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
|
/*
* SIMD-optimized IDCT-related routines
* Copyright (c) 2000, 2001 Fabrice Bellard
* Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
*
* MMX optimization by Nick Kurshev <nickols_k@mail.ru>
*
* 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
*/
#include "config.h"
#include "libavutil/cpu.h"
#include "libavutil/x86/asm.h"
#include "idctdsp.h"
#include "inline_asm.h"
#if HAVE_INLINE_ASM
void ff_put_pixels_clamped_mmx(const int16_t *block, uint8_t *pixels,
int line_size)
{
const int16_t *p;
uint8_t *pix;
/* read the pixels */
p = block;
pix = pixels;
/* unrolled loop */
__asm__ volatile (
"movq (%3), %%mm0 \n\t"
"movq 8(%3), %%mm1 \n\t"
"movq 16(%3), %%mm2 \n\t"
"movq 24(%3), %%mm3 \n\t"
"movq 32(%3), %%mm4 \n\t"
"movq 40(%3), %%mm5 \n\t"
"movq 48(%3), %%mm6 \n\t"
"movq 56(%3), %%mm7 \n\t"
"packuswb %%mm1, %%mm0 \n\t"
"packuswb %%mm3, %%mm2 \n\t"
"packuswb %%mm5, %%mm4 \n\t"
"packuswb %%mm7, %%mm6 \n\t"
"movq %%mm0, (%0) \n\t"
"movq %%mm2, (%0, %1) \n\t"
"movq %%mm4, (%0, %1, 2) \n\t"
"movq %%mm6, (%0, %2) \n\t"
:: "r" (pix), "r" ((x86_reg) line_size), "r" ((x86_reg) line_size * 3),
"r" (p)
: "memory");
pix += line_size * 4;
p += 32;
// if here would be an exact copy of the code above
// compiler would generate some very strange code
// thus using "r"
__asm__ volatile (
"movq (%3), %%mm0 \n\t"
"movq 8(%3), %%mm1 \n\t"
"movq 16(%3), %%mm2 \n\t"
"movq 24(%3), %%mm3 \n\t"
"movq 32(%3), %%mm4 \n\t"
"movq 40(%3), %%mm5 \n\t"
"movq 48(%3), %%mm6 \n\t"
"movq 56(%3), %%mm7 \n\t"
"packuswb %%mm1, %%mm0 \n\t"
"packuswb %%mm3, %%mm2 \n\t"
"packuswb %%mm5, %%mm4 \n\t"
"packuswb %%mm7, %%mm6 \n\t"
"movq %%mm0, (%0) \n\t"
"movq %%mm2, (%0, %1) \n\t"
"movq %%mm4, (%0, %1, 2) \n\t"
"movq %%mm6, (%0, %2) \n\t"
:: "r" (pix), "r" ((x86_reg) line_size), "r" ((x86_reg) line_size * 3),
"r" (p)
: "memory");
}
void ff_add_pixels_clamped_mmx(const int16_t *block, uint8_t *pixels,
int line_size)
{
const int16_t *p;
uint8_t *pix;
int i;
/* read the pixels */
p = block;
pix = pixels;
MOVQ_ZERO(mm7);
i = 4;
do {
__asm__ volatile (
"movq (%2), %%mm0 \n\t"
"movq 8(%2), %%mm1 \n\t"
"movq 16(%2), %%mm2 \n\t"
"movq 24(%2), %%mm3 \n\t"
"movq %0, %%mm4 \n\t"
"movq %1, %%mm6 \n\t"
"movq %%mm4, %%mm5 \n\t"
"punpcklbw %%mm7, %%mm4 \n\t"
"punpckhbw %%mm7, %%mm5 \n\t"
"paddsw %%mm4, %%mm0 \n\t"
"paddsw %%mm5, %%mm1 \n\t"
"movq %%mm6, %%mm5 \n\t"
"punpcklbw %%mm7, %%mm6 \n\t"
"punpckhbw %%mm7, %%mm5 \n\t"
"paddsw %%mm6, %%mm2 \n\t"
"paddsw %%mm5, %%mm3 \n\t"
"packuswb %%mm1, %%mm0 \n\t"
"packuswb %%mm3, %%mm2 \n\t"
"movq %%mm0, %0 \n\t"
"movq %%mm2, %1 \n\t"
: "+m" (*pix), "+m" (*(pix + line_size))
: "r" (p)
: "memory");
pix += line_size * 2;
p += 16;
} while (--i);
}
#endif /* HAVE_INLINE_ASM */
|
