blob: df0f9c4886d3d65ecbee95a9f4b171554557beb4 [file] [log] [blame]
Gian Marco Iodice215b4ea2018-06-28 16:29:29 +01001/*
2 * Copyright (c) 2018 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 "helpers.h"
25
26#if defined(DATA_TYPE) && defined(SRC_WIDTH) && defined(SRC_HEIGHT)
27
28/** This opencl kernel flattens the first 3 dimensions of the input tensor
29 *
30 * @note Datatype should be given as a preprocessor argument using -DDATA_TYPE=type. e.g. -DDATA_TYPE=float
31 * @note The width and height of the input tensor must be passed at compile time using -DSRC_WIDTH and -DSRC_HEIGHT. e.g. -DSRC_WIDTH=24, -DSRC_HEIGHT=24
32 *
33 * @param[in] src_ptr Pointer to the source tensor. Supported data types: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32
34 * @param[in] src_stride_x Stride of the source tensor in X dimension (in bytes)
35 * @param[in] src_step_x src_stride_x * number of elements along X processed per workitem(in bytes)
36 * @param[in] src_stride_y Stride of the source tensor in Y dimension (in bytes)
37 * @param[in] src_step_y src_stride_y * number of elements along Y processed per workitem(in bytes)
38 * @param[in] src_stride_z Stride of the source tensor in Z dimension (in bytes)
39 * @param[in] src_step_z src_stride_z * number of elements along Y processed per workitem(in bytes)
40 * @param[in] src_offset_first_element_in_bytes The offset of the first element in the source tensor
41 * @param[out] dst_ptr Pointer to the destination tensor. Same as @p src_ptr
42 * @param[in] dst_stride_x Stride of the destination tensor in X dimension (in bytes)
43 * @param[in] dst_step_x dst_stride_x * number of elements along X processed per workitem(in bytes)
44 * @param[in] dst_offset_first_element_in_bytes The offset of the first element in the destination tensor
45 */
46__kernel void flatten(
47 TENSOR3D_DECLARATION(src),
48 VECTOR_DECLARATION(dst))
49{
50 Tensor3D src = CONVERT_TO_TENSOR3D_STRUCT(src);
51
52 __global uchar *output_ptr = dst_ptr + dst_offset_first_element_in_bytes + (get_global_id(0) + get_global_id(1) * (int)SRC_WIDTH + get_global_id(2) * (int)(SRC_WIDTH * SRC_HEIGHT)) * sizeof(
53 DATA_TYPE);
54
55 *((__global DATA_TYPE *)output_ptr) = *((__global DATA_TYPE *)src.ptr);
56}
57#endif // defined(DATA_TYPE) && defined(SRC_WIDTH) && defined(SRC_HEIGHT)