xref: /aosp_15_r20/external/ComputeLibrary/src/runtime/CL/functions/CLConcatenateLayer.cpp (revision c217d954acce2dbc11938adb493fc0abd69584f3)
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 "arm_compute/runtime/CL/functions/CLConcatenateLayer.h"
25 
26 #include "arm_compute/core/CL/ICLTensor.h"
27 #include "src/core/CL/ICLKernel.h"
28 #include "src/gpu/cl/operators/ClConcatenate.h"
29 
30 #include "src/common/utils/Log.h"
31 
32 namespace arm_compute
33 {
34 struct CLConcatenateLayer::Impl
35 {
36     std::vector<const ICLTensor *>         srcs{};
37     ICLTensor                             *dst{ nullptr };
38     unsigned int                           num_inputs{ 0 };
39     unsigned int                           axis{ 0 };
40     std::unique_ptr<opencl::ClConcatenate> op{ nullptr };
41 };
42 
CLConcatenateLayer()43 CLConcatenateLayer::CLConcatenateLayer()
44     : _impl(std::make_unique<Impl>())
45 {
46 }
47 
48 CLConcatenateLayer::CLConcatenateLayer(CLConcatenateLayer &&) = default;
49 
50 CLConcatenateLayer &CLConcatenateLayer::operator=(CLConcatenateLayer &&) = default;
51 
52 CLConcatenateLayer::~CLConcatenateLayer() = default;
53 
configure(std::vector<const ICLTensor * > & inputs_vector,ICLTensor * output,size_t axis)54 void CLConcatenateLayer::configure(std::vector<const ICLTensor *> &inputs_vector, ICLTensor *output, size_t axis)
55 {
56     configure(CLKernelLibrary::get().get_compile_context(), inputs_vector, output, axis);
57 }
58 
configure(const CLCompileContext & compile_context,std::vector<const ICLTensor * > & inputs_vector,ICLTensor * output,size_t axis)59 void CLConcatenateLayer::configure(const CLCompileContext &compile_context, std::vector<const ICLTensor *> &inputs_vector, ICLTensor *output, size_t axis)
60 {
61     ARM_COMPUTE_ERROR_ON(output == nullptr);
62     ARM_COMPUTE_LOG_PARAMS(inputs_vector, output, axis);
63 
64     _impl->srcs       = inputs_vector;
65     _impl->dst        = output;
66     _impl->axis       = axis;
67     _impl->num_inputs = inputs_vector.size();
68     _impl->op         = std::make_unique<opencl::ClConcatenate>();
69 
70     std::vector<ITensorInfo *> inputs_vector_info;
71     for(unsigned int i = 0; i < inputs_vector.size(); ++i)
72     {
73         ARM_COMPUTE_ERROR_ON_NULLPTR(inputs_vector.at(i));
74         inputs_vector_info.emplace_back(inputs_vector.at(i)->info());
75     }
76     _impl->op->configure(compile_context, inputs_vector_info, _impl->dst->info(), axis);
77 }
78 
validate(const std::vector<const ITensorInfo * > & inputs_vector,const ITensorInfo * output,size_t axis)79 Status CLConcatenateLayer::validate(const std::vector<const ITensorInfo *> &inputs_vector, const ITensorInfo *output, size_t axis)
80 {
81     return opencl::ClConcatenate::validate(inputs_vector, output, axis);
82 }
83 
run()84 void CLConcatenateLayer::run()
85 {
86     ITensorPack pack;
87     for(unsigned i = 0; i < _impl->num_inputs; ++i)
88     {
89         pack.add_tensor(TensorType::ACL_SRC_VEC + i, _impl->srcs.at(i));
90     }
91     pack.add_tensor(TensorType::ACL_DST, _impl->dst);
92 
93     _impl->op->run(pack);
94 }
95 } // namespace arm_compute
96