vp8dsp-init.c
Go to the documentation of this file.
1 /*
2  * VP8 DSP functions x86-optimized
3  * Copyright (c) 2010 Ronald S. Bultje <rsbultje@gmail.com>
4  * Copyright (c) 2010 Jason Garrett-Glaser <darkshikari@gmail.com>
5  *
6  * This file is part of Libav.
7  *
8  * Libav is free software; you can redistribute it and/or
9  * modify it under the terms of the GNU Lesser General Public
10  * License as published by the Free Software Foundation; either
11  * version 2.1 of the License, or (at your option) any later version.
12  *
13  * Libav is distributed in the hope that it will be useful,
14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16  * Lesser General Public License for more details.
17  *
18  * You should have received a copy of the GNU Lesser General Public
19  * License along with Libav; if not, write to the Free Software
20  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21  */
22 
23 #include "libavutil/cpu.h"
24 #include "libavutil/x86_cpu.h"
25 #include "libavcodec/vp8dsp.h"
26 
27 #if HAVE_YASM
28 
29 /*
30  * MC functions
31  */
32 extern void ff_put_vp8_epel4_h4_mmxext(uint8_t *dst, int dststride,
33  uint8_t *src, int srcstride,
34  int height, int mx, int my);
35 extern void ff_put_vp8_epel4_h6_mmxext(uint8_t *dst, int dststride,
36  uint8_t *src, int srcstride,
37  int height, int mx, int my);
38 extern void ff_put_vp8_epel4_v4_mmxext(uint8_t *dst, int dststride,
39  uint8_t *src, int srcstride,
40  int height, int mx, int my);
41 extern void ff_put_vp8_epel4_v6_mmxext(uint8_t *dst, int dststride,
42  uint8_t *src, int srcstride,
43  int height, int mx, int my);
44 
45 extern void ff_put_vp8_epel8_h4_sse2 (uint8_t *dst, int dststride,
46  uint8_t *src, int srcstride,
47  int height, int mx, int my);
48 extern void ff_put_vp8_epel8_h6_sse2 (uint8_t *dst, int dststride,
49  uint8_t *src, int srcstride,
50  int height, int mx, int my);
51 extern void ff_put_vp8_epel8_v4_sse2 (uint8_t *dst, int dststride,
52  uint8_t *src, int srcstride,
53  int height, int mx, int my);
54 extern void ff_put_vp8_epel8_v6_sse2 (uint8_t *dst, int dststride,
55  uint8_t *src, int srcstride,
56  int height, int mx, int my);
57 
58 extern void ff_put_vp8_epel4_h4_ssse3 (uint8_t *dst, int dststride,
59  uint8_t *src, int srcstride,
60  int height, int mx, int my);
61 extern void ff_put_vp8_epel4_h6_ssse3 (uint8_t *dst, int dststride,
62  uint8_t *src, int srcstride,
63  int height, int mx, int my);
64 extern void ff_put_vp8_epel4_v4_ssse3 (uint8_t *dst, int dststride,
65  uint8_t *src, int srcstride,
66  int height, int mx, int my);
67 extern void ff_put_vp8_epel4_v6_ssse3 (uint8_t *dst, int dststride,
68  uint8_t *src, int srcstride,
69  int height, int mx, int my);
70 extern void ff_put_vp8_epel8_h4_ssse3 (uint8_t *dst, int dststride,
71  uint8_t *src, int srcstride,
72  int height, int mx, int my);
73 extern void ff_put_vp8_epel8_h6_ssse3 (uint8_t *dst, int dststride,
74  uint8_t *src, int srcstride,
75  int height, int mx, int my);
76 extern void ff_put_vp8_epel8_v4_ssse3 (uint8_t *dst, int dststride,
77  uint8_t *src, int srcstride,
78  int height, int mx, int my);
79 extern void ff_put_vp8_epel8_v6_ssse3 (uint8_t *dst, int dststride,
80  uint8_t *src, int srcstride,
81  int height, int mx, int my);
82 
83 extern void ff_put_vp8_bilinear4_h_mmxext(uint8_t *dst, int dststride,
84  uint8_t *src, int srcstride,
85  int height, int mx, int my);
86 extern void ff_put_vp8_bilinear8_h_sse2 (uint8_t *dst, int dststride,
87  uint8_t *src, int srcstride,
88  int height, int mx, int my);
89 extern void ff_put_vp8_bilinear4_h_ssse3 (uint8_t *dst, int dststride,
90  uint8_t *src, int srcstride,
91  int height, int mx, int my);
92 extern void ff_put_vp8_bilinear8_h_ssse3 (uint8_t *dst, int dststride,
93  uint8_t *src, int srcstride,
94  int height, int mx, int my);
95 
96 extern void ff_put_vp8_bilinear4_v_mmxext(uint8_t *dst, int dststride,
97  uint8_t *src, int srcstride,
98  int height, int mx, int my);
99 extern void ff_put_vp8_bilinear8_v_sse2 (uint8_t *dst, int dststride,
100  uint8_t *src, int srcstride,
101  int height, int mx, int my);
102 extern void ff_put_vp8_bilinear4_v_ssse3 (uint8_t *dst, int dststride,
103  uint8_t *src, int srcstride,
104  int height, int mx, int my);
105 extern void ff_put_vp8_bilinear8_v_ssse3 (uint8_t *dst, int dststride,
106  uint8_t *src, int srcstride,
107  int height, int mx, int my);
108 
109 
110 extern void ff_put_vp8_pixels8_mmx (uint8_t *dst, int dststride,
111  uint8_t *src, int srcstride,
112  int height, int mx, int my);
113 extern void ff_put_vp8_pixels16_mmx(uint8_t *dst, int dststride,
114  uint8_t *src, int srcstride,
115  int height, int mx, int my);
116 extern void ff_put_vp8_pixels16_sse(uint8_t *dst, int dststride,
117  uint8_t *src, int srcstride,
118  int height, int mx, int my);
119 
120 #define TAP_W16(OPT, FILTERTYPE, TAPTYPE) \
121 static void ff_put_vp8_ ## FILTERTYPE ## 16_ ## TAPTYPE ## _ ## OPT( \
122  uint8_t *dst, int dststride, uint8_t *src, \
123  int srcstride, int height, int mx, int my) \
124 { \
125  ff_put_vp8_ ## FILTERTYPE ## 8_ ## TAPTYPE ## _ ## OPT( \
126  dst, dststride, src, srcstride, height, mx, my); \
127  ff_put_vp8_ ## FILTERTYPE ## 8_ ## TAPTYPE ## _ ## OPT( \
128  dst + 8, dststride, src + 8, srcstride, height, mx, my); \
129 }
130 #define TAP_W8(OPT, FILTERTYPE, TAPTYPE) \
131 static void ff_put_vp8_ ## FILTERTYPE ## 8_ ## TAPTYPE ## _ ## OPT( \
132  uint8_t *dst, int dststride, uint8_t *src, \
133  int srcstride, int height, int mx, int my) \
134 { \
135  ff_put_vp8_ ## FILTERTYPE ## 4_ ## TAPTYPE ## _ ## OPT( \
136  dst, dststride, src, srcstride, height, mx, my); \
137  ff_put_vp8_ ## FILTERTYPE ## 4_ ## TAPTYPE ## _ ## OPT( \
138  dst + 4, dststride, src + 4, srcstride, height, mx, my); \
139 }
140 
141 TAP_W8 (mmxext, epel, h4)
142 TAP_W8 (mmxext, epel, h6)
143 TAP_W16(mmxext, epel, h6)
144 TAP_W8 (mmxext, epel, v4)
145 TAP_W8 (mmxext, epel, v6)
146 TAP_W16(mmxext, epel, v6)
147 TAP_W8 (mmxext, bilinear, h)
148 TAP_W16(mmxext, bilinear, h)
149 TAP_W8 (mmxext, bilinear, v)
150 TAP_W16(mmxext, bilinear, v)
151 
152 TAP_W16(sse2, epel, h6)
153 TAP_W16(sse2, epel, v6)
154 TAP_W16(sse2, bilinear, h)
155 TAP_W16(sse2, bilinear, v)
156 
157 TAP_W16(ssse3, epel, h6)
158 TAP_W16(ssse3, epel, v6)
159 TAP_W16(ssse3, bilinear, h)
160 TAP_W16(ssse3, bilinear, v)
161 
162 #define HVTAP(OPT, ALIGN, TAPNUMX, TAPNUMY, SIZE, MAXHEIGHT) \
163 static void ff_put_vp8_epel ## SIZE ## _h ## TAPNUMX ## v ## TAPNUMY ## _ ## OPT( \
164  uint8_t *dst, int dststride, uint8_t *src, \
165  int srcstride, int height, int mx, int my) \
166 { \
167  DECLARE_ALIGNED(ALIGN, uint8_t, tmp)[SIZE * (MAXHEIGHT + TAPNUMY - 1)]; \
168  uint8_t *tmpptr = tmp + SIZE * (TAPNUMY / 2 - 1); \
169  src -= srcstride * (TAPNUMY / 2 - 1); \
170  ff_put_vp8_epel ## SIZE ## _h ## TAPNUMX ## _ ## OPT( \
171  tmp, SIZE, src, srcstride, height + TAPNUMY - 1, mx, my); \
172  ff_put_vp8_epel ## SIZE ## _v ## TAPNUMY ## _ ## OPT( \
173  dst, dststride, tmpptr, SIZE, height, mx, my); \
174 }
175 
176 #define HVTAPMMX(x, y) \
177 HVTAP(mmxext, 8, x, y, 4, 8) \
178 HVTAP(mmxext, 8, x, y, 8, 16)
179 
180 HVTAPMMX(4, 4)
181 HVTAPMMX(4, 6)
182 HVTAPMMX(6, 4)
183 HVTAPMMX(6, 6)
184 HVTAP(mmxext, 8, 6, 6, 16, 16)
185 
186 #define HVTAPSSE2(x, y, w) \
187 HVTAP(sse2, 16, x, y, w, 16) \
188 HVTAP(ssse3, 16, x, y, w, 16)
189 
190 HVTAPSSE2(4, 4, 8)
191 HVTAPSSE2(4, 6, 8)
192 HVTAPSSE2(6, 4, 8)
193 HVTAPSSE2(6, 6, 8)
194 HVTAPSSE2(6, 6, 16)
195 
196 HVTAP(ssse3, 16, 4, 4, 4, 8)
197 HVTAP(ssse3, 16, 4, 6, 4, 8)
198 HVTAP(ssse3, 16, 6, 4, 4, 8)
199 HVTAP(ssse3, 16, 6, 6, 4, 8)
200 
201 #define HVBILIN(OPT, ALIGN, SIZE, MAXHEIGHT) \
202 static void ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT( \
203  uint8_t *dst, int dststride, uint8_t *src, \
204  int srcstride, int height, int mx, int my) \
205 { \
206  DECLARE_ALIGNED(ALIGN, uint8_t, tmp)[SIZE * (MAXHEIGHT + 2)]; \
207  ff_put_vp8_bilinear ## SIZE ## _h_ ## OPT( \
208  tmp, SIZE, src, srcstride, height + 1, mx, my); \
209  ff_put_vp8_bilinear ## SIZE ## _v_ ## OPT( \
210  dst, dststride, tmp, SIZE, height, mx, my); \
211 }
212 
213 HVBILIN(mmxext, 8, 4, 8)
214 HVBILIN(mmxext, 8, 8, 16)
215 HVBILIN(mmxext, 8, 16, 16)
216 HVBILIN(sse2, 8, 8, 16)
217 HVBILIN(sse2, 8, 16, 16)
218 HVBILIN(ssse3, 8, 4, 8)
219 HVBILIN(ssse3, 8, 8, 16)
220 HVBILIN(ssse3, 8, 16, 16)
221 
222 extern void ff_vp8_idct_dc_add_mmx(uint8_t *dst, DCTELEM block[16], int stride);
223 extern void ff_vp8_idct_dc_add_sse4(uint8_t *dst, DCTELEM block[16], int stride);
224 extern void ff_vp8_idct_dc_add4y_mmx(uint8_t *dst, DCTELEM block[4][16], int stride);
225 extern void ff_vp8_idct_dc_add4y_sse2(uint8_t *dst, DCTELEM block[4][16], int stride);
226 extern void ff_vp8_idct_dc_add4uv_mmx(uint8_t *dst, DCTELEM block[2][16], int stride);
227 extern void ff_vp8_luma_dc_wht_mmx(DCTELEM block[4][4][16], DCTELEM dc[16]);
228 extern void ff_vp8_luma_dc_wht_sse(DCTELEM block[4][4][16], DCTELEM dc[16]);
229 extern void ff_vp8_idct_add_mmx(uint8_t *dst, DCTELEM block[16], int stride);
230 extern void ff_vp8_idct_add_sse(uint8_t *dst, DCTELEM block[16], int stride);
231 
232 #define DECLARE_LOOP_FILTER(NAME)\
233 extern void ff_vp8_v_loop_filter_simple_ ## NAME(uint8_t *dst, int stride, int flim);\
234 extern void ff_vp8_h_loop_filter_simple_ ## NAME(uint8_t *dst, int stride, int flim);\
235 extern void ff_vp8_v_loop_filter16y_inner_ ## NAME (uint8_t *dst, int stride,\
236  int e, int i, int hvt);\
237 extern void ff_vp8_h_loop_filter16y_inner_ ## NAME (uint8_t *dst, int stride,\
238  int e, int i, int hvt);\
239 extern void ff_vp8_v_loop_filter8uv_inner_ ## NAME (uint8_t *dstU, uint8_t *dstV,\
240  int s, int e, int i, int hvt);\
241 extern void ff_vp8_h_loop_filter8uv_inner_ ## NAME (uint8_t *dstU, uint8_t *dstV,\
242  int s, int e, int i, int hvt);\
243 extern void ff_vp8_v_loop_filter16y_mbedge_ ## NAME(uint8_t *dst, int stride,\
244  int e, int i, int hvt);\
245 extern void ff_vp8_h_loop_filter16y_mbedge_ ## NAME(uint8_t *dst, int stride,\
246  int e, int i, int hvt);\
247 extern void ff_vp8_v_loop_filter8uv_mbedge_ ## NAME(uint8_t *dstU, uint8_t *dstV,\
248  int s, int e, int i, int hvt);\
249 extern void ff_vp8_h_loop_filter8uv_mbedge_ ## NAME(uint8_t *dstU, uint8_t *dstV,\
250  int s, int e, int i, int hvt);
251 
252 DECLARE_LOOP_FILTER(mmx)
253 DECLARE_LOOP_FILTER(mmxext)
254 DECLARE_LOOP_FILTER(sse2)
255 DECLARE_LOOP_FILTER(ssse3)
256 DECLARE_LOOP_FILTER(sse4)
257 
258 #endif
259 
260 #define VP8_LUMA_MC_FUNC(IDX, SIZE, OPT) \
261  c->put_vp8_epel_pixels_tab[IDX][0][2] = ff_put_vp8_epel ## SIZE ## _h6_ ## OPT; \
262  c->put_vp8_epel_pixels_tab[IDX][2][0] = ff_put_vp8_epel ## SIZE ## _v6_ ## OPT; \
263  c->put_vp8_epel_pixels_tab[IDX][2][2] = ff_put_vp8_epel ## SIZE ## _h6v6_ ## OPT
264 
265 #define VP8_MC_FUNC(IDX, SIZE, OPT) \
266  c->put_vp8_epel_pixels_tab[IDX][0][1] = ff_put_vp8_epel ## SIZE ## _h4_ ## OPT; \
267  c->put_vp8_epel_pixels_tab[IDX][1][0] = ff_put_vp8_epel ## SIZE ## _v4_ ## OPT; \
268  c->put_vp8_epel_pixels_tab[IDX][1][1] = ff_put_vp8_epel ## SIZE ## _h4v4_ ## OPT; \
269  c->put_vp8_epel_pixels_tab[IDX][1][2] = ff_put_vp8_epel ## SIZE ## _h6v4_ ## OPT; \
270  c->put_vp8_epel_pixels_tab[IDX][2][1] = ff_put_vp8_epel ## SIZE ## _h4v6_ ## OPT; \
271  VP8_LUMA_MC_FUNC(IDX, SIZE, OPT)
272 
273 #define VP8_BILINEAR_MC_FUNC(IDX, SIZE, OPT) \
274  c->put_vp8_bilinear_pixels_tab[IDX][0][1] = ff_put_vp8_bilinear ## SIZE ## _h_ ## OPT; \
275  c->put_vp8_bilinear_pixels_tab[IDX][0][2] = ff_put_vp8_bilinear ## SIZE ## _h_ ## OPT; \
276  c->put_vp8_bilinear_pixels_tab[IDX][1][0] = ff_put_vp8_bilinear ## SIZE ## _v_ ## OPT; \
277  c->put_vp8_bilinear_pixels_tab[IDX][1][1] = ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT; \
278  c->put_vp8_bilinear_pixels_tab[IDX][1][2] = ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT; \
279  c->put_vp8_bilinear_pixels_tab[IDX][2][0] = ff_put_vp8_bilinear ## SIZE ## _v_ ## OPT; \
280  c->put_vp8_bilinear_pixels_tab[IDX][2][1] = ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT; \
281  c->put_vp8_bilinear_pixels_tab[IDX][2][2] = ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT
282 
283 
285 {
286 #if HAVE_YASM
287  int mm_flags = av_get_cpu_flags();
288 
289  if (mm_flags & AV_CPU_FLAG_MMX) {
290  c->vp8_idct_dc_add = ff_vp8_idct_dc_add_mmx;
291  c->vp8_idct_dc_add4y = ff_vp8_idct_dc_add4y_mmx;
292  c->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_mmx;
293  c->vp8_idct_add = ff_vp8_idct_add_mmx;
294  c->vp8_luma_dc_wht = ff_vp8_luma_dc_wht_mmx;
295  c->put_vp8_epel_pixels_tab[0][0][0] =
296  c->put_vp8_bilinear_pixels_tab[0][0][0] = ff_put_vp8_pixels16_mmx;
297  c->put_vp8_epel_pixels_tab[1][0][0] =
298  c->put_vp8_bilinear_pixels_tab[1][0][0] = ff_put_vp8_pixels8_mmx;
299 
300  c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_mmx;
301  c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_mmx;
302 
303  c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_mmx;
304  c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmx;
305  c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_mmx;
306  c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_mmx;
307 
308  c->vp8_v_loop_filter16y = ff_vp8_v_loop_filter16y_mbedge_mmx;
309  c->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16y_mbedge_mmx;
310  c->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_mbedge_mmx;
311  c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_mmx;
312  }
313 
314  /* note that 4-tap width=16 functions are missing because w=16
315  * is only used for luma, and luma is always a copy or sixtap. */
316  if (mm_flags & AV_CPU_FLAG_MMX2) {
317  VP8_LUMA_MC_FUNC(0, 16, mmxext);
318  VP8_MC_FUNC(1, 8, mmxext);
319  VP8_MC_FUNC(2, 4, mmxext);
320  VP8_BILINEAR_MC_FUNC(0, 16, mmxext);
321  VP8_BILINEAR_MC_FUNC(1, 8, mmxext);
322  VP8_BILINEAR_MC_FUNC(2, 4, mmxext);
323 
324  c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_mmxext;
325  c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_mmxext;
326 
327  c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_mmxext;
328  c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmxext;
329  c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_mmxext;
330  c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_mmxext;
331 
332  c->vp8_v_loop_filter16y = ff_vp8_v_loop_filter16y_mbedge_mmxext;
333  c->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16y_mbedge_mmxext;
334  c->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_mbedge_mmxext;
335  c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_mmxext;
336  }
337 
338  if (mm_flags & AV_CPU_FLAG_SSE) {
339  c->vp8_idct_add = ff_vp8_idct_add_sse;
340  c->vp8_luma_dc_wht = ff_vp8_luma_dc_wht_sse;
341  c->put_vp8_epel_pixels_tab[0][0][0] =
342  c->put_vp8_bilinear_pixels_tab[0][0][0] = ff_put_vp8_pixels16_sse;
343  }
344 
345  if (mm_flags & (AV_CPU_FLAG_SSE2|AV_CPU_FLAG_SSE2SLOW)) {
346  VP8_LUMA_MC_FUNC(0, 16, sse2);
347  VP8_MC_FUNC(1, 8, sse2);
348  VP8_BILINEAR_MC_FUNC(0, 16, sse2);
349  VP8_BILINEAR_MC_FUNC(1, 8, sse2);
350 
351  c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_sse2;
352 
353  c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_sse2;
354  c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_sse2;
355 
356  c->vp8_v_loop_filter16y = ff_vp8_v_loop_filter16y_mbedge_sse2;
357  c->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_mbedge_sse2;
358  }
359 
360  if (mm_flags & AV_CPU_FLAG_SSE2) {
361  c->vp8_idct_dc_add4y = ff_vp8_idct_dc_add4y_sse2;
362 
363  c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_sse2;
364 
365  c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_sse2;
366  c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_sse2;
367 
368  c->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16y_mbedge_sse2;
369  c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_sse2;
370  }
371 
372  if (mm_flags & AV_CPU_FLAG_SSSE3) {
373  VP8_LUMA_MC_FUNC(0, 16, ssse3);
374  VP8_MC_FUNC(1, 8, ssse3);
375  VP8_MC_FUNC(2, 4, ssse3);
376  VP8_BILINEAR_MC_FUNC(0, 16, ssse3);
377  VP8_BILINEAR_MC_FUNC(1, 8, ssse3);
378  VP8_BILINEAR_MC_FUNC(2, 4, ssse3);
379 
380  c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_ssse3;
381  c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_ssse3;
382 
383  c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_ssse3;
384  c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_ssse3;
385  c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_ssse3;
386  c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_ssse3;
387 
388  c->vp8_v_loop_filter16y = ff_vp8_v_loop_filter16y_mbedge_ssse3;
389  c->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16y_mbedge_ssse3;
390  c->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_mbedge_ssse3;
391  c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_ssse3;
392  }
393 
394  if (mm_flags & AV_CPU_FLAG_SSE4) {
395  c->vp8_idct_dc_add = ff_vp8_idct_dc_add_sse4;
396 
397  c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_sse4;
398  c->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16y_mbedge_sse4;
399  c->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_mbedge_sse4;
400  }
401 #endif
402 }