libblur 0.24.0

// Copyright (c) Radzivon Bartoshyk. All rights reserved.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// 1.  Redistributions of source code must retain the above copyright notice, this
// list of conditions and the following disclaimer.
//
// 2.  Redistributions in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// 3.  Neither the name of the copyright holder nor the names of its
// contributors may be used to endorse or promote products derived from
// this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
// DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
// FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
// DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
// SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
// CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
// OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

use crate::EdgeMode;
use crate::edge_mode::clamp_edge;
use crate::sse::{_mm_opt_fnmlaf_ps, load_f32, store_f32};
use crate::unsafe_slice::UnsafeSlice;
use crate::util::ScratchBuffer;
#[cfg(target_arch = "x86")]
use std::arch::x86::*;
#[cfg(target_arch = "x86_64")]
use std::arch::x86_64::*;

#[repr(C, align(16))]
#[derive(Copy, Clone, Default)]
pub(crate) struct SseF32x4(pub(crate) [f32; 4]);

pub(crate) fn fg_horizontal_pass_sse_f32<const CN: usize>(
    slice: &UnsafeSlice<f32>,
    stride: u32,
    width: u32,
    height: u32,
    radius: u32,
    start: u32,
    end: u32,
    edge_mode: EdgeMode,
) {
    unsafe {
        if std::arch::is_x86_feature_detected!("fma") {
            fg_horizontal_pass_sse_f32_fma::<CN>(
                slice, stride, width, height, radius, start, end, edge_mode,
            );
        } else {
            fg_horizontal_pass_sse_f32_def::<CN>(
                slice, stride, width, height, radius, start, end, edge_mode,
            );
        }
    }
}

#[target_feature(enable = "sse4.1")]
fn fg_horizontal_pass_sse_f32_def<const CN: usize>(
    bytes: &UnsafeSlice<f32>,
    stride: u32,
    width: u32,
    height: u32,
    radius: u32,
    start: u32,
    end: u32,
    edge_mode: EdgeMode,
) {
    fg_horizontal_pass_sse_f32_impl::<CN, false>(
        bytes, stride, width, height, radius, start, end, edge_mode,
    );
}

#[target_feature(enable = "sse4.1", enable = "fma")]
fn fg_horizontal_pass_sse_f32_fma<const CN: usize>(
    bytes: &UnsafeSlice<f32>,
    stride: u32,
    width: u32,
    height: u32,
    radius: u32,
    start: u32,
    end: u32,
    edge_mode: EdgeMode,
) {
    fg_horizontal_pass_sse_f32_impl::<CN, true>(
        bytes, stride, width, height, radius, start, end, edge_mode,
    );
}

#[inline(always)]
fn fg_horizontal_pass_sse_f32_impl<const CN: usize, const FMA: bool>(
    bytes: &UnsafeSlice<f32>,
    stride: u32,
    width: u32,
    height: u32,
    radius: u32,
    start: u32,
    end: u32,
    edge_mode: EdgeMode,
) {
    unsafe {
        let mut buffer = ScratchBuffer::<[SseF32x4; 4], 1024>::new(1024);
        let buffer = buffer.as_mut_slice();

        let radius_64 = radius as i64;
        let width_wide = width as i64;

        let v_double = _mm_set1_ps(2.);
        let v_weight = _mm_set1_ps(1f32 / (radius as f32 * radius as f32));

        let mut yy = start as usize;

        while yy + 4 <= height.min(end) as usize {
            let mut diffs0 = _mm_setzero_ps();
            let mut diffs1 = _mm_setzero_ps();
            let mut diffs2 = _mm_setzero_ps();
            let mut diffs3 = _mm_setzero_ps();

            let mut sums0 = _mm_setzero_ps();
            let mut sums1 = _mm_setzero_ps();
            let mut sums2 = _mm_setzero_ps();
            let mut sums3 = _mm_setzero_ps();

            let current_y0 = ((yy as i64) * (stride as i64)) as usize;
            let current_y1 = ((yy as i64 + 1) * (stride as i64)) as usize;
            let current_y2 = ((yy as i64 + 2) * (stride as i64)) as usize;
            let current_y3 = ((yy as i64 + 3) * (stride as i64)) as usize;

            let start_x = 0 - 2 * radius_64;
            for x in start_x..(width as i64) {
                if x >= 0 {
                    let current_px = x as usize * CN;

                    let prepared_px0 = _mm_mul_ps(sums0, v_weight);
                    let prepared_px1 = _mm_mul_ps(sums1, v_weight);
                    let prepared_px2 = _mm_mul_ps(sums2, v_weight);
                    let prepared_px3 = _mm_mul_ps(sums3, v_weight);

                    store_f32::<CN>(bytes.get_ptr(current_y0 + current_px), prepared_px0);
                    store_f32::<CN>(bytes.get_ptr(current_y1 + current_px), prepared_px1);
                    store_f32::<CN>(bytes.get_ptr(current_y2 + current_px), prepared_px2);
                    store_f32::<CN>(bytes.get_ptr(current_y3 + current_px), prepared_px3);

                    let arr_index = ((x - radius_64) & 1023) as usize;
                    let d_arr_index = (x & 1023) as usize;
                    let da_b = buffer.get_unchecked(d_arr_index);
                    let da = buffer.get_unchecked(arr_index);

                    let d_s0 = _mm_load_ps(da_b.as_ptr().cast());
                    let d_s1 = _mm_load_ps(da_b[1..].as_ptr().cast());
                    let d_s2 = _mm_load_ps(da_b[2..].as_ptr().cast());
                    let d_s3 = _mm_load_ps(da_b[3..].as_ptr().cast());

                    let a_s0 = _mm_load_ps(da.as_ptr().cast());
                    let a_s1 = _mm_load_ps(da[1..].as_ptr().cast());
                    let a_s2 = _mm_load_ps(da[2..].as_ptr().cast());
                    let a_s3 = _mm_load_ps(da[3..].as_ptr().cast());

                    diffs0 = _mm_add_ps(diffs0, _mm_opt_fnmlaf_ps::<FMA>(a_s0, d_s0, v_double));
                    diffs1 = _mm_add_ps(diffs1, _mm_opt_fnmlaf_ps::<FMA>(a_s1, d_s1, v_double));
                    diffs2 = _mm_add_ps(diffs2, _mm_opt_fnmlaf_ps::<FMA>(a_s2, d_s2, v_double));
                    diffs3 = _mm_add_ps(diffs3, _mm_opt_fnmlaf_ps::<FMA>(a_s3, d_s3, v_double));
                } else if x + radius_64 >= 0 {
                    let arr_index = (x & 1023) as usize;
                    let da = buffer.get_unchecked(arr_index);
                    let s0 = _mm_loadu_ps(da.as_ptr().cast());
                    let s1 = _mm_loadu_ps(da[1..].as_ptr().cast());
                    let s2 = _mm_loadu_ps(da[2..].as_ptr().cast());
                    let s3 = _mm_loadu_ps(da[3..].as_ptr().cast());

                    diffs0 = _mm_opt_fnmlaf_ps::<FMA>(diffs0, s0, v_double);
                    diffs1 = _mm_opt_fnmlaf_ps::<FMA>(diffs1, s1, v_double);
                    diffs2 = _mm_opt_fnmlaf_ps::<FMA>(diffs2, s2, v_double);
                    diffs3 = _mm_opt_fnmlaf_ps::<FMA>(diffs3, s3, v_double);
                }

                let next_row_x = clamp_edge!(edge_mode, x + radius_64, 0, width_wide);
                let next_row_px = next_row_x * CN;

                let px0 = load_f32::<CN>(bytes.get_ptr(current_y0 + next_row_px));
                let px1 = load_f32::<CN>(bytes.get_ptr(current_y1 + next_row_px));
                let px2 = load_f32::<CN>(bytes.get_ptr(current_y2 + next_row_px));
                let px3 = load_f32::<CN>(bytes.get_ptr(current_y3 + next_row_px));

                let arr_index = ((x + radius_64) & 1023) as usize;
                let da = buffer.get_unchecked_mut(arr_index);

                _mm_store_ps(da.as_mut_ptr().cast(), px0);
                _mm_store_ps(da[1..].as_mut_ptr().cast(), px1);
                _mm_store_ps(da[2..].as_mut_ptr().cast(), px2);
                _mm_store_ps(da[3..].as_mut_ptr().cast(), px3);

                diffs0 = _mm_add_ps(diffs0, px0);
                diffs1 = _mm_add_ps(diffs1, px1);
                diffs2 = _mm_add_ps(diffs2, px2);
                diffs3 = _mm_add_ps(diffs3, px3);

                sums0 = _mm_add_ps(sums0, diffs0);
                sums1 = _mm_add_ps(sums1, diffs1);
                sums2 = _mm_add_ps(sums2, diffs2);
                sums3 = _mm_add_ps(sums3, diffs3);
            }

            yy += 4;
        }

        for y in yy..height.min(end) as usize {
            let mut diffs = _mm_setzero_ps();
            let mut summs = _mm_setzero_ps();

            let current_y = ((y as i64) * (stride as i64)) as usize;

            let start_x = 0 - 2 * radius_64;
            for x in start_x..(width as i64) {
                if x >= 0 {
                    let current_px = x as usize * CN;

                    let pixel = _mm_mul_ps(summs, v_weight);

                    let bytes_offset = current_y + current_px;

                    let dst_ptr = bytes.get_ptr(bytes_offset);
                    store_f32::<CN>(dst_ptr, pixel);

                    let arr_index = ((x - radius_64) & 1023) as usize;
                    let d_arr_index = (x & 1023) as usize;

                    let d_buf_ptr = buffer.get_unchecked(d_arr_index);
                    let d_stored = _mm_load_ps(d_buf_ptr.as_ptr().cast());

                    let buf_ptr = buffer.get_unchecked(arr_index);
                    let a_stored = _mm_load_ps(buf_ptr.as_ptr().cast());

                    diffs = _mm_add_ps(
                        diffs,
                        _mm_opt_fnmlaf_ps::<FMA>(a_stored, d_stored, v_double),
                    );
                } else if x + radius_64 >= 0 {
                    let arr_index = (x & 1023) as usize;
                    let buf_ptr = buffer.get_unchecked(arr_index);
                    let stored = _mm_load_ps(buf_ptr.as_ptr().cast());
                    diffs = _mm_opt_fnmlaf_ps::<FMA>(diffs, stored, v_double);
                }

                let next_row_y = y * (stride as usize);
                let next_row_x = clamp_edge!(edge_mode, x + radius_64, 0, width_wide);
                let next_row_px = next_row_x * CN;

                let s_ptr = bytes.get_ptr(next_row_y + next_row_px);
                let pixel_color = load_f32::<CN>(s_ptr);

                let arr_index = ((x + radius_64) & 1023) as usize;
                let buf_ptr = buffer.get_unchecked_mut(arr_index);

                diffs = _mm_add_ps(diffs, pixel_color);
                summs = _mm_add_ps(summs, diffs);
                _mm_store_ps(buf_ptr.as_mut_ptr().cast(), pixel_color);
            }
        }
    }
}

pub(crate) fn fg_vertical_pass_sse_f32<const CN: usize>(
    slice: &UnsafeSlice<f32>,
    stride: u32,
    width: u32,
    height: u32,
    radius: u32,
    start: u32,
    end: u32,
    edge_mode: EdgeMode,
) {
    unsafe {
        if std::arch::is_x86_feature_detected!("fma") {
            fg_vertical_pass_sse_f32_fma::<CN>(
                slice, stride, width, height, radius, start, end, edge_mode,
            );
        } else {
            fg_vertical_pass_sse_f32_def::<CN>(
                slice, stride, width, height, radius, start, end, edge_mode,
            );
        }
    }
}

#[target_feature(enable = "sse4.1")]
fn fg_vertical_pass_sse_f32_def<const CN: usize>(
    bytes: &UnsafeSlice<f32>,
    stride: u32,
    width: u32,
    height: u32,
    radius: u32,
    start: u32,
    end: u32,
    edge_mode: EdgeMode,
) {
    fg_vertical_pass_sse_f32_impl::<CN, false>(
        bytes, stride, width, height, radius, start, end, edge_mode,
    );
}

#[target_feature(enable = "sse4.1", enable = "fma")]
fn fg_vertical_pass_sse_f32_fma<const CN: usize>(
    bytes: &UnsafeSlice<f32>,
    stride: u32,
    width: u32,
    height: u32,
    radius: u32,
    start: u32,
    end: u32,
    edge_mode: EdgeMode,
) {
    fg_vertical_pass_sse_f32_impl::<CN, true>(
        bytes, stride, width, height, radius, start, end, edge_mode,
    );
}

#[inline(always)]
fn fg_vertical_pass_sse_f32_impl<const CN: usize, const FMA: bool>(
    bytes: &UnsafeSlice<f32>,
    stride: u32,
    width: u32,
    height: u32,
    radius: u32,
    start: u32,
    end: u32,
    edge_mode: EdgeMode,
) {
    unsafe {
        let mut buffer = ScratchBuffer::<[SseF32x4; 4], 1024>::new(1024);
        let buffer = buffer.as_mut_slice();

        let v_double = _mm_set1_ps(2.);
        let v_weight = _mm_set1_ps(1f32 / (radius as f32 * radius as f32));

        let height_wide = height as i64;

        let radius_64 = radius as i64;

        let mut xx = start as usize;

        while xx + 4 <= width.min(end) as usize {
            let mut diffs0 = _mm_setzero_ps();
            let mut diffs1 = _mm_setzero_ps();
            let mut diffs2 = _mm_setzero_ps();
            let mut diffs3 = _mm_setzero_ps();

            let mut sums0 = _mm_setzero_ps();
            let mut sums1 = _mm_setzero_ps();
            let mut sums2 = _mm_setzero_ps();
            let mut sums3 = _mm_setzero_ps();

            let current_px0 = xx * CN;
            let current_px1 = (xx + 1) * CN;
            let current_px2 = (xx + 2) * CN;
            let current_px3 = (xx + 3) * CN;

            let start_y = 0 - 2 * radius as i64;
            for y in start_y..height_wide {
                if y >= 0 {
                    let current_y = (y * (stride as i64)) as usize;

                    let prepared_px0 = _mm_mul_ps(sums0, v_weight);
                    let prepared_px1 = _mm_mul_ps(sums1, v_weight);
                    let prepared_px2 = _mm_mul_ps(sums2, v_weight);
                    let prepared_px3 = _mm_mul_ps(sums3, v_weight);

                    store_f32::<CN>(bytes.get_ptr(current_y + current_px0), prepared_px0);
                    store_f32::<CN>(bytes.get_ptr(current_y + current_px1), prepared_px1);
                    store_f32::<CN>(bytes.get_ptr(current_y + current_px2), prepared_px2);
                    store_f32::<CN>(bytes.get_ptr(current_y + current_px3), prepared_px3);

                    let arr_index = ((y - radius_64) & 1023) as usize;
                    let d_arr_index = (y & 1023) as usize;
                    let da_b = buffer.get_unchecked(d_arr_index);
                    let da = buffer.get_unchecked(arr_index);

                    let d_s0 = _mm_load_ps(da_b.as_ptr().cast());
                    let d_s1 = _mm_load_ps(da_b[1..].as_ptr().cast());
                    let d_s2 = _mm_load_ps(da_b[2..].as_ptr().cast());
                    let d_s3 = _mm_load_ps(da_b[3..].as_ptr().cast());

                    let a_s0 = _mm_load_ps(da.as_ptr().cast());
                    let a_s1 = _mm_load_ps(da[1..].as_ptr().cast());
                    let a_s2 = _mm_load_ps(da[2..].as_ptr().cast());
                    let a_s3 = _mm_load_ps(da[3..].as_ptr().cast());

                    diffs0 = _mm_add_ps(diffs0, _mm_opt_fnmlaf_ps::<FMA>(a_s0, d_s0, v_double));
                    diffs1 = _mm_add_ps(diffs1, _mm_opt_fnmlaf_ps::<FMA>(a_s1, d_s1, v_double));
                    diffs2 = _mm_add_ps(diffs2, _mm_opt_fnmlaf_ps::<FMA>(a_s2, d_s2, v_double));
                    diffs3 = _mm_add_ps(diffs3, _mm_opt_fnmlaf_ps::<FMA>(a_s3, d_s3, v_double));
                } else if y + radius_64 >= 0 {
                    let arr_index = (y & 1023) as usize;
                    let da = buffer.get_unchecked(arr_index);
                    let s0 = _mm_loadu_ps(da.as_ptr().cast());
                    let s1 = _mm_loadu_ps(da[1..].as_ptr().cast());
                    let s2 = _mm_loadu_ps(da[2..].as_ptr().cast());
                    let s3 = _mm_loadu_ps(da[3..].as_ptr().cast());

                    diffs0 = _mm_opt_fnmlaf_ps::<FMA>(diffs0, s0, v_double);
                    diffs1 = _mm_opt_fnmlaf_ps::<FMA>(diffs1, s1, v_double);
                    diffs2 = _mm_opt_fnmlaf_ps::<FMA>(diffs2, s2, v_double);
                    diffs3 = _mm_opt_fnmlaf_ps::<FMA>(diffs3, s3, v_double);
                }

                let next_row_y =
                    clamp_edge!(edge_mode, y + radius_64, 0, height_wide) * (stride as usize);

                let px0 = load_f32::<CN>(bytes.get_ptr(next_row_y + current_px0));
                let px1 = load_f32::<CN>(bytes.get_ptr(next_row_y + current_px1));
                let px2 = load_f32::<CN>(bytes.get_ptr(next_row_y + current_px2));
                let px3 = load_f32::<CN>(bytes.get_ptr(next_row_y + current_px3));

                let arr_index = ((y + radius_64) & 1023) as usize;
                let da = buffer.get_unchecked_mut(arr_index);

                _mm_store_ps(da.as_mut_ptr().cast(), px0);
                _mm_store_ps(da[1..].as_mut_ptr().cast(), px1);
                _mm_store_ps(da[2..].as_mut_ptr().cast(), px2);
                _mm_store_ps(da[3..].as_mut_ptr().cast(), px3);

                diffs0 = _mm_add_ps(diffs0, px0);
                diffs1 = _mm_add_ps(diffs1, px1);
                diffs2 = _mm_add_ps(diffs2, px2);
                diffs3 = _mm_add_ps(diffs3, px3);

                sums0 = _mm_add_ps(sums0, diffs0);
                sums1 = _mm_add_ps(sums1, diffs1);
                sums2 = _mm_add_ps(sums2, diffs2);
                sums3 = _mm_add_ps(sums3, diffs3);
            }

            xx += 4;
        }

        for x in xx..width.min(end) as usize {
            let mut diffs = _mm_setzero_ps();
            let mut summs = _mm_setzero_ps();

            let start_y = 0 - 2 * radius as i64;

            let current_px = x * CN;

            for y in start_y..height_wide {
                let current_y = (y * (stride as i64)) as usize;

                if y >= 0 {
                    let pixel = _mm_mul_ps(summs, v_weight);

                    let bytes_offset = current_y + current_px;

                    let dst_ptr = bytes.get_ptr(bytes_offset);
                    store_f32::<CN>(dst_ptr, pixel);

                    let arr_index = ((y - radius_64) & 1023) as usize;
                    let d_arr_index = (y & 1023) as usize;

                    let d_buf_ptr = buffer.get_unchecked(d_arr_index);
                    let d_stored = _mm_load_ps(d_buf_ptr.as_ptr().cast());

                    let buf_ptr = buffer.get_unchecked(arr_index);
                    let a_stored = _mm_load_ps(buf_ptr.as_ptr().cast());

                    diffs = _mm_add_ps(
                        diffs,
                        _mm_opt_fnmlaf_ps::<FMA>(a_stored, d_stored, v_double),
                    );
                } else if y + radius_64 >= 0 {
                    let arr_index = (y & 1023) as usize;
                    let buf_ptr = buffer.get_unchecked(arr_index);
                    let stored = _mm_loadu_ps(buf_ptr.as_ptr().cast());
                    diffs = _mm_opt_fnmlaf_ps::<FMA>(diffs, stored, v_double);
                }

                let next_row_y =
                    clamp_edge!(edge_mode, y + radius_64, 0, height_wide) * (stride as usize);
                let next_row_x = x * CN;

                let s_ptr = bytes.get_ptr(next_row_y + next_row_x);
                let pixel_color = load_f32::<CN>(s_ptr);

                let arr_index = ((y + radius_64) & 1023) as usize;
                let buf_ptr = buffer.get_unchecked_mut(arr_index);

                diffs = _mm_add_ps(diffs, pixel_color);
                summs = _mm_add_ps(summs, diffs);
                _mm_storeu_ps(buf_ptr.as_mut_ptr().cast(), pixel_color);
            }
        }
    }
}