type = dst->data_type();
83 
84  float scale_to_apply = qinfo.scale;
85  int32_t offset_to_apply = qinfo.offset;
86  if(is_data_type_quantized_asymmetric(src->data_type()))
87  {
88  /*
89  * In case of requantization of a quantized input tensor to an output tensor with another quantization
90  * instead of of apply dequantization and then a quantization functions, we just compute new scale and
91  * offset to apply.
92  *
93  * Assuming:
94  * - q_i as input quantized value
95  * - q_o as output quantized value
96  * - z_i as input quantization offset value
97  * - z_o as output quantization offset value
98  * - s_i as input quantization scale value
99  * - s_o as output quantization scale value
100  * - z_n as new quantization offset value
101  * - s_n as new quantization scale value
102  *
103  * q_o = ( q_i - z_i ) * s_i / s_o + z_o
104  *
105  * We can rewrite the formula as:
106  *
107  * q_o = ( q_i * s_i / s_o ) - z_i * s_i / s_o + z_o
108  *
109  * q_o = q_i / s_n + z_n
110  *
111  * Where:
112  *
113  * s_n = s_o / s_i
114  *
115  * z_n = - z_i * s_i / s_o + z_o
116  *
117  */
118  const UniformQuantizationInfo qinfo_in = src->quantization_info().uniform();
119  scale_to_apply /= qinfo_in.scale;
120  // In order to minimize flooring we convert the offset to a float,
121  // then compute the new offset in the float domain,
122  // finally we convert it back as int32_t
123  offset_to_apply -= static_cast<int32_t>(static_cast<float>(qinfo_in.offset) * qinfo_in.scale / qinfo.scale);
124  }
125 
126  // Create kernel
127  CLBuildOptions build_opts;
128  build_opts.add_option_if(is_data_type_float(src->data_type()), "-DIS_FLOAT");
129  build_opts.add_option("-DSCALE=" + float_to_string_with_full_precision(scale_to_apply));
130  build_opts.add_option("-DOFFSET=" + support::cpp11::to_string(offset_to_apply));
131  build_opts.add_option("-DVEC_SIZE=" + support::cpp11::to_string(vec_size_x));
132  build_opts.add_option("-DDATA_TYPE_IN=" + get_cl_type_from_data_type(src->data_type()));
133  build_opts.add_option("-DDATA_TYPE_OUT=" + get_cl_type_from_data_type(output_data_type));
134  build_opts.add_option_if(multi_access_x, "-DLAST_ACCESSED_X=" + support::cpp11::to_string(std::max<int>(input_width_x - vec_size_x, 0)));
135  std::pair<int, int> min_max_quant_values = quantization::get_min_max_values_from_quantized_data_type(output_data_type);
136  build_opts.add_option("-DMIN_QUANT_VAL=" + support::cpp11::to_string(min_max_quant_values.first));
137  build_opts.add_option("-DMAX_QUANT_VAL=" + support::cpp11::to_string(min_max_quant_values.second));
138 
139  _kernel = create_kernel(compile_context, "quantization_layer", build_opts.options());
140 
141  // Configure kernel window
142  Window win = calculate_max_window(*src, Steps());
143  if(multi_access_x)
144  {
145  win.set(Window::DimX, Window::Dimension(win.x().start(), ceil_to_multiple(win.x().end(), vec_size_x), vec_size_x));
146  }
147  ICLKernel::configure_internal(win);
148 
150 }
Window calculate_max_window(const ValidRegion &valid_region, const Steps &steps, bool skip_border, BorderSize border_size)
std::string to_string(T &&value)
Convert integer and float values to string.
#define ARM_COMPUTE_ERROR_ON(cond)
If the condition is true then an error message is printed and an exception thrown.
Definition: Error.h:466
#define ARM_COMPUTE_ERROR_THROW_ON(status)
Definition: Error.h:455
Status validate_arguments(const ITensorInfo *src, const ITensorInfo *weights, const ITensorInfo *dst, const PadStrideInfo &conv_info)
SimpleTensor< float > src
Definition: DFT.cpp:155
cl::Kernel create_kernel(const CLCompileContext &ctx, const std::string &kernel_name, const std::set< std::string > &build_opts=std::set< std::string >())
Creates an opencl kernel using a compile context.
Definition: CLHelpers.cpp:404
std::pair< int, int > get_min_max_values_from_quantized_data_type(DataType data_type)
Get minimum and maximum values for the input quantized data type.
static constexpr size_t DimX
Alias for dimension 0 also known as X dimension.
Definition: Window.h:43
std::string float_to_string_with_full_precision(float val)
Create a string with the float in full precision.
Definition: Utils.h:1124
auto ceil_to_multiple(S value, T divisor) -> decltype(((value+divisor - 1)/divisor) *divisor)
Computes the smallest number larger or equal to value that is a multiple of divisor.
Definition: Utils.h:71
std::string get_cl_type_from_data_type(const DataType &dt)
Translates a tensor data type to the appropriate OpenCL type.
Definition: CLHelpers.cpp:39
bool has_padding_changed(const std::unordered_map< const ITensorInfo *, PaddingSize > &padding_map)
Check if the previously stored padding info has changed after configuring a kernel.
Definition: Utils.cpp:601
bool is_data_type_quantized_asymmetric(DataType dt)
Check if a given data type is of asymmetric quantized type.
Definition: Utils.h:1052
const QuantizationInfo qinfo
Definition: Im2Col.cpp:155
std::unordered_map< const ITensorInfo *, PaddingSize > get_padding_info(std::initializer_list< const ITensorInfo *> infos)
Stores padding information before configuring a kernel.
Definition: Utils.cpp:586
#define ARM_COMPUTE_ERROR_ON_NULLPTR(...)
Definition: Validate.h:157
DataType
Available data types.
Definition: Types.h:79
bool is_data_type_float(DataType dt)
Check if a given data type is of floating point type.
Definition: Utils.h:1010

◆ run_op()

void run_op ( ITensorPack tensors,
const Window window,
cl::CommandQueue &  queue 
)
overridevirtual

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]tensorsA vector containing the tensors to operato on.
[in]windowRegion on which to execute the kernel. (Must be a valid region of the window returned by window()).
[in,out]queueCommand queue on which to enqueue the kernel.

Reimplemented from ICLKernel.

Definition at line 158 of file ClQuantizeKernel.cpp.

References arm_compute::ACL_DST, arm_compute::ACL_SRC, ICLKernel::add_3D_tensor_argument(), ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW, ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL, Window::collapse_if_possible(), arm_compute::enqueue(), ITensorPack::get_const_tensor(), ITensorPack::get_tensor(), ICLKernel::lws_hint(), arm_compute::test::validation::reference::slice(), and IKernel::window().

159 {
162 
163  auto src = utils::cast::polymorphic_downcast<const ICLTensor *>(tensors.get_const_tensor(TensorType::ACL_SRC));
164  auto dst = utils::cast::polymorphic_downcast<ICLTensor *>(tensors.get_tensor(TensorType::ACL_DST));
165 
166  Window window_collapsed = window.collapse_if_possible(ICLKernel::window(), 3);
167  Window slice = window_collapsed.first_slice_window_3D();
168 
169  do
170  {
171  unsigned int idx = 0;
172  add_3D_tensor_argument(idx, src, slice);
173  add_3D_tensor_argument(idx, dst, slice);
174  enqueue(queue, *this, slice, lws_hint());
175  }
176  while(window_collapsed.slide_window_slice_3D(slice));
177 }
const Window & window() const
The maximum window the kernel can be executed on.
Definition: IKernel.cpp:28
void enqueue(cl::CommandQueue &queue, ICLKernel &kernel, const Window &window, const cl::NDRange &lws_hint=CLKernelLibrary::get().default_ndrange(), bool use_dummy_work_items=false)
Add the kernel to the command queue with the given window.
Definition: ICLKernel.cpp:32
cl::NDRange lws_hint() const
Return the Local-Workgroup-Size hint.
Definition: ICLKernel.h:384
void add_3D_tensor_argument(unsigned int &idx, const ICLTensor *tensor, const Window &window)
Add the passed 3D tensor&#39;s parameters to the object&#39;s kernel&#39;s arguments starting from the index idx...
Definition: ICLKernel.h:227
SimpleTensor< float > src
Definition: DFT.cpp:155
Window collapse_if_possible(const Window &full_window, size_t first, size_t last, bool *has_collapsed=nullptr) const
Collapse the dimensions between first and last if possible.
Definition: Window.inl:68
#define ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(k)
Definition: Validate.h:915
#define ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(f, s)
Definition: Validate.h:201
SimpleTensor< T > slice(const SimpleTensor< T > &src, Coordinates starts, Coordinates ends)

◆ validate()

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

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

Similar to ClQuantizeKernel::configure()

Returns
a status

Definition at line 152 of file ClQuantizeKernel.cpp.

References ARM_COMPUTE_RETURN_ON_ERROR, and arm_compute::cpu::kernels::validate_arguments().

Referenced by ClQuantize::validate().

153 {
155  return Status{};
156 }
#define ARM_COMPUTE_RETURN_ON_ERROR(status)
Checks if a status contains an error and returns it.
Definition: Error.h:204
Status validate_arguments(const ITensorInfo *src, const ITensorInfo *weights, const ITensorInfo *dst, const PadStrideInfo &conv_info)
SimpleTensor< float > src
Definition: DFT.cpp:155

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