Mercurial > libavcodec.hg
comparison x86/h264dsp_mmx.c @ 8519:cc64e1343397 libavcodec
Use H264 MMX chroma functions to accelerate RV40 decoding.
Patch by Mathieu Velten (matmaul A gmail)
| author | cehoyos |
|---|---|
| date | Sun, 04 Jan 2009 01:36:11 +0000 |
| parents | cea216e44ee3 |
| children | 7a463923ecd1 |
comparison
equal
deleted
inserted
replaced
| 8518:f2c406b05158 | 8519:cc64e1343397 |
|---|---|
| 2082 #ifdef HAVE_SSSE3 | 2082 #ifdef HAVE_SSSE3 |
| 2083 H264_MC_816(H264_MC_H, ssse3) | 2083 H264_MC_816(H264_MC_H, ssse3) |
| 2084 H264_MC_816(H264_MC_HV, ssse3) | 2084 H264_MC_816(H264_MC_HV, ssse3) |
| 2085 #endif | 2085 #endif |
| 2086 | 2086 |
| 2087 /* rnd interleaved with rnd div 8, use p+1 to access rnd div 8 */ | |
| 2088 DECLARE_ALIGNED_8(static const uint64_t, h264_rnd_reg[4]) = { | |
| 2089 0x0020002000200020ULL, 0x0004000400040004ULL, 0x001C001C001C001CULL, 0x0003000300030003ULL | |
| 2090 }; | |
| 2087 | 2091 |
| 2088 #define H264_CHROMA_OP(S,D) | 2092 #define H264_CHROMA_OP(S,D) |
| 2089 #define H264_CHROMA_OP4(S,D,T) | 2093 #define H264_CHROMA_OP4(S,D,T) |
| 2090 #define H264_CHROMA_MC8_TMPL put_h264_chroma_mc8_mmx | 2094 #define H264_CHROMA_MC8_TMPL put_h264_chroma_generic_mc8_mmx |
| 2091 #define H264_CHROMA_MC4_TMPL put_h264_chroma_mc4_mmx | 2095 #define H264_CHROMA_MC4_TMPL put_h264_chroma_generic_mc4_mmx |
| 2092 #define H264_CHROMA_MC2_TMPL put_h264_chroma_mc2_mmx2 | 2096 #define H264_CHROMA_MC2_TMPL put_h264_chroma_mc2_mmx2 |
| 2093 #define H264_CHROMA_MC8_MV0 put_pixels8_mmx | 2097 #define H264_CHROMA_MC8_MV0 put_pixels8_mmx |
| 2094 #include "dsputil_h264_template_mmx.c" | 2098 #include "dsputil_h264_template_mmx.c" |
| 2095 | 2099 |
| 2096 static void put_h264_chroma_mc8_mmx_rnd(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y) | 2100 static void put_h264_chroma_mc8_mmx_rnd(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y) |
| 2097 { | 2101 { |
| 2098 put_h264_chroma_mc8_mmx(dst, src, stride, h, x, y, 1); | 2102 put_h264_chroma_generic_mc8_mmx(dst, src, stride, h, x, y, h264_rnd_reg); |
| 2099 } | 2103 } |
| 2100 static void put_h264_chroma_mc8_mmx_nornd(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y) | 2104 static void put_h264_chroma_mc8_mmx_nornd(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y) |
| 2101 { | 2105 { |
| 2102 put_h264_chroma_mc8_mmx(dst, src, stride, h, x, y, 0); | 2106 put_h264_chroma_generic_mc8_mmx(dst, src, stride, h, x, y, h264_rnd_reg+2); |
| 2107 } | |
| 2108 static void put_h264_chroma_mc4_mmx(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y) | |
| 2109 { | |
| 2110 put_h264_chroma_generic_mc4_mmx(dst, src, stride, h, x, y, h264_rnd_reg); | |
| 2103 } | 2111 } |
| 2104 | 2112 |
| 2105 #undef H264_CHROMA_OP | 2113 #undef H264_CHROMA_OP |
| 2106 #undef H264_CHROMA_OP4 | 2114 #undef H264_CHROMA_OP4 |
| 2107 #undef H264_CHROMA_MC8_TMPL | 2115 #undef H264_CHROMA_MC8_TMPL |
| 2110 #undef H264_CHROMA_MC8_MV0 | 2118 #undef H264_CHROMA_MC8_MV0 |
| 2111 | 2119 |
| 2112 #define H264_CHROMA_OP(S,D) "pavgb " #S ", " #D " \n\t" | 2120 #define H264_CHROMA_OP(S,D) "pavgb " #S ", " #D " \n\t" |
| 2113 #define H264_CHROMA_OP4(S,D,T) "movd " #S ", " #T " \n\t"\ | 2121 #define H264_CHROMA_OP4(S,D,T) "movd " #S ", " #T " \n\t"\ |
| 2114 "pavgb " #T ", " #D " \n\t" | 2122 "pavgb " #T ", " #D " \n\t" |
| 2115 #define H264_CHROMA_MC8_TMPL avg_h264_chroma_mc8_mmx2 | 2123 #define H264_CHROMA_MC8_TMPL avg_h264_chroma_generic_mc8_mmx2 |
| 2116 #define H264_CHROMA_MC4_TMPL avg_h264_chroma_mc4_mmx2 | 2124 #define H264_CHROMA_MC4_TMPL avg_h264_chroma_generic_mc4_mmx2 |
| 2117 #define H264_CHROMA_MC2_TMPL avg_h264_chroma_mc2_mmx2 | 2125 #define H264_CHROMA_MC2_TMPL avg_h264_chroma_mc2_mmx2 |
| 2118 #define H264_CHROMA_MC8_MV0 avg_pixels8_mmx2 | 2126 #define H264_CHROMA_MC8_MV0 avg_pixels8_mmx2 |
| 2119 #include "dsputil_h264_template_mmx.c" | 2127 #include "dsputil_h264_template_mmx.c" |
| 2120 static void avg_h264_chroma_mc8_mmx2_rnd(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y) | 2128 static void avg_h264_chroma_mc8_mmx2_rnd(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y) |
| 2121 { | 2129 { |
| 2122 avg_h264_chroma_mc8_mmx2(dst, src, stride, h, x, y, 1); | 2130 avg_h264_chroma_generic_mc8_mmx2(dst, src, stride, h, x, y, h264_rnd_reg); |
| 2131 } | |
| 2132 static void avg_h264_chroma_mc4_mmx2(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y) | |
| 2133 { | |
| 2134 avg_h264_chroma_generic_mc4_mmx2(dst, src, stride, h, x, y, h264_rnd_reg); | |
| 2123 } | 2135 } |
| 2124 #undef H264_CHROMA_OP | 2136 #undef H264_CHROMA_OP |
| 2125 #undef H264_CHROMA_OP4 | 2137 #undef H264_CHROMA_OP4 |
| 2126 #undef H264_CHROMA_MC8_TMPL | 2138 #undef H264_CHROMA_MC8_TMPL |
| 2127 #undef H264_CHROMA_MC4_TMPL | 2139 #undef H264_CHROMA_MC4_TMPL |
| 2129 #undef H264_CHROMA_MC8_MV0 | 2141 #undef H264_CHROMA_MC8_MV0 |
| 2130 | 2142 |
| 2131 #define H264_CHROMA_OP(S,D) "pavgusb " #S ", " #D " \n\t" | 2143 #define H264_CHROMA_OP(S,D) "pavgusb " #S ", " #D " \n\t" |
| 2132 #define H264_CHROMA_OP4(S,D,T) "movd " #S ", " #T " \n\t"\ | 2144 #define H264_CHROMA_OP4(S,D,T) "movd " #S ", " #T " \n\t"\ |
| 2133 "pavgusb " #T ", " #D " \n\t" | 2145 "pavgusb " #T ", " #D " \n\t" |
| 2134 #define H264_CHROMA_MC8_TMPL avg_h264_chroma_mc8_3dnow | 2146 #define H264_CHROMA_MC8_TMPL avg_h264_chroma_generic_mc8_3dnow |
| 2135 #define H264_CHROMA_MC4_TMPL avg_h264_chroma_mc4_3dnow | 2147 #define H264_CHROMA_MC4_TMPL avg_h264_chroma_generic_mc4_3dnow |
| 2136 #define H264_CHROMA_MC8_MV0 avg_pixels8_3dnow | 2148 #define H264_CHROMA_MC8_MV0 avg_pixels8_3dnow |
| 2137 #include "dsputil_h264_template_mmx.c" | 2149 #include "dsputil_h264_template_mmx.c" |
| 2138 static void avg_h264_chroma_mc8_3dnow_rnd(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y) | 2150 static void avg_h264_chroma_mc8_3dnow_rnd(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y) |
| 2139 { | 2151 { |
| 2140 avg_h264_chroma_mc8_3dnow(dst, src, stride, h, x, y, 1); | 2152 avg_h264_chroma_generic_mc8_3dnow(dst, src, stride, h, x, y, h264_rnd_reg); |
| 2153 } | |
| 2154 static void avg_h264_chroma_mc4_3dnow(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y) | |
| 2155 { | |
| 2156 avg_h264_chroma_generic_mc4_3dnow(dst, src, stride, h, x, y, h264_rnd_reg); | |
| 2141 } | 2157 } |
| 2142 #undef H264_CHROMA_OP | 2158 #undef H264_CHROMA_OP |
| 2143 #undef H264_CHROMA_OP4 | 2159 #undef H264_CHROMA_OP4 |
| 2144 #undef H264_CHROMA_MC8_TMPL | 2160 #undef H264_CHROMA_MC8_TMPL |
| 2145 #undef H264_CHROMA_MC4_TMPL | 2161 #undef H264_CHROMA_MC4_TMPL |