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
|
/*
* Intra predict 8x8 chroma block
* Copyright © <2010>, Intel Corporation.
*
* This program is licensed under the terms and conditions of the
* Eclipse Public License (EPL), version 1.0. The full text of the EPL is at
* http://www.opensource.org/licenses/eclipse-1.0.php.
*
*/
#if !defined(__INTRA_PRED_CHROMA__) // Make sure this is only included once
#define __INTRA_PRED_CHROMA__
// Module name: intra_Pred_Chroma.asm
//
// Intra predict 8x8 chroma block
//
shr (1) PINTRAPRED_UV<1>:w REG_INTRA_CHROMA_PRED_MODE<0;1,0>:ub INTRA_CHROMA_PRED_MODE_SHIFT:w // Bits 1:0 = intra chroma pred mode
// WA for "jmpi" restriction
mov (1) REG_INTRA_TEMP_1<1>:d r[PINTRAPRED_UV, INTRA_CHROMA_OFFSET]:b
jmpi (1) REG_INTRA_TEMP_1<0;1,0>:d
// Mode 0
INTRA_CHROMA_DC:
and.nz.f0.0 (8) NULLREG REG_INTRA_PRED_AVAIL_FLAG INTRA_PRED_UP_AVAIL_FLAG:ud // Top macroblock available for intra prediction?
// Calculate DC values for sub-block 0 and 3
//
// Rearrange reference samples for unified DC prediction code
// Need to check INTRA_PRED_LEFT_TH_AVAIL_FLAG for blk0 and INTRA_PRED_LEFT_BH_AVAIL_FLAG for blk3
//
(-f0.0.any8h) mov (8) INTRA_REF_TOP_W(0)<1> 0x8080:uw // Up not available
and.nz.f0.1 (4) NULLREG REG_INTRA_PRED_AVAIL_FLAG INTRA_PRED_LEFT_TH_AVAIL_FLAG:ud
(-f0.1.any4h) mov (4) INTRA_REF_LEFT_W(0)<2> INTRA_REF_TOP_W(0)REGION(4,1) // Left top half macroblock not available for intra prediction
and.nz.f0.1 (4) NULLREG REG_INTRA_PRED_AVAIL_FLAG INTRA_PRED_LEFT_BH_AVAIL_FLAG:ud
(-f0.1.any4h) mov (4) INTRA_REF_LEFT_W(0,8)<2> INTRA_REF_TOP_W(0,4)REGION(4,1) // Left bottom half macroblock not available for intra prediction
(-f0.0.any8h) mov (8) INTRA_REF_TOP_W(0)<1> INTRA_REF_LEFT_W(0)REGION(8,2) // Up not available
// Calculate DC prediction
//
add (16) PRED_UVW(0)<1> INTRA_REF_TOP(0)REGION(16,1) INTRA_REF_LEFT_UV(0)<4;2,1> // Sum of top and left reference
add (8) PRED_UVW(0)<1> PRED_UVW(0)<4;2,1> PRED_UVW(0,2)<4;2,1> // Sum of first half (blk #0) and second half (blk #3)
add (8) PRED_UVW(9)<1> PRED_UVW(0)<0;2,1> PRED_UVW(0,2)<0;2,1> // Sum of blk #0
add (8) PRED_UVW(11,8)<1> PRED_UVW(0,4)<0;2,1> PRED_UVW(0,6)<0;2,1> // Sum of blk #3
// Calculate DC values for sub-block 1 and 2
//
// Rearrange reference samples for unified DC prediction code
//
// Blk #2
(-f0.0.any4h) mov (4) INTRA_REF_TOP_W(0)<1> 0x8080:uw
(f0.1.any4h) mov (4) INTRA_REF_TOP_W(0)<1> INTRA_REF_LEFT_W(0,8)REGION(4,2) // Always use available left reference
(-f0.1.any4h) mov (4) INTRA_REF_LEFT_W(0,8)<2> INTRA_REF_TOP_W(0)REGION(4,1)
// Blk #1
and.nz.f0.1 (4) NULLREG REG_INTRA_PRED_AVAIL_FLAG INTRA_PRED_LEFT_TH_AVAIL_FLAG:ud
(-f0.1.any4h) mov (4) INTRA_REF_LEFT_W(0)<2> 0x8080:uw
(f0.0.any4h) mov (4) INTRA_REF_LEFT_W(0)<2> INTRA_REF_TOP_W(0,4)REGION(4,1) // Always use available top reference
(-f0.0.any4h) mov (4) INTRA_REF_TOP_W(0,4)<1> INTRA_REF_LEFT_W(0)REGION(4,2)
// Calculate DC prediction
//
add (8) PRED_UVW(0)<1> INTRA_REF_TOP(0)REGION(8,1) INTRA_REF_LEFT_UV(0,16)<4;2,1> // Sum of top and left reference for blk #2
add (8) PRED_UVW(0,8)<1> INTRA_REF_LEFT_UV(0)<4;2,1> INTRA_REF_TOP(0,8)REGION(8,1) // Sum of top and left reference for blk #1
add (8) PRED_UVW(0)<1> PRED_UVW(0)<4;2,1> PRED_UVW(0,2)<4;2,1> // Sum of first half (blk #2) and second half (blk #1)
add (8) PRED_UVW(9,8)<1> PRED_UVW(0,4)<0;2,1> PRED_UVW(0,6)<0;2,1> // Sum of blk #1
add (8) PRED_UVW(11)<1> PRED_UVW(0)<0;2,1> PRED_UVW(0,2)<0;2,1> // Sum of blk #2
// Now, PRED_UVW(9) holds sums for blks #0 and #1 and PRED_UVW(11) holds sums for blks #2 and #3
//
add (32) acc0<1>:w PRED_UVW(9)REGION(16,1) 4:w {Compr} // Add rounder
$for(0; <4; 2) {
shr (32) PRED_UVW(%1)<1> acc0:w 3:w {Compr}
}
add (32) acc0<1>:w PRED_UVW(11)REGION(16,1) 4:w {Compr} // Add rounder
$for(4; <8; 2) {
shr (32) PRED_UVW(%1)<1> acc0:w 3:w {Compr}
}
jmpi (1) End_of_intra_Pred_Chroma
// Mode 1
INTRA_CHROMA_HORIZONTAL:
mov (1) PREF_LEFT_UD<1>:ud INTRA_REF_LEFT_ID*GRFWIB*0x00010001+0x00040000:ud // Set address registers for instruction compression
$for(0,0; <8; 2,8) {
mov (32) PRED_UVW(%1)<1> r[PREF_LEFT,%2+2]<0;2,1>:ub {Compr} // Actual left column reference data start at offset 2
}
jmpi (1) End_of_intra_Pred_Chroma
// Mode 2
INTRA_CHROMA_VERTICAL:
$for(0; <8; 2) {
mov (32) PRED_UVW(%1)<1> INTRA_REF_TOP(0) {Compr}
}
jmpi (1) End_of_intra_Pred_Chroma
// Mode 3
INTRA_Chroma_PLANE:
// Refer to H.264/AVC spec Section 8.3.4.4
#undef C
#define A REG_INTRA_TEMP_2.0 // All are WORD type
#define B REG_INTRA_TEMP_3.0 // B[U] & B[V]
#define C REG_INTRA_TEMP_3.2 // C[U] & C[V]
#define YP REG_INTRA_TEMP_0 // Store intermediate results of c*(y-3). Make sure it's an even GRF
#define YP1 REG_INTRA_TEMP_1 // Store intermediate results of c*(y-3). Make sure it's an odd GRF
#define XP REG_INTRA_TEMP_5 // Store intermediate results of a+b*(x-3)+16. Make sure it's an odd GRF
// First Calculate constants H and V
// H1 = sum((x'+1)*p[4+x',-1]), x'=0,1,2,3
// H2 = sum((-x'-1)*p[2-x',-1]), x'=3,2,1,0
// H = H1 + H2
// The same calculation holds for V
//
mul (8) H1(0)<1> INTRA_REF_TOP(0,8)REGION(8,1) 0x44332211:v
mul (8) H2(0)<1> INTRA_REF_TOP(0,-2)REGION(8,1) 0xFFEEDDCC:v
mul (8) V1(0)<1> INTRA_REF_LEFT_UV(0,4*4)<4;2,1> 0x44332211:v
mul (8) V2(0)<1> INTRA_REF_LEFT_UV(0)<4;2,1> 0x00FFEEDD:v
mul (2) V2(0,6)<1> INTRA_REF_TOP(0,-2)REGION(2,1) -4:w // Replace 0*p[-1,3] with -4*p[-1,-1]
// Now, REG_INTRA_TEMP_0 holds [H2, H1] and REG_INTRA_TEMP_1 holds [V2, V1]
// Sum up [H2, H1] and [V2, V1] using instruction compression
// ExecSize = 16 is restricted by B-spec for instruction compression
// Actual intermediate results are in lower sub-registers after each summing step
add (16) H1(0)<1> H1(0) H2(0) {Compr} // Results in lower 8 WORDs
add (16) H1(0)<1> H1(0) H1(0,4) {Compr} // Results in lower 4 WORDs
add (16) H1(0)<1> H1(0) H1(0,2) {Compr} // Results in lower 2 WORDs
// Calculate a, b, c and further derivations
mov (16) acc0<1>:w 32:w
mac (4) acc0<1>:w H1(0)<16;2,1> 34:w
shr (4) B<1>:w acc0:w 6:w // Done b,c
mov (16) acc0<1>:w 16:w
mac (16) acc0<1>:w INTRA_REF_TOP(0,7*2)<0;2,1> 16:w
mac (16) A<1>:w INTRA_REF_LEFT_UV(0,7*4)<0;2,1> 16:w // A = a+16
mac (16) XP<1>:w B<0;2,1>:w XY_3<1;2,0>:b // XP = A+b*(x-3)
mul (8) YP<1>:w C<0;2,1>:w XY_3<2;2,0>:b // YP = c*(y-3), Even portion
mul (8) YP1<1>:w C<0;2,1>:w XY_3_1<2;2,0>:b // YP = c*(y-3), Odd portion
// Finally the intra_Chroma plane prediction
$for(0; <8; 2) {
add (32) acc0<1>:w XP<16;16,1>:w YP.%1<0;2,1>:w {Compr}
shr.sat (32) PRED_UV(%1)<2> acc0<16;16,1>:w 5:w {Compr}
}
End_of_intra_Pred_Chroma:
// End of intra_Pred_Chroma
#endif // !defined(__INTRA_PRED_CHROMA__)
|
