Georgios Pinitas | 4a578b9 | 2021-06-25 12:13:49 +0100 | [diff] [blame] | 1 | /* |
| 2 | * Copyright (c) 2017-2021 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_CL_GEMMLOWP_REDUCTION_KERNEL_H |
| 25 | #define ARM_COMPUTE_CL_GEMMLOWP_REDUCTION_KERNEL_H |
| 26 | |
| 27 | #include "arm_compute/core/KernelDescriptors.h" |
| 28 | #include "src/core/common/Macros.h" |
| 29 | #include "src/core/gpu/cl/ClCompileContext.h" |
| 30 | #include "src/core/gpu/cl/IClKernel.h" |
| 31 | |
| 32 | namespace arm_compute |
| 33 | { |
| 34 | namespace opencl |
| 35 | { |
| 36 | namespace kernels |
| 37 | { |
| 38 | /** Common interface for all OpenCL reduction kernels */ |
| 39 | class IClGemmLowpReductionKernel : public IClKernel |
| 40 | { |
| 41 | public: |
| 42 | IClGemmLowpReductionKernel(); |
| 43 | ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(IClGemmLowpReductionKernel); |
| 44 | /** Initialise the kernel's input and output. |
| 45 | * |
| 46 | * @param[in] compile_context The compile context to be used. |
| 47 | * @param[in] input Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8. |
| 48 | * @param[out] output Output row-vector of sums of all the entries in each row/col of input tensor. Data type supported: S32 |
| 49 | * @param[in] info Kernel metadata: |
| 50 | * - k Number of matrix columns/rows depending on the type of reduction. |
| 51 | * - is_reshaped True if the matrix has been reshaped. |
| 52 | * - scalar Scalar value to multiply each reduced column/row by. |
| 53 | * - mul_byscalar True if each reduced column/row must be multiplied by a scalar value. |
| 54 | */ |
| 55 | virtual void configure(const CLCompileContext &compile_context, const ITensorInfo *input, ITensorInfo *output, const GEMMLowpReductionKernelInfo &info) = 0; |
| 56 | }; |
| 57 | |
| 58 | /** OpenCL kernel used to compute the row-vectors of sums of all the entries in each row of Matrix A. |
| 59 | * |
| 60 | * @note This stage is needed to handle the offset of matrix product |
| 61 | * https://github.com/google/gemmlowp/blob/master/doc/low-precision.md |
| 62 | */ |
| 63 | class ClGemmLowpMatrixAReductionKernel : public IClGemmLowpReductionKernel |
| 64 | { |
| 65 | public: |
| 66 | /** Initialise the kernel's input and output. |
| 67 | * |
| 68 | * @param[in] compile_context The compile context to be used. |
| 69 | * @param[in] mtx_a Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8. |
| 70 | * @param[out] vector_sum_row Output row-vector of sums of all the entries in each row of mtx_a. Data type supported: S32 |
| 71 | * @param[in] info Kernel metadata: |
| 72 | * - k Number of matrix columns/rows depending on the type of reduction. |
| 73 | * - is_reshaped True if the matrix has been reshaped. |
| 74 | * - scalar Scalar value to multiply each reduced column/row by. |
| 75 | * - mul_byscalar True if each reduced column/row must be multiplied by a scalar value. |
| 76 | */ |
| 77 | void configure(const CLCompileContext &compile_context, const ITensorInfo *mtx_a, ITensorInfo *vector_sum_row, const GEMMLowpReductionKernelInfo &info) override; |
| 78 | /** Static function to check if given info will lead to a valid configuration |
| 79 | * |
| 80 | * Similar to @ref ClGemmLowpQuantizeDownInt32ScaleByFixedPointKernel::configure() |
| 81 | * |
| 82 | * @return a status |
| 83 | */ |
| 84 | static Status validate(const ITensorInfo *mtx_a, const ITensorInfo *vector_sum_row, const GEMMLowpReductionKernelInfo &info); |
| 85 | |
| 86 | // Inherited methods overridden: |
| 87 | void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override; |
| 88 | }; |
| 89 | |
| 90 | /** OpenCL kernel used to compute the row-vectors of sums of all the entries in each column of Matrix B. |
| 91 | * |
| 92 | * @note This stage is needed to handle the offset of matrix product |
| 93 | * https://github.com/google/gemmlowp/blob/master/doc/low-precision.md |
| 94 | */ |
| 95 | class ClGemmLowpMatrixBReductionKernel : public IClGemmLowpReductionKernel |
| 96 | { |
| 97 | public: |
| 98 | /** Initialise the kernel's input and output. |
| 99 | * |
| 100 | * @param[in] compile_context The compile context to be used. |
| 101 | * @param[in] mtx_b Input tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8/QSYMM8_PER_CHANNEL. |
| 102 | * @param[out] vector_sum_col Output row-vector of sums of all the entries in each column of mtx_b. Data type supported: S32 |
| 103 | * @param[in] info Kernel metadata: |
| 104 | * - k Number of matrix columns/rows depending on the type of reduction. |
| 105 | * - is_reshaped True if the matrix has been reshaped. |
| 106 | * - scalar Scalar value to multiply each reduced column/row by. |
| 107 | * - mul_byscalar True if each reduced column/row must be multiplied by a scalar value. |
| 108 | */ |
| 109 | void configure(const CLCompileContext &compile_context, const ITensorInfo *mtx_b, ITensorInfo *vector_sum_col, const GEMMLowpReductionKernelInfo &info) override; |
| 110 | /** Static function to check if given info will lead to a valid configuration |
| 111 | * |
| 112 | * Similar to @ref ClGemmLowpQuantizeDownInt32ScaleByFixedPointKernel::configure() |
| 113 | * |
| 114 | * @return a status |
| 115 | */ |
| 116 | static Status validate(const ITensorInfo *mtx_b, const ITensorInfo *vector_sum_col, const GEMMLowpReductionKernelInfo &info); |
| 117 | |
| 118 | // Inherited methods overridden: |
| 119 | void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override; |
| 120 | }; |
| 121 | } // namespace kernels |
| 122 | } // namespace opencl |
| 123 | } // namespace arm_compute |
| 124 | #endif /* ARM_COMPUTE_CL_GEMMLOWP_REDUCTION_KERNEL_H */ |