xref: /aosp_15_r20/external/ComputeLibrary/src/gpu/cl/kernels/ClBatchConcatenateKernel.h (revision c217d954acce2dbc11938adb493fc0abd69584f3)
1 /*
2  * Copyright (c) 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 #ifndef ARM_COMPUTE_CL_BATCH_CONCATENATE_KERNEL_H
25 #define ARM_COMPUTE_CL_BATCH_CONCATENATE_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 batch concatenate kernel.
38  *  The src tensor will be concatenated into the destination tensor.
39  */
40 class ClBatchConcatenateKernel : public IClKernel
41 {
42 public:
43     ClBatchConcatenateKernel();
44     ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClBatchConcatenateKernel);
45     /** Initialise the kernel's source and destination
46      *
47      * @param[in]     compile_context The compile context to be used.
48      * @param[in]     src             Source tensor info. Data types supported: All.
49      * @param[in]     batch_offset    The offset on axis # 3.
50      * @param[in,out] dst             Destination tensor info. Data types supported: Same as @p src.
51      *
52      * @note: The dst tensor's low two dimensions can't be smaller than the src one's.
53      * @note: The gaps between the two lowest dimensions of src and dst need to be divisible by 2.
54      *
55      */
56     void configure(const CLCompileContext &compile_context, ITensorInfo *src, unsigned int batch_offset, ITensorInfo *dst);
57     /** Static function to check if given info will lead to a valid configuration
58      *
59      * Similar to @ref ClBatchConcatenateKernel::configure()
60      *
61      * @return a status
62      */
63     static Status validate(const ITensorInfo *src, unsigned int batch_offset, const ITensorInfo *dst);
64 
65     // Inherited methods overridden:
66     void run_op(ITensorPack &tensors, const Window &window, ::cl::CommandQueue &queue) override;
67 
68 private:
69     unsigned int _batch_offset{ 0 };
70 };
71 } // namespace kernels
72 } // namespace opencl
73 } // namespace arm_compute
74 #endif /* ARM_COMPUTE_CL_BATCH_CONCATENATE_KERNEL_H */
75