xref: /aosp_15_r20/external/ComputeLibrary/src/gpu/cl/kernels/ClWidthConcatenate2TensorsKernel.h (revision c217d954acce2dbc11938adb493fc0abd69584f3)
1 /*
2  * Copyright (c) 2021-2022 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 #ifndef ARM_COMPUTE_CL_WIDTHCONCATENATE_2TENSORS_KERNEL_H
25 #define ARM_COMPUTE_CL_WIDTHCONCATENATE_2TENSORS_KERNEL_H
26 
27 #include "src/core/common/Macros.h"
28 #include "src/gpu/cl/ClCompileContext.h"
29 #include "src/gpu/cl/IClKernel.h"
30 
31 namespace arm_compute
32 {
33 namespace opencl
34 {
35 namespace kernels
36 {
37 /** Interface for the width concatenate kernel of 2 tensors.
38  *  The src1 and src2 tensors will be concatenated into the dst tensor.
39  */
40 class ClWidthConcatenate2TensorsKernel : public IClKernel
41 {
42 public:
43     ClWidthConcatenate2TensorsKernel();
44     ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClWidthConcatenate2TensorsKernel);
45     /** Initialise the kernel's sources and destination
46      *
47      * @param[in]  compile_context The compile context to be used.
48      * @param[in]  src1            First source tensor info. Data types supported: All.
49      * @param[in]  src2            Second source tensor info. Data types supported: same as @p src1
50      * @param[out] dst             Destination tensor info. Data types supported: Same as @p src1.
51      */
52     void configure(const CLCompileContext &compile_context, ITensorInfo *src1, ITensorInfo *src2, ITensorInfo *dst);
53     /** Static function to check if given info will lead to a valid configuration
54      *
55      * Similar to @ref ClWidthConcatenate2TensorsKernel::configure()
56      *
57      * @return a status
58      */
59     static Status validate(const ITensorInfo *src1, const ITensorInfo *src2, const ITensorInfo *dst);
60 
61     // Inherited methods overridden:
62     void run_op(ITensorPack &tensors, const Window &window, ::cl::CommandQueue &queue) override;
63 
64 private:
65     int32_t _depth{ 0 };
66     int32_t _input1_width{ 0 };
67 };
68 } // namespace kernels
69 } // namespace opencl
70 } // namespace arm_compute
71 #endif /* ARM_COMPUTE_CL_WIDTH_CONCATENATE_2TENSORS_KERNEL_H */
72