Compute Library
 21.02
ClCopyKernel.cpp
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2018-2021 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 
31 #include "arm_compute/core/Utils.h"
33 #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, const ITensorInfo *dst, Window *dst_window = nullptr)
48 {
50 
51  // Validate dst if initialized
52  if(dst->total_size() != 0)
53  {
56  if(dst_window == nullptr)
57  {
58  ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(src->tensor_shape(), dst->tensor_shape());
59  }
60  else
61  {
62  ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(src->tensor_shape(), dst_window->shape());
63  }
64  }
65 
66  return Status{};
67 }
68 
69 } // namespace
70 
71 void ClCopyKernel::configure(const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst, Window *dst_window)
72 {
74  ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(src, dst, dst_window));
75 
76  auto padding_info = get_padding_info({ src, dst });
77 
78  // Create kernel
79  CLBuildOptions build_opts;
80  build_opts.add_option("-DDATA_TYPE=" + get_cl_type_from_data_type(src->data_type()));
81 
82  // Output auto inizialitation if not yet initialized
83  auto_init_if_empty(*dst, *src);
84 
85  // Configure window
86  const unsigned int vec_size_x = adjust_vec_size(16 / src->element_size(), src->dimension(0));
87 
88  const Window win_config = calculate_max_window(*src, Steps(vec_size_x));
89 
90  if(dst_window != nullptr)
91  {
92  _has_dst_window = true;
93  _dst_window = Window(*dst_window);
94  const int width_x = dst_window->num_iterations(0);
95  const int vec_size_x_leftover = width_x % vec_size_x;
96  const bool multi_access_x = width_x >= static_cast<int32_t>(vec_size_x);
97 
98  if(multi_access_x)
99  {
100  _dst_window.set(Window::DimX, Window::Dimension(dst_window->x().start(), ceil_to_multiple(dst_window->x().end(), vec_size_x), vec_size_x));
101  }
102 
103  build_opts.add_option("-DVEC_SIZE_LEFTOVER=" + support::cpp11::to_string(vec_size_x_leftover));
104  }
105  else
106  {
107  const int width_x = src->tensor_shape().x();
108  const int vec_size_x_leftover = width_x % vec_size_x;
109 
110  build_opts.add_option("-DVEC_SIZE_LEFTOVER=" + support::cpp11::to_string(vec_size_x_leftover));
111  }
112 
113  build_opts.add_option("-DVEC_SIZE=" + support::cpp11::to_string(vec_size_x));
114 
115  // Build kernel
116  _kernel = create_kernel(compile_context, "copy_tensor", build_opts.options());
117 
118  // Validate and set the window
119  ICLKernel::configure_internal(win_config);
120 
122 }
123 
125 {
126  ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(src, dst, dst_window));
127 
128  return Status{};
129 }
130 
131 void ClCopyKernel::run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue)
132 {
135 
136  const auto src = utils::cast::polymorphic_downcast<const ICLTensor *>(tensors.get_const_tensor(TensorType::ACL_SRC));
137  auto dst = utils::cast::polymorphic_downcast<ICLTensor *>(tensors.get_tensor(TensorType::ACL_DST));
138 
139  Window slice;
140 
141  if(_has_dst_window)
142  {
143  slice = window.first_slice_window_3D();
144  Window out_slice = _dst_window.first_slice_window_3D();
145  do
146  {
147  unsigned int idx = 0;
148  add_3D_tensor_argument(idx, src, slice);
149  add_3D_tensor_argument(idx, dst, out_slice);
150  enqueue(queue, *this, slice, lws_hint());
151  }
152  while(window.slide_window_slice_3D(slice) && _dst_window.slide_window_slice_3D(out_slice));
153  }
154  else
155  {
157  slice = collapsed.first_slice_window_3D();
158  do
159  {
160  unsigned int idx = 0;
161  add_3D_tensor_argument(idx, src, slice);
162  add_3D_tensor_argument(idx, dst, slice);
163  enqueue(queue, *this, slice, lws_hint());
164  }
165  while(collapsed.slide_window_slice_3D(slice));
166  }
167 }
168 } // namespace kernels
169 } // namespace opencl
170 } // namespace arm_compute
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 run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override
Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue...
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.
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO(...)
Definition: Validate.h:610
cl::NDRange lws_hint() const
Return the Local-Workgroup-Size hint.
Definition: ICLKernel.h:276
#define ARM_COMPUTE_RETURN_ON_ERROR(status)
Checks if a status contains an error and returns it.
Definition: Error.h:204
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.
void configure(const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst, Window *dst_window=nullptr)
Initialize the kernel&#39;s src, dst.
#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
Store the tensor&#39;s metadata.
Definition: ITensorInfo.h:40
#define ARM_COMPUTE_ERROR_THROW_ON(status)
Definition: Error.h:455
Describe one of the image&#39;s dimensions with a start, end and step.
Definition: Window.h:77
Status class.
Definition: Error.h:52
constexpr size_t num_iterations(size_t dimension) const
Return the number of iterations needed to iterate through a given dimension.
Definition: Window.inl:182
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(...)
Definition: Validate.h:288
void add_3D_tensor_argument(unsigned int &idx, const ICLTensor *tensor, 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: ICLKernel.h:172
SimpleTensor< float > src
Definition: DFT.cpp:155
Copyright (c) 2017-2021 Arm Limited.
#define ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(...)
Definition: Validate.h:163
void add_option(std::string option)
Adds option to the existing build option list.
T x() const
Alias to access the size of the first dimension.
Definition: Dimensions.h:87
const ITensor * get_const_tensor(int id) const
Get constant tensor of a given id.
Definition: ITensorPack.cpp:40
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
static constexpr size_t DimX
Alias for dimension 0 also known as X dimension.
Definition: Window.h:43
Window collapse_if_possible(const Window &full_window, size_t first, size_t last, bool *has_collapsed=nullptr) const
Collapse the dimensions between first and last if possible.
Definition: Window.inl:68
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
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
bool auto_init_if_empty(ITensorInfo &info, const TensorShape &shape, int num_channels, DataType data_type, QuantizationInfo quantization_info=QuantizationInfo())
Auto initialize the tensor info (shape, number of channels and data type) if the current assignment i...
virtual size_t element_size() const =0
Element size in bytes calculated as data_size() * num_channels()
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
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:528
CLCompileContext class.
ITensor * get_tensor(int id)
Get tensor of a given id from the pac.
Definition: ITensorPack.cpp:50
static Status validate(const ITensorInfo *src, const ITensorInfo *dst, Window *dst_window=nullptr)
Static function to check if given info will lead to a valid configuration of ClCopyKernel.
static constexpr size_t DimZ
Alias for dimension 2 also known as Z dimension.
Definition: Window.h:47
Manages all the OpenCL kernels compilation and caching, provides accessors for the OpenCL Context...
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(...)
Definition: Validate.h:545
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:513
Status validate_arguments(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const GEMMLowpOutputStageInfo *output_stage)
Tensor packing service.
Definition: ITensorPack.h:37
#define ARM_COMPUTE_ERROR_ON_NULLPTR(...)
Definition: Validate.h:161
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&#39;s first dimension, getting rounded down to its closest valid vector size.
Definition: Utils.h:1358
constexpr int end() const
Return the end of the dimension.
Definition: Window.h:99
Window first_slice_window_3D() const
First 3D slice of the window.
Definition: Window.h:291
constexpr int start() const
Return the start of the dimension.
Definition: Window.h:94
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)
constexpr const Dimension & x() const
Alias to access the first dimension of the window.
Definition: Window.h:145