xref: /aosp_15_r20/external/ComputeLibrary/src/gpu/cl/kernels/ClWidthConcatenate4TensorsKernel.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 
25 #ifndef ARM_COMPUTE_CL_WIDTH_CONCATENATE_4TENSORS_KERNEL_H
26 #define ARM_COMPUTE_CL_WIDTH_CONCATENATE_4TENSORS_KERNEL_H
27 
28 #include "src/core/common/Macros.h"
29 #include "src/gpu/cl/ClCompileContext.h"
30 #include "src/gpu/cl/IClKernel.h"
31 
32 namespace arm_compute
33 {
34 namespace opencl
35 {
36 namespace kernels
37 {
38 /** Interface for the width concatenate kernel of 4 tensors.
39  *  All source tensors will be concatenated into the destination tensor.
40  */
41 class ClWidthConcatenate4TensorsKernel : public IClKernel
42 {
43 public:
44     ClWidthConcatenate4TensorsKernel();
45     ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClWidthConcatenate4TensorsKernel);
46     /** Initialise the kernel's sources and destination
47      *
48      * @param[in]  compile_context The compile context to be used.
49      * @param[in]  src1            First source tensor info. Data types supported: All.
50      * @param[in]  src2            Second source tensor info. Data types supported: same as @p src1
51      * @param[in]  src3            Third source tensor info. Data types supported: same as @p src1
52      * @param[in]  src4            Fourth source tensor info. Data types supported: same as @p src1
53      * @param[out] dst             Destination tensor info. Data types supported: same as @p src1.
54      */
55     void configure(const CLCompileContext &compile_context, ITensorInfo *src1, ITensorInfo *src2, ITensorInfo *src3, ITensorInfo *src4, ITensorInfo *dst);
56     /** Static function to check if given info will lead to a valid configuration
57      *
58      * Similar to @ref ClWidthConcatenate4TensorsKernel::configure()
59      *
60      * @return a status
61      */
62     static Status validate(const ITensorInfo *src1, const ITensorInfo *src2, const ITensorInfo *src3, const ITensorInfo *src4, const ITensorInfo *dst);
63 
64     // Inherited methods overridden:
65     void run_op(ITensorPack &tensors, const Window &window, ::cl::CommandQueue &queue) override;
66 
67 private:
68     int32_t _depth{ 0 };
69     int32_t _input1_width{ 0 };
70     int32_t _input2_width{ 0 };
71     int32_t _input3_width{ 0 };
72 };
73 } // namespace kernels
74 } // namespace opencl
75 } // namespace arm_compute
76 #endif /* ARM_COMPUTE_CL_WIDTH_CONCATENATE_4TENSORS_KERNEL_H */
77