[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 1.65.2.5, Wed Jun 27 18:57:42 2007 UTC revision 1.86, Fri Dec 31 11:30:38 2010 UTC
# Line 4  Line 4 
4   *  - Native API implementation  -   *  - Native API implementation  -
5   *   *
6   *  Copyright(C) 2001-2004 Peter Ross <pross@xvid.org>   *  Copyright(C) 2001-2004 Peter Ross <pross@xvid.org>
7     *               2002-2010 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 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 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 256  Line 287 
287          yv12_to_yv12    = yv12_to_yv12_c;          yv12_to_yv12    = yv12_to_yv12_c;
288          rgb555_to_yv12  = rgb555_to_yv12_c;          rgb555_to_yv12  = rgb555_to_yv12_c;
289          rgb565_to_yv12  = rgb565_to_yv12_c;          rgb565_to_yv12  = rgb565_to_yv12_c;
290            rgb_to_yv12     = rgb_to_yv12_c;
291          bgr_to_yv12     = bgr_to_yv12_c;          bgr_to_yv12     = bgr_to_yv12_c;
292          bgra_to_yv12    = bgra_to_yv12_c;          bgra_to_yv12    = bgra_to_yv12_c;
293          abgr_to_yv12    = abgr_to_yv12_c;          abgr_to_yv12    = abgr_to_yv12_c;
# Line 277  Line 309 
309          /* All colorspace transformation functions YV12->User format */          /* All colorspace transformation functions YV12->User format */
310          yv12_to_rgb555  = yv12_to_rgb555_c;          yv12_to_rgb555  = yv12_to_rgb555_c;
311          yv12_to_rgb565  = yv12_to_rgb565_c;          yv12_to_rgb565  = yv12_to_rgb565_c;
312            yv12_to_rgb     = yv12_to_rgb_c;
313          yv12_to_bgr     = yv12_to_bgr_c;          yv12_to_bgr     = yv12_to_bgr_c;
314          yv12_to_bgra    = yv12_to_bgra_c;          yv12_to_bgra    = yv12_to_bgra_c;
315          yv12_to_abgr    = yv12_to_abgr_c;          yv12_to_abgr    = yv12_to_abgr_c;
# Line 306  Line 339 
339          sse8_16bit = sse8_16bit_c;          sse8_16bit = sse8_16bit_c;
340          sse8_8bit  = sse8_8bit_c;          sse8_8bit  = sse8_8bit_c;
341    
342  #if defined(ARCH_IS_IA32)          sseh8_16bit   = sseh8_16bit_c;
343            coeff8_energy = coeff8_energy_c;
344            blocksum8     = blocksum8_c;
345    
346            init_GMC(cpu_flags);
347    
348    #if defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)
349    
350          if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||          if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||
351                  (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||                  (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||
352                  (cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2))                  (cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2) ||
353            (cpu_flags & XVID_CPU_SSE3) || (cpu_flags & XVID_CPU_SSE41))
354          {          {
355                  /* Restore FPU context : emms_c is a nop functions */                  /* Restore FPU context : emms_c is a nop functions */
356                  emms = emms_mmx;                  emms = emms_mmx;
# Line 331  Line 371 
371                  quant_h263_inter   = quant_h263_inter_mmx;                  quant_h263_inter   = quant_h263_inter_mmx;
372                  dequant_h263_intra = dequant_h263_intra_mmx;                  dequant_h263_intra = dequant_h263_intra_mmx;
373                  dequant_h263_inter = dequant_h263_inter_mmx;                  dequant_h263_inter = dequant_h263_inter_mmx;
   
374                  quant_mpeg_intra   = quant_mpeg_intra_mmx;                  quant_mpeg_intra   = quant_mpeg_intra_mmx;
375                  quant_mpeg_inter   = quant_mpeg_inter_mmx;                  quant_mpeg_inter   = quant_mpeg_inter_mmx;
376                  dequant_mpeg_intra = dequant_mpeg_intra_mmx;                  dequant_mpeg_intra = dequant_mpeg_intra_mmx;
377                  dequant_mpeg_inter = dequant_mpeg_inter_mmx;                  dequant_mpeg_inter = dequant_mpeg_inter_mmx;
378    
379    
380                  /* Block related functions */                  /* Block related functions */
381                  transfer_8to16copy = transfer_8to16copy_mmx;                  transfer_8to16copy = transfer_8to16copy_mmx;
382                  transfer_16to8copy = transfer_16to8copy_mmx;                  transfer_16to8copy = transfer_16to8copy_mmx;
# Line 374  Line 414 
414                  image_brightness = image_brightness_mmx;                  image_brightness = image_brightness_mmx;
415    
416                  /* image input xxx_to_yv12 related functions */                  /* image input xxx_to_yv12 related functions */
417    
418                  yv12_to_yv12  = yv12_to_yv12_mmx;                  yv12_to_yv12  = yv12_to_yv12_mmx;
419    
420                  bgr_to_yv12   = bgr_to_yv12_mmx;                  bgr_to_yv12   = bgr_to_yv12_mmx;
421                    rgb_to_yv12   = rgb_to_yv12_mmx;
422                  bgra_to_yv12  = bgra_to_yv12_mmx;                  bgra_to_yv12  = bgra_to_yv12_mmx;
423                    rgba_to_yv12  = rgba_to_yv12_mmx;
424                  yuyv_to_yv12  = yuyv_to_yv12_mmx;                  yuyv_to_yv12  = yuyv_to_yv12_mmx;
425                  uyvy_to_yv12  = uyvy_to_yv12_mmx;                  uyvy_to_yv12  = uyvy_to_yv12_mmx;
426    
# Line 412  Line 456 
456    
457                  yuyv_to_yv12  = yuyv_to_yv12_3dn;                  yuyv_to_yv12  = yuyv_to_yv12_3dn;
458                  uyvy_to_yv12  = uyvy_to_yv12_3dn;                  uyvy_to_yv12  = uyvy_to_yv12_3dn;
459    
460          }          }
461    
462    
# Line 436  Line 481 
481                  interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_xmm;                  interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_xmm;
482    
483                  /* Quantization */                  /* Quantization */
                 quant_mpeg_intra = quant_mpeg_intra_xmm;  
484                  quant_mpeg_inter = quant_mpeg_inter_xmm;                  quant_mpeg_inter = quant_mpeg_inter_xmm;
485    
486                  dequant_h263_intra = dequant_h263_intra_xmm;                  dequant_h263_intra = dequant_h263_intra_xmm;
# Line 447  Line 491 
491                  transfer_8to16sub2ro = transfer_8to16sub2ro_xmm;                  transfer_8to16sub2ro = transfer_8to16sub2ro_xmm;
492    
493                  /* Colorspace transformation */                  /* Colorspace transformation */
494                  yv12_to_yv12  = yv12_to_yv12_xmm;                  /* yv12_to_yv12  = yv12_to_yv12_xmm; */ /* appears to be slow on many machines */
495                  yuyv_to_yv12  = yuyv_to_yv12_xmm;                  yuyv_to_yv12  = yuyv_to_yv12_xmm;
496                  uyvy_to_yv12  = uyvy_to_yv12_xmm;                  uyvy_to_yv12  = uyvy_to_yv12_xmm;
497    
# Line 508  Line 552 
552                          dequant_h263_inter = dequant_h263_inter_3dne;                          dequant_h263_inter = dequant_h263_inter_3dne;
553    
554                          /* ME functions */                          /* ME functions */
                         calc_cbp = calc_cbp_3dne;  
   
555                          sad16 = sad16_3dne;                          sad16 = sad16_3dne;
556                          sad8 = sad8_3dne;                          sad8 = sad8_3dne;
557                          sad16bi = sad16bi_3dne;                          sad16bi = sad16bi_3dne;
# Line 532  Line 574 
574                  sad16    = sad16_sse2;                  sad16    = sad16_sse2;
575                  dev16    = dev16_sse2;                  dev16    = dev16_sse2;
576    
577                    /* PSNR-HVS-M distortion metric */
578                    sseh8_16bit   = sseh8_16bit_sse2;
579                    coeff8_energy = coeff8_energy_sse2;
580                    blocksum8     = blocksum8_sse2;
581    
582                  /* DCT operators */                  /* DCT operators */
583                  fdct = fdct_sse2_skal;                  fdct = fdct_sse2_skal;
584      /* idct = idct_sse2_skal; */   /* Is now IEEE1180 and Walken compliant. Disabled until fully tested. */                  idct = idct_sse2_skal;   /* Is now IEEE1180 and Walken compliant. */
585    
586                  /* postprocessing */                  /* postprocessing */
587                  image_brightness = image_brightness_sse2;                  image_brightness = image_brightness_sse2;
588    
589          }          }
590    
591            if ((cpu_flags & XVID_CPU_SSE3)) {
592    
593                    /* SAD operators */
594                    sad16    = sad16_sse3;
595                    dev16    = dev16_sse3;
596            }
597    
598  #endif /* ARCH_IS_IA32 */  #endif /* ARCH_IS_IA32 */
599    
600  #if defined(ARCH_IS_IA64)  #if defined(ARCH_IS_IA64)
# Line 629  Line 685 
685          }          }
686  #endif  #endif
687    
 #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_mmx;  
   
                 /* 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  
   
688  #if defined(_DEBUG)  #if defined(_DEBUG)
689      xvid_debug = init->debug;      xvid_debug = init->debug;
690  #endif  #endif
# Line 712  Line 700 
700                  return XVID_ERR_VERSION;                  return XVID_ERR_VERSION;
701    
702          info->actual_version = XVID_VERSION;          info->actual_version = XVID_VERSION;
703          info->build = "xvid-1.1.3";          info->build = "xvid-1.4.0-dev";
704          info->cpu_flags = detect_cpu_flags();          info->cpu_flags = detect_cpu_flags();
705            info->num_threads = 0; /* single-thread */
706    
707    #if defined(_WIN32)
708    
709      {
710            SYSTEM_INFO siSysInfo;
711            GetSystemInfo(&siSysInfo);
712            info->num_threads = siSysInfo.dwNumberOfProcessors; /* number of _logical_ cores */
713      }
714    
715    #elif defined(_SC_NPROCESSORS_CONF) /* should be available on Apple too actually */
716    
717      info->num_threads = sysconf(_SC_NPROCESSORS_CONF);
718    
719    #elif defined(__APPLE__) && defined(__MACH__)
720    
721      {
722        size_t len;
723        int    mib[2], ncpu;
724    
725        mib[0] = CTL_HW;
726        mib[1] = HW_NCPU;
727        len    = sizeof(ncpu);
728        if (sysctl(mib, 2, &ncpu, &len, NULL, 0) == 0)
729          info -> num_threads = ncpu;
730        else
731          info -> num_threads = 1;
732      }
733    
734    #elif defined(__amigaos4__)
735    
736      {
737        uint32_t num_threads = 1;
738        IExec->GetCPUInfoTags(GCIT_NumberOfCPUs, &num_threads, TAG_END);
739        info->num_threads = num_threads;
740      }
741    
 #if defined(_SMP) && defined(WIN32)  
         info->num_threads = pthread_num_processors_np();;  
 #else  
         info->num_threads = 0;  
742  #endif  #endif
743    
744          return 0;          return 0;
# Line 756  Line 776 
776                                                  convert->output.csp, convert->interlacing);                                                  convert->output.csp, convert->interlacing);
777                          break;                          break;
778    
779                    case XVID_CSP_INTERNAL :
780                            img.y = (uint8_t*)convert->input.plane[0];
781                            img.u = (uint8_t*)convert->input.plane[1];
782                            img.v = (uint8_t*)convert->input.plane[2];
783                            image_output(&img, width, height, convert->input.stride[0],
784                                                    (uint8_t**)convert->output.plane, convert->output.stride,
785                                                    convert->output.csp, convert->interlacing);
786                            break;
787    
788                  default :                  default :
789                          return XVID_ERR_FORMAT;                          return XVID_ERR_FORMAT;
790          }          }
# Line 766  Line 795 
795  }  }
796    
797  /*****************************************************************************  /*****************************************************************************
798   * XviD Global Entry point   * Xvid Global Entry point
799   *   *
800   * Well this function initialize all internal function pointers according   * Well this function initialize all internal function pointers according
801   * 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 799  Line 828 
828  }  }
829    
830  /*****************************************************************************  /*****************************************************************************
831   * XviD Native decoder entry point   * Xvid Native decoder entry point
832   *   *
833   * This function is just a wrapper to all the option cases.   * This function is just a wrapper to all the option cases.
834   *   *
# Line 831  Line 860 
860    
861    
862  /*****************************************************************************  /*****************************************************************************
863   * XviD Native encoder entry point   * Xvid Native encoder entry point
864   *   *
865   * This function is just a wrapper to all the option cases.   * This function is just a wrapper to all the option cases.
866   *   *

Legend:
Removed from v.1.65.2.5  
changed lines
  Added in v.1.86

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