diff options
author | Gunes Bayir <gunes.bayir@arm.com> | 2022-09-11 15:59:19 +0100 |
---|---|---|
committer | Gunes Bayir <gunes.bayir@arm.com> | 2022-09-16 12:38:59 +0000 |
commit | c4f2743951473f8d97f5a43767fdbb31a4df967c (patch) | |
tree | 26c49b1af8113bb169931f3af5e502904d455a25 /src/cpu/kernels/scale/neon/integer.cpp | |
parent | 0d05b6690fe69c57f63ca43d59b551f074613062 (diff) | |
download | ComputeLibrary-c4f2743951473f8d97f5a43767fdbb31a4df967c.tar.gz |
Optimize Quantized/Integer Bilinear Scale for Neon™
This patch introduces several performance optimizations regarding the Bilinear Scale operator with REPLICATE Border mode. Changes apply only to NHWC.
This patch
- Reduces the memory footprint by disabling precomputation of indices and weights when they're not used
- Rewrites the kernels for QASYMM8/QASYMM8_SIGNED/U8(Uint8)
- Adds S8(Int8) Bilinear Scale for Border mode REPLICATE
- Removes Bilinear Scale SVE kernels for Quantized and Integer types and adjust the heuristics to choose the Neon™ implementation
- Adds new test cases where the input and output of the Bilinear Scale operator have different quantization scale and offset
Resolves: COMPMID-5453, COMPMID-5454
Change-Id: I3d251e76e0c6978fd5a0a1795ec62ab536bec93c
Signed-off-by: Gunes Bayir <gunes.bayir@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8250
Reviewed-by: SiCong Li <sicong.li@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/cpu/kernels/scale/neon/integer.cpp')
-rw-r--r-- | src/cpu/kernels/scale/neon/integer.cpp | 474 |
1 files changed, 427 insertions, 47 deletions
diff --git a/src/cpu/kernels/scale/neon/integer.cpp b/src/cpu/kernels/scale/neon/integer.cpp index a2359aac94..2ab14cf83a 100644 --- a/src/cpu/kernels/scale/neon/integer.cpp +++ b/src/cpu/kernels/scale/neon/integer.cpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2021 Arm Limited. + * Copyright (c) 2021-2022 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -22,17 +22,12 @@ * SOFTWARE. */ #include "arm_compute/core/Helpers.h" -#include "arm_compute/core/ITensorPack.h" -#include "arm_compute/core/Window.h" -#include "src/core/NEON/NEMath.h" #include "src/core/NEON/wrapper/wrapper.h" #include "src/core/helpers/ScaleHelpers.h" #include "src/core/utils/ScaleUtils.h" #include "support/Rounding.h" #include <arm_neon.h> -#include <cmath> -#include <cstddef> namespace arm_compute { @@ -84,37 +79,39 @@ void u8_neon_scale_bilinear(const ITensor *src, ITensor *dst, const ITensor *off BorderMode border_mode, PixelValue constant_border_value, float sampling_offset, bool align_corners, const Window &window) { - // Compute the ratio between source height and destination height - const auto hr = scale_utils::calculate_resize_ratio(src->info()->dimension(2), dst->info()->dimension(2), align_corners); - - Iterator out(dst, window); - const int in_stride_c = src->info()->dimension(0) + src->info()->padding().left + src->info()->padding().right; - const int in_dim_w = src->info()->dimension(1); - const int in_dim_h = src->info()->dimension(2); - const int in_stride_wc = in_stride_c * (in_dim_w + src->info()->padding().top + src->info()->padding().bottom); + // Compute the ratio between source and destination dimensions + const float scale_x = scale_utils::calculate_resize_ratio(src->info()->dimension(1), dst->info()->dimension(1), align_corners); + const float scale_y = scale_utils::calculate_resize_ratio(src->info()->dimension(2), dst->info()->dimension(2), align_corners); - // Don't increment in Y and Z direction for the input tensor - // A pointer to the start of this plane is needed as base for the precomputed offsets - Window win_in(window); - win_in.set(Window::DimY, Window::Dimension(0, 0, 0)); - win_in.set(Window::DimZ, Window::Dimension(0, 0, 0)); - Iterator in(src, win_in); + const int input_width = src->info()->dimension(1); + const int input_height = src->info()->dimension(2); if(border_mode == BorderMode::CONSTANT) { + Iterator out(dst, window); + const int in_stride_c = src->info()->dimension(0) + src->info()->padding().left + src->info()->padding().right; + const int in_stride_wc = in_stride_c * (input_width + src->info()->padding().top + src->info()->padding().bottom); + + // Don't increment in Y and Z direction for the input tensor + // A pointer to the start of this plane is needed as base for the precomputed offsets + Window win_in(window); + win_in.set(Window::DimY, Window::Dimension(0, 0, 0)); + win_in.set(Window::DimZ, Window::Dimension(0, 0, 0)); + Iterator in(src, win_in); + const uint8_t const_border_value = static_cast<uint8_t>(constant_border_value.get<uint8_t>()); execute_window_loop(window, [&](const Coordinates & id) { const auto offset = *reinterpret_cast<const int32_t *>(offsets->ptr_to_element(Coordinates(id.y(), id.z()))); const auto dx_val = *reinterpret_cast<const float *>(dx->ptr_to_element(Coordinates(id.y(), id.z()))); const auto dy_val = *reinterpret_cast<const float *>(dy->ptr_to_element(Coordinates(id.y(), id.z()))); - const int32_t in_hi = std::floor((id.z() + sampling_offset) * hr - sampling_offset); + const int32_t in_hi = std::floor((id.z() + sampling_offset) * scale_y - sampling_offset); const uint8_t *in_ptr = reinterpret_cast<const uint8_t *>(in.ptr()) + offset * in_stride_c + in_hi * in_stride_wc; - const auto a00 = (0 <= offset && offset < in_dim_w && 0 <= in_hi && in_hi < in_dim_h) ? *in_ptr : const_border_value; - const auto a01 = (-1 <= offset && offset < in_dim_w - 1 && 0 <= in_hi && in_hi < in_dim_h) ? *(in_ptr + in_stride_c) : const_border_value; - const auto a10 = (0 <= offset && offset < in_dim_w && -1 <= in_hi && in_hi < in_dim_h - 1) ? *(in_ptr + in_stride_wc) : const_border_value; - const auto a11 = (-1 <= offset && offset < in_dim_w - 1 && -1 <= in_hi && in_hi < in_dim_h - 1) ? *(in_ptr + in_stride_c + in_stride_wc) : const_border_value; + const auto a00 = (0 <= offset && offset < input_width && 0 <= in_hi && in_hi < input_height) ? *in_ptr : const_border_value; + const auto a01 = (-1 <= offset && offset < input_width - 1 && 0 <= in_hi && in_hi < input_height) ? *(in_ptr + in_stride_c) : const_border_value; + const auto a10 = (0 <= offset && offset < input_width && -1 <= in_hi && in_hi < input_height - 1) ? *(in_ptr + in_stride_wc) : const_border_value; + const auto a11 = (-1 <= offset && offset < input_width - 1 && -1 <= in_hi && in_hi < input_height - 1) ? *(in_ptr + in_stride_c + in_stride_wc) : const_border_value; *reinterpret_cast<uint8_t *>(out.ptr()) = static_cast<uint8_t>(scale_helpers::delta_bilinear(a00, a01, a10, a11, dx_val, dy_val)); }, @@ -122,26 +119,395 @@ void u8_neon_scale_bilinear(const ITensor *src, ITensor *dst, const ITensor *off } else if(border_mode == BorderMode::REPLICATE) { - execute_window_loop(window, [&](const Coordinates & id) + using ExactTagType = typename wrapper::traits::neon_bitvector_tag_t<float, wrapper::traits::BitWidth::W128>; + + const int in_stride_x = src->info()->strides_in_bytes()[1]; + const int in_stride_y = src->info()->strides_in_bytes()[2]; + const int in_stride_b = src->info()->strides_in_bytes()[3]; + const int out_stride_x = dst->info()->strides_in_bytes()[1]; + const int out_stride_y = dst->info()->strides_in_bytes()[2]; + const int out_stride_b = dst->info()->strides_in_bytes()[3]; + + const int out_dim_ch = dst->info()->dimension(0); + constexpr int step_cout = 16; + + Window window_execution = window; + window_execution.set(Window::DimX, Window::Dimension(0, 1, 1)); + Window win_in_out(window); + win_in_out.set(Window::DimY, Window::Dimension(0, 0, 0)); + win_in_out.set(Window::DimZ, Window::Dimension(0, 0, 0)); + Iterator in(src, win_in_out); + Iterator out(dst, win_in_out); + + const int xo_start = window_execution[1].start(); + const int xo_end = window_execution[1].end(); + const int xo_step = window_execution[1].step(); + const int yo_start = window_execution[2].start(); + const int yo_end = window_execution[2].end(); + const int yo_step = window_execution[2].step(); + const int bo_start = window_execution[3].start(); + const int bo_end = window_execution[3].end(); + const int bo_step = window_execution[3].step(); + + const float fp_coord_offset_y = sampling_offset * (scale_y - 1); + const float fp_coord_offset_x = sampling_offset * (scale_x - 1); + + for(int bo = bo_start; bo < bo_end; bo += bo_step) { - const auto offset = *reinterpret_cast<const int32_t *>(offsets->ptr_to_element(Coordinates(id.y(), id.z()))); - const auto dx_val = *reinterpret_cast<const float *>(dx->ptr_to_element(Coordinates(id.y(), id.z()))); - const auto dy_val = *reinterpret_cast<const float *>(dy->ptr_to_element(Coordinates(id.y(), id.z()))); - const int in_hi = std::floor((id.z() + sampling_offset) * hr - sampling_offset); - - auto clamped_w = utility::clamp<int>(offset, 0, in_dim_w - 1); - auto clamped_w1 = utility::clamp<int>(offset + 1, 0, in_dim_w - 1); - auto clamped_h = utility::clamp<int>(in_hi, 0, in_dim_h - 1); - auto clamped_h1 = utility::clamp<int>(in_hi + 1, 0, in_dim_h - 1); - - const auto a00 = *(reinterpret_cast<const uint8_t *>(in.ptr()) + clamped_w * in_stride_c + clamped_h * in_stride_wc); - const auto a01 = *(reinterpret_cast<const uint8_t *>(in.ptr()) + clamped_w1 * in_stride_c + clamped_h * in_stride_wc); - const auto a10 = *(reinterpret_cast<const uint8_t *>(in.ptr()) + clamped_w * in_stride_c + clamped_h1 * in_stride_wc); - const auto a11 = *(reinterpret_cast<const uint8_t *>(in.ptr()) + clamped_w1 * in_stride_c + clamped_h1 * in_stride_wc); + const uint8_t *in_ptr = in.ptr() + bo * in_stride_b; + uint8_t *out_ptr = out.ptr() + bo * out_stride_b; + + for(int yo = yo_start; yo < yo_end; yo += yo_step) + { + // Floating-point coordinate + const float yi_f = yo * scale_y + fp_coord_offset_y; + // Integer coordinate + const int yi = static_cast<int>(std::floor(yi_f)); + // Weight for the y coordinate + const float a1 = (yi_f - static_cast<float>(yi)); + const float b1 = (1.f - a1); + + const int yi0 = utility::clamp<int>(yi, 0, input_height - 1); + const int yi1 = utility::clamp<int>(yi + 1, 0, input_height - 1); + + const uint8_t *in_ptr_yi0 = in_ptr + yi0 * in_stride_y; + const uint8_t *in_ptr_yi1 = in_ptr + yi1 * in_stride_y; + + uint8_t *out_ptr_yo = out_ptr + yo * out_stride_y; + for(int xo = xo_start; xo < xo_end; xo += xo_step) + { + // Floating-point coordinate + const float xi_f = xo * scale_x + fp_coord_offset_x; + // Integer coordinate + const int xi = static_cast<int>(std::floor(xi_f)); + // Weight for the x coordinate + const float a = (xi_f - static_cast<float>(xi)); + const float b = (1.f - a); + + const float s00_s = b * b1; + const float s01_s = a * b1; + const float s10_s = b * a1; + const float s11_s = a * a1; + + const auto s00 = wrapper::vdup_n(s00_s, ExactTagType{}); + const auto s01 = wrapper::vdup_n(s01_s, ExactTagType{}); + const auto s10 = wrapper::vdup_n(s10_s, ExactTagType{}); + const auto s11 = wrapper::vdup_n(s11_s, ExactTagType{}); + + const int xi0 = utility::clamp<int>(xi, 0, input_width - 1); + const int xi1 = utility::clamp<int>(xi + 1, 0, input_width - 1); + + const auto in_ptr_xi0_yi0 = in_ptr_yi0 + xi0 * in_stride_x; + const auto in_ptr_xi1_yi0 = in_ptr_yi0 + xi1 * in_stride_x; + const auto in_ptr_xi0_yi1 = in_ptr_yi1 + xi0 * in_stride_x; + const auto in_ptr_xi1_yi1 = in_ptr_yi1 + xi1 * in_stride_x; + + uint8_t *out_ptr_xo_yo = out_ptr_yo + xo * out_stride_x; + + int cout = 0; + for(; cout <= (out_dim_ch - step_cout); cout += step_cout) + { + const auto in00 = wrapper::vloadq(in_ptr_xi0_yi0 + cout * sizeof(uint8_t)); + const auto in01 = wrapper::vloadq(in_ptr_xi1_yi0 + cout * sizeof(uint8_t)); + const auto in10 = wrapper::vloadq(in_ptr_xi0_yi1 + cout * sizeof(uint8_t)); + const auto in11 = wrapper::vloadq(in_ptr_xi1_yi1 + cout * sizeof(uint8_t)); + + const uint16x8_t in00_low = wrapper::vmovl(wrapper::vgetlow(in00)); + const uint16x8_t in00_high = wrapper::vmovl(wrapper::vgethigh(in00)); + + const auto in00_0 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in00_low))); + const auto in00_1 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in00_low))); + const auto in00_2 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in00_high))); + const auto in00_3 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in00_high))); + + const uint16x8_t in01_low = wrapper::vmovl(wrapper::vgetlow(in01)); + const uint16x8_t in01_high = wrapper::vmovl(wrapper::vgethigh(in01)); + + const auto in01_0 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in01_low))); + const auto in01_1 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in01_low))); + const auto in01_2 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in01_high))); + const auto in01_3 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in01_high))); + + const uint16x8_t in10_low = wrapper::vmovl(wrapper::vgetlow(in10)); + const uint16x8_t in10_high = wrapper::vmovl(wrapper::vgethigh(in10)); + + const auto in10_0 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in10_low))); + const auto in10_1 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in10_low))); + const auto in10_2 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in10_high))); + const auto in10_3 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in10_high))); + + const uint16x8_t in11_low = wrapper::vmovl(wrapper::vgetlow(in11)); + const uint16x8_t in11_high = wrapper::vmovl(wrapper::vgethigh(in11)); + + const auto in11_0 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in11_low))); + const auto in11_1 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in11_low))); + const auto in11_2 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in11_high))); + const auto in11_3 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in11_high))); + + auto out_0 = wrapper::vmul(in00_0, s00); + out_0 = wrapper::vmla(out_0, in01_0, s01); + out_0 = wrapper::vmla(out_0, in10_0, s10); + out_0 = wrapper::vmla(out_0, in11_0, s11); + + auto out_1 = wrapper::vmul(in00_1, s00); + out_1 = wrapper::vmla(out_1, in01_1, s01); + out_1 = wrapper::vmla(out_1, in10_1, s10); + out_1 = wrapper::vmla(out_1, in11_1, s11); + + auto out_2 = wrapper::vmul(in00_2, s00); + out_2 = wrapper::vmla(out_2, in01_2, s01); + out_2 = wrapper::vmla(out_2, in10_2, s10); + out_2 = wrapper::vmla(out_2, in11_2, s11); + + auto out_3 = wrapper::vmul(in00_3, s00); + out_3 = wrapper::vmla(out_3, in01_3, s01); + out_3 = wrapper::vmla(out_3, in10_3, s10); + out_3 = wrapper::vmla(out_3, in11_3, s11); + +#if defined(__aarch64__) && !defined(BARE_METAL) + const auto out_0_int = wrapper::vcvta<uint32_t>(out_0); + const auto out_1_int = wrapper::vcvta<uint32_t>(out_1); + const auto out_2_int = wrapper::vcvta<uint32_t>(out_2); + const auto out_3_int = wrapper::vcvta<uint32_t>(out_3); +#else // defined(__aarch64__) && !defined(BARE_METAL) + const auto out_0_int = wrapper::vcvt<uint32_t>(out_0); + const auto out_1_int = wrapper::vcvt<uint32_t>(out_1); + const auto out_2_int = wrapper::vcvt<uint32_t>(out_2); + const auto out_3_int = wrapper::vcvt<uint32_t>(out_3); +#endif // defined(__aarch64__) && !defined(BARE_METAL) + const auto low_part = wrapper::vqmovn(wrapper::vcombine(wrapper::vqmovn(out_0_int), wrapper::vqmovn(out_1_int))); + const auto high_part = wrapper::vqmovn(wrapper::vcombine(wrapper::vqmovn(out_2_int), wrapper::vqmovn(out_3_int))); + const auto out = wrapper::vcombine(low_part, high_part); + + wrapper::vstore(out_ptr_xo_yo + cout * sizeof(uint8_t), out); + } + + for(; cout < out_dim_ch; ++cout) + { + const uint8_t in00 = *(in_ptr_xi0_yi0 + cout * sizeof(uint8_t)); + const uint8_t in01 = *(in_ptr_xi1_yi0 + cout * sizeof(uint8_t)); + const uint8_t in10 = *(in_ptr_xi0_yi1 + cout * sizeof(uint8_t)); + const uint8_t in11 = *(in_ptr_xi1_yi1 + cout * sizeof(uint8_t)); + + float out0 = in00 * s00_s; + out0 += in01 * s01_s; + out0 += in10 * s10_s; + out0 += in11 * s11_s; + + // Rounding modes of vector and scalar loops should match +#if defined(__aarch64__) && !defined(BARE_METAL) + *(out_ptr_xo_yo + cout * sizeof(uint8_t)) = static_cast<uint8_t>(std::round(out0)); +#else // defined(__aarch64__) && !defined(BARE_METAL) + *(out_ptr_xo_yo + cout * sizeof(uint8_t)) = static_cast<uint8_t>(out0); +#endif // defined(__aarch64__) && !defined(BARE_METAL) + } + } + } + } + } + else + { + ARM_COMPUTE_ERROR("Not implemented"); + } +} - *reinterpret_cast<uint8_t *>(out.ptr()) = static_cast<uint8_t>(scale_helpers::delta_bilinear(a00, a01, a10, a11, dx_val, dy_val)); - }, - in, out); +void s8_neon_scale_bilinear(const ITensor *src, ITensor *dst, const ITensor *offsets, const ITensor *dx, const ITensor *dy, + BorderMode border_mode, PixelValue constant_border_value, float sampling_offset, + bool align_corners, const Window &window) +{ + ARM_COMPUTE_UNUSED(dx, dy, offsets, constant_border_value); + if(border_mode == BorderMode::REPLICATE) + { + using ExactTagType = typename wrapper::traits::neon_bitvector_tag_t<float, wrapper::traits::BitWidth::W128>; + + // Compute the ratio between source and destination dimensions + const float scale_x = scale_utils::calculate_resize_ratio(src->info()->dimension(1), dst->info()->dimension(1), align_corners); + const float scale_y = scale_utils::calculate_resize_ratio(src->info()->dimension(2), dst->info()->dimension(2), align_corners); + + const int in_stride_x = src->info()->strides_in_bytes()[1]; + const int in_stride_y = src->info()->strides_in_bytes()[2]; + const int in_stride_b = src->info()->strides_in_bytes()[3]; + const int out_stride_x = dst->info()->strides_in_bytes()[1]; + const int out_stride_y = dst->info()->strides_in_bytes()[2]; + const int out_stride_b = dst->info()->strides_in_bytes()[3]; + const int input_width = src->info()->dimension(1); + const int input_height = src->info()->dimension(2); + const int out_dim_ch = dst->info()->dimension(0); + constexpr int step_cout = 16; + + Window window_execution = window; + window_execution.set(Window::DimX, Window::Dimension(0, 1, 1)); + Window win_in_out(window); + win_in_out.set(Window::DimY, Window::Dimension(0, 0, 0)); + win_in_out.set(Window::DimZ, Window::Dimension(0, 0, 0)); + Iterator in(src, win_in_out); + Iterator out(dst, win_in_out); + + const int xo_start = window_execution[1].start(); + const int xo_end = window_execution[1].end(); + const int xo_step = window_execution[1].step(); + const int yo_start = window_execution[2].start(); + const int yo_end = window_execution[2].end(); + const int yo_step = window_execution[2].step(); + const int bo_start = window_execution[3].start(); + const int bo_end = window_execution[3].end(); + const int bo_step = window_execution[3].step(); + + const float fp_coord_offset_y = sampling_offset * (scale_y - 1); + const float fp_coord_offset_x = sampling_offset * (scale_x - 1); + + for(int bo = bo_start; bo < bo_end; bo += bo_step) + { + const int8_t *in_ptr = reinterpret_cast<int8_t *>(in.ptr() + bo * in_stride_b); + int8_t *out_ptr = reinterpret_cast<int8_t *>(out.ptr() + bo * out_stride_b); + + for(int yo = yo_start; yo < yo_end; yo += yo_step) + { + // Floating-point coordinate + const float yi_f = yo * scale_y + fp_coord_offset_y; + // Integer coordinate + const int yi = static_cast<int>(std::floor(yi_f)); + // Weight for the y coordinate + const float a1 = (yi_f - static_cast<float>(yi)); + const float b1 = (1.f - a1); + + const int yi0 = utility::clamp<int>(yi, 0, input_height - 1); + const int yi1 = utility::clamp<int>(yi + 1, 0, input_height - 1); + + const int8_t *in_ptr_yi0 = in_ptr + yi0 * in_stride_y; + const int8_t *in_ptr_yi1 = in_ptr + yi1 * in_stride_y; + + int8_t *out_ptr_yo = out_ptr + yo * out_stride_y; + for(int xo = xo_start; xo < xo_end; xo += xo_step) + { + // Floating-point coordinate + const float xi_f = xo * scale_x + fp_coord_offset_x; + // Integer coordinate + const int xi = static_cast<int>(std::floor(xi_f)); + // Weight for the x coordinate + const float a = (xi_f - static_cast<float>(xi)); + const float b = (1.f - a); + + const float s00_s = b * b1; + const float s01_s = a * b1; + const float s10_s = b * a1; + const float s11_s = a * a1; + + const auto s00 = wrapper::vdup_n(s00_s, ExactTagType{}); + const auto s01 = wrapper::vdup_n(s01_s, ExactTagType{}); + const auto s10 = wrapper::vdup_n(s10_s, ExactTagType{}); + const auto s11 = wrapper::vdup_n(s11_s, ExactTagType{}); + + const int xi0 = utility::clamp<int>(xi, 0, input_width - 1); + const int xi1 = utility::clamp<int>(xi + 1, 0, input_width - 1); + + const auto in_ptr_xi0_yi0 = in_ptr_yi0 + xi0 * in_stride_x; + const auto in_ptr_xi1_yi0 = in_ptr_yi0 + xi1 * in_stride_x; + const auto in_ptr_xi0_yi1 = in_ptr_yi1 + xi0 * in_stride_x; + const auto in_ptr_xi1_yi1 = in_ptr_yi1 + xi1 * in_stride_x; + + int8_t *out_ptr_xo_yo = out_ptr_yo + xo * out_stride_x; + + int cout = 0; + for(; cout <= (out_dim_ch - step_cout); cout += step_cout) + { + const auto in00 = wrapper::vloadq(in_ptr_xi0_yi0 + cout * sizeof(int8_t)); + const auto in01 = wrapper::vloadq(in_ptr_xi1_yi0 + cout * sizeof(int8_t)); + const auto in10 = wrapper::vloadq(in_ptr_xi0_yi1 + cout * sizeof(int8_t)); + const auto in11 = wrapper::vloadq(in_ptr_xi1_yi1 + cout * sizeof(int8_t)); + + const int16x8_t in00_low = wrapper::vmovl(wrapper::vgetlow(in00)); + const int16x8_t in00_high = wrapper::vmovl(wrapper::vgethigh(in00)); + + const auto in00_0 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in00_low))); + const auto in00_1 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in00_low))); + const auto in00_2 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in00_high))); + const auto in00_3 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in00_high))); + + const int16x8_t in01_low = wrapper::vmovl(wrapper::vgetlow(in01)); + const int16x8_t in01_high = wrapper::vmovl(wrapper::vgethigh(in01)); + + const auto in01_0 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in01_low))); + const auto in01_1 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in01_low))); + const auto in01_2 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in01_high))); + const auto in01_3 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in01_high))); + + const int16x8_t in10_low = wrapper::vmovl(wrapper::vgetlow(in10)); + const int16x8_t in10_high = wrapper::vmovl(wrapper::vgethigh(in10)); + + const auto in10_0 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in10_low))); + const auto in10_1 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in10_low))); + const auto in10_2 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in10_high))); + const auto in10_3 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in10_high))); + + const int16x8_t in11_low = wrapper::vmovl(wrapper::vgetlow(in11)); + const int16x8_t in11_high = wrapper::vmovl(wrapper::vgethigh(in11)); + + const auto in11_0 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in11_low))); + const auto in11_1 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in11_low))); + const auto in11_2 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgetlow(in11_high))); + const auto in11_3 = wrapper::vcvt<float>(wrapper::vmovl(wrapper::vgethigh(in11_high))); + + auto out_0 = wrapper::vmul(in00_0, s00); + out_0 = wrapper::vmla(out_0, in01_0, s01); + out_0 = wrapper::vmla(out_0, in10_0, s10); + out_0 = wrapper::vmla(out_0, in11_0, s11); + + auto out_1 = wrapper::vmul(in00_1, s00); + out_1 = wrapper::vmla(out_1, in01_1, s01); + out_1 = wrapper::vmla(out_1, in10_1, s10); + out_1 = wrapper::vmla(out_1, in11_1, s11); + + auto out_2 = wrapper::vmul(in00_2, s00); + out_2 = wrapper::vmla(out_2, in01_2, s01); + out_2 = wrapper::vmla(out_2, in10_2, s10); + out_2 = wrapper::vmla(out_2, in11_2, s11); + + auto out_3 = wrapper::vmul(in00_3, s00); + out_3 = wrapper::vmla(out_3, in01_3, s01); + out_3 = wrapper::vmla(out_3, in10_3, s10); + out_3 = wrapper::vmla(out_3, in11_3, s11); + +#if defined(__aarch64__) && !defined(BARE_METAL) + const auto out_0_int = wrapper::vcvta<int32_t>(out_0); + const auto out_1_int = wrapper::vcvta<int32_t>(out_1); + const auto out_2_int = wrapper::vcvta<int32_t>(out_2); + const auto out_3_int = wrapper::vcvta<int32_t>(out_3); +#else // defined(__aarch64__) && !defined(BARE_METAL) + const auto out_0_int = wrapper::vcvt<int32_t>(out_0); + const auto out_1_int = wrapper::vcvt<int32_t>(out_1); + const auto out_2_int = wrapper::vcvt<int32_t>(out_2); + const auto out_3_int = wrapper::vcvt<int32_t>(out_3); +#endif // defined(__aarch64__) && !defined(BARE_METAL) + const auto low_part = wrapper::vqmovn(wrapper::vcombine(wrapper::vqmovn(out_0_int), wrapper::vqmovn(out_1_int))); + const auto high_part = wrapper::vqmovn(wrapper::vcombine(wrapper::vqmovn(out_2_int), wrapper::vqmovn(out_3_int))); + const auto out = wrapper::vcombine(low_part, high_part); + + wrapper::vstore(out_ptr_xo_yo + cout * sizeof(int8_t), out); + } + + for(; cout < out_dim_ch; ++cout) + { + const int8_t in00 = *(in_ptr_xi0_yi0 + cout * sizeof(int8_t)); + const int8_t in01 = *(in_ptr_xi1_yi0 + cout * sizeof(int8_t)); + const int8_t in10 = *(in_ptr_xi0_yi1 + cout * sizeof(int8_t)); + const int8_t in11 = *(in_ptr_xi1_yi1 + cout * sizeof(int8_t)); + + float out0 = in00 * s00_s; + out0 += in01 * s01_s; + out0 += in10 * s10_s; + out0 += in11 * s11_s; + + // Rounding modes of vector and scalar loops should match +#if defined(__aarch64__) && !defined(BARE_METAL) + *(out_ptr_xo_yo + cout * sizeof(int8_t)) = static_cast<int8_t>(std::round(out0)); +#else // defined(__aarch64__) && !defined(BARE_METAL) + *(out_ptr_xo_yo + cout * sizeof(int8_t)) = static_cast<int8_t>(out0); +#endif // defined(__aarch64__) && !defined(BARE_METAL) + } + } + } + } } else { @@ -240,10 +606,10 @@ void s16_neon_scale_bilinear(const ITensor *src, ITensor *dst, const ITensor *of const auto dy_val = *reinterpret_cast<const float *>(dy->ptr_to_element(Coordinates(id.y(), id.z()))); const int in_hi = std::floor((id.z() + sampling_offset) * hr - sampling_offset); - auto clamped_w = utility::clamp<int>(offset, 0, in_dim_w - 1); - auto clamped_w1 = utility::clamp<int>(offset + 1, 0, in_dim_w - 1); - auto clamped_h = utility::clamp<int>(in_hi, 0, in_dim_h - 1); - auto clamped_h1 = utility::clamp<int>(in_hi + 1, 0, in_dim_h - 1); + const auto clamped_w = utility::clamp<int>(offset, 0, in_dim_w - 1); + const auto clamped_w1 = utility::clamp<int>(offset + 1, 0, in_dim_w - 1); + const auto clamped_h = utility::clamp<int>(in_hi, 0, in_dim_h - 1); + const auto clamped_h1 = utility::clamp<int>(in_hi + 1, 0, in_dim_h - 1); const auto a00 = *(reinterpret_cast<const int16_t *>(in.ptr()) + clamped_w * in_stride_c + clamped_h * in_stride_wc); const auto a01 = *(reinterpret_cast<const int16_t *>(in.ptr()) + clamped_w1 * in_stride_c + clamped_h * in_stride_wc); @@ -262,6 +628,20 @@ void s16_neon_scale_bilinear(const ITensor *src, ITensor *dst, const ITensor *of } namespace cpu { +void s8_neon_scale(const ITensor *src, ITensor *dst, const ITensor *offsets, const ITensor *dx, const ITensor *dy, + InterpolationPolicy policy, BorderMode border_mode, PixelValue constant_border_value, float sampling_offset, + bool align_corners, const Window &window) +{ + if(policy == InterpolationPolicy::BILINEAR) + { + s8_neon_scale_bilinear(src, dst, offsets, dx, dy, border_mode, constant_border_value, sampling_offset, align_corners, window); + } + else + { + ARM_COMPUTE_ERROR("Not implemented"); + } +} + void u8_neon_scale(const ITensor *src, ITensor *dst, const ITensor *offsets, const ITensor *dx, const ITensor *dy, InterpolationPolicy policy, BorderMode border_mode, PixelValue constant_border_value, float sampling_offset, bool align_corners, const Window &window) |