2 * Copyright (c) 2017-2018 ARM Limited.
4 * SPDX-License-Identifier: MIT
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:
13 * The above copyright notice and this permission notice shall be included in all
14 * copies or substantial portions of the Software.
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
24 #ifndef __ARM_COMPUTE_CLGEMMMATRIXMULTIPLYKERNEL_H__
25 #define __ARM_COMPUTE_CLGEMMMATRIXMULTIPLYKERNEL_H__
27 #include "arm_compute/core/CL/ICLKernel.h"
33 /** OpenCL kernel to multiply two input matrices "A" and "B" . All elements of the output matrix will be multiplied by alpha
35 * @note If the input tensors @p input0 and @p input1 have been reshaped respectively with @ref CLGEMMInterleave4x4Kernel" and @ref CLGEMMTranspose1xWKernel,
36 * the flag @p is_interleaved_transposed must be set to true
38 * @attention The second input tensor must have at least 2 dimensions (matrix)
41 class CLGEMMMatrixMultiplyKernel : public ICLKernel
44 /** Default constructor */
45 CLGEMMMatrixMultiplyKernel();
46 /** Prevent instances of this class from being copied (As this class contains pointers) */
47 CLGEMMMatrixMultiplyKernel(const CLGEMMMatrixMultiplyKernel &) = delete;
48 /** Prevent instances of this class from being copied (As this class contains pointers) */
49 CLGEMMMatrixMultiplyKernel &operator=(const CLGEMMMatrixMultiplyKernel &) = delete;
50 /** Allow instances of this class to be moved */
51 CLGEMMMatrixMultiplyKernel(CLGEMMMatrixMultiplyKernel &&) = default;
52 /** Allow instances of this class to be moved */
53 CLGEMMMatrixMultiplyKernel &operator=(CLGEMMMatrixMultiplyKernel &&) = default;
54 /** Initialise the kernel's input, output and alpha
56 * @param[in] input0 Input tensor containing the Matrix A. Data types supported: QS8/QS16/F16/F32
57 * @param[in] input1 Input tensor containing the Matrix B. Data type supported: same as @p input0
58 * @param[out] output Output tensor to store the result of matrix multiplication. Data type supported: same as @p input0
59 * @param[in] alpha Weight of the matrix product
60 * @param[in] is_interleaved_transposed (Optional) True if input0 and input1 have been reshaped respectively using @ref CLGEMMInterleave4x4Kernel and @ref CLGEMMTranspose1xWKernel
61 * @param[in] reshape_info (Optional) GEMM reshape info. If is_interleaved_transposed = true, this object must contain the information to understand how the matrix A and matrix B have been reshaped
64 void configure(const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output, float alpha, bool is_interleaved_transposed = true, const GEMMReshapeInfo &reshape_info = GEMMReshapeInfo());
65 /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMMatrixMultiplyKernel
67 * @param[in] input0 Input tensor containing the Matrix A. Data types supported: QS8/QS16/F16/F32
68 * @param[in] input1 Input tensor containing the Matrix B. Data type supported: same as @p input0
69 * @param[in] output Output tensor to store the result of matrix multiplication. Data type supported: same as @p input0
70 * @param[in] alpha Weight of the matrix product
71 * @param[in] is_interleaved_transposed True if input0 and input1 have been reshaped respectively using @ref CLGEMMInterleave4x4Kernel and @ref CLGEMMTranspose1xWKernel
72 * @param[in] reshape_info GEMM reshape info. If is_interleaved_transposed = true, this object must contain the information to understand how the matrix A and matrix B have been reshaped
73 * @param[in] gpu_target GPU Target
77 static Status validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *output, float alpha, bool is_interleaved_transposed, const GEMMReshapeInfo &reshape_info,
78 GPUTarget gpu_target);
80 // Inherited methods overridden:
81 void run(const Window &window, cl::CommandQueue &queue) override;
84 const ICLTensor *_input0;
85 const ICLTensor *_input1;
88 } // namespace arm_compute
89 #endif /* __ARM_COMPUTE_CLGEMMMATRIXMULTIPLYKERNEL_H__ */