1 /* 2 * Copyright (c) 2016-2021 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_MUL_KERNEL_H 25 #define ARM_COMPUTE_CL_MUL_KERNEL_H 26 27 #include "src/core/common/Macros.h" 28 #include "src/gpu/cl/ClCompileContext.h" 29 #include "src/gpu/cl/IClKernel.h" 30 31 namespace arm_compute 32 { 33 namespace opencl 34 { 35 namespace kernels 36 { 37 /** Interface for the pixelwise multiplication kernel. 38 * 39 * For binary elementwise ops in-place cannot be enabled by passing nullptr to dst, it can only be enabled by passing either src1 or src2 to dst instead. 40 * 41 */ 42 class ClMulKernel : public IClKernel 43 { 44 public: 45 ClMulKernel(); 46 ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClMulKernel); 47 /** Initialise the kernel's src and dst. 48 * 49 * Valid configurations (Input1,Input2) -> Output : 50 * 51 * - (U8,U8) -> U8 52 * - (U8,U8) -> S16 53 * - (U8,S16) -> S16 54 * - (S16,U8) -> S16 55 * - (S16,S16) -> S16 56 * - (S32,S32) -> S32 57 * - (F16,F16) -> F16 58 * - (F32,F32) -> F32 59 * - (QASYMM8,QASYMM8) -> QASYMM8 60 * - (QASYMM8_SIGNED,QASYMM8_SIGNED) -> QASYMM8_SIGNED 61 * - (QSYMM16,QSYMM16) -> QSYMM16 62 * - (QSYMM16,QSYMM16) -> S32 63 * 64 * @param[in] compile_context The compile context to be used. 65 * @param[in] src1 An src tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/F32/S32 66 * @param[in] src2 An src tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/F32/S32 67 * @param[out] dst The dst tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/F32/S32 68 * @param[in] scale Scale to apply after multiplication. 69 * Scale must be positive and its value must be either 1/255 or 1/2^n where n is between 0 and 15. 70 * @param[in] overflow_policy Overflow policy. Supported overflow policies: Wrap, Saturate 71 * @param[in] rounding_policy Rounding policy. Supported rounding modes: to zero, to nearest even. 72 * @param[in] act_info (Optional) Activation layer information in case of a fused activation. 73 */ 74 void configure(const CLCompileContext &compile_context, ITensorInfo *src1, ITensorInfo *src2, ITensorInfo *dst, float scale, 75 ConvertPolicy overflow_policy, RoundingPolicy rounding_policy, const ActivationLayerInfo &act_info = ActivationLayerInfo()); 76 /** Static function to check if given info will lead to a valid configuration 77 * 78 * Similar to @ref ClMulKernel::configure() 79 * 80 * @return a status 81 */ 82 static Status validate(const ITensorInfo *src1, const ITensorInfo *src2, const ITensorInfo *dst, float scale, 83 ConvertPolicy overflow_policy, RoundingPolicy rounding_policy, const ActivationLayerInfo &act_info = ActivationLayerInfo()); 84 85 // Inherited methods overridden: 86 void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override; 87 }; 88 89 /** Interface for the complex pixelwise multiplication kernel. */ 90 class ClComplexMulKernel : public ICLKernel 91 { 92 public: 93 ClComplexMulKernel(); 94 ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClComplexMulKernel); 95 /** Initialise the kernel's src and dst. 96 * 97 * @param[in] compile_context The compile context to be used. 98 * @param[in] src1 An src tensor info. Data types supported: F32. Number of channels supported: 2. 99 * @param[in] src2 An src tensor info. Data types supported: same as @p src1. Number of channels supported: same as @p src1. 100 * @param[out] dst The dst tensor info. Data types supported: same as @p src1. Number of channels supported: same as @p src1. 101 * @param[in] act_info (Optional) Activation layer information in case of a fused activation. 102 */ 103 void configure(const CLCompileContext &compile_context, ITensorInfo *src1, ITensorInfo *src2, ITensorInfo *dst, const ActivationLayerInfo &act_info = ActivationLayerInfo()); 104 /** Static function to check if given info will lead to a valid configuration 105 * 106 * Similar to @ref ClComplexMulKernel::configure() 107 * 108 * @return a status 109 */ 110 static Status validate(const ITensorInfo *src1, const ITensorInfo *src2, const ITensorInfo *dst, const ActivationLayerInfo &act_info = ActivationLayerInfo()); 111 112 // Inherited methods overridden: 113 void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override; 114 }; 115 } // namespace kernels 116 } // namespace opencl 117 } // namespace arm_compute 118 #endif /* ARM_COMPUTE_CL_MUL_KERNEL_H */ 119