/src/vvenc/source/Lib/CommonLib/x86/InitX86.cpp
Line | Count | Source |
1 | | /* ----------------------------------------------------------------------------- |
2 | | The copyright in this software is being made available under the Clear BSD |
3 | | License, included below. No patent rights, trademark rights and/or |
4 | | other Intellectual Property Rights other than the copyrights concerning |
5 | | the Software are granted under this license. |
6 | | |
7 | | The Clear BSD License |
8 | | |
9 | | Copyright (c) 2019-2026, Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. & The VVenC Authors. |
10 | | All rights reserved. |
11 | | |
12 | | Redistribution and use in source and binary forms, with or without modification, |
13 | | are permitted (subject to the limitations in the disclaimer below) provided that |
14 | | the following conditions are met: |
15 | | |
16 | | * Redistributions of source code must retain the above copyright notice, |
17 | | this list of conditions and the following disclaimer. |
18 | | |
19 | | * Redistributions in binary form must reproduce the above copyright |
20 | | notice, this list of conditions and the following disclaimer in the |
21 | | documentation and/or other materials provided with the distribution. |
22 | | |
23 | | * Neither the name of the copyright holder nor the names of its |
24 | | contributors may be used to endorse or promote products derived from this |
25 | | software without specific prior written permission. |
26 | | |
27 | | NO EXPRESS OR IMPLIED LICENSES TO ANY PARTY'S PATENT RIGHTS ARE GRANTED BY |
28 | | THIS LICENSE. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND |
29 | | CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT |
30 | | LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A |
31 | | PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR |
32 | | CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, |
33 | | EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, |
34 | | PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR |
35 | | BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER |
36 | | IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) |
37 | | ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE |
38 | | POSSIBILITY OF SUCH DAMAGE. |
39 | | |
40 | | |
41 | | ------------------------------------------------------------------------------------------- */ |
42 | | |
43 | | |
44 | | /* |
45 | | * \ingroup CommonLib |
46 | | * \file InitX86.cpp |
47 | | * \brief Initialize encoder SIMD functions. |
48 | | */ |
49 | | |
50 | | #include "CommonDefX86.h" |
51 | | #include "InterpolationFilter.h" |
52 | | #include "TrQuant.h" |
53 | | #include "RdCost.h" |
54 | | #include "Unit.h" |
55 | | #include "LoopFilter.h" |
56 | | #include "AdaptiveLoopFilter.h" |
57 | | #include "SampleAdaptiveOffset.h" |
58 | | #include "InterPrediction.h" |
59 | | #include "IntraPrediction.h" |
60 | | #include "AffineGradientSearch.h" |
61 | | #include "MCTF.h" |
62 | | #include "TrQuant_EMT.h" |
63 | | #include "QuantRDOQ2.h" |
64 | | #include "SEIFilmGrainAnalyzer.h" |
65 | | |
66 | | #ifdef TARGET_SIMD_X86 |
67 | | |
68 | | //! \ingroup CommonLib |
69 | | //! \{ |
70 | | |
71 | | # if defined( REAL_TARGET_X86 ) \ |
72 | | || ( defined( SIMD_EVERYWHERE_EXTENSION_LEVEL_ID ) && SIMD_EVERYWHERE_EXTENSION_LEVEL_ID >= X86_SIMD_AVX2 ) |
73 | | # define ENABLE_AVX2_IMPLEMENTATIONS 1 |
74 | | # else |
75 | | # define ENABLE_AVX2_IMPLEMENTATIONS 0 |
76 | | # endif |
77 | | |
78 | | namespace vvenc { |
79 | | |
80 | | #if ENABLE_SIMD_OPT_MCIF |
81 | | void InterpolationFilter::initInterpolationFilterX86( /*int iBitDepthY, int iBitDepthC*/ ) |
82 | 0 | { |
83 | 0 | auto vext = read_x86_extension_flags(); |
84 | 0 | switch (vext){ |
85 | 0 | case AVX512: |
86 | 0 | case AVX2: |
87 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
88 | 0 | _initInterpolationFilterX86<AVX2>(/*iBitDepthY, iBitDepthC*/); |
89 | 0 | break; |
90 | 0 | #endif |
91 | 0 | case AVX: |
92 | | //_initInterpolationFilterX86<AVX>(/*iBitDepthY, iBitDepthC*/); |
93 | | //break; |
94 | 0 | case SSE42: |
95 | 0 | case SSE41: |
96 | 0 | _initInterpolationFilterX86<SSE41>(/*iBitDepthY, iBitDepthC*/); |
97 | 0 | break; |
98 | 0 | default: |
99 | 0 | break; |
100 | 0 | } |
101 | 0 | } |
102 | | #endif |
103 | | |
104 | | #if ENABLE_SIMD_OPT_BUFFER |
105 | | void PelBufferOps::initPelBufOpsX86() |
106 | 0 | { |
107 | 0 | if( isInitX86Done ) |
108 | 0 | return; |
109 | | |
110 | 0 | isInitX86Done = true; |
111 | |
|
112 | 0 | auto vext = read_x86_extension_flags(); |
113 | 0 | switch (vext){ |
114 | 0 | case AVX512: |
115 | 0 | case AVX2: |
116 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
117 | 0 | _initPelBufOpsX86<AVX2>(); |
118 | 0 | break; |
119 | 0 | #endif |
120 | 0 | case AVX: |
121 | | //_initPelBufOpsX86<AVX>(); |
122 | | //break; |
123 | 0 | case SSE42: |
124 | 0 | case SSE41: |
125 | 0 | _initPelBufOpsX86<SSE41>(); |
126 | 0 | break; |
127 | 0 | default: |
128 | 0 | break; |
129 | 0 | } |
130 | 0 | } |
131 | | #endif |
132 | | |
133 | | |
134 | | #if ENABLE_SIMD_DBLF |
135 | | void LoopFilter::initLoopFilterX86() |
136 | 0 | { |
137 | 0 | auto vext = read_x86_extension_flags(); |
138 | 0 | switch (vext) |
139 | 0 | { |
140 | 0 | case AVX512: |
141 | 0 | case AVX2: |
142 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
143 | 0 | _initLoopFilterX86<AVX2>(); |
144 | 0 | break; |
145 | 0 | #endif |
146 | 0 | case AVX: |
147 | | //_initLoopFilterX86<AVX>(); |
148 | | //break; |
149 | 0 | case SSE42: |
150 | 0 | case SSE41: |
151 | 0 | _initLoopFilterX86<SSE41>(); |
152 | 0 | break; |
153 | 0 | default: |
154 | 0 | break; |
155 | 0 | } |
156 | 0 | } |
157 | | #endif |
158 | | |
159 | | |
160 | | #if ENABLE_SIMD_OPT_DIST |
161 | | void RdCost::initRdCostX86() |
162 | 0 | { |
163 | 0 | auto vext = read_x86_extension_flags(); |
164 | 0 | switch (vext){ |
165 | 0 | case AVX512: |
166 | 0 | case AVX2: |
167 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
168 | | #if defined( _MSC_VER ) && _MSC_VER >= 1938 && _MSC_VER < 1939 |
169 | | #else |
170 | 0 | _initRdCostX86<AVX2>(); |
171 | 0 | break; |
172 | 0 | #endif |
173 | 0 | #endif |
174 | 0 | case AVX: |
175 | | //_initRdCostX86<AVX>(); |
176 | | //break; |
177 | 0 | case SSE42: |
178 | 0 | case SSE41: |
179 | 0 | _initRdCostX86<SSE41>(); |
180 | 0 | break; |
181 | 0 | default: |
182 | 0 | break; |
183 | 0 | } |
184 | 0 | } |
185 | | #endif |
186 | | |
187 | | #if ENABLE_SIMD_OPT_ALF |
188 | | void AdaptiveLoopFilter::initAdaptiveLoopFilterX86() |
189 | 0 | { |
190 | 0 | auto vext = read_x86_extension_flags(); |
191 | 0 | switch( vext ) |
192 | 0 | { |
193 | 0 | case AVX512: |
194 | 0 | case AVX2: |
195 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
196 | 0 | _initAdaptiveLoopFilterX86<AVX2>(); |
197 | 0 | break; |
198 | 0 | #endif |
199 | 0 | case AVX: |
200 | | //_initAdaptiveLoopFilterX86<AVX>(); |
201 | | //break; |
202 | 0 | case SSE42: |
203 | 0 | case SSE41: |
204 | 0 | _initAdaptiveLoopFilterX86<SSE41>(); |
205 | 0 | break; |
206 | 0 | default: |
207 | 0 | break; |
208 | 0 | } |
209 | 0 | } |
210 | | #endif |
211 | | |
212 | | #if ENABLE_SIMD_OPT_SAO |
213 | | void SampleAdaptiveOffset::initSampleAdaptiveOffsetX86() |
214 | 0 | { |
215 | 0 | auto vext = read_x86_extension_flags(); |
216 | 0 | switch (vext){ |
217 | 0 | case AVX512: |
218 | 0 | case AVX2: |
219 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
220 | 0 | _initSampleAdaptiveOffsetX86<AVX2>(); |
221 | 0 | break; |
222 | 0 | #endif |
223 | 0 | case AVX: |
224 | | //_initSampleAdaptiveOffsetX86<AVX>(); |
225 | | //break; |
226 | 0 | case SSE42: |
227 | 0 | case SSE41: |
228 | 0 | _initSampleAdaptiveOffsetX86<SSE41>(); |
229 | 0 | break; |
230 | 0 | default: |
231 | 0 | break; |
232 | 0 | } |
233 | 0 | } |
234 | | |
235 | | #endif |
236 | | |
237 | | #if ENABLE_SIMD_OPT_BDOF |
238 | | void InterPredInterpolation::initInterPredictionX86() |
239 | 0 | { |
240 | 0 | auto vext = read_x86_extension_flags(); |
241 | 0 | switch (vext){ |
242 | 0 | case AVX512: |
243 | 0 | case AVX2: |
244 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
245 | 0 | _initInterPredictionX86<AVX2>(); |
246 | 0 | break; |
247 | 0 | #endif |
248 | 0 | case AVX: |
249 | | //_initInterPredictionX86<AVX>(); |
250 | | //break; |
251 | 0 | case SSE42: |
252 | 0 | case SSE41: |
253 | 0 | _initInterPredictionX86<SSE41>(); |
254 | 0 | break; |
255 | 0 | default: |
256 | 0 | break; |
257 | 0 | } |
258 | 0 | } |
259 | | #endif |
260 | | |
261 | | #if ENABLE_SIMD_OPT_AFFINE_ME |
262 | | void AffineGradientSearch::initAffineGradientSearchX86() |
263 | 0 | { |
264 | 0 | auto vext = read_x86_extension_flags(); |
265 | 0 | switch (vext) { |
266 | 0 | case AVX512: |
267 | 0 | case AVX2: |
268 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
269 | 0 | _initAffineGradientSearchX86<AVX2>(); |
270 | 0 | break; |
271 | 0 | #endif |
272 | 0 | case AVX: |
273 | | //_initAffineGradientSearchX86<AVX>(); |
274 | | //break; |
275 | 0 | case SSE42: |
276 | 0 | case SSE41: |
277 | 0 | _initAffineGradientSearchX86<SSE41>(); |
278 | 0 | break; |
279 | 0 | default: |
280 | 0 | break; |
281 | 0 | } |
282 | 0 | } |
283 | | #endif |
284 | | |
285 | | #if ENABLE_SIMD_OPT_INTRAPRED |
286 | | void IntraPrediction::initIntraPredictionX86() |
287 | 0 | { |
288 | 0 | auto vext = read_x86_extension_flags(); |
289 | 0 | switch (vext){ |
290 | 0 | case AVX512: |
291 | 0 | case AVX2: |
292 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
293 | 0 | _initIntraPredictionX86<AVX2>(); |
294 | 0 | break; |
295 | 0 | #endif |
296 | 0 | case AVX: |
297 | | //_initIntraPredictionX86<AVX>(); |
298 | | //break; |
299 | 0 | case SSE42: |
300 | 0 | case SSE41: |
301 | 0 | _initIntraPredictionX86<SSE41>(); |
302 | 0 | break; |
303 | 0 | default: |
304 | 0 | break; |
305 | 0 | } |
306 | 0 | } |
307 | | |
308 | | #endif |
309 | | #if ENABLE_SIMD_OPT_MCTF |
310 | | void MCTF::initMCTF_X86() |
311 | 0 | { |
312 | 0 | auto vext = read_x86_extension_flags(); |
313 | 0 | switch (vext){ |
314 | 0 | case AVX512: |
315 | 0 | case AVX2: |
316 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
317 | 0 | _initMCTF_X86<AVX2 >(); |
318 | 0 | break; |
319 | 0 | #endif |
320 | 0 | case AVX: |
321 | | //_initMCTF_X86<AVX >(); |
322 | | //break; |
323 | 0 | case SSE42: |
324 | | //_initMCTF_X86<SSE42>(); |
325 | | //break; |
326 | 0 | case SSE41: |
327 | 0 | _initMCTF_X86<SSE41>(); |
328 | 0 | break; |
329 | 0 | default: |
330 | 0 | break; |
331 | 0 | } |
332 | 0 | } |
333 | | |
334 | | #endif |
335 | | #if ENABLE_SIMD_TRAFO |
336 | | void TCoeffOps::initTCoeffOpsX86() |
337 | 0 | { |
338 | 0 | auto vext = read_x86_extension_flags(); |
339 | 0 | switch (vext){ |
340 | 0 | case AVX512: |
341 | 0 | case AVX2: |
342 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
343 | 0 | _initTCoeffOpsX86<AVX2 >(); |
344 | 0 | break; |
345 | 0 | #endif |
346 | 0 | case AVX: |
347 | | //_initTCoeffOpsX86<AVX >(); |
348 | | //break; |
349 | 0 | case SSE42: |
350 | | //_initTCoeffOpsX86<SSE42>(); |
351 | | //break; |
352 | 0 | case SSE41: |
353 | 0 | _initTCoeffOpsX86<SSE41>(); |
354 | 0 | break; |
355 | 0 | default: |
356 | 0 | break; |
357 | 0 | } |
358 | 0 | } |
359 | | |
360 | | void TrQuant::initTrQuantX86() |
361 | 0 | { |
362 | 0 | auto vext = read_x86_extension_flags(); |
363 | 0 | switch( vext ) |
364 | 0 | { |
365 | 0 | case AVX512: |
366 | 0 | case AVX2: |
367 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
368 | 0 | _initTrQuantX86<AVX2 >(); |
369 | 0 | break; |
370 | 0 | #endif |
371 | 0 | case AVX: |
372 | | //_initTrQuantX86<AVX >(); |
373 | | //break; |
374 | 0 | case SSE42: |
375 | | //_initTrQuantX86<SSE42>(); |
376 | | //break; |
377 | 0 | case SSE41: |
378 | 0 | _initTrQuantX86<SSE41>(); |
379 | 0 | break; |
380 | 0 | default: |
381 | 0 | break; |
382 | 0 | } |
383 | 0 | } |
384 | | |
385 | | #endif |
386 | | |
387 | | #if ENABLE_SIMD_OPT_QUANT |
388 | | |
389 | | void Quant::initQuantX86() |
390 | 0 | { |
391 | 0 | auto vext = read_x86_extension_flags(); |
392 | 0 | switch (vext){ |
393 | 0 | case AVX512: |
394 | 0 | case AVX2: |
395 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
396 | 0 | _initQuantX86<AVX2>(); |
397 | 0 | break; |
398 | 0 | #endif |
399 | 0 | case AVX: |
400 | | //_initQuantX86<AVX>(); |
401 | | //break; |
402 | 0 | case SSE42: |
403 | 0 | case SSE41: |
404 | 0 | _initQuantX86<SSE41>(); |
405 | 0 | break; |
406 | 0 | default: |
407 | 0 | break; |
408 | 0 | } |
409 | 0 | } |
410 | | |
411 | | void DepQuant::initDepQuantX86() |
412 | 0 | { |
413 | 0 | auto vext = read_x86_extension_flags(); |
414 | 0 | switch (vext){ |
415 | 0 | case AVX512: |
416 | 0 | case AVX2: |
417 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
418 | 0 | _initDepQuantX86<AVX2>(); |
419 | 0 | break; |
420 | 0 | #endif |
421 | 0 | case AVX: |
422 | 0 | case SSE42: |
423 | 0 | _initDepQuantX86<SSE42>(); |
424 | 0 | break; |
425 | 0 | case SSE41: |
426 | 0 | _initDepQuantX86<SSE41>(); |
427 | 0 | break; |
428 | 0 | default: |
429 | 0 | break; |
430 | 0 | } |
431 | 0 | } |
432 | | |
433 | | #endif |
434 | | |
435 | | #if ENABLE_SIMD_OPT_FGA |
436 | | void Canny::initFGACannyX86() |
437 | 0 | { |
438 | 0 | auto vext = read_x86_extension_flags(); |
439 | 0 | switch (vext){ |
440 | 0 | case AVX512: |
441 | 0 | case AVX2: |
442 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
443 | 0 | _initFGACannyX86<AVX2>(); |
444 | 0 | break; |
445 | 0 | #endif |
446 | 0 | case SSE42: |
447 | 0 | case SSE41: |
448 | 0 | _initFGACannyX86<SSE41>(); |
449 | 0 | break; |
450 | 0 | default: |
451 | 0 | break; |
452 | 0 | } |
453 | 0 | } |
454 | | void Morph::initFGAMorphX86() |
455 | 0 | { |
456 | 0 | auto vext = read_x86_extension_flags(); |
457 | 0 | switch (vext){ |
458 | 0 | case AVX512: |
459 | 0 | case AVX2: |
460 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
461 | 0 | _initFGAMorphX86<AVX2>(); |
462 | 0 | break; |
463 | 0 | #endif |
464 | 0 | case SSE42: |
465 | 0 | case SSE41: |
466 | 0 | _initFGAMorphX86<SSE41>(); |
467 | 0 | break; |
468 | 0 | default: |
469 | 0 | break; |
470 | 0 | } |
471 | 0 | } |
472 | | |
473 | | void FGAnalyzer::initFGAnalyzerX86() |
474 | 0 | { |
475 | 0 | auto vext = read_x86_extension_flags(); |
476 | 0 | switch (vext){ |
477 | 0 | case AVX512: |
478 | 0 | case AVX2: |
479 | 0 | #if ENABLE_AVX2_IMPLEMENTATIONS |
480 | 0 | _initFGAnalyzerX86<AVX2>(); |
481 | 0 | break; |
482 | 0 | #endif |
483 | 0 | case SSE42: |
484 | 0 | case SSE41: |
485 | 0 | _initFGAnalyzerX86<SSE41>(); |
486 | 0 | break; |
487 | 0 | default: |
488 | 0 | break; |
489 | 0 | } |
490 | 0 | } |
491 | | #endif |
492 | | |
493 | | } // namespace vvenc |
494 | | |
495 | | //! \} |
496 | | |
497 | | #endif // TARGET_SIMD_X86 |
498 | | |