Compute Library
 21.02
CLScharr3x3Kernel.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 
32 
33 #include <set>
34 #include <string>
35 
36 using namespace arm_compute;
37 
39  : _run_scharr_x(false), _run_scharr_y(false), _input(nullptr), _output_x(nullptr), _output_y(nullptr)
40 {
41 }
42 
44 {
45  return BorderSize(1);
46 }
47 
48 void CLScharr3x3Kernel::configure(const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined)
49 {
50  configure(CLKernelLibrary::get().get_compile_context(), input, output_x, output_y, border_undefined);
51 }
52 
53 void CLScharr3x3Kernel::configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined)
54 {
56  ARM_COMPUTE_ERROR_ON((output_x == nullptr) && (output_y == nullptr));
57 
58  _run_scharr_x = output_x != nullptr;
59  _run_scharr_y = output_y != nullptr;
60 
61  if(_run_scharr_x)
62  {
64  }
65 
66  if(_run_scharr_y)
67  {
69  }
70 
71  _input = input;
72  _output_x = output_x;
73  _output_y = output_y;
74 
75  // Set build options
76  std::set<std::string> build_opts;
77 
78  if(_run_scharr_x)
79  {
80  build_opts.insert("-DGRAD_X");
81  }
82 
83  if(_run_scharr_y)
84  {
85  build_opts.insert("-DGRAD_Y");
86  }
87 
88  // Create kernel
89  _kernel = create_kernel(compile_context, "scharr3x3", build_opts);
90 
91  // Configure kernel window
92  constexpr unsigned int num_elems_processed_per_iteration = 8;
93  constexpr unsigned int num_elems_read_per_iteration = 16;
94  constexpr unsigned int num_elems_written_per_iteration = 8;
95  constexpr unsigned int num_rows_read_per_iteration = 3;
96 
97  Window win = calculate_max_window(*input->info(), Steps(num_elems_processed_per_iteration), border_undefined, border_size());
98 
99  AccessWindowRectangle input_access(input->info(), -border_size().left, -border_size().top, num_elems_read_per_iteration, num_rows_read_per_iteration);
100  AccessWindowHorizontal output_x_access(output_x == nullptr ? nullptr : output_x->info(), 0, num_elems_written_per_iteration);
101  AccessWindowHorizontal output_y_access(output_y == nullptr ? nullptr : output_y->info(), 0, num_elems_written_per_iteration);
102 
103  update_window_and_padding(win, input_access, output_x_access, output_y_access);
104 
105  output_x_access.set_valid_region(win, input->info()->valid_region(), border_undefined, border_size());
106  output_y_access.set_valid_region(win, input->info()->valid_region(), border_undefined, border_size());
107 
108  ICLKernel::configure_internal(win);
109 }
110 
111 void CLScharr3x3Kernel::run(const Window &window, cl::CommandQueue &queue)
112 {
115 
117  do
118  {
119  unsigned int idx = 0;
120  add_2D_tensor_argument(idx, _input, slice);
121  add_2D_tensor_argument_if((_run_scharr_x), idx, _output_x, slice);
122  add_2D_tensor_argument_if((_run_scharr_y), idx, _output_y, slice);
123 
124  enqueue(queue, *this, slice, lws_hint());
125  }
126  while(window.slide_window_slice_2D(slice));
127 }
Window first_slice_window_2D() const
First 2D slice of the window.
Definition: Window.h:283
unsigned int top
top of the border
Definition: Types.h:375
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 add_2D_tensor_argument_if(bool cond, 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:159
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
Container for 2D border size.
Definition: Types.h:273
cl::NDRange lws_hint() const
Return the Local-Workgroup-Size hint.
Definition: ICLKernel.h:276
1 channel, 1 U8 per channel
void run(const Window &window, cl::CommandQueue &queue) override
Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue...
#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
static CLKernelLibrary & get()
Access the KernelLibrary singleton.
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
bool update_window_and_padding(Window &win, Ts &&... patterns)
Update window and padding size for each of the access patterns.
Definition: WindowHelpers.h:46
Class to describe a number of elements in each dimension.
Definition: Steps.h:40
Implementation of a row access pattern.
virtual ITensorInfo * info() const =0
Interface to be implemented by the child class to return the tensor&#39;s metadata.
CLScharr3x3Kernel()
Default constructor: initialize all the pointers to nullptr and parameters to zero.
unsigned int left
left of the border
Definition: Types.h:378
#define ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(k)
Definition: Validate.h:941
1 channel, 1 S16 per channel
#define ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(t, c,...)
Definition: Validate.h:790
CLCompileContext class.
BorderSize border_size() const override
The size of the border for that kernel.
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(const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined)
Initialise the kernel&#39;s source, destination and border.
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)