1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
|
// Copyright (c) 2017-2018 Intel Corporation
//
// Permission is hereby granted, free of charge, to any person obtaining a copy
// of this software and associated documentation files (the "Software"), to deal
// in the Software without restriction, including without limitation the rights
// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
// copies of the Software, and to permit persons to whom the Software is
// furnished to do so, subject to the following conditions:
//
// The above copyright notice and this permission notice shall be included in all
// copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
// SOFTWARE.
#pragma once
#ifndef _ASC_COMMON_IMPL_H_
#define _ASC_COMMON_IMPL_H_
#include "asc_defs.h"
#include "asc_structures.h"
static const int HIST_THRESH_LO = 1;
static const int HIST_THRESH_HI = 12;
#define SAD_SEARCH_VSTEP 2 // 1=FS 2=FHS
ASC_ALIGN_DECL(16) static const mfxU16 tab_twostep[8] = {
0x0000, 0xffff, 0x0000, 0xffff, 0x0000, 0xffff, 0x0000, 0xffff,
};
ASC_ALIGN_DECL(16) static const mfxU16 tab_killmask[8][8] = {
{ 0xffff, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff },
{ 0x0000, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff },
{ 0x0000, 0x0000, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff },
{ 0x0000, 0x0000, 0x0000, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff },
{ 0x0000, 0x0000, 0x0000, 0x0000, 0xffff, 0xffff, 0xffff, 0xffff },
{ 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xffff, 0xffff, 0xffff },
{ 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xffff, 0xffff },
{ 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xffff },
};
#define _mm_loadh_epi64(a, ptr) _mm_castps_si128(_mm_loadh_pi(_mm_castsi128_ps(a), (__m64 *)(ptr)))
#define _mm_movehl_epi64(a, b) _mm_castps_si128(_mm_movehl_ps(_mm_castsi128_ps(a), _mm_castsi128_ps(b)))
// Load 0..3 floats to XMM register from memory
// NOTE: elements of XMM are permuted [ 2 - 1 ]
static inline __m128 LoadPartialXmm(float *pSrc, mfxI32 len)
{
__m128 xmm = _mm_setzero_ps();
if (len & 2)
{
xmm = _mm_loadh_pi(xmm, (__m64 *)pSrc);
pSrc += 2;
}
if (len & 1)
{
xmm = _mm_move_ss(xmm, _mm_load_ss(pSrc));
}
return xmm;
}
// Store 0..3 floats from XMM register to memory
// NOTE: elements of XMM are permuted [ 2 - 1 ]
static inline void StorePartialXmm(float *pDst, __m128 xmm, mfxI32 len)
{
if (len & 2)
{
_mm_storeh_pi((__m64 *)pDst, xmm);
pDst += 2;
}
if (len & 1)
{
_mm_store_ss(pDst, xmm);
}
}
#if defined(__AVX2__)
// Load 0..7 floats to YMM register from memory
// NOTE: elements of YMM are permuted [ 4 2 - 1 ]
static inline __m256 LoadPartialYmm(float *pSrc, mfxI32 len)
{
__m128 xlo = _mm_setzero_ps();
__m128 xhi = _mm_setzero_ps();
if (len & 4) {
xhi = _mm_loadu_ps(pSrc);
pSrc += 4;
}
if (len & 2) {
xlo = _mm_loadh_pi(xlo, (__m64 *)pSrc);
pSrc += 2;
}
if (len & 1) {
xlo = _mm_move_ss(xlo, _mm_load_ss(pSrc));
}
return _mm256_insertf128_ps(_mm256_castps128_ps256(xlo), xhi, 1);
}
// Store 0..7 floats from YMM register to memory
// NOTE: elements of YMM are permuted [ 4 2 - 1 ]
static inline void StorePartialYmm(float *pDst, __m256 ymm, mfxI32 len)
{
__m128 xlo = _mm256_castps256_ps128(ymm);
__m128 xhi = _mm256_extractf128_ps(ymm, 1);
if (len & 4) {
_mm_storeu_ps(pDst, xhi);
pDst += 4;
}
if (len & 2) {
_mm_storeh_pi((__m64 *)pDst, xlo);
pDst += 2;
}
if (len & 1) {
_mm_store_ss(pDst, xlo);
}
}
#endif //defined(__AVX2__)
// Load 0..15 bytes to XMM register from memory
// NOTE: elements of XMM are permuted [ 8 4 2 - 1 ]
template <char init>
static inline __m128i LoadPartialXmm(unsigned char *pSrc, mfxI32 len)
{
__m128i xmm = _mm_set1_epi8(init);
if (len & 8) {
xmm = _mm_loadh_epi64(xmm, (__m64 *)pSrc);
pSrc += 8;
}
if (len & 4) {
xmm = _mm_insert_epi32(xmm, *((int *)pSrc), 1);
pSrc += 4;
}
if (len & 2) {
xmm = _mm_insert_epi16(xmm, *((short *)pSrc), 1);
pSrc += 2;
}
if (len & 1) {
xmm = _mm_insert_epi8(xmm, *pSrc, 0);
}
return xmm;
}
#if defined(__AVX2__)
// Load 0..31 bytes to YMM register from memory
// NOTE: elements of YMM are permuted [ 16 8 4 2 - 1 ]
template <char init>
static inline __m256i LoadPartialYmm(unsigned char *pSrc, mfxI32 len)
{
__m128i xlo = _mm_set1_epi8(init);
__m128i xhi = _mm_set1_epi8(init);
if (len & 16) {
xhi = _mm_loadu_si128((__m128i *)pSrc);
pSrc += 16;
}
if (len & 8) {
xlo = _mm_loadh_epi64(xlo, (__m64 *)pSrc);
pSrc += 8;
}
if (len & 4) {
xlo = _mm_insert_epi32(xlo, *((int *)pSrc), 1);
pSrc += 4;
}
if (len & 2) {
xlo = _mm_insert_epi16(xlo, *((short *)pSrc), 1);
pSrc += 2;
}
if (len & 1) {
xlo = _mm_insert_epi8(xlo, *pSrc, 0);
}
return _mm256_inserti128_si256(_mm256_castsi128_si256(xlo), xhi, 1);
}
#endif //defined(__AVX2__)
static inline void calc_RACA_4x4_C(mfxU8 *pSrc, mfxI32 pitch, mfxI32 *RS, mfxI32 *CS) {
mfxI32 i, j;
mfxU8 *pS = pSrc;
mfxU8 *pS2 = pSrc + pitch;
mfxI32 Rs, Cs;
Cs = 0;
Rs = 0;
for (i = 0; i < 4; i++)
{
for (j = 0; j < 4; j++)
{
Cs += (pS[j] > pS[j + 1]) ? (pS[j] - pS[j + 1]) : (pS[j + 1] - pS[j]);
Rs += (pS[j] > pS2[j]) ? (pS[j] - pS2[j]) : (pS2[j] - pS[j]);
}
pS += pitch;
pS2 += pitch;
}
*CS += Cs >> 4;
*RS += Rs >> 4;
}
#endif //_ASC_COMMON_IMPL_H_
|
