Compute Library
 23.08
CLStackLayer.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  */
24 #include <complex>
25 
27 
29 #include "arm_compute/core/Error.h"
32 #include "arm_compute/core/Types.h"
36 
37 #include "src/common/utils/Log.h"
38 
39 namespace arm_compute
40 {
42  : _input(),
43  _stack_kernels(),
44  _num_inputs(0)
45 {
46 }
47 
48 CLStackLayer::~CLStackLayer() = default;
49 
50 void CLStackLayer::configure(const std::vector<ICLTensor *> &input, int axis, ICLTensor *output)
51 {
52  configure(CLKernelLibrary::get().get_compile_context(), input, axis, output);
53 }
54 
55 void CLStackLayer::configure(const CLCompileContext &compile_context, const std::vector<ICLTensor *> &input, int axis, ICLTensor *output)
56 {
57  ARM_COMPUTE_LOG_PARAMS(input, axis, output);
58  _num_inputs = input.size();
59  _stack_kernels.reserve(_num_inputs);
60 
61  // Wrap around negative values
62  const unsigned int axis_u = wrap_around(axis, static_cast<int>(input[0]->info()->num_dimensions() + 1));
63 
64  for(unsigned int i = 0; i < _num_inputs; i++)
65  {
66  _stack_kernels.emplace_back(std::make_unique<CLStackLayerKernel>());
67  _stack_kernels.back()->configure(compile_context, input[i], axis_u, i, _num_inputs, output);
68  }
69 }
70 
71 Status CLStackLayer::validate(const std::vector<ITensorInfo *> &input, int axis, const ITensorInfo *output)
72 {
75 
76  // Wrap around negative values
77  const size_t rank = input[0]->num_dimensions();
78  const unsigned int axis_u = wrap_around(axis, static_cast<int>(rank + 1));
79 
80  const unsigned int num_inputs = input.size();
81 
82  for(unsigned int i = 0; i < num_inputs; i++)
83  {
84  // All the tensors must have the same rank
85  ARM_COMPUTE_RETURN_ERROR_ON(input[i]->num_dimensions() != rank);
86  // Validate Kernel
87  ARM_COMPUTE_RETURN_ON_ERROR(CLStackLayerKernel::validate(input[i], axis_u, i, num_inputs, output));
88  }
89 
90  return Status{};
91 }
92 
94 {
95  for(unsigned i = 0; i < _num_inputs; i++)
96  {
97  CLScheduler::get().enqueue(*_stack_kernels[i], false);
98  }
99 }
100 } // namespace arm_compute
ICLTensor.h
Helpers.h
arm_compute::CLStackLayer::~CLStackLayer
~CLStackLayer()
Default destructor.
arm_compute::CLScheduler::enqueue
void enqueue(ICLKernel &kernel, bool flush=true)
Schedule the execution of the passed kernel if possible.
Definition: CLScheduler.cpp:205
CLStackLayerKernel.h
Types.h
arm_compute::ICLTensor
Interface for OpenCL tensor.
Definition: ICLTensor.h:42
TensorInfo.h
arm_compute::wrap_around
T wrap_around(T x, T m)
Wrap-around a number within the range 0 <= x < m.
Definition: Helpers.h:268
Error.h
arm_compute::CLKernelLibrary::get
static CLKernelLibrary & get()
Access the KernelLibrary singleton.
Definition: CLKernelLibrary.cpp:39
CLStackLayer.h
ARM_COMPUTE_RETURN_ON_ERROR
#define ARM_COMPUTE_RETURN_ON_ERROR(status)
Checks if a status contains an error and returns it.
Definition: Error.h:204
arm_compute::CLStackLayer::configure
void configure(const std::vector< ICLTensor * > &input, int axis, ICLTensor *output)
Initialise the kernel's inputs vector and output.
Definition: CLStackLayer.cpp:50
arm_compute::CLCompileContext
CLCompileContext class.
Definition: CLCompileContext.h:204
ARM_COMPUTE_RETURN_ERROR_ON
#define ARM_COMPUTE_RETURN_ERROR_ON(cond)
If the condition is true, an error is returned.
Definition: Error.h:297
arm_compute::Status
Status class.
Definition: Error.h:52
arm_compute::CLStackLayer::CLStackLayer
CLStackLayer()
Default constructor.
Definition: CLStackLayer.cpp:41
CLScheduler.h
Interface to enqueue OpenCL kernels and get/set the OpenCL CommandQueue and ICLTuner.
arm_compute::CLScheduler::get
static CLScheduler & get()
Access the scheduler singleton.
Definition: CLScheduler.cpp:103
arm_compute::CLStackLayer::run
void run() override
Run the kernels contained in the function.
Definition: CLStackLayer.cpp:93
ShapeCalculator.h
arm_compute::CLStackLayer::validate
static Status validate(const std::vector< ITensorInfo * > &input, int axis, const ITensorInfo *output)
Static function to check if given info will lead to a valid configuration of CLStackLayerKernel.
Definition: CLStackLayer.cpp:71
arm_compute
Copyright (c) 2017-2023 Arm Limited.
Definition: introduction.dox:24
Log.h
ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR
#define ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(...)
Definition: Validate.h:163
arm_compute::ITensorInfo
Store the tensor's metadata.
Definition: ITensorInfo.h:43
arm_compute::test::validation::info
ScaleKernelInfo info(interpolation_policy, default_border_mode, PixelValue(), sampling_policy, false)
arm_compute::CLStackLayerKernel::validate
static Status validate(const ITensorInfo *input, unsigned int axis, unsigned int idx_input, unsigned int num_tensors, const ITensorInfo *output)
Static function to check if given info will lead to a valid configuration of CLStackLayerKernel.
Definition: CLStackLayerKernel.cpp:115
ARM_COMPUTE_LOG_PARAMS
#define ARM_COMPUTE_LOG_PARAMS(...)
Definition: Log.h:35
arm_compute::test::validation::input
auto input
Definition: LSTMLayerQuantized.cpp:486