blob: 84b82d0ffca72efb4192d927931c332598864939 [file] [log] [blame]
Anthony Barbier6ff3b192017-09-04 18:44:23 +01001/*
2 * Copyright (c) 2016, 2017 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_NEGEMMINTERLEAVE4x4KERNEL_H__
25#define __ARM_COMPUTE_NEGEMMINTERLEAVE4x4KERNEL_H__
26
27#include "arm_compute/core/NEON/INESimpleKernel.h"
28
29namespace arm_compute
30{
31class ITensor;
32
33/** NEON kernel to interleave the elements of a matrix
34 *
35 * This function puts the values in a 4x4 block of Matrix A on the same row (Interleaved values)
36 *
37 * @f[
38 * \left( \begin{array}{cccc}
39 * a00 & a01 & a02 & a03 \\
40 * a10 & a11 & a12 & a13 \\
41 * a20 & a21 & a22 & a23 \\
42 * a30 & a31 & a32 & a33 \\
43 * \end{array} \right)
44 * \rightarrow
45 * \left( \begin{array}{ccccccccccccccccc}
46 * a00 & a10 & a20 & a30 & a01 & a11 & a21 & a31 & a02 & a12 & a22 & a32 & a03 & a13 & a23 & a33 \\
47 * \end{array} \right)
48 * @f]
49 *
50 * After this operation, the output matrix will have the following shape: [ height * 4, ceil(width / 4.0f) ]
51 */
52class NEGEMMInterleave4x4Kernel : public INESimpleKernel
53{
54public:
55 /* Constructor */
56 NEGEMMInterleave4x4Kernel();
57 /** Initialise the kernel's input and output.
58 *
Gian Marco Iodicebdb6b0b2017-06-30 12:21:00 +010059 * @param[in] input Input tensor. Data types supported: U8/S8/QS8/QS16/U16/S16/F16/U32/S32/F32
Anthony Barbier6ff3b192017-09-04 18:44:23 +010060 * @param[out] output Output tensor which stores the interleaved matrix. Data type supported: same as @p input.
61 */
62 void configure(const ITensor *input, ITensor *output);
63
64 // Inherited methods overridden:
65 void run(const Window &window) override;
66
67private:
68 /** Common signature for all the transpose functions
69 *
Gian Marco Iodicebdb6b0b2017-06-30 12:21:00 +010070 * @param[in] input An input tensor. Data types supported: U8/S8/QS8/U16/S16/QS16/F16/U32/S32/F32
Anthony Barbier6ff3b192017-09-04 18:44:23 +010071 * @param[out] output The output tensor. Data type supported: same as @p input
72 * @param[in] window Region on which to execute the kernel.
73 */
74 using GEMMInterleaveFunction = void(const ITensor *input, ITensor *output, const Window &window);
75
76 GEMMInterleaveFunction *_func; /**< GEMM interleave function to use for the particular tensor types passed to configure() */
77};
78}
79#endif /*__ARM_COMPUTE_NEGEMMINTERLEAVE4x4KERNEL_H__*/