Compute Library
 21.02
CLRemapKernel.cpp
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2017-2020 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 
34 
35 #include <algorithm>
36 
37 using namespace arm_compute;
38 
40  : _input(nullptr), _output(nullptr), _map_x(nullptr), _map_y(nullptr)
41 {
42 }
43 
45 {
46  return BorderSize(1);
47 }
48 
49 void CLRemapKernel::configure(const ICLTensor *input, const ICLTensor *map_x, const ICLTensor *map_y, ICLTensor *output, InterpolationPolicy policy, bool border_undefined)
50 {
51  configure(CLKernelLibrary::get().get_compile_context(), input, map_x, map_y, output, policy, border_undefined);
52 }
53 
54 void CLRemapKernel::configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *map_x, const ICLTensor *map_y, ICLTensor *output, InterpolationPolicy policy,
55  bool border_undefined)
56 {
61  ARM_COMPUTE_ERROR_ON_MSG(policy == InterpolationPolicy::AREA, "Area interpolation is not supported!");
62  ARM_COMPUTE_UNUSED(border_undefined);
63 
64  _input = input;
65  _output = output;
66  _map_x = map_x;
67  _map_y = map_y;
68 
69  // Create kernel
70  std::set<std::string> build_opts = { ("-DDATA_TYPE=" + get_cl_type_from_data_type(input->info()->data_type())) };
71  std::string interpolation_name = string_from_interpolation_policy(policy);
72  std::transform(interpolation_name.begin(), interpolation_name.end(), interpolation_name.begin(), ::tolower);
73  std::string kernel_name = "remap_" + interpolation_name;
74  _kernel = create_kernel(compile_context, kernel_name, build_opts);
75 
76  // Configure window
77  constexpr unsigned int num_elems_processed_per_iteration = 4;
78 
79  const int total_right = ceil_to_multiple(input->info()->dimension(0), num_elems_processed_per_iteration);
80  const int access_right = total_right + (((total_right - input->info()->dimension(0)) == 0) ? border_size().right : 0);
81 
82  Window win = calculate_max_window(*_output->info(), Steps(num_elems_processed_per_iteration));
83  AccessWindowStatic input_access(input->info(), -border_size().left, -border_size().top, access_right, input->info()->dimension(1) + border_size().bottom);
84 
86 
87  update_window_and_padding(win, input_access, output_access);
88 
89  output_access.set_valid_region(win, ValidRegion(Coordinates(), output->info()->tensor_shape()));
90 
91  ICLKernel::configure_internal(win);
92 
93  // Set static arguments
94  unsigned int idx = 4 * num_arguments_per_2D_tensor(); //Skip the input and output parameters
95  _kernel.setArg<cl_float>(idx++, input->info()->dimension(0));
96  _kernel.setArg<cl_float>(idx++, input->info()->dimension(1));
97 }
98 
99 void CLRemapKernel::run(const Window &window, cl::CommandQueue &queue)
100 {
103 
105 
106  do
107  {
108  unsigned int idx = 0;
109  add_2D_tensor_argument(idx, _input, slice);
110  add_2D_tensor_argument(idx, _output, slice);
111  add_2D_tensor_argument(idx, _map_x, slice);
112  add_2D_tensor_argument(idx, _map_y, slice);
113  enqueue(queue, *this, slice, lws_hint());
114  }
115  while(window.slide_window_slice_2D(slice));
116 }
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)
InterpolationPolicy
Interpolation method.
Definition: Types.h:392
const Window & window() const
The maximum window the kernel can be executed on.
Definition: IKernel.cpp:28
void configure(const ICLTensor *input, const ICLTensor *map_x, const ICLTensor *map_y, ICLTensor *output, InterpolationPolicy policy, bool border_undefined)
Initialize the kernel&#39;s input, output and border mode.
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.
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
virtual DataType data_type() const =0
Data type used for each element of the tensor.
1 channel, 1 F32 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...
static CLKernelLibrary & get()
Access the KernelLibrary singleton.
unsigned int bottom
bottom of the border
Definition: Types.h:377
bool slide_window_slice_2D(Window &slice) const
Slide the passed 2D window slice.
Definition: Window.h:323
Copyright (c) 2017-2021 Arm Limited.
Implementation of a static rectangular access pattern.
std::string tolower(std::string string)
Convert string to lower case.
Definition: Utility.h:203
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
#define ARM_COMPUTE_UNUSED(...)
To avoid unused variables warnings.
Definition: Error.h:152
virtual const TensorShape & tensor_shape() const =0
Size for each dimension of the tensor.
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
#define ARM_COMPUTE_ERROR_ON_MSG(cond, msg)
Definition: Error.h:456
Coordinates of an item.
Definition: Coordinates.h:37
Implementation of a row access pattern.
std::string kernel_name
const std::string & string_from_interpolation_policy(InterpolationPolicy policy)
Translates a given interpolation policy to a string.
Definition: Utils.cpp:212
std::string get_cl_type_from_data_type(const DataType &dt)
Translates a tensor data type to the appropriate OpenCL type.
Definition: CLHelpers.cpp:37
virtual ITensorInfo * info() const =0
Interface to be implemented by the child class to return the tensor&#39;s metadata.
static constexpr unsigned int num_arguments_per_2D_tensor()
Returns the number of arguments enqueued per 2D tensor object.
Definition: ICLKernel.h:206
unsigned int left
left of the border
Definition: Types.h:378
unsigned int right
right of the border
Definition: Types.h:376
#define ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(k)
Definition: Validate.h:941
Output values are determined by averaging the source pixels whose areas fall under the area of the de...
#define ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(t, c,...)
Definition: Validate.h:790
CLRemapKernel()
Default constructor.
CLCompileContext class.
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...
unsigned int num_elems_processed_per_iteration
Container for valid region of a window.
Definition: Types.h:188
BorderSize border_size() const override
The size of the border for that kernel.
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)