FFmpeg coverage

Directory: ../../../ffmpeg/
File: src/libavcodec/x86/h264_qpel.c
Date: 2026-09-13 04:46:26
Exec Total Coverage
Lines: 79 79 100.0%
Functions: 146 146 100.0%
Branches: 20 20 100.0%
Line Branch Exec Source
1 /*
2 * Copyright (c) 2004-2005 Michael Niedermayer, Loren Merritt
3 * Copyright (c) 2011 Daniel Kang
4 *
5 * This file is part of FFmpeg.
6 *
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.
11 *
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.
16 *
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
20 */
21
22 #include <stddef.h>
23
24 #include "libavutil/attributes.h"
25 #include "libavutil/avassert.h"
26 #include "libavutil/cpu.h"
27 #include "libavutil/mem_internal.h"
28 #include "libavutil/x86/cpu.h"
29 #include "libavcodec/h264qpel.h"
30 #include "fpel.h"
31 #include "qpel.h"
32
33 void ff_avg_pixels4_mmxext(uint8_t *dst, const uint8_t *src, ptrdiff_t stride);
34 void ff_put_pixels4x4_l2_mmxext(uint8_t *dst, const uint8_t *src1, const uint8_t *src2,
35 ptrdiff_t stride);
36 void ff_avg_pixels4x4_l2_mmxext(uint8_t *dst, const uint8_t *src1, const uint8_t *src2,
37 ptrdiff_t stride);
38 #define ff_put_pixels4x4_l2_mmxext(dst, src1, src2, dststride, src1stride) \
39 ff_put_pixels4x4_l2_mmxext((dst), (src1), (src2), (dststride))
40 #define ff_avg_pixels4x4_l2_mmxext(dst, src1, src2, dststride, src1stride) \
41 ff_avg_pixels4x4_l2_mmxext((dst), (src1), (src2), (dststride))
42
43 #define DEF_QPEL(OPNAME)\
44 void ff_ ## OPNAME ## _h264_qpel4_h_lowpass_mmxext(uint8_t *dst, const uint8_t *src, ptrdiff_t dstStride, ptrdiff_t srcStride);\
45 void ff_ ## OPNAME ## _h264_qpel8_h_lowpass_ssse3(uint8_t *dst, const uint8_t *src, ptrdiff_t dstStride, ptrdiff_t srcStride);\
46 void ff_ ## OPNAME ## _h264_qpel4_h_lowpass_l2_mmxext(uint8_t *dst, const uint8_t *src, const uint8_t *src2, ptrdiff_t dstStride, ptrdiff_t src2Stride);\
47 void ff_ ## OPNAME ## _h264_qpel8_h_lowpass_l2_sse2(uint8_t *dst, const uint8_t *src, const uint8_t *src2, ptrdiff_t dstStride, ptrdiff_t src2Stride);\
48 void ff_ ## OPNAME ## _h264_qpel16_h_lowpass_l2_sse2(uint8_t *dst, const uint8_t *src, const uint8_t *src2, ptrdiff_t dstStride, ptrdiff_t src2Stride);\
49 void ff_ ## OPNAME ## _h264_qpel8_h_lowpass_l2_ssse3(uint8_t *dst, const uint8_t *src, const uint8_t *src2, ptrdiff_t dstStride, ptrdiff_t src2Stride);\
50 void ff_ ## OPNAME ## _h264_qpel4_v_lowpass_mmxext(uint8_t *dst, const uint8_t *src, ptrdiff_t dstStride, ptrdiff_t srcStride);\
51 void ff_ ## OPNAME ## _h264_qpel8or16_v_lowpass_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t dstStride, ptrdiff_t srcStride, int h);\
52 void ff_ ## OPNAME ## _h264_qpel4_hv_lowpass_h_mmxext(int16_t *tmp, uint8_t *dst, ptrdiff_t dstStride);\
53 void ff_ ## OPNAME ## _h264_qpel8or16_hv1_lowpass_op_sse2(const uint8_t *src, int16_t *tmp, ptrdiff_t srcStride, int size);\
54 void ff_ ## OPNAME ## _h264_qpel8_hv2_lowpass_sse2(uint8_t *dst, int16_t *tmp, ptrdiff_t dstStride);\
55 void ff_ ## OPNAME ## _h264_qpel16_hv2_lowpass_sse2(uint8_t *dst, int16_t *tmp, ptrdiff_t dstStride);\
56 void ff_ ## OPNAME ## _h264_qpel8_hv2_lowpass_ssse3(uint8_t *dst, int16_t *tmp, ptrdiff_t dstStride);\
57 void ff_ ## OPNAME ## _h264_qpel16_hv2_lowpass_ssse3(uint8_t *dst, int16_t *tmp, ptrdiff_t dstStride);\
58 void ff_ ## OPNAME ## _pixels4_l2_shift5_mmxext(uint8_t *dst, const int16_t *src16, const uint8_t *src8, ptrdiff_t dstStride);\
59 void ff_ ## OPNAME ## _pixels8_l2_shift5_sse2(uint8_t *dst, const int16_t *src16, const uint8_t *src8, ptrdiff_t dstStride);\
60 void ff_ ## OPNAME ## _pixels16_l2_shift5_sse2(uint8_t *dst, const int16_t *src16, const uint8_t *src8, ptrdiff_t dstStride);\
61
62 void ff_put_h264_qpel4_hv_lowpass_v_mmxext(const uint8_t *src, int16_t *tmp, ptrdiff_t srcStride);
63
64 DEF_QPEL(avg)
65 DEF_QPEL(put)
66
67 #define QPEL_H264(OPNAME, MMX)\
68 static av_always_inline void OPNAME ## h264_qpel4_hv_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, const uint8_t *src, ptrdiff_t dstStride, ptrdiff_t srcStride)\
69 {\
70 src -= 2*srcStride+2;\
71 ff_put_h264_qpel4_hv_lowpass_v_mmxext(src, tmp, srcStride);\
72 ff_ ## OPNAME ## h264_qpel4_hv_lowpass_h_mmxext(tmp, dst, dstStride);\
73 }\
74
75 #define QPEL_H264_H16(OPNAME, EXT) \
76 static av_always_inline void ff_ ## OPNAME ## h264_qpel16_h_lowpass_l2_ ## EXT(uint8_t *dst, const uint8_t *src, const uint8_t *src2, ptrdiff_t dstStride, ptrdiff_t src2Stride)\
77 {\
78 ff_ ## OPNAME ## h264_qpel8_h_lowpass_l2_ ## EXT(dst , src , src2 , dstStride, src2Stride);\
79 ff_ ## OPNAME ## h264_qpel8_h_lowpass_l2_ ## EXT(dst+8, src+8, src2+8, dstStride, src2Stride);\
80 src += 8*dstStride;\
81 dst += 8*dstStride;\
82 src2 += 8*src2Stride;\
83 ff_ ## OPNAME ## h264_qpel8_h_lowpass_l2_ ## EXT(dst , src , src2 , dstStride, src2Stride);\
84 ff_ ## OPNAME ## h264_qpel8_h_lowpass_l2_ ## EXT(dst+8, src+8, src2+8, dstStride, src2Stride);\
85 }\
86
87
88 #if ARCH_X86_64
89 #define QPEL_H264_H16_XMM(OPNAME, MMX)\
90
91 void ff_avg_h264_qpel16_h_lowpass_l2_ssse3(uint8_t *dst, const uint8_t *src, const uint8_t *src2, ptrdiff_t dstStride, ptrdiff_t src2Stride);
92 void ff_put_h264_qpel16_h_lowpass_l2_ssse3(uint8_t *dst, const uint8_t *src, const uint8_t *src2, ptrdiff_t dstStride, ptrdiff_t src2Stride);
93
94 #else // ARCH_X86_64
95 #define QPEL_H264_H16_XMM(OPNAME, EXT) QPEL_H264_H16(OPNAME, EXT)
96 #endif // ARCH_X86_64
97
98 #define QPEL_H264_H_XMM(OPNAME, MMX)\
99 QPEL_H264_H16_XMM(OPNAME, MMX)\
100 static av_always_inline void ff_ ## OPNAME ## h264_qpel16_h_lowpass_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t dstStride, ptrdiff_t srcStride)\
101 {\
102 ff_ ## OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst , src , dstStride, srcStride);\
103 ff_ ## OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst+8, src+8, dstStride, srcStride);\
104 src += 8*srcStride;\
105 dst += 8*dstStride;\
106 ff_ ## OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst , src , dstStride, srcStride);\
107 ff_ ## OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst+8, src+8, dstStride, srcStride);\
108 }\
109
110 #define QPEL_H264_V_XMM(OPNAME, XMM, XMM2)\
111 static av_always_inline void ff_ ## OPNAME ## h264_qpel8_v_lowpass_ ## XMM(uint8_t *dst, const uint8_t *src, ptrdiff_t dstStride, ptrdiff_t srcStride)\
112 {\
113 ff_ ## OPNAME ## h264_qpel8or16_v_lowpass_ ## XMM2(dst , src , dstStride, srcStride, 8);\
114 }\
115 static av_always_inline void ff_ ## OPNAME ## h264_qpel16_v_lowpass_ ## XMM(uint8_t *dst, const uint8_t *src, ptrdiff_t dstStride, ptrdiff_t srcStride)\
116 {\
117 ff_ ## OPNAME ## h264_qpel8or16_v_lowpass_ ## XMM2(dst , src , dstStride, srcStride, 16);\
118 ff_ ## OPNAME ## h264_qpel8or16_v_lowpass_ ## XMM2(dst+8, src+8, dstStride, srcStride, 16);\
119 }
120
121 161798 static av_always_inline void put_h264_qpel8or16_hv1_lowpass_sse2(int16_t *tmp,
122 const uint8_t *src,
123 ptrdiff_t srcStride,
124 int size)
125 {
126 161798 int w = (size+8)>>3;
127 161798 src -= 2*srcStride+2;
128
2/2
✓ Branch 0 taken 390381 times.
✓ Branch 1 taken 161798 times.
 552179 while(w--){
129 390381 ff_put_h264_qpel8or16_hv1_lowpass_op_sse2(src, tmp, srcStride, size);
130 390381 tmp += 8;
131 390381 src += 8;
132 }
133 161798 }
134
135 #define QPEL_H264_HV_XMM(OPNAME, MMX)\
136 static av_always_inline void OPNAME ## h264_qpel8_hv_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, const uint8_t *src, ptrdiff_t dstStride, ptrdiff_t srcStride)\
137 {\
138 put_h264_qpel8or16_hv1_lowpass_sse2(tmp, src, srcStride, 8);\
139 ff_ ## OPNAME ## h264_qpel8_hv2_lowpass_ ## MMX(dst, tmp, dstStride);\
140 }\
141 static av_always_inline void OPNAME ## h264_qpel16_hv_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, const uint8_t *src, ptrdiff_t dstStride, ptrdiff_t srcStride)\
142 {\
143 put_h264_qpel8or16_hv1_lowpass_sse2(tmp, src, srcStride, 16);\
144 ff_ ## OPNAME ## h264_qpel16_hv2_lowpass_ ## MMX(dst, tmp, dstStride);\
145 }\
146
147 #define H264_MC_V_H_HV(OPNAME, SIZE, MMX, ALIGN, SHIFT5_EXT) \
148 H264_MC_V(OPNAME, SIZE, MMX, ALIGN, SHIFT5_EXT)\
149 H264_MC_H(OPNAME, SIZE, MMX, ALIGN, SHIFT5_EXT)\
150 H264_MC_HV(OPNAME, SIZE, MMX, ALIGN, SHIFT5_EXT)\
151
152 #define H264_MC_H(OPNAME, SIZE, MMX, ALIGN, UNUSED) \
153 static void OPNAME ## h264_qpel ## SIZE ## _mc10_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
154 {\
155 ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src, src, stride, stride);\
156 }\
157 \
158 static void OPNAME ## h264_qpel ## SIZE ## _mc20_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
159 {\
160 ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_ ## MMX(dst, src, stride, stride);\
161 }\
162 \
163 static void OPNAME ## h264_qpel ## SIZE ## _mc30_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
164 {\
165 ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src, src+1, stride, stride);\
166 }\
167
168 #define H264_MC_V(OPNAME, SIZE, MMX, ALIGN, UNUSED) \
169 static void OPNAME ## h264_qpel ## SIZE ## _mc01_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
170 {\
171 LOCAL_ALIGNED(ALIGN, uint8_t, temp, [SIZE*SIZE]);\
172 ff_put_h264_qpel ## SIZE ## _v_lowpass_ ## MMX(temp, src, SIZE, stride);\
173 ff_ ## OPNAME ## pixels ## SIZE ## x ## SIZE ## _l2_ ## MMX(dst, src, temp, stride, stride);\
174 }\
175 \
176 static void OPNAME ## h264_qpel ## SIZE ## _mc02_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
177 {\
178 ff_ ## OPNAME ## h264_qpel ## SIZE ## _v_lowpass_ ## MMX(dst, src, stride, stride);\
179 }\
180 \
181 static void OPNAME ## h264_qpel ## SIZE ## _mc03_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
182 {\
183 LOCAL_ALIGNED(ALIGN, uint8_t, temp, [SIZE*SIZE]);\
184 ff_put_h264_qpel ## SIZE ## _v_lowpass_ ## MMX(temp, src, SIZE, stride);\
185 ff_ ## OPNAME ## pixels ## SIZE ## x ## SIZE ## _l2_ ## MMX(dst, src+stride, temp, stride, stride);\
186 }\
187
188 #define H264_MC_HV(OPNAME, SIZE, MMX, ALIGN, SHIFT5_EXT) \
189 static void OPNAME ## h264_qpel ## SIZE ## _mc11_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
190 {\
191 LOCAL_ALIGNED(ALIGN, uint8_t, temp, [SIZE*SIZE]);\
192 ff_put_h264_qpel ## SIZE ## _v_lowpass_ ## MMX(temp, src, SIZE, stride);\
193 ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src, temp, stride, SIZE);\
194 }\
195 \
196 static void OPNAME ## h264_qpel ## SIZE ## _mc31_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
197 {\
198 LOCAL_ALIGNED(ALIGN, uint8_t, temp, [SIZE*SIZE]);\
199 ff_put_h264_qpel ## SIZE ## _v_lowpass_ ## MMX(temp, src+1, SIZE, stride);\
200 ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src, temp, stride, SIZE);\
201 }\
202 \
203 static void OPNAME ## h264_qpel ## SIZE ## _mc13_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
204 {\
205 LOCAL_ALIGNED(ALIGN, uint8_t, temp, [SIZE*SIZE]);\
206 ff_put_h264_qpel ## SIZE ## _v_lowpass_ ## MMX(temp, src, SIZE, stride);\
207 ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src+stride, temp, stride, SIZE);\
208 }\
209 \
210 static void OPNAME ## h264_qpel ## SIZE ## _mc33_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
211 {\
212 LOCAL_ALIGNED(ALIGN, uint8_t, temp, [SIZE*SIZE]);\
213 ff_put_h264_qpel ## SIZE ## _v_lowpass_ ## MMX(temp, src+1, SIZE, stride);\
214 ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src+stride, temp, stride, SIZE);\
215 }\
216 \
217 static void OPNAME ## h264_qpel ## SIZE ## _mc22_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
218 {\
219 LOCAL_ALIGNED(ALIGN, uint16_t, temp, [SIZE*(SIZE<8?12:24)]);\
220 OPNAME ## h264_qpel ## SIZE ## _hv_lowpass_ ## MMX(dst, temp, src, stride, stride);\
221 }\
222 \
223 static void OPNAME ## h264_qpel ## SIZE ## _mc21_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
224 {\
225 LOCAL_ALIGNED(ALIGN, uint8_t, temp, [SIZE*(SIZE<8?12:24)*2 + SIZE*SIZE]);\
226 uint8_t * const halfHV= temp;\
227 int16_t * const halfV= (int16_t*)(temp + SIZE*SIZE);\
228 av_assert2(((uintptr_t)temp & 7) == 0);\
229 put_h264_qpel ## SIZE ## _hv_lowpass_ ## MMX(halfHV, halfV, src, SIZE, stride);\
230 ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src, halfHV, stride, SIZE);\
231 }\
232 \
233 static void OPNAME ## h264_qpel ## SIZE ## _mc23_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
234 {\
235 LOCAL_ALIGNED(ALIGN, uint8_t, temp, [SIZE*(SIZE<8?12:24)*2 + SIZE*SIZE]);\
236 uint8_t * const halfHV= temp;\
237 int16_t * const halfV= (int16_t*)(temp + SIZE*SIZE);\
238 av_assert2(((uintptr_t)temp & 7) == 0);\
239 put_h264_qpel ## SIZE ## _hv_lowpass_ ## MMX(halfHV, halfV, src, SIZE, stride);\
240 ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src+stride, halfHV, stride, SIZE);\
241 }\
242 \
243 static void OPNAME ## h264_qpel ## SIZE ## _mc12_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
244 {\
245 LOCAL_ALIGNED(ALIGN, uint8_t, temp, [SIZE*(SIZE<8?12:24)*2 + SIZE*SIZE]);\
246 uint8_t * const halfHV= temp;\
247 int16_t * const halfV= (int16_t*)(temp + SIZE*SIZE);\
248 av_assert2(((uintptr_t)temp & 7) == 0);\
249 put_h264_qpel ## SIZE ## _hv_lowpass_ ## MMX(halfHV, halfV, src, SIZE, stride);\
250 ff_ ## OPNAME ## pixels ## SIZE ## _l2_shift5_ ## SHIFT5_EXT(dst, halfV+2, halfHV, stride);\
251 }\
252 \
253 static void OPNAME ## h264_qpel ## SIZE ## _mc32_ ## MMX(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
254 {\
255 LOCAL_ALIGNED(ALIGN, uint8_t, temp, [SIZE*(SIZE<8?12:24)*2 + SIZE*SIZE]);\
256 uint8_t * const halfHV= temp;\
257 int16_t * const halfV= (int16_t*)(temp + SIZE*SIZE);\
258 av_assert2(((uintptr_t)temp & 7) == 0);\
259 put_h264_qpel ## SIZE ## _hv_lowpass_ ## MMX(halfHV, halfV, src, SIZE, stride);\
260 ff_ ## OPNAME ## pixels ## SIZE ## _l2_shift5_ ## SHIFT5_EXT(dst, halfV+3, halfHV, stride);\
261 }\
262
263 #define H264_MC(QPEL, SIZE, MMX, ALIGN, SHIFT5_EXT)\
264 QPEL(put_, SIZE, MMX, ALIGN, SHIFT5_EXT) \
265 QPEL(avg_, SIZE, MMX, ALIGN, SHIFT5_EXT) \
266
267 #define H264_MC_816(QPEL, XMM, SHIFT5_EXT)\
268 QPEL(put_, 8, XMM, 16, SHIFT5_EXT)\
269 QPEL(put_, 16,XMM, 16, SHIFT5_EXT)\
270 QPEL(avg_, 8, XMM, 16, SHIFT5_EXT)\
271 QPEL(avg_, 16,XMM, 16, SHIFT5_EXT)\
272
273 346 QPEL_H264(put_, mmxext)
274 2 QPEL_H264(avg_, mmxext)
275 179130 QPEL_H264_V_XMM(put_, sse2, sse2)
276 3760 QPEL_H264_V_XMM(avg_, sse2, sse2)
277 108 QPEL_H264_HV_XMM(put_, sse2)
278 12 QPEL_H264_HV_XMM(avg_, sse2)
279 14055 QPEL_H264_H_XMM(put_, ssse3)
280 3352 QPEL_H264_H_XMM(avg_, ssse3)
281 322024 QPEL_H264_V_XMM(put_, ssse3, sse2)
282 317896 QPEL_H264_HV_XMM(put_, ssse3)
283 5580 QPEL_H264_HV_XMM(avg_, ssse3)
284
285 2344 H264_MC(H264_MC_V_H_HV, 4, mmxext, 8, mmxext)
286 182794 H264_MC_816(H264_MC_V, sse2, sse2)
287 216 H264_MC_816(H264_MC_HV, sse2, sse2)
288 235970 H264_MC_816(H264_MC_H, ssse3, sse2)
289 645500 H264_MC_816(H264_MC_HV, ssse3, sse2)
290
291
292 //10bit
293 #define LUMA_MC_OP(OP, NUM, DEPTH, TYPE, OPT) \
294 void ff_ ## OP ## _h264_qpel ## NUM ## _ ## TYPE ## _ ## DEPTH ## _ ## OPT \
295 (uint8_t *dst, const uint8_t *src, ptrdiff_t stride);
296
297 #define LUMA_MC_4(DEPTH, TYPE, OPT) \
298 LUMA_MC_OP(put, 4, DEPTH, TYPE, OPT) \
299 LUMA_MC_OP(avg, 4, DEPTH, TYPE, OPT)
300
301 #define LUMA_MC_816(DEPTH, TYPE, OPT) \
302 LUMA_MC_OP(put, 8, DEPTH, TYPE, OPT) \
303 LUMA_MC_OP(avg, 8, DEPTH, TYPE, OPT) \
304 LUMA_MC_OP(put, 16, DEPTH, TYPE, OPT) \
305 LUMA_MC_OP(avg, 16, DEPTH, TYPE, OPT)
306
307 LUMA_MC_4(10, mc00, mmxext)
308 LUMA_MC_4(10, mc10, mmxext)
309 LUMA_MC_4(10, mc20, mmxext)
310 LUMA_MC_4(10, mc30, mmxext)
311 LUMA_MC_4(10, mc01, mmxext)
312 LUMA_MC_4(10, mc11, mmxext)
313 LUMA_MC_4(10, mc21, mmxext)
314 LUMA_MC_4(10, mc31, mmxext)
315 LUMA_MC_4(10, mc02, mmxext)
316 LUMA_MC_4(10, mc12, mmxext)
317 LUMA_MC_4(10, mc22, mmxext)
318 LUMA_MC_4(10, mc32, mmxext)
319 LUMA_MC_4(10, mc03, mmxext)
320 LUMA_MC_4(10, mc13, mmxext)
321 LUMA_MC_4(10, mc23, mmxext)
322 LUMA_MC_4(10, mc33, mmxext)
323
324 LUMA_MC_816(10, mc00, sse2)
325 LUMA_MC_816(10, mc10, sse2)
326 LUMA_MC_816(10, mc10, ssse3_cache64)
327 LUMA_MC_816(10, mc20, sse2)
328 LUMA_MC_816(10, mc20, ssse3_cache64)
329 LUMA_MC_816(10, mc30, sse2)
330 LUMA_MC_816(10, mc30, ssse3_cache64)
331 LUMA_MC_816(10, mc01, sse2)
332 LUMA_MC_816(10, mc11, sse2)
333 LUMA_MC_816(10, mc21, sse2)
334 LUMA_MC_816(10, mc31, sse2)
335 LUMA_MC_816(10, mc02, sse2)
336 LUMA_MC_816(10, mc12, sse2)
337 LUMA_MC_816(10, mc22, sse2)
338 LUMA_MC_816(10, mc32, sse2)
339 LUMA_MC_816(10, mc03, sse2)
340 LUMA_MC_816(10, mc13, sse2)
341 LUMA_MC_816(10, mc23, sse2)
342 LUMA_MC_816(10, mc33, sse2)
343
344 #define SET_QPEL_FUNCS_1PP(PFX, IDX, SIZE, CPU, PREFIX) \
345 do { \
346 c->PFX ## _pixels_tab[IDX][ 1] = PREFIX ## PFX ## SIZE ## _mc10_ ## CPU; \
347 c->PFX ## _pixels_tab[IDX][ 2] = PREFIX ## PFX ## SIZE ## _mc20_ ## CPU; \
348 c->PFX ## _pixels_tab[IDX][ 3] = PREFIX ## PFX ## SIZE ## _mc30_ ## CPU; \
349 c->PFX ## _pixels_tab[IDX][ 4] = PREFIX ## PFX ## SIZE ## _mc01_ ## CPU; \
350 c->PFX ## _pixels_tab[IDX][ 5] = PREFIX ## PFX ## SIZE ## _mc11_ ## CPU; \
351 c->PFX ## _pixels_tab[IDX][ 6] = PREFIX ## PFX ## SIZE ## _mc21_ ## CPU; \
352 c->PFX ## _pixels_tab[IDX][ 7] = PREFIX ## PFX ## SIZE ## _mc31_ ## CPU; \
353 c->PFX ## _pixels_tab[IDX][ 8] = PREFIX ## PFX ## SIZE ## _mc02_ ## CPU; \
354 c->PFX ## _pixels_tab[IDX][ 9] = PREFIX ## PFX ## SIZE ## _mc12_ ## CPU; \
355 c->PFX ## _pixels_tab[IDX][10] = PREFIX ## PFX ## SIZE ## _mc22_ ## CPU; \
356 c->PFX ## _pixels_tab[IDX][11] = PREFIX ## PFX ## SIZE ## _mc32_ ## CPU; \
357 c->PFX ## _pixels_tab[IDX][12] = PREFIX ## PFX ## SIZE ## _mc03_ ## CPU; \
358 c->PFX ## _pixels_tab[IDX][13] = PREFIX ## PFX ## SIZE ## _mc13_ ## CPU; \
359 c->PFX ## _pixels_tab[IDX][14] = PREFIX ## PFX ## SIZE ## _mc23_ ## CPU; \
360 c->PFX ## _pixels_tab[IDX][15] = PREFIX ## PFX ## SIZE ## _mc33_ ## CPU; \
361 } while (0)
362 #define SET_QPEL_FUNCS(PFX, IDX, SIZE, CPU, PREFIX) \
363 do { \
364 c->PFX ## _pixels_tab[IDX][ 0] = PREFIX ## PFX ## SIZE ## _mc00_ ## CPU; \
365 SET_QPEL_FUNCS_1PP(PFX, IDX, SIZE, CPU, PREFIX); \
366 } while (0)
367
368 #define H264_QPEL_FUNCS(x, y, CPU) \
369 do { \
370 c->put_h264_qpel_pixels_tab[0][x + y * 4] = put_h264_qpel16_mc ## x ## y ## _ ## CPU; \
371 c->put_h264_qpel_pixels_tab[1][x + y * 4] = put_h264_qpel8_mc ## x ## y ## _ ## CPU; \
372 c->avg_h264_qpel_pixels_tab[0][x + y * 4] = avg_h264_qpel16_mc ## x ## y ## _ ## CPU; \
373 c->avg_h264_qpel_pixels_tab[1][x + y * 4] = avg_h264_qpel8_mc ## x ## y ## _ ## CPU; \
374 } while (0)
375
376 #define H264_QPEL_FUNCS_10(x, y, CPU) \
377 do { \
378 c->put_h264_qpel_pixels_tab[0][x + y * 4] = ff_put_h264_qpel16_mc ## x ## y ## _10_ ## CPU; \
379 c->put_h264_qpel_pixels_tab[1][x + y * 4] = ff_put_h264_qpel8_mc ## x ## y ## _10_ ## CPU; \
380 c->avg_h264_qpel_pixels_tab[0][x + y * 4] = ff_avg_h264_qpel16_mc ## x ## y ## _10_ ## CPU; \
381 c->avg_h264_qpel_pixels_tab[1][x + y * 4] = ff_avg_h264_qpel8_mc ## x ## y ## _10_ ## CPU; \
382 } while (0)
383
384 797 av_cold void ff_h264qpel_init_x86(H264QpelContext *c, int bit_depth)
385 {
386 797 int high_bit_depth = bit_depth > 8;
387 797 int cpu_flags = av_get_cpu_flags();
388
389
2/2
✓ Branch 0 taken 152 times.
✓ Branch 1 taken 645 times.
 797 if (EXTERNAL_MMXEXT(cpu_flags)) {
390
2/2
✓ Branch 0 taken 92 times.
✓ Branch 1 taken 60 times.
 152 if (!high_bit_depth) {
391 92 SET_QPEL_FUNCS_1PP(put_h264_qpel, 2, 4, mmxext, );
392 92 SET_QPEL_FUNCS_1PP(avg_h264_qpel, 2, 4, mmxext, );
393 92 c->avg_h264_qpel_pixels_tab[2][0] = ff_avg_pixels4_mmxext;
394
2/2
✓ Branch 0 taken 36 times.
✓ Branch 1 taken 24 times.
 60 } else if (bit_depth == 10) {
395 36 SET_QPEL_FUNCS(put_h264_qpel, 2, 4, 10_mmxext, ff_);
396 36 SET_QPEL_FUNCS(avg_h264_qpel, 2, 4, 10_mmxext, ff_);
397 }
398 }
399
400
2/2
✓ Branch 0 taken 138 times.
✓ Branch 1 taken 659 times.
 797 if (EXTERNAL_SSE2(cpu_flags)) {
401
2/2
✓ Branch 0 taken 86 times.
✓ Branch 1 taken 52 times.
 138 if (!high_bit_depth) {
402 86 H264_QPEL_FUNCS(0, 1, sse2);
403 86 H264_QPEL_FUNCS(0, 2, sse2);
404 86 H264_QPEL_FUNCS(0, 3, sse2);
405 86 H264_QPEL_FUNCS(1, 1, sse2);
406 86 H264_QPEL_FUNCS(1, 2, sse2);
407 86 H264_QPEL_FUNCS(1, 3, sse2);
408 86 H264_QPEL_FUNCS(2, 1, sse2);
409 86 H264_QPEL_FUNCS(2, 2, sse2);
410 86 H264_QPEL_FUNCS(2, 3, sse2);
411 86 H264_QPEL_FUNCS(3, 1, sse2);
412 86 H264_QPEL_FUNCS(3, 2, sse2);
413 86 H264_QPEL_FUNCS(3, 3, sse2);
414 86 c->put_h264_qpel_pixels_tab[0][0] = ff_put_pixels16x16_sse2;
415 86 c->avg_h264_qpel_pixels_tab[0][0] = ff_avg_pixels16x16_sse2;
416 86 c->avg_h264_qpel_pixels_tab[1][0] = ff_avg_pixels8x8_sse2;
417 }
418
419
2/2
✓ Branch 0 taken 32 times.
✓ Branch 1 taken 106 times.
 138 if (bit_depth == 10) {
420 32 SET_QPEL_FUNCS(put_h264_qpel, 0, 16, 10_sse2, ff_);
421 32 SET_QPEL_FUNCS(put_h264_qpel, 1, 8, 10_sse2, ff_);
422 32 SET_QPEL_FUNCS(avg_h264_qpel, 0, 16, 10_sse2, ff_);
423 32 SET_QPEL_FUNCS(avg_h264_qpel, 1, 8, 10_sse2, ff_);
424 32 H264_QPEL_FUNCS_10(1, 0, sse2);
425 32 H264_QPEL_FUNCS_10(2, 0, sse2);
426 32 H264_QPEL_FUNCS_10(3, 0, sse2);
427 }
428 }
429
430
2/2
✓ Branch 0 taken 124 times.
✓ Branch 1 taken 673 times.
 797 if (EXTERNAL_SSSE3(cpu_flags)) {
431
2/2
✓ Branch 0 taken 80 times.
✓ Branch 1 taken 44 times.
 124 if (!high_bit_depth) {
432 80 H264_QPEL_FUNCS(1, 0, ssse3);
433 80 H264_QPEL_FUNCS(1, 1, ssse3);
434 80 H264_QPEL_FUNCS(1, 2, ssse3);
435 80 H264_QPEL_FUNCS(1, 3, ssse3);
436 80 H264_QPEL_FUNCS(2, 0, ssse3);
437 80 H264_QPEL_FUNCS(2, 1, ssse3);
438 80 H264_QPEL_FUNCS(2, 2, ssse3);
439 80 H264_QPEL_FUNCS(2, 3, ssse3);
440 80 H264_QPEL_FUNCS(3, 0, ssse3);
441 80 H264_QPEL_FUNCS(3, 1, ssse3);
442 80 H264_QPEL_FUNCS(3, 2, ssse3);
443 80 H264_QPEL_FUNCS(3, 3, ssse3);
444 }
445
446
2/2
✓ Branch 0 taken 28 times.
✓ Branch 1 taken 96 times.
 124 if (bit_depth == 10) {
447 28 H264_QPEL_FUNCS_10(1, 0, ssse3_cache64);
448 28 H264_QPEL_FUNCS_10(2, 0, ssse3_cache64);
449 28 H264_QPEL_FUNCS_10(3, 0, ssse3_cache64);
450 }
451 }
452 797 }
453