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
|
/*
* Copyright (c) 2017 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_NEGEMM_H__
#define __ARM_COMPUTE_NEGEMM_H__
#include "arm_compute/core/NEON/kernels/NEFillBorderKernel.h"
#include "arm_compute/core/NEON/kernels/NEGEMMInterleave4x4Kernel.h"
#include "arm_compute/core/NEON/kernels/NEGEMMMatrixAdditionKernel.h"
#include "arm_compute/core/NEON/kernels/NEGEMMMatrixMultiplyKernel.h"
#include "arm_compute/core/NEON/kernels/NEGEMMTranspose1xWKernel.h"
#include "arm_compute/runtime/IFunction.h"
#include "arm_compute/runtime/Tensor.h"
namespace arm_compute
{
/** Basic function to execute GEMM on NEON. This function calls the following NEON kernels:
*
* -# @ref NEGEMMInterleave4x4Kernel (if the output tensor is a matrix)
* -# @ref NEGEMMTranspose1xWKernel (if the output tensor is a matrix)
* -# @ref NEGEMMMatrixMultiplyKernel
* -# @ref NEGEMMMatrixAdditionKernel (if c != nullptr and beta != 0.0)
*
*/
class NEGEMM : public IFunction
{
public:
/** Constructor */
NEGEMM();
/** Initialise the kernel's inputs, output
*
* @note GEMM: General Matrix Multiply - [alpha * A * B + beta * C].
* @note GEMM: The tensors a, b, c, d must have the same data type. You should not mix data types when calling this function.
*
* @param[in] a First input tensor (Matrix A or Vector A). Data type supported: QS8/QS16/F16/F32
* @param[in] b Second input tensor (Matrix B). Data type supported: same as @p a
* @param[in] c Third input tensor (Matrix C). It can be a nullptr if just the multiplication between @p a and @p b is needed. Data type supported: same as @p a
* @param[out] d Output tensor. Data type supported: same as @p a
* @param[in] alpha Weight of the matrix product
* @param[in] beta Weight of matrix C
*/
void configure(const ITensor *a, const ITensor *b, const ITensor *c, ITensor *d, float alpha, float beta);
// Inherited methods overridden:
void run() override;
private:
NEGEMMInterleave4x4Kernel _interleave_kernel;
NEGEMMTranspose1xWKernel _transpose_kernel;
NEGEMMMatrixMultiplyKernel _mm_kernel;
NEGEMMMatrixAdditionKernel _ma_kernel;
Tensor _tmp_a;
Tensor _tmp_b;
bool _run_vector_matrix_multiplication;
bool _run_addition;
};
}
#endif /*__ARM_COMPUTE_NEGEMM_H__ */
|
