From 9f2403ff462c3725a03df68484dc43df6f011ab0 Mon Sep 17 00:00:00 2001 From: Michele Di Giorgio Date: Fri, 27 Mar 2020 10:23:44 +0000 Subject: COMPMID-3237: Add support for QSYMM16 ArithmeticSubtraction on NEON Change-Id: Ib38796e52665233351b181bf3417eb5650ad7ca7 Signed-off-by: Michele Di Giorgio Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/2939 Comments-Addressed: Arm Jenkins Reviewed-by: Michalis Spyrou Tested-by: Arm Jenkins --- .../NEON/kernels/NEArithmeticSubtractionKernel.cpp | 188 +++++++++++++++------ 1 file changed, 140 insertions(+), 48 deletions(-) (limited to 'src/core/NEON/kernels/NEArithmeticSubtractionKernel.cpp') diff --git a/src/core/NEON/kernels/NEArithmeticSubtractionKernel.cpp b/src/core/NEON/kernels/NEArithmeticSubtractionKernel.cpp index 0695c94927..9b7b235c9f 100644 --- a/src/core/NEON/kernels/NEArithmeticSubtractionKernel.cpp +++ b/src/core/NEON/kernels/NEArithmeticSubtractionKernel.cpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2016-2019 ARM Limited. + * Copyright (c) 2016-2020 ARM Limited. * * SPDX-License-Identifier: MIT * @@ -24,27 +24,13 @@ #include "arm_compute/core/NEON/kernels/NEArithmeticSubtractionKernel.h" #include "arm_compute/core/CPP/Validate.h" -#include "arm_compute/core/Error.h" -#include "arm_compute/core/Helpers.h" -#include "arm_compute/core/ITensor.h" #include "arm_compute/core/NEON/NEAsymm.h" -#include "arm_compute/core/NEON/NEFixedPoint.h" +#include "arm_compute/core/NEON/NESymm.h" #include "arm_compute/core/TensorInfo.h" #include "arm_compute/core/Validate.h" -#include -#include -#include -#include -#include - -using namespace arm_compute; - namespace arm_compute { -class Coordinates; -} // namespace arm_compute - namespace { constexpr unsigned int num_elems_processed_per_iteration = 16; @@ -145,6 +131,53 @@ void sub_saturate_QASYMM8_SIGNED_QASYMM8_SIGNED_QASYMM8_SIGNED(const ITensor *in input1, input2, output); } +void sub_saturate_QSYMM16_QSYMM16_QSYMM16(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window) +{ + Iterator input1(in1, window.broadcast_if_dimension_le_one(in1->info()->tensor_shape())); + Iterator input2(in2, window.broadcast_if_dimension_le_one(in2->info()->tensor_shape())); + Iterator output(out, window); + + const UniformQuantizationInfo iq1_info = in1->info()->quantization_info().uniform(); + const UniformQuantizationInfo iq2_info = in2->info()->quantization_info().uniform(); + const UniformQuantizationInfo oq_info = out->info()->quantization_info().uniform(); + + execute_window_loop(window, [&](const Coordinates &) + { + const int16x8x2_t in1_s16 = + { + { + vld1q_s16(reinterpret_cast(input1.ptr())), + vld1q_s16(reinterpret_cast(input1.ptr()) + 8), + } + }; + const int16x8x2_t in2_s16 = + { + { + vld1q_s16(reinterpret_cast(input2.ptr())), + vld1q_s16(reinterpret_cast(input2.ptr()) + 8), + } + }; + const float32x4x4_t ta1 = vdequantize(in1_s16, iq1_info); + const float32x4x4_t ta2 = vdequantize(in2_s16, iq2_info); + + const float32x4x4_t ta3 = + { + { + vsubq_f32(ta1.val[0], ta2.val[0]), + vsubq_f32(ta1.val[1], ta2.val[1]), + vsubq_f32(ta1.val[2], ta2.val[2]), + vsubq_f32(ta1.val[3], ta2.val[3]), + } + }; + + const int16x8x2_t result = vquantize_qsymm16(ta3, oq_info); + + vst1q_s16(reinterpret_cast(output.ptr()), result.val[0]); + vst1q_s16(reinterpret_cast(output.ptr()) + 8, result.val[1]); + }, + input1, input2, output); +} + void sub_wrap_S16_S16_S16(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window) { Iterator input1(in1, window.broadcast_if_dimension_le_one(in1->info()->tensor_shape())); @@ -153,8 +186,20 @@ void sub_wrap_S16_S16_S16(const ITensor *in1, const ITensor *in2, ITensor *out, execute_window_loop(window, [&](const Coordinates &) { - const int16x8x2_t ta1 = vld2q_s16(reinterpret_cast(input1.ptr())); - const int16x8x2_t ta2 = vld2q_s16(reinterpret_cast(input2.ptr())); + const int16x8x2_t ta1 = + { + { + vld1q_s16(reinterpret_cast(input1.ptr())), + vld1q_s16(reinterpret_cast(input1.ptr()) + 8), + } + }; + const int16x8x2_t ta2 = + { + { + vld1q_s16(reinterpret_cast(input2.ptr())), + vld1q_s16(reinterpret_cast(input2.ptr()) + 8), + } + }; const int16x8x2_t ta3 = { @@ -164,7 +209,8 @@ void sub_wrap_S16_S16_S16(const ITensor *in1, const ITensor *in2, ITensor *out, } }; - vst2q_s16(reinterpret_cast(output.ptr()), ta3); + vst1q_s16(reinterpret_cast(output.ptr()), ta3.val[0]); + vst1q_s16(reinterpret_cast(output.ptr()) + 8, ta3.val[1]); }, input1, input2, output); } @@ -177,8 +223,20 @@ void sub_saturate_S16_S16_S16(const ITensor *in1, const ITensor *in2, ITensor *o execute_window_loop(window, [&](const Coordinates &) { - const int16x8x2_t ta1 = vld2q_s16(reinterpret_cast(input1.ptr())); - const int16x8x2_t ta2 = vld2q_s16(reinterpret_cast(input2.ptr())); + const int16x8x2_t ta1 = + { + { + vld1q_s16(reinterpret_cast(input1.ptr())), + vld1q_s16(reinterpret_cast(input1.ptr()) + 8), + } + }; + const int16x8x2_t ta2 = + { + { + vld1q_s16(reinterpret_cast(input2.ptr())), + vld1q_s16(reinterpret_cast(input2.ptr()) + 8), + } + }; const int16x8x2_t ta3 = { @@ -188,26 +246,12 @@ void sub_saturate_S16_S16_S16(const ITensor *in1, const ITensor *in2, ITensor *o } }; - vst2q_s16(reinterpret_cast(output.ptr()), ta3); + vst1q_s16(reinterpret_cast(output.ptr()), ta3.val[0]); + vst1q_s16(reinterpret_cast(output.ptr()) + 8, ta3.val[1]); }, input1, input2, output); } -#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC -inline float16x8x2_t vsub2q_f16(const float16x8x2_t &a, const float16x8x2_t &b) -{ - const float16x8x2_t res = - { - { - vsubq_f16(a.val[0], b.val[0]), - vsubq_f16(a.val[1], b.val[1]) - } - }; - - return res; -} -#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ - void sub_F16_F16_F16(const ITensor *in1, const ITensor *in2, ITensor *out, const Window &window) { #ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC @@ -217,10 +261,30 @@ void sub_F16_F16_F16(const ITensor *in1, const ITensor *in2, ITensor *out, const execute_window_loop(window, [&](const Coordinates &) { - const float16x8x2_t a = vld2q_f16(reinterpret_cast(input1.ptr())); - const float16x8x2_t b = vld2q_f16(reinterpret_cast(input2.ptr())); + const float16x8x2_t a = + { + { + vld1q_f16(reinterpret_cast(input1.ptr())), + vld1q_f16(reinterpret_cast(input1.ptr()) + 8), + } + }; + const float16x8x2_t b = + { + { + vld1q_f16(reinterpret_cast(input2.ptr())), + vld1q_f16(reinterpret_cast(input2.ptr()) + 8), + } + }; + const float16x8x2_t res = + { + { + vsubq_f16(a.val[0], b.val[0]), + vsubq_f16(a.val[1], b.val[1]), + } + }; - vst2q_f16(reinterpret_cast(output.ptr()), vsub2q_f16(a, b)); + vst1q_f16(reinterpret_cast(output.ptr()), res.val[0]); + vst1q_f16(reinterpret_cast(output.ptr()) + 8, res.val[1]); }, input1, input2, output); #else /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ @@ -240,8 +304,24 @@ void sub_F32_F32_F32(const ITensor *in1, const ITensor *in2, ITensor *out, const execute_window_loop(window, [&](const Coordinates &) { - const float32x4x4_t ta1 = vld4q_f32(reinterpret_cast(input1.ptr())); - const float32x4x4_t ta2 = vld4q_f32(reinterpret_cast(input2.ptr())); + const float32x4x4_t ta1 = + { + { + vld1q_f32(reinterpret_cast(input1.ptr())), + vld1q_f32(reinterpret_cast(input1.ptr()) + 4), + vld1q_f32(reinterpret_cast(input1.ptr()) + 8), + vld1q_f32(reinterpret_cast(input1.ptr()) + 12), + } + }; + const float32x4x4_t ta2 = + { + { + vld1q_f32(reinterpret_cast(input2.ptr())), + vld1q_f32(reinterpret_cast(input2.ptr()) + 4), + vld1q_f32(reinterpret_cast(input2.ptr()) + 8), + vld1q_f32(reinterpret_cast(input2.ptr()) + 12), + } + }; const float32x4x4_t ta3 = { @@ -253,7 +333,10 @@ void sub_F32_F32_F32(const ITensor *in1, const ITensor *in2, ITensor *out, const } }; - vst4q_f32(reinterpret_cast(output.ptr()), ta3); + vst1q_f32(reinterpret_cast(output.ptr()), ta3.val[0]); + vst1q_f32(reinterpret_cast(output.ptr()) + 4, ta3.val[1]); + vst1q_f32(reinterpret_cast(output.ptr()) + 8, ta3.val[2]); + vst1q_f32(reinterpret_cast(output.ptr()) + 12, ta3.val[3]); }, input1, input2, output); } @@ -389,9 +472,9 @@ inline Status validate_arguments(const ITensorInfo &input1, const ITensorInfo &i { ARM_COMPUTE_UNUSED(policy); ARM_COMPUTE_RETURN_ERROR_ON_CPU_F16_UNSUPPORTED(&input1); - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&input1, 1, DataType::U8, DataType::QASYMM8, DataType::QASYMM8_SIGNED, DataType::S16, DataType::F16, DataType::F32); - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&input2, 1, DataType::U8, DataType::QASYMM8, DataType::QASYMM8_SIGNED, DataType::S16, DataType::F16, DataType::F32); - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&output, 1, DataType::U8, DataType::QASYMM8, DataType::QASYMM8_SIGNED, DataType::S16, DataType::F16, DataType::F32); + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&input1, 1, DataType::U8, DataType::QASYMM8, DataType::QASYMM8_SIGNED, DataType::QSYMM16, DataType::S16, DataType::F16, DataType::F32); + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&input2, 1, DataType::U8, DataType::QASYMM8, DataType::QASYMM8_SIGNED, DataType::QSYMM16, DataType::S16, DataType::F16, DataType::F32); + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&output, 1, DataType::U8, DataType::QASYMM8, DataType::QASYMM8_SIGNED, DataType::QSYMM16, DataType::S16, DataType::F16, DataType::F32); const TensorShape out_shape = TensorShape::broadcast_shape(input1.tensor_shape(), input2.tensor_shape()); ARM_COMPUTE_RETURN_ERROR_ON_MSG(out_shape.total_size() == 0, "Inputs are not broadcast compatible"); @@ -400,6 +483,7 @@ inline Status validate_arguments(const ITensorInfo &input1, const ITensorInfo &i !(input1.data_type() == DataType::U8 && input2.data_type() == DataType::U8) && !(input1.data_type() == DataType::QASYMM8 && input2.data_type() == DataType::QASYMM8) && !(input1.data_type() == DataType::QASYMM8_SIGNED && input2.data_type() == DataType::QASYMM8_SIGNED) + && !(input1.data_type() == DataType::QSYMM16 && input2.data_type() == DataType::QSYMM16) && !(input1.data_type() == DataType::U8 && input2.data_type() == DataType::U8) && !(input1.data_type() == DataType::U8 && input2.data_type() == DataType::S16) && !(input1.data_type() == DataType::S16 && input2.data_type() == DataType::U8) @@ -410,7 +494,8 @@ inline Status validate_arguments(const ITensorInfo &input1, const ITensorInfo &i ARM_COMPUTE_RETURN_ERROR_ON_MSG( input1.data_type() == DataType::QASYMM8_SIGNED && input2.data_type() == DataType::QASYMM8_SIGNED && policy == ConvertPolicy::WRAP - && input1.data_type() == DataType::QASYMM8 && input2.data_type() == DataType::QASYMM8 && policy == ConvertPolicy::WRAP, + && input1.data_type() == DataType::QASYMM8 && input2.data_type() == DataType::QASYMM8 && policy == ConvertPolicy::WRAP + && input1.data_type() == DataType::QSYMM16 && input2.data_type() == DataType::QSYMM16 && policy == ConvertPolicy::WRAP, "Convert policy cannot be WRAP if datatype is QASYMM8 or QASYMM8_SIGNED"); // Validate in case of configured output @@ -420,6 +505,7 @@ inline Status validate_arguments(const ITensorInfo &input1, const ITensorInfo &i !(input1.data_type() == DataType::U8 && input2.data_type() == DataType::U8 && output.data_type() == DataType::U8) && !(input1.data_type() == DataType::QASYMM8 && input2.data_type() == DataType::QASYMM8 && output.data_type() == DataType::QASYMM8) && !(input1.data_type() == DataType::QASYMM8_SIGNED && input2.data_type() == DataType::QASYMM8_SIGNED && output.data_type() == DataType::QASYMM8_SIGNED) + && !(input1.data_type() == DataType::QSYMM16 && input2.data_type() == DataType::QSYMM16 && output.data_type() == DataType::QSYMM16) && !(input1.data_type() == DataType::U8 && input2.data_type() == DataType::U8 && output.data_type() == DataType::S16) && !(input1.data_type() == DataType::U8 && input2.data_type() == DataType::S16 && output.data_type() == DataType::S16) && !(input1.data_type() == DataType::S16 && input2.data_type() == DataType::U8 && output.data_type() == DataType::S16) @@ -464,6 +550,10 @@ inline std::pair validate_and_configure_window(ITensorInfo &inpu { set_data_type_if_unknown(output, DataType::QASYMM8_SIGNED); } + else if(input1.data_type() == DataType::QSYMM16 || input2.data_type() == DataType::QSYMM16) + { + set_data_type_if_unknown(output, DataType::QSYMM16); + } } Window win = calculate_max_window(valid_region, Steps(num_elems_processed_per_iteration)); @@ -507,6 +597,7 @@ void NEArithmeticSubtractionKernel::configure(const ITensor *input1, const ITens { "sub_saturate_U8_U8_S16", &sub_saturate_U8_U8_S16 }, { "sub_saturate_QASYMM8_QASYMM8_QASYMM8", &sub_saturate_QAYSMM8_QAYSMM8_QAYSMM8 }, { "sub_saturate_QASYMM8_SIGNED_QASYMM8_SIGNED_QASYMM8_SIGNED", &sub_saturate_QASYMM8_SIGNED_QASYMM8_SIGNED_QASYMM8_SIGNED }, + { "sub_saturate_QSYMM16_QSYMM16_QSYMM16", &sub_saturate_QSYMM16_QSYMM16_QSYMM16 }, { "sub_wrap_U8_S16_S16", &sub_wrap_U8_S16_S16 }, { "sub_wrap_S16_U8_S16", &sub_wrap_S16_U8_S16 }, { "sub_saturate_U8_S16_S16", &sub_saturate_U8_S16_S16 }, @@ -564,4 +655,5 @@ BorderSize NEArithmeticSubtractionKernel::border_size() const const unsigned int replicateSize = _output->info()->dimension(0) - std::min(_input1->info()->dimension(0), _input2->info()->dimension(0)); const unsigned int border = std::min(num_elems_processed_per_iteration - 1U, replicateSize); return BorderSize{ 0, border, 0, 0 }; -} \ No newline at end of file +} +} // namespace arm_compute \ No newline at end of file -- cgit v1.2.1