222 |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_c; |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_c; |
223 |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c; |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c; |
224 |
|
|
225 |
|
interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_c; |
226 |
|
interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_c; |
227 |
|
interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_c; |
228 |
|
interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_c; |
229 |
|
|
230 |
interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c; |
interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c; |
231 |
interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c; |
interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c; |
232 |
interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c; |
interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c; |
307 |
dev16 = dev16_c; |
dev16 = dev16_c; |
308 |
sad16v = sad16v_c; |
sad16v = sad16v_c; |
309 |
sse8_16bit = sse8_16bit_c; |
sse8_16bit = sse8_16bit_c; |
310 |
|
sse8_8bit = sse8_8bit_c; |
311 |
|
|
312 |
#if defined(ARCH_IS_IA32) |
#if defined(ARCH_IS_IA32) |
313 |
|
|
362 |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_mmx; |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_mmx; |
363 |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx; |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx; |
364 |
|
|
365 |
|
interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_mmx; |
366 |
|
interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_mmx; |
367 |
|
interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_mmx; |
368 |
|
interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_mmx; |
369 |
|
|
370 |
interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx; |
interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx; |
371 |
interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx; |
interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx; |
372 |
|
|
408 |
dev16 = dev16_mmx; |
dev16 = dev16_mmx; |
409 |
sad16v = sad16v_mmx; |
sad16v = sad16v_mmx; |
410 |
sse8_16bit = sse8_16bit_mmx; |
sse8_16bit = sse8_16bit_mmx; |
411 |
|
sse8_8bit = sse8_8bit_mmx; |
412 |
} |
} |
413 |
|
|
414 |
/* these 3dnow functions are faster than mmx, but slower than xmm. */ |
/* these 3dnow functions are faster than mmx, but slower than xmm. */ |
436 |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_xmm; |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_xmm; |
437 |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm; |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm; |
438 |
|
|
439 |
|
interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_xmm; |
440 |
|
interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_xmm; |
441 |
|
interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_xmm; |
442 |
|
interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_xmm; |
443 |
|
|
444 |
/* reduced resolution */ |
/* reduced resolution */ |
445 |
copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_xmm; |
copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_xmm; |
446 |
add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_xmm; |
add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_xmm; |
479 |
|
|
480 |
if ((cpu_flags & XVID_CPU_3DNOWEXT)) { |
if ((cpu_flags & XVID_CPU_3DNOWEXT)) { |
481 |
|
|
|
/* Inverse DCT */ |
|
|
idct = idct_3dne; |
|
|
|
|
482 |
/* Buffer transfer */ |
/* Buffer transfer */ |
483 |
transfer_8to16copy = transfer_8to16copy_3dne; |
transfer_8to16copy = transfer_8to16copy_3dne; |
484 |
transfer_16to8copy = transfer_16to8copy_3dne; |
transfer_16to8copy = transfer_16to8copy_3dne; |
485 |
transfer_8to16sub = transfer_8to16sub_3dne; |
transfer_8to16sub = transfer_8to16sub_3dne; |
486 |
transfer_8to16subro = transfer_8to16subro_3dne; |
transfer_8to16subro = transfer_8to16subro_3dne; |
|
transfer_8to16sub2 = transfer_8to16sub2_3dne; |
|
487 |
transfer_16to8add = transfer_16to8add_3dne; |
transfer_16to8add = transfer_16to8add_3dne; |
488 |
transfer8x8_copy = transfer8x8_copy_3dne; |
transfer8x8_copy = transfer8x8_copy_3dne; |
489 |
|
|
490 |
|
if ((cpu_flags & XVID_CPU_MMXEXT)) { |
491 |
|
/* Inverse DCT */ |
492 |
|
idct = idct_3dne; |
493 |
|
|
494 |
|
/* Buffer transfer */ |
495 |
|
transfer_8to16sub2 = transfer_8to16sub2_3dne; |
496 |
|
|
497 |
|
/* Interpolation */ |
498 |
|
interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dne; |
499 |
|
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne; |
500 |
|
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne; |
501 |
|
|
502 |
/* Quantization */ |
/* Quantization */ |
503 |
quant_h263_intra = quant_h263_intra_3dne; |
quant_h263_intra = quant_h263_intra_3dne; /* cmov only */ |
504 |
quant_h263_inter = quant_h263_inter_3dne; |
quant_h263_inter = quant_h263_inter_3dne; |
505 |
dequant_mpeg_intra = dequant_mpeg_intra_3dne; |
dequant_mpeg_intra = dequant_mpeg_intra_3dne; /* cmov only */ |
506 |
dequant_mpeg_inter = dequant_mpeg_inter_3dne; |
dequant_mpeg_inter = dequant_mpeg_inter_3dne; |
507 |
dequant_h263_intra = dequant_h263_intra_3dne; |
dequant_h263_intra = dequant_h263_intra_3dne; |
508 |
dequant_h263_inter = dequant_h263_inter_3dne; |
dequant_h263_inter = dequant_h263_inter_3dne; |
509 |
|
|
510 |
/* ME functions */ |
/* ME functions */ |
511 |
calc_cbp = calc_cbp_3dne; |
calc_cbp = calc_cbp_3dne; |
512 |
|
|
513 |
sad16 = sad16_3dne; |
sad16 = sad16_3dne; |
514 |
sad8 = sad8_3dne; |
sad8 = sad8_3dne; |
515 |
sad16bi = sad16bi_3dne; |
sad16bi = sad16bi_3dne; |
516 |
sad8bi = sad8bi_3dne; |
sad8bi = sad8bi_3dne; |
517 |
dev16 = dev16_3dne; |
dev16 = dev16_3dne; |
518 |
|
} |
|
/* Interpolation */ |
|
|
interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dne; |
|
|
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne; |
|
|
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne; |
|
519 |
} |
} |
520 |
|
|
521 |
if ((cpu_flags & XVID_CPU_SSE2)) { |
if ((cpu_flags & XVID_CPU_SSE2)) { |
535 |
/* DCT operators |
/* DCT operators |
536 |
* no iDCT because it's not "Walken matching" */ |
* no iDCT because it's not "Walken matching" */ |
537 |
fdct = fdct_sse2_skal; |
fdct = fdct_sse2_skal; |
538 |
|
|
539 |
|
/* postprocessing */ |
540 |
|
image_brightness = image_brightness_sse2; |
541 |
} |
} |
542 |
#endif /* ARCH_IS_IA32 */ |
#endif /* ARCH_IS_IA32 */ |
543 |
|
|
634 |
return XVID_ERR_VERSION; |
return XVID_ERR_VERSION; |
635 |
|
|
636 |
info->actual_version = XVID_VERSION; |
info->actual_version = XVID_VERSION; |
637 |
info->build = "xvid-1.0.0"; |
info->build = "xvid-1.1-cvshead"; |
638 |
info->cpu_flags = detect_cpu_flags(); |
info->cpu_flags = detect_cpu_flags(); |
639 |
|
|
640 |
#if defined(_SMP) && defined(WIN32) |
#if defined(_SMP) && defined(WIN32) |