2 * MMX optimized DSP utils
3 * Copyright (c) 2007 Aurelien Jacobs <aurel@gnuage.org>
5 * This file is part of FFmpeg.
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
22 #ifndef AVCODEC_X86_DSPUTIL_MMX_H
23 #define AVCODEC_X86_DSPUTIL_MMX_H
26 #include "libavcodec/dsputil.h"
27 #include "libavutil/x86/asm.h"
29 typedef struct xmm_reg { uint64_t a, b; } xmm_reg;
31 extern const uint64_t ff_bone;
32 extern const uint64_t ff_wtwo;
34 extern const xmm_reg ff_pw_3;
35 extern const xmm_reg ff_pw_4;
36 extern const xmm_reg ff_pw_5;
37 extern const xmm_reg ff_pw_8;
38 extern const uint64_t ff_pw_15;
39 extern const xmm_reg ff_pw_16;
40 extern const xmm_reg ff_pw_18;
41 extern const uint64_t ff_pw_20;
42 extern const xmm_reg ff_pw_27;
43 extern const xmm_reg ff_pw_28;
44 extern const xmm_reg ff_pw_32;
45 extern const uint64_t ff_pw_42;
46 extern const uint64_t ff_pw_53;
47 extern const xmm_reg ff_pw_63;
48 extern const xmm_reg ff_pw_64;
49 extern const uint64_t ff_pw_96;
50 extern const uint64_t ff_pw_128;
51 extern const uint64_t ff_pw_255;
53 extern const xmm_reg ff_pb_1;
54 extern const xmm_reg ff_pb_3;
55 extern const uint64_t ff_pb_7;
56 extern const uint64_t ff_pb_1F;
57 extern const uint64_t ff_pb_3F;
58 extern const uint64_t ff_pb_81;
59 extern const xmm_reg ff_pb_A1;
60 extern const xmm_reg ff_pb_F8;
61 extern const uint64_t ff_pb_FC;
62 extern const xmm_reg ff_pb_FE;
64 extern const double ff_pd_1[2];
65 extern const double ff_pd_2[2];
67 #define SBUTTERFLY(a,b,t,n,m)\
68 "mov" #m " " #a ", " #t " \n\t" /* abcd */\
69 "punpckl" #n " " #b ", " #a " \n\t" /* aebf */\
70 "punpckh" #n " " #b ", " #t " \n\t" /* cgdh */\
72 #define TRANSPOSE4(a,b,c,d,t)\
73 SBUTTERFLY(a,b,t,wd,q) /* a=aebf t=cgdh */\
74 SBUTTERFLY(c,d,b,wd,q) /* c=imjn b=kolp */\
75 SBUTTERFLY(a,c,d,dq,q) /* a=aeim d=bfjn */\
76 SBUTTERFLY(t,b,c,dq,q) /* t=cgko c=dhlp */
78 #define MOVQ_WONE(regd) \
80 "pcmpeqd %%" #regd ", %%" #regd " \n\t" \
81 "psrlw $15, %%" #regd ::)
83 void ff_dsputilenc_init_mmx(DSPContext* c, AVCodecContext *avctx);
84 void ff_dsputil_init_pix_mmx(DSPContext* c, AVCodecContext *avctx);
86 void ff_add_pixels_clamped_mmx(const int16_t *block, uint8_t *pixels, int line_size);
87 void ff_put_pixels_clamped_mmx(const int16_t *block, uint8_t *pixels, int line_size);
88 void ff_put_signed_pixels_clamped_mmx(const int16_t *block, uint8_t *pixels, int line_size);
90 void ff_put_cavs_qpel8_mc00_mmxext(uint8_t *dst, uint8_t *src, int stride);
91 void ff_avg_cavs_qpel8_mc00_mmxext(uint8_t *dst, uint8_t *src, int stride);
92 void ff_put_cavs_qpel16_mc00_mmxext(uint8_t *dst, uint8_t *src, int stride);
93 void ff_avg_cavs_qpel16_mc00_mmxext(uint8_t *dst, uint8_t *src, int stride);
95 void ff_put_vc1_mspel_mc00_mmx(uint8_t *dst, const uint8_t *src, int stride, int rnd);
96 void ff_avg_vc1_mspel_mc00_mmxext(uint8_t *dst, const uint8_t *src, int stride, int rnd);
98 void ff_put_rv40_qpel8_mc33_mmx(uint8_t *block, uint8_t *pixels, int line_size);
99 void ff_put_rv40_qpel16_mc33_mmx(uint8_t *block, uint8_t *pixels, int line_size);
100 void ff_avg_rv40_qpel8_mc33_mmx(uint8_t *block, uint8_t *pixels, int line_size);
101 void ff_avg_rv40_qpel16_mc33_mmx(uint8_t *block, uint8_t *pixels, int line_size);
103 void ff_mmx_idct(int16_t *block);
104 void ff_mmxext_idct(int16_t *block);
107 void ff_deinterlace_line_mmx(uint8_t *dst,
108 const uint8_t *lum_m4, const uint8_t *lum_m3,
109 const uint8_t *lum_m2, const uint8_t *lum_m1,
113 void ff_deinterlace_line_inplace_mmx(const uint8_t *lum_m4,
114 const uint8_t *lum_m3,
115 const uint8_t *lum_m2,
116 const uint8_t *lum_m1,
117 const uint8_t *lum, int size);
119 #endif /* AVCODEC_X86_DSPUTIL_MMX_H */