Compute Library
 23.11
ClDepthConcatenateKernel.cpp
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2017-2023 Arm Limited.
3  *
4  * SPDX-License-Identifier: MIT
5  *
6  * Permission is hereby granted, free of charge, to any person obtaining a copy
7  * of this software and associated documentation files (the "Software"), to
8  * deal in the Software without restriction, including without limitation the
9  * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
10  * sell copies of the Software, and to permit persons to whom the Software is
11  * furnished to do so, subject to the following conditions:
12  *
13  * The above copyright notice and this permission notice shall be included in all
14  * copies or substantial portions of the Software.
15  *
16  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
19  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
21  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
22  * SOFTWARE.
23  */
25 
30 #include "arm_compute/core/Utils.h"
33 
34 #include "src/core/CL/CLValidate.h"
36 #include "support/Cast.h"
37 #include "support/StringSupport.h"
38 
39 namespace arm_compute
40 {
41 namespace opencl
42 {
43 namespace kernels
44 {
45 namespace
46 {
47 Status validate_arguments(const ITensorInfo *src, unsigned int depth_offset, const ITensorInfo *dst)
48 {
54 
55  ARM_COMPUTE_RETURN_ERROR_ON(src->dimension(Window::DimX) != dst->dimension(Window::DimX));
56  ARM_COMPUTE_RETURN_ERROR_ON(src->dimension(Window::DimY) != dst->dimension(Window::DimY));
57  ARM_COMPUTE_RETURN_ERROR_ON(src->dimension(2) + depth_offset > dst->dimension(2));
59 
60  return Status{};
61 }
62 } // namespace
63 
65 {
67 }
68 
71  unsigned int depth_offset,
73 {
76 
77  auto padding_info = get_padding_info({src, dst});
78 
79  _depth_offset = depth_offset;
80 
81  const unsigned int num_elems_processed_per_iteration = adjust_vec_size(16 / src->element_size(), src->dimension(0));
82 
83  // Add build options
84  CLBuildOptions build_opts;
85  build_opts.add_option("-DDATA_TYPE=" + get_cl_type_from_data_type(src->data_type()));
87  build_opts.add_option("-DVEC_SIZE_LEFTOVER=" +
89  if (is_data_type_quantized_asymmetric(src->data_type()) && src->quantization_info() != dst->quantization_info())
90  {
91  const UniformQuantizationInfo iq_info = src->quantization_info().uniform();
92  const UniformQuantizationInfo oq_info = dst->quantization_info().uniform();
93 
94  build_opts.add_option("-DOFFSET_IN1=" + float_to_string_with_full_precision(iq_info.offset));
95  build_opts.add_option("-DOFFSET_OUT=" + float_to_string_with_full_precision(oq_info.offset));
96  build_opts.add_option("-DSCALE_IN1=" + float_to_string_with_full_precision(iq_info.scale));
97  build_opts.add_option("-DSCALE_OUT=" + float_to_string_with_full_precision(oq_info.scale));
98  }
99 
100  std::string kernel_name = "concatenate";
101 
102  // A macro guard to compile ONLY the kernel of interest
103  build_opts.add_option("-D" + upper_string(kernel_name));
104 
105  // Create kernel
106  _kernel = create_kernel(compile_context, kernel_name, build_opts.options());
107 
108  // Configure kernel window
110  win.set(Window::DimZ, Window::Dimension(0, src->tensor_shape().z(), 1));
111  ICLKernel::configure_internal(win);
112 
114 }
115 
117  unsigned int depth_offset,
119 {
121  return Status{};
122 }
123 
124 void ClDepthConcatenateKernel::run_op(ITensorPack &tensors, const Window &window, ::cl::CommandQueue &queue)
125 {
128 
129  const auto src =
130  utils::cast::polymorphic_downcast<const ICLTensor *>(tensors.get_const_tensor(TensorType::ACL_SRC));
131  auto dst = utils::cast::polymorphic_downcast<ICLTensor *>(tensors.get_tensor(TensorType::ACL_DST));
132 
134 
135  const int offset_to_first_elements_in_bytes = _depth_offset * dst->info()->strides_in_bytes()[2];
136 
137  unsigned int idx = 2 * num_arguments_per_3D_tensor(); // Skip the src and dst parameters
138  _kernel.setArg<cl_int>(idx, offset_to_first_elements_in_bytes);
139 
140  do
141  {
142  unsigned int idx = 0;
145  enqueue(queue, *this, slice, lws_hint());
147 }
148 } // namespace kernels
149 } // namespace opencl
150 } // namespace arm_compute
arm_compute::Steps
Class to describe a number of elements in each dimension.
Definition: Steps.h:40
arm_compute::support::cpp11::to_string
std::string to_string(T &&value)
Convert integer and float values to string.
Definition: StringSupport.h:168
Cast.h
StringSupport.h
ClDepthConcatenateKernel.h
arm_compute::UniformQuantizationInfo::offset
int32_t offset
Definition: QuantizationInfo.h:63
arm_compute::test::validation::src
SimpleTensor< float > src
Definition: DFT.cpp:155
AdjustVecSize.h
ICLTensor.h
Helpers.h
arm_compute::CLBuildOptions::options
const StringSet & options() const
Gets the current options list set.
Definition: CLCompileContext.cpp:72
arm_compute::calculate_max_window
Window calculate_max_window(const ValidRegion &valid_region, const Steps &steps, bool skip_border, BorderSize border_size)
Definition: WindowHelpers.cpp:29
arm_compute::DataType::QASYMM8
@ QASYMM8
quantized, asymmetric fixed-point 8-bit number unsigned
arm_compute::test::validation::dst
auto dst
Definition: DFT.cpp:170
arm_compute::cpu::kernels::validate_arguments
Status validate_arguments(const ITensorInfo *src, const ITensorInfo *weights, const ITensorInfo *dst, const PadStrideInfo &conv_info)
Definition: CpuDirectConv2dKernel.cpp:57
ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL
#define ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(k)
Definition: Validate.h:1079
ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(...)
Definition: Validate.h:574
arm_compute::Window::DimX
static constexpr size_t DimX
Alias for dimension 0 also known as X dimension.
Definition: Window.h:43
arm_compute::ITensorPack::get_tensor
ITensor * get_tensor(int id)
Get tensor of a given id from the pac.
Definition: ITensorPack.cpp:63
arm_compute::upper_string
std::string upper_string(const std::string &val)
Raise a given string to upper case.
Definition: StringUtils.cpp:45
arm_compute::UniformQuantizationInfo
Quantization info when assuming per layer quantization.
Definition: QuantizationInfo.h:42
ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(...)
Definition: Validate.h:677
ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN
#define ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(t, c,...)
Definition: Validate.h:952
CLKernelLibrary.h
Manages all the OpenCL kernels compilation and caching, provides accessors for the OpenCL Context.
StringUtils.h
ARM_COMPUTE_RETURN_ON_ERROR
#define ARM_COMPUTE_RETURN_ON_ERROR(status)
Checks if a status contains an error and returns it.
Definition: Error.h:205
ARM_COMPUTE_ERROR_ON_NULLPTR
#define ARM_COMPUTE_ERROR_ON_NULLPTR(...)
Definition: Validate.h:159
ARM_COMPUTE_ERROR_ON
#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
arm_compute::ITensorPack::get_const_tensor
const ITensor * get_const_tensor(int id) const
Get constant tensor of a given id.
Definition: ITensorPack.cpp:53
ARM_COMPUTE_ERROR_THROW_ON
#define ARM_COMPUTE_ERROR_THROW_ON(status)
Definition: Error.h:455
arm_compute::ITensorPack
Tensor packing service.
Definition: ITensorPack.h:39
arm_compute::CLBuildOptions::add_option
void add_option(std::string option)
Adds option to the existing build option list.
Definition: CLCompileContext.cpp:41
arm_compute::CLCompileContext
CLCompileContext class.
Definition: CLCompileContext.h:204
ARM_COMPUTE_RETURN_ERROR_ON
#define ARM_COMPUTE_RETURN_ERROR_ON(cond)
If the condition is true, an error is returned.
Definition: Error.h:298
arm_compute::ACL_DST
@ ACL_DST
Definition: Types.h:55
ARM_COMPUTE_RETURN_ERROR_ON_F16_UNSUPPORTED
#define ARM_COMPUTE_RETURN_ERROR_ON_F16_UNSUPPORTED(tensor)
Definition: CLValidate.h:36
arm_compute::create_kernel
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:409
arm_compute::Status
Status class.
Definition: Error.h:52
arm_compute::opencl::kernels::ClDepthConcatenateKernel::run_op
void run_op(ITensorPack &tensors, const Window &window, ::cl::CommandQueue &queue) override
Definition: ClDepthConcatenateKernel.cpp:124
arm_compute::DataType::QASYMM8_SIGNED
@ QASYMM8_SIGNED
quantized, asymmetric fixed-point 8-bit number signed
WindowHelpers.h
ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW
#define ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(f, s)
Definition: Validate.h:203
arm_compute::float_to_string_with_full_precision
std::string float_to_string_with_full_precision(float val)
Create a string with the float in full precision.
Definition: StringUtils.cpp:52
arm_compute::Window::slide_window_slice_3D
bool slide_window_slice_3D(Window &slice) const
Slide the passed 3D window slice.
Definition: Window.h:350
arm_compute::Window::Dimension
Describe one of the image's dimensions with a start, end and step.
Definition: Window.h:79
arm_compute::Window::DimY
static constexpr size_t DimY
Alias for dimension 1 also known as Y dimension.
Definition: Window.h:45
arm_compute::opencl::kernels::ClDepthConcatenateKernel::ClDepthConcatenateKernel
ClDepthConcatenateKernel()
Definition: ClDepthConcatenateKernel.cpp:64
arm_compute::opencl::kernels::ClDepthConcatenateKernel::configure
void configure(const CLCompileContext &compile_context, ITensorInfo *src, unsigned int depth_offset, ITensorInfo *dst)
Initialise the kernel's source and destination.
Definition: ClDepthConcatenateKernel.cpp:69
arm_compute::Window::first_slice_window_3D
Window first_slice_window_3D() const
First 3D slice of the window.
Definition: Window.h:306
CLValidate.h
arm_compute::opencl::kernels::ClDepthConcatenateKernel::validate
static Status validate(const ITensorInfo *src, unsigned int depth_offset, const ITensorInfo *dst)
Static function to check if given info will lead to a valid configuration.
Definition: ClDepthConcatenateKernel.cpp:116
arm_compute::IKernel::window
const Window & window() const
The maximum window the kernel can be executed on.
Definition: IKernel.cpp:28
arm_compute::get_cl_type_from_data_type
std::string get_cl_type_from_data_type(const DataType &dt)
Translates a tensor data type to the appropriate OpenCL type.
Definition: CLHelpers.cpp:41
arm_compute::ICLKernel::add_3D_tensor_argument
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.
Definition: ICLKernel.h:234
Utils.h
arm_compute::UniformQuantizationInfo::scale
float scale
Definition: QuantizationInfo.h:62
arm_compute::Window
Describe a multidimensional execution window.
Definition: Window.h:39
arm_compute::ELEMENTWISE
@ ELEMENTWISE
Elementwise CL kernel type.
Definition: CLTypes.h:83
arm_compute::ICLKernel::num_arguments_per_3D_tensor
constexpr static unsigned int num_arguments_per_3D_tensor()
Returns the number of arguments enqueued per 3D tensor object.
Definition: ICLKernel.h:321
arm_compute
Copyright (c) 2017-2023 Arm Limited.
Definition: introduction.dox:24
arm_compute::DataType::F16
@ F16
16-bit floating-point number
arm_compute::adjust_vec_size
unsigned int adjust_vec_size(unsigned int vec_size, size_t dim0)
Returns the adjusted vector size in case it is less than the input's first dimension,...
Definition: AdjustVecSize.h:38
arm_compute::has_padding_changed
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:491
arm_compute::Window::DimZ
static constexpr size_t DimZ
Alias for dimension 2 also known as Z dimension.
Definition: Window.h:47
arm_compute::ICLKernel::lws_hint
cl::NDRange lws_hint() const
Return the Local-Workgroup-Size hint.
Definition: ICLKernel.h:383
arm_compute::is_data_type_quantized_asymmetric
bool is_data_type_quantized_asymmetric(DataType dt)
Check if a given data type is of asymmetric quantized type.
Definition: DataTypeUtils.h:346
ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR
#define ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(...)
Definition: Validate.h:161
arm_compute::ACL_SRC
@ ACL_SRC
Definition: Types.h:44
arm_compute::ITensorInfo
Store the tensor's metadata.
Definition: ITensorInfo.h:44
arm_compute::DataType::F32
@ F32
32-bit floating-point number
arm_compute::CLBuildOptions
Build options.
Definition: CLCompileContext.h:38
num_elems_processed_per_iteration
unsigned int num_elems_processed_per_iteration
Definition: ClIm2ColKernel.cpp:60
arm_compute::get_padding_info
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:476
arm_compute::test::validation::reference::slice
SimpleTensor< T > slice(const SimpleTensor< T > &src, Coordinates starts, Coordinates ends)
Definition: SliceOperations.cpp:38
CLHelpers.h
kernel_name
std::string kernel_name
Definition: ClIm2ColKernel.cpp:58
arm_compute::enqueue
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:33