Debian Codegraph @

xvmc/shader/mc/ipicture_igd.g4a

raw
1 /*
2 GRF allocation:
3 g1~g30: constant buffer
4 g1~g2:intra IQ matrix
5 g3~g4:non intra IQ matrix
6 g5~g20:IDCT tab
7 g31: read and write message descriptor
8 g32~g55:DCT data
9 g58~g81:reference data
10 g82: thread payload
11 g83~g106:IDCT data
12 */
13 mov (8) g82.0<1>UD g31.0<8,8,1>UD {align1};
14
15 mov.sat (16) g33.0<2>UB g33.0<16,16,1>W {align1};
16 mov.sat (16) g34.0<2>UB g34.0<16,16,1>W {align1};
17 mov.sat (16) g35.0<2>UB g35.0<16,16,1>W {align1};
18 mov.sat (16) g36.0<2>UB g36.0<16,16,1>W {align1};
19 mov.sat (16) g37.0<2>UB g37.0<16,16,1>W {align1};
20 mov.sat (16) g38.0<2>UB g38.0<16,16,1>W {align1};
21 mov.sat (16) g39.0<2>UB g39.0<16,16,1>W {align1};
22 mov.sat (16) g40.0<2>UB g40.0<16,16,1>W {align1};
23 mov.sat (16) g41.0<2>UB g41.0<16,16,1>W {align1};
24 mov.sat (16) g42.0<2>UB g42.0<16,16,1>W {align1};
25 mov.sat (16) g43.0<2>UB g43.0<16,16,1>W {align1};
26 mov.sat (16) g44.0<2>UB g44.0<16,16,1>W {align1};
27 mov.sat (16) g45.0<2>UB g45.0<16,16,1>W {align1};
28 mov.sat (16) g46.0<2>UB g46.0<16,16,1>W {align1};
29 mov.sat (16) g47.0<2>UB g47.0<16,16,1>W {align1};
30 mov.sat (16) g48.0<2>UB g48.0<16,16,1>W {align1};
31
32 mov.sat (16) g49.0<2>UB g49.0<16,16,1>W {align1};
33 mov.sat (16) g50.0<2>UB g50.0<16,16,1>W {align1};
34 mov.sat (16) g51.0<2>UB g51.0<16,16,1>W {align1};
35 mov.sat (16) g52.0<2>UB g52.0<16,16,1>W {align1};
36 mov.sat (16) g53.0<2>UB g53.0<16,16,1>W {align1};
37 mov.sat (16) g54.0<2>UB g54.0<16,16,1>W {align1};
38 mov.sat (16) g55.0<2>UB g55.0<16,16,1>W {align1};
39 mov.sat (16) g56.0<2>UB g56.0<16,16,1>W {align1};
40
41 mov (1) g31.8<1>UD 0x00F000FUD {align1};
42
43 and.nz (1) null g82.30<1,1,1>UB 0x1UW{align1};
44 (f0) jmpi field_dct_y;
45
46 mov (16) m1.0<1>UB g33.0<16,16,2>UB {align1};
47 mov (16) m1.16<1>UB g34.0<16,16,2>UB {align1};
48 mov (16) m2.0<1>UB g35.0<16,16,2>UB {align1};
49 mov (16) m2.16<1>UB g36.0<16,16,2>UB {align1};
50 mov (16) m3.0<1>UB g37.0<16,16,2>UB {align1};
51 mov (16) m3.16<1>UB g38.0<16,16,2>UB {align1};
52 mov (16) m4.0<1>UB g39.0<16,16,2>UB {align1};
53 mov (16) m4.16<1>UB g40.0<16,16,2>UB {align1};
54 mov (16) m5.0<1>UB g41.0<16,16,2>UB {align1};
55 mov (16) m5.16<1>UB g42.0<16,16,2>UB {align1};
56 mov (16) m6.0<1>UB g43.0<16,16,2>UB {align1};
57 mov (16) m6.16<1>UB g44.0<16,16,2>UB {align1};
58 mov (16) m7.0<1>UB g45.0<16,16,2>UB {align1};
59 mov (16) m7.16<1>UB g46.0<16,16,2>UB {align1};
60 mov (16) m8.0<1>UB g47.0<16,16,2>UB {align1};
61 mov (16) m8.16<1>UB g48.0<16,16,2>UB {align1};
62 jmpi write_back_y;
63
64 field_dct_y:
65 mov (16) m1.0<1>UB g33.0<16,16,2>UB {align1};
66 mov (16) m1.16<1>UB g41.0<16,16,2>UB {align1};
67 mov (16) m2.0<1>UB g34.0<16,16,2>UB {align1};
68 mov (16) m2.16<1>UB g42.0<16,16,2>UB {align1};
69 mov (16) m3.0<1>UB g35.0<16,16,2>UB {align1};
70 mov (16) m3.16<1>UB g43.0<16,16,2>UB {align1};
71 mov (16) m4.0<1>UB g36.0<16,16,2>UB {align1};
72 mov (16) m4.16<1>UB g44.0<16,16,2>UB {align1};
73 mov (16) m5.0<1>UB g37.0<16,16,2>UB {align1};
74 mov (16) m5.16<1>UB g45.0<16,16,2>UB {align1};
75 mov (16) m6.0<1>UB g38.0<16,16,2>UB {align1};
76 mov (16) m6.16<1>UB g46.0<16,16,2>UB {align1};
77 mov (16) m7.0<1>UB g39.0<16,16,2>UB {align1};
78 mov (16) m7.16<1>UB g47.0<16,16,2>UB {align1};
79 mov (16) m8.0<1>UB g40.0<16,16,2>UB {align1};
80 mov (16) m8.16<1>UB g48.0<16,16,2>UB {align1};
81
82 write_back_y:
83 send (16) 0 acc0<1>UW g31<8,8,1>UW write(0,0,2,0) mlen 9 rlen 0 {align1};
84
85 //U
86 mov (1) g31.8<1>UD 0x0070007UD { align1 };
87 shr (2) g31.0<1>UD g82.0<2,2,1>UD 1D {align1};
88
89 mov (16) m1.0<1>UB g49.0<16,16,2>UB {align1};
90 mov (16) m1.16<1>UB g50.0<16,16,2>UB {align1};
91 mov (16) m2.0<1>UB g51.0<16,16,2>UB {align1};
92 mov (16) m2.16<1>UB g52.0<16,16,2>UB {align1};
93 send (16) 0 acc0<1>UW g31<8,8,1>UW write(1, 0, 2, 0) mlen 3 rlen 0 { align1 };
94
95 //V
96 mov (16) m1.0<1>UB g53.0<16,16,2>UB {align1};
97 mov (16) m1.16<1>UB g54.0<16,16,2>UB {align1};
98 mov (16) m2.0<1>UB g55.0<16,16,2>UB {align1};
99 mov (16) m2.16<1>UB g56.0<16,16,2>UB {align1};
100 send (16) 0 acc0<1>UW g31<8,8,1>UW write(2, 0, 2, 0) mlen 3 rlen 0 { align1 };
101
102 OUT:
103 send (16) 0 acc0<1>UW g0<8,8,1>UW
104 thread_spawner(0, 0, 0) mlen 1 rlen 0 { align1 EOT};
105