aboutsummaryrefslogtreecommitdiff
path: root/arm_compute/runtime/CL/functions/CLElementwiseOperations.h
blob: 4a0911ec4e4bd1e0beafd19f3ad6520c8a3a428c (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
/*
 * Copyright (c) 2018 ARM Limited.
 *
 * SPDX-License-Identifier: MIT
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to
 * deal in the Software without restriction, including without limitation the
 * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
 * sell copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in all
 * copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARI SING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 */
#ifndef __ARM_COMPUTE_CLELEMENTWISEOPERATIONS_H__
#define __ARM_COMPUTE_CLELEMENTWISEOPERATIONS_H__

#include "arm_compute/core/Types.h"
#include "arm_compute/runtime/CL/ICLSimpleFunction.h"

namespace arm_compute
{
class ICLTensor;

/** Basic function to run @ref CLSaturatedArithmeticOperationKernel for addition
 *
 * @note The tensor data type for the inputs must be U8/QASYMM8/S16/S32/U32/F16/F32.
 * @note The function performs an arithmetic addition between two tensors.
 */
class CLArithmeticAddition : public ICLSimpleFunction
{
public:
    /** Initialise the kernel's inputs, output and conversion policy.
     *
     * @param[in, out] input1 First tensor input. Data types supported: U8/QASYMM8/S16/S32/U32/F16/F32.
     *                        The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
     * @param[in, out] input2 Second tensor input. Data types supported: U8, QASYMM8 (only if @p input1 is QASYMM8), S16/F16/F32.
     *                        The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
     * @param[out]     output Output tensor. Data types supported: U8 (Only if both inputs are U8), QASYMM8 (only if both inputs are QASYMM8), S16/F16/F32.
     * @param[in]      policy Policy to use to handle overflow.
     */
    void configure(ICLTensor *input1, ICLTensor *input2, ICLTensor *output, ConvertPolicy policy);
    /** Static function to check if given info will lead to a valid configuration of @ref CLSaturatedArithmeticOperationKernel for addition
     *
     * @param[in] input1 First tensor input info. Data types supported: U8/QASYMM8/S16/S32/U32/F16/F32.
     * @param[in] input2 Second tensor input info. Data types supported: U8, QASYMM8 (only if @p input1 is QASYMM8), S16/F16/F32.
     * @param[in] output Output tensor info. Data types supported: U8 (Only if both inputs are U8), QASYMM8 ( only if both inputs are QASYMM8), S16/F16/F32.
     * @param[in] policy Policy to use to handle overflow.
     *
     * @return a status
     */
    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, ConvertPolicy policy);
};

/** Basic function to run @ref CLSaturatedArithmeticOperationKernel for subtraction
 *
 * @note The tensor data type for the inputs must be U8/QASYMM8/S16/S32/U32/F16/F32.
 * @note The function performs an arithmetic subtraction between two tensors.
 */
class CLArithmeticSubtraction : public ICLSimpleFunction
{
public:
    /** Initialise the kernel's inputs, output and conversion policy.
     *
     * @param[in, out] input1 First tensor input. Data types supported: U8/QASYMM8/S16/S32/U32/F16/F32.
     *                        The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
     * @param[in, out] input2 Second tensor input. Data types supported: U8, QASYMM8 (only if @p input1 is QASYMM8), S16/F16/F32.
     *                        The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
     * @param[out]     output Output tensor. Data types supported: U8 (Only if both inputs are U8), QASYMM8 (only if both inputs are QASYMM8), S16/F16/F32.
     * @param[in]      policy Policy to use to handle overflow.
     */
    void configure(ICLTensor *input1, ICLTensor *input2, ICLTensor *output, ConvertPolicy policy);
    /** Static function to check if given info will lead to a valid configuration of @ref CLSaturatedArithmeticOperationKernel for subtraction
     *
     * @param[in] input1 First tensor input info. Data types supported: U8/QASYMM8/S16/S32/U32/F16/F32.
     * @param[in] input2 Second tensor input info. Data types supported: U8, QASYMM8 (only if @p input1 is QASYMM8), S16/F16/F32.
     * @param[in] output Output tensor info. Data types supported: U8 (Only if both inputs are U8), QASYMM8 ( only if both inputs are QASYMM8), S16/F16/F32.
     * @param[in] policy Policy to use to handle overflow.
     *
     * @return a status
     */
    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, ConvertPolicy policy);
};

/** Basic function to run @ref CLSaturatedArithmeticOperationKernel for division
 *
 * @note The tensor data type for the inputs must be F16/F32.
 * @note The function performs an arithmetic division between two tensors.
 */
class CLArithmeticDivision : public ICLSimpleFunction
{
public:
    /** Initialise the kernel's inputs, output.
     *
     * @param[in, out] input1 First tensor input. Data types supported: F16/F32.
     *                        The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
     * @param[in, out] input2 Second tensor input. Same as @p input1.
     *                        The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
     * @param[out]     output Output tensor. Data types supported: Same as @p input1.
     */
    void configure(ICLTensor *input1, ICLTensor *input2, ICLTensor *output);
    /** Static function to check if given info will lead to a valid configuration of @ref CLArithmeticDivision
     *
     * @param[in] input1 First tensor input info. Data types supported: F16/F32.
     * @param[in] input2 Second tensor input info. Data types supported: Same as @p input1.
     * @param[in] output Output tensor info. Data types supported: Same as @p input1.
     *
     * @return a status
     */
    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output);
};

/** Basic function to run @ref CLArithmeticOperationKernel for max
 *
 * @note The tensor data type for the inputs must be U8/QASYMM8/S16/S32/U32/F16/F32.
 * @note The function performs a max operation between two tensors.
 */
class CLElementwiseMax : public ICLSimpleFunction
{
public:
    /** Initialise the kernel's inputs, output and conversion policy.
     *
     * @param[in, out] input1 First tensor input. Data types supported: U8/QASYMM8/S16/S32/U32/F16/F32.
     *                        The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
     * @param[in, out] input2 Second tensor input. Data types supported: U8, QASYMM8 (only if @p input1 is QASYMM8), S16/F16/F32.
     *                        The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
     * @param[out]     output Output tensor. Data types supported: U8 (Only if both inputs are U8), QASYMM8 (only if both inputs are QASYMM8), S16/F16/F32.
     */
    void configure(ICLTensor *input1, ICLTensor *input2, ICLTensor *output);
    /** Static function to check if given info will lead to a valid configuration of @ref CLArithmeticOperationKernel for max
     *
     * @param[in] input1 First tensor input info. Data types supported: U8/QASYMM8/S16/S32/U32/F16/F32.
     * @param[in] input2 Second tensor input info. Data types supported: U8, QASYMM8 (only if @p input1 is QASYMM8), S16/F16/F32.
     * @param[in] output Output tensor info. Data types supported: U8 (Only if both inputs are U8), QASYMM8 ( only if both inputs are QASYMM8), S16/F16/F32.
     *
     * @return a status
     */
    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output);
};

/** Basic function to run @ref CLArithmeticOperationKernel for min
 *
 * @note The tensor data type for the inputs must be U8/QASYMM8/S16/S32/U32/F16/F32.
 * @note The function performs a max operation between two tensors.
 */
class CLElementwiseMin : public ICLSimpleFunction
{
public:
    /** Initialise the kernel's inputs, output and conversion policy.
     *
     * @param[in, out] input1 First tensor input. Data types supported: U8/QASYMM8/S16/S32/U32/F16/F32.
     *                        The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
     * @param[in, out] input2 Second tensor input. Data types supported: U8, QASYMM8 (only if @p input1 is QASYMM8), S16/F16/F32.
     *                        The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
     * @param[out]     output Output tensor. Data types supported: U8 (Only if both inputs are U8), QASYMM8 (only if both inputs are QASYMM8), S16/F16/F32.
     */
    void configure(ICLTensor *input1, ICLTensor *input2, ICLTensor *output);
    /** Static function to check if given info will lead to a valid configuration of @ref CLArithmeticOperationKernel for min
     *
     * @param[in] input1 First tensor input info. Data types supported: U8/QASYMM8/S16/S32/U32/F16/F32.
     * @param[in] input2 Second tensor input info. Data types supported: U8, QASYMM8 (only if @p input1 is QASYMM8), S16/F16/F32.
     * @param[in] output Output tensor info. Data types supported: U8 (Only if both inputs are U8), QASYMM8 ( only if both inputs are QASYMM8), S16/F16/F32.
     *
     * @return a status
     */
    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output);
};

/** Basic function to run @ref CLArithmeticOperationKernel for squared difference
 *
 * @note The tensor data type for the inputs must be QASYMM8/U8/S16/F16/F32.
 * @note The function performs a squared different operation between two tensors (i.e., out[i] = (in1[i] - in2[i])^2
 */
class CLElementwiseSquaredDiff : public ICLSimpleFunction
{
public:
    /** Initialise the kernel's inputs, output and conversion policy.
     *
     * @param[in, out] input1 First tensor input. Data types supported: U8/QASYMM8/S16/F16/F32.
     *                        The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
     * @param[in, out] input2 Second tensor input. Data types supported: U8, QASYMM8 (only if @p input1 is QASYMM8), S16/F16/F32.
     *                        The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0.
     * @param[out]     output Output tensor. Data types supported: U8 (Only if both inputs are U8), QASYMM8 (only if both inputs are QASYMM8), S16/F16/F32.
     */
    void configure(ICLTensor *input1, ICLTensor *input2, ICLTensor *output);
    /** Static function to check if given info will lead to a valid configuration of @ref CLArithmeticOperationKernel for squared difference
     *
     * @param[in] input1 First tensor input info. Data types supported: U8/QASYMM8/S16/F16/F32.
     * @param[in] input2 Second tensor input info. Data types supported: U8, QASYMM8 (only if @p input1 is QASYMM8), S16/F16/F32.
     * @param[in] output Output tensor info. Data types supported: U8 (Only if both inputs are U8), QASYMM8 ( only if both inputs are QASYMM8), S16/F16/F32.
     *
     * @return a status
     */
    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output);
};
} // namespace arm_compute
#endif /* __ARM_COMPUTE_CLELEMENTWISEOPERATIONS_H__ */