[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 1596, Sun Jan 16 10:34:52 2005 UTC branches/release-1_3-branch/xvidcore/src/xvid.c revision 2011, Mon May 23 07:47:37 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) 2001-2004 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 1.61 2005-01-16 10:34:52 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 40  Line 61 
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 53  Line 75 
75    
76  #if (defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)) && defined(_MSC_VER)  #if (defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)) && defined(_MSC_VER)
77  #       include <windows.h>  #       include <windows.h>
78  #elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || defined(ARCH_IS_PPC)  #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 66  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 89  Line 110 
110          }          }
111          return(0);          return(0);
112  }  }
113  #elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || defined(ARCH_IS_PPC)  #elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || (defined(ARCH_IS_PPC) && !defined(__amigaos4__))
114  static int  static int
115  sigill_check(void (*func)())  sigill_check(void (*func)())
116  {  {
# Line 126  Line 147 
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;
# Line 136  Line 157 
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(__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))          if (!sigill_check(altivec_trigger))
175                  cpu_flags |= XVID_CPU_ALTIVEC;                  cpu_flags |= XVID_CPU_ALTIVEC;
176  #endif  #endif
177    #endif
178    
179          return cpu_flags;          return cpu_flags;
180  }  }
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 175  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 183  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 213  Line 243 
243          transfer_8to16sub2ro = transfer_8to16sub2ro_c;          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 222  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;          interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_c;
261          interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_c;          interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_c;
262          interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_c;          interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_c;
# Line 251  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 272  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 301  Line 338 
338          sse8_16bit = sse8_16bit_c;          sse8_16bit = sse8_16bit_c;
339          sse8_8bit  = sse8_8bit_c;          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            init_GMC(cpu_flags);
346    
347    #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 326  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 340  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 349  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;                  interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_mmx;
402                  interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_mmx;                  interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_mmx;
403                  interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_mmx;                  interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_mmx;
# Line 364  Line 413 
413                  image_brightness = image_brightness_mmx;                  image_brightness = image_brightness_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 402  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 416  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                    interpolate8x4_halfpel_h  = interpolate8x4_halfpel_h_xmm;
474                    interpolate8x4_halfpel_v  = interpolate8x4_halfpel_v_xmm;
475                    interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_xmm;
476    
477                  interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_xmm;                  interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_xmm;
478                  interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_xmm;                  interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_xmm;
479                  interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_xmm;                  interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_xmm;
480                  interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_xmm;                  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 433  Line 490 
490                  transfer_8to16sub2ro = transfer_8to16sub2ro_xmm;                  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 452  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)) {
# Line 463  Line 524 
524                  transfer_8to16subro =  transfer_8to16subro_3dne;                  transfer_8to16subro =  transfer_8to16subro_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)) {                  if ((cpu_flags & XVID_CPU_MMXEXT)) {
530                          /* Inverse DCT */                          /* Inverse DCT */
# Line 476  Line 538 
538                          interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne;                          interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne;
539                          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne;                          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;               /* cmov only */                          quant_h263_intra = quant_h263_intra_3dne;               /* cmov only */
547                          quant_h263_inter = quant_h263_inter_3dne;                          quant_h263_inter = quant_h263_inter_3dne;
# Line 485  Line 551 
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;
# Line 509  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 */                  /* postprocessing */
586                  image_brightness = image_brightness_sse2;                  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 597  Line 675 
675            dequant_h263_intra = dequant_h263_intra_altivec_c;            dequant_h263_intra = dequant_h263_intra_altivec_c;
676            dequant_h263_inter = dequant_h263_inter_altivec_c;            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 */                    /* Qpel stuff */
682                    xvid_QP_Funcs = &xvid_QP_Funcs_Altivec_C;                    xvid_QP_Funcs = &xvid_QP_Funcs_Altivec_C;
683                    xvid_QP_Add_Funcs = &xvid_QP_Add_Funcs_Altivec_C;                    xvid_QP_Add_Funcs = &xvid_QP_Add_Funcs_Altivec_C;
684          }          }
685  #endif  #endif
686    
 #if defined(ARCH_IS_X86_64)  
         /* For now, only XVID_CPU_ASM is looked for, so user can still  
          * disable asm usage the usual way. When Intel EMT64 cpus will  
          * be out, maybe we'll have to check more precisely what cpu  
          * features there really are. */  
         if (cpu_flags & XVID_CPU_ASM) {  
                 /* SIMD state flusher */  
                 emms = emms_3dn;  
   
                 /* DCT operators */  
                 fdct = fdct_skal_x86_64;  
                 idct = idct_x86_64;  
   
                 /* SAD operators */  
                 sad16      = sad16_x86_64;  
                 sad8       = sad8_x86_64;  
                 sad16bi    = sad16bi_x86_64;  
                 sad8bi     = sad8bi_x86_64;  
                 dev16      = dev16_x86_64;  
                 sad16v     = sad16v_x86_64;  
                 sse8_16bit = sse8_16bit_x86_64;  
                 sse8_8bit  = sse8_8bit_x86_64;  
   
                 /* Interpolation operators */  
                 interpolate8x8_halfpel_h  = interpolate8x8_halfpel_h_x86_64;  
                 interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_x86_64;  
                 interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_x86_64;  
   
                 interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_x86_64;  
                 interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_x86_64;  
                 interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_x86_64;  
                 interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_x86_64;  
   
                 interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_x86_64;  
                 interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_x86_64;  
   
                 interpolate8x8_avg2 = interpolate8x8_avg2_x86_64;  
                 interpolate8x8_avg4 = interpolate8x8_avg4_x86_64;  
   
                 /* Quantization related functions */  
                 quant_h263_intra   = quant_h263_intra_x86_64;  
                 quant_h263_inter   = quant_h263_inter_x86_64;  
                 dequant_h263_intra = dequant_h263_intra_x86_64;  
                 dequant_h263_inter = dequant_h263_inter_x86_64;  
                 quant_mpeg_intra   = quant_mpeg_intra_x86_64;  
                 quant_mpeg_inter   = quant_mpeg_inter_x86_64;  
                 dequant_mpeg_intra   = dequant_mpeg_intra_x86_64;  
                 dequant_mpeg_inter   = dequant_mpeg_inter_x86_64;  
   
                 /* Block related functions */  
                 transfer_8to16copy  = transfer_8to16copy_x86_64;  
                 transfer_16to8copy  = transfer_16to8copy_x86_64;  
                 transfer_8to16sub   = transfer_8to16sub_x86_64;  
                 transfer_8to16subro = transfer_8to16subro_x86_64;  
                 transfer_8to16sub2  = transfer_8to16sub2_x86_64;  
                 transfer_8to16sub2ro= transfer_8to16sub2ro_x86_64;  
                 transfer_16to8add   = transfer_16to8add_x86_64;  
                 transfer8x8_copy    = transfer8x8_copy_x86_64;  
   
                 /* Qpel stuff */  
                 xvid_QP_Funcs = &xvid_QP_Funcs_x86_64;  
                 xvid_QP_Add_Funcs = &xvid_QP_Add_Funcs_x86_64;  
   
                 /* Interlacing Functions */  
                 MBFieldTest = MBFieldTest_x86_64;  
         }  
 #endif  
   
687  #if defined(_DEBUG)  #if defined(_DEBUG)
688      xvid_debug = init->debug;      xvid_debug = init->debug;
689  #endif  #endif
# Line 686  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-1.1.0-beta1";          info->build = "xvid-1.3.2";
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 730  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 740  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 773  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 805  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.1596  
changed lines
  Added in v.2011
No admin address has been configured
 ViewVC Help
Powered by ViewVC 1.0.4