ref: 50a90744407be7b88c47737d1bba79e7ffb5c2d3
dir: /vpx_dsp/x86/vpx_subpixel_4t_intrin_sse2.c/
/*
* Copyright (c) 2018 The WebM project authors. All Rights Reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
#include <emmintrin.h>
#include "./vpx_dsp_rtcd.h"
#include "vpx/vpx_integer.h"
#include "vpx_dsp/x86/convolve.h"
#include "vpx_dsp/x86/convolve_sse2.h"
#include "vpx_ports/mem.h"
void vpx_filter_block1d16_h4_sse2(const uint8_t *src_ptr, ptrdiff_t src_stride,
uint8_t *dst_ptr, ptrdiff_t dst_stride,
uint32_t height, const int16_t *kernel) {
__m128i kernel_reg; // Kernel
__m128i kernel_reg_23, kernel_reg_45; // Segments of the kernel used
const __m128i reg_32 = _mm_set1_epi16(32); // Used for rounding
int h;
__m128i src_reg, src_reg_shift_1, src_reg_shift_2, src_reg_shift_3;
__m128i dst_first, dst_second;
__m128i even, odd;
// Start one pixel before as we need tap/2 - 1 = 1 sample from the past
src_ptr -= 1;
// Load Kernel
kernel_reg = _mm_loadu_si128((const __m128i *)kernel);
kernel_reg = _mm_srai_epi16(kernel_reg, 1);
kernel_reg_23 = extract_quarter_2_epi16_sse2(&kernel_reg);
kernel_reg_45 = extract_quarter_3_epi16_sse2(&kernel_reg);
for (h = height; h > 0; --h) {
// We will load multiple shifted versions of the row and shuffle them into
// 16-bit words of the form
// ... s[2] s[1] s[0] s[-1]
// ... s[4] s[3] s[2] s[1]
// Then we call multiply and add to get partial results
// s[2]k[3]+s[1]k[2] s[0]k[3]s[-1]k[2]
// s[4]k[5]+s[3]k[4] s[2]k[5]s[1]k[4]
// The two results are then added together for the first half of even
// output.
// Repeat multiple times to get the whole outoput
src_reg = _mm_loadu_si128((const __m128i *)src_ptr);
src_reg_shift_1 = _mm_srli_si128(src_reg, 1);
src_reg_shift_2 = _mm_srli_si128(src_reg, 2);
src_reg_shift_3 = _mm_srli_si128(src_reg, 3);
// Output 6 4 2 0
even = multiply_add_epi8_sse2(&src_reg, &src_reg_shift_2, &kernel_reg_23,
&kernel_reg_45);
// Output 7 5 3 1
odd = multiply_add_epi8_sse2(&src_reg_shift_1, &src_reg_shift_3,
&kernel_reg_23, &kernel_reg_45);
// Combine to get the first half of the dst
dst_first = combine_epi32_sse2(&even, &odd);
// Do again to get the second half of dst
src_reg = _mm_loadu_si128((const __m128i *)(src_ptr + 8));
src_reg_shift_1 = _mm_srli_si128(src_reg, 1);
src_reg_shift_2 = _mm_srli_si128(src_reg, 2);
src_reg_shift_3 = _mm_srli_si128(src_reg, 3);
// Output 14 12 10 8
even = multiply_add_epi8_sse2(&src_reg, &src_reg_shift_2, &kernel_reg_23,
&kernel_reg_45);
// Output 15 13 11 9
odd = multiply_add_epi8_sse2(&src_reg_shift_1, &src_reg_shift_3,
&kernel_reg_23, &kernel_reg_45);
// Combine to get the second half of the dst
dst_second = combine_epi32_sse2(&even, &odd);
// Round each result
dst_first = round_epi16_sse2(&dst_first, ®_32, 6);
dst_second = round_epi16_sse2(&dst_second, ®_32, 6);
// Finally combine to get the final dst
dst_first = _mm_packus_epi16(dst_first, dst_second);
_mm_store_si128((__m128i *)dst_ptr, dst_first);
src_ptr += src_stride;
dst_ptr += dst_stride;
}
}
/* The macro used to generate functions shifts the src_ptr up by 3 rows already
* */
void vpx_filter_block1d16_v4_sse2(const uint8_t *src_ptr, ptrdiff_t src_stride,
uint8_t *dst_ptr, ptrdiff_t dst_stride,
uint32_t height, const int16_t *kernel) {
// Register for source s[-1:3, :]
__m128i src_reg_m1, src_reg_0, src_reg_1, src_reg_2, src_reg_3;
// Interleaved rows of the source. lo is first half, hi second
__m128i src_reg_m10_lo, src_reg_m10_hi, src_reg_01_lo, src_reg_01_hi;
__m128i src_reg_12_lo, src_reg_12_hi, src_reg_23_lo, src_reg_23_hi;
// Half of half of the interleaved rows
__m128i src_reg_m10_lo_1, src_reg_m10_lo_2, src_reg_m10_hi_1,
src_reg_m10_hi_2;
__m128i src_reg_01_lo_1, src_reg_01_lo_2, src_reg_01_hi_1, src_reg_01_hi_2;
__m128i src_reg_12_lo_1, src_reg_12_lo_2, src_reg_12_hi_1, src_reg_12_hi_2;
__m128i src_reg_23_lo_1, src_reg_23_lo_2, src_reg_23_hi_1, src_reg_23_hi_2;
__m128i kernel_reg; // Kernel
__m128i kernel_reg_23, kernel_reg_45; // Segments of the kernel used
// Result after multiply and add
__m128i res_reg_m10_lo, res_reg_01_lo, res_reg_12_lo, res_reg_23_lo;
__m128i res_reg_m10_hi, res_reg_01_hi, res_reg_12_hi, res_reg_23_hi;
__m128i res_reg_m1012, res_reg_0123;
__m128i res_reg_m1012_lo, res_reg_0123_lo, res_reg_m1012_hi, res_reg_0123_hi;
const __m128i reg_32 = _mm_set1_epi16(32); // Used for rounding
// We will compute the result two rows at a time
const ptrdiff_t src_stride_unrolled = src_stride << 1;
const ptrdiff_t dst_stride_unrolled = dst_stride << 1;
int h;
// We only need to go num_taps/2 - 1 row above the souce, so we move
// 3 - (num_taps/2 - 1) = 4 - num_taps/2 = 2 back down
src_ptr += src_stride_unrolled;
// Load Kernel
kernel_reg = _mm_loadu_si128((const __m128i *)kernel);
kernel_reg = _mm_srai_epi16(kernel_reg, 1);
kernel_reg_23 = extract_quarter_2_epi16_sse2(&kernel_reg);
kernel_reg_45 = extract_quarter_3_epi16_sse2(&kernel_reg);
// We will load two rows of pixels as 8-bit words, rearrange them as 16-bit
// words,
// shuffle the data into the form
// ... s[0,1] s[-1,1] s[0,0] s[-1,0]
// ... s[0,7] s[-1,7] s[0,6] s[-1,6]
// ... s[0,9] s[-1,9] s[0,8] s[-1,8]
// ... s[0,13] s[-1,13] s[0,12] s[-1,12]
// so that we can call multiply and add with the kernel to get 32-bit words of
// the form
// ... s[0,1]k[3]+s[-1,1]k[2] s[0,0]k[3]+s[-1,0]k[2]
// Finally, we can add multiple rows together to get the desired output.
// First shuffle the data
src_reg_m1 = _mm_loadu_si128((const __m128i *)src_ptr);
src_reg_0 = _mm_loadu_si128((const __m128i *)(src_ptr + src_stride));
src_reg_m10_lo = _mm_unpacklo_epi8(src_reg_m1, src_reg_0);
src_reg_m10_hi = _mm_unpackhi_epi8(src_reg_m1, src_reg_0);
src_reg_m10_lo_1 = _mm_unpacklo_epi8(src_reg_m10_lo, _mm_setzero_si128());
src_reg_m10_lo_2 = _mm_unpackhi_epi8(src_reg_m10_lo, _mm_setzero_si128());
src_reg_m10_hi_1 = _mm_unpacklo_epi8(src_reg_m10_hi, _mm_setzero_si128());
src_reg_m10_hi_2 = _mm_unpackhi_epi8(src_reg_m10_hi, _mm_setzero_si128());
// More shuffling
src_reg_1 = _mm_loadu_si128((const __m128i *)(src_ptr + src_stride * 2));
src_reg_01_lo = _mm_unpacklo_epi8(src_reg_0, src_reg_1);
src_reg_01_hi = _mm_unpackhi_epi8(src_reg_0, src_reg_1);
src_reg_01_lo_1 = _mm_unpacklo_epi8(src_reg_01_lo, _mm_setzero_si128());
src_reg_01_lo_2 = _mm_unpackhi_epi8(src_reg_01_lo, _mm_setzero_si128());
src_reg_01_hi_1 = _mm_unpacklo_epi8(src_reg_01_hi, _mm_setzero_si128());
src_reg_01_hi_2 = _mm_unpackhi_epi8(src_reg_01_hi, _mm_setzero_si128());
for (h = height; h > 1; h -= 2) {
src_reg_2 = _mm_loadu_si128((const __m128i *)(src_ptr + src_stride * 3));
src_reg_12_lo = _mm_unpacklo_epi8(src_reg_1, src_reg_2);
src_reg_12_hi = _mm_unpackhi_epi8(src_reg_1, src_reg_2);
src_reg_3 = _mm_loadu_si128((const __m128i *)(src_ptr + src_stride * 4));
src_reg_23_lo = _mm_unpacklo_epi8(src_reg_2, src_reg_3);
src_reg_23_hi = _mm_unpackhi_epi8(src_reg_2, src_reg_3);
// Partial output from first half
res_reg_m10_lo = multiply_add_packs_epi16_sse2(
&src_reg_m10_lo_1, &src_reg_m10_lo_2, &kernel_reg_23);
res_reg_01_lo = multiply_add_packs_epi16_sse2(
&src_reg_01_lo_1, &src_reg_01_lo_2, &kernel_reg_23);
src_reg_12_lo_1 = _mm_unpacklo_epi8(src_reg_12_lo, _mm_setzero_si128());
src_reg_12_lo_2 = _mm_unpackhi_epi8(src_reg_12_lo, _mm_setzero_si128());
res_reg_12_lo = multiply_add_packs_epi16_sse2(
&src_reg_12_lo_1, &src_reg_12_lo_2, &kernel_reg_45);
src_reg_23_lo_1 = _mm_unpacklo_epi8(src_reg_23_lo, _mm_setzero_si128());
src_reg_23_lo_2 = _mm_unpackhi_epi8(src_reg_23_lo, _mm_setzero_si128());
res_reg_23_lo = multiply_add_packs_epi16_sse2(
&src_reg_23_lo_1, &src_reg_23_lo_2, &kernel_reg_45);
// Add to get first half of the results
res_reg_m1012_lo = _mm_adds_epi16(res_reg_m10_lo, res_reg_12_lo);
res_reg_0123_lo = _mm_adds_epi16(res_reg_01_lo, res_reg_23_lo);
// Now repeat everything again for the second half
// Partial output for second half
res_reg_m10_hi = multiply_add_packs_epi16_sse2(
&src_reg_m10_hi_1, &src_reg_m10_hi_2, &kernel_reg_23);
res_reg_01_hi = multiply_add_packs_epi16_sse2(
&src_reg_01_hi_1, &src_reg_01_hi_2, &kernel_reg_23);
src_reg_12_hi_1 = _mm_unpacklo_epi8(src_reg_12_hi, _mm_setzero_si128());
src_reg_12_hi_2 = _mm_unpackhi_epi8(src_reg_12_hi, _mm_setzero_si128());
res_reg_12_hi = multiply_add_packs_epi16_sse2(
&src_reg_12_hi_1, &src_reg_12_hi_2, &kernel_reg_45);
src_reg_23_hi_1 = _mm_unpacklo_epi8(src_reg_23_hi, _mm_setzero_si128());
src_reg_23_hi_2 = _mm_unpackhi_epi8(src_reg_23_hi, _mm_setzero_si128());
res_reg_23_hi = multiply_add_packs_epi16_sse2(
&src_reg_23_hi_1, &src_reg_23_hi_2, &kernel_reg_45);
// Second half of the results
res_reg_m1012_hi = _mm_adds_epi16(res_reg_m10_hi, res_reg_12_hi);
res_reg_0123_hi = _mm_adds_epi16(res_reg_01_hi, res_reg_23_hi);
// Round the words
res_reg_m1012_lo = round_epi16_sse2(&res_reg_m1012_lo, ®_32, 6);
res_reg_0123_lo = round_epi16_sse2(&res_reg_0123_lo, ®_32, 6);
res_reg_m1012_hi = round_epi16_sse2(&res_reg_m1012_hi, ®_32, 6);
res_reg_0123_hi = round_epi16_sse2(&res_reg_0123_hi, ®_32, 6);
// Combine to get the result
res_reg_m1012 = _mm_packus_epi16(res_reg_m1012_lo, res_reg_m1012_hi);
res_reg_0123 = _mm_packus_epi16(res_reg_0123_lo, res_reg_0123_hi);
_mm_store_si128((__m128i *)dst_ptr, res_reg_m1012);
_mm_store_si128((__m128i *)(dst_ptr + dst_stride), res_reg_0123);
// Update the source by two rows
src_ptr += src_stride_unrolled;
dst_ptr += dst_stride_unrolled;
src_reg_m10_lo_1 = src_reg_12_lo_1;
src_reg_m10_lo_2 = src_reg_12_lo_2;
src_reg_m10_hi_1 = src_reg_12_hi_1;
src_reg_m10_hi_2 = src_reg_12_hi_2;
src_reg_01_lo_1 = src_reg_23_lo_1;
src_reg_01_lo_2 = src_reg_23_lo_2;
src_reg_01_hi_1 = src_reg_23_hi_1;
src_reg_01_hi_2 = src_reg_23_hi_2;
src_reg_1 = src_reg_3;
}
}
void vpx_filter_block1d8_h4_sse2(const uint8_t *src_ptr, ptrdiff_t src_stride,
uint8_t *dst_ptr, ptrdiff_t dst_stride,
uint32_t height, const int16_t *kernel) {
__m128i kernel_reg; // Kernel
__m128i kernel_reg_23, kernel_reg_45; // Segments of the kernel used
const __m128i reg_32 = _mm_set1_epi16(32); // Used for rounding
int h;
__m128i src_reg, src_reg_shift_1, src_reg_shift_2, src_reg_shift_3;
__m128i dst_first;
__m128i even, odd;
// Start one pixel before as we need tap/2 - 1 = 1 sample from the past
src_ptr -= 1;
// Load Kernel
kernel_reg = _mm_loadu_si128((const __m128i *)kernel);
kernel_reg = _mm_srai_epi16(kernel_reg, 1);
kernel_reg_23 = extract_quarter_2_epi16_sse2(&kernel_reg);
kernel_reg_45 = extract_quarter_3_epi16_sse2(&kernel_reg);
for (h = height; h > 0; --h) {
// We will load multiple shifted versions of the row and shuffle them into
// 16-bit words of the form
// ... s[2] s[1] s[0] s[-1]
// ... s[4] s[3] s[2] s[1]
// Then we call multiply and add to get partial results
// s[2]k[3]+s[1]k[2] s[0]k[3]s[-1]k[2]
// s[4]k[5]+s[3]k[4] s[2]k[5]s[1]k[4]
// The two results are then added together to get the even output
src_reg = _mm_loadu_si128((const __m128i *)src_ptr);
src_reg_shift_1 = _mm_srli_si128(src_reg, 1);
src_reg_shift_2 = _mm_srli_si128(src_reg, 2);
src_reg_shift_3 = _mm_srli_si128(src_reg, 3);
// Output 6 4 2 0
even = multiply_add_epi8_sse2(&src_reg, &src_reg_shift_2, &kernel_reg_23,
&kernel_reg_45);
// Output 7 5 3 1
odd = multiply_add_epi8_sse2(&src_reg_shift_1, &src_reg_shift_3,
&kernel_reg_23, &kernel_reg_45);
// Combine to get the first half of the dst
dst_first = combine_epi32_sse2(&even, &odd);
dst_first = round_epi16_sse2(&dst_first, ®_32, 6);
// Saturate and convert to 8-bit words
dst_first = _mm_packus_epi16(dst_first, _mm_setzero_si128());
_mm_storel_epi64((__m128i *)dst_ptr, dst_first);
src_ptr += src_stride;
dst_ptr += dst_stride;
}
}
void vpx_filter_block1d8_v4_sse2(const uint8_t *src_ptr, ptrdiff_t src_stride,
uint8_t *dst_ptr, ptrdiff_t dst_stride,
uint32_t height, const int16_t *kernel) {
// Register for source s[-1:3, :]
__m128i src_reg_m1, src_reg_0, src_reg_1, src_reg_2, src_reg_3;
// Interleaved rows of the source. lo is first half, hi second
__m128i src_reg_m10_lo, src_reg_01_lo;
__m128i src_reg_12_lo, src_reg_23_lo;
// Half of half of the interleaved rows
__m128i src_reg_m10_lo_1, src_reg_m10_lo_2;
__m128i src_reg_01_lo_1, src_reg_01_lo_2;
__m128i src_reg_12_lo_1, src_reg_12_lo_2;
__m128i src_reg_23_lo_1, src_reg_23_lo_2;
__m128i kernel_reg; // Kernel
__m128i kernel_reg_23, kernel_reg_45; // Segments of the kernel used
// Result after multiply and add
__m128i res_reg_m10_lo, res_reg_01_lo, res_reg_12_lo, res_reg_23_lo;
__m128i res_reg_m1012, res_reg_0123;
__m128i res_reg_m1012_lo, res_reg_0123_lo;
const __m128i reg_32 = _mm_set1_epi16(32); // Used for rounding
// We will compute the result two rows at a time
const ptrdiff_t src_stride_unrolled = src_stride << 1;
const ptrdiff_t dst_stride_unrolled = dst_stride << 1;
int h;
// We only need to go num_taps/2 - 1 row above the souce, so we move
// 3 - (num_taps/2 - 1) = 4 - num_taps/2 = 2 back down
src_ptr += src_stride_unrolled;
// Load Kernel
kernel_reg = _mm_loadu_si128((const __m128i *)kernel);
kernel_reg = _mm_srai_epi16(kernel_reg, 1);
kernel_reg_23 = extract_quarter_2_epi16_sse2(&kernel_reg);
kernel_reg_45 = extract_quarter_3_epi16_sse2(&kernel_reg);
// We will load two rows of pixels as 8-bit words, rearrange them as 16-bit
// words,
// shuffle the data into the form
// ... s[0,1] s[-1,1] s[0,0] s[-1,0]
// ... s[0,7] s[-1,7] s[0,6] s[-1,6]
// ... s[0,9] s[-1,9] s[0,8] s[-1,8]
// ... s[0,13] s[-1,13] s[0,12] s[-1,12]
// so that we can call multiply and add with the kernel to get 32-bit words of
// the form
// ... s[0,1]k[3]+s[-1,1]k[2] s[0,0]k[3]+s[-1,0]k[2]
// Finally, we can add multiple rows together to get the desired output.
// First shuffle the data
src_reg_m1 = _mm_loadu_si128((const __m128i *)src_ptr);
src_reg_0 = _mm_loadu_si128((const __m128i *)(src_ptr + src_stride));
src_reg_m10_lo = _mm_unpacklo_epi8(src_reg_m1, src_reg_0);
src_reg_m10_lo_1 = _mm_unpacklo_epi8(src_reg_m10_lo, _mm_setzero_si128());
src_reg_m10_lo_2 = _mm_unpackhi_epi8(src_reg_m10_lo, _mm_setzero_si128());
// More shuffling
src_reg_1 = _mm_loadu_si128((const __m128i *)(src_ptr + src_stride * 2));
src_reg_01_lo = _mm_unpacklo_epi8(src_reg_0, src_reg_1);
src_reg_01_lo_1 = _mm_unpacklo_epi8(src_reg_01_lo, _mm_setzero_si128());
src_reg_01_lo_2 = _mm_unpackhi_epi8(src_reg_01_lo, _mm_setzero_si128());
for (h = height; h > 1; h -= 2) {
src_reg_2 = _mm_loadu_si128((const __m128i *)(src_ptr + src_stride * 3));
src_reg_12_lo = _mm_unpacklo_epi8(src_reg_1, src_reg_2);
src_reg_3 = _mm_loadu_si128((const __m128i *)(src_ptr + src_stride * 4));
src_reg_23_lo = _mm_unpacklo_epi8(src_reg_2, src_reg_3);
// Partial output
res_reg_m10_lo = multiply_add_packs_epi16_sse2(
&src_reg_m10_lo_1, &src_reg_m10_lo_2, &kernel_reg_23);
res_reg_01_lo = multiply_add_packs_epi16_sse2(
&src_reg_01_lo_1, &src_reg_01_lo_2, &kernel_reg_23);
src_reg_12_lo_1 = _mm_unpacklo_epi8(src_reg_12_lo, _mm_setzero_si128());
src_reg_12_lo_2 = _mm_unpackhi_epi8(src_reg_12_lo, _mm_setzero_si128());
res_reg_12_lo = multiply_add_packs_epi16_sse2(
&src_reg_12_lo_1, &src_reg_12_lo_2, &kernel_reg_45);
src_reg_23_lo_1 = _mm_unpacklo_epi8(src_reg_23_lo, _mm_setzero_si128());
src_reg_23_lo_2 = _mm_unpackhi_epi8(src_reg_23_lo, _mm_setzero_si128());
res_reg_23_lo = multiply_add_packs_epi16_sse2(
&src_reg_23_lo_1, &src_reg_23_lo_2, &kernel_reg_45);
// Add to get results
res_reg_m1012_lo = _mm_adds_epi16(res_reg_m10_lo, res_reg_12_lo);
res_reg_0123_lo = _mm_adds_epi16(res_reg_01_lo, res_reg_23_lo);
// Round the words
res_reg_m1012_lo = round_epi16_sse2(&res_reg_m1012_lo, ®_32, 6);
res_reg_0123_lo = round_epi16_sse2(&res_reg_0123_lo, ®_32, 6);
// Convert to 8-bit words
res_reg_m1012 = _mm_packus_epi16(res_reg_m1012_lo, _mm_setzero_si128());
res_reg_0123 = _mm_packus_epi16(res_reg_0123_lo, _mm_setzero_si128());
// Save only half of the register (8 words)
_mm_storel_epi64((__m128i *)dst_ptr, res_reg_m1012);
_mm_storel_epi64((__m128i *)(dst_ptr + dst_stride), res_reg_0123);
// Update the source by two rows
src_ptr += src_stride_unrolled;
dst_ptr += dst_stride_unrolled;
src_reg_m10_lo_1 = src_reg_12_lo_1;
src_reg_m10_lo_2 = src_reg_12_lo_2;
src_reg_01_lo_1 = src_reg_23_lo_1;
src_reg_01_lo_2 = src_reg_23_lo_2;
src_reg_1 = src_reg_3;
}
}
void vpx_filter_block1d4_h4_sse2(const uint8_t *src_ptr, ptrdiff_t src_stride,
uint8_t *dst_ptr, ptrdiff_t dst_stride,
uint32_t height, const int16_t *kernel) {
__m128i kernel_reg; // Kernel
__m128i kernel_reg_23, kernel_reg_45; // Segments of the kernel used
const __m128i reg_32 = _mm_set1_epi16(32); // Used for rounding
int h;
__m128i src_reg, src_reg_shift_1, src_reg_shift_2, src_reg_shift_3;
__m128i dst_first;
__m128i tmp_0, tmp_1;
// Start one pixel before as we need tap/2 - 1 = 1 sample from the past
src_ptr -= 1;
// Load Kernel
kernel_reg = _mm_loadu_si128((const __m128i *)kernel);
kernel_reg = _mm_srai_epi16(kernel_reg, 1);
kernel_reg_23 = extract_quarter_2_epi16_sse2(&kernel_reg);
kernel_reg_45 = extract_quarter_3_epi16_sse2(&kernel_reg);
for (h = height; h > 0; --h) {
// We will load multiple shifted versions of the row and shuffle them into
// 16-bit words of the form
// ... s[1] s[0] s[0] s[-1]
// ... s[3] s[2] s[2] s[1]
// Then we call multiply and add to get partial results
// s[1]k[3]+s[0]k[2] s[0]k[3]s[-1]k[2]
// s[3]k[5]+s[2]k[4] s[2]k[5]s[1]k[4]
// The two results are then added together to get the output
src_reg = _mm_loadu_si128((const __m128i *)src_ptr);
src_reg_shift_1 = _mm_srli_si128(src_reg, 1);
src_reg_shift_2 = _mm_srli_si128(src_reg, 2);
src_reg_shift_3 = _mm_srli_si128(src_reg, 3);
// Convert to 16-bit words
src_reg = _mm_unpacklo_epi8(src_reg, _mm_setzero_si128());
src_reg_shift_1 = _mm_unpacklo_epi8(src_reg_shift_1, _mm_setzero_si128());
src_reg_shift_2 = _mm_unpacklo_epi8(src_reg_shift_2, _mm_setzero_si128());
src_reg_shift_3 = _mm_unpacklo_epi8(src_reg_shift_3, _mm_setzero_si128());
// Shuffle into the right format
tmp_0 = _mm_unpacklo_epi32(src_reg, src_reg_shift_1);
tmp_1 = _mm_unpacklo_epi32(src_reg_shift_2, src_reg_shift_3);
// Partial output
tmp_0 = _mm_madd_epi16(tmp_0, kernel_reg_23);
tmp_1 = _mm_madd_epi16(tmp_1, kernel_reg_45);
// Output
dst_first = _mm_add_epi32(tmp_0, tmp_1);
dst_first = _mm_packs_epi32(dst_first, _mm_setzero_si128());
dst_first = round_epi16_sse2(&dst_first, ®_32, 6);
// Saturate and convert to 8-bit words
dst_first = _mm_packus_epi16(dst_first, _mm_setzero_si128());
*((uint32_t *)(dst_ptr)) = _mm_cvtsi128_si32(dst_first);
src_ptr += src_stride;
dst_ptr += dst_stride;
}
}
void vpx_filter_block1d4_v4_sse2(const uint8_t *src_ptr, ptrdiff_t src_stride,
uint8_t *dst_ptr, ptrdiff_t dst_stride,
uint32_t height, const int16_t *kernel) {
// Register for source s[-1:3, :]
__m128i src_reg_m1, src_reg_0, src_reg_1, src_reg_2, src_reg_3;
// Interleaved rows of the source. lo is first half, hi second
__m128i src_reg_m10_lo, src_reg_01_lo;
__m128i src_reg_12_lo, src_reg_23_lo;
// Half of half of the interleaved rows
__m128i src_reg_m10_lo_1;
__m128i src_reg_01_lo_1;
__m128i src_reg_12_lo_1;
__m128i src_reg_23_lo_1;
__m128i kernel_reg; // Kernel
__m128i kernel_reg_23, kernel_reg_45; // Segments of the kernel used
// Result after multiply and add
__m128i res_reg_m10_lo, res_reg_01_lo, res_reg_12_lo, res_reg_23_lo;
__m128i res_reg_m1012, res_reg_0123;
__m128i res_reg_m1012_lo, res_reg_0123_lo;
const __m128i reg_32 = _mm_set1_epi16(32); // Used for rounding
const __m128i reg_zero = _mm_setzero_si128();
// We will compute the result two rows at a time
const ptrdiff_t src_stride_unrolled = src_stride << 1;
const ptrdiff_t dst_stride_unrolled = dst_stride << 1;
int h;
// We only need to go num_taps/2 - 1 row above the souce, so we move
// 3 - (num_taps/2 - 1) = 4 - num_taps/2 = 2 back down
src_ptr += src_stride_unrolled;
// Load Kernel
kernel_reg = _mm_loadu_si128((const __m128i *)kernel);
kernel_reg = _mm_srai_epi16(kernel_reg, 1);
kernel_reg_23 = extract_quarter_2_epi16_sse2(&kernel_reg);
kernel_reg_45 = extract_quarter_3_epi16_sse2(&kernel_reg);
// We will load two rows of pixels as 8-bit words, rearrange them as 16-bit
// words,
// shuffle the data into the form
// ... s[0,1] s[-1,1] s[0,0] s[-1,0]
// ... s[0,7] s[-1,7] s[0,6] s[-1,6]
// ... s[0,9] s[-1,9] s[0,8] s[-1,8]
// ... s[0,13] s[-1,13] s[0,12] s[-1,12]
// so that we can call multiply and add with the kernel to get 32-bit words of
// the form
// ... s[0,1]k[3]+s[-1,1]k[2] s[0,0]k[3]+s[-1,0]k[2]
// Finally, we can add multiple rows together to get the desired output.
// First shuffle the data
src_reg_m1 = _mm_loadu_si128((const __m128i *)src_ptr);
src_reg_0 = _mm_loadu_si128((const __m128i *)(src_ptr + src_stride));
src_reg_m10_lo = _mm_unpacklo_epi8(src_reg_m1, src_reg_0);
src_reg_m10_lo_1 = _mm_unpacklo_epi8(src_reg_m10_lo, _mm_setzero_si128());
// More shuffling
src_reg_1 = _mm_loadu_si128((const __m128i *)(src_ptr + src_stride * 2));
src_reg_01_lo = _mm_unpacklo_epi8(src_reg_0, src_reg_1);
src_reg_01_lo_1 = _mm_unpacklo_epi8(src_reg_01_lo, _mm_setzero_si128());
for (h = height; h > 1; h -= 2) {
src_reg_2 = _mm_loadu_si128((const __m128i *)(src_ptr + src_stride * 3));
src_reg_12_lo = _mm_unpacklo_epi8(src_reg_1, src_reg_2);
src_reg_3 = _mm_loadu_si128((const __m128i *)(src_ptr + src_stride * 4));
src_reg_23_lo = _mm_unpacklo_epi8(src_reg_2, src_reg_3);
// Partial output
res_reg_m10_lo = multiply_add_packs_epi16_sse2(&src_reg_m10_lo_1, ®_zero,
&kernel_reg_23);
res_reg_01_lo = multiply_add_packs_epi16_sse2(&src_reg_01_lo_1, ®_zero,
&kernel_reg_23);
src_reg_12_lo_1 = _mm_unpacklo_epi8(src_reg_12_lo, _mm_setzero_si128());
res_reg_12_lo = multiply_add_packs_epi16_sse2(&src_reg_12_lo_1, ®_zero,
&kernel_reg_45);
src_reg_23_lo_1 = _mm_unpacklo_epi8(src_reg_23_lo, _mm_setzero_si128());
res_reg_23_lo = multiply_add_packs_epi16_sse2(&src_reg_23_lo_1, ®_zero,
&kernel_reg_45);
// Add to get results
res_reg_m1012_lo = _mm_adds_epi16(res_reg_m10_lo, res_reg_12_lo);
res_reg_0123_lo = _mm_adds_epi16(res_reg_01_lo, res_reg_23_lo);
// Round the words
res_reg_m1012_lo = round_epi16_sse2(&res_reg_m1012_lo, ®_32, 6);
res_reg_0123_lo = round_epi16_sse2(&res_reg_0123_lo, ®_32, 6);
// Convert to 8-bit words
res_reg_m1012 = _mm_packus_epi16(res_reg_m1012_lo, reg_zero);
res_reg_0123 = _mm_packus_epi16(res_reg_0123_lo, reg_zero);
// Save only half of the register (8 words)
*((uint32_t *)(dst_ptr)) = _mm_cvtsi128_si32(res_reg_m1012);
*((uint32_t *)(dst_ptr + dst_stride)) = _mm_cvtsi128_si32(res_reg_0123);
// Update the source by two rows
src_ptr += src_stride_unrolled;
dst_ptr += dst_stride_unrolled;
src_reg_m10_lo_1 = src_reg_12_lo_1;
src_reg_01_lo_1 = src_reg_23_lo_1;
src_reg_1 = src_reg_3;
}
}