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

Diff of /xvidcore/src/xvid.c

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

revision 1.49, Thu Apr 1 11:11:28 2004 UTC revision 1.68, Sat Feb 25 04:41:12 2006 UTC
# 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"
# Line 52  Line 51 
51  unsigned int xvid_debug = 0; /* xvid debug mask */  unsigned int xvid_debug = 0; /* xvid debug mask */
52  #endif  #endif
53    
54  #if defined(ARCH_IS_IA32)  #if (defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)) && defined(_MSC_VER)
 #if defined(_MSC_VER)  
55  #       include <windows.h>  #       include <windows.h>
56  #else  #elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || defined(ARCH_IS_PPC)
57  #       include <signal.h>  #       include <signal.h>
58  #       include <setjmp.h>  #       include <setjmp.h>
59    
# Line 78  Line 76 
76   *   0 : SIGILL was *not* signalled   *   0 : SIGILL was *not* signalled
77   *   1 : SIGILL was signalled   *   1 : SIGILL was signalled
78   */   */
79    #if (defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)) && defined(_MSC_VER)
80  int  static int
81  sigill_check(void (*func)())  sigill_check(void (*func)())
82  {  {
 #if defined(_MSC_VER)  
83          _try {          _try {
84                  func();                  func();
85          }          } _except(EXCEPTION_EXECUTE_HANDLER) {
         _except(EXCEPTION_EXECUTE_HANDLER) {  
86    
87                  if (_exception_code() == STATUS_ILLEGAL_INSTRUCTION)                  if (_exception_code() == STATUS_ILLEGAL_INSTRUCTION)
88                          return 1;                          return(1);
89          }          }
90          return 0;          return(0);
91  #else  }
92    #elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || defined(ARCH_IS_PPC)
93    static int
94    sigill_check(void (*func)())
95    {
96      void * old_handler;      void * old_handler;
97      int jmpret;      int jmpret;
98    
99        /* Set our SIGILL handler */
100      old_handler = signal(SIGILL, sigill_handler);      old_handler = signal(SIGILL, sigill_handler);
101      if (old_handler == SIG_ERR)  
102      {      /* Check for error */
103          return -1;      if (old_handler == SIG_ERR) {
104            return(-1);
105      }      }
106    
107        /* Save stack context, so if func triggers a SIGILL, we can still roll
108             * back to a valid CPU state */
109      jmpret = setjmp(mark);      jmpret = setjmp(mark);
110      if (jmpret == 0)  
111      {          /* If setjmp returned directly, then its returned value is 0, and we still
112             * have to test the passed func. Otherwise it means the stack context has
113             * been restored by a longjmp() call, which in our case happens only in the
114             * signal handler */
115        if (jmpret == 0) {
116          func();          func();
117      }      }
118    
119        /* Restore old signal handler */
120      signal(SIGILL, old_handler);      signal(SIGILL, old_handler);
121    
122      return jmpret;      return(jmpret);
 #endif  
123  }  }
124  #endif  #endif
125    
126    
127  /* detect cpu flags  */  /* detect cpu flags  */
128  static unsigned int  static unsigned int
129  detect_cpu_flags()  detect_cpu_flags(void)
130  {  {
131          /* enable native assembly optimizations by default */          /* enable native assembly optimizations by default */
132          unsigned int cpu_flags = XVID_CPU_ASM;          unsigned int cpu_flags = XVID_CPU_ASM;
133    
134  #if defined(ARCH_IS_IA32)  #if defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)
135          cpu_flags |= check_cpu_features();          cpu_flags |= check_cpu_features();
136          if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger))          if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger))
137                  cpu_flags &= ~XVID_CPU_SSE;                  cpu_flags &= ~XVID_CPU_SSE;
# Line 134  Line 141 
141  #endif  #endif
142    
143  #if defined(ARCH_IS_PPC)  #if defined(ARCH_IS_PPC)
144  #if defined(ARCH_IS_PPC_ALTIVEC)          if (!sigill_check(altivec_trigger))
145          cpu_flags |= XVID_CPU_ALTIVEC;          cpu_flags |= XVID_CPU_ALTIVEC;
146  #endif  #endif
 #endif  
147    
148          return cpu_flags;          return cpu_flags;
149  }  }
# Line 177  Line 183 
183          idct = idct_int32;          idct = idct_int32;
184    
185          /* Only needed on PPC Altivec archs */          /* Only needed on PPC Altivec archs */
186          sadInit = 0;          sadInit = NULL;
187    
188          /* Restore FPU context : emms_c is a nop functions */          /* Restore FPU context : emms_c is a nop functions */
189          emms = emms_c;          emms = emms_c;
# Line 204  Line 210 
210          transfer_8to16sub  = transfer_8to16sub_c;          transfer_8to16sub  = transfer_8to16sub_c;
211          transfer_8to16subro  = transfer_8to16subro_c;          transfer_8to16subro  = transfer_8to16subro_c;
212          transfer_8to16sub2 = transfer_8to16sub2_c;          transfer_8to16sub2 = transfer_8to16sub2_c;
213            transfer_8to16sub2ro = transfer_8to16sub2ro_c;
214          transfer_16to8add  = transfer_16to8add_c;          transfer_16to8add  = transfer_16to8add_c;
215          transfer8x8_copy   = transfer8x8_copy_c;          transfer8x8_copy   = transfer8x8_copy_c;
216            transfer8x4_copy   = transfer8x4_copy_c;
217    
218          /* Interlacing functions */          /* Interlacing functions */
219          MBFieldTest = MBFieldTest_c;          MBFieldTest = MBFieldTest_c;
# Line 215  Line 223 
223          interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_c;          interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_c;
224          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;
225    
226            interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_c;
227            interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_c;
228            interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_c;
229    
230            interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_c;
231            interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_c;
232            interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_c;
233            interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_c;
234    
235          interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c;          interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c;
236          interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c;          interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c;
237          interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c;          interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c;
# Line 232  Line 249 
249          /* postprocessing */          /* postprocessing */
250          image_brightness = image_brightness_c;          image_brightness = image_brightness_c;
251    
         /* reduced resolution */  
         copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_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          /* Initialize internal colorspace transformation tables */          /* Initialize internal colorspace transformation tables */
253          colorspace_init();          colorspace_init();
254    
# Line 295  Line 304 
304          dev16      = dev16_c;          dev16      = dev16_c;
305          sad16v     = sad16v_c;          sad16v     = sad16v_c;
306          sse8_16bit = sse8_16bit_c;          sse8_16bit = sse8_16bit_c;
307            sse8_8bit  = sse8_8bit_c;
308    
309  #if defined(ARCH_IS_IA32)  #if defined(ARCH_IS_IA32)
310    
         if ((cpu_flags & XVID_CPU_ASM)) {  
                 vfilter_31 = xvid_VFilter_31_x86;  
                 hfilter_31 = xvid_HFilter_31_x86;  
         }  
   
311          if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||          if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||
312                  (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||                  (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||
313                  (cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2))                  (cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2))
# Line 340  Line 345 
345                  transfer_8to16sub2 = transfer_8to16sub2_mmx;                  transfer_8to16sub2 = transfer_8to16sub2_mmx;
346                  transfer_16to8add  = transfer_16to8add_mmx;                  transfer_16to8add  = transfer_16to8add_mmx;
347                  transfer8x8_copy   = transfer8x8_copy_mmx;                  transfer8x8_copy   = transfer8x8_copy_mmx;
348                    transfer8x4_copy   = transfer8x4_copy_mmx;
349    
350                  /* Interlacing Functions */                  /* Interlacing Functions */
351                  MBFieldTest = MBFieldTest_mmx;                  MBFieldTest = MBFieldTest_mmx;
# Line 349  Line 355 
355                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_mmx;                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_mmx;
356                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;
357    
358                    interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_mmx;
359                    interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_mmx;
360                    interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_mmx;
361    
362                    interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_mmx;
363                    interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_mmx;
364                    interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_mmx;
365                    interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_mmx;
366    
367                  interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx;                  interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx;
368                  interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx;                  interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx;
369    
# Line 358  Line 373 
373                  /* postprocessing */                  /* postprocessing */
374                  image_brightness = image_brightness_mmx;                  image_brightness = image_brightness_mmx;
375    
                 /* reduced resolution */  
                 copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_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;  
   
376                  /* image input xxx_to_yv12 related functions */                  /* image input xxx_to_yv12 related functions */
377                  yv12_to_yv12  = yv12_to_yv12_mmx;                  yv12_to_yv12  = yv12_to_yv12_mmx;
378                  bgr_to_yv12   = bgr_to_yv12_mmx;                  bgr_to_yv12   = bgr_to_yv12_mmx;
# Line 390  Line 398 
398                  dev16    = dev16_mmx;                  dev16    = dev16_mmx;
399                  sad16v   = sad16v_mmx;                  sad16v   = sad16v_mmx;
400                  sse8_16bit = sse8_16bit_mmx;                  sse8_16bit = sse8_16bit_mmx;
401                    sse8_8bit  = sse8_8bit_mmx;
402          }          }
403    
404          /* these 3dnow functions are faster than mmx, but slower than xmm. */          /* these 3dnow functions are faster than mmx, but slower than xmm. */
# Line 417  Line 426 
426                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_xmm;                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_xmm;
427                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;
428    
429                  /* reduced resolution */                  interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_xmm;
430                  copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_xmm;                  interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_xmm;
431                  add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_xmm;                  interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_xmm;
432    
433                    interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_xmm;
434                    interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_xmm;
435                    interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_xmm;
436                    interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_xmm;
437    
438                  /* Quantization */                  /* Quantization */
439                  quant_mpeg_intra = quant_mpeg_intra_xmm;                  quant_mpeg_intra = quant_mpeg_intra_xmm;
# Line 430  Line 444 
444    
445                  /* Buffer transfer */                  /* Buffer transfer */
446                  transfer_8to16sub2 = transfer_8to16sub2_xmm;                  transfer_8to16sub2 = transfer_8to16sub2_xmm;
447                    transfer_8to16sub2ro = transfer_8to16sub2ro_xmm;
448    
449                  /* Colorspace transformation */                  /* Colorspace transformation */
450                  yv12_to_yv12  = yv12_to_yv12_xmm;                  yv12_to_yv12  = yv12_to_yv12_xmm;
# Line 451  Line 466 
466                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn;                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn;
467                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn;                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn;
468                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;
469    
470                    interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_3dn;
471                    interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_3dn;
472                    interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_3dn;
473          }          }
474    
475          if ((cpu_flags & XVID_CPU_3DNOWEXT)) {          if ((cpu_flags & XVID_CPU_3DNOWEXT)) {
476    
                 /* Inverse DCT */  
                 idct =  idct_3dne;  
   
477                  /* Buffer transfer */                  /* Buffer transfer */
478                  transfer_8to16copy =  transfer_8to16copy_3dne;                  transfer_8to16copy =  transfer_8to16copy_3dne;
479                  transfer_16to8copy = transfer_16to8copy_3dne;                  transfer_16to8copy = transfer_16to8copy_3dne;
480                  transfer_8to16sub =  transfer_8to16sub_3dne;                  transfer_8to16sub =  transfer_8to16sub_3dne;
481                  transfer_8to16subro =  transfer_8to16subro_3dne;                  transfer_8to16subro =  transfer_8to16subro_3dne;
                 transfer_8to16sub2 =  transfer_8to16sub2_3dne;  
482                  transfer_16to8add = transfer_16to8add_3dne;                  transfer_16to8add = transfer_16to8add_3dne;
483                  transfer8x8_copy = transfer8x8_copy_3dne;                  transfer8x8_copy = transfer8x8_copy_3dne;
484                    transfer8x4_copy = transfer8x4_copy_3dne;
485    
486                    if ((cpu_flags & XVID_CPU_MMXEXT)) {
487                            /* Inverse DCT */
488                            idct =  idct_3dne;
489    
490                            /* Buffer transfer */
491                            transfer_8to16sub2 =  transfer_8to16sub2_3dne;
492    
493                            /* Interpolation */
494                            interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dne;
495                            interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne;
496                            interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne;
497    
498                            interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_3dne;
499                            interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_3dne;
500                            interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_3dne;
501    
502                  /* Quantization */                  /* Quantization */
503                  quant_h263_intra = quant_h263_intra_3dne;                          quant_h263_intra = quant_h263_intra_3dne;               /* cmov only */
504                  quant_h263_inter = quant_h263_inter_3dne;                  quant_h263_inter = quant_h263_inter_3dne;
505                  dequant_mpeg_intra = dequant_mpeg_intra_3dne;                          dequant_mpeg_intra = dequant_mpeg_intra_3dne;   /* cmov only */
506                  dequant_mpeg_inter = dequant_mpeg_inter_3dne;                  dequant_mpeg_inter = dequant_mpeg_inter_3dne;
507                  dequant_h263_intra = dequant_h263_intra_3dne;                  dequant_h263_intra = dequant_h263_intra_3dne;
508                  dequant_h263_inter = dequant_h263_inter_3dne;                  dequant_h263_inter = dequant_h263_inter_3dne;
509    
510                  /* ME functions */                  /* ME functions */
511                  calc_cbp = calc_cbp_3dne;                  calc_cbp = calc_cbp_3dne;
512    
513                  sad16 = sad16_3dne;                  sad16 = sad16_3dne;
514                  sad8 = sad8_3dne;                  sad8 = sad8_3dne;
515                  sad16bi = sad16bi_3dne;                  sad16bi = sad16bi_3dne;
516                  sad8bi = sad8bi_3dne;                  sad8bi = sad8bi_3dne;
517                  dev16 = dev16_3dne;                  dev16 = dev16_3dne;
518                    }
                 /* Interpolation */  
                 interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dne;  
                 interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne;  
                 interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne;  
519          }          }
520    
521          if ((cpu_flags & XVID_CPU_SSE2)) {          if ((cpu_flags & XVID_CPU_SSE2)) {
# Line 503  Line 532 
532                  sad16    = sad16_sse2;                  sad16    = sad16_sse2;
533                  dev16    = dev16_sse2;                  dev16    = dev16_sse2;
534    
535                  /* DCT operators                  /* DCT operators */
                  * no iDCT because it's not "Walken matching" */  
536                  fdct = fdct_sse2_skal;                  fdct = fdct_sse2_skal;
537        /* idct = idct_sse2_skal; */   /* Is now IEEE1180 and Walken compliant. Disabled until fully tested. */
538    
539                    /* postprocessing */
540                    image_brightness = image_brightness_sse2;
541          }          }
542  #endif /* ARCH_IS_IA32 */  #endif /* ARCH_IS_IA32 */
543    
# Line 536  Line 568 
568  #endif  #endif
569    
570  #if defined(ARCH_IS_PPC)  #if defined(ARCH_IS_PPC)
571          if ((cpu_flags & XVID_CPU_ASM))          if ((cpu_flags & XVID_CPU_ALTIVEC)) {
572          {            /* sad operators */
573                  calc_cbp = calc_cbp_ppc;                    sad16 = sad16_altivec_c;
574                      sad16bi = sad16bi_altivec_c;
575                      sad8 = sad8_altivec_c;
576                      dev16 = dev16_altivec_c;
577    
578              sse8_16bit = sse8_16bit_altivec_c;
579    
580              /* mem transfer */
581              transfer_8to16copy = transfer_8to16copy_altivec_c;
582              transfer_16to8copy = transfer_16to8copy_altivec_c;
583              transfer_8to16sub = transfer_8to16sub_altivec_c;
584              transfer_8to16subro = transfer_8to16subro_altivec_c;
585              transfer_8to16sub2 = transfer_8to16sub2_altivec_c;
586              transfer_16to8add = transfer_16to8add_altivec_c;
587              transfer8x8_copy = transfer8x8_copy_altivec_c;
588    
589              /* Inverse DCT */
590              idct = idct_altivec_c;
591    
592              /* Interpolation */
593              interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_altivec_c;
594              interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_altivec_c;
595              interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_altivec_c;
596    
597              interpolate8x8_avg2 = interpolate8x8_avg2_altivec_c;
598              interpolate8x8_avg4 = interpolate8x8_avg4_altivec_c;
599    
600                      interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_altivec_c;
601                      interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_altivec_c;
602                      interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_altivec_c;
603                      interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_altivec_c;
604    
605              /* Colorspace conversion */
606              bgra_to_yv12 = bgra_to_yv12_altivec_c;
607              abgr_to_yv12 = abgr_to_yv12_altivec_c;
608              rgba_to_yv12 = rgba_to_yv12_altivec_c;
609              argb_to_yv12 = argb_to_yv12_altivec_c;
610    
611              yuyv_to_yv12 = yuyv_to_yv12_altivec_c;
612              uyvy_to_yv12 = uyvy_to_yv12_altivec_c;
613    
614              yv12_to_yuyv = yv12_to_yuyv_altivec_c;
615              yv12_to_uyvy = yv12_to_uyvy_altivec_c;
616    
617              /* Quantization */
618              quant_h263_intra = quant_h263_intra_altivec_c;
619              quant_h263_inter = quant_h263_inter_altivec_c;
620              dequant_h263_intra = dequant_h263_intra_altivec_c;
621              dequant_h263_inter = dequant_h263_inter_altivec_c;
622    
623                      dequant_mpeg_intra = dequant_mpeg_intra_altivec_c;
624                      dequant_mpeg_inter = dequant_mpeg_inter_altivec_c;
625    
626                      /* Qpel stuff */
627                      xvid_QP_Funcs = &xvid_QP_Funcs_Altivec_C;
628                      xvid_QP_Add_Funcs = &xvid_QP_Add_Funcs_Altivec_C;
629          }          }
630    #endif
631    
632          if ((cpu_flags & XVID_CPU_ALTIVEC))  #if defined(ARCH_IS_X86_64)
633          {          /* For now, only XVID_CPU_ASM is looked for, so user can still
634                  calc_cbp = calc_cbp_altivec;           * disable asm usage the usual way. When Intel EMT64 cpus will
635                  fdct = fdct_altivec;           * be out, maybe we'll have to check more precisely what cpu
636                  idct = idct_altivec;           * features there really are. */
637                  sadInit = sadInit_altivec;          if (cpu_flags & XVID_CPU_ASM) {
638                  sad16 = sad16_altivec;                  /* SIMD state flusher */
639                  sad8 = sad8_altivec;                  emms = emms_mmx;
640                  dev16 = dev16_altivec;  
641                    /* DCT operators */
642                    fdct = fdct_skal_x86_64;
643                    idct = idct_x86_64;
644    
645                    /* SAD operators */
646                    sad16      = sad16_x86_64;
647                    sad8       = sad8_x86_64;
648                    sad16bi    = sad16bi_x86_64;
649                    sad8bi     = sad8bi_x86_64;
650                    dev16      = dev16_x86_64;
651                    sad16v     = sad16v_x86_64;
652                    sse8_16bit = sse8_16bit_x86_64;
653                    sse8_8bit  = sse8_8bit_x86_64;
654    
655                    /* Interpolation operators */
656                    interpolate8x8_halfpel_h  = interpolate8x8_halfpel_h_x86_64;
657                    interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_x86_64;
658                    interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_x86_64;
659    
660                    interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_x86_64;
661                    interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_x86_64;
662                    interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_x86_64;
663                    interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_x86_64;
664    
665                    interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_x86_64;
666                    interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_x86_64;
667    
668                    interpolate8x8_avg2 = interpolate8x8_avg2_x86_64;
669                    interpolate8x8_avg4 = interpolate8x8_avg4_x86_64;
670    
671                    /* Quantization related functions */
672                    quant_h263_intra   = quant_h263_intra_x86_64;
673                    quant_h263_inter   = quant_h263_inter_x86_64;
674                    dequant_h263_intra = dequant_h263_intra_x86_64;
675                    dequant_h263_inter = dequant_h263_inter_x86_64;
676                    quant_mpeg_intra   = quant_mpeg_intra_x86_64;
677                    quant_mpeg_inter   = quant_mpeg_inter_x86_64;
678                    dequant_mpeg_intra   = dequant_mpeg_intra_x86_64;
679                    dequant_mpeg_inter   = dequant_mpeg_inter_x86_64;
680    
681                    /* Block related functions */
682                    transfer_8to16copy  = transfer_8to16copy_x86_64;
683                    transfer_16to8copy  = transfer_16to8copy_x86_64;
684                    transfer_8to16sub   = transfer_8to16sub_x86_64;
685                    transfer_8to16subro = transfer_8to16subro_x86_64;
686                    transfer_8to16sub2  = transfer_8to16sub2_x86_64;
687                    transfer_8to16sub2ro= transfer_8to16sub2ro_x86_64;
688                    transfer_16to8add   = transfer_16to8add_x86_64;
689                    transfer8x8_copy    = transfer8x8_copy_x86_64;
690    
691                    /* Qpel stuff */
692                    xvid_QP_Funcs = &xvid_QP_Funcs_x86_64;
693                    xvid_QP_Add_Funcs = &xvid_QP_Add_Funcs_x86_64;
694    
695                    /* Interlacing Functions */
696                    MBFieldTest = MBFieldTest_x86_64;
697          }          }
698  #endif  #endif
699    
# Line 557  Line 701 
701      xvid_debug = init->debug;      xvid_debug = init->debug;
702  #endif  #endif
703    
704      return 0;      return(0);
705  }  }
706    
707    
# Line 568  Line 712 
712                  return XVID_ERR_VERSION;                  return XVID_ERR_VERSION;
713    
714          info->actual_version = XVID_VERSION;          info->actual_version = XVID_VERSION;
715          info->build = "xvid-1.0.0";          info->build = "xvid-1.2.0-dev";
716          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  
717          info->num_threads = 0;          info->num_threads = 0;
718    
719    #if defined(WIN32)
720      {
721        DWORD dwProcessAffinityMask, dwSystemAffinityMask;
722        if (GetProcessAffinityMask(GetCurrentProcess(), &dwProcessAffinityMask, &dwSystemAffinityMask)) {
723          int i;
724          for(i=0; i<32; i++) {
725            if ((dwProcessAffinityMask & (1<<i)))
726              info->num_threads++;
727          }
728        }
729      }
730  #endif  #endif
731    
732          return 0;          return 0;

Legend:
Removed from v.1.49  
changed lines
  Added in v.1.68

No admin address has been configured
ViewVC Help
Powered by ViewVC 1.0.4