Compute Library
 21.02
CLIntegralImageKernel.cpp
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2016-2020 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 
33 #include "support/StringSupport.h"
34 
35 #include <cstddef>
36 
37 using namespace arm_compute;
38 
40 {
41  configure(CLKernelLibrary::get().get_compile_context(), input, output);
42 }
43 
44 void CLIntegralImageHorKernel::configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output)
45 {
48 
49  _input = input;
50  _output = output;
51 
52  // Create kernel
53  const std::string kernel_name = std::string("integral_horizontal");
54  _kernel = create_kernel(compile_context, kernel_name);
55 
56  // Configure kernel window
57  const unsigned int num_elems_processed_per_iteration = input->info()->dimension(0);
58  const unsigned int num_elems_accessed_per_iteration = ceil_to_multiple(num_elems_processed_per_iteration, 16);
59 
60  Window win = calculate_max_window(*input->info(), Steps(num_elems_processed_per_iteration));
61  AccessWindowHorizontal output_access(output->info(), 0, num_elems_accessed_per_iteration);
62 
64  AccessWindowHorizontal(input->info(), 0, num_elems_accessed_per_iteration),
65  output_access);
66 
67  output_access.set_valid_region(win, input->info()->valid_region());
68 
69  ICLKernel::configure_internal(win);
70 
71  // Set config_id for enabling LWS tuning
72  _config_id = kernel_name;
73  _config_id += "_";
74  _config_id += lower_string(string_from_data_type(input->info()->data_type()));
75  _config_id += "_";
76  _config_id += support::cpp11::to_string(input->info()->dimension(0));
77  _config_id += "_";
78  _config_id += support::cpp11::to_string(input->info()->dimension(1));
79  _config_id += "_";
80  _config_id += support::cpp11::to_string(output->info()->dimension(0));
81  _config_id += "_";
82  _config_id += support::cpp11::to_string(output->info()->dimension(1));
83 }
84 
86  : _in_out(nullptr)
87 {
88 }
89 
91 {
92  configure(CLKernelLibrary::get().get_compile_context(), in_out);
93 }
94 
96 {
98 
99  _in_out = in_out;
100 
101  // Create kernel
102  const std::string kernel_name = std::string("integral_vertical");
103  _kernel = create_kernel(compile_context, kernel_name);
104 
105  // Configure kernel window
106  constexpr unsigned int num_elems_processed_per_iteration_x = 8;
107  const unsigned int num_elems_processed_per_iteration_y = in_out->info()->dimension(Window::DimY);
108 
109  Window win = calculate_max_window(*in_out->info(), Steps(num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y));
110 
111  AccessWindowRectangle in_out_access(in_out->info(), 0, 0, num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y);
112 
113  update_window_and_padding(win, in_out_access);
114 
115  in_out_access.set_valid_region(win, in_out->info()->valid_region());
116 
117  ICLKernel::configure_internal(win);
118 
119  // Set config_id for enabling LWS tuning
120  _config_id = kernel_name;
121  _config_id += "_";
122  _config_id += lower_string(string_from_data_type(in_out->info()->data_type()));
123  _config_id += "_";
124  _config_id += support::cpp11::to_string(in_out->info()->dimension(0));
125  _config_id += "_";
126  _config_id += support::cpp11::to_string(in_out->info()->dimension(1));
127 }
128 
129 void CLIntegralImageVertKernel::run(const Window &window, cl::CommandQueue &queue)
130 {
133 
134  const size_t height = _in_out->info()->dimension(1);
135 
137 
138  do
139  {
140  unsigned int idx = 0;
141  add_2D_tensor_argument(idx, _in_out, slice);
142  _kernel.setArg<cl_uint>(idx++, height);
143  enqueue(queue, *this, slice, lws_hint());
144  }
145  while(window.slide_window_slice_2D(slice));
146 }
Window first_slice_window_2D() const
First 2D slice of the window.
Definition: Window.h:283
Window calculate_max_window(const ValidRegion &valid_region, const Steps &steps, bool skip_border, BorderSize border_size)
const Window & window() const
The maximum window the kernel can be executed on.
Definition: IKernel.cpp:28
void enqueue(IGCKernel &kernel, const Window &window, const gles::NDRange &lws=gles::NDRange(1U, 1U, 1U))
Add the kernel to the command queue with the given window.
Definition: IGCKernel.cpp:41
virtual size_t dimension(size_t index) const =0
Return the size of the requested dimension.
cl::NDRange lws_hint() const
Return the Local-Workgroup-Size hint.
Definition: ICLKernel.h:276
1 channel, 1 U8 per channel
std::string to_string(T &&value)
Convert integer and float values to string.
virtual DataType data_type() const =0
Data type used for each element of the tensor.
static CLKernelLibrary & get()
Access the KernelLibrary singleton.
std::string lower_string(const std::string &val)
Lower a given string.
Definition: Utils.cpp:350
bool slide_window_slice_2D(Window &slice) const
Slide the passed 2D window slice.
Definition: Window.h:323
Copyright (c) 2017-2021 Arm Limited.
virtual ValidRegion valid_region() const =0
Valid region of the tensor.
Implementation of a rectangular access pattern.
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:403
const std::string & string_from_data_type(DataType dt)
Convert a data type identity into a string.
Definition: Utils.cpp:135
bool update_window_and_padding(Window &win, Ts &&... patterns)
Update window and padding size for each of the access patterns.
Definition: WindowHelpers.h:46
1 channel, 1 U32 per channel
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
Class to describe a number of elements in each dimension.
Definition: Steps.h:40
void run(const Window &window, cl::CommandQueue &queue) override
Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue...
Implementation of a row access pattern.
std::string kernel_name
virtual ITensorInfo * info() const =0
Interface to be implemented by the child class to return the tensor&#39;s metadata.
void configure(const ICLTensor *input, ICLTensor *output)
Initialise the kernel&#39;s input and output.
#define ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(k)
Definition: Validate.h:941
#define ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(t, c,...)
Definition: Validate.h:790
CLCompileContext class.
static constexpr size_t DimY
Alias for dimension 1 also known as Y dimension.
Definition: Window.h:45
void add_2D_tensor_argument(unsigned int &idx, const ICLTensor *tensor, const Window &window)
Add the passed 2D tensor&#39;s parameters to the object&#39;s kernel&#39;s arguments starting from the index idx...
Definition: ICLKernel.h:148
Interface for OpenCL tensor.
Definition: ICLTensor.h:42
Manages all the OpenCL kernels compilation and caching, provides accessors for the OpenCL Context...
Wrapper to configure the Khronos OpenCL C++ header.
unsigned int num_elems_processed_per_iteration
void configure(ICLTensor *in_out)
Initialise the kernel&#39;s input and output.
Describe a multidimensional execution window.
Definition: Window.h:39
#define ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(f, s)
Definition: Validate.h:205
SimpleTensor< T > slice(const SimpleTensor< T > &src, Coordinates starts, Coordinates ends)