[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 1.81, Tue Jun 2 13:06:49 2009 UTC
# Line 28  Line 28 
28  #include <string.h>  #include <string.h>
29  #include <time.h>  #include <time.h>
30    
31    #if !defined(_WIN32)
32    #include <unistd.h>
33    #endif
34    
35    #if defined(__APPLE__) && defined(__MACH__) && !defined(_SC_NPROCESSORS_CONF)
36    #include <sys/types.h>
37    #include <sys/sysctl.h>
38    #ifdef MAX
39    #undef MAX
40    #endif
41    #ifdef MIN
42    #undef MIN
43    #endif
44    #endif
45    
46  #include "xvid.h"  #include "xvid.h"
47  #include "decoder.h"  #include "decoder.h"
48  #include "encoder.h"  #include "encoder.h"
# Line 36  Line 51 
51  #include "dct/fdct.h"  #include "dct/fdct.h"
52  #include "image/colorspace.h"  #include "image/colorspace.h"
53  #include "image/interpolate8x8.h"  #include "image/interpolate8x8.h"
 #include "image/reduced.h"  
54  #include "utils/mem_transfer.h"  #include "utils/mem_transfer.h"
55  #include "utils/mbfunctions.h"  #include "utils/mbfunctions.h"
56  #include "quant/quant.h"  #include "quant/quant.h"
57  #include "motion/motion.h"  #include "motion/motion.h"
58    #include "motion/gmc.h"
59  #include "motion/sad.h"  #include "motion/sad.h"
60  #include "utils/emms.h"  #include "utils/emms.h"
61  #include "utils/timer.h"  #include "utils/timer.h"
# Line 52  Line 67 
67  unsigned int xvid_debug = 0; /* xvid debug mask */  unsigned int xvid_debug = 0; /* xvid debug mask */
68  #endif  #endif
69    
70  #if defined(ARCH_IS_IA32)  #if (defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)) && defined(_MSC_VER)
 #if defined(_MSC_VER)  
71  #       include <windows.h>  #       include <windows.h>
72  #else  #elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || defined(ARCH_IS_PPC)
73  #       include <signal.h>  #       include <signal.h>
74  #       include <setjmp.h>  #       include <setjmp.h>
75    
# Line 78  Line 92 
92   *   0 : SIGILL was *not* signalled   *   0 : SIGILL was *not* signalled
93   *   1 : SIGILL was signalled   *   1 : SIGILL was signalled
94   */   */
95    #if (defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)) && defined(_MSC_VER)
96  int  static int
97  sigill_check(void (*func)())  sigill_check(void (*func)())
98  {  {
 #if defined(_MSC_VER)  
99          _try {          _try {
100                  func();                  func();
101          }          } _except(EXCEPTION_EXECUTE_HANDLER) {
         _except(EXCEPTION_EXECUTE_HANDLER) {  
102    
103                  if (_exception_code() == STATUS_ILLEGAL_INSTRUCTION)                  if (_exception_code() == STATUS_ILLEGAL_INSTRUCTION)
104                          return 1;                          return(1);
105          }          }
106          return 0;          return(0);
107  #else  }
108    #elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || defined(ARCH_IS_PPC)
109    static int
110    sigill_check(void (*func)())
111    {
112      void * old_handler;      void * old_handler;
113      int jmpret;      int jmpret;
114    
115        /* Set our SIGILL handler */
116      old_handler = signal(SIGILL, sigill_handler);      old_handler = signal(SIGILL, sigill_handler);
117      if (old_handler == SIG_ERR)  
118      {      /* Check for error */
119          return -1;      if (old_handler == SIG_ERR) {
120            return(-1);
121      }      }
122    
123        /* Save stack context, so if func triggers a SIGILL, we can still roll
124             * back to a valid CPU state */
125      jmpret = setjmp(mark);      jmpret = setjmp(mark);
126      if (jmpret == 0)  
127      {          /* If setjmp returned directly, then its returned value is 0, and we still
128             * have to test the passed func. Otherwise it means the stack context has
129             * been restored by a longjmp() call, which in our case happens only in the
130             * signal handler */
131        if (jmpret == 0) {
132          func();          func();
133      }      }
134    
135        /* Restore old signal handler */
136      signal(SIGILL, old_handler);      signal(SIGILL, old_handler);
137    
138      return jmpret;      return(jmpret);
 #endif  
139  }  }
140  #endif  #endif
141    
142    
143  /* detect cpu flags  */  /* detect cpu flags  */
144  static unsigned int  static unsigned int
145  detect_cpu_flags()  detect_cpu_flags(void)
146  {  {
147          /* enable native assembly optimizations by default */          /* enable native assembly optimizations by default */
148          unsigned int cpu_flags = XVID_CPU_ASM;          unsigned int cpu_flags = XVID_CPU_ASM;
149    
150  #if defined(ARCH_IS_IA32)  #if defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)
151          cpu_flags |= check_cpu_features();          cpu_flags |= check_cpu_features();
152          if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger))          if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger))
153                  cpu_flags &= ~XVID_CPU_SSE;                  cpu_flags &= ~XVID_CPU_SSE;
154    
155          if ((cpu_flags & XVID_CPU_SSE2) && sigill_check(sse2_os_trigger))          if ((cpu_flags & (XVID_CPU_SSE2|XVID_CPU_SSE3|XVID_CPU_SSE41)) && sigill_check(sse2_os_trigger))
156                  cpu_flags &= ~XVID_CPU_SSE2;                  cpu_flags &= ~(XVID_CPU_SSE2|XVID_CPU_SSE3|XVID_CPU_SSE41);
157  #endif  #endif
158    
159  #if defined(ARCH_IS_PPC)  #if defined(ARCH_IS_PPC)
160  #if defined(ARCH_IS_PPC_ALTIVEC)          if (!sigill_check(altivec_trigger))
161          cpu_flags |= XVID_CPU_ALTIVEC;          cpu_flags |= XVID_CPU_ALTIVEC;
162  #endif  #endif
 #endif  
163    
164          return cpu_flags;          return cpu_flags;
165  }  }
# Line 177  Line 199 
199          idct = idct_int32;          idct = idct_int32;
200    
201          /* Only needed on PPC Altivec archs */          /* Only needed on PPC Altivec archs */
202          sadInit = 0;          sadInit = NULL;
203    
204          /* Restore FPU context : emms_c is a nop functions */          /* Restore FPU context : emms_c is a nop functions */
205          emms = emms_c;          emms = emms_c;
# Line 204  Line 226 
226          transfer_8to16sub  = transfer_8to16sub_c;          transfer_8to16sub  = transfer_8to16sub_c;
227          transfer_8to16subro  = transfer_8to16subro_c;          transfer_8to16subro  = transfer_8to16subro_c;
228          transfer_8to16sub2 = transfer_8to16sub2_c;          transfer_8to16sub2 = transfer_8to16sub2_c;
229            transfer_8to16sub2ro = transfer_8to16sub2ro_c;
230          transfer_16to8add  = transfer_16to8add_c;          transfer_16to8add  = transfer_16to8add_c;
231          transfer8x8_copy   = transfer8x8_copy_c;          transfer8x8_copy   = transfer8x8_copy_c;
232            transfer8x4_copy   = transfer8x4_copy_c;
233    
234          /* Interlacing functions */          /* Interlacing functions */
235          MBFieldTest = MBFieldTest_c;          MBFieldTest = MBFieldTest_c;
# Line 215  Line 239 
239          interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_c;          interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_c;
240          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;
241    
242            interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_c;
243            interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_c;
244            interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_c;
245    
246            interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_c;
247            interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_c;
248            interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_c;
249            interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_c;
250    
251          interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c;          interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c;
252          interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c;          interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c;
253          interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c;          interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c;
# Line 232  Line 265 
265          /* postprocessing */          /* postprocessing */
266          image_brightness = image_brightness_c;          image_brightness = image_brightness_c;
267    
         /* 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;  
   
268          /* Initialize internal colorspace transformation tables */          /* Initialize internal colorspace transformation tables */
269          colorspace_init();          colorspace_init();
270    
# Line 247  Line 272 
272          yv12_to_yv12    = yv12_to_yv12_c;          yv12_to_yv12    = yv12_to_yv12_c;
273          rgb555_to_yv12  = rgb555_to_yv12_c;          rgb555_to_yv12  = rgb555_to_yv12_c;
274          rgb565_to_yv12  = rgb565_to_yv12_c;          rgb565_to_yv12  = rgb565_to_yv12_c;
275            rgb_to_yv12     = rgb_to_yv12_c;
276          bgr_to_yv12     = bgr_to_yv12_c;          bgr_to_yv12     = bgr_to_yv12_c;
277          bgra_to_yv12    = bgra_to_yv12_c;          bgra_to_yv12    = bgra_to_yv12_c;
278          abgr_to_yv12    = abgr_to_yv12_c;          abgr_to_yv12    = abgr_to_yv12_c;
# Line 268  Line 294 
294          /* All colorspace transformation functions YV12->User format */          /* All colorspace transformation functions YV12->User format */
295          yv12_to_rgb555  = yv12_to_rgb555_c;          yv12_to_rgb555  = yv12_to_rgb555_c;
296          yv12_to_rgb565  = yv12_to_rgb565_c;          yv12_to_rgb565  = yv12_to_rgb565_c;
297            yv12_to_rgb     = yv12_to_rgb_c;
298          yv12_to_bgr     = yv12_to_bgr_c;          yv12_to_bgr     = yv12_to_bgr_c;
299          yv12_to_bgra    = yv12_to_bgra_c;          yv12_to_bgra    = yv12_to_bgra_c;
300          yv12_to_abgr    = yv12_to_abgr_c;          yv12_to_abgr    = yv12_to_abgr_c;
# Line 295  Line 322 
322          dev16      = dev16_c;          dev16      = dev16_c;
323          sad16v     = sad16v_c;          sad16v     = sad16v_c;
324          sse8_16bit = sse8_16bit_c;          sse8_16bit = sse8_16bit_c;
325            sse8_8bit  = sse8_8bit_c;
326    
327  #if defined(ARCH_IS_IA32)          init_GMC(cpu_flags);
328    
329          if ((cpu_flags & XVID_CPU_ASM)) {  #if defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)
                 vfilter_31 = xvid_VFilter_31_x86;  
                 hfilter_31 = xvid_HFilter_31_x86;  
         }  
330    
331          if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||          if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||
332                  (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||                  (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||
333                  (cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2))                  (cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2) ||
334            (cpu_flags & XVID_CPU_SSE3) || (cpu_flags & XVID_CPU_SSE41))
335          {          {
336                  /* Restore FPU context : emms_c is a nop functions */                  /* Restore FPU context : emms_c is a nop functions */
337                  emms = emms_mmx;                  emms = emms_mmx;
# Line 326  Line 352 
352                  quant_h263_inter   = quant_h263_inter_mmx;                  quant_h263_inter   = quant_h263_inter_mmx;
353                  dequant_h263_intra = dequant_h263_intra_mmx;                  dequant_h263_intra = dequant_h263_intra_mmx;
354                  dequant_h263_inter = dequant_h263_inter_mmx;                  dequant_h263_inter = dequant_h263_inter_mmx;
   
355                  quant_mpeg_intra   = quant_mpeg_intra_mmx;                  quant_mpeg_intra   = quant_mpeg_intra_mmx;
356                  quant_mpeg_inter   = quant_mpeg_inter_mmx;                  quant_mpeg_inter   = quant_mpeg_inter_mmx;
357                  dequant_mpeg_intra = dequant_mpeg_intra_mmx;                  dequant_mpeg_intra = dequant_mpeg_intra_mmx;
358                  dequant_mpeg_inter = dequant_mpeg_inter_mmx;                  dequant_mpeg_inter = dequant_mpeg_inter_mmx;
359    
360    
361                  /* Block related functions */                  /* Block related functions */
362                  transfer_8to16copy = transfer_8to16copy_mmx;                  transfer_8to16copy = transfer_8to16copy_mmx;
363                  transfer_16to8copy = transfer_16to8copy_mmx;                  transfer_16to8copy = transfer_16to8copy_mmx;
# Line 340  Line 366 
366                  transfer_8to16sub2 = transfer_8to16sub2_mmx;                  transfer_8to16sub2 = transfer_8to16sub2_mmx;
367                  transfer_16to8add  = transfer_16to8add_mmx;                  transfer_16to8add  = transfer_16to8add_mmx;
368                  transfer8x8_copy   = transfer8x8_copy_mmx;                  transfer8x8_copy   = transfer8x8_copy_mmx;
369                    transfer8x4_copy   = transfer8x4_copy_mmx;
370    
371                  /* Interlacing Functions */                  /* Interlacing Functions */
372                  MBFieldTest = MBFieldTest_mmx;                  MBFieldTest = MBFieldTest_mmx;
# Line 349  Line 376 
376                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_mmx;                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_mmx;
377                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;
378    
379                    interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_mmx;
380                    interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_mmx;
381                    interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_mmx;
382    
383                    interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_mmx;
384                    interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_mmx;
385                    interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_mmx;
386                    interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_mmx;
387    
388                  interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx;                  interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx;
389                  interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx;                  interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx;
390    
# Line 358  Line 394 
394                  /* postprocessing */                  /* postprocessing */
395                  image_brightness = image_brightness_mmx;                  image_brightness = image_brightness_mmx;
396    
                 /* 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;  
   
397                  /* image input xxx_to_yv12 related functions */                  /* image input xxx_to_yv12 related functions */
398    
399                  yv12_to_yv12  = yv12_to_yv12_mmx;                  yv12_to_yv12  = yv12_to_yv12_mmx;
400    
401                  bgr_to_yv12   = bgr_to_yv12_mmx;                  bgr_to_yv12   = bgr_to_yv12_mmx;
402                    rgb_to_yv12   = rgb_to_yv12_mmx;
403                  bgra_to_yv12  = bgra_to_yv12_mmx;                  bgra_to_yv12  = bgra_to_yv12_mmx;
404                    rgba_to_yv12  = rgba_to_yv12_mmx;
405                  yuyv_to_yv12  = yuyv_to_yv12_mmx;                  yuyv_to_yv12  = yuyv_to_yv12_mmx;
406                  uyvy_to_yv12  = uyvy_to_yv12_mmx;                  uyvy_to_yv12  = uyvy_to_yv12_mmx;
407    
# Line 390  Line 423 
423                  dev16    = dev16_mmx;                  dev16    = dev16_mmx;
424                  sad16v   = sad16v_mmx;                  sad16v   = sad16v_mmx;
425                  sse8_16bit = sse8_16bit_mmx;                  sse8_16bit = sse8_16bit_mmx;
426                    sse8_8bit  = sse8_8bit_mmx;
427          }          }
428    
429          /* these 3dnow functions are faster than mmx, but slower than xmm. */          /* these 3dnow functions are faster than mmx, but slower than xmm. */
# Line 403  Line 437 
437    
438                  yuyv_to_yv12  = yuyv_to_yv12_3dn;                  yuyv_to_yv12  = yuyv_to_yv12_3dn;
439                  uyvy_to_yv12  = uyvy_to_yv12_3dn;                  uyvy_to_yv12  = uyvy_to_yv12_3dn;
440    
441          }          }
442    
443    
# Line 417  Line 452 
452                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_xmm;                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_xmm;
453                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;
454    
455                  /* reduced resolution */                  interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_xmm;
456                  copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_xmm;                  interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_xmm;
457                  add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_xmm;                  interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_xmm;
458    
459                    interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_xmm;
460                    interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_xmm;
461                    interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_xmm;
462                    interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_xmm;
463    
464                  /* Quantization */                  /* Quantization */
                 quant_mpeg_intra = quant_mpeg_intra_xmm;  
465                  quant_mpeg_inter = quant_mpeg_inter_xmm;                  quant_mpeg_inter = quant_mpeg_inter_xmm;
466    
467                  dequant_h263_intra = dequant_h263_intra_xmm;                  dequant_h263_intra = dequant_h263_intra_xmm;
# Line 430  Line 469 
469    
470                  /* Buffer transfer */                  /* Buffer transfer */
471                  transfer_8to16sub2 = transfer_8to16sub2_xmm;                  transfer_8to16sub2 = transfer_8to16sub2_xmm;
472                    transfer_8to16sub2ro = transfer_8to16sub2ro_xmm;
473    
474                  /* Colorspace transformation */                  /* Colorspace transformation */
475                  yv12_to_yv12  = yv12_to_yv12_xmm;                  /* yv12_to_yv12  = yv12_to_yv12_xmm; */ /* appears to be slow on many machines */
476                  yuyv_to_yv12  = yuyv_to_yv12_xmm;                  yuyv_to_yv12  = yuyv_to_yv12_xmm;
477                  uyvy_to_yv12  = uyvy_to_yv12_xmm;                  uyvy_to_yv12  = uyvy_to_yv12_xmm;
478    
# Line 451  Line 491 
491                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn;                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn;
492                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn;                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn;
493                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;
494    
495                    interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_3dn;
496                    interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_3dn;
497                    interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_3dn;
498          }          }
499    
500          if ((cpu_flags & XVID_CPU_3DNOWEXT)) {          if ((cpu_flags & XVID_CPU_3DNOWEXT)) {
501    
                 /* Inverse DCT */  
                 idct =  idct_3dne;  
   
502                  /* Buffer transfer */                  /* Buffer transfer */
503                  transfer_8to16copy =  transfer_8to16copy_3dne;                  transfer_8to16copy =  transfer_8to16copy_3dne;
504                  transfer_16to8copy = transfer_16to8copy_3dne;                  transfer_16to8copy = transfer_16to8copy_3dne;
505                  transfer_8to16sub =  transfer_8to16sub_3dne;                  transfer_8to16sub =  transfer_8to16sub_3dne;
506                  transfer_8to16subro =  transfer_8to16subro_3dne;                  transfer_8to16subro =  transfer_8to16subro_3dne;
                 transfer_8to16sub2 =  transfer_8to16sub2_3dne;  
507                  transfer_16to8add = transfer_16to8add_3dne;                  transfer_16to8add = transfer_16to8add_3dne;
508                  transfer8x8_copy = transfer8x8_copy_3dne;                  transfer8x8_copy = transfer8x8_copy_3dne;
509                    transfer8x4_copy = transfer8x4_copy_3dne;
510    
511                    if ((cpu_flags & XVID_CPU_MMXEXT)) {
512                            /* Inverse DCT */
513                            idct =  idct_3dne;
514    
515                            /* Buffer transfer */
516                            transfer_8to16sub2 =  transfer_8to16sub2_3dne;
517    
518                            /* Interpolation */
519                            interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dne;
520                            interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne;
521                            interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne;
522    
523                            interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_3dne;
524                            interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_3dne;
525                            interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_3dne;
526    
527                  /* Quantization */                  /* Quantization */
528                  quant_h263_intra = quant_h263_intra_3dne;                          quant_h263_intra = quant_h263_intra_3dne;               /* cmov only */
529                  quant_h263_inter = quant_h263_inter_3dne;                  quant_h263_inter = quant_h263_inter_3dne;
530                  dequant_mpeg_intra = dequant_mpeg_intra_3dne;                          dequant_mpeg_intra = dequant_mpeg_intra_3dne;   /* cmov only */
531                  dequant_mpeg_inter = dequant_mpeg_inter_3dne;                  dequant_mpeg_inter = dequant_mpeg_inter_3dne;
532                  dequant_h263_intra = dequant_h263_intra_3dne;                  dequant_h263_intra = dequant_h263_intra_3dne;
533                  dequant_h263_inter = dequant_h263_inter_3dne;                  dequant_h263_inter = dequant_h263_inter_3dne;
534    
535                  /* ME functions */                  /* ME functions */
                 calc_cbp = calc_cbp_3dne;  
536                  sad16 = sad16_3dne;                  sad16 = sad16_3dne;
537                  sad8 = sad8_3dne;                  sad8 = sad8_3dne;
538                  sad16bi = sad16bi_3dne;                  sad16bi = sad16bi_3dne;
539                  sad8bi = sad8bi_3dne;                  sad8bi = sad8bi_3dne;
540                  dev16 = dev16_3dne;                  dev16 = dev16_3dne;
541                    }
                 /* Interpolation */  
                 interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dne;  
                 interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne;  
                 interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne;  
542          }          }
543    
544          if ((cpu_flags & XVID_CPU_SSE2)) {          if ((cpu_flags & XVID_CPU_SSE2)) {
# Line 503  Line 555 
555                  sad16    = sad16_sse2;                  sad16    = sad16_sse2;
556                  dev16    = dev16_sse2;                  dev16    = dev16_sse2;
557    
558                  /* DCT operators                  /* DCT operators */
                  * no iDCT because it's not "Walken matching" */  
559                  fdct = fdct_sse2_skal;                  fdct = fdct_sse2_skal;
560                    idct = idct_sse2_skal;   /* Is now IEEE1180 and Walken compliant. */
561    
562                    /* postprocessing */
563                    image_brightness = image_brightness_sse2;
564    
565          }          }
566    
567            if ((cpu_flags & XVID_CPU_SSE3)) {
568    
569                    /* SAD operators */
570                    sad16    = sad16_sse3;
571                    dev16    = dev16_sse3;
572            }
573    
574  #endif /* ARCH_IS_IA32 */  #endif /* ARCH_IS_IA32 */
575    
576  #if defined(ARCH_IS_IA64)  #if defined(ARCH_IS_IA64)
# Line 536  Line 600 
600  #endif  #endif
601    
602  #if defined(ARCH_IS_PPC)  #if defined(ARCH_IS_PPC)
603          if ((cpu_flags & XVID_CPU_ASM))          if ((cpu_flags & XVID_CPU_ALTIVEC)) {
604          {            /* sad operators */
605                  calc_cbp = calc_cbp_ppc;                    sad16 = sad16_altivec_c;
606          }                    sad16bi = sad16bi_altivec_c;
607                      sad8 = sad8_altivec_c;
608                      dev16 = dev16_altivec_c;
609    
610              sse8_16bit = sse8_16bit_altivec_c;
611    
612              /* mem transfer */
613              transfer_8to16copy = transfer_8to16copy_altivec_c;
614              transfer_16to8copy = transfer_16to8copy_altivec_c;
615              transfer_8to16sub = transfer_8to16sub_altivec_c;
616              transfer_8to16subro = transfer_8to16subro_altivec_c;
617              transfer_8to16sub2 = transfer_8to16sub2_altivec_c;
618              transfer_16to8add = transfer_16to8add_altivec_c;
619              transfer8x8_copy = transfer8x8_copy_altivec_c;
620    
621          if ((cpu_flags & XVID_CPU_ALTIVEC))            /* Inverse DCT */
622          {            idct = idct_altivec_c;
623                  calc_cbp = calc_cbp_altivec;  
624                  fdct = fdct_altivec;            /* Interpolation */
625                  idct = idct_altivec;            interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_altivec_c;
626                  sadInit = sadInit_altivec;            interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_altivec_c;
627                  sad16 = sad16_altivec;            interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_altivec_c;
628                  sad8 = sad8_altivec;  
629                  dev16 = dev16_altivec;            interpolate8x8_avg2 = interpolate8x8_avg2_altivec_c;
630              interpolate8x8_avg4 = interpolate8x8_avg4_altivec_c;
631    
632                      interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_altivec_c;
633                      interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_altivec_c;
634                      interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_altivec_c;
635                      interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_altivec_c;
636    
637              /* Colorspace conversion */
638              bgra_to_yv12 = bgra_to_yv12_altivec_c;
639              abgr_to_yv12 = abgr_to_yv12_altivec_c;
640              rgba_to_yv12 = rgba_to_yv12_altivec_c;
641              argb_to_yv12 = argb_to_yv12_altivec_c;
642    
643              yuyv_to_yv12 = yuyv_to_yv12_altivec_c;
644              uyvy_to_yv12 = uyvy_to_yv12_altivec_c;
645    
646              yv12_to_yuyv = yv12_to_yuyv_altivec_c;
647              yv12_to_uyvy = yv12_to_uyvy_altivec_c;
648    
649              /* Quantization */
650              quant_h263_intra = quant_h263_intra_altivec_c;
651              quant_h263_inter = quant_h263_inter_altivec_c;
652              dequant_h263_intra = dequant_h263_intra_altivec_c;
653              dequant_h263_inter = dequant_h263_inter_altivec_c;
654    
655                      dequant_mpeg_intra = dequant_mpeg_intra_altivec_c;
656                      dequant_mpeg_inter = dequant_mpeg_inter_altivec_c;
657    
658                      /* Qpel stuff */
659                      xvid_QP_Funcs = &xvid_QP_Funcs_Altivec_C;
660                      xvid_QP_Add_Funcs = &xvid_QP_Add_Funcs_Altivec_C;
661          }          }
662  #endif  #endif
663    
# Line 557  Line 665 
665      xvid_debug = init->debug;      xvid_debug = init->debug;
666  #endif  #endif
667    
668      return 0;      return(0);
669  }  }
670    
671    
# Line 568  Line 676 
676                  return XVID_ERR_VERSION;                  return XVID_ERR_VERSION;
677    
678          info->actual_version = XVID_VERSION;          info->actual_version = XVID_VERSION;
679          info->build = "xvid-1.0.0";          info->build = "xvid-1.3.0-dev";
680          info->cpu_flags = detect_cpu_flags();          info->cpu_flags = detect_cpu_flags();
681            info->num_threads = 0; /* single-thread */
682    
683    #if defined(_WIN32)
684    
685      {
686            SYSTEM_INFO siSysInfo;
687            GetSystemInfo(&siSysInfo);
688            info->num_threads = siSysInfo.dwNumberOfProcessors; /* number of _logical_ cores */
689      }
690    
691    #elif defined(_SC_NPROCESSORS_CONF) /* should be available on Apple too actually */
692    
693      info->num_threads = sysconf(_SC_NPROCESSORS_CONF);
694    
695    #elif defined(__APPLE__) && defined(__MACH__)
696    
697      {
698        size_t len;
699        int    mib[2], ncpu;
700    
701        mib[0] = CTL_HW;
702        mib[1] = HW_NCPU;
703        len    = sizeof(ncpu);
704        if (sysctl(mib, 2, &ncpu, &len, NULL, 0) == 0)
705          info -> num_threads = ncpu;
706        else
707          info -> num_threads = 1;
708      }
709    
 #if defined(_SMP) && defined(WIN32)  
         info->num_threads = pthread_num_processors_np();;  
 #else  
         info->num_threads = 0;  
710  #endif  #endif
711    
712          return 0;          return 0;
Legend:
Removed from v.1.49  
changed lines
  Added in v.1.81
No admin address has been configured
 ViewVC Help
Powered by ViewVC 1.0.4