55387f623ebe45ed8a8402a0ae20bddb4e6880d4
[libav.git] / libavcodec / x86 / h264_intrapred_init.c
1 /*
2 * Copyright (c) 2010 Jason Garrett-Glaser
3 *
4 * This file is part of Libav.
5 *
6 * Libav is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
10 *
11 * Libav is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
15 *
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with Libav; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19 */
20
21 #include "libavutil/cpu.h"
22 #include "libavcodec/h264pred.h"
23
24 #define PRED4x4(TYPE, DEPTH, OPT) \
25 void ff_pred4x4_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, const uint8_t *topright, int stride);
26
27 PRED4x4(dc, 10, mmxext)
28 PRED4x4(down_left, 10, sse2)
29 PRED4x4(down_left, 10, avx)
30 PRED4x4(down_right, 10, sse2)
31 PRED4x4(down_right, 10, ssse3)
32 PRED4x4(down_right, 10, avx)
33 PRED4x4(vertical_left, 10, sse2)
34 PRED4x4(vertical_left, 10, avx)
35 PRED4x4(vertical_right, 10, sse2)
36 PRED4x4(vertical_right, 10, ssse3)
37 PRED4x4(vertical_right, 10, avx)
38 PRED4x4(horizontal_up, 10, mmxext)
39 PRED4x4(horizontal_down, 10, sse2)
40 PRED4x4(horizontal_down, 10, ssse3)
41 PRED4x4(horizontal_down, 10, avx)
42
43 #define PRED8x8(TYPE, DEPTH, OPT) \
44 void ff_pred8x8_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, int stride);
45
46 PRED8x8(dc, 10, mmxext)
47 PRED8x8(dc, 10, sse2)
48 PRED8x8(top_dc, 10, sse2)
49 PRED8x8(plane, 10, sse2)
50 PRED8x8(vertical, 10, sse2)
51 PRED8x8(horizontal, 10, sse2)
52
53 #define PRED8x8L(TYPE, DEPTH, OPT)\
54 void ff_pred8x8l_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, int has_topleft, int has_topright, int stride);
55
56 PRED8x8L(dc, 10, sse2)
57 PRED8x8L(dc, 10, avx)
58 PRED8x8L(128_dc, 10, mmxext)
59 PRED8x8L(128_dc, 10, sse2)
60 PRED8x8L(top_dc, 10, sse2)
61 PRED8x8L(top_dc, 10, avx)
62 PRED8x8L(vertical, 10, sse2)
63 PRED8x8L(vertical, 10, avx)
64 PRED8x8L(horizontal, 10, sse2)
65 PRED8x8L(horizontal, 10, ssse3)
66 PRED8x8L(horizontal, 10, avx)
67 PRED8x8L(down_left, 10, sse2)
68 PRED8x8L(down_left, 10, ssse3)
69 PRED8x8L(down_left, 10, avx)
70 PRED8x8L(down_right, 10, sse2)
71 PRED8x8L(down_right, 10, ssse3)
72 PRED8x8L(down_right, 10, avx)
73 PRED8x8L(vertical_right, 10, sse2)
74 PRED8x8L(vertical_right, 10, ssse3)
75 PRED8x8L(vertical_right, 10, avx)
76 PRED8x8L(horizontal_up, 10, sse2)
77 PRED8x8L(horizontal_up, 10, ssse3)
78 PRED8x8L(horizontal_up, 10, avx)
79
80 #define PRED16x16(TYPE, DEPTH, OPT)\
81 void ff_pred16x16_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, int stride);
82
83 PRED16x16(dc, 10, mmxext)
84 PRED16x16(dc, 10, sse2)
85 PRED16x16(top_dc, 10, mmxext)
86 PRED16x16(top_dc, 10, sse2)
87 PRED16x16(128_dc, 10, mmxext)
88 PRED16x16(128_dc, 10, sse2)
89 PRED16x16(left_dc, 10, mmxext)
90 PRED16x16(left_dc, 10, sse2)
91 PRED16x16(vertical, 10, mmxext)
92 PRED16x16(vertical, 10, sse2)
93 PRED16x16(horizontal, 10, mmxext)
94 PRED16x16(horizontal, 10, sse2)
95
96 void ff_pred16x16_vertical_mmx (uint8_t *src, int stride);
97 void ff_pred16x16_vertical_sse (uint8_t *src, int stride);
98 void ff_pred16x16_horizontal_mmx (uint8_t *src, int stride);
99 void ff_pred16x16_horizontal_mmxext(uint8_t *src, int stride);
100 void ff_pred16x16_horizontal_ssse3 (uint8_t *src, int stride);
101 void ff_pred16x16_dc_mmxext (uint8_t *src, int stride);
102 void ff_pred16x16_dc_sse2 (uint8_t *src, int stride);
103 void ff_pred16x16_dc_ssse3 (uint8_t *src, int stride);
104 void ff_pred16x16_plane_h264_mmx (uint8_t *src, int stride);
105 void ff_pred16x16_plane_h264_mmx2 (uint8_t *src, int stride);
106 void ff_pred16x16_plane_h264_sse2 (uint8_t *src, int stride);
107 void ff_pred16x16_plane_h264_ssse3 (uint8_t *src, int stride);
108 void ff_pred16x16_plane_rv40_mmx (uint8_t *src, int stride);
109 void ff_pred16x16_plane_rv40_mmx2 (uint8_t *src, int stride);
110 void ff_pred16x16_plane_rv40_sse2 (uint8_t *src, int stride);
111 void ff_pred16x16_plane_rv40_ssse3 (uint8_t *src, int stride);
112 void ff_pred16x16_plane_svq3_mmx (uint8_t *src, int stride);
113 void ff_pred16x16_plane_svq3_mmx2 (uint8_t *src, int stride);
114 void ff_pred16x16_plane_svq3_sse2 (uint8_t *src, int stride);
115 void ff_pred16x16_plane_svq3_ssse3 (uint8_t *src, int stride);
116 void ff_pred16x16_tm_vp8_mmx (uint8_t *src, int stride);
117 void ff_pred16x16_tm_vp8_mmxext (uint8_t *src, int stride);
118 void ff_pred16x16_tm_vp8_sse2 (uint8_t *src, int stride);
119 void ff_pred8x8_top_dc_mmxext (uint8_t *src, int stride);
120 void ff_pred8x8_dc_rv40_mmxext (uint8_t *src, int stride);
121 void ff_pred8x8_dc_mmxext (uint8_t *src, int stride);
122 void ff_pred8x8_vertical_mmx (uint8_t *src, int stride);
123 void ff_pred8x8_horizontal_mmx (uint8_t *src, int stride);
124 void ff_pred8x8_horizontal_mmxext (uint8_t *src, int stride);
125 void ff_pred8x8_horizontal_ssse3 (uint8_t *src, int stride);
126 void ff_pred8x8_plane_mmx (uint8_t *src, int stride);
127 void ff_pred8x8_plane_mmx2 (uint8_t *src, int stride);
128 void ff_pred8x8_plane_sse2 (uint8_t *src, int stride);
129 void ff_pred8x8_plane_ssse3 (uint8_t *src, int stride);
130 void ff_pred8x8_tm_vp8_mmx (uint8_t *src, int stride);
131 void ff_pred8x8_tm_vp8_mmxext (uint8_t *src, int stride);
132 void ff_pred8x8_tm_vp8_sse2 (uint8_t *src, int stride);
133 void ff_pred8x8_tm_vp8_ssse3 (uint8_t *src, int stride);
134 void ff_pred8x8l_top_dc_mmxext (uint8_t *src, int has_topleft, int has_topright, int stride);
135 void ff_pred8x8l_top_dc_ssse3 (uint8_t *src, int has_topleft, int has_topright, int stride);
136 void ff_pred8x8l_dc_mmxext (uint8_t *src, int has_topleft, int has_topright, int stride);
137 void ff_pred8x8l_dc_ssse3 (uint8_t *src, int has_topleft, int has_topright, int stride);
138 void ff_pred8x8l_horizontal_mmxext (uint8_t *src, int has_topleft, int has_topright, int stride);
139 void ff_pred8x8l_horizontal_ssse3 (uint8_t *src, int has_topleft, int has_topright, int stride);
140 void ff_pred8x8l_vertical_mmxext (uint8_t *src, int has_topleft, int has_topright, int stride);
141 void ff_pred8x8l_vertical_ssse3 (uint8_t *src, int has_topleft, int has_topright, int stride);
142 void ff_pred8x8l_down_left_mmxext (uint8_t *src, int has_topleft, int has_topright, int stride);
143 void ff_pred8x8l_down_left_sse2 (uint8_t *src, int has_topleft, int has_topright, int stride);
144 void ff_pred8x8l_down_left_ssse3 (uint8_t *src, int has_topleft, int has_topright, int stride);
145 void ff_pred8x8l_down_right_mmxext (uint8_t *src, int has_topleft, int has_topright, int stride);
146 void ff_pred8x8l_down_right_sse2 (uint8_t *src, int has_topleft, int has_topright, int stride);
147 void ff_pred8x8l_down_right_ssse3 (uint8_t *src, int has_topleft, int has_topright, int stride);
148 void ff_pred8x8l_vertical_right_mmxext(uint8_t *src, int has_topleft, int has_topright, int stride);
149 void ff_pred8x8l_vertical_right_sse2(uint8_t *src, int has_topleft, int has_topright, int stride);
150 void ff_pred8x8l_vertical_right_ssse3(uint8_t *src, int has_topleft, int has_topright, int stride);
151 void ff_pred8x8l_vertical_left_sse2(uint8_t *src, int has_topleft, int has_topright, int stride);
152 void ff_pred8x8l_vertical_left_ssse3(uint8_t *src, int has_topleft, int has_topright, int stride);
153 void ff_pred8x8l_horizontal_up_mmxext(uint8_t *src, int has_topleft, int has_topright, int stride);
154 void ff_pred8x8l_horizontal_up_ssse3(uint8_t *src, int has_topleft, int has_topright, int stride);
155 void ff_pred8x8l_horizontal_down_mmxext(uint8_t *src, int has_topleft, int has_topright, int stride);
156 void ff_pred8x8l_horizontal_down_sse2(uint8_t *src, int has_topleft, int has_topright, int stride);
157 void ff_pred8x8l_horizontal_down_ssse3(uint8_t *src, int has_topleft, int has_topright, int stride);
158 void ff_pred4x4_dc_mmxext (uint8_t *src, const uint8_t *topright, int stride);
159 void ff_pred4x4_down_left_mmxext (uint8_t *src, const uint8_t *topright, int stride);
160 void ff_pred4x4_down_right_mmxext (uint8_t *src, const uint8_t *topright, int stride);
161 void ff_pred4x4_vertical_left_mmxext(uint8_t *src, const uint8_t *topright, int stride);
162 void ff_pred4x4_vertical_right_mmxext(uint8_t *src, const uint8_t *topright, int stride);
163 void ff_pred4x4_horizontal_up_mmxext(uint8_t *src, const uint8_t *topright, int stride);
164 void ff_pred4x4_horizontal_down_mmxext(uint8_t *src, const uint8_t *topright, int stride);
165 void ff_pred4x4_tm_vp8_mmx (uint8_t *src, const uint8_t *topright, int stride);
166 void ff_pred4x4_tm_vp8_mmxext (uint8_t *src, const uint8_t *topright, int stride);
167 void ff_pred4x4_tm_vp8_ssse3 (uint8_t *src, const uint8_t *topright, int stride);
168 void ff_pred4x4_vertical_vp8_mmxext(uint8_t *src, const uint8_t *topright, int stride);
169
170 void ff_h264_pred_init_x86(H264PredContext *h, int codec_id, const int bit_depth)
171 {
172 int mm_flags = av_get_cpu_flags();
173
174 #if HAVE_YASM
175 if (bit_depth == 8) {
176 if (mm_flags & AV_CPU_FLAG_MMX) {
177 h->pred16x16[VERT_PRED8x8 ] = ff_pred16x16_vertical_mmx;
178 h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_mmx;
179 h->pred8x8 [VERT_PRED8x8 ] = ff_pred8x8_vertical_mmx;
180 h->pred8x8 [HOR_PRED8x8 ] = ff_pred8x8_horizontal_mmx;
181 if (codec_id == CODEC_ID_VP8) {
182 h->pred16x16[PLANE_PRED8x8 ] = ff_pred16x16_tm_vp8_mmx;
183 h->pred8x8 [PLANE_PRED8x8 ] = ff_pred8x8_tm_vp8_mmx;
184 h->pred4x4 [TM_VP8_PRED ] = ff_pred4x4_tm_vp8_mmx;
185 } else {
186 h->pred8x8 [PLANE_PRED8x8] = ff_pred8x8_plane_mmx;
187 if (codec_id == CODEC_ID_SVQ3) {
188 h->pred16x16[PLANE_PRED8x8] = ff_pred16x16_plane_svq3_mmx;
189 } else if (codec_id == CODEC_ID_RV40) {
190 h->pred16x16[PLANE_PRED8x8] = ff_pred16x16_plane_rv40_mmx;
191 } else {
192 h->pred16x16[PLANE_PRED8x8] = ff_pred16x16_plane_h264_mmx;
193 }
194 }
195 }
196
197 if (mm_flags & AV_CPU_FLAG_MMX2) {
198 h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_mmxext;
199 h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_mmxext;
200 h->pred8x8 [HOR_PRED8x8 ] = ff_pred8x8_horizontal_mmxext;
201 h->pred8x8l [TOP_DC_PRED ] = ff_pred8x8l_top_dc_mmxext;
202 h->pred8x8l [DC_PRED ] = ff_pred8x8l_dc_mmxext;
203 h->pred8x8l [HOR_PRED ] = ff_pred8x8l_horizontal_mmxext;
204 h->pred8x8l [VERT_PRED ] = ff_pred8x8l_vertical_mmxext;
205 h->pred8x8l [DIAG_DOWN_RIGHT_PRED ] = ff_pred8x8l_down_right_mmxext;
206 h->pred8x8l [VERT_RIGHT_PRED ] = ff_pred8x8l_vertical_right_mmxext;
207 h->pred8x8l [HOR_UP_PRED ] = ff_pred8x8l_horizontal_up_mmxext;
208 h->pred8x8l [DIAG_DOWN_LEFT_PRED ] = ff_pred8x8l_down_left_mmxext;
209 h->pred8x8l [HOR_DOWN_PRED ] = ff_pred8x8l_horizontal_down_mmxext;
210 h->pred4x4 [DIAG_DOWN_RIGHT_PRED ] = ff_pred4x4_down_right_mmxext;
211 h->pred4x4 [VERT_RIGHT_PRED ] = ff_pred4x4_vertical_right_mmxext;
212 h->pred4x4 [HOR_DOWN_PRED ] = ff_pred4x4_horizontal_down_mmxext;
213 h->pred4x4 [DC_PRED ] = ff_pred4x4_dc_mmxext;
214 if (codec_id == CODEC_ID_VP8 || codec_id == CODEC_ID_H264) {
215 h->pred4x4 [DIAG_DOWN_LEFT_PRED] = ff_pred4x4_down_left_mmxext;
216 }
217 if (codec_id == CODEC_ID_SVQ3 || codec_id == CODEC_ID_H264) {
218 h->pred4x4 [VERT_LEFT_PRED ] = ff_pred4x4_vertical_left_mmxext;
219 }
220 if (codec_id != CODEC_ID_RV40) {
221 h->pred4x4 [HOR_UP_PRED ] = ff_pred4x4_horizontal_up_mmxext;
222 }
223 if (codec_id == CODEC_ID_SVQ3 || codec_id == CODEC_ID_H264) {
224 h->pred8x8 [TOP_DC_PRED8x8 ] = ff_pred8x8_top_dc_mmxext;
225 h->pred8x8 [DC_PRED8x8 ] = ff_pred8x8_dc_mmxext;
226 }
227 if (codec_id == CODEC_ID_VP8) {
228 h->pred16x16[PLANE_PRED8x8 ] = ff_pred16x16_tm_vp8_mmxext;
229 h->pred8x8 [DC_PRED8x8 ] = ff_pred8x8_dc_rv40_mmxext;
230 h->pred8x8 [PLANE_PRED8x8 ] = ff_pred8x8_tm_vp8_mmxext;
231 h->pred4x4 [TM_VP8_PRED ] = ff_pred4x4_tm_vp8_mmxext;
232 h->pred4x4 [VERT_PRED ] = ff_pred4x4_vertical_vp8_mmxext;
233 } else {
234 h->pred8x8 [PLANE_PRED8x8] = ff_pred8x8_plane_mmx2;
235 if (codec_id == CODEC_ID_SVQ3) {
236 h->pred16x16[PLANE_PRED8x8 ] = ff_pred16x16_plane_svq3_mmx2;
237 } else if (codec_id == CODEC_ID_RV40) {
238 h->pred16x16[PLANE_PRED8x8 ] = ff_pred16x16_plane_rv40_mmx2;
239 } else {
240 h->pred16x16[PLANE_PRED8x8 ] = ff_pred16x16_plane_h264_mmx2;
241 }
242 }
243 }
244
245 if (mm_flags & AV_CPU_FLAG_SSE) {
246 h->pred16x16[VERT_PRED8x8] = ff_pred16x16_vertical_sse;
247 }
248
249 if (mm_flags & AV_CPU_FLAG_SSE2) {
250 h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_sse2;
251 h->pred8x8l [DIAG_DOWN_LEFT_PRED ] = ff_pred8x8l_down_left_sse2;
252 h->pred8x8l [DIAG_DOWN_RIGHT_PRED ] = ff_pred8x8l_down_right_sse2;
253 h->pred8x8l [VERT_RIGHT_PRED ] = ff_pred8x8l_vertical_right_sse2;
254 h->pred8x8l [VERT_LEFT_PRED ] = ff_pred8x8l_vertical_left_sse2;
255 h->pred8x8l [HOR_DOWN_PRED ] = ff_pred8x8l_horizontal_down_sse2;
256 if (codec_id == CODEC_ID_VP8) {
257 h->pred16x16[PLANE_PRED8x8 ] = ff_pred16x16_tm_vp8_sse2;
258 h->pred8x8 [PLANE_PRED8x8 ] = ff_pred8x8_tm_vp8_sse2;
259 } else {
260 h->pred8x8 [PLANE_PRED8x8 ] = ff_pred8x8_plane_sse2;
261 if (codec_id == CODEC_ID_SVQ3) {
262 h->pred16x16[PLANE_PRED8x8] = ff_pred16x16_plane_svq3_sse2;
263 } else if (codec_id == CODEC_ID_RV40) {
264 h->pred16x16[PLANE_PRED8x8] = ff_pred16x16_plane_rv40_sse2;
265 } else {
266 h->pred16x16[PLANE_PRED8x8] = ff_pred16x16_plane_h264_sse2;
267 }
268 }
269 }
270
271 if (mm_flags & AV_CPU_FLAG_SSSE3) {
272 h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_ssse3;
273 h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_ssse3;
274 h->pred8x8 [HOR_PRED8x8 ] = ff_pred8x8_horizontal_ssse3;
275 h->pred8x8l [TOP_DC_PRED ] = ff_pred8x8l_top_dc_ssse3;
276 h->pred8x8l [DC_PRED ] = ff_pred8x8l_dc_ssse3;
277 h->pred8x8l [HOR_PRED ] = ff_pred8x8l_horizontal_ssse3;
278 h->pred8x8l [VERT_PRED ] = ff_pred8x8l_vertical_ssse3;
279 h->pred8x8l [DIAG_DOWN_LEFT_PRED ] = ff_pred8x8l_down_left_ssse3;
280 h->pred8x8l [DIAG_DOWN_RIGHT_PRED ] = ff_pred8x8l_down_right_ssse3;
281 h->pred8x8l [VERT_RIGHT_PRED ] = ff_pred8x8l_vertical_right_ssse3;
282 h->pred8x8l [VERT_LEFT_PRED ] = ff_pred8x8l_vertical_left_ssse3;
283 h->pred8x8l [HOR_UP_PRED ] = ff_pred8x8l_horizontal_up_ssse3;
284 h->pred8x8l [HOR_DOWN_PRED ] = ff_pred8x8l_horizontal_down_ssse3;
285 if (codec_id == CODEC_ID_VP8) {
286 h->pred8x8 [PLANE_PRED8x8 ] = ff_pred8x8_tm_vp8_ssse3;
287 h->pred4x4 [TM_VP8_PRED ] = ff_pred4x4_tm_vp8_ssse3;
288 } else {
289 h->pred8x8 [PLANE_PRED8x8] = ff_pred8x8_plane_ssse3;
290 if (codec_id == CODEC_ID_SVQ3) {
291 h->pred16x16[PLANE_PRED8x8] = ff_pred16x16_plane_svq3_ssse3;
292 } else if (codec_id == CODEC_ID_RV40) {
293 h->pred16x16[PLANE_PRED8x8] = ff_pred16x16_plane_rv40_ssse3;
294 } else {
295 h->pred16x16[PLANE_PRED8x8] = ff_pred16x16_plane_h264_ssse3;
296 }
297 }
298 }
299 } else if (bit_depth == 10) {
300 if (mm_flags & AV_CPU_FLAG_MMX2) {
301 h->pred4x4[DC_PRED ] = ff_pred4x4_dc_10_mmxext;
302 h->pred4x4[HOR_UP_PRED ] = ff_pred4x4_horizontal_up_10_mmxext;
303
304 h->pred8x8[DC_PRED8x8 ] = ff_pred8x8_dc_10_mmxext;
305
306 h->pred8x8l[DC_128_PRED ] = ff_pred8x8l_128_dc_10_mmxext;
307
308 h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_10_mmxext;
309 h->pred16x16[TOP_DC_PRED8x8 ] = ff_pred16x16_top_dc_10_mmxext;
310 h->pred16x16[DC_128_PRED8x8 ] = ff_pred16x16_128_dc_10_mmxext;
311 h->pred16x16[LEFT_DC_PRED8x8 ] = ff_pred16x16_left_dc_10_mmxext;
312 h->pred16x16[VERT_PRED8x8 ] = ff_pred16x16_vertical_10_mmxext;
313 h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_10_mmxext;
314 }
315 if (mm_flags & AV_CPU_FLAG_SSE2) {
316 h->pred4x4[DIAG_DOWN_LEFT_PRED ] = ff_pred4x4_down_left_10_sse2;
317 h->pred4x4[DIAG_DOWN_RIGHT_PRED] = ff_pred4x4_down_right_10_sse2;
318 h->pred4x4[VERT_LEFT_PRED ] = ff_pred4x4_vertical_left_10_sse2;
319 h->pred4x4[VERT_RIGHT_PRED ] = ff_pred4x4_vertical_right_10_sse2;
320 h->pred4x4[HOR_DOWN_PRED ] = ff_pred4x4_horizontal_down_10_sse2;
321
322 h->pred8x8[DC_PRED8x8 ] = ff_pred8x8_dc_10_sse2;
323 h->pred8x8[TOP_DC_PRED8x8 ] = ff_pred8x8_top_dc_10_sse2;
324 h->pred8x8[PLANE_PRED8x8 ] = ff_pred8x8_plane_10_sse2;
325 h->pred8x8[VERT_PRED8x8 ] = ff_pred8x8_vertical_10_sse2;
326 h->pred8x8[HOR_PRED8x8 ] = ff_pred8x8_horizontal_10_sse2;
327
328 h->pred8x8l[VERT_PRED ] = ff_pred8x8l_vertical_10_sse2;
329 h->pred8x8l[HOR_PRED ] = ff_pred8x8l_horizontal_10_sse2;
330 h->pred8x8l[DC_PRED ] = ff_pred8x8l_dc_10_sse2;
331 h->pred8x8l[DC_128_PRED ] = ff_pred8x8l_128_dc_10_sse2;
332 h->pred8x8l[TOP_DC_PRED ] = ff_pred8x8l_top_dc_10_sse2;
333 h->pred8x8l[DIAG_DOWN_LEFT_PRED ] = ff_pred8x8l_down_left_10_sse2;
334 h->pred8x8l[DIAG_DOWN_RIGHT_PRED] = ff_pred8x8l_down_right_10_sse2;
335 h->pred8x8l[VERT_RIGHT_PRED ] = ff_pred8x8l_vertical_right_10_sse2;
336 h->pred8x8l[HOR_UP_PRED ] = ff_pred8x8l_horizontal_up_10_sse2;
337
338 h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_10_sse2;
339 h->pred16x16[TOP_DC_PRED8x8 ] = ff_pred16x16_top_dc_10_sse2;
340 h->pred16x16[DC_128_PRED8x8 ] = ff_pred16x16_128_dc_10_sse2;
341 h->pred16x16[LEFT_DC_PRED8x8 ] = ff_pred16x16_left_dc_10_sse2;
342 h->pred16x16[VERT_PRED8x8 ] = ff_pred16x16_vertical_10_sse2;
343 h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_10_sse2;
344 }
345 if (mm_flags & AV_CPU_FLAG_SSSE3) {
346 h->pred4x4[DIAG_DOWN_RIGHT_PRED] = ff_pred4x4_down_right_10_ssse3;
347 h->pred4x4[VERT_RIGHT_PRED ] = ff_pred4x4_vertical_right_10_ssse3;
348 h->pred4x4[HOR_DOWN_PRED ] = ff_pred4x4_horizontal_down_10_ssse3;
349
350 h->pred8x8l[HOR_PRED ] = ff_pred8x8l_horizontal_10_ssse3;
351 h->pred8x8l[DIAG_DOWN_LEFT_PRED ] = ff_pred8x8l_down_left_10_ssse3;
352 h->pred8x8l[DIAG_DOWN_RIGHT_PRED] = ff_pred8x8l_down_right_10_ssse3;
353 h->pred8x8l[VERT_RIGHT_PRED ] = ff_pred8x8l_vertical_right_10_ssse3;
354 h->pred8x8l[HOR_UP_PRED ] = ff_pred8x8l_horizontal_up_10_ssse3;
355 }
356 #if HAVE_AVX
357 if (mm_flags & AV_CPU_FLAG_AVX) {
358 h->pred4x4[DIAG_DOWN_LEFT_PRED ] = ff_pred4x4_down_left_10_avx;
359 h->pred4x4[DIAG_DOWN_RIGHT_PRED] = ff_pred4x4_down_right_10_avx;
360 h->pred4x4[VERT_LEFT_PRED ] = ff_pred4x4_vertical_left_10_avx;
361 h->pred4x4[VERT_RIGHT_PRED ] = ff_pred4x4_vertical_right_10_avx;
362 h->pred4x4[HOR_DOWN_PRED ] = ff_pred4x4_horizontal_down_10_avx;
363
364 h->pred8x8l[VERT_PRED ] = ff_pred8x8l_vertical_10_avx;
365 h->pred8x8l[HOR_PRED ] = ff_pred8x8l_horizontal_10_avx;
366 h->pred8x8l[DC_PRED ] = ff_pred8x8l_dc_10_avx;
367 h->pred8x8l[TOP_DC_PRED ] = ff_pred8x8l_top_dc_10_avx;
368 h->pred8x8l[DIAG_DOWN_RIGHT_PRED] = ff_pred8x8l_down_right_10_avx;
369 h->pred8x8l[DIAG_DOWN_LEFT_PRED ] = ff_pred8x8l_down_left_10_avx;
370 h->pred8x8l[VERT_RIGHT_PRED ] = ff_pred8x8l_vertical_right_10_avx;
371 h->pred8x8l[HOR_UP_PRED ] = ff_pred8x8l_horizontal_up_10_avx;
372 }
373 #endif /* HAVE_AVX */
374 }
375 #endif /* HAVE_YASM */
376 }