hashirama/uvg266
1
0
Fork 0
mirror of https://github.com/ultravideo/uvg266.git synced 2024-11-24 18:34:06 +00:00
Code Issues Actions Projects Releases Wiki Activity
uvg266/src/strategies/sse41/reg_sad_pow2_widths-sse41.h

886 lines
37 KiB
C
Raw Normal View History

Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
#ifndef REG_SAD_POW2_WIDTHS_SSE41_H_
#define REG_SAD_POW2_WIDTHS_SSE41_H_
#include <immintrin.h>
#include "kvazaar.h"
Create strategy for directly obtaining pointer to constant-width SAD function
2019-01-18 15:32:52 +00:00
static INLINE uint32_t reg_sad_w0(const kvz_pixel * const data1, const kvz_pixel * const data2,
const int32_t height, const uint32_t stride1,
const uint32_t stride2)
{
return 0;
}
Inline fixed width vectorized SAD functions
2019-01-16 17:57:53 +00:00
static INLINE uint32_t reg_sad_w4(const kvz_pixel * const data1, const kvz_pixel * const data2,
Add SSE4.1 strategies for SAD on widths 4 and 12 and AVX2 strategies for 32 and 64
2019-01-15 19:09:18 +00:00
const int32_t height, const uint32_t stride1,
const uint32_t stride2)
{
__m128i sse_inc = _mm_setzero_si128();
int32_t y;
Unify unrolling naming conventions
2019-01-16 19:48:18 +00:00
const int32_t height_fourline_groups = height & ~3;
const int32_t height_residual_lines = height & 3;
Add SSE4.1 strategies for SAD on widths 4 and 12 and AVX2 strategies for 32 and 64
2019-01-15 19:09:18 +00:00
Unify unrolling naming conventions
2019-01-16 19:48:18 +00:00
for (y = 0; y < height_fourline_groups; y += 4) {
Add SSE4.1 strategies for SAD on widths 4 and 12 and AVX2 strategies for 32 and 64
2019-01-15 19:09:18 +00:00
__m128i a = _mm_cvtsi32_si128(*(uint32_t *)(data1 + y * stride1));
__m128i b = _mm_cvtsi32_si128(*(uint32_t *)(data2 + y * stride2));
Cast all pointers to const types in vector SAD funcs Also tidy up the pointer arithmetic
2019-01-16 18:39:51 +00:00
a = _mm_insert_epi32(a, *(const uint32_t *)(data1 + (y + 1) * stride1), 1);
b = _mm_insert_epi32(b, *(const uint32_t *)(data2 + (y + 1) * stride2), 1);
a = _mm_insert_epi32(a, *(const uint32_t *)(data1 + (y + 2) * stride1), 2);
b = _mm_insert_epi32(b, *(const uint32_t *)(data2 + (y + 2) * stride2), 2);
a = _mm_insert_epi32(a, *(const uint32_t *)(data1 + (y + 3) * stride1), 3);
b = _mm_insert_epi32(b, *(const uint32_t *)(data2 + (y + 3) * stride2), 3);
Add SSE4.1 strategies for SAD on widths 4 and 12 and AVX2 strategies for 32 and 64
2019-01-15 19:09:18 +00:00
__m128i curr_sads = _mm_sad_epu8(a, b);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
Unify unrolling naming conventions
2019-01-16 19:48:18 +00:00
if (height_residual_lines) {
Add SSE4.1 strategies for SAD on widths 4 and 12 and AVX2 strategies for 32 and 64
2019-01-15 19:09:18 +00:00
for (; y < height; y++) {
Cast all pointers to const types in vector SAD funcs Also tidy up the pointer arithmetic
2019-01-16 18:39:51 +00:00
__m128i a = _mm_cvtsi32_si128(*(const uint32_t *)(data1 + y * stride1));
__m128i b = _mm_cvtsi32_si128(*(const uint32_t *)(data2 + y * stride2));
Add SSE4.1 strategies for SAD on widths 4 and 12 and AVX2 strategies for 32 and 64
2019-01-15 19:09:18 +00:00
__m128i curr_sads = _mm_sad_epu8(a, b);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
Inline fixed width vectorized SAD functions
2019-01-16 17:57:53 +00:00
static INLINE uint32_t reg_sad_w8(const kvz_pixel * const data1, const kvz_pixel * const data2,
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
const int32_t height, const uint32_t stride1,
const uint32_t stride2)
{
__m128i sse_inc = _mm_setzero_si128();
uint64_t result = 0;
int32_t y;
Unroll SSE SAD for 8-wide blocks to process 4 lines at once
2019-01-16 19:41:42 +00:00
const int32_t height_fourline_groups = height & ~3;
const int32_t height_residual_lines = height & 3;
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
Unroll SSE SAD for 8-wide blocks to process 4 lines at once
2019-01-16 19:41:42 +00:00
for (y = 0; y < height_fourline_groups; y += 4) {
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
__m128d a_d = _mm_setzero_pd();
__m128d b_d = _mm_setzero_pd();
Unroll SSE SAD for 8-wide blocks to process 4 lines at once
2019-01-16 19:41:42 +00:00
__m128d c_d = _mm_setzero_pd();
__m128d d_d = _mm_setzero_pd();
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
a_d = _mm_loadl_pd(a_d, (const double *)(data1 + (y + 0) * stride1));
b_d = _mm_loadl_pd(b_d, (const double *)(data2 + (y + 0) * stride2));
a_d = _mm_loadh_pd(a_d, (const double *)(data1 + (y + 1) * stride1));
b_d = _mm_loadh_pd(b_d, (const double *)(data2 + (y + 1) * stride2));
Unroll SSE SAD for 8-wide blocks to process 4 lines at once
2019-01-16 19:41:42 +00:00
c_d = _mm_loadl_pd(c_d, (const double *)(data1 + (y + 2) * stride1));
d_d = _mm_loadl_pd(d_d, (const double *)(data2 + (y + 2) * stride2));
c_d = _mm_loadh_pd(c_d, (const double *)(data1 + (y + 3) * stride1));
d_d = _mm_loadh_pd(d_d, (const double *)(data2 + (y + 3) * stride2));
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
__m128i a = _mm_castpd_si128(a_d);
__m128i b = _mm_castpd_si128(b_d);
Unroll SSE SAD for 8-wide blocks to process 4 lines at once
2019-01-16 19:41:42 +00:00
__m128i c = _mm_castpd_si128(c_d);
__m128i d = _mm_castpd_si128(d_d);
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
Unroll SSE SAD for 8-wide blocks to process 4 lines at once
2019-01-16 19:41:42 +00:00
__m128i curr_sads_ab = _mm_sad_epu8(a, b);
__m128i curr_sads_cd = _mm_sad_epu8(c, d);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_cd);
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
}
Unroll SSE SAD for 8-wide blocks to process 4 lines at once
2019-01-16 19:41:42 +00:00
if (height_residual_lines) {
for (; y < height; y++) {
__m64 a = *(__m64 *)(data1 + y * stride1);
__m64 b = *(__m64 *)(data2 + y * stride2);
__m64 sads = _mm_sad_pu8(a, b);
result += (uint64_t)sads;
}
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
result += _mm_cvtsi128_si32(sad);
return result;
}
Inline fixed width vectorized SAD functions
2019-01-16 17:57:53 +00:00
static INLINE uint32_t reg_sad_w12(const kvz_pixel * const data1, const kvz_pixel * const data2,
Add SSE4.1 strategies for SAD on widths 4 and 12 and AVX2 strategies for 32 and 64
2019-01-15 19:09:18 +00:00
const int32_t height, const uint32_t stride1,
const uint32_t stride2)
{
__m128i sse_inc = _mm_setzero_si128();
int32_t y;
for (y = 0; y < height; y++) {
Cast all pointers to const types in vector SAD funcs Also tidy up the pointer arithmetic
2019-01-16 18:39:51 +00:00
__m128i a = _mm_loadu_si128((const __m128i *)(data1 + y * stride1));
__m128i b = _mm_loadu_si128((const __m128i *)(data2 + y * stride2));
Add SSE4.1 strategies for SAD on widths 4 and 12 and AVX2 strategies for 32 and 64
2019-01-15 19:09:18 +00:00
__m128i b_masked = _mm_blend_epi16(a, b, 0x3f);
__m128i curr_sads = _mm_sad_epu8 (a, b_masked);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
Inline fixed width vectorized SAD functions
2019-01-16 17:57:53 +00:00
static INLINE uint32_t reg_sad_w16(const kvz_pixel * const data1, const kvz_pixel * const data2,
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
const int32_t height, const uint32_t stride1,
const uint32_t stride2)
{
__m128i sse_inc = _mm_setzero_si128();
int32_t y;
Unroll 32 and 16 wide SAD vector implementations by 4
2019-01-16 18:54:34 +00:00
const int32_t height_fourline_groups = height & ~3;
const int32_t height_residual_lines = height & 3;
for (y = 0; y < height_fourline_groups; y += 4) {
__m128i a = _mm_loadu_si128((const __m128i *)(data1 + (y + 0) * stride1));
__m128i b = _mm_loadu_si128((const __m128i *)(data2 + (y + 0) * stride2));
__m128i c = _mm_loadu_si128((const __m128i *)(data1 + (y + 1) * stride1));
__m128i d = _mm_loadu_si128((const __m128i *)(data2 + (y + 1) * stride2));
__m128i e = _mm_loadu_si128((const __m128i *)(data1 + (y + 2) * stride1));
__m128i f = _mm_loadu_si128((const __m128i *)(data2 + (y + 2) * stride2));
__m128i g = _mm_loadu_si128((const __m128i *)(data1 + (y + 3) * stride1));
__m128i h = _mm_loadu_si128((const __m128i *)(data2 + (y + 3) * stride2));
__m128i curr_sads_ab = _mm_sad_epu8(a, b);
__m128i curr_sads_cd = _mm_sad_epu8(c, d);
__m128i curr_sads_ef = _mm_sad_epu8(e, f);
__m128i curr_sads_gh = _mm_sad_epu8(g, h);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_cd);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ef);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_gh);
}
if (height_residual_lines) {
for (; y < height; y++) {
__m128i a = _mm_loadu_si128((const __m128i *)(data1 + (y + 0) * stride1));
__m128i b = _mm_loadu_si128((const __m128i *)(data2 + (y + 0) * stride2));
__m128i curr_sads = _mm_sad_epu8(a, b);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
}
Unroll 32 and 16 wide SAD vector implementations by 4
2019-01-16 18:54:34 +00:00
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
Inline fixed width vectorized SAD functions
2019-01-16 17:57:53 +00:00
static INLINE uint32_t reg_sad_w24(const kvz_pixel * const data1, const kvz_pixel * const data2,
Add SSE4.1 strategy for 24px wide SAD and an AVX2 strategy for 16
2019-01-15 20:08:50 +00:00
const int32_t height, const uint32_t stride1,
const uint32_t stride2)
{
__m128i sse_inc = _mm_setzero_si128();
int32_t y;
const int32_t height_doublelines = height & ~1;
const int32_t height_parity = height & 1;
for (y = 0; y < height_doublelines; y += 2) {
__m128i a = _mm_loadu_si128((const __m128i *)(data1 + (y + 0) * stride1));
__m128i b = _mm_loadu_si128((const __m128i *)(data2 + (y + 0) * stride2));
__m128i c = _mm_loadu_si128((const __m128i *)(data1 + (y + 1) * stride1));
__m128i d = _mm_loadu_si128((const __m128i *)(data2 + (y + 1) * stride2));
__m128d e_d = _mm_setzero_pd();
__m128d f_d = _mm_setzero_pd();
e_d = _mm_loadl_pd(e_d, (const double *)(data1 + (y + 0) * stride1 + 16));
f_d = _mm_loadl_pd(f_d, (const double *)(data2 + (y + 0) * stride2 + 16));
e_d = _mm_loadh_pd(e_d, (const double *)(data1 + (y + 1) * stride1 + 16));
f_d = _mm_loadh_pd(f_d, (const double *)(data2 + (y + 1) * stride2 + 16));
__m128i e = _mm_castpd_si128(e_d);
__m128i f = _mm_castpd_si128(f_d);
__m128i curr_sads_1 = _mm_sad_epu8(a, b);
__m128i curr_sads_2 = _mm_sad_epu8(c, d);
__m128i curr_sads_3 = _mm_sad_epu8(e, f);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_1);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_2);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_3);
}
if (height_parity) {
Cast all pointers to const types in vector SAD funcs Also tidy up the pointer arithmetic
2019-01-16 18:39:51 +00:00
__m128i a = _mm_loadu_si128 ((const __m128i *) (data1 + y * stride1));
__m128i b = _mm_loadu_si128 ((const __m128i *) (data2 + y * stride2));
Fix bug in 24-wide SAD *facepalm*
2019-01-24 13:25:45 +00:00
__m128i c = _mm_cvtsi64_si128(*(const uint64_t *)(data1 + y * stride1 + 16));
__m128i d = _mm_cvtsi64_si128(*(const uint64_t *)(data2 + y * stride2 + 16));
Add SSE4.1 strategy for 24px wide SAD and an AVX2 strategy for 16
2019-01-15 20:08:50 +00:00
__m128i curr_sads_1 = _mm_sad_epu8(a, b);
__m128i curr_sads_2 = _mm_sad_epu8(c, d);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_1);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_2);
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
Inline fixed width vectorized SAD functions
2019-01-16 17:57:53 +00:00
static INLINE uint32_t reg_sad_arbitrary(const kvz_pixel * const data1, const kvz_pixel * const data2,
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
const int32_t width, const int32_t height, const uint32_t stride1,
const uint32_t stride2)
{
int32_t y, x;
__m128i sse_inc = _mm_setzero_si128();
// Bytes in block in 128-bit blocks per each scanline, and remainder
Unify unrolling naming conventions
2019-01-16 19:48:18 +00:00
const int32_t width_xmms = width & ~15;
const int32_t width_residual_pixels = width & 15;
Unroll arbitrary width SSE4.1 SAD by 4
2019-01-16 19:35:33 +00:00
const int32_t height_fourline_groups = height & ~3;
const int32_t height_residual_lines = height & 3;
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
Unify unrolling naming conventions
2019-01-16 19:48:18 +00:00
const __m128i rds = _mm_set1_epi8 (width_residual_pixels);
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
const __m128i ns = _mm_setr_epi8 (0, 1, 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15);
const __m128i rdmask = _mm_cmpgt_epi8(rds, ns);
Unify unrolling naming conventions
2019-01-16 19:48:18 +00:00
for (x = 0; x < width_xmms; x += 16) {
Unroll arbitrary width SSE4.1 SAD by 4
2019-01-16 19:35:33 +00:00
for (y = 0; y < height_fourline_groups; y += 4) {
__m128i a = _mm_loadu_si128((const __m128i *)(data1 + (y + 0) * stride1 + x));
__m128i b = _mm_loadu_si128((const __m128i *)(data2 + (y + 0) * stride2 + x));
__m128i c = _mm_loadu_si128((const __m128i *)(data1 + (y + 1) * stride1 + x));
__m128i d = _mm_loadu_si128((const __m128i *)(data2 + (y + 1) * stride2 + x));
__m128i e = _mm_loadu_si128((const __m128i *)(data1 + (y + 2) * stride1 + x));
__m128i f = _mm_loadu_si128((const __m128i *)(data2 + (y + 2) * stride2 + x));
__m128i g = _mm_loadu_si128((const __m128i *)(data1 + (y + 3) * stride1 + x));
__m128i h = _mm_loadu_si128((const __m128i *)(data2 + (y + 3) * stride2 + x));
__m128i curr_sads_ab = _mm_sad_epu8(a, b);
__m128i curr_sads_cd = _mm_sad_epu8(c, d);
__m128i curr_sads_ef = _mm_sad_epu8(e, f);
__m128i curr_sads_gh = _mm_sad_epu8(g, h);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_cd);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ef);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_gh);
}
if (height_residual_lines) {
for (; y < height; y++) {
__m128i a = _mm_loadu_si128((const __m128i *)(data1 + y * stride1 + x));
__m128i b = _mm_loadu_si128((const __m128i *)(data2 + y * stride2 + x));
__m128i curr_sads = _mm_sad_epu8(a, b);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
}
}
Unify unrolling naming conventions
2019-01-16 19:48:18 +00:00
if (width_residual_pixels) {
Unroll arbitrary width SSE4.1 SAD by 4
2019-01-16 19:35:33 +00:00
for (y = 0; y < height_fourline_groups; y += 4) {
__m128i a = _mm_loadu_si128((const __m128i *)(data1 + (y + 0) * stride1 + x));
__m128i b = _mm_loadu_si128((const __m128i *)(data2 + (y + 0) * stride2 + x));
__m128i c = _mm_loadu_si128((const __m128i *)(data1 + (y + 1) * stride1 + x));
__m128i d = _mm_loadu_si128((const __m128i *)(data2 + (y + 1) * stride2 + x));
__m128i e = _mm_loadu_si128((const __m128i *)(data1 + (y + 2) * stride1 + x));
__m128i f = _mm_loadu_si128((const __m128i *)(data2 + (y + 2) * stride2 + x));
__m128i g = _mm_loadu_si128((const __m128i *)(data1 + (y + 3) * stride1 + x));
__m128i h = _mm_loadu_si128((const __m128i *)(data2 + (y + 3) * stride2 + x));
__m128i b_masked = _mm_blendv_epi8(a, b, rdmask);
__m128i d_masked = _mm_blendv_epi8(c, d, rdmask);
__m128i f_masked = _mm_blendv_epi8(e, f, rdmask);
__m128i h_masked = _mm_blendv_epi8(g, h, rdmask);
__m128i curr_sads_ab = _mm_sad_epu8 (a, b_masked);
__m128i curr_sads_cd = _mm_sad_epu8 (c, d_masked);
__m128i curr_sads_ef = _mm_sad_epu8 (e, f_masked);
__m128i curr_sads_gh = _mm_sad_epu8 (g, h_masked);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_cd);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ef);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_gh);
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
}
Unroll arbitrary width SSE4.1 SAD by 4
2019-01-16 19:35:33 +00:00
if (height_residual_lines) {
for (; y < height; y++) {
__m128i a = _mm_loadu_si128((const __m128i *)(data1 + y * stride1 + x));
__m128i b = _mm_loadu_si128((const __m128i *)(data2 + y * stride2 + x));
__m128i b_masked = _mm_blendv_epi8(a, b, rdmask);
__m128i curr_sads = _mm_sad_epu8 (a, b_masked);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
}
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
Create strategy for ver_sad Easy to vectorize
2019-01-22 13:57:16 +00:00
static uint32_t ver_sad_w4(const kvz_pixel *pic_data, const kvz_pixel *ref_data,
int32_t height, uint32_t stride)
{
__m128i ref_row = _mm_set1_epi32(*(const uint32_t *)ref_data);
__m128i sse_inc = _mm_setzero_si128();
int32_t y;
const int32_t height_fourline_groups = height & ~3;
const int32_t height_residual_lines = height & 3;
for (y = 0; y < height_fourline_groups; y += 4) {
__m128i a = _mm_cvtsi32_si128(*(uint32_t *)(pic_data + y * stride));
a = _mm_insert_epi32(a, *(const uint32_t *)(pic_data + (y + 1) * stride), 1);
a = _mm_insert_epi32(a, *(const uint32_t *)(pic_data + (y + 2) * stride), 2);
a = _mm_insert_epi32(a, *(const uint32_t *)(pic_data + (y + 3) * stride), 3);
__m128i curr_sads = _mm_sad_epu8(a, ref_row);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
if (height_residual_lines) {
// Only pick the last dword, because we're comparing single dwords (lines)
ref_row = _mm_bsrli_si128(ref_row, 12);
for (; y < height; y++) {
__m128i a = _mm_cvtsi32_si128(*(const uint32_t *)(pic_data + y * stride));
__m128i curr_sads = _mm_sad_epu8(a, ref_row);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
static uint32_t ver_sad_w8(const kvz_pixel *pic_data, const kvz_pixel *ref_data,
int32_t height, uint32_t stride)
{
const __m128i ref_row = _mm_set1_epi64x(*(const uint64_t *)ref_data);
__m128i sse_inc = _mm_setzero_si128();
uint64_t result = 0;
int32_t y;
const int32_t height_fourline_groups = height & ~3;
const int32_t height_residual_lines = height & 3;
for (y = 0; y < height_fourline_groups; y += 4) {
__m128d a_d = _mm_setzero_pd();
__m128d c_d = _mm_setzero_pd();
a_d = _mm_loadl_pd(a_d, (const double *)(pic_data + (y + 0) * stride));
a_d = _mm_loadh_pd(a_d, (const double *)(pic_data + (y + 1) * stride));
c_d = _mm_loadl_pd(c_d, (const double *)(pic_data + (y + 2) * stride));
c_d = _mm_loadh_pd(c_d, (const double *)(pic_data + (y + 3) * stride));
__m128i a = _mm_castpd_si128(a_d);
__m128i c = _mm_castpd_si128(c_d);
__m128i curr_sads_ab = _mm_sad_epu8(a, ref_row);
__m128i curr_sads_cd = _mm_sad_epu8(c, ref_row);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_cd);
}
if (height_residual_lines) {
__m64 b = (__m64)_mm_cvtsi128_si64(ref_row);
for (; y < height; y++) {
__m64 a = *(__m64 *)(pic_data + y * stride);
__m64 sads = _mm_sad_pu8(a, b);
result += (uint64_t)sads;
}
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
result += _mm_cvtsi128_si32(sad);
return result;
}
static uint32_t ver_sad_w12(const kvz_pixel *pic_data, const kvz_pixel *ref_data,
int32_t height, uint32_t stride)
{
const __m128i ref_row = _mm_loadu_si128((__m128i *)ref_data);
__m128i sse_inc = _mm_setzero_si128();
int32_t y;
for (y = 0; y < height; y++) {
__m128i a = _mm_loadu_si128((const __m128i *)(pic_data + y * stride));
__m128i a_masked = _mm_blend_epi16(ref_row, a, 0x3f);
__m128i curr_sads = _mm_sad_epu8 (ref_row, a_masked);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
static uint32_t ver_sad_w16(const kvz_pixel *pic_data, const kvz_pixel *ref_data,
int32_t height, uint32_t stride)
{
const __m128i ref_row = _mm_loadu_si128((__m128i *)ref_data);
__m128i sse_inc = _mm_setzero_si128();
int32_t y;
const int32_t height_fourline_groups = height & ~3;
const int32_t height_residual_lines = height & 3;
for (y = 0; y < height_fourline_groups; y += 4) {
__m128i pic_row_1 = _mm_loadu_si128((__m128i *)(pic_data + (y + 0) * stride));
__m128i pic_row_2 = _mm_loadu_si128((__m128i *)(pic_data + (y + 1) * stride));
__m128i pic_row_3 = _mm_loadu_si128((__m128i *)(pic_data + (y + 2) * stride));
__m128i pic_row_4 = _mm_loadu_si128((__m128i *)(pic_data + (y + 3) * stride));
__m128i curr_sads_1 = _mm_sad_epu8 (pic_row_1, ref_row);
__m128i curr_sads_2 = _mm_sad_epu8 (pic_row_2, ref_row);
__m128i curr_sads_3 = _mm_sad_epu8 (pic_row_3, ref_row);
__m128i curr_sads_4 = _mm_sad_epu8 (pic_row_4, ref_row);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_1);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_2);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_3);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_4);
}
if (height_residual_lines) {
for (; y < height; y++) {
__m128i pic_row = _mm_loadu_si128((__m128i *)(pic_data + (y + 0) * stride));
__m128i curr_sads = _mm_sad_epu8 (pic_row, ref_row);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
static uint32_t ver_sad_arbitrary(const kvz_pixel *pic_data, const kvz_pixel *ref_data,
int32_t width, int32_t height, uint32_t stride)
{
int32_t y, x;
__m128i sse_inc = _mm_setzero_si128();
// Bytes in block in 128-bit blocks per each scanline, and remainder
const int32_t width_xmms = width & ~15;
const int32_t width_residual_pixels = width & 15;
const int32_t height_fourline_groups = height & ~3;
const int32_t height_residual_lines = height & 3;
const __m128i rds = _mm_set1_epi8 (width_residual_pixels);
const __m128i ns = _mm_setr_epi8 (0, 1, 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15);
const __m128i rdmask = _mm_cmpgt_epi8(rds, ns);
for (x = 0; x < width_xmms; x += 16) {
const __m128i ref_row = _mm_loadu_si128((__m128i *)(ref_data + x));
for (y = 0; y < height_fourline_groups; y += 4) {
__m128i a = _mm_loadu_si128((const __m128i *)(pic_data + (y + 0) * stride + x));
__m128i c = _mm_loadu_si128((const __m128i *)(pic_data + (y + 1) * stride + x));
__m128i e = _mm_loadu_si128((const __m128i *)(pic_data + (y + 2) * stride + x));
__m128i g = _mm_loadu_si128((const __m128i *)(pic_data + (y + 3) * stride + x));
__m128i curr_sads_ab = _mm_sad_epu8(ref_row, a);
__m128i curr_sads_cd = _mm_sad_epu8(ref_row, c);
__m128i curr_sads_ef = _mm_sad_epu8(ref_row, e);
__m128i curr_sads_gh = _mm_sad_epu8(ref_row, g);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_cd);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ef);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_gh);
}
if (height_residual_lines) {
for (; y < height; y++) {
__m128i a = _mm_loadu_si128((const __m128i *)(pic_data + y * stride + x));
__m128i curr_sads = _mm_sad_epu8(a, ref_row);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
}
}
if (width_residual_pixels) {
const __m128i ref_row = _mm_loadu_si128((__m128i *)(ref_data + x));
for (y = 0; y < height_fourline_groups; y += 4) {
__m128i a = _mm_loadu_si128((const __m128i *)(pic_data + (y + 0) * stride + x));
__m128i c = _mm_loadu_si128((const __m128i *)(pic_data + (y + 1) * stride + x));
__m128i e = _mm_loadu_si128((const __m128i *)(pic_data + (y + 2) * stride + x));
__m128i g = _mm_loadu_si128((const __m128i *)(pic_data + (y + 3) * stride + x));
__m128i a_masked = _mm_blendv_epi8(ref_row, a, rdmask);
__m128i c_masked = _mm_blendv_epi8(ref_row, c, rdmask);
__m128i e_masked = _mm_blendv_epi8(ref_row, e, rdmask);
__m128i g_masked = _mm_blendv_epi8(ref_row, g, rdmask);
__m128i curr_sads_ab = _mm_sad_epu8 (ref_row, a_masked);
__m128i curr_sads_cd = _mm_sad_epu8 (ref_row, c_masked);
__m128i curr_sads_ef = _mm_sad_epu8 (ref_row, e_masked);
__m128i curr_sads_gh = _mm_sad_epu8 (ref_row, g_masked);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_cd);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ef);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_gh);
}
if (height_residual_lines) {
for (; y < height; y++) {
__m128i a = _mm_loadu_si128((const __m128i *)(pic_data + y * stride + x));
__m128i a_masked = _mm_blendv_epi8(ref_row, a, rdmask);
__m128i curr_sads = _mm_sad_epu8 (ref_row, a_masked);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
}
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
Start work on left and right border aware hor_sad Comes with 4, 8, 16 and 32 pixel wide implementations now, at some point investigate if this can start to thrash icache
2019-01-22 19:27:21 +00:00
static uint32_t hor_sad_left_sse41_w4(const kvz_pixel *pic_data, const kvz_pixel *ref_data,
int32_t width, int32_t height, uint32_t pic_stride,
uint32_t ref_stride, uint32_t overhang)
{
// Dualword (ie. line) base indexes, ie. the edges the lines read will be
// clamped towards
const __m128i dwbaseids = _mm_setr_epi8(0, 0, 0, 0, 4, 4, 4, 4,
8, 8, 8, 8, 12, 12, 12, 12);
const __m128i ns = _mm_setr_epi8(0, 1, 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15);
const __m128i excess = _mm_set1_epi8(overhang);
const __m128i mask1 = _mm_sub_epi8 (ns, excess);
const __m128i epol_mask = _mm_max_epi8 (mask1, dwbaseids);
const int32_t height_fourline_groups = height & ~3;
const int32_t height_residual_lines = height & 3;
__m128i sse_inc = _mm_setzero_si128();
int32_t y;
for (y = 0; y < height_fourline_groups; y += 4) {
__m128i a = _mm_cvtsi32_si128(*(const uint32_t *)(pic_data + y * pic_stride));
__m128i b = _mm_cvtsi32_si128(*(const uint32_t *)(ref_data + y * ref_stride + overhang));
a = _mm_insert_epi32(a, *(const uint32_t *)(pic_data + (y + 1) * pic_stride), 1);
b = _mm_insert_epi32(b, *(const uint32_t *)(ref_data + (y + 1) * ref_stride + overhang), 1);
a = _mm_insert_epi32(a, *(const uint32_t *)(pic_data + (y + 2) * pic_stride), 2);
b = _mm_insert_epi32(b, *(const uint32_t *)(ref_data + (y + 2) * ref_stride + overhang), 2);
a = _mm_insert_epi32(a, *(const uint32_t *)(pic_data + (y + 3) * pic_stride), 3);
b = _mm_insert_epi32(b, *(const uint32_t *)(ref_data + (y + 3) * ref_stride + overhang), 3);
__m128i b_epol = _mm_shuffle_epi8(b, epol_mask);
__m128i curr_sads = _mm_sad_epu8 (a, b_epol);
sse_inc = _mm_add_epi64 (sse_inc, curr_sads);
}
if (height_residual_lines) {
for (; y < height; y++) {
__m128i a = _mm_cvtsi32_si128(*(const uint32_t *)(pic_data + y * pic_stride));
__m128i b = _mm_cvtsi32_si128(*(const uint32_t *)(ref_data + y * ref_stride + overhang));
__m128i b_epol = _mm_shuffle_epi8(b, epol_mask);
__m128i curr_sads = _mm_sad_epu8 (a, b_epol);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
static uint32_t hor_sad_left_sse41_w8(const kvz_pixel *pic_data, const kvz_pixel *ref_data,
int32_t width, int32_t height, uint32_t pic_stride,
uint32_t ref_stride, uint32_t overhang)
{
// Quadword (ie. line) base indexes, ie. the edges the lines read will be
// clamped towards; higher qword (lower line) bytes tend towards 8 and lower
// qword (higher line) bytes towards 0
const __m128i qwbaseids = _mm_setr_epi8(0, 0, 0, 0, 0, 0, 0, 0,
8, 8, 8, 8, 8, 8, 8, 8);
const __m128i ns = _mm_setr_epi8(0, 1, 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15);
const __m128i excess = _mm_set1_epi8(overhang);
const __m128i mask1 = _mm_sub_epi8 (ns, excess);
const __m128i epol_mask = _mm_max_epi8 (mask1, qwbaseids);
const __m64 epol_mask_64 = (__m64)_mm_cvtsi128_si64(epol_mask);
const int32_t height_fourline_groups = height & ~3;
const int32_t height_residual_lines = height & 3;
uint64_t result = 0;
__m128i sse_inc = _mm_setzero_si128();
int32_t y;
for (y = 0; y < height_fourline_groups; y += 4) {
__m128d a_d = _mm_setzero_pd();
__m128d b_d = _mm_setzero_pd();
__m128d c_d = _mm_setzero_pd();
__m128d d_d = _mm_setzero_pd();
a_d = _mm_loadl_pd(a_d, (const double *)(pic_data + (y + 0) * pic_stride));
b_d = _mm_loadl_pd(b_d, (const double *)(ref_data + (y + 0) * ref_stride + overhang));
a_d = _mm_loadh_pd(a_d, (const double *)(pic_data + (y + 1) * pic_stride));
b_d = _mm_loadh_pd(b_d, (const double *)(ref_data + (y + 1) * ref_stride + overhang));
c_d = _mm_loadl_pd(c_d, (const double *)(pic_data + (y + 2) * pic_stride));
d_d = _mm_loadl_pd(d_d, (const double *)(ref_data + (y + 2) * ref_stride + overhang));
c_d = _mm_loadh_pd(c_d, (const double *)(pic_data + (y + 3) * pic_stride));
d_d = _mm_loadh_pd(d_d, (const double *)(ref_data + (y + 3) * ref_stride + overhang));
__m128i a = _mm_castpd_si128(a_d);
__m128i b = _mm_castpd_si128(b_d);
__m128i c = _mm_castpd_si128(c_d);
__m128i d = _mm_castpd_si128(d_d);
__m128i b_epol = _mm_shuffle_epi8(b, epol_mask);
__m128i d_epol = _mm_shuffle_epi8(d, epol_mask);
__m128i curr_sads_ab = _mm_sad_epu8(a, b_epol);
__m128i curr_sads_cd = _mm_sad_epu8(c, d_epol);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_cd);
}
if (height_residual_lines) {
for (; y < height; y++) {
__m64 a = *(__m64 *)(pic_data + y * pic_stride);
__m64 b = *(__m64 *)(ref_data + y * ref_stride + overhang);
__m64 b_epol = _mm_shuffle_pi8(b, epol_mask_64);
__m64 sads = _mm_sad_pu8(a, b_epol);
result += (uint64_t)sads;
}
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
result += _mm_cvtsi128_si32(sad);
return result;
}
/*
* overhang is a measure of how many pixels the intended starting X coordinate
* is right from the one pointed to by pic_data, ie. abs(x0). We can read
* starting from X = 0, and to preserve pixel alignment, shift all the bytes
* right "left" places while duplicating the leftmost pixel (extrapolating it
* to the left of the image buffer).
*/
static uint32_t hor_sad_left_sse41_w16(const kvz_pixel *pic_data, const kvz_pixel *ref_data,
int32_t width, int32_t height, uint32_t pic_stride,
uint32_t ref_stride, uint32_t overhang)
{
const __m128i excess = _mm_set1_epi8(overhang);
const __m128i ns = _mm_setr_epi8(0, 1, 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15);
const __m128i mask1 = _mm_sub_epi8 (ns, excess);
const __m128i epol_mask = _mm_max_epi8 (mask1, _mm_setzero_si128());
const int32_t height_fourline_groups = height & ~3;
const int32_t height_residual_lines = height & 3;
__m128i sse_inc = _mm_setzero_si128();
int32_t y;
for (y = 0; y < height_fourline_groups; y += 4) {
__m128i a = _mm_loadu_si128((__m128i *)(pic_data + (y + 0) * pic_stride));
__m128i b = _mm_loadu_si128((__m128i *)(ref_data + (y + 0) * ref_stride + overhang));
__m128i c = _mm_loadu_si128((__m128i *)(pic_data + (y + 1) * pic_stride));
__m128i d = _mm_loadu_si128((__m128i *)(ref_data + (y + 1) * ref_stride + overhang));
__m128i e = _mm_loadu_si128((__m128i *)(pic_data + (y + 2) * pic_stride));
__m128i f = _mm_loadu_si128((__m128i *)(ref_data + (y + 2) * ref_stride + overhang));
__m128i g = _mm_loadu_si128((__m128i *)(pic_data + (y + 3) * pic_stride));
__m128i h = _mm_loadu_si128((__m128i *)(ref_data + (y + 3) * ref_stride + overhang));
__m128i b_epol = _mm_shuffle_epi8(b, epol_mask);
__m128i d_epol = _mm_shuffle_epi8(d, epol_mask);
__m128i f_epol = _mm_shuffle_epi8(f, epol_mask);
__m128i h_epol = _mm_shuffle_epi8(h, epol_mask);
__m128i curr_sads_ab = _mm_sad_epu8(a, b_epol);
__m128i curr_sads_cd = _mm_sad_epu8(c, d_epol);
__m128i curr_sads_ef = _mm_sad_epu8(e, f_epol);
__m128i curr_sads_gh = _mm_sad_epu8(g, h_epol);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_cd);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ef);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_gh);
}
if (height_residual_lines) {
for (; y < height; y++) {
__m128i a = _mm_loadu_si128((__m128i *)(pic_data + (y + 0) * pic_stride));
__m128i b = _mm_loadu_si128((__m128i *)(ref_data + (y + 0) * ref_stride + overhang));
__m128i b_epol = _mm_shuffle_epi8(b, epol_mask);
__m128i curr_sads = _mm_sad_epu8(a, b_epol);
sse_inc = _mm_add_epi64(sse_inc, curr_sads);
}
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
static uint32_t hor_sad_sse41_w32(const kvz_pixel *pic_data, const kvz_pixel *ref_data,
First version of arbitrary-width SSE4.1 hor_sad
2019-01-30 20:57:06 +00:00
int32_t width, int32_t height, uint32_t pic_stride,
uint32_t ref_stride, uint32_t left, uint32_t right)
Start work on left and right border aware hor_sad Comes with 4, 8, 16 and 32 pixel wide implementations now, at some point investigate if this can start to thrash icache
2019-01-22 19:27:21 +00:00
{
const int32_t height_twoline_groups = height & ~1;
const int32_t height_residual_lines = height & 1;
__m128i nslo = _mm_setr_epi8(0, 1, 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15);
__m128i nshi = _mm_setr_epi8(16, 17, 18, 19, 20, 21, 22, 23,
24, 25, 26, 27, 28, 29, 30, 31);
__m128i epol_masklo, epol_maskhi;
int32_t border_pix_off;
if (left) {
border_pix_off = left;
__m128i first_valid_idx = _mm_set1_epi8(left);
epol_masklo = _mm_cmpgt_epi8(first_valid_idx, nslo);
epol_maskhi = _mm_cmpgt_epi8(first_valid_idx, nshi);
} else {
border_pix_off = width - (right + 1);
__m128i last_valid_idx = _mm_set1_epi8(border_pix_off);
epol_masklo = _mm_cmpgt_epi8(nslo, last_valid_idx);
epol_maskhi = _mm_cmpgt_epi8(nshi, last_valid_idx);
}
__m128i sse_inc = _mm_setzero_si128();
int32_t y;
for (y = 0; y < height_twoline_groups; y += 2) {
__m128i a = _mm_loadu_si128((__m128i *)(pic_data + (y + 0) * pic_stride + 0));
__m128i b = _mm_loadu_si128((__m128i *)(ref_data + (y + 0) * ref_stride + 0));
__m128i c = _mm_loadu_si128((__m128i *)(pic_data + (y + 0) * pic_stride + 16));
__m128i d = _mm_loadu_si128((__m128i *)(ref_data + (y + 0) * ref_stride + 16));
__m128i e = _mm_loadu_si128((__m128i *)(pic_data + (y + 1) * pic_stride + 0));
__m128i f = _mm_loadu_si128((__m128i *)(ref_data + (y + 1) * ref_stride + 0));
__m128i g = _mm_loadu_si128((__m128i *)(pic_data + (y + 1) * pic_stride + 16));
__m128i h = _mm_loadu_si128((__m128i *)(ref_data + (y + 1) * ref_stride + 16));
__m128i border_px_lo = _mm_set1_epi8 (*(uint8_t *)(ref_data + (y + 0) * ref_stride + border_pix_off));
__m128i border_px_hi = _mm_set1_epi8 (*(uint8_t *)(ref_data + (y + 1) * ref_stride + border_pix_off));
__m128i b_epol = _mm_blendv_epi8(b, border_px_lo, epol_masklo);
__m128i d_epol = _mm_blendv_epi8(d, border_px_lo, epol_maskhi);
__m128i f_epol = _mm_blendv_epi8(f, border_px_hi, epol_masklo);
__m128i h_epol = _mm_blendv_epi8(h, border_px_hi, epol_maskhi);
__m128i curr_sads_ab = _mm_sad_epu8(a, b_epol);
__m128i curr_sads_cd = _mm_sad_epu8(c, d_epol);
__m128i curr_sads_ef = _mm_sad_epu8(e, f_epol);
__m128i curr_sads_gh = _mm_sad_epu8(g, h_epol);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_cd);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ef);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_gh);
}
if (height_residual_lines) {
__m128i a = _mm_loadu_si128((__m128i *)(pic_data + (y + 0) * pic_stride + 0));
__m128i b = _mm_loadu_si128((__m128i *)(ref_data + (y + 0) * ref_stride + 0));
__m128i c = _mm_loadu_si128((__m128i *)(pic_data + (y + 0) * pic_stride + 16));
__m128i d = _mm_loadu_si128((__m128i *)(ref_data + (y + 0) * ref_stride + 16));
__m128i border_px = _mm_set1_epi8 (*(uint8_t *)(ref_data + (y + 0) * ref_stride + border_pix_off));
__m128i b_epol = _mm_blendv_epi8(b, border_px, epol_masklo);
__m128i d_epol = _mm_blendv_epi8(d, border_px, epol_maskhi);
__m128i curr_sads_ab = _mm_sad_epu8(a, b_epol);
__m128i curr_sads_cd = _mm_sad_epu8(c, d_epol);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_cd);
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
First version of arbitrary-width SSE4.1 hor_sad
2019-01-30 20:57:06 +00:00
static uint32_t hor_sad_sse41_arbitrary(const kvz_pixel *pic_data, const kvz_pixel *ref_data,
int32_t width, int32_t height, uint32_t pic_stride,
uint32_t ref_stride, uint32_t left, uint32_t right)
{
const size_t xmm_width = 16;
const __m128i xmm_widths = _mm_set1_epi8(xmm_width);
// Bytes in block in 128-bit blocks per each scanline, and remainder
const int32_t width_xmms = width & ~(xmm_width - 1);
const int32_t width_residual_pixels = width & (xmm_width - 1);
const int32_t height_fourline_groups = height & ~3;
const int32_t height_residual_lines = height & 3;
__m128i ns = _mm_setr_epi8(0, 1, 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15);
const __m128i rds = _mm_set1_epi8 (width_residual_pixels);
const __m128i rdmask = _mm_cmpgt_epi8(rds, ns);
int32_t border_idx;
__m128i is_right_border = _mm_setzero_si128();
if (left) {
border_idx = left;
} else {
border_idx = width - (right + 1);
is_right_border = _mm_cmpeq_epi8(is_right_border, is_right_border);
}
const __m128i epol_src_idx = _mm_set1_epi8(border_idx);
int32_t x, y;
__m128i sse_inc = _mm_setzero_si128();
__m128i epol_mask;
for (x = 0; x < width_xmms; x += xmm_width) {
// This is a dirty hack, but it saves us an easily predicted branch! It
// also marks the first or last valid pixel (the border one) for
// extrapolating, but that makes no difference since the pixels marked
// for extrapolation will always be written over with that exact pixel's
// value.
epol_mask = _mm_cmpgt_epi8(epol_src_idx, ns);
epol_mask = _mm_xor_si128 (epol_mask, is_right_border);
for (y = 0; y < height; y++) {
__m128i a = _mm_loadu_si128((__m128i *)(pic_data + (y + 0) * pic_stride + x));
__m128i b = _mm_loadu_si128((__m128i *)(ref_data + (y + 0) * ref_stride + x));
__m128i border_px_b = _mm_set1_epi8 (*(uint8_t *)(ref_data + (y + 0) * ref_stride + border_idx));
__m128i b_epol = _mm_blendv_epi8(b, border_px_b, epol_mask);
__m128i curr_sads_ab = _mm_sad_epu8(a, b_epol);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
}
ns = _mm_add_epi8(ns, xmm_widths);
}
if (width_residual_pixels) {
epol_mask = _mm_cmpgt_epi8(epol_src_idx, ns);
epol_mask = _mm_xor_si128 (epol_mask, is_right_border);
for (y = 0; y < height; y++) {
__m128i a = _mm_loadu_si128((__m128i *)(pic_data + (y + 0) * pic_stride + x));
__m128i b = _mm_loadu_si128((__m128i *)(ref_data + (y + 0) * ref_stride + x));
__m128i border_px_b = _mm_set1_epi8 (*(uint8_t *)(ref_data + (y + 0) * ref_stride + border_idx));
__m128i b_epol_1 = _mm_blendv_epi8(b, border_px_b, epol_mask);
__m128i b_epol_2 = _mm_blendv_epi8(a, b_epol_1, rdmask);
__m128i curr_sads_ab = _mm_sad_epu8(a, b_epol_2);
sse_inc = _mm_add_epi64(sse_inc, curr_sads_ab);
}
}
__m128i sse_inc_2 = _mm_shuffle_epi32(sse_inc, _MM_SHUFFLE(1, 0, 3, 2));
__m128i sad = _mm_add_epi64 (sse_inc, sse_inc_2);
return _mm_cvtsi128_si32(sad);
}
Move SSE4.1 SAD functions to separate header
2019-01-15 17:37:53 +00:00
#endif
Reference in a new issue Copy permalink