Compute Library
 21.02
GCDepthConcatenateLayerKernel.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 
26 #include "arm_compute/core/Error.h"
32 #include "arm_compute/core/Utils.h"
37 
38 #include "support/StringSupport.h"
39 
40 using namespace arm_compute;
41 
43  : _input(nullptr), _output(nullptr), _depth_offset(0)
44 {
45 }
46 void GCDepthConcatenateLayerKernel::configure(const IGCTensor *input, unsigned int depth_offset, IGCTensor *output)
47 {
50 
53  ARM_COMPUTE_ERROR_ON(input->info()->dimension(2) + depth_offset > output->info()->dimension(2));
55 
56  _input = input;
57  _output = output;
58  _depth_offset = depth_offset;
59 
60  // Add build options
61  std::set<std::string> build_opts;
62  std::string dt_name = (input->info()->data_type() == DataType::F32) ? "DATA_TYPE_FP32" : "DATA_TYPE_FP16";
63  build_opts.emplace(("#define " + dt_name));
64  build_opts.emplace("#define LOCAL_SIZE_X " + support::cpp11::to_string(1));
65  build_opts.emplace("#define LOCAL_SIZE_Y " + support::cpp11::to_string(1));
66  build_opts.emplace("#define LOCAL_SIZE_Z " + support::cpp11::to_string(1));
67 
68  // Create kernel
69  _kernel = static_cast<GCKernel>(GCKernelLibrary::get().create_kernel("concatenate_depth", build_opts));
70 
71  unsigned int num_elems_processed_per_iteration = 1;
72  if(input->info()->data_type() == DataType::F16)
73  {
74  num_elems_processed_per_iteration = 4;
75  }
76 
77  // The window needs to be based on input as we copy all the depths of input
78  Window win = calculate_max_window(*output->info(), Steps(num_elems_processed_per_iteration));
79  win.set(Window::DimZ, Window::Dimension(0, input->info()->tensor_shape().z(), 1));
80 
83  update_window_and_padding(win, input_access, output_access);
84  output_access.set_valid_region(win, ValidRegion(Coordinates(), output->info()->tensor_shape()));
85 
86  IGCKernel::configure(win);
87 }
88 
90 {
93 
94  _kernel.use();
95 
96  _output->set_needs_shifting(true);
97 
98  Window slice_in = window.first_slice_window_3D();
99  Window slice_out = window.first_slice_window_3D();
100 
101  slice_out.set(Window::DimZ, Window::Dimension(_depth_offset));
102 
103  do
104  {
105  unsigned int idx = 0;
106  add_3D_tensor_argument(idx, _input, 1, slice_in);
107  add_3D_tensor_argument(idx, _output, 2, slice_out);
108 
109  _kernel.update_shader_params();
110 
111  enqueue(*this, slice_in);
112  }
113  while(window.slide_window_slice_3D(slice_in));
114 }
Window calculate_max_window(const ValidRegion &valid_region, const Steps &steps, bool skip_border, BorderSize border_size)
void run(const Window &window) override
Enqueue the OpenGL ES shader to process the given window.
const Window & window() const
The maximum window the kernel can be executed on.
Definition: IKernel.cpp:28
void add_3D_tensor_argument(unsigned int &idx, const IGCTensor *tensor, const unsigned int binding_point, 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: IGCKernel.cpp:132
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.
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.
1 channel, 1 F32 per channel
#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
Interface for GLES Compute tensor.
Definition: IGCTensor.h:35
Describe one of the image&#39;s dimensions with a start, end and step.
Definition: Window.h:77
GCKernel class.
Copyright (c) 2017-2021 Arm Limited.
1 channel, 1 F16 per channel
static constexpr size_t DimX
Alias for dimension 0 also known as X dimension.
Definition: Window.h:43
bool update_window_and_padding(Window &win, Ts &&... patterns)
Update window and padding size for each of the access patterns.
Definition: WindowHelpers.h:46
virtual const TensorShape & tensor_shape() const =0
Size for each dimension of the tensor.
#define ARM_COMPUTE_ERROR_ON_MISMATCHING_DATA_TYPES(...)
Definition: Validate.h:543
Manages all the GLES kernels compilation and caching, provides accessors for the GLES Context...
Class to describe a number of elements in each dimension.
Definition: Steps.h:40
T z() const
Alias to access the size of the third dimension.
Definition: Dimensions.h:97
void set_needs_shifting(bool needs_shifting)
Set the flag indicating whether or not a tensor needs shifting.
Definition: IGCTensor.cpp:61
Coordinates of an item.
Definition: Coordinates.h:37
Implementation of a row access pattern.
#define ARM_COMPUTE_ERROR_ON_MISMATCHING_SHAPES(...)
Definition: Validate.h:441
virtual ITensorInfo * info() const =0
Interface to be implemented by the child class to return the tensor&#39;s metadata.
void set(size_t dimension, const Dimension &dim)
Set the values of a given dimension.
Definition: Window.inl:49
bool slide_window_slice_3D(Window &slice) const
Slide the passed 3D window slice.
Definition: Window.h:335
#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
static constexpr size_t DimY
Alias for dimension 1 also known as Y dimension.
Definition: Window.h:45
Wrapper to configure the Khronos EGL and OpenGL ES C header.
static GCKernelLibrary & get()
Get the static instance of GCKernelLibrary.
static constexpr size_t DimZ
Alias for dimension 2 also known as Z dimension.
Definition: Window.h:47
GCKernel create_kernel(const std::string &shader_name, const StringSet &build_options_set={}) const
Creates a kernel from the kernel library.
unsigned int num_elems_processed_per_iteration
Container for valid region of a window.
Definition: Types.h:188
Window first_slice_window_3D() const
First 3D slice of the window.
Definition: Window.h:291
void configure(const IGCTensor *input, unsigned int depth_offset, IGCTensor *output)
Initialise the kernel&#39;s inputs and output.
Describe a multidimensional execution window.
Definition: Window.h:39
#define ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(f, s)
Definition: Validate.h:205