[svn] / trunk / xvidcore / src / xvid.c Repository:
ViewVC logotype

Diff of /trunk/xvidcore/src/xvid.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1382, Mon Mar 22 22:36:25 2004 UTC revision 1709, Wed Jun 14 21:44:07 2006 UTC
# Line 3  Line 3 
3   *  XVID MPEG-4 VIDEO CODEC   *  XVID MPEG-4 VIDEO CODEC
4   *  - Native API implementation  -   *  - Native API implementation  -
5   *   *
6   *  Copyright(C) 2001-2003 Peter Ross <pross@xvid.org>   *  Copyright(C) 2001-2004 Peter Ross <pross@xvid.org>
7   *   *
8   *  This program is free software ; you can redistribute it and/or modify   *  This program is free software ; you can redistribute it and/or modify
9   *  it under the terms of the GNU General Public License as published by   *  it under the terms of the GNU General Public License as published by
# Line 19  Line 19 
19   *  along with this program ; if not, write to the Free Software   *  along with this program ; if not, write to the Free Software
20   *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA   *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
21   *   *
22   * $Id: xvid.c,v 1.48 2004-03-22 22:36:23 edgomez Exp $   * $Id: xvid.c,v 1.69 2006-06-14 21:44:07 Skal Exp $
23   *   *
24   ****************************************************************************/   ****************************************************************************/
25    
# Line 36  Line 36 
36  #include "dct/fdct.h"  #include "dct/fdct.h"
37  #include "image/colorspace.h"  #include "image/colorspace.h"
38  #include "image/interpolate8x8.h"  #include "image/interpolate8x8.h"
 #include "image/reduced.h"  
39  #include "utils/mem_transfer.h"  #include "utils/mem_transfer.h"
40  #include "utils/mbfunctions.h"  #include "utils/mbfunctions.h"
41  #include "quant/quant.h"  #include "quant/quant.h"
42  #include "motion/motion.h"  #include "motion/motion.h"
43    #include "motion/gmc.h"
44  #include "motion/sad.h"  #include "motion/sad.h"
45  #include "utils/emms.h"  #include "utils/emms.h"
46  #include "utils/timer.h"  #include "utils/timer.h"
# Line 52  Line 52 
52  unsigned int xvid_debug = 0; /* xvid debug mask */  unsigned int xvid_debug = 0; /* xvid debug mask */
53  #endif  #endif
54    
55  #if defined(ARCH_IS_IA32)  #if (defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)) && defined(_MSC_VER)
 #if defined(_MSC_VER)  
56  #       include <windows.h>  #       include <windows.h>
57  #else  #elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || defined(ARCH_IS_PPC)
58  #       include <signal.h>  #       include <signal.h>
59  #       include <setjmp.h>  #       include <setjmp.h>
60    
# Line 78  Line 77 
77   *   0 : SIGILL was *not* signalled   *   0 : SIGILL was *not* signalled
78   *   1 : SIGILL was signalled   *   1 : SIGILL was signalled
79   */   */
80    #if (defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)) && defined(_MSC_VER)
81  int  static int
82  sigill_check(void (*func)())  sigill_check(void (*func)())
83  {  {
 #if defined(_MSC_VER)  
84          _try {          _try {
85                  func();                  func();
86          }          } _except(EXCEPTION_EXECUTE_HANDLER) {
         _except(EXCEPTION_EXECUTE_HANDLER) {  
87    
88                  if (_exception_code() == STATUS_ILLEGAL_INSTRUCTION)                  if (_exception_code() == STATUS_ILLEGAL_INSTRUCTION)
89                          return 1;                          return(1);
90          }          }
91          return 0;          return(0);
92  #else  }
93    #elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || defined(ARCH_IS_PPC)
94    static int
95    sigill_check(void (*func)())
96    {
97      void * old_handler;      void * old_handler;
98      int jmpret;      int jmpret;
99    
100        /* Set our SIGILL handler */
101      old_handler = signal(SIGILL, sigill_handler);      old_handler = signal(SIGILL, sigill_handler);
102      if (old_handler == SIG_ERR)  
103      {      /* Check for error */
104          return -1;      if (old_handler == SIG_ERR) {
105            return(-1);
106      }      }
107    
108        /* Save stack context, so if func triggers a SIGILL, we can still roll
109             * back to a valid CPU state */
110      jmpret = setjmp(mark);      jmpret = setjmp(mark);
111      if (jmpret == 0)  
112      {          /* If setjmp returned directly, then its returned value is 0, and we still
113             * have to test the passed func. Otherwise it means the stack context has
114             * been restored by a longjmp() call, which in our case happens only in the
115             * signal handler */
116        if (jmpret == 0) {
117          func();          func();
118      }      }
119    
120        /* Restore old signal handler */
121      signal(SIGILL, old_handler);      signal(SIGILL, old_handler);
122    
123      return jmpret;      return(jmpret);
 #endif  
124  }  }
125  #endif  #endif
126    
127    
128  /* detect cpu flags  */  /* detect cpu flags  */
129  static unsigned int  static unsigned int
130  detect_cpu_flags()  detect_cpu_flags(void)
131  {  {
132          /* enable native assembly optimizations by default */          /* enable native assembly optimizations by default */
133          unsigned int cpu_flags = XVID_CPU_ASM;          unsigned int cpu_flags = XVID_CPU_ASM;
134    
135  #if defined(ARCH_IS_IA32)  #if defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)
136          cpu_flags |= check_cpu_features();          cpu_flags |= check_cpu_features();
137          if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger))          if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger))
138                  cpu_flags &= ~XVID_CPU_SSE;                  cpu_flags &= ~XVID_CPU_SSE;
# Line 134  Line 142 
142  #endif  #endif
143    
144  #if defined(ARCH_IS_PPC)  #if defined(ARCH_IS_PPC)
145  #if defined(ARCH_IS_PPC_ALTIVEC)          if (!sigill_check(altivec_trigger))
146          cpu_flags |= XVID_CPU_ALTIVEC;          cpu_flags |= XVID_CPU_ALTIVEC;
147  #endif  #endif
 #endif  
148    
149          return cpu_flags;          return cpu_flags;
150  }  }
# Line 177  Line 184 
184          idct = idct_int32;          idct = idct_int32;
185    
186          /* Only needed on PPC Altivec archs */          /* Only needed on PPC Altivec archs */
187          sadInit = 0;          sadInit = NULL;
188    
189          /* Restore FPU context : emms_c is a nop functions */          /* Restore FPU context : emms_c is a nop functions */
190          emms = emms_c;          emms = emms_c;
# Line 204  Line 211 
211          transfer_8to16sub  = transfer_8to16sub_c;          transfer_8to16sub  = transfer_8to16sub_c;
212          transfer_8to16subro  = transfer_8to16subro_c;          transfer_8to16subro  = transfer_8to16subro_c;
213          transfer_8to16sub2 = transfer_8to16sub2_c;          transfer_8to16sub2 = transfer_8to16sub2_c;
214            transfer_8to16sub2ro = transfer_8to16sub2ro_c;
215          transfer_16to8add  = transfer_16to8add_c;          transfer_16to8add  = transfer_16to8add_c;
216          transfer8x8_copy   = transfer8x8_copy_c;          transfer8x8_copy   = transfer8x8_copy_c;
217            transfer8x4_copy   = transfer8x4_copy_c;
218    
219          /* Interlacing functions */          /* Interlacing functions */
220          MBFieldTest = MBFieldTest_c;          MBFieldTest = MBFieldTest_c;
# Line 215  Line 224 
224          interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_c;          interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_c;
225          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;
226    
227            interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_c;
228            interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_c;
229            interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_c;
230    
231            interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_c;
232            interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_c;
233            interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_c;
234            interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_c;
235    
236          interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c;          interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c;
237          interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c;          interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c;
238          interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c;          interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c;
# Line 229  Line 247 
247          interpolate8x8_avg2 = interpolate8x8_avg2_c;          interpolate8x8_avg2 = interpolate8x8_avg2_c;
248          interpolate8x8_avg4 = interpolate8x8_avg4_c;          interpolate8x8_avg4 = interpolate8x8_avg4_c;
249    
250          /* reduced resolution */          /* postprocessing */
251          copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_C;          image_brightness = image_brightness_c;
         add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_C;  
         vfilter_31 = xvid_VFilter_31_C;  
         hfilter_31 = xvid_HFilter_31_C;  
         filter_18x18_to_8x8 = xvid_Filter_18x18_To_8x8_C;  
         filter_diff_18x18_to_8x8 = xvid_Filter_Diff_18x18_To_8x8_C;  
252    
253          /* Initialize internal colorspace transformation tables */          /* Initialize internal colorspace transformation tables */
254          colorspace_init();          colorspace_init();
# Line 292  Line 305 
305          dev16      = dev16_c;          dev16      = dev16_c;
306          sad16v     = sad16v_c;          sad16v     = sad16v_c;
307          sse8_16bit = sse8_16bit_c;          sse8_16bit = sse8_16bit_c;
308            sse8_8bit  = sse8_8bit_c;
309    
310  #if defined(ARCH_IS_IA32)          init_GMC(cpu_flags);
311    
312          if ((cpu_flags & XVID_CPU_ASM)) {  #if defined(ARCH_IS_IA32)
                 vfilter_31 = xvid_VFilter_31_x86;  
                 hfilter_31 = xvid_HFilter_31_x86;  
         }  
313    
314          if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||          if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||
315                  (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||                  (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||
# Line 337  Line 348 
348                  transfer_8to16sub2 = transfer_8to16sub2_mmx;                  transfer_8to16sub2 = transfer_8to16sub2_mmx;
349                  transfer_16to8add  = transfer_16to8add_mmx;                  transfer_16to8add  = transfer_16to8add_mmx;
350                  transfer8x8_copy   = transfer8x8_copy_mmx;                  transfer8x8_copy   = transfer8x8_copy_mmx;
351                    transfer8x4_copy   = transfer8x4_copy_mmx;
352    
353                  /* Interlacing Functions */                  /* Interlacing Functions */
354                  MBFieldTest = MBFieldTest_mmx;                  MBFieldTest = MBFieldTest_mmx;
# Line 346  Line 358 
358                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_mmx;                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_mmx;
359                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;
360    
361                    interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_mmx;
362                    interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_mmx;
363                    interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_mmx;
364    
365                    interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_mmx;
366                    interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_mmx;
367                    interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_mmx;
368                    interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_mmx;
369    
370                  interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx;                  interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx;
371                  interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx;                  interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx;
372    
373                  interpolate8x8_avg2 = interpolate8x8_avg2_mmx;                  interpolate8x8_avg2 = interpolate8x8_avg2_mmx;
374                  interpolate8x8_avg4 = interpolate8x8_avg4_mmx;                  interpolate8x8_avg4 = interpolate8x8_avg4_mmx;
375    
376                  /* reduced resolution */                  /* postprocessing */
377                  copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_mmx;                  image_brightness = image_brightness_mmx;
                 add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_mmx;  
                 hfilter_31 = xvid_HFilter_31_mmx;  
                 filter_18x18_to_8x8 = xvid_Filter_18x18_To_8x8_mmx;  
                 filter_diff_18x18_to_8x8 = xvid_Filter_Diff_18x18_To_8x8_mmx;  
378    
379                  /* image input xxx_to_yv12 related functions */                  /* image input xxx_to_yv12 related functions */
380                  yv12_to_yv12  = yv12_to_yv12_mmx;                  yv12_to_yv12  = yv12_to_yv12_mmx;
# Line 384  Line 401 
401                  dev16    = dev16_mmx;                  dev16    = dev16_mmx;
402                  sad16v   = sad16v_mmx;                  sad16v   = sad16v_mmx;
403                  sse8_16bit = sse8_16bit_mmx;                  sse8_16bit = sse8_16bit_mmx;
404                    sse8_8bit  = sse8_8bit_mmx;
405          }          }
406    
407          /* these 3dnow functions are faster than mmx, but slower than xmm. */          /* these 3dnow functions are faster than mmx, but slower than xmm. */
# Line 411  Line 429 
429                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_xmm;                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_xmm;
430                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;
431    
432                  /* reduced resolution */                  interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_xmm;
433                  copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_xmm;                  interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_xmm;
434                  add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_xmm;                  interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_xmm;
435    
436                    interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_xmm;
437                    interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_xmm;
438                    interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_xmm;
439                    interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_xmm;
440    
441                  /* Quantization */                  /* Quantization */
442                  quant_mpeg_intra = quant_mpeg_intra_xmm;                  quant_mpeg_intra = quant_mpeg_intra_xmm;
# Line 424  Line 447 
447    
448                  /* Buffer transfer */                  /* Buffer transfer */
449                  transfer_8to16sub2 = transfer_8to16sub2_xmm;                  transfer_8to16sub2 = transfer_8to16sub2_xmm;
450                    transfer_8to16sub2ro = transfer_8to16sub2ro_xmm;
451    
452                  /* Colorspace transformation */                  /* Colorspace transformation */
453                  yv12_to_yv12  = yv12_to_yv12_xmm;                  yv12_to_yv12  = yv12_to_yv12_xmm;
# Line 445  Line 469 
469                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn;                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn;
470                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn;                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn;
471                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;
472    
473                    interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_3dn;
474                    interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_3dn;
475                    interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_3dn;
476          }          }
477    
478          if ((cpu_flags & XVID_CPU_3DNOWEXT)) {          if ((cpu_flags & XVID_CPU_3DNOWEXT)) {
479    
                 /* Inverse DCT */  
                 idct =  idct_3dne;  
   
480                  /* Buffer transfer */                  /* Buffer transfer */
481                  transfer_8to16copy =  transfer_8to16copy_3dne;                  transfer_8to16copy =  transfer_8to16copy_3dne;
482                  transfer_16to8copy = transfer_16to8copy_3dne;                  transfer_16to8copy = transfer_16to8copy_3dne;
483                  transfer_8to16sub =  transfer_8to16sub_3dne;                  transfer_8to16sub =  transfer_8to16sub_3dne;
484                  transfer_8to16subro =  transfer_8to16subro_3dne;                  transfer_8to16subro =  transfer_8to16subro_3dne;
                 transfer_8to16sub2 =  transfer_8to16sub2_3dne;  
485                  transfer_16to8add = transfer_16to8add_3dne;                  transfer_16to8add = transfer_16to8add_3dne;
486                  transfer8x8_copy = transfer8x8_copy_3dne;                  transfer8x8_copy = transfer8x8_copy_3dne;
487                    transfer8x4_copy = transfer8x4_copy_3dne;
488    
489                    if ((cpu_flags & XVID_CPU_MMXEXT)) {
490                            /* Inverse DCT */
491                            idct =  idct_3dne;
492    
493                            /* Buffer transfer */
494                            transfer_8to16sub2 =  transfer_8to16sub2_3dne;
495    
496                            /* Interpolation */
497                            interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dne;
498                            interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne;
499                            interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne;
500    
501                            interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_3dne;
502                            interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_3dne;
503                            interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_3dne;
504    
505                  /* Quantization */                  /* Quantization */
506                  quant_h263_intra = quant_h263_intra_3dne;                          quant_h263_intra = quant_h263_intra_3dne;               /* cmov only */
507                  quant_h263_inter = quant_h263_inter_3dne;                  quant_h263_inter = quant_h263_inter_3dne;
508                  dequant_mpeg_intra = dequant_mpeg_intra_3dne;                          dequant_mpeg_intra = dequant_mpeg_intra_3dne;   /* cmov only */
509                  dequant_mpeg_inter = dequant_mpeg_inter_3dne;                  dequant_mpeg_inter = dequant_mpeg_inter_3dne;
510                  dequant_h263_intra = dequant_h263_intra_3dne;                  dequant_h263_intra = dequant_h263_intra_3dne;
511                  dequant_h263_inter = dequant_h263_inter_3dne;                  dequant_h263_inter = dequant_h263_inter_3dne;
512    
513                  /* ME functions */                  /* ME functions */
514                  calc_cbp = calc_cbp_3dne;                  calc_cbp = calc_cbp_3dne;
515    
516                  sad16 = sad16_3dne;                  sad16 = sad16_3dne;
517                  sad8 = sad8_3dne;                  sad8 = sad8_3dne;
518                  sad16bi = sad16bi_3dne;                  sad16bi = sad16bi_3dne;
519                  sad8bi = sad8bi_3dne;                  sad8bi = sad8bi_3dne;
520                  dev16 = dev16_3dne;                  dev16 = dev16_3dne;
521                    }
                 /* Interpolation */  
                 interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dne;  
                 interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne;  
                 interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne;  
522          }          }
523    
524          if ((cpu_flags & XVID_CPU_SSE2)) {          if ((cpu_flags & XVID_CPU_SSE2)) {
# Line 497  Line 535 
535                  sad16    = sad16_sse2;                  sad16    = sad16_sse2;
536                  dev16    = dev16_sse2;                  dev16    = dev16_sse2;
537    
538                  /* DCT operators                  /* DCT operators */
                  * no iDCT because it's not "Walken matching" */  
539                  fdct = fdct_sse2_skal;                  fdct = fdct_sse2_skal;
540        /* idct = idct_sse2_skal; */   /* Is now IEEE1180 and Walken compliant. Disabled until fully tested. */
541    
542                    /* postprocessing */
543                    image_brightness = image_brightness_sse2;
544          }          }
545  #endif /* ARCH_IS_IA32 */  #endif /* ARCH_IS_IA32 */
546    
# Line 530  Line 571 
571  #endif  #endif
572    
573  #if defined(ARCH_IS_PPC)  #if defined(ARCH_IS_PPC)
574          if ((cpu_flags & XVID_CPU_ASM))          if ((cpu_flags & XVID_CPU_ALTIVEC)) {
575          {            /* sad operators */
576                  calc_cbp = calc_cbp_ppc;                    sad16 = sad16_altivec_c;
577                      sad16bi = sad16bi_altivec_c;
578                      sad8 = sad8_altivec_c;
579                      dev16 = dev16_altivec_c;
580    
581              sse8_16bit = sse8_16bit_altivec_c;
582    
583              /* mem transfer */
584              transfer_8to16copy = transfer_8to16copy_altivec_c;
585              transfer_16to8copy = transfer_16to8copy_altivec_c;
586              transfer_8to16sub = transfer_8to16sub_altivec_c;
587              transfer_8to16subro = transfer_8to16subro_altivec_c;
588              transfer_8to16sub2 = transfer_8to16sub2_altivec_c;
589              transfer_16to8add = transfer_16to8add_altivec_c;
590              transfer8x8_copy = transfer8x8_copy_altivec_c;
591    
592              /* Inverse DCT */
593              idct = idct_altivec_c;
594    
595              /* Interpolation */
596              interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_altivec_c;
597              interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_altivec_c;
598              interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_altivec_c;
599    
600              interpolate8x8_avg2 = interpolate8x8_avg2_altivec_c;
601              interpolate8x8_avg4 = interpolate8x8_avg4_altivec_c;
602    
603                      interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_altivec_c;
604                      interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_altivec_c;
605                      interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_altivec_c;
606                      interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_altivec_c;
607    
608              /* Colorspace conversion */
609              bgra_to_yv12 = bgra_to_yv12_altivec_c;
610              abgr_to_yv12 = abgr_to_yv12_altivec_c;
611              rgba_to_yv12 = rgba_to_yv12_altivec_c;
612              argb_to_yv12 = argb_to_yv12_altivec_c;
613    
614              yuyv_to_yv12 = yuyv_to_yv12_altivec_c;
615              uyvy_to_yv12 = uyvy_to_yv12_altivec_c;
616    
617              yv12_to_yuyv = yv12_to_yuyv_altivec_c;
618              yv12_to_uyvy = yv12_to_uyvy_altivec_c;
619    
620              /* Quantization */
621              quant_h263_intra = quant_h263_intra_altivec_c;
622              quant_h263_inter = quant_h263_inter_altivec_c;
623              dequant_h263_intra = dequant_h263_intra_altivec_c;
624              dequant_h263_inter = dequant_h263_inter_altivec_c;
625    
626                      dequant_mpeg_intra = dequant_mpeg_intra_altivec_c;
627                      dequant_mpeg_inter = dequant_mpeg_inter_altivec_c;
628    
629                      /* Qpel stuff */
630                      xvid_QP_Funcs = &xvid_QP_Funcs_Altivec_C;
631                      xvid_QP_Add_Funcs = &xvid_QP_Add_Funcs_Altivec_C;
632          }          }
633    #endif
634    
635          if ((cpu_flags & XVID_CPU_ALTIVEC))  #if defined(ARCH_IS_X86_64)
636          {          /* For now, only XVID_CPU_ASM is looked for, so user can still
637                  calc_cbp = calc_cbp_altivec;           * disable asm usage the usual way. When Intel EMT64 cpus will
638                  fdct = fdct_altivec;           * be out, maybe we'll have to check more precisely what cpu
639                  idct = idct_altivec;           * features there really are. */
640                  sadInit = sadInit_altivec;          if (cpu_flags & XVID_CPU_ASM) {
641                  sad16 = sad16_altivec;                  /* SIMD state flusher */
642                  sad8 = sad8_altivec;                  emms = emms_mmx;
643                  dev16 = dev16_altivec;  
644                    /* DCT operators */
645                    fdct = fdct_skal_x86_64;
646                    idct = idct_x86_64;
647    
648                    /* SAD operators */
649                    sad16      = sad16_x86_64;
650                    sad8       = sad8_x86_64;
651                    sad16bi    = sad16bi_x86_64;
652                    sad8bi     = sad8bi_x86_64;
653                    dev16      = dev16_x86_64;
654                    sad16v     = sad16v_x86_64;
655                    sse8_16bit = sse8_16bit_x86_64;
656                    sse8_8bit  = sse8_8bit_x86_64;
657    
658                    /* Interpolation operators */
659                    interpolate8x8_halfpel_h  = interpolate8x8_halfpel_h_x86_64;
660                    interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_x86_64;
661                    interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_x86_64;
662    
663                    interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_x86_64;
664                    interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_x86_64;
665                    interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_x86_64;
666                    interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_x86_64;
667    
668                    interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_x86_64;
669                    interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_x86_64;
670    
671                    interpolate8x8_avg2 = interpolate8x8_avg2_x86_64;
672                    interpolate8x8_avg4 = interpolate8x8_avg4_x86_64;
673    
674                    /* Quantization related functions */
675                    quant_h263_intra   = quant_h263_intra_x86_64;
676                    quant_h263_inter   = quant_h263_inter_x86_64;
677                    dequant_h263_intra = dequant_h263_intra_x86_64;
678                    dequant_h263_inter = dequant_h263_inter_x86_64;
679                    quant_mpeg_intra   = quant_mpeg_intra_x86_64;
680                    quant_mpeg_inter   = quant_mpeg_inter_x86_64;
681                    dequant_mpeg_intra   = dequant_mpeg_intra_x86_64;
682                    dequant_mpeg_inter   = dequant_mpeg_inter_x86_64;
683    
684                    /* Block related functions */
685                    transfer_8to16copy  = transfer_8to16copy_x86_64;
686                    transfer_16to8copy  = transfer_16to8copy_x86_64;
687                    transfer_8to16sub   = transfer_8to16sub_x86_64;
688                    transfer_8to16subro = transfer_8to16subro_x86_64;
689                    transfer_8to16sub2  = transfer_8to16sub2_x86_64;
690                    transfer_8to16sub2ro= transfer_8to16sub2ro_x86_64;
691                    transfer_16to8add   = transfer_16to8add_x86_64;
692                    transfer8x8_copy    = transfer8x8_copy_x86_64;
693    
694                    /* Qpel stuff */
695                    xvid_QP_Funcs = &xvid_QP_Funcs_x86_64;
696                    xvid_QP_Add_Funcs = &xvid_QP_Add_Funcs_x86_64;
697    
698                    /* Interlacing Functions */
699                    MBFieldTest = MBFieldTest_x86_64;
700          }          }
701  #endif  #endif
702    
# Line 551  Line 704 
704      xvid_debug = init->debug;      xvid_debug = init->debug;
705  #endif  #endif
706    
707      return 0;      return(0);
708  }  }
709    
710    
# Line 562  Line 715 
715                  return XVID_ERR_VERSION;                  return XVID_ERR_VERSION;
716    
717          info->actual_version = XVID_VERSION;          info->actual_version = XVID_VERSION;
718          info->build = "xvid-1.0.0";          info->build = "xvid-1.2.0-dev";
719          info->cpu_flags = detect_cpu_flags();          info->cpu_flags = detect_cpu_flags();
   
 #if defined(_SMP) && defined(WIN32)  
         info->num_threads = pthread_num_processors_np();;  
 #else  
720          info->num_threads = 0;          info->num_threads = 0;
721    
722    #if defined(WIN32)
723      {
724        DWORD dwProcessAffinityMask, dwSystemAffinityMask;
725        if (GetProcessAffinityMask(GetCurrentProcess(), &dwProcessAffinityMask, &dwSystemAffinityMask)) {
726          int i;
727          for(i=0; i<32; i++) {
728            if ((dwProcessAffinityMask & (1<<i)))
729              info->num_threads++;
730          }
731        }
732      }
733  #endif  #endif
734    
735          return 0;          return 0;
Legend:
Removed from v.1382  
changed lines
  Added in v.1709
No admin address has been configured
 ViewVC Help
Powered by ViewVC 1.0.4