From ae58bdf3b58739e105a24e3640d0245e81cea5ee Mon Sep 17 00:00:00 2001 From: Manuel Bottini Date: Thu, 17 Jun 2021 17:18:45 +0100 Subject: Port NEGEMMLowp Part 1 Details: Port NEGEMMLowpQuantizeDownInt32ScaleKernel to CpuGemmLowpQuantizeDownInt32ScaleKernel Port NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel to CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel Port NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel to CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel Port NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel to CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel Port NEGEMMLowpOutputStage functions to CpuGemmLowpOutputStage operators Partially Resolves: COMPMID-4403 Change-Id: I6d5f45e43f35d731d564ed3b5c0e804d2a318fb1 Signed-off-by: Manuel Bottini Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5833 Tested-by: Arm Jenkins Reviewed-by: Michele Di Giorgio Comments-Addressed: Arm Jenkins --- Android.bp | 9 +- .../runtime/NEON/functions/NEFullyConnectedLayer.h | 2 +- .../NEON/functions/NEGEMMConvolutionLayer.h | 2 +- .../runtime/NEON/functions/NEGEMMLowpOutputStage.h | 237 +----------- .../runtime/NEON/functions/NELSTMLayerQuantized.h | 50 +-- arm_compute/runtime/NEON/functions/NEQLSTMLayer.h | 2 +- docs/user_guide/release_version_and_change_log.dox | 8 +- examples/neon_gemm_qasymm8.cpp | 40 +- filelist.json | 5 + src/core/NEON/NEKernels.h | 4 - .../NEGEMMLowpQuantizeDownInt32ScaleKernel.cpp | 320 ---------------- .../NEGEMMLowpQuantizeDownInt32ScaleKernel.h | 114 ------ ...tizeDownInt32ToInt16ScaleByFixedPointKernel.cpp | 234 ------------ ...antizeDownInt32ToInt16ScaleByFixedPointKernel.h | 118 ------ ...ntizeDownInt32ToInt8ScaleByFixedPointKernel.cpp | 243 ------------ ...uantizeDownInt32ToInt8ScaleByFixedPointKernel.h | 121 ------ ...tizeDownInt32ToUint8ScaleByFixedPointKernel.cpp | 245 ------------ ...antizeDownInt32ToUint8ScaleByFixedPointKernel.h | 121 ------ .../CpuGemmLowpQuantizeDownInt32ScaleKernel.cpp | 326 ++++++++++++++++ .../CpuGemmLowpQuantizeDownInt32ScaleKernel.h | 107 ++++++ ...tizeDownInt32ToInt16ScaleByFixedPointKernel.cpp | 227 ++++++++++++ ...antizeDownInt32ToInt16ScaleByFixedPointKernel.h | 111 ++++++ ...ntizeDownInt32ToInt8ScaleByFixedPointKernel.cpp | 239 ++++++++++++ ...uantizeDownInt32ToInt8ScaleByFixedPointKernel.h | 114 ++++++ ...tizeDownInt32ToUint8ScaleByFixedPointKernel.cpp | 236 ++++++++++++ ...antizeDownInt32ToUint8ScaleByFixedPointKernel.h | 108 ++++++ .../NEON/functions/NEGEMMLowpOutputStage.cpp | 165 ++------- .../NEON/functions/NELSTMLayerQuantized.cpp | 15 +- .../cpu/operators/CpuGemmLowpOutputStage.cpp | 147 ++++++++ src/runtime/cpu/operators/CpuGemmLowpOutputStage.h | 90 +++++ tests/validation/NEON/GEMMLowp.cpp | 409 +-------------------- 31 files changed, 1826 insertions(+), 2343 deletions(-) delete mode 100644 src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.cpp delete mode 100644 src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.h delete mode 100644 src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp delete mode 100644 src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h delete mode 100644 src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp delete mode 100644 src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h delete mode 100644 src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp delete mode 100644 src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h create mode 100644 src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ScaleKernel.cpp create mode 100644 src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ScaleKernel.h create mode 100644 src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp create mode 100644 src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h create mode 100644 src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp create mode 100644 src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h create mode 100644 src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp create mode 100644 src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h create mode 100644 src/runtime/cpu/operators/CpuGemmLowpOutputStage.cpp create mode 100644 src/runtime/cpu/operators/CpuGemmLowpOutputStage.h diff --git a/Android.bp b/Android.bp index 19645c0c26..7dc764de93 100644 --- a/Android.bp +++ b/Android.bp @@ -169,10 +169,6 @@ cc_library_static { "src/core/NEON/kernels/NEGEMMLowpMatrixMultiplyKernel.cpp", "src/core/NEON/kernels/NEGEMMLowpOffsetContributionKernel.cpp", "src/core/NEON/kernels/NEGEMMLowpOffsetContributionOutputStageKernel.cpp", - "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.cpp", - "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp", - "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp", - "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp", "src/core/NEON/kernels/NEGEMMLowpReductionKernel.cpp", "src/core/NEON/kernels/NEGEMMMatrixAdditionKernel.cpp", "src/core/NEON/kernels/NEGEMMMatrixMultiplyKernel.cpp", @@ -286,6 +282,10 @@ cc_library_static { "src/core/cpu/kernels/CpuElementwiseUnaryKernel.cpp", "src/core/cpu/kernels/CpuFillKernel.cpp", "src/core/cpu/kernels/CpuFloorKernel.cpp", + "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ScaleKernel.cpp", + "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp", + "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp", + "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp", "src/core/cpu/kernels/CpuMulKernel.cpp", "src/core/cpu/kernels/CpuPermuteKernel.cpp", "src/core/cpu/kernels/CpuPool2dKernel.cpp", @@ -645,6 +645,7 @@ cc_library_static { "src/runtime/cpu/operators/CpuFlatten.cpp", "src/runtime/cpu/operators/CpuFloor.cpp", "src/runtime/cpu/operators/CpuGemmDirectConv2d.cpp", + "src/runtime/cpu/operators/CpuGemmLowpOutputStage.cpp", "src/runtime/cpu/operators/CpuMul.cpp", "src/runtime/cpu/operators/CpuPermute.cpp", "src/runtime/cpu/operators/CpuPool2d.cpp", diff --git a/arm_compute/runtime/NEON/functions/NEFullyConnectedLayer.h b/arm_compute/runtime/NEON/functions/NEFullyConnectedLayer.h index 9727e108a5..d2cd60e576 100644 --- a/arm_compute/runtime/NEON/functions/NEFullyConnectedLayer.h +++ b/arm_compute/runtime/NEON/functions/NEFullyConnectedLayer.h @@ -80,7 +80,7 @@ private: * -# @ref NEIm2ColKernel (called when the input comes from a convolutional layer) * -# @ref NETranspose (if @p are_weights_reshaped is set to false and transpose_weights is set to true ) (called once) * -# @ref NEGEMMMatrixMultiplyKernel or @ref NEGEMMLowpMatrixMultiplyCore (if quantized asymmetric) - * -# @ref NEGEMMMatrixAdditionKernel or @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint (if quantized asymmetric) (if @p biases is not equal to nullptr) + * -# @ref NEGEMMMatrixAdditionKernel or @ref NEGEMMLowpOutputStage (if quantized asymmetric) (if @p biases is not equal to nullptr) * * @note The fully connected layer accepts "weights" tensors only with 2 dimensions. */ diff --git a/arm_compute/runtime/NEON/functions/NEGEMMConvolutionLayer.h b/arm_compute/runtime/NEON/functions/NEGEMMConvolutionLayer.h index e89eae1d31..edb58e956a 100644 --- a/arm_compute/runtime/NEON/functions/NEGEMMConvolutionLayer.h +++ b/arm_compute/runtime/NEON/functions/NEGEMMConvolutionLayer.h @@ -155,7 +155,7 @@ private: * -# @ref NEIm2ColKernel * -# @ref NEGEMM (if the data type is BFLOAT16/FP16/FP32) * -# @ref NEGEMMLowpMatrixMultiplyCore (if the data type is QASYMM8/QASYMM8_SIGNED) - * -# @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint (if the data type is QASYMM8/QASYMM8_SIGNED) + * -# @ref NEGEMMLowpOutputStage (if the data type is QASYMM8/QASYMM8_SIGNED) * -# @ref NEArithmeticAddition (if biases != nullptr and we have a 1x1 convolution with the NHWC data layout) * -# @ref NECol2ImKernel (if NCHW data layout) * diff --git a/arm_compute/runtime/NEON/functions/NEGEMMLowpOutputStage.h b/arm_compute/runtime/NEON/functions/NEGEMMLowpOutputStage.h index fa5f5e3826..232344e5c2 100644 --- a/arm_compute/runtime/NEON/functions/NEGEMMLowpOutputStage.h +++ b/arm_compute/runtime/NEON/functions/NEGEMMLowpOutputStage.h @@ -25,7 +25,7 @@ #define ARM_COMPUTE_NEGEMMLOWPOUTPUTSTAGE_H #include "arm_compute/core/Types.h" -#include "arm_compute/runtime/NEON/INESimpleFunctionNoBorder.h" +#include "arm_compute/runtime/IFunction.h" /** This file contains all available output stages for GEMMLowp. * @@ -39,237 +39,17 @@ namespace arm_compute { class ITensor; class ITensorInfo; - -/** Basic function to execute NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint. - * - * NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint depends on 3 parameters: - * - * result_fixedpoint_multiplier, result_shift, result_offset_after_shift - * - * The final result is: - * - * (FixedPointMul(input[i][k], result_fixedpoint_multiplier) >> result_shift) + result_offset_after_shift - * - * where FixedPointMul(x, y) is the nearest integer to the following - * mathematical expression, evaluated without overflow or intermediate rounding: - * - * (x * y) / 2^31 - * - * For more information: https://github.com/google/gemmlowp/blob/master/public/output_stages.h#L68 - * - * In case the bias tensor is provided, the final result is: - * - * ((FixedPointMul(input[i][k] + bias[k], result_fixedpoint_multiplier)) >> result_shift) + result_offset_after_shift - * - * This function calls the following kernels: - * - * -# @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel - * - * @note The function accepts also 2 optional input arguments (min and max) which can be used to implement "rectified linear unit" activation functions - * after the result is shifted right by result_shift -*/ -class NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint : public INESimpleFunctionNoBorder -{ -public: - /** Constructor */ - NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint() = default; - /** Prevent instances of this class from being copied (As this class contains pointers) */ - NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint(const NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint &) = delete; - /** Prevent instances of this class from being copied (As this class contains pointers) */ - NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint &operator=(const NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint &) = delete; - /** Prevent instances of this class from being moved (As this class contains non movable objects) */ - NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint(NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint &&) = delete; - /** Prevent instances of this class from being moved (As this class contains non movable objects) */ - NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint &operator=(NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint &&) = delete; - /** Default destructor */ - ~NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint(); - /** Initialise the kernel's inputs, output - * - * @param[in] input Input tensor. Data type supported: S32 - * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[out] output Output tensor. Data type supported: Data type supported: QASYMM8 - * @param[in] result_fixedpoint_multiplier Fixed point value to be multiplied to each element of the input matrix when once the result_offset has been add - * @param[in] result_shift Number of bits to shift right the result after the fixed point multiplication - * @param[in] result_offset_after_shift Offset to be applied to result before converting it back to QASYMM8 - * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8. Defaults to the minimum possible 32-bit signed integer. - * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8, - * Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer. - */ - void configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift, int result_offset_after_shift, - int min = std::numeric_limits::lowest(), int max = std::numeric_limits::max()); - /** Static function to check if given info will lead to a valid configuration of @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint - * - * @param[in] input Input tensor. It is the output of @ref NEGEMMLowpMatrixMultiplyCore function. Data type supported: S32 - * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[in] output Output tensor. Data type supported: Data type supported: QASYMM8 - * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8. Defaults to the minimum possible 32-bit signed integer. - * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8, - * Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer. - * - * @return a status - */ - static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min = std::numeric_limits::lowest(), int max = std::numeric_limits::max()); -}; -/** Basic function to execute NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint. - * - * NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint depends on 3 parameters: - * - * result_fixedpoint_multiplier, result_shift, result_offset_after_shift - * - * The final result is: - * - * (FixedPointMul(input[i][k], result_fixedpoint_multiplier) >> result_shift) + result_offset_after_shift - * - * where FixedPointMul(x, y) is the nearest integer to the following - * mathematical expression, evaluated without overflow or intermediate rounding: - * - * (x * y) / 2^31 - * - * For more information: https://github.com/google/gemmlowp/blob/master/public/output_stages.h#L68 - * - * In case the bias tensor is provided, the final result is: - * - * ((FixedPointMul(input[i][k] + bias[k], result_fixedpoint_multiplier)) >> result_shift) + result_offset_after_shift - * - * This function calls the following kernels: - * - * -# @ref NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel - * - * @note The function accepts also 2 optional input arguments (min and max) which can be used to implement "rectified linear unit" activation functions - * after the result is shifted right by result_shift -*/ -class NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint : public INESimpleFunctionNoBorder -{ -public: - /** Constructor */ - NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint() = default; - /** Prevent instances of this class from being copied (As this class contains pointers) */ - NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint(const NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint &) = delete; - /** Prevent instances of this class from being copied (As this class contains pointers) */ - NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint &operator=(const NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint &) = delete; - /** Prevent instances of this class from being moved (As this class contains non movable objects) */ - NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint(NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint &&) = delete; - /** Prevent instances of this class from being moved (As this class contains non movable objects) */ - NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint &operator=(NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint &&) = delete; - /** Default destructor */ - ~NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint(); - /** Initialise the kernel's inputs, output - * - * @param[in] input Input tensor. Data type supported: S32 - * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[out] output Output tensor. Data type supported: Data type supported: QASYMM8_SIGNED - * @param[in] result_fixedpoint_multiplier Fixed point value to be multiplied to each element of the input matrix when once the result_offset has been add - * @param[in] result_shift Number of bits to shift right the result after the fixed point multiplication - * @param[in] result_offset_after_shift Offset to be applied to result before converting it back to QASYMM8_SIGNED - * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8_SIGNED. Defaults to the minimum possible 32-bit signed integer. - * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8_SIGNED, - * Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer. - */ - void configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift, int result_offset_after_shift, - int min = std::numeric_limits::lowest(), int max = std::numeric_limits::max()); - /** Static function to check if given info will lead to a valid configuration of @ref NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint - * - * @param[in] input Input tensor. It is the output of @ref NEGEMMLowpMatrixMultiplyCore function. Data type supported: S32 - * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[in] output Output tensor. Data type supported: Data type supported: QASYMM8_SIGNED - * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8_SIGNED. Defaults to the minimum possible 32-bit signed integer. - * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8_SIGNED, - * Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer. - * - * @return a status - */ - static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min = std::numeric_limits::lowest(), int max = std::numeric_limits::max()); -}; -/** Basic function to execute NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint. - * - * NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint depends on 2 parameters: - * - * result_fixedpoint_multiplier, result_shift - * - * The final result is: - * - * (FixedPointMul(input[i][k], result_fixedpoint_multiplier) >> result_shift) - * - * where FixedPointMul(x, y) is the nearest integer to the following - * mathematical expression, evaluated without overflow or intermediate rounding: - * - * (x * y) / 2^31 - * - * For more information: https://github.com/google/gemmlowp/blob/master/public/output_stages.h#L68 - * - * In case the bias tensor is provided, the final result is: - * - * ((FixedPointMul(input[i][k] + bias[k], result_fixedpoint_multiplier)) >> result_shift) + result_offset_after_shift - * - * This function calls the following kernels: - * - * -# @ref NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel - * - * @note The function accepts also 2 optional input arguments (min and max) which can be used to implement "rectified linear unit" activation functions - * after the result is shifted right by result_shift -*/ -class NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint : public INESimpleFunctionNoBorder -{ -public: - /** Constructor */ - NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint() = default; - /** Prevent instances of this class from being copied (As this class contains pointers) */ - NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint(const NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint &) = delete; - /** Prevent instances of this class from being copied (As this class contains pointers) */ - NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint &operator=(const NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint &) = delete; - /** Prevent instances of this class from being moved (As this class contains non movable objects) */ - NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint(NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint &&) = delete; - /** Prevent instances of this class from being moved (As this class contains non movable objects) */ - NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint &operator=(NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint &&) = delete; - /** Default destructor */ - ~NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint(); - /** Initialise the kernel's inputs, output - * - * @param[in] input Input tensor. Data type supported: S32 - * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[out] output Output tensor. Data type supported: Data type supported: QSYMM16 - * @param[in] result_fixedpoint_multiplier Fixed point value to be multiplied to each element of the input matrix when once the result_offset has been add - * @param[in] result_shift Number of bits to shift right the result after the fixed point multiplication - * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QSYMM16. Defaults to the minimum possible 32-bit signed integer. - * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QSYMM16. - * Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer. - */ - void configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift, int min = std::numeric_limits::lowest(), - int max = std::numeric_limits::max()); - /** Static function to check if given info will lead to a valid configuration of @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint - * - * @param[in] input Input tensor info. It is the output of @ref NEGEMMLowpMatrixMultiplyCore function. Data type supported: S32 - * @param[in] bias Biases tensor info. Only shared biases supported and it can be a nullptr if the addition of biases is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[in] output Output tensor info. Data type supported: Data type supported: QSYMM16 - * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QSYMM16. Defaults to the minimum possible 32-bit signed integer. - * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QSYMM16, - * Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to the maximum possible 32-bit signed integer. - * - * @return a status - */ - static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min = std::numeric_limits::lowest(), int max = std::numeric_limits::max()); -}; - /** Basic function to execute GEMMLowpQuantizeDown kernels. * - * This function calls the following kernels: + * This function calls the following operators: * - * -# @ref NEGEMMLowpQuantizeDownInt32ScaleKernel - * -# @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel - * -# @ref NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel - * -# @ref NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel + * -# @ref cpu::CpuGemmLowpOutputStage */ -class NEGEMMLowpOutputStage : public INESimpleFunctionNoBorder +class NEGEMMLowpOutputStage : public IFunction { public: /** Constructor */ - NEGEMMLowpOutputStage() = default; + NEGEMMLowpOutputStage(); /** Prevent instances of this class from being copied (As this class contains pointers) */ NEGEMMLowpOutputStage(const NEGEMMLowpOutputStage &) = delete; /** Prevent instances of this class from being copied (As this class contains pointers) */ @@ -310,6 +90,13 @@ public: * @return a status */ static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const GEMMLowpOutputStageInfo &info); + + // Inherited methods overridden: + void run() override; + +private: + struct Impl; + std::unique_ptr _impl; }; } // namespace arm_compute #endif /*ARM_COMPUTE_NEGEMMLOWPOUTPUTSTAGE_H */ diff --git a/arm_compute/runtime/NEON/functions/NELSTMLayerQuantized.h b/arm_compute/runtime/NEON/functions/NELSTMLayerQuantized.h index 2f0c753691..bcb89d997d 100644 --- a/arm_compute/runtime/NEON/functions/NELSTMLayerQuantized.h +++ b/arm_compute/runtime/NEON/functions/NELSTMLayerQuantized.h @@ -50,7 +50,7 @@ class ITensor; * This function calls the following functions/kernels: * * -# @ref NEGEMMLowpMatrixMultiplyCore Quantized matrix multiplication core. Accumulators are 32-bit integers - * -# @ref NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint Convert 32-bit integers into QSYMM16 + * -# @ref NEGEMMLowpOutputStage Convert 32-bit integers into QSYMM16 * -# @ref NETranspose Matrix transpose * -# @ref NEConcatenateLayer Tensor concatenation * -# @ref NEActivationLayer Activation functions (tanh and logistic) @@ -147,30 +147,30 @@ private: MemoryGroup _memory_group; // Functions used - NEGEMMLowpMatrixMultiplyCore _gemmlowp; - NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint _output_stage; - NETranspose _transpose_weights; - NEConcatenateLayer _concat_input_weights; - NEConcatenateLayer _concat_recurrent_weights; - NEConcatenateLayer _concat_weights; - NEConcatenateLayer _concat_inputs; - NEConcatenateLayer _concat_bias; - NEActivationLayer _sigmoid_forget_gate; - NEActivationLayer _sigmoid_input_gate; - NEActivationLayer _sigmoid_output_gate; - NEActivationLayer _tanh_modulation_gate; - NEActivationLayer _tanh_output_state; - NEArithmeticAddition _add1; - NEArithmeticAddition _add2; - NEPixelWiseMultiplication _mul1; - NEPixelWiseMultiplication _mul2; - NEPixelWiseMultiplication _mul3; - NESlice _slice_input_tensor; - NESlice _slice_forget_tensor; - NESlice _slice_cell_tensor; - NESlice _slice_output_tensor; - NEDequantizationLayer _dequantize; - NEQuantizationLayer _quantize; + NEGEMMLowpMatrixMultiplyCore _gemmlowp; + NEGEMMLowpOutputStage _output_stage; + NETranspose _transpose_weights; + NEConcatenateLayer _concat_input_weights; + NEConcatenateLayer _concat_recurrent_weights; + NEConcatenateLayer _concat_weights; + NEConcatenateLayer _concat_inputs; + NEConcatenateLayer _concat_bias; + NEActivationLayer _sigmoid_forget_gate; + NEActivationLayer _sigmoid_input_gate; + NEActivationLayer _sigmoid_output_gate; + NEActivationLayer _tanh_modulation_gate; + NEActivationLayer _tanh_output_state; + NEArithmeticAddition _add1; + NEArithmeticAddition _add2; + NEPixelWiseMultiplication _mul1; + NEPixelWiseMultiplication _mul2; + NEPixelWiseMultiplication _mul3; + NESlice _slice_input_tensor; + NESlice _slice_forget_tensor; + NESlice _slice_cell_tensor; + NESlice _slice_output_tensor; + NEDequantizationLayer _dequantize; + NEQuantizationLayer _quantize; // Tensor pointers const ITensor *_input_to_input_weights; diff --git a/arm_compute/runtime/NEON/functions/NEQLSTMLayer.h b/arm_compute/runtime/NEON/functions/NEQLSTMLayer.h index 7c2e9bc5a1..77adffd543 100644 --- a/arm_compute/runtime/NEON/functions/NEQLSTMLayer.h +++ b/arm_compute/runtime/NEON/functions/NEQLSTMLayer.h @@ -54,7 +54,7 @@ class NEGEMMLowpMatrixAReductionKernel; * -# @ref NEArithmeticSubtraction Elementwise subtraction * -# @ref NECopy Copy kernel for copying output_state_out to output * -# @ref NEGEMMLowpMatrixMultiplyCore Quantized matrix multiplication core. Accumulators are 32-bit integers - * -# @ref NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint Convert 32-bit integers into QSYMM16 + * -# @ref NEGEMMLowpOutputStage Convert 32-bit integers into QSYMM16 * -# @ref NEGEMMLowpMatrixAReductionKernel For precomputing effective biases to use * -# @ref NEPixelWiseMultiplication Elementwise multiplication * -# @ref NETranspose Transpose function for reshaping the weights diff --git a/docs/user_guide/release_version_and_change_log.dox b/docs/user_guide/release_version_and_change_log.dox index 3ffa11b045..b847761b8c 100644 --- a/docs/user_guide/release_version_and_change_log.dox +++ b/docs/user_guide/release_version_and_change_log.dox @@ -609,7 +609,7 @@ v20.05 Public major release - @ref NEReduceMean - @ref NEArgMinMaxLayer - @ref NEDeconvolutionLayer - - @ref NEGEMMLowpQuantizeDownInt32ScaleKernel + - NEGEMMLowpQuantizeDownInt32ScaleKernel - @ref CPPBoxWithNonMaximaSuppressionLimit - @ref CPPDetectionPostProcessLayer - @ref CPPPermuteKernel @@ -683,7 +683,7 @@ v20.02 Public major release - CLGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel / @ref CLGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint - New Arm® Neon™ kernels / functions: - @ref NEFill - - @ref NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel / @ref NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint + - NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel / NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint - Deprecated Arm® Neon™ functions / interfaces: - CLDepthwiseConvolutionLayer3x3 - NEDepthwiseConvolutionLayerOptimized @@ -800,7 +800,7 @@ v19.08 Public major release - NEBatchConcatenateLayerKernel - @ref NEDepthToSpaceLayerKernel / @ref NEDepthToSpaceLayer - NEDepthwiseConvolutionLayerNativeKernel - - @ref NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel + - NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel - @ref NEMeanStdDevNormalizationKernel / @ref NEMeanStdDevNormalizationLayer - @ref NESpaceToDepthLayerKernel / @ref NESpaceToDepthLayer - New OpenCL kernels / functions: @@ -1243,7 +1243,7 @@ v17.12 Public major release - arm_compute::NEHGEMMAArch64FP16Kernel - NEDepthwiseConvolutionLayer3x3Kernel / NEDepthwiseIm2ColKernel / NEGEMMMatrixVectorMultiplyKernel / NEDepthwiseVectorToTensorKernel / @ref NEDepthwiseConvolutionLayer - @ref NEGEMMLowpOffsetContributionKernel / @ref NEGEMMLowpMatrixAReductionKernel / @ref NEGEMMLowpMatrixBReductionKernel / @ref NEGEMMLowpMatrixMultiplyCore - - @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel / @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint + - NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel / NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint - NEWinogradLayer / NEWinogradLayerKernel - New OpenCL kernels / functions diff --git a/examples/neon_gemm_qasymm8.cpp b/examples/neon_gemm_qasymm8.cpp index efe1655d45..159636db1d 100644 --- a/examples/neon_gemm_qasymm8.cpp +++ b/examples/neon_gemm_qasymm8.cpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2020 Arm Limited. + * Copyright (c) 2020-2021 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -26,8 +26,8 @@ #include "arm_compute/core/utils/quantization/AsymmHelpers.h" #include "arm_compute/runtime/NEON/NEFunctions.h" #include "arm_compute/runtime/NEON/NEScheduler.h" -#include "utils/Utils.h" #include "support/ToolchainSupport.h" +#include "utils/Utils.h" #include @@ -102,10 +102,10 @@ int main(int argc, char **argv) Tensor q_dst0; Tensor q_res; Tensor q_res_output; - size_t M = 4; - size_t N = 4; - size_t K = 4; - bool default_input = true; + size_t M = 4; + size_t N = 4; + size_t K = 4; + bool default_input = true; // Parse args if(argc < 3) /* case default matrix sizes */ @@ -144,15 +144,18 @@ int main(int argc, char **argv) // Fill in: one is the identity matrix, other is sequential values // src1: Identity matrix - for(size_t i = 0; i < M * K; i++) { + for(size_t i = 0; i < M * K; i++) + { src1_ptr[i] = 0; } - for(size_t i = 0; i < M; i++) { + for(size_t i = 0; i < M; i++) + { src1_ptr[i * K + i] = 1.0f; } // src2: Sequential values matrix - for(size_t i = 0; i < K * N; i++) { + for(size_t i = 0; i < K * N; i++) + { src2_ptr[i] = i * 1.123f; } @@ -217,13 +220,22 @@ int main(int argc, char **argv) qgemm.configure(&q_src1, &q_src2, nullptr, &q_res); // Configure output stage after computing shift and multiplier parameters - NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint gemmlowp_output_stage; - int output_multiplier; - int output_shift; - float multiplier = (src1_qinfo.uniform().scale * src2_qinfo.uniform().scale) / dst0_qinfo.uniform().scale; + NEGEMMLowpOutputStage gemmlowp_output_stage; + int output_multiplier; + int output_shift; + float multiplier = (src1_qinfo.uniform().scale * src2_qinfo.uniform().scale) / dst0_qinfo.uniform().scale; quantization::calculate_quantized_multiplier_less_than_one(multiplier, &output_multiplier, &output_shift); std::cout << "(q_multiplier, q_shift) = (" << output_multiplier << ", " << output_shift << ")\n\n"; - gemmlowp_output_stage.configure(&q_res, nullptr, &q_res_output, output_multiplier, output_shift, dst0_qinfo.uniform().offset); + + GEMMLowpOutputStageInfo info; + info.type = GEMMLowpOutputStageType::QUANTIZE_DOWN_FIXEDPOINT; + info.gemmlowp_multiplier = output_multiplier; + info.gemmlowp_shift = output_shift; + info.gemmlowp_offset = dst0_qinfo.uniform().offset; + info.output_data_type = DataType::QASYMM8; + q_res_output.info()->set_data_type(DataType::QASYMM8); + q_res_output.info()->set_num_channels(1); + gemmlowp_output_stage.configure(&q_res, nullptr, &q_res_output, info); // Allocate all tensors q_src1.allocator()->allocate(); diff --git a/filelist.json b/filelist.json index c375b1daef..e30426bd19 100644 --- a/filelist.json +++ b/filelist.json @@ -160,6 +160,7 @@ "src/runtime/cpu/operators/CpuFlatten.cpp", "src/runtime/cpu/operators/CpuFloor.cpp", "src/runtime/cpu/operators/CpuGemmDirectConv2d.cpp", + "src/runtime/cpu/operators/CpuGemmLowpOutputStage.cpp", "src/runtime/cpu/operators/CpuMul.cpp", "src/runtime/cpu/operators/CpuQuantize.cpp", "src/runtime/cpu/operators/CpuReshape.cpp", @@ -197,6 +198,10 @@ "src/core/cpu/kernels/CpuElementwiseUnaryKernel.cpp", "src/core/cpu/kernels/CpuFillKernel.cpp", "src/core/cpu/kernels/CpuFloorKernel.cpp", + "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ScaleKernel.cpp", + "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp", + "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp", + "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp", "src/core/cpu/kernels/CpuMulKernel.cpp", "src/core/cpu/kernels/CpuQuantizeKernel.cpp", "src/core/cpu/kernels/CpuScaleKernel.cpp", diff --git a/src/core/NEON/NEKernels.h b/src/core/NEON/NEKernels.h index ea15f4eddd..268871a4e8 100644 --- a/src/core/NEON/NEKernels.h +++ b/src/core/NEON/NEKernels.h @@ -45,10 +45,6 @@ #include "src/core/NEON/kernels/NEGEMMLowpMatrixMultiplyKernel.h" #include "src/core/NEON/kernels/NEGEMMLowpOffsetContributionKernel.h" #include "src/core/NEON/kernels/NEGEMMLowpOffsetContributionOutputStageKernel.h" -#include "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.h" -#include "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h" -#include "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h" -#include "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h" #include "src/core/NEON/kernels/NEGEMMLowpReductionKernel.h" #include "src/core/NEON/kernels/NEGEMMMatrixAdditionKernel.h" #include "src/core/NEON/kernels/NEGEMMMatrixMultiplyKernel.h" diff --git a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.cpp b/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.cpp deleted file mode 100644 index 84365ba25b..0000000000 --- a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.cpp +++ /dev/null @@ -1,320 +0,0 @@ -/* - * Copyright (c) 2020-2021 Arm Limited. - * - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a copy - * of this software and associated documentation files (the "Software"), to - * deal in the Software without restriction, including without limitation the - * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or - * sell copies of the Software, and to permit persons to whom the Software is - * furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in all - * copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE - * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, - * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE - * SOFTWARE. - */ -#include "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.h" - -#include "arm_compute/core/Error.h" -#include "arm_compute/core/Helpers.h" -#include "arm_compute/core/ITensor.h" -#include "arm_compute/core/Types.h" -#include "arm_compute/core/Utils.h" -#include "arm_compute/core/Validate.h" -#include "arm_compute/core/Window.h" -#include "arm_compute/core/utils/quantization/AsymmHelpers.h" -#include "src/core/AccessWindowStatic.h" -#include "src/core/NEON/wrapper/wrapper.h" -#include "src/core/helpers/AutoConfiguration.h" -#include "src/core/helpers/WindowHelpers.h" - -#include -#include -#include - -namespace arm_compute -{ -Status validate_arguments(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const GEMMLowpOutputStageInfo *output_stage) -{ - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::S32); - - ARM_COMPUTE_RETURN_ERROR_ON(output_stage->gemmlowp_max_bound > std::get<1>(quantization::get_min_max_values_from_quantized_data_type(output_stage->output_data_type))); - ARM_COMPUTE_RETURN_ERROR_ON(output_stage->gemmlowp_min_bound < std::get<0>(quantization::get_min_max_values_from_quantized_data_type(output_stage->output_data_type)) - || output_stage->gemmlowp_min_bound > output_stage->gemmlowp_max_bound); - - // Check biases if exist - if(bias != nullptr) - { - ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, bias); - ARM_COMPUTE_RETURN_ERROR_ON(bias->num_dimensions() > 1); - ARM_COMPUTE_RETURN_ERROR_ON(input->dimension(0) != bias->dimension(0)); - } - - if(output->total_size() != 0) - { - if(output->data_type() != output_stage->output_data_type && (output_stage->output_data_type == DataType::QASYMM8 || output_stage->output_data_type == DataType::QASYMM8_SIGNED)) - { - ARM_COMPUTE_RETURN_ERROR_MSG("Mismatching data types"); - } - - ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(input, output); - } - - return Status{}; -} - -inline void scale_input(int32x4x4_t &in_s32, int32x4_t result_offset_s32, int32_t result_mult_int) -{ - // Add the offset terms to GEMM's result - in_s32.val[0] = vaddq_s32(in_s32.val[0], result_offset_s32); - in_s32.val[1] = vaddq_s32(in_s32.val[1], result_offset_s32); - in_s32.val[2] = vaddq_s32(in_s32.val[2], result_offset_s32); - in_s32.val[3] = vaddq_s32(in_s32.val[3], result_offset_s32); - - // Multiply by result_mult_int - in_s32.val[0] = vmulq_n_s32(in_s32.val[0], result_mult_int); - in_s32.val[1] = vmulq_n_s32(in_s32.val[1], result_mult_int); - in_s32.val[2] = vmulq_n_s32(in_s32.val[2], result_mult_int); - in_s32.val[3] = vmulq_n_s32(in_s32.val[3], result_mult_int); -} - -template -inline typename std::enable_if::value, - typename wrapper::traits::neon_vector::type>::type - convert_to_8bit(const int16x8x2_t in_s16) -{ - return wrapper::vcombine(wrapper::vqmovun(in_s16.val[0]), wrapper::vqmovun(in_s16.val[1])); -} - -template -inline typename std::enable_if::value, - typename wrapper::traits::neon_vector::type>::type - convert_to_8bit(const int16x8x2_t in_s16) -{ - return wrapper::vcombine(wrapper::vqmovn(in_s16.val[0]), wrapper::vqmovn(in_s16.val[1])); -} - -template -inline typename wrapper::traits::neon_vector::type finalize_quantization(int32x4x4_t &in_s32, int32x4_t result_shift_s32, typename wrapper::traits::neon_vector::type min, - typename wrapper::traits::neon_vector::type max) -{ - // Shift final result (negative value shift right) - in_s32.val[0] = vshlq_s32(in_s32.val[0], result_shift_s32); - in_s32.val[1] = vshlq_s32(in_s32.val[1], result_shift_s32); - in_s32.val[2] = vshlq_s32(in_s32.val[2], result_shift_s32); - in_s32.val[3] = vshlq_s32(in_s32.val[3], result_shift_s32); - - // Convert S32 to S16 - const int16x8x2_t in_s16 = - { - { - vcombine_s16(vqmovn_s32(in_s32.val[0]), vqmovn_s32(in_s32.val[1])), - vcombine_s16(vqmovn_s32(in_s32.val[2]), vqmovn_s32(in_s32.val[3])) - } - }; - - // Convert S16 to S8 or U8 - typename wrapper::traits::neon_vector::type out = convert_to_8bit(in_s16); - - out = wrapper::vmax(out, min); - out = wrapper::vmin(out, max); - - return out; -} - -class Coordinates; - -template -void NEGEMMLowpQuantizeDownInt32ScaleKernel::run(const Window &window) -{ - using VectorType = typename wrapper::traits::neon_vector::type; - - const int32x4_t result_offset_s32 = vdupq_n_s32(_output_stage->gemmlowp_offset); - const int32x4_t result_shift_s32 = vdupq_n_s32(-_output_stage->gemmlowp_shift); - const int window_step_x = 16; - const auto window_start_x = static_cast(window.x().start()); - const auto window_end_x = static_cast(window.x().end()); - - const int clamp_min = (_is_bounded_relu) ? _output_stage->gemmlowp_min_bound : std::numeric_limits::lowest(); - const int clamp_max = (_is_bounded_relu) ? _output_stage->gemmlowp_max_bound : std::numeric_limits::max(); - - VectorType min = wrapper::vdup_n(static_cast(clamp_min), wrapper::traits::vector_128_tag{}); - VectorType max = wrapper::vdup_n(static_cast(clamp_max), wrapper::traits::vector_128_tag{}); - - Window win(window); - win.set(Window::DimX, Window::Dimension(0, 1, 1)); - - Iterator in(_input, win); - Iterator out(_output, win); - - if(_bias != nullptr) - { - Window win_biases; - win_biases.set(Window::DimX, Window::Dimension(0, 1, 1)); - win_biases.set(Window::DimY, Window::Dimension(0, 1, 1)); - - Iterator bias(_bias, win_biases); - execute_window_loop(win, [&](const Coordinates &) - { - // Compute 16 elements per iteration - int x = window_start_x; - for(; x <= (window_end_x - window_step_x); x += window_step_x) - { - int32x4x4_t in_s32 = - { - { - vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 4), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 8), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 12) - } - }; - - const int32x4x4_t bias_s32 = - { - { - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 0), - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 4), - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 8), - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 12) - } - }; - - // Add the bias to GEMM's result - in_s32.val[0] = vaddq_s32(in_s32.val[0], bias_s32.val[0]); - in_s32.val[1] = vaddq_s32(in_s32.val[1], bias_s32.val[1]); - in_s32.val[2] = vaddq_s32(in_s32.val[2], bias_s32.val[2]); - in_s32.val[3] = vaddq_s32(in_s32.val[3], bias_s32.val[3]); - - // Add the offset terms to GEMM's result and multiply by result_mult_int - scale_input(in_s32, result_offset_s32, _output_stage->gemmlowp_multiplier); - - wrapper::vstore(reinterpret_cast(out.ptr() + x), finalize_quantization(in_s32, result_shift_s32, min, max)); - } - - // Compute left-over elements - for(; x < window_end_x; ++x) - { - const int bias_value = *(reinterpret_cast(bias.ptr()) + x); - int in_value = *(reinterpret_cast(in.ptr()) + x); - - // Quantize - in_value = ((in_value + bias_value + _output_stage->gemmlowp_offset) * _output_stage->gemmlowp_multiplier) >> _output_stage->gemmlowp_shift; - - // Store the result - *(out.ptr() + x) = static_cast(utility::clamp(in_value, clamp_min, clamp_max)); - } - }, - in, bias, out); - } - else - { - execute_window_loop(win, [&](const Coordinates &) - { - // Compute 16 elements per iteration - int x = window_start_x; - for(; x <= (window_end_x - window_step_x); x += window_step_x) - { - int32x4x4_t in_s32 = - { - { - vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 4), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 8), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 12) - } - }; - - // Add the offset terms to GEMM's result and multiply by result_mult_int - scale_input(in_s32, result_offset_s32, _output_stage->gemmlowp_multiplier); - - wrapper::vstore(reinterpret_cast(out.ptr() + x), finalize_quantization(in_s32, result_shift_s32, min, max)); - } - - // Compute left-over elements - for(; x < window_end_x; ++x) - { - int in_value = *(reinterpret_cast(in.ptr()) + x); - - // Quantize - in_value = ((in_value + _output_stage->gemmlowp_offset) * _output_stage->gemmlowp_multiplier) >> _output_stage->gemmlowp_shift; - - // Store the result - *(out.ptr() + x) = static_cast(utility::clamp(in_value, clamp_min, clamp_max)); - } - }, - in, out); - } -} - -NEGEMMLowpQuantizeDownInt32ScaleKernel::NEGEMMLowpQuantizeDownInt32ScaleKernel() - : _func(nullptr), _input(nullptr), _bias(nullptr), _output(nullptr), _output_stage(nullptr), _is_bounded_relu(false) -{ -} - -void NEGEMMLowpQuantizeDownInt32ScaleKernel::configure(const ITensor *input, const ITensor *bias, ITensor *output, const GEMMLowpOutputStageInfo *output_stage) -{ - // Perform validate step - ARM_COMPUTE_ERROR_ON_NULLPTR(input, output, output_stage); - - // Output auto inizialitation if not yet initialized - auto_init_if_empty(*output->info(), input->info()->clone()->set_data_type(output_stage->output_data_type)); - - ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input->info(), - (bias != nullptr) ? bias->info() : nullptr, - output->info(), - output_stage)); - - _input = input; - _bias = bias; - _output = output; - _output_stage = output_stage; - - // Configure kernel window - Window win = calculate_max_window(*input->info(), Steps()); - - INEKernel::configure(win); - - // Check if we need to clamp the result using min and max - _is_bounded_relu = ((_output_stage->gemmlowp_min_bound != _output_stage->gemmlowp_max_bound) - && !(_output_stage->gemmlowp_min_bound == std::get<0>(quantization::get_min_max_values_from_quantized_data_type(output_stage->output_data_type)) - && _output_stage->gemmlowp_max_bound == std::get<1>(quantization::get_min_max_values_from_quantized_data_type(output_stage->output_data_type)))); - if(_output_stage->output_data_type == DataType::QASYMM8) - { - _func = &NEGEMMLowpQuantizeDownInt32ScaleKernel::run; - } - else if(_output_stage->output_data_type == DataType::QASYMM8_SIGNED) - { - _func = &NEGEMMLowpQuantizeDownInt32ScaleKernel::run; - } - else - { - ARM_COMPUTE_ERROR("Data type not supported"); - } -} - -Status NEGEMMLowpQuantizeDownInt32ScaleKernel::validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const GEMMLowpOutputStageInfo *output_stage) -{ - ARM_COMPUTE_ERROR_ON_NULLPTR(input, output); - ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, bias, output, output_stage)); - - return Status{}; -} - -void NEGEMMLowpQuantizeDownInt32ScaleKernel::run(const Window &window, const ThreadInfo &info) -{ - ARM_COMPUTE_UNUSED(info); - ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); - ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(INEKernel::window(), window); - - (this->*_func)(window); -} -} // namespace arm_compute \ No newline at end of file diff --git a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.h b/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.h deleted file mode 100644 index 021ff8e2e0..0000000000 --- a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.h +++ /dev/null @@ -1,114 +0,0 @@ -/* - * Copyright (c) 2020-2021 Arm Limited. - * - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a copy - * of this software and associated documentation files (the "Software"), to - * deal in the Software without restriction, including without limitation the - * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or - * sell copies of the Software, and to permit persons to whom the Software is - * furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in all - * copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE - * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, - * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE - * SOFTWARE. - */ -#ifndef ARM_COMPUTE_NEGEMMLOWPQUANTIZEDOWNINT32SCALEKERNEL_H -#define ARM_COMPUTE_NEGEMMLOWPQUANTIZEDOWNINT32SCALEKERNEL_H - -#include "src/core/NEON/INEKernel.h" - -namespace arm_compute -{ -class ITensor; - -/** Kernel used to quantize down the int32 accumulator values of GEMMLowp to QASYMM8/QASYMM8_SIGNED - * - * This kernel takes a final int32 accumulator value (the output of @ref NEGEMMLowpMatrixMultiplyKernel), and processes it to obtain the final QASYMM8/QASYMM8_SIGNED value. - * The following computations will be performed by the kernel: - * - * -# Add offset terms to final result - * -# Multiply each entry of result by result_mult_int - * -# Add bias to final result if bias tensor is not a nullptr - * -# Shift the int32 accumulator by result_shift - * -# Clamp the value between the specified min and max bounds - * -# Clamp the resulting int32 values: - * -# -to the [0..255] range and cast to QASYMM8. - * -# -to the [-128..127] range and cast to QASYMM8_SIGNED. - * - */ -class NEGEMMLowpQuantizeDownInt32ScaleKernel : public INEKernel -{ -public: - const char *name() const override - { - return "NEGEMMLowpQuantizeDownInt32ScaleKernel"; - } - /** Constructor */ - NEGEMMLowpQuantizeDownInt32ScaleKernel(); - /** Prevent instances of this class from being copied (As this class contains pointers)*/ - NEGEMMLowpQuantizeDownInt32ScaleKernel(const NEGEMMLowpQuantizeDownInt32ScaleKernel &) = delete; - /** Prevent instances of this class from being copied (As this class contains pointers)*/ - NEGEMMLowpQuantizeDownInt32ScaleKernel &operator=(const NEGEMMLowpQuantizeDownInt32ScaleKernel &) = delete; - /** Allow instances of this class to be moved */ - NEGEMMLowpQuantizeDownInt32ScaleKernel(NEGEMMLowpQuantizeDownInt32ScaleKernel &&) = default; - /** Allow instances of this class to be moved */ - NEGEMMLowpQuantizeDownInt32ScaleKernel &operator=(NEGEMMLowpQuantizeDownInt32ScaleKernel &&) = default; - /** Default destructor */ - ~NEGEMMLowpQuantizeDownInt32ScaleKernel() = default; - /** Initialise the kernel's input and output. - * - * @param[in] input Input tensor. Data type supported: S32 - * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[out] output Output tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED - * @param[out] output_stage GEMMLowp output stage metadata. - */ - void configure(const ITensor *input, const ITensor *bias, ITensor *output, const GEMMLowpOutputStageInfo *output_stage); - /** Static function to check if given info will lead to a valid configuration of @ref NEGEMMLowpQuantizeDownInt32ScaleKernel - * - * @param[in] input Input tensor. Data type supported: S32 - * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[in] output Output tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED - * @param[out] output_stage GEMMLowp output stage metadata. - * - * @return a status - */ - static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const GEMMLowpOutputStageInfo *output_stage); - - // Inherited methods overridden: - void run(const Window &window, const ThreadInfo &info) override; - -private: - /** Template function to run the NEGEMMLowpQuantizeDownInt32ScaleKernel - * - * @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()). - */ - template - void run(const Window &window); - - /** Common signature for all the specialised NEGEMMLowpQuantizeDownInt32ScaleKernel functions - * - * @param[in] window Region on which to execute the kernel. - */ - using QuantizeDownFunctionPtr = void (NEGEMMLowpQuantizeDownInt32ScaleKernel::*)(const Window &window); - - QuantizeDownFunctionPtr _func; - const ITensor *_input; - const ITensor *_bias; - ITensor *_output; - const GEMMLowpOutputStageInfo *_output_stage; - bool _is_bounded_relu; -}; -} // namespace arm_compute - -#endif /* ARM_COMPUTE_NEGEMMLOWPQUANTIZEDOWNINT32SCALEKERNEL_H */ diff --git a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp b/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp deleted file mode 100644 index aa54b80436..0000000000 --- a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp +++ /dev/null @@ -1,234 +0,0 @@ -/* - * Copyright (c) 2019-2021 Arm Limited. - * - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a copy - * of this software and associated documentation files (the "Software"), to - * deal in the Software without restriction, including without limitation the - * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or - * sell copies of the Software, and to permit persons to whom the Software is - * furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in all - * copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE - * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, - * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE - * SOFTWARE. - */ -#include "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h" - -#include "arm_compute/core/Error.h" -#include "arm_compute/core/Helpers.h" -#include "arm_compute/core/ITensor.h" -#include "arm_compute/core/TensorInfo.h" -#include "arm_compute/core/Types.h" -#include "arm_compute/core/Utils.h" -#include "arm_compute/core/Validate.h" -#include "arm_compute/core/Window.h" -#include "arm_compute/core/utils/misc/ShapeCalculator.h" -#include "src/core/NEON/NESymm.h" -#include "src/core/helpers/AutoConfiguration.h" -#include "src/core/helpers/WindowHelpers.h" - -#include -#include -#include - -namespace arm_compute -{ -namespace -{ -Status validate_arguments(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min, int max) -{ - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::S32); - ARM_COMPUTE_RETURN_ERROR_ON(min > max); - - // Check biases if exist - if(bias != nullptr) - { - ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, bias); - ARM_COMPUTE_RETURN_ERROR_ON(bias->num_dimensions() > 1); - ARM_COMPUTE_RETURN_ERROR_ON(input->dimension(0) != bias->dimension(0)); - } - - if(output->total_size() != 0) - { - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(output, 1, DataType::QSYMM16); - ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(output, input); - } - - return Status{}; -} - -std::pair validate_and_configure_window(ITensorInfo *input, ITensorInfo *output) -{ - // Output auto inizialitation if not yet initialized - auto_init_if_empty(*output, input->clone()->set_data_type(DataType::QSYMM16)); - - // Configure kernel window - Window win = calculate_max_window(*input, Steps()); - - // NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel doesn't need padding so update_window_and_padding() can be skipped - - return std::make_pair(Status{}, win); -} -} // namespace - -class Coordinates; - -template -void NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::run(const Window &window) -{ - const int16x8_t min_s16 = vdupq_n_s16(static_cast(_min)); - const int16x8_t max_s16 = vdupq_n_s16(static_cast(_max)); - - ARM_COMPUTE_UNUSED(min_s16); - ARM_COMPUTE_UNUSED(max_s16); - - const int window_step_x = 8; - const auto window_start_x = static_cast(window.x().start()); - const auto window_end_x = static_cast(window.x().end()); - - Window win_collapsed = window.collapse_if_possible(window, Window::DimZ); - win_collapsed.set(Window::DimX, Window::Dimension(0, 1, 1)); - - Iterator in(_input, win_collapsed); - Iterator out(_output, win_collapsed); - if(_bias != nullptr) - { - Window win_biases; - win_biases.set(Window::DimX, Window::Dimension(0, 1, 1)); - win_biases.set(Window::DimY, Window::Dimension(0, 1, 1)); - - Iterator bias(_bias, win_biases); - execute_window_loop(win_collapsed, [&](const Coordinates &) - { - // Compute 16 elements per iteration - int x = window_start_x; - for(; x <= (window_end_x - window_step_x); x += window_step_x) - { - int32x4x2_t in_s32 = - { - { - vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 4) - } - }; - - const int32x4x2_t bias_s32 = - { - { - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 0), - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 4) - } - }; - - // Add the bias to GEMM's result - in_s32.val[0] = vaddq_s32(in_s32.val[0], bias_s32.val[0]); - in_s32.val[1] = vaddq_s32(in_s32.val[1], bias_s32.val[1]); - - vst1q_s16(reinterpret_cast(out.ptr()) + x, finalize_quantization_int16(in_s32, _result_fixedpoint_multiplier, _result_shift, min_s16, max_s16)); - } - - // Compute left-over elements - for(; x < window_end_x; ++x) - { - const int32_t bias_value = *(reinterpret_cast(bias.ptr()) + x); - int32_t in_value = *(reinterpret_cast(in.ptr()) + x); - - // Add bias - in_value += bias_value; - // Finalize and store the result - *(reinterpret_cast(out.ptr()) + x) = finalize_quantization_int16(in_value, _result_fixedpoint_multiplier, _result_shift, static_cast(_min), - static_cast(_max)); - } - }, - in, out, bias); - } - else - { - execute_window_loop(win_collapsed, [&](const Coordinates &) - { - // Compute 16 elements per iteration - int x = window_start_x; - for(; x <= (window_end_x - window_step_x); x += window_step_x) - { - int32x4x2_t in_s32 = - { - { - vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 4) - } - }; - - vst1q_s16(reinterpret_cast(out.ptr()) + x, finalize_quantization_int16(in_s32, _result_fixedpoint_multiplier, _result_shift, min_s16, max_s16)); - } - - // Compute left-over elements - for(; x < window_end_x; ++x) - { - const int32_t in_value = *(reinterpret_cast(in.ptr()) + x); - ARM_COMPUTE_UNUSED(in_value); - // Finalize and store the result - *(reinterpret_cast(out.ptr()) + x) = finalize_quantization_int16(in_value, _result_fixedpoint_multiplier, _result_shift, static_cast(_min), - static_cast(_max)); - } - }, - in, out); - } -} - -NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel() - : _func(nullptr), _input(nullptr), _bias(nullptr), _output(nullptr), _result_fixedpoint_multiplier(0), _result_shift(0), _min(0), _max(0) -{ -} - -void NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift, - int min, int max) -{ - // Perform validate step - ARM_COMPUTE_ERROR_ON_NULLPTR(input, output); - ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input->info(), (bias != nullptr) ? bias->info() : nullptr, output->info(), min, max)); - - _input = input; - _bias = bias; - _output = output; - _result_fixedpoint_multiplier = result_fixedpoint_multiplier; - _result_shift = result_shift; - _min = min; - _max = max; - - // Configure kernel window - auto win_config = validate_and_configure_window(input->info(), output->info()); - ARM_COMPUTE_ERROR_THROW_ON(win_config.first); - INEKernel::configure(win_config.second); - - // Check if we need to clamp the result using min and max - const bool is_bounded_relu = !(min <= -32768 && max >= 32767); - _func = is_bounded_relu ? &NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::run : &NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::run; -} - -Status NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min, int max) -{ - ARM_COMPUTE_ERROR_ON_NULLPTR(input, output); - ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, bias, output, min, max)); - ARM_COMPUTE_RETURN_ON_ERROR(validate_and_configure_window(input->clone().get(), output->clone().get()).first); - - return Status{}; -} - -void NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::run(const Window &window, const ThreadInfo &info) -{ - ARM_COMPUTE_UNUSED(info); - ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); - ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(INEKernel::window(), window); - - (this->*_func)(window); -} -} // namespace arm_compute diff --git a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h b/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h deleted file mode 100644 index b01b204a6f..0000000000 --- a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h +++ /dev/null @@ -1,118 +0,0 @@ -/* - * Copyright (c) 2019-2021 Arm Limited. - * - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a copy - * of this software and associated documentation files (the "Software"), to - * deal in the Software without restriction, including without limitation the - * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or - * sell copies of the Software, and to permit persons to whom the Software is - * furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in all - * copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE - * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, - * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE - * SOFTWARE. - */ -#ifndef ARM_COMPUTE_NEGEMMLOWPQUANTIZEDOWNINT32TOINT16SCALEBYFIXEDPOINTKERNEL_H -#define ARM_COMPUTE_NEGEMMLOWPQUANTIZEDOWNINT32TOINT16SCALEBYFIXEDPOINTKERNEL_H - -#include "src/core/NEON/INEKernel.h" - -namespace arm_compute -{ -class ITensor; - -/** Kernel used to quantize down the int32 accumulator values of GEMMLowp to QSYMM16 - * - * This kernel takes a final int32 accumulator value (the output of @ref NEGEMMLowpMatrixMultiplyKernel), and processes it to obtain the final QSYMM16 value. - * The following computations will be performed by the kernel: - * - * -# Compute fixed point multiplication between each entry of input by result_fixedpoint_multiplier - * -# Add bias to final result if bias tensor is not a nullptr - * -# Round to nearest division by a power-of-two using result_shift - * -# Clamp the value between the specified min and max bounds - * -# Clamp the resulting int32 values to the [-32768, 32767] range and cast to QSYMM16. - * - */ -class NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel : public INEKernel -{ -public: - const char *name() const override - { - return "NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel"; - } - /** Constructor */ - NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel(); - /** Prevent instances of this class from being copied (As this class contains pointers)*/ - NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel(const NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel &) = delete; - /** Prevent instances of this class from being copied (As this class contains pointers)*/ - NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel &operator=(const NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel &) = delete; - /** Allow instances of this class to be moved */ - NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel(NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel &&) = default; - /** Allow instances of this class to be moved */ - NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel &operator=(NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel &&) = default; - /** Default destructor */ - ~NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel() = default; - /** Initialise the kernel's input and output. - * - * @param[in] input Input tensor. Data type supported: S32 - * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[out] output Output tensor. Data type supported: Data type supported: QSYMM16 - * @param[in] result_fixedpoint_multiplier Fixed point value to be multiplied to each element of the input matrix when once the result_offset has been add - * @param[in] result_shift Integer value used to round to nearest division by a power-of-two the result after the fixed point multiplication - * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QSYMM16. Defaults to 0. - * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QSYMM16. - * Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to 0. - */ - void configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift, int min = 0, int max = 0); - /** Static function to check if given info will lead to a valid configuration of @ref NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel - * - * @param[in] input Input tensor info. Data type supported: S32 - * @param[in] bias Biases tensor info. Only shared biases supported and it can be a nullptr if the biases addition is not required. - * Biases are 1D tensor info with dimensions [OFM]. Data type supported: Same as @p input. - * @param[in] output Output tensor info. Data type supported: Data type supported: QSYMM16 - * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QSYMM16. Defaults to 0. - * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QSYMM16, - * Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to 0. - * - * @return a status - */ - static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min = 0, int max = 0); - - // Inherited methods overridden: - void run(const Window &window, const ThreadInfo &info) override; - -private: - /** Template function to run the NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel - * - * @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()). - */ - template - void run(const Window &window); - - /** Common signature for all the specialised NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel functions - * - * @param[in] window Region on which to execute the kernel. - */ - using QuantizeDownFunctionPtr = void (NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::*)(const Window &window); - - QuantizeDownFunctionPtr _func; - const ITensor *_input; - const ITensor *_bias; - ITensor *_output; - int _result_fixedpoint_multiplier; - int _result_shift; - int _min; - int _max; -}; -} // namespace arm_compute -#endif /* ARM_COMPUTE_NEGEMMLOWPQUANTIZEDOWNINT32TOINT16SCALEBYFIXEDPOINTKERNEL_H */ diff --git a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp b/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp deleted file mode 100644 index 9ed85e62aa..0000000000 --- a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp +++ /dev/null @@ -1,243 +0,0 @@ -/* - * Copyright (c) 2019-2021 Arm Limited. - * - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a copy - * of this software and associated documentation files (the "Software"), to - * deal in the Software without restriction, including without limitation the - * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or - * sell copies of the Software, and to permit persons to whom the Software is - * furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in all - * copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE - * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, - * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE - * SOFTWARE. - */ -#include "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h" - -#include "arm_compute/core/Error.h" -#include "arm_compute/core/Helpers.h" -#include "arm_compute/core/ITensor.h" -#include "arm_compute/core/TensorInfo.h" -#include "arm_compute/core/Types.h" -#include "arm_compute/core/Utils.h" -#include "arm_compute/core/Validate.h" -#include "arm_compute/core/Window.h" -#include "arm_compute/core/utils/misc/ShapeCalculator.h" -#include "src/core/NEON/NEAsymm.h" -#include "src/core/helpers/AutoConfiguration.h" -#include "src/core/helpers/WindowHelpers.h" - -#include -#include -#include - -namespace arm_compute -{ -namespace -{ -Status validate_arguments(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min, int max) -{ - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::S32); - ARM_COMPUTE_RETURN_ERROR_ON(min > max); - - // Check biases if exist - if(bias != nullptr) - { - ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, bias); - ARM_COMPUTE_RETURN_ERROR_ON(bias->num_dimensions() > 1); - ARM_COMPUTE_RETURN_ERROR_ON(input->dimension(0) != bias->dimension(0)); - } - - if(output->total_size() != 0) - { - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(output, 1, DataType::QASYMM8_SIGNED); - ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(output, input); - } - - return Status{}; -} - -std::pair validate_and_configure_window(ITensorInfo *input, ITensorInfo *output) -{ - // Output auto initialization if not yet initialized - auto_init_if_empty(*output, input->clone()->set_data_type(DataType::QASYMM8_SIGNED)); - - // Configure kernel window - Window win = calculate_max_window(*input, Steps()); - - // NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel doesn't need padding so update_window_and_padding() can be skipped - - return std::make_pair(Status{}, win); -} -} // namespace - -template -void NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::run(const Window &window) -{ - const int32x4_t result_offset_after_shift_s32 = vdupq_n_s32(_result_offset_after_shift); - const int8x16_t min_s8 = vdupq_n_s8(static_cast(_min)); - const int8x16_t max_s8 = vdupq_n_s8(static_cast(_max)); - - ARM_COMPUTE_UNUSED(min_s8, max_s8); - - const int window_step_x = 16; - const auto window_start_x = static_cast(window.x().start()); - const auto window_end_x = static_cast(window.x().end()); - - Window win_collapsed = window.collapse_if_possible(window, Window::DimZ); - win_collapsed.set(Window::DimX, Window::Dimension(0, 1, 1)); - - Iterator in(_input, win_collapsed); - Iterator out(_output, win_collapsed); - if(_bias != nullptr) - { - Window win_biases; - win_biases.set(Window::DimX, Window::Dimension(0, 1, 1)); - win_biases.set(Window::DimY, Window::Dimension(0, 1, 1)); - - Iterator bias(_bias, win_biases); - execute_window_loop(win_collapsed, [&](const Coordinates &) - { - // Compute 16 elements per iteration - int x = window_start_x; - for(; x <= (window_end_x - window_step_x); x += window_step_x) - { - int32x4x4_t in_s32 = - { - { - vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 4), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 8), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 12) - } - }; - - const int32x4x4_t bias_s32 = - { - { - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 0), - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 4), - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 8), - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 12) - } - }; - - // Add the bias to GEMM's result - in_s32.val[0] = vaddq_s32(in_s32.val[0], bias_s32.val[0]); - in_s32.val[1] = vaddq_s32(in_s32.val[1], bias_s32.val[1]); - in_s32.val[2] = vaddq_s32(in_s32.val[2], bias_s32.val[2]); - in_s32.val[3] = vaddq_s32(in_s32.val[3], bias_s32.val[3]); - - vst1q_s8(reinterpret_cast(out.ptr() + x), - finalize_quantization(in_s32, _result_fixedpoint_multiplier, _result_shift, result_offset_after_shift_s32, min_s8, max_s8, is_bounded_relu)); - } - - // Compute left-over elements - for(; x < window_end_x; ++x) - { - const int32_t bias_value = *(reinterpret_cast(bias.ptr()) + x); - int32_t in_value = *(reinterpret_cast(in.ptr()) + x); - - // Add bias - in_value += bias_value; - // Finalize and store the result - *reinterpret_cast(out.ptr() + x) = finalize_quantization(in_value, _result_fixedpoint_multiplier, _result_shift, _result_offset_after_shift, - static_cast(_min), static_cast(_max), is_bounded_relu); - } - }, - in, out, bias); - } - else - { - execute_window_loop(win_collapsed, [&](const Coordinates &) - { - // Compute 16 elements per iteration - int x = window_start_x; - for(; x <= (window_end_x - window_step_x); x += window_step_x) - { - int32x4x4_t in_s32 = - { - { - vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 4), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 8), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 12) - } - }; - - vst1q_s8(reinterpret_cast(out.ptr() + x), - finalize_quantization(in_s32, _result_fixedpoint_multiplier, _result_shift, result_offset_after_shift_s32, min_s8, max_s8, is_bounded_relu)); - } - - // Compute left-over elements - for(; x < window_end_x; ++x) - { - const int32_t in_value = *(reinterpret_cast(in.ptr()) + x); - - // Finalize and store the result - *reinterpret_cast(out.ptr() + x) = finalize_quantization(in_value, _result_fixedpoint_multiplier, _result_shift, _result_offset_after_shift, - static_cast(_min), static_cast(_max), is_bounded_relu); - } - }, - in, out); - } -} - -NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel() - : _func(nullptr), _input(nullptr), _bias(nullptr), _output(nullptr), _result_fixedpoint_multiplier(0), _result_shift(0), _result_offset_after_shift(0), _min(0), _max(0) -{ -} - -void NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift, - int result_offset_after_shift, int min, int max) -{ - // Perform validate step - ARM_COMPUTE_ERROR_ON_NULLPTR(input, output); - ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input->info(), (bias != nullptr) ? bias->info() : nullptr, output->info(), min, max)); - - _input = input; - _bias = bias; - _output = output; - _result_fixedpoint_multiplier = result_fixedpoint_multiplier; - _result_shift = result_shift; - _result_offset_after_shift = result_offset_after_shift; - _min = min; - _max = max; - - // Configure kernel window - auto win_config = validate_and_configure_window(input->info(), output->info()); - ARM_COMPUTE_ERROR_THROW_ON(win_config.first); - INEKernel::configure(win_config.second); - - // Check if we need to clamp the result using min and max - const bool is_bounded_relu = !(min <= -128 && max >= 127); - _func = is_bounded_relu ? &NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::run : &NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::run; -} - -Status NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min, int max) -{ - ARM_COMPUTE_ERROR_ON_NULLPTR(input, output); - ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, bias, output, min, max)); - ARM_COMPUTE_RETURN_ON_ERROR(validate_and_configure_window(input->clone().get(), output->clone().get()).first); - - return Status{}; -} - -void NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::run(const Window &window, const ThreadInfo &info) -{ - ARM_COMPUTE_UNUSED(info); - ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); - ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(INEKernel::window(), window); - - (this->*_func)(window); -} -} // namespace arm_compute diff --git a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h b/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h deleted file mode 100644 index 9e7dc2f599..0000000000 --- a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h +++ /dev/null @@ -1,121 +0,0 @@ -/* - * Copyright (c) 2019-2021 Arm Limited. - * - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a copy - * of this software and associated documentation files (the "Software"), to - * deal in the Software without restriction, including without limitation the - * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or - * sell copies of the Software, and to permit persons to whom the Software is - * furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in all - * copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE - * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, - * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE - * SOFTWARE. - */ -#ifndef ARM_COMPUTE_NEGEMMLOWPQUANTIZEDOWNINT32TOINT8SCALEBYFIXEDPOINTKERNEL_H -#define ARM_COMPUTE_NEGEMMLOWPQUANTIZEDOWNINT32TOINT8SCALEBYFIXEDPOINTKERNEL_H - -#include "src/core/NEON/INEKernel.h" - -namespace arm_compute -{ -class ITensor; - -/** Kernel used to quantize down the int32 accumulator values of GEMMLowp to QASYMM8_SIGNED - * - * This kernel takes a final int32 accumulator value (the output of @ref NEGEMMLowpMatrixMultiplyKernel), and processes it to obtain the final QASYMM8_SIGNED value. - * The following computations will be performed by the kernel: - * - * -# Compute fixed point multiplication between each entry of input by result_fixedpoint_multiplier - * -# Add bias to final result if bias tensor is not a nullptr - * -# Round to nearest division by a power-of-two using result_shift - * -# Add offset to each result - * -# Clamp the value between the specified min and max bounds - * -# Clamp the resulting int32 values to the [-128..127] range and cast to QASYMM8_SIGNED. - * - */ -class NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel : public INEKernel -{ -public: - const char *name() const override - { - return "NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel"; - } - /** Constructor */ - NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel(); - /** Prevent instances of this class from being copied (As this class contains pointers)*/ - NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel(const NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel &) = delete; - /** Prevent instances of this class from being copied (As this class contains pointers)*/ - NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel &operator=(const NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel &) = delete; - /** Allow instances of this class to be moved */ - NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel(NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel &&) = default; - /** Allow instances of this class to be moved */ - NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel &operator=(NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel &&) = default; - /** Default destructor */ - ~NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel() = default; - /** Initialise the kernel's input and output. - * - * @param[in] input Input tensor. Data type supported: S32 - * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[out] output Output tensor. Data type supported: Data type supported: QASYMM8_SIGNED - * @param[in] result_fixedpoint_multiplier Fixed point value to be multiplied to each element of the input matrix when once the result_offset has been add - * @param[in] result_shift Integer value used to round to nearest division by a power-of-two the result after the fixed point multiplication - * @param[in] result_offset_after_shift Offset to be applied to result before converting it back to QASYMM8_SIGNED - * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8_SIGNED - * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8_SIGNED, - * Along with @p min, this value can be used to implement "rectified linear unit" activation functions - */ - void configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift, int result_offset_after_shift, int min = 0, int max = 0); - /** Static function to check if given info will lead to a valid configuration of @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel - * - * @param[in] input Input tensor. Data type supported: S32 - * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[in] output Output tensor. Data type supported: Data type supported: QASYMM8_SIGNED - * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8_SIGNED - * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8_SIGNED, - * Along with @p min, this value can be used to implement "rectified linear unit" activation functions - * - * @return a status - */ - static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min = 0, int max = 0); - - // Inherited methods overridden: - void run(const Window &window, const ThreadInfo &info) override; - -private: - /** Template function to run the NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel - * - * @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()). - */ - template - void run(const Window &window); - - /** Common signature for all the specialised NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel functions - * - * @param[in] window Region on which to execute the kernel. - */ - using QuantizeDownFunctionPtr = void (NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::*)(const Window &window); - - QuantizeDownFunctionPtr _func; - const ITensor *_input; - const ITensor *_bias; - ITensor *_output; - int _result_fixedpoint_multiplier; - int _result_shift; - int _result_offset_after_shift; - int _min; - int _max; -}; -} // namespace arm_compute -#endif /* ARM_COMPUTE_NEGEMMLOWPQUANTIZEDOWNINT32TOINT8SCALEBYFIXEDPOINTKERNEL_H */ diff --git a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp b/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp deleted file mode 100644 index 83ca6f944d..0000000000 --- a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp +++ /dev/null @@ -1,245 +0,0 @@ -/* - * Copyright (c) 2017-2021 Arm Limited. - * - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a copy - * of this software and associated documentation files (the "Software"), to - * deal in the Software without restriction, including without limitation the - * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or - * sell copies of the Software, and to permit persons to whom the Software is - * furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in all - * copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE - * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, - * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE - * SOFTWARE. - */ -#include "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h" - -#include "arm_compute/core/Error.h" -#include "arm_compute/core/Helpers.h" -#include "arm_compute/core/ITensor.h" -#include "arm_compute/core/TensorInfo.h" -#include "arm_compute/core/Types.h" -#include "arm_compute/core/Utils.h" -#include "arm_compute/core/Validate.h" -#include "arm_compute/core/Window.h" -#include "arm_compute/core/utils/misc/ShapeCalculator.h" -#include "src/core/NEON/NEAsymm.h" -#include "src/core/helpers/AutoConfiguration.h" -#include "src/core/helpers/WindowHelpers.h" - -#include -#include -#include - -namespace arm_compute -{ -namespace -{ -Status validate_arguments(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min, int max) -{ - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::S32); - ARM_COMPUTE_RETURN_ERROR_ON(min > max); - - // Check biases if exist - if(bias != nullptr) - { - ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, bias); - ARM_COMPUTE_RETURN_ERROR_ON(bias->num_dimensions() > 1); - ARM_COMPUTE_RETURN_ERROR_ON(input->dimension(0) != bias->dimension(0)); - } - - if(output->total_size() != 0) - { - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(output, 1, DataType::QASYMM8); - ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(output, input); - } - - return Status{}; -} - -std::pair validate_and_configure_window(ITensorInfo *input, ITensorInfo *output) -{ - // Output auto inizialitation if not yet initialized - auto_init_if_empty(*output, input->clone()->set_data_type(DataType::QASYMM8)); - - // Configure kernel window - Window win = calculate_max_window(*input, Steps()); - - // NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel doesn't need padding so update_window_and_padding() can be skipped - - return std::make_pair(Status{}, win); -} -} // namespace - -namespace arm_compute -{ -class Coordinates; -} // namespace arm_compute - -template -void NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::run(const Window &window) -{ - const int32x4_t result_offset_after_shift_s32 = vdupq_n_s32(_result_offset_after_shift); - const uint8x16_t min_u8 = vdupq_n_u8(static_cast(_min)); - const uint8x16_t max_u8 = vdupq_n_u8(static_cast(_max)); - - ARM_COMPUTE_UNUSED(min_u8); - ARM_COMPUTE_UNUSED(max_u8); - - const int window_step_x = 16; - const auto window_start_x = static_cast(window.x().start()); - const auto window_end_x = static_cast(window.x().end()); - - Window win_collapsed = window.collapse_if_possible(window, Window::DimZ); - win_collapsed.set(Window::DimX, Window::Dimension(0, 1, 1)); - - Iterator in(_input, win_collapsed); - Iterator out(_output, win_collapsed); - if(_bias != nullptr) - { - Window win_biases; - win_biases.set(Window::DimX, Window::Dimension(0, 1, 1)); - win_biases.set(Window::DimY, Window::Dimension(0, 1, 1)); - - Iterator bias(_bias, win_biases); - execute_window_loop(win_collapsed, [&](const Coordinates &) - { - // Compute 16 elements per iteration - int x = window_start_x; - for(; x <= (window_end_x - window_step_x); x += window_step_x) - { - int32x4x4_t in_s32 = - { - { - vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 4), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 8), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 12) - } - }; - - const int32x4x4_t bias_s32 = - { - { - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 0), - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 4), - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 8), - vld1q_s32(reinterpret_cast(bias.ptr()) + x + 12) - } - }; - - // Add the bias to GEMM's result - in_s32.val[0] = vaddq_s32(in_s32.val[0], bias_s32.val[0]); - in_s32.val[1] = vaddq_s32(in_s32.val[1], bias_s32.val[1]); - in_s32.val[2] = vaddq_s32(in_s32.val[2], bias_s32.val[2]); - in_s32.val[3] = vaddq_s32(in_s32.val[3], bias_s32.val[3]); - - vst1q_u8(out.ptr() + x, finalize_quantization(in_s32, _result_fixedpoint_multiplier, _result_shift, result_offset_after_shift_s32, min_u8, max_u8, is_bounded_relu)); - } - - // Compute left-over elements - for(; x < window_end_x; ++x) - { - const int32_t bias_value = *(reinterpret_cast(bias.ptr()) + x); - int32_t in_value = *(reinterpret_cast(in.ptr()) + x); - - // Add bias - in_value += bias_value; - // Finalize and store the result - *(out.ptr() + x) = finalize_quantization(in_value, _result_fixedpoint_multiplier, _result_shift, _result_offset_after_shift, static_cast(_min), static_cast(_max), is_bounded_relu); - } - }, - in, out, bias); - } - else - { - execute_window_loop(win_collapsed, [&](const Coordinates &) - { - // Compute 16 elements per iteration - int x = window_start_x; - for(; x <= (window_end_x - window_step_x); x += window_step_x) - { - int32x4x4_t in_s32 = - { - { - vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 4), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 8), - vld1q_s32(reinterpret_cast(in.ptr()) + x + 12) - } - }; - - vst1q_u8(out.ptr() + x, finalize_quantization(in_s32, _result_fixedpoint_multiplier, _result_shift, result_offset_after_shift_s32, min_u8, max_u8, is_bounded_relu)); - } - - // Compute left-over elements - for(; x < window_end_x; ++x) - { - const int32_t in_value = *(reinterpret_cast(in.ptr()) + x); - - // Finalize and store the result - *(out.ptr() + x) = finalize_quantization(in_value, _result_fixedpoint_multiplier, _result_shift, _result_offset_after_shift, static_cast(_min), static_cast(_max), is_bounded_relu); - } - }, - in, out); - } -} - -NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel() - : _func(nullptr), _input(nullptr), _bias(nullptr), _output(nullptr), _result_fixedpoint_multiplier(0), _result_shift(0), _result_offset_after_shift(0), _min(0), _max(0) -{ -} - -void NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift, - int result_offset_after_shift, int min, int max) -{ - // Perform validate step - ARM_COMPUTE_ERROR_ON_NULLPTR(input, output); - ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input->info(), (bias != nullptr) ? bias->info() : nullptr, output->info(), min, max)); - - _input = input; - _bias = bias; - _output = output; - _result_fixedpoint_multiplier = result_fixedpoint_multiplier; - _result_shift = result_shift; - _result_offset_after_shift = result_offset_after_shift; - _min = min; - _max = max; - - // Configure kernel window - auto win_config = validate_and_configure_window(input->info(), output->info()); - ARM_COMPUTE_ERROR_THROW_ON(win_config.first); - INEKernel::configure(win_config.second); - - // Check if we need to clamp the result using min and max - const bool is_bounded_relu = !(min <= 0 && max >= 255); - _func = is_bounded_relu ? &NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::run : &NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::run; -} - -Status NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min, int max) -{ - ARM_COMPUTE_ERROR_ON_NULLPTR(input, output); - ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, bias, output, min, max)); - ARM_COMPUTE_RETURN_ON_ERROR(validate_and_configure_window(input->clone().get(), output->clone().get()).first); - - return Status{}; -} - -void NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::run(const Window &window, const ThreadInfo &info) -{ - ARM_COMPUTE_UNUSED(info); - ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); - ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(INEKernel::window(), window); - - (this->*_func)(window); -} -} // namespace arm_compute diff --git a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h b/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h deleted file mode 100644 index def0573967..0000000000 --- a/src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h +++ /dev/null @@ -1,121 +0,0 @@ -/* - * Copyright (c) 2017-2021 Arm Limited. - * - * SPDX-License-Identifier: MIT - * - * Permission is hereby granted, free of charge, to any person obtaining a copy - * of this software and associated documentation files (the "Software"), to - * deal in the Software without restriction, including without limitation the - * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or - * sell copies of the Software, and to permit persons to whom the Software is - * furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in all - * copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE - * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, - * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE - * SOFTWARE. - */ -#ifndef ARM_COMPUTE_NEGEMMLOWPQUANTIZEDOWNINT32TOUINT8SCALEBYFIXEDPOINTKERNEL_H -#define ARM_COMPUTE_NEGEMMLOWPQUANTIZEDOWNINT32TOUINT8SCALEBYFIXEDPOINTKERNEL_H - -#include "src/core/NEON/INEKernel.h" - -namespace arm_compute -{ -class ITensor; - -/** Kernel used to quantize down the int32 accumulator values of GEMMLowp to QASYMM8 - * - * This kernel takes a final int32 accumulator value (the output of @ref NEGEMMLowpMatrixMultiplyKernel), and processes it to obtain the final QASYMM8 value. - * The following computations will be performed by the kernel: - * - * -# Compute fixed point multiplication between each entry of input by result_fixedpoint_multiplier - * -# Add bias to final result if bias tensor is not a nullptr - * -# Round to nearest division by a power-of-two using result_shift - * -# Add offset to each result - * -# Clamp the value between the specified min and max bounds - * -# Clamp the resulting int32 values to the [0..255] range and cast to QASYMM8. - * - */ -class NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel : public INEKernel -{ -public: - const char *name() const override - { - return "NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel"; - } - /** Constructor */ - NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel(); - /** Prevent instances of this class from being copied (As this class contains pointers)*/ - NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel(const NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel &) = delete; - /** Prevent instances of this class from being copied (As this class contains pointers)*/ - NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel &operator=(const NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel &) = delete; - /** Allow instances of this class to be moved */ - NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel(NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel &&) = default; - /** Allow instances of this class to be moved */ - NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel &operator=(NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel &&) = default; - /** Default destructor */ - ~NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel() = default; - /** Initialise the kernel's input and output. - * - * @param[in] input Input tensor. Data type supported: S32 - * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[out] output Output tensor. Data type supported: Data type supported: QASYMM8 - * @param[in] result_fixedpoint_multiplier Fixed point value to be multiplied to each element of the input matrix when once the result_offset has been add - * @param[in] result_shift Integer value used to round to nearest division by a power-of-two the result after the fixed point multiplication - * @param[in] result_offset_after_shift Offset to be applied to result before converting it back to QASYMM8 - * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8 - * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8, - * Along with @p min, this value can be used to implement "rectified linear unit" activation functions - */ - void configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift, int result_offset_after_shift, int min = 0, int max = 0); - /** Static function to check if given info will lead to a valid configuration of @ref NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel - * - * @param[in] input Input tensor. Data type supported: S32 - * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required. - * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. - * @param[in] output Output tensor. Data type supported: Data type supported: QASYMM8 - * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8 - * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8, - * Along with @p min, this value can be used to implement "rectified linear unit" activation functions - * - * @return a status - */ - static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min = 0, int max = 0); - - // Inherited methods overridden: - void run(const Window &window, const ThreadInfo &info) override; - -private: - /** Template function to run the NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel - * - * @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()). - */ - template - void run(const Window &window); - - /** Common signature for all the specialised NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel functions - * - * @param[in] window Region on which to execute the kernel. - */ - using QuantizeDownFunctionPtr = void (NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::*)(const Window &window); - - QuantizeDownFunctionPtr _func; - const ITensor *_input; - const ITensor *_bias; - ITensor *_output; - int _result_fixedpoint_multiplier; - int _result_shift; - int _result_offset_after_shift; - int _min; - int _max; -}; -} // namespace arm_compute -#endif /* ARM_COMPUTE_NEGEMMLOWPQUANTIZEDOWNINT32TOUINT8SCALEBYFIXEDPOINTKERNEL_H */ diff --git a/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ScaleKernel.cpp b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ScaleKernel.cpp new file mode 100644 index 0000000000..f1c797244a --- /dev/null +++ b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ScaleKernel.cpp @@ -0,0 +1,326 @@ +/* + * Copyright (c) 2020-2021 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#include "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ScaleKernel.h" + +#include "arm_compute/core/Error.h" +#include "arm_compute/core/Helpers.h" +#include "arm_compute/core/ITensor.h" +#include "arm_compute/core/Types.h" +#include "arm_compute/core/Utils.h" +#include "arm_compute/core/Validate.h" +#include "arm_compute/core/Window.h" +#include "arm_compute/core/utils/quantization/AsymmHelpers.h" +#include "src/core/AccessWindowStatic.h" +#include "src/core/NEON/wrapper/wrapper.h" +#include "src/core/helpers/AutoConfiguration.h" +#include "src/core/helpers/WindowHelpers.h" + +#include + +namespace arm_compute +{ +namespace cpu +{ +namespace kernels +{ +namespace +{ +Status validate_arguments(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, const GEMMLowpOutputStageInfo *output_stage) +{ + ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst); + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(src, 1, DataType::S32); + + ARM_COMPUTE_RETURN_ERROR_ON(output_stage->gemmlowp_max_bound > std::get<1>(quantization::get_min_max_values_from_quantized_data_type(output_stage->output_data_type))); + ARM_COMPUTE_RETURN_ERROR_ON(output_stage->gemmlowp_min_bound < std::get<0>(quantization::get_min_max_values_from_quantized_data_type(output_stage->output_data_type)) + || output_stage->gemmlowp_min_bound > output_stage->gemmlowp_max_bound); + + // Check biases if exist + if(bias != nullptr) + { + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(src, bias); + ARM_COMPUTE_RETURN_ERROR_ON(bias->num_dimensions() > 1); + ARM_COMPUTE_RETURN_ERROR_ON(src->dimension(0) != bias->dimension(0)); + } + + if(dst->total_size() != 0) + { + if(dst->data_type() != output_stage->output_data_type && (output_stage->output_data_type == DataType::QASYMM8 || output_stage->output_data_type == DataType::QASYMM8_SIGNED)) + { + ARM_COMPUTE_RETURN_ERROR_MSG("Mismatching data types"); + } + + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(src, dst); + } + + return Status{}; +} + +inline void scale_input(int32x4x4_t &in_s32, int32x4_t result_offset_s32, int32_t result_mult_int) +{ + // Add the offset terms to GEMM's result + in_s32.val[0] = vaddq_s32(in_s32.val[0], result_offset_s32); + in_s32.val[1] = vaddq_s32(in_s32.val[1], result_offset_s32); + in_s32.val[2] = vaddq_s32(in_s32.val[2], result_offset_s32); + in_s32.val[3] = vaddq_s32(in_s32.val[3], result_offset_s32); + + // Multiply by result_mult_int + in_s32.val[0] = vmulq_n_s32(in_s32.val[0], result_mult_int); + in_s32.val[1] = vmulq_n_s32(in_s32.val[1], result_mult_int); + in_s32.val[2] = vmulq_n_s32(in_s32.val[2], result_mult_int); + in_s32.val[3] = vmulq_n_s32(in_s32.val[3], result_mult_int); +} + +template +inline typename std::enable_if::value, + typename wrapper::traits::neon_vector::type>::type + convert_to_8bit(const int16x8x2_t in_s16) +{ + return wrapper::vcombine(wrapper::vqmovun(in_s16.val[0]), wrapper::vqmovun(in_s16.val[1])); +} + +template +inline typename std::enable_if::value, + typename wrapper::traits::neon_vector::type>::type + convert_to_8bit(const int16x8x2_t in_s16) +{ + return wrapper::vcombine(wrapper::vqmovn(in_s16.val[0]), wrapper::vqmovn(in_s16.val[1])); +} + +template +inline typename wrapper::traits::neon_vector::type finalize_quantization(int32x4x4_t &in_s32, int32x4_t result_shift_s32, typename wrapper::traits::neon_vector::type min, + typename wrapper::traits::neon_vector::type max) +{ + // Shift final result (negative value shift right) + in_s32.val[0] = vshlq_s32(in_s32.val[0], result_shift_s32); + in_s32.val[1] = vshlq_s32(in_s32.val[1], result_shift_s32); + in_s32.val[2] = vshlq_s32(in_s32.val[2], result_shift_s32); + in_s32.val[3] = vshlq_s32(in_s32.val[3], result_shift_s32); + + // Convert S32 to S16 + const int16x8x2_t in_s16 = + { + { + vcombine_s16(vqmovn_s32(in_s32.val[0]), vqmovn_s32(in_s32.val[1])), + vcombine_s16(vqmovn_s32(in_s32.val[2]), vqmovn_s32(in_s32.val[3])) + } + }; + + // Convert S16 to S8 or U8 + typename wrapper::traits::neon_vector::type out = convert_to_8bit(in_s16); + + out = wrapper::vmax(out, min); + out = wrapper::vmin(out, max); + + return out; +} +} // namespace + +template +void CpuGemmLowpQuantizeDownInt32ScaleKernel::run_internal(const ITensor *src, const ITensor *bias, ITensor *dst, const Window &window) +{ + using VectorType = typename wrapper::traits::neon_vector::type; + + const int32x4_t result_offset_s32 = vdupq_n_s32(_output_stage->gemmlowp_offset); + const int32x4_t result_shift_s32 = vdupq_n_s32(-_output_stage->gemmlowp_shift); + const int window_step_x = 16; + const auto window_start_x = static_cast(window.x().start()); + const auto window_end_x = static_cast(window.x().end()); + + const int clamp_min = (_is_bounded_relu) ? _output_stage->gemmlowp_min_bound : std::numeric_limits::lowest(); + const int clamp_max = (_is_bounded_relu) ? _output_stage->gemmlowp_max_bound : std::numeric_limits::max(); + + VectorType min = wrapper::vdup_n(static_cast(clamp_min), wrapper::traits::vector_128_tag{}); + VectorType max = wrapper::vdup_n(static_cast(clamp_max), wrapper::traits::vector_128_tag{}); + + Window win(window); + win.set(Window::DimX, Window::Dimension(0, 1, 1)); + + Iterator in(src, win); + Iterator out(dst, win); + + if(bias != nullptr) + { + Window win_biases; + win_biases.set(Window::DimX, Window::Dimension(0, 1, 1)); + win_biases.set(Window::DimY, Window::Dimension(0, 1, 1)); + + Iterator bias_i(bias, win_biases); + execute_window_loop(win, [&](const Coordinates &) + { + // Compute 16 elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) + { + int32x4x4_t in_s32 = + { + { + vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 4), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 8), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 12) + } + }; + + const int32x4x4_t bias_s32 = + { + { + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 0), + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 4), + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 8), + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 12) + } + }; + + // Add the bias to GEMM's result + in_s32.val[0] = vaddq_s32(in_s32.val[0], bias_s32.val[0]); + in_s32.val[1] = vaddq_s32(in_s32.val[1], bias_s32.val[1]); + in_s32.val[2] = vaddq_s32(in_s32.val[2], bias_s32.val[2]); + in_s32.val[3] = vaddq_s32(in_s32.val[3], bias_s32.val[3]); + + // Add the offset terms to GEMM's result and multiply by result_mult_int + scale_input(in_s32, result_offset_s32, _output_stage->gemmlowp_multiplier); + + wrapper::vstore(reinterpret_cast(out.ptr() + x), finalize_quantization(in_s32, result_shift_s32, min, max)); + } + + // Compute left-over elements + for(; x < window_end_x; ++x) + { + const int bias_value = *(reinterpret_cast(bias_i.ptr()) + x); + int in_value = *(reinterpret_cast(in.ptr()) + x); + + // Quantize + in_value = ((in_value + bias_value + _output_stage->gemmlowp_offset) * _output_stage->gemmlowp_multiplier) >> _output_stage->gemmlowp_shift; + + // Store the result + *(out.ptr() + x) = static_cast(utility::clamp(in_value, clamp_min, clamp_max)); + } + }, + in, bias_i, out); + } + else + { + execute_window_loop(win, [&](const Coordinates &) + { + // Compute 16 elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) + { + int32x4x4_t in_s32 = + { + { + vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 4), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 8), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 12) + } + }; + + // Add the offset terms to GEMM's result and multiply by result_mult_int + scale_input(in_s32, result_offset_s32, _output_stage->gemmlowp_multiplier); + + wrapper::vstore(reinterpret_cast(out.ptr() + x), finalize_quantization(in_s32, result_shift_s32, min, max)); + } + + // Compute left-over elements + for(; x < window_end_x; ++x) + { + int in_value = *(reinterpret_cast(in.ptr()) + x); + + // Quantize + in_value = ((in_value + _output_stage->gemmlowp_offset) * _output_stage->gemmlowp_multiplier) >> _output_stage->gemmlowp_shift; + + // Store the result + *(out.ptr() + x) = static_cast(utility::clamp(in_value, clamp_min, clamp_max)); + } + }, + in, out); + } +} + +void CpuGemmLowpQuantizeDownInt32ScaleKernel::configure(ITensorInfo *src, ITensorInfo *bias, ITensorInfo *dst, const GEMMLowpOutputStageInfo *output_stage) +{ + ARM_COMPUTE_UNUSED(bias); + // Perform validate step + ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst, output_stage); + + // Output auto inizialitation if not yet initialized + auto_init_if_empty(*dst, src->clone()->set_data_type(output_stage->output_data_type)); + + ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(src, + bias, + dst, + output_stage)); + + _output_stage = output_stage; + + // Configure kernel window + Window win = calculate_max_window(*src, Steps()); + + ICpuKernel::configure(win); + + // Check if we need to clamp the result using min and max + _is_bounded_relu = ((_output_stage->gemmlowp_min_bound != _output_stage->gemmlowp_max_bound) + && !(_output_stage->gemmlowp_min_bound == std::get<0>(quantization::get_min_max_values_from_quantized_data_type(output_stage->output_data_type)) + && _output_stage->gemmlowp_max_bound == std::get<1>(quantization::get_min_max_values_from_quantized_data_type(output_stage->output_data_type)))); + if(_output_stage->output_data_type == DataType::QASYMM8) + { + _func = &CpuGemmLowpQuantizeDownInt32ScaleKernel::run_internal; + } + else if(_output_stage->output_data_type == DataType::QASYMM8_SIGNED) + { + _func = &CpuGemmLowpQuantizeDownInt32ScaleKernel::run_internal; + } + else + { + ARM_COMPUTE_ERROR("Data type not supported"); + } +} + +Status CpuGemmLowpQuantizeDownInt32ScaleKernel::validate(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, const GEMMLowpOutputStageInfo *output_stage) +{ + ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(src, bias, dst, output_stage)); + return Status{}; +} + +void CpuGemmLowpQuantizeDownInt32ScaleKernel::run_op(ITensorPack &tensors, const Window &window, const ThreadInfo &info) +{ + ARM_COMPUTE_UNUSED(info); + ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); + ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(ICpuKernel::window(), window); + ARM_COMPUTE_ERROR_ON_MSG(tensors.empty(), "No inputs provided"); + + auto src = tensors.get_const_tensor(TensorType::ACL_SRC); + auto bias = tensors.get_const_tensor(TensorType::ACL_BIAS); + auto dst = tensors.get_tensor(TensorType::ACL_DST); + (this->*_func)(src, bias, dst, window); +} + +const char *CpuGemmLowpQuantizeDownInt32ScaleKernel::name() const +{ + return "CpuGemmLowpQuantizeDownInt32ScaleKernel"; +} +} // namespace kernels +} // namespace cpu +} // namespace arm_compute \ No newline at end of file diff --git a/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ScaleKernel.h b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ScaleKernel.h new file mode 100644 index 0000000000..646242dc7e --- /dev/null +++ b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ScaleKernel.h @@ -0,0 +1,107 @@ +/* + * Copyright (c) 2020-2021 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#ifndef ARM_COMPUTE_CPU_GEMMLOWP_QUANTIZEDOWNINT32_SCALE_KERNEL_H +#define ARM_COMPUTE_CPU_GEMMLOWP_QUANTIZEDOWNINT32_SCALE_KERNEL_H + +#include "arm_compute/core/KernelDescriptors.h" +#include "src/core/common/Macros.h" +#include "src/core/cpu/ICpuKernel.h" + +namespace arm_compute +{ +class ITensor; +namespace cpu +{ +namespace kernels +{ +/** Kernel used to quantize down the int32 accumulator values of GEMMLowp to QASYMM8/QASYMM8_SIGNED + * + * This kernel takes a final int32 accumulator value (the output of @ref NEGEMMLowpMatrixMultiplyKernel), and processes it to obtain the final QASYMM8/QASYMM8_SIGNED value. + * The following computations will be performed by the kernel: + * + * -# Add offset terms to final result + * -# Multiply each entry of result by result_mult_int + * -# Add bias to final result if bias tensor is not a nullptr + * -# Shift the int32 accumulator by result_shift + * -# Clamp the value between the specified min and max bounds + * -# Clamp the resulting int32 values: + * -# -to the [0..255] range and cast to QASYMM8. + * -# -to the [-128..127] range and cast to QASYMM8_SIGNED. + * + */ +class CpuGemmLowpQuantizeDownInt32ScaleKernel : public ICpuKernel +{ +public: + /** Default constructor */ + CpuGemmLowpQuantizeDownInt32ScaleKernel() = default; + ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(CpuGemmLowpQuantizeDownInt32ScaleKernel); + /** Initialise the kernel's input and output. + * + * @param[in] src Input tensor info. Data type supported: S32 + * @param[in] bias Biases tensor info. Only shared biases supported and it can be a nullptr if the biases addition is not required. + * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. + * @param[out] dst Output tensor info. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED + * @param[out] output_stage GEMMLowp output stage metadata. + */ + void configure(ITensorInfo *src, ITensorInfo *bias, ITensorInfo *dst, const GEMMLowpOutputStageInfo *output_stage); + /** Static function to check if given info will lead to a valid configuration + * + * Similar to CpuGemmLowpQuantizeDownInt32ScaleKernel::configure() + * + * @return a status + */ + static Status validate(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, const GEMMLowpOutputStageInfo *output_stage); + + // Inherited methods overridden: + void run_op(ITensorPack &tensors, const Window &window, const ThreadInfo &info) override; + const char *name() const override; + +private: + /** Template function to run the NEGEMMLowpQuantizeDownInt32ScaleKernel + * + * @param[in] src Input tensor info + * @param[in] bias Biases tensor info + * @param[out] dst Output tensor info + * @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()) + */ + template + void run_internal(const ITensor *src, const ITensor *bias, ITensor *dst, const Window &window); + + /** Common signature for all the specialised CpuGemmLowpQuantizeDownInt32ScaleKernel functions + * + * @param[in] src Input tensor info + * @param[in] bias Biases tensor info + * @param[out] dst Output tensor info + * @param[in] window Region on which to execute the kernel. + */ + using QuantizeDownFunctionPtr = void (CpuGemmLowpQuantizeDownInt32ScaleKernel::*)(const ITensor *src, const ITensor *bias, ITensor *dst, const Window &window); + + QuantizeDownFunctionPtr _func{ nullptr }; + const GEMMLowpOutputStageInfo *_output_stage{ nullptr }; + bool _is_bounded_relu{ false }; +}; +} // namespace kernels +} // namespace cpu +} // namespace arm_compute +#endif /* ARM_COMPUTE_CPU_GEMMLOWP_QUANTIZEDOWNINT32_SCALE_KERNEL_H */ diff --git a/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp new file mode 100644 index 0000000000..390e269cbb --- /dev/null +++ b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp @@ -0,0 +1,227 @@ +/* + * Copyright (c) 2019-2021 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#include "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h" + +#include "arm_compute/core/Error.h" +#include "arm_compute/core/Helpers.h" +#include "arm_compute/core/ITensor.h" +#include "arm_compute/core/TensorInfo.h" +#include "arm_compute/core/Types.h" +#include "arm_compute/core/Utils.h" +#include "arm_compute/core/Validate.h" +#include "arm_compute/core/Window.h" +#include "arm_compute/core/utils/misc/ShapeCalculator.h" +#include "src/core/NEON/NESymm.h" +#include "src/core/helpers/AutoConfiguration.h" +#include "src/core/helpers/WindowHelpers.h" + +#include + +namespace arm_compute +{ +namespace cpu +{ +namespace kernels +{ +namespace +{ +Status validate_arguments(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, int min, int max) +{ + ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst); + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(src, 1, DataType::S32); + ARM_COMPUTE_RETURN_ERROR_ON(min > max); + + // Check biases if exist + if(bias != nullptr) + { + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(src, bias); + ARM_COMPUTE_RETURN_ERROR_ON(bias->num_dimensions() > 1); + ARM_COMPUTE_RETURN_ERROR_ON(src->dimension(0) != bias->dimension(0)); + } + + if(dst->total_size() != 0) + { + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(dst, 1, DataType::QSYMM16); + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(dst, src); + } + + return Status{}; +} +} // namespace + +template +void CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::run_internal(const ITensor *src, const ITensor *bias, ITensor *dst, const Window &window) +{ + const int16x8_t min_s16 = vdupq_n_s16(static_cast(_min)); + const int16x8_t max_s16 = vdupq_n_s16(static_cast(_max)); + + ARM_COMPUTE_UNUSED(min_s16); + ARM_COMPUTE_UNUSED(max_s16); + + const int window_step_x = 8; + const auto window_start_x = static_cast(window.x().start()); + const auto window_end_x = static_cast(window.x().end()); + + Window win_collapsed = window.collapse_if_possible(window, Window::DimZ); + win_collapsed.set(Window::DimX, Window::Dimension(0, 1, 1)); + + Iterator in(src, win_collapsed); + Iterator out(dst, win_collapsed); + if(bias != nullptr) + { + Window win_biases; + win_biases.set(Window::DimX, Window::Dimension(0, 1, 1)); + win_biases.set(Window::DimY, Window::Dimension(0, 1, 1)); + + Iterator bias_i(bias, win_biases); + execute_window_loop(win_collapsed, [&](const Coordinates &) + { + // Compute 16 elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) + { + int32x4x2_t in_s32 = + { + { + vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 4) + } + }; + + const int32x4x2_t bias_s32 = + { + { + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 0), + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 4) + } + }; + + // Add the bias to GEMM's result + in_s32.val[0] = vaddq_s32(in_s32.val[0], bias_s32.val[0]); + in_s32.val[1] = vaddq_s32(in_s32.val[1], bias_s32.val[1]); + + vst1q_s16(reinterpret_cast(out.ptr()) + x, finalize_quantization_int16(in_s32, _result_fixedpoint_multiplier, _result_shift, min_s16, max_s16)); + } + + // Compute left-over elements + for(; x < window_end_x; ++x) + { + const int32_t bias_value = *(reinterpret_cast(bias_i.ptr()) + x); + int32_t in_value = *(reinterpret_cast(in.ptr()) + x); + + // Add bias + in_value += bias_value; + // Finalize and store the result + *(reinterpret_cast(out.ptr()) + x) = finalize_quantization_int16(in_value, _result_fixedpoint_multiplier, _result_shift, static_cast(_min), + static_cast(_max)); + } + }, + in, out, bias_i); + } + else + { + execute_window_loop(win_collapsed, [&](const Coordinates &) + { + // Compute 16 elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) + { + int32x4x2_t in_s32 = + { + { + vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 4) + } + }; + + vst1q_s16(reinterpret_cast(out.ptr()) + x, finalize_quantization_int16(in_s32, _result_fixedpoint_multiplier, _result_shift, min_s16, max_s16)); + } + + // Compute left-over elements + for(; x < window_end_x; ++x) + { + const int32_t in_value = *(reinterpret_cast(in.ptr()) + x); + ARM_COMPUTE_UNUSED(in_value); + // Finalize and store the result + *(reinterpret_cast(out.ptr()) + x) = finalize_quantization_int16(in_value, _result_fixedpoint_multiplier, _result_shift, static_cast(_min), + static_cast(_max)); + } + }, + in, out); + } +} + +void CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::configure(ITensorInfo *src, ITensorInfo *bias, ITensorInfo *dst, int result_fixedpoint_multiplier, int result_shift, + int min, int max) +{ + // Perform validate step + ARM_COMPUTE_UNUSED(bias, dst); + ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst); + ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(src, bias, dst, min, max)); + + _result_fixedpoint_multiplier = result_fixedpoint_multiplier; + _result_shift = result_shift; + _min = min; + _max = max; + + // Output auto inizialitation if not yet initialized + auto_init_if_empty(*src, src->clone()->set_data_type(DataType::QSYMM16)); + // Configure kernel window + Window win_config = calculate_max_window(*src, Steps()); + ICpuKernel::configure(win_config); + + // Check if we need to clamp the result using min and max + const bool is_bounded_relu = !(min <= -32768 && max >= 32767); + _func = is_bounded_relu ? &CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::run_internal : + &CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::run_internal; +} + +Status CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min, int max) +{ + ARM_COMPUTE_ERROR_ON_NULLPTR(input, output); + ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, bias, output, min, max)); + return Status{}; +} + +void CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::run_op(ITensorPack &tensors, const Window &window, const ThreadInfo &info) +{ + ARM_COMPUTE_UNUSED(info); + ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); + ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(ICpuKernel::window(), window); + ARM_COMPUTE_ERROR_ON_MSG(tensors.empty(), "No inputs provided"); + + auto src = tensors.get_const_tensor(TensorType::ACL_SRC); + auto bias = tensors.get_const_tensor(TensorType::ACL_BIAS); + auto dst = tensors.get_tensor(TensorType::ACL_DST); + + (this->*_func)(src, bias, dst, window); +} + +const char *CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::name() const +{ + return "CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel"; +} +} // namespace kernels +} // namespace cpu +} // namespace arm_compute diff --git a/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h new file mode 100644 index 0000000000..4d743e9a0a --- /dev/null +++ b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h @@ -0,0 +1,111 @@ +/* + * Copyright (c) 2019-2021 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#ifndef ARM_COMPUTE_CPU_GEMMLOWP_QUANTIZEDOWNINT32TOINT16_SCALEBYFIXEDPOINT_KERNEL_H +#define ARM_COMPUTE_CPU_GEMMLOWP_QUANTIZEDOWNINT32TOINT16_SCALEBYFIXEDPOINT_KERNEL_H + +#include "arm_compute/core/KernelDescriptors.h" +#include "src/core/common/Macros.h" +#include "src/core/cpu/ICpuKernel.h" + +namespace arm_compute +{ +class ITensor; +namespace cpu +{ +namespace kernels +{ +/** Kernel used to quantize down the int32 accumulator values of GEMMLowp to QSYMM16 + * + * This kernel takes a final int32 accumulator value (the output of @ref NEGEMMLowpMatrixMultiplyKernel), and processes it to obtain the final QSYMM16 value. + * The following computations will be performed by the kernel: + * + * -# Compute fixed point multiplication between each entry of input by result_fixedpoint_multiplier + * -# Add bias to final result if bias tensor is not a nullptr + * -# Round to nearest division by a power-of-two using result_shift + * -# Clamp the value between the specified min and max bounds + * -# Clamp the resulting int32 values to the [-32768, 32767] range and cast to QSYMM16. + * + */ +class CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel : public ICpuKernel +{ +public: + /** Default constructor */ + CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel() = default; + ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel); + /** Initialise the kernel's input and output. + * + * @param[in] src Input tensor info. Data type supported: S32 + * @param[in] bias Biases tensor info. Only shared biases supported and it can be a nullptr if the biases addition is not required. + * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. + * @param[out] dst Output tensor info. Data type supported: Data type supported: QSYMM16 + * @param[in] result_fixedpoint_multiplier Fixed point value to be multiplied to each element of the input matrix when once the result_offset has been add + * @param[in] result_shift Integer value used to round to nearest division by a power-of-two the result after the fixed point multiplication + * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QSYMM16. Defaults to 0. + * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QSYMM16. + * Along with @p min, this value can be used to implement "rectified linear unit" activation functions. Defaults to 0. + */ + void configure(ITensorInfo *src, ITensorInfo *bias, ITensorInfo *dst, int result_fixedpoint_multiplier, int result_shift, int min = 0, int max = 0); + /** Static function to check if given info will lead to a valid configuration + * + * Similar to CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::configure() + * + * @return a status + */ + static Status validate(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, int min = 0, int max = 0); + + // Inherited methods overridden: + void run_op(ITensorPack &tensors, const Window &window, const ThreadInfo &info) override; + const char *name() const override; + +private: + /** Template function to run the CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel + * + * @param[in] src Input tensor info + * @param[in] bias Bias tensor info + * @param[out] dst Output tensor info + * @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()). + */ + template + void run_internal(const ITensor *src, const ITensor *bias, ITensor *dst, const Window &window); + + /** Common signature for all the specialised CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel functions + * + * @param[in] src Input tensor info + * @param[in] bias Bias tensor info + * @param[out] dst Output tensor info + * @param[in] window Region on which to execute the kernel. + */ + using QuantizeDownFunctionPtr = void (CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::*)( + const ITensor *src, const ITensor *bias, ITensor *dst, const Window &window); + + QuantizeDownFunctionPtr _func{ nullptr }; + int _result_fixedpoint_multiplier{ 0 }; + int _result_shift{ 0 }; + int _min{ 0 }; + int _max{ 0 }; +}; +} // namespace kernels +} // namespace cpu +} // namespace arm_compute +#endif /* ARM_COMPUTE_CPU_GEMMLOWP_QUANTIZEDOWNINT32TOINT16_SCALEBYFIXEDPOINT_KERNEL_H */ diff --git a/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp new file mode 100644 index 0000000000..318b6a06f8 --- /dev/null +++ b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp @@ -0,0 +1,239 @@ +/* + * Copyright (c) 2019-2021 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#include "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h" + +#include "arm_compute/core/Error.h" +#include "arm_compute/core/Helpers.h" +#include "arm_compute/core/ITensor.h" +#include "arm_compute/core/TensorInfo.h" +#include "arm_compute/core/Types.h" +#include "arm_compute/core/Utils.h" +#include "arm_compute/core/Validate.h" +#include "arm_compute/core/Window.h" +#include "arm_compute/core/utils/misc/ShapeCalculator.h" +#include "src/core/NEON/NEAsymm.h" +#include "src/core/helpers/AutoConfiguration.h" +#include "src/core/helpers/WindowHelpers.h" + +#include + +namespace arm_compute +{ +namespace cpu +{ +namespace kernels +{ +namespace +{ +Status validate_arguments(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, int min, int max) +{ + ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst); + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(src, 1, DataType::S32); + ARM_COMPUTE_RETURN_ERROR_ON(min > max); + + // Check biases if exist + if(bias != nullptr) + { + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(src, bias); + ARM_COMPUTE_RETURN_ERROR_ON(bias->num_dimensions() > 1); + ARM_COMPUTE_RETURN_ERROR_ON(src->dimension(0) != bias->dimension(0)); + } + + if(dst->total_size() != 0) + { + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(dst, 1, DataType::QASYMM8_SIGNED); + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(dst, src); + } + + return Status{}; +} +} // namespace + +template +void CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::run_internal(const ITensor *src, const ITensor *bias, ITensor *dst, const Window &window) +{ + const int32x4_t result_offset_after_shift_s32 = vdupq_n_s32(_result_offset_after_shift); + const int8x16_t min_s8 = vdupq_n_s8(static_cast(_min)); + const int8x16_t max_s8 = vdupq_n_s8(static_cast(_max)); + + ARM_COMPUTE_UNUSED(min_s8, max_s8); + + const int window_step_x = 16; + const auto window_start_x = static_cast(window.x().start()); + const auto window_end_x = static_cast(window.x().end()); + + Window win_collapsed = window.collapse_if_possible(window, Window::DimZ); + win_collapsed.set(Window::DimX, Window::Dimension(0, 1, 1)); + + Iterator in(src, win_collapsed); + Iterator out(dst, win_collapsed); + if(bias != nullptr) + { + Window win_biases; + win_biases.set(Window::DimX, Window::Dimension(0, 1, 1)); + win_biases.set(Window::DimY, Window::Dimension(0, 1, 1)); + + Iterator bias_i(bias, win_biases); + execute_window_loop(win_collapsed, [&](const Coordinates &) + { + // Compute 16 elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) + { + int32x4x4_t in_s32 = + { + { + vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 4), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 8), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 12) + } + }; + + const int32x4x4_t bias_s32 = + { + { + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 0), + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 4), + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 8), + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 12) + } + }; + + // Add the bias to GEMM's result + in_s32.val[0] = vaddq_s32(in_s32.val[0], bias_s32.val[0]); + in_s32.val[1] = vaddq_s32(in_s32.val[1], bias_s32.val[1]); + in_s32.val[2] = vaddq_s32(in_s32.val[2], bias_s32.val[2]); + in_s32.val[3] = vaddq_s32(in_s32.val[3], bias_s32.val[3]); + + vst1q_s8(reinterpret_cast(out.ptr() + x), + finalize_quantization(in_s32, _result_fixedpoint_multiplier, _result_shift, result_offset_after_shift_s32, min_s8, max_s8, is_bounded_relu)); + } + + // Compute left-over elements + for(; x < window_end_x; ++x) + { + const int32_t bias_value = *(reinterpret_cast(bias_i.ptr()) + x); + int32_t in_value = *(reinterpret_cast(in.ptr()) + x); + + // Add bias + in_value += bias_value; + // Finalize and store the result + *reinterpret_cast(out.ptr() + x) = finalize_quantization(in_value, _result_fixedpoint_multiplier, _result_shift, _result_offset_after_shift, + static_cast(_min), static_cast(_max), is_bounded_relu); + } + }, + in, out, bias_i); + } + else + { + execute_window_loop(win_collapsed, [&](const Coordinates &) + { + // Compute 16 elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) + { + int32x4x4_t in_s32 = + { + { + vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 4), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 8), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 12) + } + }; + + vst1q_s8(reinterpret_cast(out.ptr() + x), + finalize_quantization(in_s32, _result_fixedpoint_multiplier, _result_shift, result_offset_after_shift_s32, min_s8, max_s8, is_bounded_relu)); + } + + // Compute left-over elements + for(; x < window_end_x; ++x) + { + const int32_t in_value = *(reinterpret_cast(in.ptr()) + x); + + // Finalize and store the result + *reinterpret_cast(out.ptr() + x) = finalize_quantization(in_value, _result_fixedpoint_multiplier, _result_shift, _result_offset_after_shift, + static_cast(_min), static_cast(_max), is_bounded_relu); + } + }, + in, out); + } +} + +void CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::configure(ITensorInfo *src, ITensorInfo *bias, ITensorInfo *dst, int result_fixedpoint_multiplier, int result_shift, + int result_offset_after_shift, int min, int max) +{ + ARM_COMPUTE_UNUSED(bias); + // Perform validate step + ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst); + ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(src, bias, dst, min, max)); + + _result_fixedpoint_multiplier = result_fixedpoint_multiplier; + _result_shift = result_shift; + _result_offset_after_shift = result_offset_after_shift; + _min = min; + _max = max; + + // Output auto initialization if not yet initialized + auto_init_if_empty(*dst, src->clone()->set_data_type(DataType::QASYMM8_SIGNED)); + + // Configure kernel window + Window win_config = calculate_max_window(*src, Steps()); + ICpuKernel::configure(win_config); + + // Check if we need to clamp the result using min and max + const bool is_bounded_relu = !(min <= -128 && max >= 127); + _func = is_bounded_relu ? &CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::run_internal : + &CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::run_internal; +} + +Status CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::validate(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, int min, int max) +{ + ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst); + ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(src, bias, dst, min, max)); + return Status{}; +} + +void CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::run_op(ITensorPack &tensors, const Window &window, const ThreadInfo &info) +{ + ARM_COMPUTE_UNUSED(info); + ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); + ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(ICpuKernel::window(), window); + ARM_COMPUTE_ERROR_ON_MSG(tensors.empty(), "No inputs provided"); + + auto src = tensors.get_const_tensor(TensorType::ACL_SRC); + auto bias = tensors.get_const_tensor(TensorType::ACL_BIAS); + auto dst = tensors.get_tensor(TensorType::ACL_DST); + + (this->*_func)(src, bias, dst, window); +} + +const char *CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::name() const +{ + return "CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel"; +} +} // namespace kernels +} // namespace cpu +} // namespace arm_compute diff --git a/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h new file mode 100644 index 0000000000..a941f1f542 --- /dev/null +++ b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h @@ -0,0 +1,114 @@ +/* + * Copyright (c) 2019-2021 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#ifndef ARM_COMPUTE_CPU_GEMMLOWP_QUANTIZEDOWNINT32TOINT8_SCALEBYFIXEDPOINT_KERNEL_H +#define ARM_COMPUTE_CPU_GEMMLOWP_QUANTIZEDOWNINT32TOINT8_SCALEBYFIXEDPOINT_KERNEL_H + +#include "arm_compute/core/KernelDescriptors.h" +#include "src/core/common/Macros.h" +#include "src/core/cpu/ICpuKernel.h" + +namespace arm_compute +{ +class ITensor; +namespace cpu +{ +namespace kernels +{ +/** Kernel used to quantize down the int32 accumulator values of GEMMLowp to QASYMM8_SIGNED + * + * This kernel takes a final int32 accumulator value (the output of @ref NEGEMMLowpMatrixMultiplyKernel), and processes it to obtain the final QASYMM8_SIGNED value. + * The following computations will be performed by the kernel: + * + * -# Compute fixed point multiplication between each entry of input by result_fixedpoint_multiplier + * -# Add bias to final result if bias tensor is not a nullptr + * -# Round to nearest division by a power-of-two using result_shift + * -# Add offset to each result + * -# Clamp the value between the specified min and max bounds + * -# Clamp the resulting int32 values to the [-128..127] range and cast to QASYMM8_SIGNED. + * + */ +class CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel : public ICpuKernel +{ +public: + /** Default constructor */ + CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel() = default; + ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel); + /** Initialise the kernel's input and output. + * + * @param[in] src Input tensor info. Data type supported: S32 + * @param[in] bias Biases tensor info. Only shared biases supported and it can be a nullptr if the biases addition is not required. + * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. + * @param[out] dst Output tensor info. Data type supported: Data type supported: QASYMM8_SIGNED + * @param[in] result_fixedpoint_multiplier Fixed point value to be multiplied to each element of the input matrix when once the result_offset has been add + * @param[in] result_shift Integer value used to round to nearest division by a power-of-two the result after the fixed point multiplication + * @param[in] result_offset_after_shift Offset to be applied to result before converting it back to QASYMM8_SIGNED + * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8_SIGNED + * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8_SIGNED, + * Along with @p min, this value can be used to implement "rectified linear unit" activation functions + */ + void configure(ITensorInfo *src, ITensorInfo *bias, ITensorInfo *dst, int result_fixedpoint_multiplier, int result_shift, int result_offset_after_shift, int min = 0, int max = 0); + /** Static function to check if given info will lead to a valid configuration + * + * Similar to CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::configure() + * + * @return a status + */ + static Status validate(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, int min = 0, int max = 0); + + // Inherited methods overridden: + void run_op(ITensorPack &tensors, const Window &window, const ThreadInfo &info) override; + const char *name() const override; + +private: + /** Template function to run the CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel + * + * @param[in] src Input tensor info + * @param[in] bias Bias tensor info + * @param[out] dst Output tensor info + * @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()). + */ + template + void run_internal(const ITensor *src, const ITensor *bias, ITensor *dst, const Window &window); + + /** Common signature for all the specialised CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel functions + * + * @param[in] src Input tensor info + * @param[in] bias Bias tensor info + * @param[out] dst Output tensor info + * @param[in] window Region on which to execute the kernel. + */ + using QuantizeDownFunctionPtr = void (CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::*)( + const ITensor *src, const ITensor *bias, ITensor *dst, const Window &window); + + QuantizeDownFunctionPtr _func{ nullptr }; + int _result_fixedpoint_multiplier{ 0 }; + int _result_shift{ 0 }; + int _result_offset_after_shift{ 0 }; + int _min{ 0 }; + int _max{ 0 }; +}; +} // namespace kernels +} // namespace cpu +} // namespace arm_compute +#endif /* ARM_COMPUTE_CPU_GEMMLOWP_QUANTIZEDOWNINT32TOINT8_SCALEBYFIXEDPOINT_KERNEL_H */ diff --git a/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp new file mode 100644 index 0000000000..6631a4fc67 --- /dev/null +++ b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp @@ -0,0 +1,236 @@ +/* + * Copyright (c) 2017-2021 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#include "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h" + +#include "arm_compute/core/Error.h" +#include "arm_compute/core/Helpers.h" +#include "arm_compute/core/ITensor.h" +#include "arm_compute/core/TensorInfo.h" +#include "arm_compute/core/Types.h" +#include "arm_compute/core/Utils.h" +#include "arm_compute/core/Validate.h" +#include "arm_compute/core/Window.h" +#include "arm_compute/core/utils/misc/ShapeCalculator.h" +#include "src/core/NEON/NEAsymm.h" +#include "src/core/helpers/AutoConfiguration.h" +#include "src/core/helpers/WindowHelpers.h" + +#include + +namespace arm_compute +{ +namespace cpu +{ +namespace kernels +{ +namespace +{ +Status validate_arguments(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, int min, int max) +{ + ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst); + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(src, 1, DataType::S32); + ARM_COMPUTE_RETURN_ERROR_ON(min > max); + + // Check biases if exist + if(bias != nullptr) + { + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(src, bias); + ARM_COMPUTE_RETURN_ERROR_ON(bias->num_dimensions() > 1); + ARM_COMPUTE_RETURN_ERROR_ON(src->dimension(0) != bias->dimension(0)); + } + + if(dst->total_size() != 0) + { + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(dst, 1, DataType::QASYMM8); + ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(dst, src); + } + + return Status{}; +} +} // namespace + +template +void CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::run_internal(const ITensor *src, const ITensor *bias, ITensor *dst, const Window &window) +{ + const int32x4_t result_offset_after_shift_s32 = vdupq_n_s32(_result_offset_after_shift); + const uint8x16_t min_u8 = vdupq_n_u8(static_cast(_min)); + const uint8x16_t max_u8 = vdupq_n_u8(static_cast(_max)); + + ARM_COMPUTE_UNUSED(min_u8); + ARM_COMPUTE_UNUSED(max_u8); + + const int window_step_x = 16; + const auto window_start_x = static_cast(window.x().start()); + const auto window_end_x = static_cast(window.x().end()); + + Window win_collapsed = window.collapse_if_possible(window, Window::DimZ); + win_collapsed.set(Window::DimX, Window::Dimension(0, 1, 1)); + + Iterator in(src, win_collapsed); + Iterator out(dst, win_collapsed); + if(bias != nullptr) + { + Window win_biases; + win_biases.set(Window::DimX, Window::Dimension(0, 1, 1)); + win_biases.set(Window::DimY, Window::Dimension(0, 1, 1)); + + Iterator bias_i(bias, win_biases); + execute_window_loop(win_collapsed, [&](const Coordinates &) + { + // Compute 16 elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) + { + int32x4x4_t in_s32 = + { + { + vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 4), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 8), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 12) + } + }; + + const int32x4x4_t bias_s32 = + { + { + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 0), + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 4), + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 8), + vld1q_s32(reinterpret_cast(bias_i.ptr()) + x + 12) + } + }; + + // Add the bias to GEMM's result + in_s32.val[0] = vaddq_s32(in_s32.val[0], bias_s32.val[0]); + in_s32.val[1] = vaddq_s32(in_s32.val[1], bias_s32.val[1]); + in_s32.val[2] = vaddq_s32(in_s32.val[2], bias_s32.val[2]); + in_s32.val[3] = vaddq_s32(in_s32.val[3], bias_s32.val[3]); + + vst1q_u8(out.ptr() + x, finalize_quantization(in_s32, _result_fixedpoint_multiplier, _result_shift, result_offset_after_shift_s32, min_u8, max_u8, is_bounded_relu)); + } + + // Compute left-over elements + for(; x < window_end_x; ++x) + { + const int32_t bias_value = *(reinterpret_cast(bias_i.ptr()) + x); + int32_t in_value = *(reinterpret_cast(in.ptr()) + x); + + // Add bias + in_value += bias_value; + // Finalize and store the result + *(out.ptr() + x) = finalize_quantization(in_value, _result_fixedpoint_multiplier, _result_shift, _result_offset_after_shift, static_cast(_min), static_cast(_max), is_bounded_relu); + } + }, + in, out, bias_i); + } + else + { + execute_window_loop(win_collapsed, [&](const Coordinates &) + { + // Compute 16 elements per iteration + int x = window_start_x; + for(; x <= (window_end_x - window_step_x); x += window_step_x) + { + int32x4x4_t in_s32 = + { + { + vld1q_s32(reinterpret_cast(in.ptr()) + x + 0), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 4), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 8), + vld1q_s32(reinterpret_cast(in.ptr()) + x + 12) + } + }; + + vst1q_u8(out.ptr() + x, finalize_quantization(in_s32, _result_fixedpoint_multiplier, _result_shift, result_offset_after_shift_s32, min_u8, max_u8, is_bounded_relu)); + } + + // Compute left-over elements + for(; x < window_end_x; ++x) + { + const int32_t in_value = *(reinterpret_cast(in.ptr()) + x); + + // Finalize and store the result + *(out.ptr() + x) = finalize_quantization(in_value, _result_fixedpoint_multiplier, _result_shift, _result_offset_after_shift, static_cast(_min), static_cast(_max), is_bounded_relu); + } + }, + in, out); + } +} + +void CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::configure(ITensorInfo *src, ITensorInfo *bias, ITensorInfo *dst, int result_fixedpoint_multiplier, int result_shift, + int result_offset_after_shift, int min, int max) +{ + ARM_COMPUTE_UNUSED(bias); + // Perform validate step + ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst); + ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(src, bias, dst, min, max)); + + _result_fixedpoint_multiplier = result_fixedpoint_multiplier; + _result_shift = result_shift; + _result_offset_after_shift = result_offset_after_shift; + _min = min; + _max = max; + + // Output auto inizialitation if not yet initialized + auto_init_if_empty(*dst, src->clone()->set_data_type(DataType::QASYMM8)); + + // Configure kernel window + auto win_config = calculate_max_window(*src, Steps()); + ICpuKernel::configure(win_config); + + // Check if we need to clamp the result using min and max + const bool is_bounded_relu = !(min <= 0 && max >= 255); + _func = is_bounded_relu ? &CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::run_internal : + &CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::run_internal; +} + +Status CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::validate(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, int min, int max) +{ + ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst); + ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(src, bias, dst, min, max)); + return Status{}; +} + +void CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::run_op(ITensorPack &tensors, const Window &window, const ThreadInfo &info) +{ + ARM_COMPUTE_UNUSED(info); + ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); + ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(ICpuKernel::window(), window); + ARM_COMPUTE_ERROR_ON_MSG(tensors.empty(), "No inputs provided"); + + auto src = tensors.get_const_tensor(TensorType::ACL_SRC); + auto bias = tensors.get_const_tensor(TensorType::ACL_BIAS); + auto dst = tensors.get_tensor(TensorType::ACL_DST); + + (this->*_func)(src, bias, dst, window); +} + +const char *CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::name() const +{ + return "CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel"; +} +} // namespace kernels +} // namespace cpu +} // namespace arm_compute \ No newline at end of file diff --git a/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h new file mode 100644 index 0000000000..9b4c056419 --- /dev/null +++ b/src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h @@ -0,0 +1,108 @@ +/* + * Copyright (c) 2017-2021 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#ifndef ARM_COMPUTE_CPU_GEMMLOWP_QUANTIZEDOWNINT32TOUINT8_SCALEBYFIXEDPOINT_KERNEL_H +#define ARM_COMPUTE_CPU_GEMMLOWP_QUANTIZEDOWNINT32TOUINT8_SCALEBYFIXEDPOINT_KERNEL_H + +#include "arm_compute/core/KernelDescriptors.h" +#include "src/core/common/Macros.h" +#include "src/core/cpu/ICpuKernel.h" + +namespace arm_compute +{ +class ITensor; +namespace cpu +{ +namespace kernels +{ +/** Kernel used to quantize down the int32 accumulator values of GEMMLowp to QASYMM8 + * + * This kernel takes a final int32 accumulator value (the output of @ref NEGEMMLowpMatrixMultiplyKernel), and processes it to obtain the final QASYMM8 value. + * The following computations will be performed by the kernel: + * + * -# Compute fixed point multiplication between each entry of input by result_fixedpoint_multiplier + * -# Add bias to final result if bias tensor is not a nullptr + * -# Round to nearest division by a power-of-two using result_shift + * -# Add offset to each result + * -# Clamp the value between the specified min and max bounds + * -# Clamp the resulting int32 values to the [0..255] range and cast to QASYMM8. + * + */ +class CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel : public ICpuKernel +{ +public: + /** Default constructor */ + CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel() = default; + ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel); + /** Initialise the kernel's input and output. + * + * @param[in] src Input tensor info. Data type supported: S32 + * @param[in] bias Biases tensor info. Only shared biases supported and it can be a nullptr if the biases addition is not required. + * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. + * @param[out] dst Output tensor info. Data type supported: Data type supported: QASYMM8 + * @param[in] result_fixedpoint_multiplier Fixed point value to be multiplied to each element of the input matrix when once the result_offset has been add + * @param[in] result_shift Integer value used to round to nearest division by a power-of-two the result after the fixed point multiplication + * @param[in] result_offset_after_shift Offset to be applied to result before converting it back to QASYMM8 + * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8 + * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8, + * Along with @p min, this value can be used to implement "rectified linear unit" activation functions + */ + void configure(ITensorInfo *src, ITensorInfo *bias, ITensorInfo *dst, int result_fixedpoint_multiplier, int result_shift, int result_offset_after_shift, int min = 0, int max = 0); + /** Static function to check if given info will lead to a valid configuration + * + * Similar to CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::configure() + * + * @return a status + */ + static Status validate(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, int min = 0, int max = 0); + + // Inherited methods overridden: + void run_op(ITensorPack &tensors, const Window &window, const ThreadInfo &info) override; + const char *name() const override; + +private: + /** Template function to run the CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel + * + * @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()). + */ + template + void run_internal(const ITensor *src, const ITensor *bias, ITensor *dst, const Window &window); + + /** Common signature for all the specialised CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel functions + * + * @param[in] window Region on which to execute the kernel. + */ + using QuantizeDownFunctionPtr = void (CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::*)( + const ITensor *src, const ITensor *bias, ITensor *dst, const Window &window); + + QuantizeDownFunctionPtr _func{ nullptr }; + int _result_fixedpoint_multiplier{ 0 }; + int _result_shift{ 0 }; + int _result_offset_after_shift{ 0 }; + int _min{ 0 }; + int _max{ 0 }; +}; +} // namespace kernels +} // namespace cpu +} // namespace arm_compute +#endif /* ARM_COMPUTE_CPU_GEMMLOWP_QUANTIZEDOWNINT32TOUINT8_SCALEBYFIXEDPOINT_KERNEL_H */ diff --git a/src/runtime/NEON/functions/NEGEMMLowpOutputStage.cpp b/src/runtime/NEON/functions/NEGEMMLowpOutputStage.cpp index 807785a534..8351cc66d0 100644 --- a/src/runtime/NEON/functions/NEGEMMLowpOutputStage.cpp +++ b/src/runtime/NEON/functions/NEGEMMLowpOutputStage.cpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2017-2020 Arm Limited. + * Copyright (c) 2017-2021 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -25,57 +25,23 @@ #include "arm_compute/core/ITensor.h" #include "arm_compute/core/Validate.h" -#include "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.h" -#include "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h" -#include "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h" -#include "src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h" +#include "src/runtime/cpu/operators/CpuGemmLowpOutputStage.h" namespace arm_compute { -NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint::~NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint() = default; - -void NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint::configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift, - int result_offset_after_shift, int min, int max) -{ - auto k = std::make_unique(); - k->configure(input, bias, output, result_fixedpoint_multiplier, result_shift, result_offset_after_shift, min, max); - _kernel = std::move(k); -} - -Status NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint::validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min, int max) -{ - return NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::validate(input, bias, output, min, max); -} - -NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint::~NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint() = default; - -void NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint::configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift, - int result_offset_after_shift, int min, int max) -{ - auto k = std::make_unique(); - k->configure(input, bias, output, result_fixedpoint_multiplier, result_shift, result_offset_after_shift, min, max); - _kernel = std::move(k); -} - -Status NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint::validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min, int max) +struct NEGEMMLowpOutputStage::Impl { - return NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::validate(input, bias, output, min, max); -} - -NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint::~NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint() = default; - -void NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint::configure(const ITensor *input, const ITensor *bias, ITensor *output, int result_fixedpoint_multiplier, int result_shift, int min, int max) -{ - auto k = std::make_unique(); - k->configure(input, bias, output, result_fixedpoint_multiplier, result_shift, min, max); - _kernel = std::move(k); -} - -Status NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint::validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min, int max) + const ITensor *src{ nullptr }; + const ITensor *bias{ nullptr }; + ITensor *dst{ nullptr }; + ITensorPack run_pack{}; + std::unique_ptr op{ nullptr }; +}; + +NEGEMMLowpOutputStage::NEGEMMLowpOutputStage() + : _impl(std::make_unique()) { - return NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::validate(input, bias, output, min, max); } - NEGEMMLowpOutputStage::~NEGEMMLowpOutputStage() = default; void NEGEMMLowpOutputStage::configure(const ITensor *input, const ITensor *bias, ITensor *output, const GEMMLowpOutputStageInfo &info) @@ -83,104 +49,27 @@ void NEGEMMLowpOutputStage::configure(const ITensor *input, const ITensor *bias, // Perform validate step ARM_COMPUTE_ERROR_ON_NULLPTR(input, output); ARM_COMPUTE_ERROR_THROW_ON(NEGEMMLowpOutputStage::validate(input->info(), bias != nullptr ? bias->info() : nullptr, output->info(), info)); + _impl->src = input; + _impl->bias = bias; + _impl->dst = output; + _impl->op = std::make_unique(); + _impl->op->configure(input->info(), (bias == nullptr) ? nullptr : bias->info(), output->info(), info); - switch(info.type) + _impl->run_pack = { - case GEMMLowpOutputStageType::QUANTIZE_DOWN_FIXEDPOINT: - { - switch(info.output_data_type) - { - case DataType::QASYMM8: - { - auto k = std::make_unique(); - k->configure(input, bias, output, info.gemmlowp_multiplier, info.gemmlowp_shift, info.gemmlowp_offset, info.gemmlowp_min_bound, info.gemmlowp_max_bound); - _kernel = std::move(k); - break; - } - case DataType::QASYMM8_SIGNED: - { - auto k = std::make_unique(); - k->configure(input, bias, output, info.gemmlowp_multiplier, info.gemmlowp_shift, info.gemmlowp_offset, info.gemmlowp_min_bound, info.gemmlowp_max_bound); - _kernel = std::move(k); - break; - } - case DataType::QSYMM16: - { - auto k = std::make_unique(); - k->configure(input, bias, output, info.gemmlowp_multiplier, info.gemmlowp_shift, info.gemmlowp_min_bound, info.gemmlowp_max_bound); - _kernel = std::move(k); - break; - } - default: - { - ARM_COMPUTE_ERROR("Unsupported output data type."); - break; - } - } - break; - } - case GEMMLowpOutputStageType::QUANTIZE_DOWN: - { - switch(info.output_data_type) - { - case DataType::QASYMM8: - case DataType::QASYMM8_SIGNED: - { - auto k = std::make_unique(); - k->configure(input, bias, output, &info); - _kernel = std::move(k); - break; - } - default: - { - ARM_COMPUTE_ERROR("Unsupported output data type."); - break; - } - } - break; - } - default: - ARM_COMPUTE_ERROR("Unsupported GEMMLowpOutputStage type."); - } + { TensorType::ACL_SRC, _impl->src }, + { TensorType::ACL_BIAS, _impl->bias }, + { TensorType::ACL_DST, _impl->dst } + }; } Status NEGEMMLowpOutputStage::validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const GEMMLowpOutputStageInfo &info) { - ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(output); - ARM_COMPUTE_RETURN_ERROR_ON_MSG(output->data_type() == DataType::UNKNOWN, "NEGEMMLowpQuantizeDownScaleByFixedPoint cannot be used with UNKNOWN output data type."); - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(output, 1, DataType::QASYMM8, DataType::QASYMM8_SIGNED, DataType::QSYMM16); - - ARM_COMPUTE_RETURN_ERROR_ON((info.type != GEMMLowpOutputStageType::QUANTIZE_DOWN) && (info.type != GEMMLowpOutputStageType::QUANTIZE_DOWN_FIXEDPOINT)); + return cpu::CpuGemmLowpOutputStage::validate(input, bias, output, info); +} - switch(info.type) - { - case GEMMLowpOutputStageType::QUANTIZE_DOWN_FIXEDPOINT: - { - switch(output->data_type()) - { - case DataType::QASYMM8: - return NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::validate(input, bias, output, info.gemmlowp_min_bound, info.gemmlowp_max_bound); - case DataType::QASYMM8_SIGNED: - return NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::validate(input, bias, output, info.gemmlowp_min_bound, info.gemmlowp_max_bound); - case DataType::QSYMM16: - return NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::validate(input, bias, output, info.gemmlowp_min_bound, info.gemmlowp_max_bound); - default: - return ARM_COMPUTE_CREATE_ERROR(ErrorCode::RUNTIME_ERROR, "Unsupported output data type."); - } - } - case GEMMLowpOutputStageType::QUANTIZE_DOWN: - { - switch(output->data_type()) - { - case DataType::QASYMM8: - case DataType::QASYMM8_SIGNED: - return NEGEMMLowpQuantizeDownInt32ScaleKernel::validate(input, bias, output, &info); - default: - return ARM_COMPUTE_CREATE_ERROR(ErrorCode::RUNTIME_ERROR, "Unsupported output data type."); - } - } - default: - return ARM_COMPUTE_CREATE_ERROR(ErrorCode::RUNTIME_ERROR, "Unsupported GEMMLowpOutputStage type."); - } +void NEGEMMLowpOutputStage::run() +{ + _impl->op->run(_impl->run_pack); } } // namespace arm_compute diff --git a/src/runtime/NEON/functions/NELSTMLayerQuantized.cpp b/src/runtime/NEON/functions/NELSTMLayerQuantized.cpp index 5c0f19a15c..a56e16864a 100644 --- a/src/runtime/NEON/functions/NELSTMLayerQuantized.cpp +++ b/src/runtime/NEON/functions/NELSTMLayerQuantized.cpp @@ -152,7 +152,13 @@ void NELSTMLayerQuantized::configure(const ITensor *input, quantization::calculate_quantized_multiplier(multiplier, &output_multiplier, &output_shift); _memory_group.manage(&_output_lowp); - _output_stage.configure(&_output_highp, &_bias, &_output_lowp, output_multiplier, output_shift); + + GEMMLowpOutputStageInfo info; + info.type = GEMMLowpOutputStageType::QUANTIZE_DOWN_FIXEDPOINT; + info.gemmlowp_multiplier = output_multiplier; + info.gemmlowp_shift = output_shift; + info.output_data_type = DataType::QSYMM16; + _output_stage.configure(&_output_highp, &_bias, &_output_lowp, info); _output_highp.allocator()->allocate(); _bias.allocator()->allocate(); @@ -358,7 +364,12 @@ Status NELSTMLayerQuantized::validate(const ITensorInfo *input, ARM_COMPUTE_RETURN_ON_ERROR(quantization::calculate_quantized_multiplier(multiplier, &output_multiplier, &output_shift)); // _output_stage - ARM_COMPUTE_RETURN_ON_ERROR(NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint::validate(&output_highp, &bias_concatenated, &output_lowp)); + GEMMLowpOutputStageInfo info; + info.type = GEMMLowpOutputStageType::QUANTIZE_DOWN_FIXEDPOINT; + info.gemmlowp_multiplier = output_multiplier; + info.gemmlowp_shift = output_shift; + info.output_data_type = DataType::QSYMM16; + ARM_COMPUTE_RETURN_ON_ERROR(NEGEMMLowpOutputStage::validate(&output_highp, &bias_concatenated, &output_lowp, info)); TensorInfo input_gate_input; TensorInfo forget_gate_input; diff --git a/src/runtime/cpu/operators/CpuGemmLowpOutputStage.cpp b/src/runtime/cpu/operators/CpuGemmLowpOutputStage.cpp new file mode 100644 index 0000000000..e17f854a21 --- /dev/null +++ b/src/runtime/cpu/operators/CpuGemmLowpOutputStage.cpp @@ -0,0 +1,147 @@ +/* + * Copyright (c) 2021 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#include "src/runtime/cpu/operators/CpuGemmLowpOutputStage.h" + +#include "arm_compute/core/ITensor.h" +#include "arm_compute/core/Validate.h" +#include "arm_compute/runtime/NEON/NEScheduler.h" +#include "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ScaleKernel.h" +#include "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.h" +#include "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.h" +#include "src/core/cpu/kernels/CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h" + +namespace arm_compute +{ +namespace cpu +{ +void CpuGemmLowpOutputStage::configure(ITensorInfo *src, ITensorInfo *bias, ITensorInfo *dst, const GEMMLowpOutputStageInfo &info) +{ + // Perform validate step + ARM_COMPUTE_ERROR_THROW_ON(CpuGemmLowpOutputStage::validate(src, bias, dst, info)); + + switch(info.type) + { + case GEMMLowpOutputStageType::QUANTIZE_DOWN_FIXEDPOINT: + { + switch(info.output_data_type) + { + case DataType::QASYMM8: + { + auto k = std::make_unique(); + k->configure(src, bias, dst, info.gemmlowp_multiplier, info.gemmlowp_shift, info.gemmlowp_offset, info.gemmlowp_min_bound, info.gemmlowp_max_bound); + _kernel = std::move(k); + break; + } + case DataType::QASYMM8_SIGNED: + { + auto k = std::make_unique(); + k->configure(src, bias, dst, info.gemmlowp_multiplier, info.gemmlowp_shift, info.gemmlowp_offset, info.gemmlowp_min_bound, info.gemmlowp_max_bound); + _kernel = std::move(k); + break; + } + case DataType::QSYMM16: + { + auto k = std::make_unique(); + k->configure(src, bias, dst, info.gemmlowp_multiplier, info.gemmlowp_shift, info.gemmlowp_min_bound, info.gemmlowp_max_bound); + _kernel = std::move(k); + break; + } + default: + { + ARM_COMPUTE_ERROR("Unsupported output data type."); + break; + } + } + break; + } + case GEMMLowpOutputStageType::QUANTIZE_DOWN: + { + switch(info.output_data_type) + { + case DataType::QASYMM8: + case DataType::QASYMM8_SIGNED: + { + auto k = std::make_unique(); + k->configure(src, bias, dst, &info); + _kernel = std::move(k); + break; + } + default: + { + ARM_COMPUTE_ERROR("Unsupported output data type."); + break; + } + } + break; + } + default: + ARM_COMPUTE_ERROR("Unsupported GEMMLowpOutputStage type."); + } +} + +Status CpuGemmLowpOutputStage::validate(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, const GEMMLowpOutputStageInfo &info) +{ + ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(src, dst); + ARM_COMPUTE_RETURN_ERROR_ON_MSG(dst->data_type() == DataType::UNKNOWN, "CpuGemmLowpOutputStage cannot be used with UNKNOWN output data type."); + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(dst, 1, DataType::QASYMM8, DataType::QASYMM8_SIGNED, DataType::QSYMM16); + ARM_COMPUTE_RETURN_ERROR_ON((info.type != GEMMLowpOutputStageType::QUANTIZE_DOWN) && (info.type != GEMMLowpOutputStageType::QUANTIZE_DOWN_FIXEDPOINT)); + + switch(info.type) + { + case GEMMLowpOutputStageType::QUANTIZE_DOWN_FIXEDPOINT: + { + switch(dst->data_type()) + { + case DataType::QASYMM8: + return kernels::CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel::validate(src, bias, dst, info.gemmlowp_min_bound, info.gemmlowp_max_bound); + case DataType::QASYMM8_SIGNED: + return kernels::CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel::validate(src, bias, dst, info.gemmlowp_min_bound, info.gemmlowp_max_bound); + case DataType::QSYMM16: + return kernels::CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel::validate(src, bias, dst, info.gemmlowp_min_bound, info.gemmlowp_max_bound); + default: + return ARM_COMPUTE_CREATE_ERROR(ErrorCode::RUNTIME_ERROR, "Unsupported output data type."); + } + } + case GEMMLowpOutputStageType::QUANTIZE_DOWN: + { + switch(dst->data_type()) + { + case DataType::QASYMM8: + case DataType::QASYMM8_SIGNED: + return kernels::CpuGemmLowpQuantizeDownInt32ScaleKernel::validate(src, bias, dst, &info); + default: + return ARM_COMPUTE_CREATE_ERROR(ErrorCode::RUNTIME_ERROR, "Unsupported output data type."); + } + } + default: + return ARM_COMPUTE_CREATE_ERROR(ErrorCode::RUNTIME_ERROR, "Unsupported GEMMLowpOutputStage type."); + } +} + +void CpuGemmLowpOutputStage::run(ITensorPack &tensors) +{ + NEScheduler::get().schedule_op(_kernel.get(), Window::DimY, _kernel->window(), tensors); +} +} // namespace cpu +} // namespace arm_compute \ No newline at end of file diff --git a/src/runtime/cpu/operators/CpuGemmLowpOutputStage.h b/src/runtime/cpu/operators/CpuGemmLowpOutputStage.h new file mode 100644 index 0000000000..848aaea3e7 --- /dev/null +++ b/src/runtime/cpu/operators/CpuGemmLowpOutputStage.h @@ -0,0 +1,90 @@ +/* + * Copyright (c) 2021 Arm Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#ifndef ARM_COMPUTE_CPU_GEMMLOWP_OUTPUTSTAGE_H +#define ARM_COMPUTE_CPU_GEMMLOWP_OUTPUTSTAGE_H + +#include "arm_compute/core/Types.h" +#include "src/runtime/cpu/ICpuOperator.h" + +/** This file contains all available output stages for GEMMLowp. + * + * In gemmlowp, the "output stage" is the process that takes a final int32 accumulator value (the output of @ref NEGEMMLowpMatrixMultiplyCore), + * and processes it to obtain the final ASYMM8 value. + * + * More information about the GEMMLowp output stage can be found at https://github.com/google/gemmlowp/blob/master/doc/output.md + */ + +namespace arm_compute +{ +namespace cpu +{ +/** Basic function to execute GEMMLowpQuantizeDown kernels. + * + * This function calls the following kernels: + * + * -# @ref kernels::CpuGemmLowpQuantizeDownInt32ScaleKernel + * -# @ref kernels::CpuGemmLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel + * -# @ref kernels::CpuGemmLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel + * -# @ref kernels::CpuGemmLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel +*/ +class CpuGemmLowpOutputStage : public ICpuOperator +{ +public: + /** Constructor */ + CpuGemmLowpOutputStage() = default; + /** Default destructor */ + ~CpuGemmLowpOutputStage() = default; + /** Initialise the kernel's inputs, output + * + * Valid data layouts: + * - All + * + * Valid data type configurations: + * |src0 |src1 |dst | + * |:--------------|:-------------|:-------------| + * |S32 |S32 |QASYMM8 | + * |S32 |S32 |QASYMM8_SIGNED| + * |S32 |S32 |QSYMM16 | + * + * @param[in] src Input tensor info. Data type supported: S32 + * @param[in] bias Biases tensor info. Only shared biases supported and it can be a nullptr if the biases addition is not required. + * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input. + * @param[out] dst Output tensor info. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM16 + * @param[in] info GEMMLowp output stage metadata. + */ + void configure(ITensorInfo *src, ITensorInfo *bias, ITensorInfo *dst, const GEMMLowpOutputStageInfo &info); + /** Static function to check if given info will lead to a valid configuration + * + * Similar to CpuGemmLowpOutputStage::configure() + * + * @return a status + */ + static Status validate(const ITensorInfo *src, const ITensorInfo *bias, const ITensorInfo *dst, const GEMMLowpOutputStageInfo &info); + + // Inherited methods overridden: + void run(ITensorPack &tensors) override; +}; +} // namespace cpu +} // namespace arm_compute +#endif /* ARM_COMPUTE_CPU_GEMMLOWP_OUTPUTSTAGE_H */ diff --git a/tests/validation/NEON/GEMMLowp.cpp b/tests/validation/NEON/GEMMLowp.cpp index 9d075e12c1..5d87330982 100644 --- a/tests/validation/NEON/GEMMLowp.cpp +++ b/tests/validation/NEON/GEMMLowp.cpp @@ -138,415 +138,8 @@ FIXTURE_DATA_TEST_CASE(RunLarge, NEGEMMLowpMatrixMultiplyCoreFusedOffsetOutputFi } TEST_SUITE_END() // FusedOffsetOutput TEST_SUITE_END() // MatrixMultiplyCore - -TEST_SUITE(OutputStage) - -TEST_SUITE(QuantizeDownInt32Scale) - -TEST_SUITE(QASYMM8) - -const auto quantize_down_int32_to_uint8_scale_cases = framework::dataset::make("result_offset", -2, 1) * framework::dataset::make("result_mult_int", 1, 2) * framework::dataset::make("result_shift", 2, - 3) - * framework::dataset::make("min", 0) * framework::dataset::make("max", 255) * framework::dataset::make("addBias", { false, true }); - -const auto quantize_down_int32_to_uint8_scale_relu_cases = framework::dataset::make("result_offset", -2, 1) * framework::dataset::make("result_mult_int", 1, - 2) - * framework::dataset::make("result_shift", 2, 3) * framework::dataset::make("min", 0, 2) * framework::dataset::make("max", 171, 174) * framework::dataset::make("addBias", { false, true }); - -using NEGEMMLowpQuantizeDownInt32ScaleFixture = GEMMLowpQuantizeDownInt32ToUint8ScaleValidationFixture; - -// *INDENT-OFF* -// clang-format off -DATA_TEST_CASE(Validate, framework::DatasetMode::ALL, zip(zip(zip(zip(zip( - framework::dataset::make("InputAInfo", { TensorInfo(TensorShape(21U, 13U), 1, DataType::S32), // Input not a multiple of 16 - TensorInfo(TensorShape(20U, 13U), 1, DataType::S32), // Wrong output data type - }), - framework::dataset::make("InputBInfo",{ TensorInfo(TensorShape(21U), 1, DataType::S32), - TensorInfo(TensorShape(20U), 1, DataType::S32), - })), - framework::dataset::make("OutputInfo",{ TensorInfo(TensorShape(21U, 13U), 1, DataType::QASYMM8), - TensorInfo(TensorShape(20U, 13U), 1, DataType::S32), - })), - framework::dataset::make("Min",{ 0, - 13, - })), - framework::dataset::make("Max",{ 205, - 180, - })), - framework::dataset::make("Expected", { true, false })), - a_info, b_info, output_info, min, max, expected) -{ - - GEMMLowpOutputStageInfo output_stage = GEMMLowpOutputStageInfo(); - output_stage.type = GEMMLowpOutputStageType::QUANTIZE_DOWN; - output_stage.gemmlowp_min_bound = min; - output_stage.gemmlowp_max_bound = max; - output_stage.output_data_type = DataType::QASYMM8; - - // Lock tensors - Status status = NEGEMMLowpOutputStage::validate(&a_info.clone()->set_is_resizable(false), - &b_info.clone()->set_is_resizable(false), - &output_info.clone()->set_is_resizable(false), - output_stage); - ARM_COMPUTE_EXPECT(bool(status) == expected, framework::LogLevel::ERRORS); -} -// clang-format on -// *INDENT-ON* - -TEST_CASE(NoPaddingAdded, framework::DatasetMode::PRECOMMIT) -{ - Tensor input1 = create_tensor(TensorShape(21U, 13U), DataType::S32); - Tensor input2 = create_tensor(TensorShape(21U, 1U), DataType::S32); - Tensor output = create_tensor(TensorShape(21U, 13U), DataType::QASYMM8); - - GEMMLowpOutputStageInfo output_stage = GEMMLowpOutputStageInfo(); - output_stage.type = GEMMLowpOutputStageType::QUANTIZE_DOWN; - output_stage.gemmlowp_min_bound = 0; - output_stage.gemmlowp_max_bound = 205; - output_stage.output_data_type = DataType::QASYMM8; - - NEGEMMLowpOutputStage f; - f.configure(&input1, &input2, &output, output_stage); - - // Validate padding is zero - validate(input1.info()->padding(), PaddingSize()); - validate(input2.info()->padding(), PaddingSize()); - validate(output.info()->padding(), PaddingSize()); -} - -FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMLowpQuantizeDownInt32ScaleFixture, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), quantize_down_int32_to_uint8_scale_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} - -TEST_SUITE(BoundedReLu) -FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMLowpQuantizeDownInt32ScaleFixture, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), quantize_down_int32_to_uint8_scale_relu_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} - -TEST_SUITE_END() // BoundedReLu - -TEST_SUITE_END() // QASYMM8 - -TEST_SUITE(QASYMM8_SIGNED) - -const auto quantize_down_int32_to_int8_scale_cases = framework::dataset::make("result_offset", -2, 1) * framework::dataset::make("result_mult_int", 1, 2) * framework::dataset::make("result_shift", 2, - 3) - * framework::dataset::make("min", 0) * framework::dataset::make("max", 0) * framework::dataset::make("addBias", { false, true }); - -const auto quantize_down_int32_to_int8_scale_relu_cases = framework::dataset::make("result_offset", -2, 1) * framework::dataset::make("result_mult_int", 1, - 2) - * framework::dataset::make("result_shift", 2, 3) * framework::dataset::make("min", -100, -98) * framework::dataset::make("max", 71, 74) * framework::dataset::make("addBias", { false, true }); - -using NEGEMMLowpQuantizeDownInt32ScaleFixture = GEMMLowpQuantizeDownInt32ToInt8ScaleValidationFixture; - -// *INDENT-OFF* -// clang-format off -DATA_TEST_CASE(Validate, framework::DatasetMode::ALL, zip(zip(zip(zip(zip( - framework::dataset::make("InputAInfo", { TensorInfo(TensorShape(21U, 13U), 1, DataType::S32), // Input not a multiple of 16 - TensorInfo(TensorShape(21U, 13U), 1, DataType::S32), // Invalid min and max - TensorInfo(TensorShape(20U, 13U), 1, DataType::S32), // Wrong output data type - }), - framework::dataset::make("InputBInfo",{ TensorInfo(TensorShape(21U), 1, DataType::S32), - TensorInfo(TensorShape(21U), 1, DataType::S32), - TensorInfo(TensorShape(20U), 1, DataType::S32), - })), - framework::dataset::make("OutputInfo",{ TensorInfo(TensorShape(21U, 13U), 1, DataType::QASYMM8_SIGNED), - TensorInfo(TensorShape(21U, 13U), 1, DataType::QASYMM8_SIGNED), - TensorInfo(TensorShape(20U, 13U), 1, DataType::S32), - })), - framework::dataset::make("Min",{ -10, - -200, - -113, - })), - framework::dataset::make("Max",{ 105, - 300, - -18, - })), - framework::dataset::make("Expected", { true, false, false })), - a_info, b_info, output_info, min, max, expected) -{ - GEMMLowpOutputStageInfo output_stage = GEMMLowpOutputStageInfo(); - output_stage.type = GEMMLowpOutputStageType::QUANTIZE_DOWN; - output_stage.gemmlowp_min_bound = min; - output_stage.gemmlowp_max_bound = max; - output_stage.output_data_type = DataType::QASYMM8_SIGNED; - - // Lock tensors - Status status = NEGEMMLowpOutputStage::validate(&a_info.clone()->set_is_resizable(false), - &b_info.clone()->set_is_resizable(false), - &output_info.clone()->set_is_resizable(false), - output_stage); - ARM_COMPUTE_EXPECT(bool(status) == expected, framework::LogLevel::ERRORS); -} -// clang-format on -// *INDENT-ON* - -FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMLowpQuantizeDownInt32ScaleFixture, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), quantize_down_int32_to_int8_scale_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} - -TEST_SUITE(BoundedReLu) -FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMLowpQuantizeDownInt32ScaleFixture, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), quantize_down_int32_to_int8_scale_relu_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} - -TEST_SUITE_END() // BoundedReLu - -TEST_SUITE_END() // QASYMM8_SIGNED - -TEST_SUITE_END() // QuantizeDownInt32Scale - -TEST_SUITE(QuantizeDownInt32ToUint8ScaleByFixedPoint) - -const auto quantize_down_int32_to_uint8_scale_by_fixedpoint_cases = framework::dataset::make("result_fixedpoint_multiplier", 254601600, 254601602) * framework::dataset::make("result_shift", 1, - 2) - * framework::dataset::make("result_offset_after_shift", 2, 3) * framework::dataset::make("min", 0) * framework::dataset::make("max", 255) * framework::dataset::make("addBias", { false, true }); - -const auto quantize_down_int32_to_uint8_scale_by_fixedpoint_relu_cases = framework::dataset::make("result_fixedpoint_multiplier", 254601600, 254601602) * framework::dataset::make("result_shift", 1, - 2) - * framework::dataset::make("result_offset_after_shift", 2, 3) * framework::dataset::make("min", 0, 2) * framework::dataset::make("max", 171, 174) * framework::dataset::make("addBias", { false, true }); - -using NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointFixture = - GEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointValidationFixture; - -using NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointFixture = - GEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointValidationFixture; - -// *INDENT-OFF* -// clang-format off -DATA_TEST_CASE(Validate, framework::DatasetMode::ALL, zip(zip(zip(zip(zip( - framework::dataset::make("InputAInfo", { TensorInfo(TensorShape(21U, 13U), 1, DataType::S32), // Input not a multiple of 16 - TensorInfo(TensorShape(20U, 13U), 1, DataType::S32), // Wrong output data type - }), - framework::dataset::make("InputBInfo",{ TensorInfo(TensorShape(21U), 1, DataType::S32), - TensorInfo(TensorShape(20U), 1, DataType::S32), - })), - framework::dataset::make("OutputInfo",{ TensorInfo(TensorShape(21U, 13U), 1, DataType::QASYMM8), - TensorInfo(TensorShape(20U, 13U), 1, DataType::S32), - })), - framework::dataset::make("Min",{ 0, - 13, - })), - framework::dataset::make("Max",{ 205, - 180, - })), - framework::dataset::make("Expected", { true, false })), - a_info, b_info, output_info, min, max, expected) -{ - // Lock tensors - Status status = NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPoint::validate(&a_info.clone()->set_is_resizable(false), - &b_info.clone()->set_is_resizable(false), - &output_info.clone()->set_is_resizable(false), - min, - max); - ARM_COMPUTE_EXPECT(bool(status) == expected, framework::LogLevel::ERRORS); -} -// clang-format on -// *INDENT-ON* - -FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointFixture, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), - quantize_down_int32_to_uint8_scale_by_fixedpoint_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} - -FIXTURE_DATA_TEST_CASE(RunLarge, NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointFixture, framework::DatasetMode::NIGHTLY, combine(datasets::LargeShapes(), - quantize_down_int32_to_uint8_scale_by_fixedpoint_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} - -TEST_SUITE(BoundedReLu) -FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointFixture, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), - quantize_down_int32_to_uint8_scale_by_fixedpoint_relu_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} - -FIXTURE_DATA_TEST_CASE(RunLarge, NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointFixture, framework::DatasetMode::NIGHTLY, combine(datasets::LargeShapes(), - quantize_down_int32_to_uint8_scale_by_fixedpoint_relu_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} -TEST_SUITE_END() // BoundedReLu - -TEST_SUITE_END() // QuantizeDownInt32ToUint8ScaleByFixedPoint - -TEST_SUITE(QuantizeDownInt32ToInt8ScaleByFixedPoint) - -const auto quantize_down_int32_to_int8_scale_by_fixedpoint_cases = framework::dataset::make("result_fixedpoint_multiplier", 254601600, 254601602) * framework::dataset::make("result_shift", 1, - 2) - * framework::dataset::make("result_offset_after_shift", 2, 3) * framework::dataset::make("min", -128) * framework::dataset::make("max", 128) * framework::dataset::make("addBias", { false, true }); - -const auto quantize_down_int32_to_int8_scale_by_fixedpoint_relu_cases = framework::dataset::make("result_fixedpoint_multiplier", 254601600, 254601602) * framework::dataset::make("result_shift", 1, - 2) - * framework::dataset::make("result_offset_after_shift", 2, 3) * framework::dataset::make("min", -2, 0) * framework::dataset::make("max", 1, 3) * framework::dataset::make("addBias", { false, true }); - -using NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointFixture = - GEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointValidationFixture; - -// *INDENT-OFF* -// clang-format off -DATA_TEST_CASE(Validate, framework::DatasetMode::ALL, zip(zip(zip(zip(zip( - framework::dataset::make("InputAInfo", { TensorInfo(TensorShape(21U, 13U), 1, DataType::F32), // Invalid input data type - TensorInfo(TensorShape(20U, 13U), 1, DataType::S32), // Wrong output data type - TensorInfo(TensorShape(21U, 13U), 1, DataType::S32), - }), - framework::dataset::make("InputBInfo",{ TensorInfo(TensorShape(21U), 1, DataType::S32), - TensorInfo(TensorShape(20U), 1, DataType::S32), - TensorInfo(TensorShape(21U), 1, DataType::S32), - })), - framework::dataset::make("OutputInfo",{ TensorInfo(TensorShape(21U, 13U), 1, DataType::QASYMM8_SIGNED), - TensorInfo(TensorShape(20U, 13U), 1, DataType::S32), - TensorInfo(TensorShape(21U, 13U), 1, DataType::QASYMM8_SIGNED), - })), - framework::dataset::make("Min",{ -110, - -113, - -113, - })), - framework::dataset::make("Max",{ 87, - 97, - 97, - })), - framework::dataset::make("Expected", { false, false, true })), - a_info, b_info, output_info, min, max, expected) -{ - // Lock tensors - Status status = NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPoint::validate(&a_info.clone()->set_is_resizable(false), - &b_info.clone()->set_is_resizable(false), - &output_info.clone()->set_is_resizable(false), - min, - max); - ARM_COMPUTE_EXPECT(bool(status) == expected, framework::LogLevel::ERRORS); -} -// clang-format on -// *INDENT-ON* - -FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointFixture, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), - quantize_down_int32_to_int8_scale_by_fixedpoint_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} - -TEST_SUITE(BoundedReLu) -FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointFixture, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), - quantize_down_int32_to_int8_scale_by_fixedpoint_relu_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} -TEST_SUITE_END() // BoundedReLu -TEST_SUITE_END() // QuantizeDownInt32ToInt8ScaleByFixedPoint - -TEST_SUITE(QuantizeDownInt32ToInt16ScaleByFixedPoint) - -const auto quantize_down_int32_to_int16_scale_by_fixedpoint_cases = framework::dataset::make("result_fixedpoint_multiplier", 254601600, 254601602) * framework::dataset::make("result_shift", 1, - 2) - * framework::dataset::make("min", -32768) * framework::dataset::make("max", 32767) * framework::dataset::make("addBias", { false, true }); - -const auto quantize_down_int32_to_int16_scale_by_fixedpoint_relu_cases = framework::dataset::make("result_fixedpoint_multiplier", 254601600, 254601602) * framework::dataset::make("result_shift", 1, - 2) - * framework::dataset::make("min", -2, 0) * framework::dataset::make("max", 1, 3) * framework::dataset::make("addBias", { false, true }); -const auto quantize_down_int32_to_int16_scale_by_fixedpoint_multgreat1_cases = framework::dataset::make("result_fixedpoint_multiplier", 1073741823, - 1073741825) - * framework::dataset::make("result_shift", -3, - -2) - * framework::dataset::make("min", -32768) * framework::dataset::make("max", 32767) * framework::dataset::make("addBias", { false, true }); - -const auto quantize_down_int32_to_int16_scale_by_fixedpoint_multgreat1_relu_cases = framework::dataset::make("result_fixedpoint_multiplier", 254601600, - 254601602) - * framework::dataset::make("result_shift", -3, - -1) - * framework::dataset::make("min", -2, 0) * framework::dataset::make("max", 1, 3) * framework::dataset::make("addBias", { false, true }); - -using NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointFixture = - GEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointValidationFixture; - -// *INDENT-OFF* -// clang-format off -DATA_TEST_CASE(Validate, framework::DatasetMode::ALL, zip(zip(zip(zip(zip( - framework::dataset::make("InputAInfo", { TensorInfo(TensorShape(21U, 13U), 1, DataType::S32), // Input not a multiple of 16 - TensorInfo(TensorShape(20U, 13U), 1, DataType::S32), // Wrong output data type - }), - framework::dataset::make("InputBInfo",{ TensorInfo(TensorShape(21U), 1, DataType::S32), - TensorInfo(TensorShape(20U), 1, DataType::S32), - })), - framework::dataset::make("OutputInfo",{ TensorInfo(TensorShape(21U, 13U), 1, DataType::QSYMM16), - TensorInfo(TensorShape(20U, 13U), 1, DataType::S32), - })), - framework::dataset::make("Min",{ -205, - -180, - })), - framework::dataset::make("Max",{ 205, - 180, - })), - framework::dataset::make("Expected", { true, false })), - a_info, b_info, output_info, min, max, expected) -{ - // Lock tensors - Status status = NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPoint::validate(&a_info.clone()->set_is_resizable(false), - &b_info.clone()->set_is_resizable(false), - &output_info.clone()->set_is_resizable(false), - min, - max); - ARM_COMPUTE_EXPECT(bool(status) == expected, framework::LogLevel::ERRORS); -} -// clang-format on -// *INDENT-ON* - -TEST_SUITE(NoRelu) -TEST_SUITE(MultSmallerEq1) -FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointFixture, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), - quantize_down_int32_to_int16_scale_by_fixedpoint_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} -TEST_SUITE_END() // MultSmallerEq1 -TEST_SUITE(MultGreater1) -FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointFixture, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), - quantize_down_int32_to_int16_scale_by_fixedpoint_multgreat1_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} -TEST_SUITE_END() // MultGreater1 -TEST_SUITE_END() // NoRelu -TEST_SUITE(BoundedReLu) -TEST_SUITE(MultSmallerEq1) -FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointFixture, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), - quantize_down_int32_to_int16_scale_by_fixedpoint_relu_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} -TEST_SUITE_END() // MultSmallerEq1 -TEST_SUITE(MultGreater1) -FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointFixture, framework::DatasetMode::ALL, combine(datasets::SmallShapes(), - quantize_down_int32_to_int16_scale_by_fixedpoint_multgreat1_relu_cases)) -{ - // Validate output - validate(Accessor(_target), _reference); -} -TEST_SUITE_END() // MultGreater1 -TEST_SUITE_END() // BoundedReLu -TEST_SUITE_END() // QuantizeDownInt32ToInt16ScaleByFixedPoint -TEST_SUITE_END() // OutputStage TEST_SUITE_END() // GEMMLowp -TEST_SUITE_END() // Neon +TEST_SUITE_END() // NEON } // namespace validation } // namespace test } // namespace arm_compute -- cgit v1.2.1