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
|
/*
* Copyright (c) 2017-2020 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, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#ifndef ARM_COMPUTE_CLGEMMLOWREDUCTIONKERNEL_H
#define ARM_COMPUTE_CLGEMMLOWREDUCTIONKERNEL_H
#include "arm_compute/core/CL/ICLKernel.h"
namespace arm_compute
{
class ICLTensor;
struct GEMMLowpReductionKernelInfo;
/** Common interface for all OpenCL reduction kernels */
class ICLGEMMLowpReductionKernel : public ICLKernel
{
public:
/** Constructor */
ICLGEMMLowpReductionKernel();
/** Prevent instances of this class from being copied (As this class contains pointers)*/
ICLGEMMLowpReductionKernel(const ICLGEMMLowpReductionKernel &) = delete;
/** Prevent instances of this class from being copied (As this class contains pointers)*/
ICLGEMMLowpReductionKernel &operator=(const ICLGEMMLowpReductionKernel &) = delete;
/** Allow instances of this class to be moved */
ICLGEMMLowpReductionKernel(ICLGEMMLowpReductionKernel &&) = default;
/** Allow instances of this class to be moved */
ICLGEMMLowpReductionKernel &operator=(ICLGEMMLowpReductionKernel &&) = default;
/** Initialise the kernel's input and output.
*
* @param[in] input Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8.
* @param[out] output Output row-vector of sums of all the entries in each row/col of input tensor. Data type supported: S32
* @param[in] info Kernel metadata:
* - k Number of matrix columns/rows depending on the type of reduction.
* - is_reshaped True if the matrix has been reshaped.
* - scalar Scalar value to multiply each reduced column/row by.
* - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
*/
virtual void configure(const ICLTensor *input, ICLTensor *output, const GEMMLowpReductionKernelInfo &info) = 0;
/** Initialise the kernel's input and output.
*
* @param[in] compile_context The compile context to be used.
* @param[in] input Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8.
* @param[out] output Output row-vector of sums of all the entries in each row/col of input tensor. Data type supported: S32
* @param[in] info Kernel metadata:
* - k Number of matrix columns/rows depending on the type of reduction.
* - is_reshaped True if the matrix has been reshaped.
* - scalar Scalar value to multiply each reduced column/row by.
* - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
*/
virtual void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const GEMMLowpReductionKernelInfo &info) = 0;
protected:
const ICLTensor *_input;
ICLTensor *_output;
};
/** OpenCL kernel used to compute the row-vectors of sums of all the entries in each row of Matrix A.
*
* @note This stage is needed to handle the offset of matrix product
* https://github.com/google/gemmlowp/blob/master/doc/low-precision.md
*/
class CLGEMMLowpMatrixAReductionKernel : public ICLGEMMLowpReductionKernel
{
public:
/** Initialise the kernel's input and output.
*
* @param[in] mtx_a Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8.
* @param[out] vector_sum_row Output row-vector of sums of all the entries in each row of mtx_a. Data type supported: S32
* @param[in] info Kernel metadata:
* - k Number of matrix columns/rows depending on the type of reduction.
* - is_reshaped True if the matrix has been reshaped.
* - scalar Scalar value to multiply each reduced column/row by.
* - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
*/
void configure(const ICLTensor *mtx_a, ICLTensor *vector_sum_row, const GEMMLowpReductionKernelInfo &info) override;
/** Initialise the kernel's input and output.
*
* @param[in] compile_context The compile context to be used.
* @param[in] mtx_a Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8.
* @param[out] vector_sum_row Output row-vector of sums of all the entries in each row of mtx_a. Data type supported: S32
* @param[in] info Kernel metadata:
* - k Number of matrix columns/rows depending on the type of reduction.
* - is_reshaped True if the matrix has been reshaped.
* - scalar Scalar value to multiply each reduced column/row by.
* - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
*/
void configure(const CLCompileContext &compile_context, const ICLTensor *mtx_a, ICLTensor *vector_sum_row, const GEMMLowpReductionKernelInfo &info) override;
/** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpMatrixAReductionKernel
*
* @param[in] mtx_a Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8.
* @param[in] vector_sum_row Output row-vector of sums of all the entries in each row of mtx_a. Data type supported: S32
* @param[in] info Kernel metadata:
* - k Number of matrix columns/rows depending on the type of reduction.
* - is_reshaped True if the matrix has been reshaped.
* - scalar Scalar value to multiply each reduced column/row by.
* - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
*
* @return a status
*/
static Status validate(const ITensorInfo *mtx_a, const ITensorInfo *vector_sum_row, const GEMMLowpReductionKernelInfo &info);
// Inherited methods overridden:
void run(const Window &window, cl::CommandQueue &queue) override;
};
/** OpenCL kernel used to compute the row-vectors of sums of all the entries in each column of Matrix B.
*
* @note This stage is needed to handle the offset of matrix product
* https://github.com/google/gemmlowp/blob/master/doc/low-precision.md
*/
class CLGEMMLowpMatrixBReductionKernel : public ICLGEMMLowpReductionKernel
{
public:
/** Initialise the kernel's input and output.
*
* @param[in] mtx_b Input tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8/QSYMM8_PER_CHANNEL.
* @param[out] vector_sum_col Output row-vector of sums of all the entries in each column of mtx_b. Data type supported: S32
* @param[in] info Kernel metadata:
* - k Number of matrix columns/rows depending on the type of reduction.
* - is_reshaped True if the matrix has been reshaped.
* - scalar Scalar value to multiply each reduced column/row by.
* - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
*/
void configure(const ICLTensor *mtx_b, ICLTensor *vector_sum_col, const GEMMLowpReductionKernelInfo &info) override;
/** Initialise the kernel's input and output.
*
* @param[in] compile_context The compile context to be used.
* @param[in] mtx_b Input tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8/QSYMM8_PER_CHANNEL.
* @param[out] vector_sum_col Output row-vector of sums of all the entries in each column of mtx_b. Data type supported: S32
* @param[in] info Kernel metadata:
* - k Number of matrix columns/rows depending on the type of reduction.
* - is_reshaped True if the matrix has been reshaped.
* - scalar Scalar value to multiply each reduced column/row by.
* - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
*/
void configure(const CLCompileContext &compile_context, const ICLTensor *mtx_b, ICLTensor *vector_sum_col, const GEMMLowpReductionKernelInfo &info) override;
/** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpMatrixBReductionKernel
*
* @param[in] mtx_b Input tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8/QSYMM8_PER_CHANNEL.
* @param[in] vector_sum_col Output row-vector of sums of all the entries in each column of mtx_b. Data type supported: S32
* @param[in] info Kernel metadata:
* - k Number of matrix columns/rows depending on the type of reduction.
* - is_reshaped True if the matrix has been reshaped.
* - scalar Scalar value to multiply each reduced column/row by.
* - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
*
* @return a status
*/
static Status validate(const ITensorInfo *mtx_b, const ITensorInfo *vector_sum_col, const GEMMLowpReductionKernelInfo &info);
// Inherited methods overridden:
void run(const Window &window, cl::CommandQueue &queue) override;
};
} // namespace arm_compute
#endif /* ARM_COMPUTE_CLGEMMLOWREDUCTIONKERNEL_H */
|
