summaryrefslogtreecommitdiff
path: root/libavcodec/i386/dsputil_mmx.h
blob: a4b3bfcdacacd946a43ee779e6249e89f45e5cdb (plain)
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
/*
 * MMX optimized DSP utils
 * Copyright (c) 2007  Aurelien Jacobs <aurel@gnuage.org>
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * FFmpeg is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#ifndef FFMPEG_DSPUTIL_MMX_H
#define FFMPEG_DSPUTIL_MMX_H

#include <stdint.h>
#include "libavcodec/dsputil.h"

typedef struct { uint64_t a, b; } xmm_t;

extern const uint64_t ff_bone;
extern const uint64_t ff_wtwo;

extern const uint64_t ff_pdw_80000000[2];

extern const uint64_t ff_pw_3;
extern const uint64_t ff_pw_4;
extern const xmm_t    ff_pw_5;
extern const xmm_t    ff_pw_8;
extern const uint64_t ff_pw_15;
extern const xmm_t    ff_pw_16;
extern const uint64_t ff_pw_20;
extern const xmm_t    ff_pw_28;
extern const xmm_t    ff_pw_32;
extern const uint64_t ff_pw_42;
extern const uint64_t ff_pw_64;
extern const uint64_t ff_pw_96;
extern const uint64_t ff_pw_128;
extern const uint64_t ff_pw_255;

extern const uint64_t ff_pb_1;
extern const uint64_t ff_pb_3;
extern const uint64_t ff_pb_7;
extern const uint64_t ff_pb_3F;
extern const uint64_t ff_pb_A1;
extern const uint64_t ff_pb_FC;

extern const double ff_pd_1[2];
extern const double ff_pd_2[2];

#define LOAD4(stride,in,a,b,c,d)\
    "movq 0*"#stride"+"#in", "#a"\n\t"\
    "movq 1*"#stride"+"#in", "#b"\n\t"\
    "movq 2*"#stride"+"#in", "#c"\n\t"\
    "movq 3*"#stride"+"#in", "#d"\n\t"

#define STORE4(stride,out,a,b,c,d)\
    "movq "#a", 0*"#stride"+"#out"\n\t"\
    "movq "#b", 1*"#stride"+"#out"\n\t"\
    "movq "#c", 2*"#stride"+"#out"\n\t"\
    "movq "#d", 3*"#stride"+"#out"\n\t"

/* in/out: mma=mma+mmb, mmb=mmb-mma */
#define SUMSUB_BA( a, b ) \
    "paddw "#b", "#a" \n\t"\
    "paddw "#b", "#b" \n\t"\
    "psubw "#a", "#b" \n\t"

#define SBUTTERFLY(a,b,t,n,m)\
    "mov" #m " " #a ", " #t "         \n\t" /* abcd */\
    "punpckl" #n " " #b ", " #a "     \n\t" /* aebf */\
    "punpckh" #n " " #b ", " #t "     \n\t" /* cgdh */\

#define TRANSPOSE4(a,b,c,d,t)\
    SBUTTERFLY(a,b,t,wd,q) /* a=aebf t=cgdh */\
    SBUTTERFLY(c,d,b,wd,q) /* c=imjn b=kolp */\
    SBUTTERFLY(a,c,d,dq,q) /* a=aeim d=bfjn */\
    SBUTTERFLY(t,b,c,dq,q) /* t=cgko c=dhlp */

#ifdef ARCH_X86_64
// permutes 01234567 -> 05736421
#define TRANSPOSE8(a,b,c,d,e,f,g,h,t)\
    SBUTTERFLY(a,b,%%xmm8,wd,dqa)\
    SBUTTERFLY(c,d,b,wd,dqa)\
    SBUTTERFLY(e,f,d,wd,dqa)\
    SBUTTERFLY(g,h,f,wd,dqa)\
    SBUTTERFLY(a,c,h,dq,dqa)\
    SBUTTERFLY(%%xmm8,b,c,dq,dqa)\
    SBUTTERFLY(e,g,b,dq,dqa)\
    SBUTTERFLY(d,f,g,dq,dqa)\
    SBUTTERFLY(a,e,f,qdq,dqa)\
    SBUTTERFLY(%%xmm8,d,e,qdq,dqa)\
    SBUTTERFLY(h,b,d,qdq,dqa)\
    SBUTTERFLY(c,g,b,qdq,dqa)\
    "movdqa %%xmm8, "#g"              \n\t"
#else
#define TRANSPOSE8(a,b,c,d,e,f,g,h,t)\
    "movdqa "#h", "#t"                \n\t"\
    SBUTTERFLY(a,b,h,wd,dqa)\
    "movdqa "#h", 16"#t"              \n\t"\
    "movdqa "#t", "#h"                \n\t"\
    SBUTTERFLY(c,d,b,wd,dqa)\
    SBUTTERFLY(e,f,d,wd,dqa)\
    SBUTTERFLY(g,h,f,wd,dqa)\
    SBUTTERFLY(a,c,h,dq,dqa)\
    "movdqa "#h", "#t"                \n\t"\
    "movdqa 16"#t", "#h"              \n\t"\
    SBUTTERFLY(h,b,c,dq,dqa)\
    SBUTTERFLY(e,g,b,dq,dqa)\
    SBUTTERFLY(d,f,g,dq,dqa)\
    SBUTTERFLY(a,e,f,qdq,dqa)\
    SBUTTERFLY(h,d,e,qdq,dqa)\
    "movdqa "#h", 16"#t"              \n\t"\
    "movdqa "#t", "#h"                \n\t"\
    SBUTTERFLY(h,b,d,qdq,dqa)\
    SBUTTERFLY(c,g,b,qdq,dqa)\
    "movdqa 16"#t", "#g"              \n\t"
#endif

#define MOVQ_WONE(regd) \
    asm volatile ( \
    "pcmpeqd %%" #regd ", %%" #regd " \n\t" \
    "psrlw $15, %%" #regd ::)

void dsputilenc_init_mmx(DSPContext* c, AVCodecContext *avctx);

#endif /* FFMPEG_DSPUTIL_MMX_H */