[svn] / branches / release-1_3-branch / xvidcore / src / xvid.c Repository:
ViewVC logotype

Diff of /branches/release-1_3-branch/xvidcore/src/xvid.c

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

trunk/xvidcore/src/xvid.c revision 1382, Mon Mar 22 22:36:25 2004 UTC branches/release-1_3-branch/xvidcore/src/xvid.c revision 2053, Fri Nov 4 15:23:46 2011 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) 2001-2011 Peter Ross <pross@xvid.org>
7     *               2002-2011 Michael Militzer <isibaar@xvid.org>
8   *   *
9   *  This program is free software ; you can redistribute it and/or modify   *  This program is free software ; you can redistribute it and/or modify
10   *  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 20 
20   *  along with this program ; if not, write to the Free Software   *  along with this program ; if not, write to the Free Software
21   *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA   *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
22   *   *
23   * $Id: xvid.c,v 1.48 2004-03-22 22:36:23 edgomez Exp $   * $Id$
24   *   *
25   ****************************************************************************/   ****************************************************************************/
26    
# Line 28  Line 29 
29  #include <string.h>  #include <string.h>
30  #include <time.h>  #include <time.h>
31    
32    #if !defined(_WIN32)
33    #include <unistd.h>
34    #endif
35    
36    #if defined(__APPLE__) && defined(__MACH__) && !defined(_SC_NPROCESSORS_CONF)
37    #include <sys/types.h>
38    #include <sys/sysctl.h>
39    #ifdef MAX
40    #undef MAX
41    #endif
42    #ifdef MIN
43    #undef MIN
44    #endif
45    #endif
46    
47    #if defined(__amigaos4__)
48    #include <exec/exec.h>
49    #include <proto/exec.h>
50    #endif
51    
52  #include "xvid.h"  #include "xvid.h"
53  #include "decoder.h"  #include "decoder.h"
54  #include "encoder.h"  #include "encoder.h"
# Line 36  Line 57 
57  #include "dct/fdct.h"  #include "dct/fdct.h"
58  #include "image/colorspace.h"  #include "image/colorspace.h"
59  #include "image/interpolate8x8.h"  #include "image/interpolate8x8.h"
 #include "image/reduced.h"  
60  #include "utils/mem_transfer.h"  #include "utils/mem_transfer.h"
61  #include "utils/mbfunctions.h"  #include "utils/mbfunctions.h"
62  #include "quant/quant.h"  #include "quant/quant.h"
63  #include "motion/motion.h"  #include "motion/motion.h"
64    #include "motion/gmc.h"
65  #include "motion/sad.h"  #include "motion/sad.h"
66  #include "utils/emms.h"  #include "utils/emms.h"
67  #include "utils/timer.h"  #include "utils/timer.h"
# Line 52  Line 73 
73  unsigned int xvid_debug = 0; /* xvid debug mask */  unsigned int xvid_debug = 0; /* xvid debug mask */
74  #endif  #endif
75    
76  #if defined(ARCH_IS_IA32)  #if (defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)) && defined(_MSC_VER)
 #if defined(_MSC_VER)  
77  #       include <windows.h>  #       include <windows.h>
78  #else  #elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || (defined(ARCH_IS_PPC) && !defined(__amigaos4__))
79  #       include <signal.h>  #       include <signal.h>
80  #       include <setjmp.h>  #       include <setjmp.h>
81    
# Line 68  Line 88 
88          }          }
89  #endif  #endif
90    
   
91  /*  /*
92   * Calls the funcptr, and returns whether SIGILL (illegal instruction) was   * Calls the funcptr, and returns whether SIGILL (illegal instruction) was
93   * signalled   * signalled
# Line 78  Line 97 
97   *   0 : SIGILL was *not* signalled   *   0 : SIGILL was *not* signalled
98   *   1 : SIGILL was signalled   *   1 : SIGILL was signalled
99   */   */
100    #if (defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)) && defined(_MSC_VER)
101  int  static int
102  sigill_check(void (*func)())  sigill_check(void (*func)())
103  {  {
 #if defined(_MSC_VER)  
104          _try {          _try {
105                  func();                  func();
106          }          } _except(EXCEPTION_EXECUTE_HANDLER) {
         _except(EXCEPTION_EXECUTE_HANDLER) {  
107    
108                  if (_exception_code() == STATUS_ILLEGAL_INSTRUCTION)                  if (_exception_code() == STATUS_ILLEGAL_INSTRUCTION)
109                          return 1;                          return(1);
110          }          }
111          return 0;          return(0);
112  #else  }
113    #elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || (defined(ARCH_IS_PPC) && !defined(__amigaos4__))
114    static int
115    sigill_check(void (*func)())
116    {
117      void * old_handler;      void * old_handler;
118      int jmpret;      int jmpret;
119    
120        /* Set our SIGILL handler */
121      old_handler = signal(SIGILL, sigill_handler);      old_handler = signal(SIGILL, sigill_handler);
122      if (old_handler == SIG_ERR)  
123      {      /* Check for error */
124          return -1;      if (old_handler == SIG_ERR) {
125            return(-1);
126      }      }
127    
128        /* Save stack context, so if func triggers a SIGILL, we can still roll
129             * back to a valid CPU state */
130      jmpret = setjmp(mark);      jmpret = setjmp(mark);
131      if (jmpret == 0)  
132      {          /* If setjmp returned directly, then its returned value is 0, and we still
133             * have to test the passed func. Otherwise it means the stack context has
134             * been restored by a longjmp() call, which in our case happens only in the
135             * signal handler */
136        if (jmpret == 0) {
137          func();          func();
138      }      }
139    
140        /* Restore old signal handler */
141      signal(SIGILL, old_handler);      signal(SIGILL, old_handler);
142    
143      return jmpret;      return(jmpret);
 #endif  
144  }  }
145  #endif  #endif
146    
147    
148  /* detect cpu flags  */  /* detect cpu flags  */
149  static unsigned int  static unsigned int
150  detect_cpu_flags()  detect_cpu_flags(void)
151  {  {
152          /* enable native assembly optimizations by default */          /* enable native assembly optimizations by default */
153          unsigned int cpu_flags = XVID_CPU_ASM;          unsigned int cpu_flags = XVID_CPU_ASM;
154    
155  #if defined(ARCH_IS_IA32)  #if defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)
156          cpu_flags |= check_cpu_features();          cpu_flags |= check_cpu_features();
157          if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger))          if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger))
158                  cpu_flags &= ~XVID_CPU_SSE;                  cpu_flags &= ~XVID_CPU_SSE;
159    
160          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))
161                  cpu_flags &= ~XVID_CPU_SSE2;                  cpu_flags &= ~(XVID_CPU_SSE2|XVID_CPU_SSE3|XVID_CPU_SSE41);
162  #endif  #endif
163    
164  #if defined(ARCH_IS_PPC)  #if defined(ARCH_IS_PPC)
165  #if defined(ARCH_IS_PPC_ALTIVEC)  #if defined(__amigaos4__)
166            {
167                    uint32_t vector_unit = VECTORTYPE_NONE;
168                    IExec->GetCPUInfoTags(GCIT_VectorUnit, &vector_unit, TAG_END);
169                    if (vector_unit == VECTORTYPE_ALTIVEC) {
170                            cpu_flags |= XVID_CPU_ALTIVEC;
171                    }
172            }
173    #else
174            if (!sigill_check(altivec_trigger))
175          cpu_flags |= XVID_CPU_ALTIVEC;          cpu_flags |= XVID_CPU_ALTIVEC;
176  #endif  #endif
177  #endif  #endif
# Line 144  Line 181 
181    
182    
183  /*****************************************************************************  /*****************************************************************************
184   * XviD Init Entry point   * Xvid Init Entry point
185   *   *
186   * Well this function initialize all internal function pointers according   * Well this function initialize all internal function pointers according
187   * to the CPU features forced by the library client or autodetected (depending   * to the CPU features forced by the library client or autodetected (depending
# Line 169  Line 206 
206          cpu_flags = (init->cpu_flags & XVID_CPU_FORCE) ? init->cpu_flags : detect_cpu_flags();          cpu_flags = (init->cpu_flags & XVID_CPU_FORCE) ? init->cpu_flags : detect_cpu_flags();
207    
208          /* Initialize the function pointers */          /* Initialize the function pointers */
         idct_int32_init();  
209          init_vlc_tables();          init_vlc_tables();
210    
211          /* Fixed Point Forward/Inverse DCT transformations */          /* Fixed Point Forward/Inverse DCT transformations */
# Line 177  Line 213 
213          idct = idct_int32;          idct = idct_int32;
214    
215          /* Only needed on PPC Altivec archs */          /* Only needed on PPC Altivec archs */
216          sadInit = 0;          sadInit = NULL;
217    
218          /* Restore FPU context : emms_c is a nop functions */          /* Restore FPU context : emms_c is a nop functions */
219          emms = emms_c;          emms = emms_c;
# Line 204  Line 240 
240          transfer_8to16sub  = transfer_8to16sub_c;          transfer_8to16sub  = transfer_8to16sub_c;
241          transfer_8to16subro  = transfer_8to16subro_c;          transfer_8to16subro  = transfer_8to16subro_c;
242          transfer_8to16sub2 = transfer_8to16sub2_c;          transfer_8to16sub2 = transfer_8to16sub2_c;
243            transfer_8to16sub2ro = transfer_8to16sub2ro_c;
244          transfer_16to8add  = transfer_16to8add_c;          transfer_16to8add  = transfer_16to8add_c;
245          transfer8x8_copy   = transfer8x8_copy_c;          transfer8x8_copy   = transfer8x8_copy_c;
246            transfer8x4_copy   = transfer8x4_copy_c;
247    
248          /* Interlacing functions */          /* Interlacing functions */
249          MBFieldTest = MBFieldTest_c;          MBFieldTest = MBFieldTest_c;
# Line 215  Line 253 
253          interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_c;          interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_c;
254          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;
255    
256            interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_c;
257            interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_c;
258            interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_c;
259    
260            interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_c;
261            interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_c;
262            interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_c;
263            interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_c;
264    
265          interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c;          interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c;
266          interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c;          interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c;
267          interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c;          interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c;
# Line 229  Line 276 
276          interpolate8x8_avg2 = interpolate8x8_avg2_c;          interpolate8x8_avg2 = interpolate8x8_avg2_c;
277          interpolate8x8_avg4 = interpolate8x8_avg4_c;          interpolate8x8_avg4 = interpolate8x8_avg4_c;
278    
279          /* reduced resolution */          /* postprocessing */
280          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;  
281    
282          /* Initialize internal colorspace transformation tables */          /* Initialize internal colorspace transformation tables */
283          colorspace_init();          colorspace_init();
# Line 244  Line 286 
286          yv12_to_yv12    = yv12_to_yv12_c;          yv12_to_yv12    = yv12_to_yv12_c;
287          rgb555_to_yv12  = rgb555_to_yv12_c;          rgb555_to_yv12  = rgb555_to_yv12_c;
288          rgb565_to_yv12  = rgb565_to_yv12_c;          rgb565_to_yv12  = rgb565_to_yv12_c;
289            rgb_to_yv12     = rgb_to_yv12_c;
290          bgr_to_yv12     = bgr_to_yv12_c;          bgr_to_yv12     = bgr_to_yv12_c;
291          bgra_to_yv12    = bgra_to_yv12_c;          bgra_to_yv12    = bgra_to_yv12_c;
292          abgr_to_yv12    = abgr_to_yv12_c;          abgr_to_yv12    = abgr_to_yv12_c;
# Line 265  Line 308 
308          /* All colorspace transformation functions YV12->User format */          /* All colorspace transformation functions YV12->User format */
309          yv12_to_rgb555  = yv12_to_rgb555_c;          yv12_to_rgb555  = yv12_to_rgb555_c;
310          yv12_to_rgb565  = yv12_to_rgb565_c;          yv12_to_rgb565  = yv12_to_rgb565_c;
311            yv12_to_rgb     = yv12_to_rgb_c;
312          yv12_to_bgr     = yv12_to_bgr_c;          yv12_to_bgr     = yv12_to_bgr_c;
313          yv12_to_bgra    = yv12_to_bgra_c;          yv12_to_bgra    = yv12_to_bgra_c;
314          yv12_to_abgr    = yv12_to_abgr_c;          yv12_to_abgr    = yv12_to_abgr_c;
# Line 292  Line 336 
336          dev16      = dev16_c;          dev16      = dev16_c;
337          sad16v     = sad16v_c;          sad16v     = sad16v_c;
338          sse8_16bit = sse8_16bit_c;          sse8_16bit = sse8_16bit_c;
339            sse8_8bit     = sse8_8bit_c;
340    
341  #if defined(ARCH_IS_IA32)          sseh8_16bit   = sseh8_16bit_c;
342            coeff8_energy = coeff8_energy_c;
343            blocksum8     = blocksum8_c;
344    
345          if ((cpu_flags & XVID_CPU_ASM)) {          init_GMC(cpu_flags);
346                  vfilter_31 = xvid_VFilter_31_x86;  
347                  hfilter_31 = xvid_HFilter_31_x86;  #if defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)
         }  
348    
349          if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||          if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||
350                  (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||                  (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||
351                  (cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2))                  (cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2) ||
352            (cpu_flags & XVID_CPU_SSE3) || (cpu_flags & XVID_CPU_SSE41))
353          {          {
354                  /* Restore FPU context : emms_c is a nop functions */                  /* Restore FPU context : emms_c is a nop functions */
355                  emms = emms_mmx;                  emms = emms_mmx;
# Line 323  Line 370 
370                  quant_h263_inter   = quant_h263_inter_mmx;                  quant_h263_inter   = quant_h263_inter_mmx;
371                  dequant_h263_intra = dequant_h263_intra_mmx;                  dequant_h263_intra = dequant_h263_intra_mmx;
372                  dequant_h263_inter = dequant_h263_inter_mmx;                  dequant_h263_inter = dequant_h263_inter_mmx;
   
373                  quant_mpeg_intra   = quant_mpeg_intra_mmx;                  quant_mpeg_intra   = quant_mpeg_intra_mmx;
374                  quant_mpeg_inter   = quant_mpeg_inter_mmx;                  quant_mpeg_inter   = quant_mpeg_inter_mmx;
375                  dequant_mpeg_intra = dequant_mpeg_intra_mmx;                  dequant_mpeg_intra = dequant_mpeg_intra_mmx;
376                  dequant_mpeg_inter = dequant_mpeg_inter_mmx;                  dequant_mpeg_inter = dequant_mpeg_inter_mmx;
377    
378    
379                  /* Block related functions */                  /* Block related functions */
380                  transfer_8to16copy = transfer_8to16copy_mmx;                  transfer_8to16copy = transfer_8to16copy_mmx;
381                  transfer_16to8copy = transfer_16to8copy_mmx;                  transfer_16to8copy = transfer_16to8copy_mmx;
# Line 337  Line 384 
384                  transfer_8to16sub2 = transfer_8to16sub2_mmx;                  transfer_8to16sub2 = transfer_8to16sub2_mmx;
385                  transfer_16to8add  = transfer_16to8add_mmx;                  transfer_16to8add  = transfer_16to8add_mmx;
386                  transfer8x8_copy   = transfer8x8_copy_mmx;                  transfer8x8_copy   = transfer8x8_copy_mmx;
387                    transfer8x4_copy   = transfer8x4_copy_mmx;
388    
389                  /* Interlacing Functions */                  /* Interlacing Functions */
390                  MBFieldTest = MBFieldTest_mmx;                  MBFieldTest = MBFieldTest_mmx;
# Line 346  Line 394 
394                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_mmx;                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_mmx;
395                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;
396    
397                    interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_mmx;
398                    interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_mmx;
399                    interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_mmx;
400    
401                    interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_mmx;
402                    interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_mmx;
403                    interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_mmx;
404                    interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_mmx;
405    
406                  interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx;                  interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx;
407                  interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx;                  interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx;
408    
409                  interpolate8x8_avg2 = interpolate8x8_avg2_mmx;                  interpolate8x8_avg2 = interpolate8x8_avg2_mmx;
410                  interpolate8x8_avg4 = interpolate8x8_avg4_mmx;                  interpolate8x8_avg4 = interpolate8x8_avg4_mmx;
411    
412                  /* reduced resolution */                  /* postprocessing */
413                  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;  
414    
415                  /* image input xxx_to_yv12 related functions */                  /* image input xxx_to_yv12 related functions */
416    
417                  yv12_to_yv12  = yv12_to_yv12_mmx;                  yv12_to_yv12  = yv12_to_yv12_mmx;
418    
419                  bgr_to_yv12   = bgr_to_yv12_mmx;                  bgr_to_yv12   = bgr_to_yv12_mmx;
420                    rgb_to_yv12   = rgb_to_yv12_mmx;
421                  bgra_to_yv12  = bgra_to_yv12_mmx;                  bgra_to_yv12  = bgra_to_yv12_mmx;
422                    rgba_to_yv12  = rgba_to_yv12_mmx;
423                  yuyv_to_yv12  = yuyv_to_yv12_mmx;                  yuyv_to_yv12  = yuyv_to_yv12_mmx;
424                  uyvy_to_yv12  = uyvy_to_yv12_mmx;                  uyvy_to_yv12  = uyvy_to_yv12_mmx;
425    
# Line 384  Line 441 
441                  dev16    = dev16_mmx;                  dev16    = dev16_mmx;
442                  sad16v   = sad16v_mmx;                  sad16v   = sad16v_mmx;
443                  sse8_16bit = sse8_16bit_mmx;                  sse8_16bit = sse8_16bit_mmx;
444                    sse8_8bit  = sse8_8bit_mmx;
445          }          }
446    
447          /* these 3dnow functions are faster than mmx, but slower than xmm. */          /* these 3dnow functions are faster than mmx, but slower than xmm. */
# Line 397  Line 455 
455    
456                  yuyv_to_yv12  = yuyv_to_yv12_3dn;                  yuyv_to_yv12  = yuyv_to_yv12_3dn;
457                  uyvy_to_yv12  = uyvy_to_yv12_3dn;                  uyvy_to_yv12  = uyvy_to_yv12_3dn;
458    
459          }          }
460    
461    
# Line 411  Line 470 
470                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_xmm;                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_xmm;
471                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;
472    
473                  /* reduced resolution */                  interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_xmm;
474                  copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_xmm;                  interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_xmm;
475                  add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_xmm;                  interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_xmm;
476    
477                    interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_xmm;
478                    interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_xmm;
479                    interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_xmm;
480                    interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_xmm;
481    
482                  /* Quantization */                  /* Quantization */
                 quant_mpeg_intra = quant_mpeg_intra_xmm;  
483                  quant_mpeg_inter = quant_mpeg_inter_xmm;                  quant_mpeg_inter = quant_mpeg_inter_xmm;
484    
485                  dequant_h263_intra = dequant_h263_intra_xmm;                  dequant_h263_intra = dequant_h263_intra_xmm;
# Line 424  Line 487 
487    
488                  /* Buffer transfer */                  /* Buffer transfer */
489                  transfer_8to16sub2 = transfer_8to16sub2_xmm;                  transfer_8to16sub2 = transfer_8to16sub2_xmm;
490                    transfer_8to16sub2ro = transfer_8to16sub2ro_xmm;
491    
492                  /* Colorspace transformation */                  /* Colorspace transformation */
493                  yv12_to_yv12  = yv12_to_yv12_xmm;                  /* yv12_to_yv12  = yv12_to_yv12_xmm; */ /* appears to be slow on many machines */
494                  yuyv_to_yv12  = yuyv_to_yv12_xmm;                  yuyv_to_yv12  = yuyv_to_yv12_xmm;
495                  uyvy_to_yv12  = uyvy_to_yv12_xmm;                  uyvy_to_yv12  = uyvy_to_yv12_xmm;
496    
# Line 445  Line 509 
509                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn;                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn;
510                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn;                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn;
511                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;
512    
513                    interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_3dn;
514                    interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_3dn;
515                    interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_3dn;
516          }          }
517    
518          if ((cpu_flags & XVID_CPU_3DNOWEXT)) {          if ((cpu_flags & XVID_CPU_3DNOWEXT)) {
519    
                 /* Inverse DCT */  
                 idct =  idct_3dne;  
   
520                  /* Buffer transfer */                  /* Buffer transfer */
521                  transfer_8to16copy =  transfer_8to16copy_3dne;                  transfer_8to16copy =  transfer_8to16copy_3dne;
522                  transfer_16to8copy = transfer_16to8copy_3dne;                  transfer_16to8copy = transfer_16to8copy_3dne;
523                  transfer_8to16sub =  transfer_8to16sub_3dne;                  transfer_8to16sub =  transfer_8to16sub_3dne;
524                  transfer_8to16subro =  transfer_8to16subro_3dne;                  transfer_8to16subro =  transfer_8to16subro_3dne;
                 transfer_8to16sub2 =  transfer_8to16sub2_3dne;  
525                  transfer_16to8add = transfer_16to8add_3dne;                  transfer_16to8add = transfer_16to8add_3dne;
526                  transfer8x8_copy = transfer8x8_copy_3dne;                  transfer8x8_copy = transfer8x8_copy_3dne;
527                    transfer8x4_copy = transfer8x4_copy_3dne;
528    
529                    if ((cpu_flags & XVID_CPU_MMXEXT)) {
530                            /* Inverse DCT */
531                            idct =  idct_3dne;
532    
533                            /* Buffer transfer */
534                            transfer_8to16sub2 =  transfer_8to16sub2_3dne;
535    
536                            /* Interpolation */
537                            interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dne;
538                            interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne;
539                            interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne;
540    
541                            interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_3dne;
542                            interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_3dne;
543                            interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_3dne;
544    
545                  /* Quantization */                  /* Quantization */
546                  quant_h263_intra = quant_h263_intra_3dne;                          quant_h263_intra = quant_h263_intra_3dne;               /* cmov only */
547                  quant_h263_inter = quant_h263_inter_3dne;                  quant_h263_inter = quant_h263_inter_3dne;
548                  dequant_mpeg_intra = dequant_mpeg_intra_3dne;                          dequant_mpeg_intra = dequant_mpeg_intra_3dne;   /* cmov only */
549                  dequant_mpeg_inter = dequant_mpeg_inter_3dne;                  dequant_mpeg_inter = dequant_mpeg_inter_3dne;
550                  dequant_h263_intra = dequant_h263_intra_3dne;                  dequant_h263_intra = dequant_h263_intra_3dne;
551                  dequant_h263_inter = dequant_h263_inter_3dne;                  dequant_h263_inter = dequant_h263_inter_3dne;
552    
553                  /* ME functions */                  /* ME functions */
                 calc_cbp = calc_cbp_3dne;  
554                  sad16 = sad16_3dne;                  sad16 = sad16_3dne;
555                  sad8 = sad8_3dne;                  sad8 = sad8_3dne;
556                  sad16bi = sad16bi_3dne;                  sad16bi = sad16bi_3dne;
557                  sad8bi = sad8bi_3dne;                  sad8bi = sad8bi_3dne;
558                  dev16 = dev16_3dne;                  dev16 = dev16_3dne;
559                    }
                 /* Interpolation */  
                 interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dne;  
                 interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne;  
                 interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne;  
560          }          }
561    
562          if ((cpu_flags & XVID_CPU_SSE2)) {          if ((cpu_flags & XVID_CPU_SSE2)) {
# Line 497  Line 573 
573                  sad16    = sad16_sse2;                  sad16    = sad16_sse2;
574                  dev16    = dev16_sse2;                  dev16    = dev16_sse2;
575    
576                  /* DCT operators                  /* PSNR-HVS-M distortion metric */
577                   * no iDCT because it's not "Walken matching" */                  sseh8_16bit   = sseh8_16bit_sse2;
578                    coeff8_energy = coeff8_energy_sse2;
579                    blocksum8     = blocksum8_sse2;
580    
581                    /* DCT operators */
582                  fdct = fdct_sse2_skal;                  fdct = fdct_sse2_skal;
583                    idct = idct_sse2_skal;   /* Is now IEEE1180 and Walken compliant. */
584    
585                    /* postprocessing */
586                    image_brightness = image_brightness_sse2;
587    
588          }          }
589    
590            if ((cpu_flags & XVID_CPU_SSE3)) {
591    
592                    /* SAD operators */
593                    sad16    = sad16_sse3;
594                    dev16    = dev16_sse3;
595            }
596    
597  #endif /* ARCH_IS_IA32 */  #endif /* ARCH_IS_IA32 */
598    
599  #if defined(ARCH_IS_IA64)  #if defined(ARCH_IS_IA64)
# Line 530  Line 623 
623  #endif  #endif
624    
625  #if defined(ARCH_IS_PPC)  #if defined(ARCH_IS_PPC)
626          if ((cpu_flags & XVID_CPU_ASM))          if ((cpu_flags & XVID_CPU_ALTIVEC)) {
627          {            /* sad operators */
628                  calc_cbp = calc_cbp_ppc;                    sad16 = sad16_altivec_c;
629          }                    sad16bi = sad16bi_altivec_c;
630                      sad8 = sad8_altivec_c;
631                      dev16 = dev16_altivec_c;
632    
633              sse8_16bit = sse8_16bit_altivec_c;
634    
635              /* mem transfer */
636              transfer_8to16copy = transfer_8to16copy_altivec_c;
637              transfer_16to8copy = transfer_16to8copy_altivec_c;
638              transfer_8to16sub = transfer_8to16sub_altivec_c;
639              transfer_8to16subro = transfer_8to16subro_altivec_c;
640              transfer_8to16sub2 = transfer_8to16sub2_altivec_c;
641              transfer_16to8add = transfer_16to8add_altivec_c;
642              transfer8x8_copy = transfer8x8_copy_altivec_c;
643    
644          if ((cpu_flags & XVID_CPU_ALTIVEC))            /* Inverse DCT */
645          {            idct = idct_altivec_c;
646                  calc_cbp = calc_cbp_altivec;  
647                  fdct = fdct_altivec;            /* Interpolation */
648                  idct = idct_altivec;            interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_altivec_c;
649                  sadInit = sadInit_altivec;            interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_altivec_c;
650                  sad16 = sad16_altivec;            interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_altivec_c;
651                  sad8 = sad8_altivec;  
652                  dev16 = dev16_altivec;            interpolate8x8_avg2 = interpolate8x8_avg2_altivec_c;
653              interpolate8x8_avg4 = interpolate8x8_avg4_altivec_c;
654    
655                      interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_altivec_c;
656                      interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_altivec_c;
657                      interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_altivec_c;
658                      interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_altivec_c;
659    
660              /* Colorspace conversion */
661              bgra_to_yv12 = bgra_to_yv12_altivec_c;
662              abgr_to_yv12 = abgr_to_yv12_altivec_c;
663              rgba_to_yv12 = rgba_to_yv12_altivec_c;
664              argb_to_yv12 = argb_to_yv12_altivec_c;
665    
666              yuyv_to_yv12 = yuyv_to_yv12_altivec_c;
667              uyvy_to_yv12 = uyvy_to_yv12_altivec_c;
668    
669              yv12_to_yuyv = yv12_to_yuyv_altivec_c;
670              yv12_to_uyvy = yv12_to_uyvy_altivec_c;
671    
672              /* Quantization */
673              quant_h263_intra = quant_h263_intra_altivec_c;
674              quant_h263_inter = quant_h263_inter_altivec_c;
675              dequant_h263_intra = dequant_h263_intra_altivec_c;
676              dequant_h263_inter = dequant_h263_inter_altivec_c;
677    
678                      dequant_mpeg_intra = dequant_mpeg_intra_altivec_c;
679                      dequant_mpeg_inter = dequant_mpeg_inter_altivec_c;
680    
681                      /* Qpel stuff */
682                      xvid_QP_Funcs = &xvid_QP_Funcs_Altivec_C;
683                      xvid_QP_Add_Funcs = &xvid_QP_Add_Funcs_Altivec_C;
684          }          }
685  #endif  #endif
686    
# Line 551  Line 688 
688      xvid_debug = init->debug;      xvid_debug = init->debug;
689  #endif  #endif
690    
691      return 0;      return(0);
692  }  }
693    
694    
# Line 562  Line 699 
699                  return XVID_ERR_VERSION;                  return XVID_ERR_VERSION;
700    
701          info->actual_version = XVID_VERSION;          info->actual_version = XVID_VERSION;
702          info->build = "xvid-1.0.0";          info->build = "xvid-1.3.3";
703          info->cpu_flags = detect_cpu_flags();          info->cpu_flags = detect_cpu_flags();
704            info->num_threads = 0; /* single-thread */
705    
706    #if defined(_WIN32)
707    
708      {
709            SYSTEM_INFO siSysInfo;
710            GetSystemInfo(&siSysInfo);
711            info->num_threads = siSysInfo.dwNumberOfProcessors; /* number of _logical_ cores */
712      }
713    
714    #elif defined(_SC_NPROCESSORS_CONF) /* should be available on Apple too actually */
715    
716      info->num_threads = sysconf(_SC_NPROCESSORS_CONF);
717    
718    #elif defined(__APPLE__) && defined(__MACH__)
719    
720      {
721        size_t len;
722        int    mib[2], ncpu;
723    
724        mib[0] = CTL_HW;
725        mib[1] = HW_NCPU;
726        len    = sizeof(ncpu);
727        if (sysctl(mib, 2, &ncpu, &len, NULL, 0) == 0)
728          info -> num_threads = ncpu;
729        else
730          info -> num_threads = 1;
731      }
732    
733    #elif defined(__amigaos4__)
734    
735      {
736        uint32_t num_threads = 1;
737        IExec->GetCPUInfoTags(GCIT_NumberOfCPUs, &num_threads, TAG_END);
738        info->num_threads = num_threads;
739      }
740    
 #if defined(_SMP) && defined(WIN32)  
         info->num_threads = pthread_num_processors_np();;  
 #else  
         info->num_threads = 0;  
741  #endif  #endif
742    
743          return 0;          return 0;
# Line 606  Line 775 
775                                                  convert->output.csp, convert->interlacing);                                                  convert->output.csp, convert->interlacing);
776                          break;                          break;
777    
778                    case XVID_CSP_INTERNAL :
779                            img.y = (uint8_t*)convert->input.plane[0];
780                            img.u = (uint8_t*)convert->input.plane[1];
781                            img.v = (uint8_t*)convert->input.plane[2];
782                            image_output(&img, width, height, convert->input.stride[0],
783                                                    (uint8_t**)convert->output.plane, convert->output.stride,
784                                                    convert->output.csp, convert->interlacing);
785                            break;
786    
787                  default :                  default :
788                          return XVID_ERR_FORMAT;                          return XVID_ERR_FORMAT;
789          }          }
# Line 616  Line 794 
794  }  }
795    
796  /*****************************************************************************  /*****************************************************************************
797   * XviD Global Entry point   * Xvid Global Entry point
798   *   *
799   * Well this function initialize all internal function pointers according   * Well this function initialize all internal function pointers according
800   * to the CPU features forced by the library client or autodetected (depending   * to the CPU features forced by the library client or autodetected (depending
# Line 649  Line 827 
827  }  }
828    
829  /*****************************************************************************  /*****************************************************************************
830   * XviD Native decoder entry point   * Xvid Native decoder entry point
831   *   *
832   * This function is just a wrapper to all the option cases.   * This function is just a wrapper to all the option cases.
833   *   *
# Line 681  Line 859 
859    
860    
861  /*****************************************************************************  /*****************************************************************************
862   * XviD Native encoder entry point   * Xvid Native encoder entry point
863   *   *
864   * This function is just a wrapper to all the option cases.   * This function is just a wrapper to all the option cases.
865   *   *
Legend:
Removed from v.1382  
changed lines
  Added in v.2053
No admin address has been configured
 ViewVC Help
Powered by ViewVC 1.0.4