1 /* 2 * Copyright (c) 2019-2021 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_GEMM_MATRIXMULTIPLY_NATIVE_KERNEL_H 25 #define ARM_COMPUTE_CL_GEMM_MATRIXMULTIPLY_NATIVE_KERNEL_H 26 27 #include "arm_compute/core/KernelDescriptors.h" 28 #include "src/core/common/Macros.h" 29 #include "src/gpu/cl/ClCompileContext.h" 30 #include "src/gpu/cl/IClKernel.h" 31 32 namespace arm_compute 33 { 34 namespace opencl 35 { 36 namespace kernels 37 { 38 /** OpenCL kernel to multiply matrices when neither of the input matrices have been reshaped */ 39 class ClGemmMatrixMultiplyNativeKernel : public IClKernel 40 { 41 public: 42 ClGemmMatrixMultiplyNativeKernel(); 43 ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClGemmMatrixMultiplyNativeKernel); 44 /** Initialise the kernel's input and dst. 45 * 46 * @param[in] compile_context The compile context to be used. 47 * @param[in] src0 Input tensor for the LHS matrix. Data type supported: F32/F16. The number of dimensions for the LHS matrix must be less or equal than 4. 48 * @param[in] src1 Input tensor for the RHS matrix. Data type supported: same as @p src0. The number of dimensions for the RHS matrix must be less or equal than 3. 49 * @param[in] src2 Input tensor containing the bias matrix. Data type supported: same as @p src0. 50 * @param[out] dst dst tensor info. Data type supported: same as @p src0 51 * @param[in] alpha Weight of the matrix product 52 * @param[in] beta Weight of the matrix bias 53 * @param[in] lhs_info LHS matrix information used to retrieve the number of rows and accumulations to be processed by each thread. Only the following values are supported: 54 * lhs_info.m0: 1,2,3,4,5,6,7,8 55 * lhs_info.k0: 2,3,4,8,16 56 * @param[in] rhs_info RHS matrix information used to retrieve the number of columns and accumulations to be processed by each thread. Only the following values are supported: 57 * rhs_info.n0: 2,3,4,8,16 58 * rhs_info.k0: same of lhs_info.k0 59 * @param[in] gemm_info GEMM information used to retrieve the original dimensions of the input matrices 60 */ 61 void configure(const ClCompileContext &compile_context, ITensorInfo *src0, ITensorInfo *src1, ITensorInfo *src2, ITensorInfo *dst, float alpha, float beta, 62 const GEMMLHSMatrixInfo &lhs_info, 63 const GEMMRHSMatrixInfo &rhs_info, 64 const GEMMKernelInfo &gemm_info); 65 /** Static function to check if given info will lead to a valid configuration 66 * 67 * Similar to @ref ClGemmMatrixMultiplyNativeKernel::configure() 68 * 69 * @return a status 70 */ 71 static Status validate(const ITensorInfo *src0, const ITensorInfo *src1, const ITensorInfo *src2, const ITensorInfo *dst, float alpha, float beta, const GEMMLHSMatrixInfo &lhs_info, 72 const GEMMRHSMatrixInfo &rhs_info, 73 const GEMMKernelInfo &gemm_info); 74 75 // Inherited methods overridden: 76 void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override; 77 78 private: 79 bool _slide_matrix_b{ true }; 80 bool _reinterpret_input_as_3d{ false }; 81 bool _reinterpret_output_as_3d{ false }; 82 bool _use_dummy_work_items{ false }; 83 bool _add_bias{ false }; 84 signed int _m{ 1 }; 85 signed int _n{ 1 }; 86 signed int _k{ 1 }; 87 unsigned int _num_post_op_args{ 0 }; // (EXPERIMENTAL_POST_OPS) total number of post op arguments 88 }; 89 } // namespace kernels 90 } // namespace opencl 91 } // namespace arm_compute 92 #endif /* ARM_COMPUTE_CL_GEMM_MATRIXMULTIPLY_NATIVE_KERNEL_H */ 93