• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /*
2  * Copyright (c) 2019-2020 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_CORE_KERNEL_DESCRIPTORS_H
25 #define ARM_COMPUTE_CORE_KERNEL_DESCRIPTORS_H
26 
27 #include "arm_compute/core/PixelValue.h"
28 #include "arm_compute/core/Types.h"
29 
30 namespace arm_compute
31 {
32 /** Descriptor for FFT scale kernels */
33 struct FFTScaleKernelInfo
34 {
35     float scale{ 0.f };      /**< Axis to perform the kernel on. */
36     bool  conjugate{ true }; /**< Flag to conjugate the output/ */
37 };
38 
39 /** Descriptor for FFT digit reverse kernels */
40 struct FFTDigitReverseKernelInfo
41 {
42     unsigned int axis{ 0 };          /**< Axis to perform the kernel on. */
43     bool         conjugate{ false }; /**< Flag to conjugate the output/ */
44 };
45 
46 /** Descriptor used by the FFT core kernels */
47 struct FFTRadixStageKernelInfo
48 {
49     unsigned int axis{ 0 };               /**< Axis to run the kernel on. */
50     unsigned int radix{ 0 };              /**< Radix to use. */
51     unsigned int Nx{ 0 };                 /**< Nx coefficient. */
52     bool         is_first_stage{ false }; /**< Flags if the FFT kernels is the first stage of a decomposed FFT. */
53 };
54 
55 /** Descriptor used by the GEMM kernels */
56 struct GEMMKernelInfo
57 {
58     GEMMKernelInfo() = default;
GEMMKernelInfoGEMMKernelInfo59     GEMMKernelInfo(
60         unsigned int        im,
61         unsigned int        in,
62         unsigned int        ik,
63         unsigned int        idepth_output_gemm3d,
64         bool                ireinterpret_input_as_3d,
65         bool                ibroadcast_bias,
66         bool                ifp_mixed_precision,
67         bool                ihas_pad_y,
68         ActivationLayerInfo iactivation_info,
69         int                 inmult_transpose1xW_width,
70         int                 imult_interleave4x4_height,
71         GEMMLHSMatrixInfo   ilhs_info,
72         GEMMRHSMatrixInfo   irhs_info,
73         int32_t             ina_offset,
74         int32_t             inb_offset)
75         : m(im), n(in), k(ik), depth_output_gemm3d(idepth_output_gemm3d), reinterpret_input_as_3d(ireinterpret_input_as_3d), broadcast_bias(ibroadcast_bias), fp_mixed_precision(ifp_mixed_precision),
76           has_pad_y(ihas_pad_y), activation_info(iactivation_info), mult_transpose1xW_width(inmult_transpose1xW_width), mult_interleave4x4_height(imult_interleave4x4_height), lhs_info(ilhs_info), rhs_info(irhs_info),
77           a_offset(ina_offset), b_offset(inb_offset)
78     {
79     }
80 
81     unsigned int            m{ 0 };                           /**< Number of LHS rows*/
82     unsigned int            n{ 0 };                           /**< Number of RHS columns*/
83     unsigned int            k{ 0 };                           /**< Number of LHS columns or RHS rows */
84     unsigned int            depth_output_gemm3d{ 0 };         /**< Depth of the output tensor in case is reinterpreted as 3D */
85     bool                    reinterpret_input_as_3d{ false }; /**< Flag used to reinterpret the input as 3D */
86     bool                    broadcast_bias{ false };          /**< Flag used to broadcast the bias addition */
87     bool                    fp_mixed_precision{ false };      /**< Flag used to indicate wider accumulators (32 bit instead of 16 for FP16). */
88     bool                    has_pad_y{ false };               /**< Flag used to indicate if the input/output tensors have internal pad on the y direction */
89     ActivationLayerInfo     activation_info{};                /**< Activation function to perform after the matrix multiplication */
90     int                     mult_transpose1xW_width{ 1 };     /**< Multiplication factor for the width of the 1xW transposed block */
91     int                     mult_interleave4x4_height{ 1 };   /**< Multiplication factor for the height of the 4x4 interleaved block */
92     GEMMLHSMatrixInfo       lhs_info{};                       /**< LHS matrix information used to retrieve the number of rows processed by each thread */
93     GEMMRHSMatrixInfo       rhs_info{};                       /**< RHS matrix information used for reshaping the RHS matrix */
94     int32_t                 a_offset{ 0 };                    /**< Offset to be added to each element of the matrix A */
95     int32_t                 b_offset{ 0 };                    /**< Offset to be added to each element of the matrix B */
96     GEMMLowpOutputStageInfo output_stage{};                   /**< GEMMLowp output stage information */
97 };
98 
99 /** Descriptor used by the depthwise convolution kernels */
100 struct DWCKernelInfo
101 {
102     ActivationLayerInfo activation_info{}; /**< Activation function to perform after the depthwise convolution */
103 };
104 
105 /** Descriptor used by the depthwise convolution kernels to retrieve the number of output elements processed by each thread */
106 struct DWCWeightsKernelInfo
107 {
108     unsigned int n0{ 0 }; /**< Number of columns processed by each thread */
109 };
110 
111 /** Descriptor used by the softmax kernels */
112 struct SoftmaxKernelInfo
113 {
114     float    beta{ 1.f };                          /**< A scaling factor for the exponent with default value 1.0 */
115     bool     is_log{ false };                      /**< Flag used to perform Log Softmax operation */
116     DataType input_data_type{ DataType::UNKNOWN }; /**< Input tensor data type */
117 };
118 
119 /** Descriptor used by the direct convolution layer output stage kernels */
120 struct DirectConvolutionLayerOutputStageKernelInfo
121 {
122     int32_t  result_fixedpoint_multiplier{ 0 };     /**< Result output stage multiplier used for quantizing */
123     int32_t  result_shift{ 0 };                     /**< Result output stage shift used for quantizing */
124     int32_t  result_offset_after_shift{ 0 };        /**< Result offset used for quantizing */
125     DataType output_data_type{ DataType::UNKNOWN }; /**< Output tensor data type to use if the output is not initialized */
126 };
127 
128 struct InstanceNormalizationLayerKernelInfo
129 {
130     /** Default constructor */
InstanceNormalizationLayerKernelInfoInstanceNormalizationLayerKernelInfo131     InstanceNormalizationLayerKernelInfo()
132         : InstanceNormalizationLayerKernelInfo(1.f, 0.f, 1e-12, true)
133     {
134     }
135     /** Constructor
136      *
137      * @param[in] gamma               The scale scalar value applied to the normalized tensor.
138      * @param[in] beta                The offset scalar value applied to the normalized tensor
139      * @param[in] epsilon             Lower bound value for the normalization.
140      * @param[in] use_mixed_precision Use mixed precision in case of FP16 execution.
141      */
InstanceNormalizationLayerKernelInfoInstanceNormalizationLayerKernelInfo142     InstanceNormalizationLayerKernelInfo(float gamma, float beta, float epsilon, bool use_mixed_precision)
143         : gamma(gamma), beta(beta), epsilon(epsilon), use_mixed_precision(use_mixed_precision)
144     {
145     }
146 
147     float gamma;               /**< The scale scalar value applied to the normalized tensor. Defaults to 1.0 */
148     float beta;                /**< The offset scalar value applied to the normalized tensor. Defaults to 0.0 */
149     float epsilon;             /**< Lower bound value for the normalization. Defaults to 1e-12 */
150     bool  use_mixed_precision; /**< Use mixed precision in case of FP16 execution. Defaults to true */
151 };
152 
153 struct GEMMLowpReductionKernelInfo
154 {
155     /** Default constructor */
156     GEMMLowpReductionKernelInfo() = default;
157     /** Constructor
158      *
159      * @param[in] k             Number of matrix columns/rows.
160      * @param[in] is_reshaped   True if the input tensor has been reshaped.
161      * @param[in] scalar        Scalar value to multiply each reduced column/row by.
162      * @param[in] mul_by_scalar True if each column/row reduction has to be multiplied by a scalar value.
163      */
GEMMLowpReductionKernelInfoGEMMLowpReductionKernelInfo164     GEMMLowpReductionKernelInfo(int32_t k, bool is_reshaped, int32_t scalar, bool mul_by_scalar)
165         : k(k), is_reshaped(is_reshaped), scalar(scalar), mul_by_scalar(mul_by_scalar)
166     {
167     }
168 
169     int32_t k{ 0 };                 /**< Number of matrix columns/rows */
170     bool    is_reshaped{ false };   /**< True if the input tensor has been reshaped */
171     int32_t scalar{ 0 };            /**< Scalar value to multiply each reduced column/row by */
172     bool    mul_by_scalar{ false }; /**< True if each column/row reduction has to be multiplied by a scalar value */
173 };
174 
175 struct ScaleKernelInfo
176 {
177     /** Constructor
178      *
179      * @param[in] interpolation_policy  Interpolation type to use
180      * @param[in] border_mode           Border mode policy
181      * @param[in] constant_border_value (Optional) Constant value to use for borders if border_mode is set to CONSTANT and use_padding is set to false. Defaults to default @ref PixelValue
182      * @param[in] sampling_policy       (Optional) Sampling policy used by the interpolation. Defaults to @ref SamplingPolicy::CENTER
183      * @param[in] use_padding           (Optional) Is padding in use or not. Defaults to true.
184      * @param[in] align_corners         (Optional) Align corners of input and output, only affecting bilinear policy with TOP_LEFT sampling policy. Defaults to false.
185      */
186     ScaleKernelInfo(InterpolationPolicy interpolation_policy,
187                     BorderMode          border_mode,
188                     PixelValue          constant_border_value = PixelValue(),
189                     SamplingPolicy      sampling_policy       = SamplingPolicy::CENTER,
190                     bool                use_padding           = true,
191                     bool                align_corners         = false)
192         : interpolation_policy{ interpolation_policy },
193           border_mode{ border_mode },
194           constant_border_value{ constant_border_value },
195           sampling_policy{ sampling_policy },
196           use_padding{ use_padding },
197           align_corners{ align_corners }
198     {
199     }
200 
201     InterpolationPolicy interpolation_policy;  /**< Interpolation type to use */
202     BorderMode          border_mode;           /**< Border mode policy */
203     PixelValue          constant_border_value; /**< Constant value to use for constant border mode policy */
204     SamplingPolicy      sampling_policy;       /**< Sampling policy used by the interpolation. */
205     bool                use_padding;           /**< Indication of using padding */
206     bool                align_corners;         /**< Align corners of input and output */
207 };
208 
209 struct ThresholdKernelInfo
210 {
211     /** Default constructor */
212     ThresholdKernelInfo() = default;
213     /** Constructor
214      *
215      * @param[in] threshold   Threshold. When the threshold type is RANGE, this is used as the lower threshold.
216      * @param[in] false_value value to set when the condition is not respected.
217      * @param[in] true_value  value to set when the condition is respected.
218      * @param[in] type        Thresholding type. Either RANGE or BINARY.
219      * @param[in] upper       Upper threshold. Only used when the thresholding type is RANGE.
220      */
ThresholdKernelInfoThresholdKernelInfo221     ThresholdKernelInfo(uint8_t threshold, uint8_t false_value, uint8_t true_value, ThresholdType type, uint8_t upper)
222         : threshold(threshold), false_value(false_value), true_value(true_value), type(type), upper(upper)
223     {
224     }
225 
226     uint8_t       threshold{ 0 };
227     uint8_t       false_value{ 0 };
228     uint8_t       true_value{ 0 };
229     ThresholdType type{ ThresholdType::BINARY };
230     uint8_t       upper{ 0 };
231 };
232 } // namespace arm_compute
233 #endif /* ARM_COMPUTE_CORE_KERNEL_DESCRIPTORS_H */
234