From a84faffd290139be54e4a52ab11da0369262e889 Mon Sep 17 00:00:00 2001 From: Georgios Pinitas Date: Wed, 5 Dec 2018 18:17:24 +0000 Subject: COMPMID-1826: Add support for QASYMM8 in NEArithmeticAdditionKernel Change-Id: Ia7fb128e1f3944d0d831e1d125a6db3e1d257106 Reviewed-on: https://review.mlplatform.org/355 Tested-by: Arm Jenkins Reviewed-by: Isabella Gottardi Reviewed-by: Anthony Barbier --- .../NEON/kernels/NEArithmeticAdditionKernel.cpp | 68 +++++++++++++++++++++- 1 file changed, 65 insertions(+), 3 deletions(-) (limited to 'src') diff --git a/src/core/NEON/kernels/NEArithmeticAdditionKernel.cpp b/src/core/NEON/kernels/NEArithmeticAdditionKernel.cpp index 169554f87a..954a2c1754 100644 --- a/src/core/NEON/kernels/NEArithmeticAdditionKernel.cpp +++ b/src/core/NEON/kernels/NEArithmeticAdditionKernel.cpp @@ -171,6 +171,61 @@ void add_F32_F32_F32(const ITensor *in1, const ITensor *in2, ITensor *out, const input1, input2, output); } +void add_QASYMM8_QASYMM8_QASYMM8(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 float32x4_t vscale1 = vdupq_n_f32(in1->info()->quantization_info().scale); + const float32x4_t vscale2 = vdupq_n_f32(in2->info()->quantization_info().scale); + const float32x4_t invvscaleo = vdupq_n_f32(1.f / out->info()->quantization_info().scale); + const int32x4_t voffset1 = vdupq_n_s32(in1->info()->quantization_info().offset); + const int32x4_t voffset2 = vdupq_n_s32(in2->info()->quantization_info().offset); + const float32x4_t voffseto = vdupq_n_f32(out->info()->quantization_info().offset); + + execute_window_loop(window, [&](const Coordinates & id) + { + const uint8x16_t a = vld1q_u8(input1.ptr()); + const uint8x16_t b = vld1q_u8(input2.ptr()); + + const float32x4x4_t af = + { + { + vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(vmovl_u8(vget_low_u8(a))))), voffset1)), vscale1), + vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(vmovl_u8(vget_low_u8(a))))), voffset1)), vscale1), + vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(vmovl_u8(vget_high_u8(a))))), voffset1)), vscale1), + vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(vmovl_u8(vget_high_u8(a))))), voffset1)), vscale1), + } + }; + + const float32x4x4_t bf = + { + { + vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(vmovl_u8(vget_low_u8(b))))), voffset2)), vscale2), + vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(vmovl_u8(vget_low_u8(b))))), voffset2)), vscale2), + vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(vmovl_u8(vget_high_u8(b))))), voffset2)), vscale2), + vmulq_f32(vcvtq_f32_s32(vsubq_s32(vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(vmovl_u8(vget_high_u8(b))))), voffset2)), vscale2), + } + }; + + const int32x4x4_t rf = + { + { + vcvtq_s32_f32(vmlaq_f32(voffseto, vaddq_f32(af.val[0], bf.val[0]), invvscaleo)), + vcvtq_s32_f32(vmlaq_f32(voffseto, vaddq_f32(af.val[1], bf.val[1]), invvscaleo)), + vcvtq_s32_f32(vmlaq_f32(voffseto, vaddq_f32(af.val[2], bf.val[2]), invvscaleo)), + vcvtq_s32_f32(vmlaq_f32(voffseto, vaddq_f32(af.val[3], bf.val[3]), invvscaleo)), + } + }; + + const uint8x8_t pa = vqmovun_s16(vcombine_s16(vqmovn_s32(rf.val[0]), vqmovn_s32(rf.val[1]))); + const uint8x8_t pb = vqmovun_s16(vcombine_s16(vqmovn_s32(rf.val[2]), vqmovn_s32(rf.val[3]))); + vst1q_u8(output.ptr(), vcombine_u8(pa, pb)); + }, + input1, input2, output); +} + void add_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())); @@ -332,8 +387,8 @@ Status validate_arguments(const ITensorInfo &input1, const ITensorInfo &input2, 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::S16, DataType::F16, DataType::F32); - ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&input2, 1, DataType::U8, DataType::S16, DataType::F16, DataType::F32); + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&input1, 1, DataType::U8, DataType::QASYMM8, DataType::S16, DataType::F16, DataType::F32); + ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(&input2, 1, DataType::U8, DataType::QASYMM8, DataType::S16, DataType::F16, DataType::F32); const TensorShape out_shape = TensorShape::broadcast_shape(input1.tensor_shape(), input2.tensor_shape()); @@ -349,7 +404,8 @@ Status validate_arguments(const ITensorInfo &input1, const ITensorInfo &input2, && !(input1.data_type() == DataType::S16 && input2.data_type() == DataType::U8 && output.data_type() == DataType::S16) && !(input1.data_type() == DataType::S16 && input2.data_type() == DataType::S16 && output.data_type() == DataType::S16) && !(input1.data_type() == DataType::F32 && input2.data_type() == DataType::F32 && output.data_type() == DataType::F32) - && !(input1.data_type() == DataType::F16 && input2.data_type() == DataType::F16 && output.data_type() == DataType::F16), + && !(input1.data_type() == DataType::F16 && input2.data_type() == DataType::F16 && output.data_type() == DataType::F16) + && !(input1.data_type() == DataType::QASYMM8 && input2.data_type() == DataType::QASYMM8 && output.data_type() == DataType::QASYMM8), "You called addition with the wrong image formats"); ARM_COMPUTE_RETURN_ERROR_ON_MSG(detail::have_different_dimensions(out_shape, output.tensor_shape(), 0), @@ -381,6 +437,10 @@ std::pair validate_and_configure_window(ITensorInfo &input1, ITe { set_format_if_unknown(output, Format::F32); } + else if(input1.data_type() == DataType::QASYMM8 || input2.data_type() == DataType::QASYMM8) + { + set_data_type_if_unknown(output, DataType::QASYMM8); + } } Window win = calculate_max_window(valid_region, Steps(num_elems_processed_per_iteration)); @@ -432,6 +492,8 @@ void NEArithmeticAdditionKernel::configure(const ITensor *input1, const ITensor { "add_saturate_F32_F32_F32", &add_F32_F32_F32 }, { "add_wrap_F16_F16_F16", &add_F16_F16_F16 }, { "add_saturate_F16_F16_F16", &add_F16_F16_F16 }, + { "add_wrap_QASYMM8_QASYMM8_QASYMM8", &add_QASYMM8_QASYMM8_QASYMM8 }, + { "add_saturate_QASYMM8_QASYMM8_QASYMM8", &add_QASYMM8_QASYMM8_QASYMM8 }, }; _input1 = input1; -- cgit v1.2.1