28 |
#include <string.h> |
#include <string.h> |
29 |
#include <time.h> |
#include <time.h> |
30 |
|
|
31 |
|
#if !defined(_WIN32) |
32 |
|
#include <unistd.h> |
33 |
|
#endif |
34 |
|
|
35 |
#include "xvid.h" |
#include "xvid.h" |
36 |
#include "decoder.h" |
#include "decoder.h" |
37 |
#include "encoder.h" |
#include "encoder.h" |
40 |
#include "dct/fdct.h" |
#include "dct/fdct.h" |
41 |
#include "image/colorspace.h" |
#include "image/colorspace.h" |
42 |
#include "image/interpolate8x8.h" |
#include "image/interpolate8x8.h" |
|
#include "image/reduced.h" |
|
43 |
#include "utils/mem_transfer.h" |
#include "utils/mem_transfer.h" |
44 |
#include "utils/mbfunctions.h" |
#include "utils/mbfunctions.h" |
45 |
#include "quant/quant.h" |
#include "quant/quant.h" |
46 |
#include "motion/motion.h" |
#include "motion/motion.h" |
47 |
|
#include "motion/gmc.h" |
48 |
#include "motion/sad.h" |
#include "motion/sad.h" |
49 |
#include "utils/emms.h" |
#include "utils/emms.h" |
50 |
#include "utils/timer.h" |
#include "utils/timer.h" |
56 |
unsigned int xvid_debug = 0; /* xvid debug mask */ |
unsigned int xvid_debug = 0; /* xvid debug mask */ |
57 |
#endif |
#endif |
58 |
|
|
59 |
#if defined(ARCH_IS_IA32) && defined(_MSC_VER) |
#if (defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)) && defined(_MSC_VER) |
60 |
# include <windows.h> |
# include <windows.h> |
61 |
#elif defined(ARCH_IS_IA32) || defined(ARCH_IS_PPC) |
#elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || defined(ARCH_IS_PPC) |
62 |
# include <signal.h> |
# include <signal.h> |
63 |
# include <setjmp.h> |
# include <setjmp.h> |
64 |
|
|
81 |
* 0 : SIGILL was *not* signalled |
* 0 : SIGILL was *not* signalled |
82 |
* 1 : SIGILL was signalled |
* 1 : SIGILL was signalled |
83 |
*/ |
*/ |
84 |
#if defined(ARCH_IS_IA32) && defined(_MSC_VER) |
#if (defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64)) && defined(_MSC_VER) |
85 |
static int |
static int |
86 |
sigill_check(void (*func)()) |
sigill_check(void (*func)()) |
87 |
{ |
{ |
94 |
} |
} |
95 |
return(0); |
return(0); |
96 |
} |
} |
97 |
#elif defined(ARCH_IS_IA32) || defined(ARCH_IS_PPC) |
#elif defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) || defined(ARCH_IS_PPC) |
98 |
static int |
static int |
99 |
sigill_check(void (*func)()) |
sigill_check(void (*func)()) |
100 |
{ |
{ |
131 |
|
|
132 |
/* detect cpu flags */ |
/* detect cpu flags */ |
133 |
static unsigned int |
static unsigned int |
134 |
detect_cpu_flags() |
detect_cpu_flags(void) |
135 |
{ |
{ |
136 |
/* enable native assembly optimizations by default */ |
/* enable native assembly optimizations by default */ |
137 |
unsigned int cpu_flags = XVID_CPU_ASM; |
unsigned int cpu_flags = XVID_CPU_ASM; |
138 |
|
|
139 |
#if defined(ARCH_IS_IA32) |
#if defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) |
140 |
cpu_flags |= check_cpu_features(); |
cpu_flags |= check_cpu_features(); |
141 |
if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger)) |
if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger)) |
142 |
cpu_flags &= ~XVID_CPU_SSE; |
cpu_flags &= ~XVID_CPU_SSE; |
143 |
|
|
144 |
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)) |
145 |
cpu_flags &= ~XVID_CPU_SSE2; |
cpu_flags &= ~(XVID_CPU_SSE2|XVID_CPU_SSE3|XVID_CPU_SSE41); |
146 |
#endif |
#endif |
147 |
|
|
148 |
#if defined(ARCH_IS_PPC) |
#if defined(ARCH_IS_PPC) |
188 |
idct = idct_int32; |
idct = idct_int32; |
189 |
|
|
190 |
/* Only needed on PPC Altivec archs */ |
/* Only needed on PPC Altivec archs */ |
191 |
sadInit = 0; |
sadInit = NULL; |
192 |
|
|
193 |
/* Restore FPU context : emms_c is a nop functions */ |
/* Restore FPU context : emms_c is a nop functions */ |
194 |
emms = emms_c; |
emms = emms_c; |
215 |
transfer_8to16sub = transfer_8to16sub_c; |
transfer_8to16sub = transfer_8to16sub_c; |
216 |
transfer_8to16subro = transfer_8to16subro_c; |
transfer_8to16subro = transfer_8to16subro_c; |
217 |
transfer_8to16sub2 = transfer_8to16sub2_c; |
transfer_8to16sub2 = transfer_8to16sub2_c; |
218 |
|
transfer_8to16sub2ro = transfer_8to16sub2ro_c; |
219 |
transfer_16to8add = transfer_16to8add_c; |
transfer_16to8add = transfer_16to8add_c; |
220 |
transfer8x8_copy = transfer8x8_copy_c; |
transfer8x8_copy = transfer8x8_copy_c; |
221 |
|
transfer8x4_copy = transfer8x4_copy_c; |
222 |
|
|
223 |
/* Interlacing functions */ |
/* Interlacing functions */ |
224 |
MBFieldTest = MBFieldTest_c; |
MBFieldTest = MBFieldTest_c; |
228 |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_c; |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_c; |
229 |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c; |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c; |
230 |
|
|
231 |
|
interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_c; |
232 |
|
interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_c; |
233 |
|
interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_c; |
234 |
|
|
235 |
|
interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_c; |
236 |
|
interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_c; |
237 |
|
interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_c; |
238 |
|
interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_c; |
239 |
|
|
240 |
interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c; |
interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c; |
241 |
interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c; |
interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c; |
242 |
interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c; |
interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c; |
254 |
/* postprocessing */ |
/* postprocessing */ |
255 |
image_brightness = image_brightness_c; |
image_brightness = image_brightness_c; |
256 |
|
|
|
/* reduced resolution */ |
|
|
copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_C; |
|
|
add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_C; |
|
|
vfilter_31 = xvid_VFilter_31_C; |
|
|
hfilter_31 = xvid_HFilter_31_C; |
|
|
filter_18x18_to_8x8 = xvid_Filter_18x18_To_8x8_C; |
|
|
filter_diff_18x18_to_8x8 = xvid_Filter_Diff_18x18_To_8x8_C; |
|
|
|
|
257 |
/* Initialize internal colorspace transformation tables */ |
/* Initialize internal colorspace transformation tables */ |
258 |
colorspace_init(); |
colorspace_init(); |
259 |
|
|
261 |
yv12_to_yv12 = yv12_to_yv12_c; |
yv12_to_yv12 = yv12_to_yv12_c; |
262 |
rgb555_to_yv12 = rgb555_to_yv12_c; |
rgb555_to_yv12 = rgb555_to_yv12_c; |
263 |
rgb565_to_yv12 = rgb565_to_yv12_c; |
rgb565_to_yv12 = rgb565_to_yv12_c; |
264 |
|
rgb_to_yv12 = rgb_to_yv12_c; |
265 |
bgr_to_yv12 = bgr_to_yv12_c; |
bgr_to_yv12 = bgr_to_yv12_c; |
266 |
bgra_to_yv12 = bgra_to_yv12_c; |
bgra_to_yv12 = bgra_to_yv12_c; |
267 |
abgr_to_yv12 = abgr_to_yv12_c; |
abgr_to_yv12 = abgr_to_yv12_c; |
283 |
/* All colorspace transformation functions YV12->User format */ |
/* All colorspace transformation functions YV12->User format */ |
284 |
yv12_to_rgb555 = yv12_to_rgb555_c; |
yv12_to_rgb555 = yv12_to_rgb555_c; |
285 |
yv12_to_rgb565 = yv12_to_rgb565_c; |
yv12_to_rgb565 = yv12_to_rgb565_c; |
286 |
|
yv12_to_rgb = yv12_to_rgb_c; |
287 |
yv12_to_bgr = yv12_to_bgr_c; |
yv12_to_bgr = yv12_to_bgr_c; |
288 |
yv12_to_bgra = yv12_to_bgra_c; |
yv12_to_bgra = yv12_to_bgra_c; |
289 |
yv12_to_abgr = yv12_to_abgr_c; |
yv12_to_abgr = yv12_to_abgr_c; |
311 |
dev16 = dev16_c; |
dev16 = dev16_c; |
312 |
sad16v = sad16v_c; |
sad16v = sad16v_c; |
313 |
sse8_16bit = sse8_16bit_c; |
sse8_16bit = sse8_16bit_c; |
314 |
|
sse8_8bit = sse8_8bit_c; |
315 |
|
|
316 |
#if defined(ARCH_IS_IA32) |
init_GMC(cpu_flags); |
317 |
|
|
318 |
if ((cpu_flags & XVID_CPU_ASM)) { |
#if defined(ARCH_IS_IA32) || defined(ARCH_IS_X86_64) |
|
vfilter_31 = xvid_VFilter_31_x86; |
|
|
hfilter_31 = xvid_HFilter_31_x86; |
|
|
} |
|
319 |
|
|
320 |
if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) || |
if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) || |
321 |
(cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) || |
(cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) || |
322 |
(cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2)) |
(cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2) || |
323 |
|
(cpu_flags & XVID_CPU_SSE3) || (cpu_flags & XVID_CPU_SSE41)) |
324 |
{ |
{ |
325 |
/* Restore FPU context : emms_c is a nop functions */ |
/* Restore FPU context : emms_c is a nop functions */ |
326 |
emms = emms_mmx; |
emms = emms_mmx; |
341 |
quant_h263_inter = quant_h263_inter_mmx; |
quant_h263_inter = quant_h263_inter_mmx; |
342 |
dequant_h263_intra = dequant_h263_intra_mmx; |
dequant_h263_intra = dequant_h263_intra_mmx; |
343 |
dequant_h263_inter = dequant_h263_inter_mmx; |
dequant_h263_inter = dequant_h263_inter_mmx; |
|
|
|
344 |
quant_mpeg_intra = quant_mpeg_intra_mmx; |
quant_mpeg_intra = quant_mpeg_intra_mmx; |
345 |
quant_mpeg_inter = quant_mpeg_inter_mmx; |
quant_mpeg_inter = quant_mpeg_inter_mmx; |
346 |
dequant_mpeg_intra = dequant_mpeg_intra_mmx; |
dequant_mpeg_intra = dequant_mpeg_intra_mmx; |
347 |
dequant_mpeg_inter = dequant_mpeg_inter_mmx; |
dequant_mpeg_inter = dequant_mpeg_inter_mmx; |
348 |
|
|
349 |
|
|
350 |
/* Block related functions */ |
/* Block related functions */ |
351 |
transfer_8to16copy = transfer_8to16copy_mmx; |
transfer_8to16copy = transfer_8to16copy_mmx; |
352 |
transfer_16to8copy = transfer_16to8copy_mmx; |
transfer_16to8copy = transfer_16to8copy_mmx; |
355 |
transfer_8to16sub2 = transfer_8to16sub2_mmx; |
transfer_8to16sub2 = transfer_8to16sub2_mmx; |
356 |
transfer_16to8add = transfer_16to8add_mmx; |
transfer_16to8add = transfer_16to8add_mmx; |
357 |
transfer8x8_copy = transfer8x8_copy_mmx; |
transfer8x8_copy = transfer8x8_copy_mmx; |
358 |
|
transfer8x4_copy = transfer8x4_copy_mmx; |
359 |
|
|
360 |
/* Interlacing Functions */ |
/* Interlacing Functions */ |
361 |
MBFieldTest = MBFieldTest_mmx; |
MBFieldTest = MBFieldTest_mmx; |
365 |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_mmx; |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_mmx; |
366 |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx; |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx; |
367 |
|
|
368 |
|
interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_mmx; |
369 |
|
interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_mmx; |
370 |
|
interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_mmx; |
371 |
|
|
372 |
|
interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_mmx; |
373 |
|
interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_mmx; |
374 |
|
interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_mmx; |
375 |
|
interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_mmx; |
376 |
|
|
377 |
interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx; |
interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx; |
378 |
interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx; |
interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx; |
379 |
|
|
383 |
/* postprocessing */ |
/* postprocessing */ |
384 |
image_brightness = image_brightness_mmx; |
image_brightness = image_brightness_mmx; |
385 |
|
|
|
/* reduced resolution */ |
|
|
copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_mmx; |
|
|
add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_mmx; |
|
|
hfilter_31 = xvid_HFilter_31_mmx; |
|
|
filter_18x18_to_8x8 = xvid_Filter_18x18_To_8x8_mmx; |
|
|
filter_diff_18x18_to_8x8 = xvid_Filter_Diff_18x18_To_8x8_mmx; |
|
|
|
|
386 |
/* image input xxx_to_yv12 related functions */ |
/* image input xxx_to_yv12 related functions */ |
387 |
|
|
388 |
yv12_to_yv12 = yv12_to_yv12_mmx; |
yv12_to_yv12 = yv12_to_yv12_mmx; |
389 |
|
|
390 |
bgr_to_yv12 = bgr_to_yv12_mmx; |
bgr_to_yv12 = bgr_to_yv12_mmx; |
391 |
|
rgb_to_yv12 = rgb_to_yv12_mmx; |
392 |
bgra_to_yv12 = bgra_to_yv12_mmx; |
bgra_to_yv12 = bgra_to_yv12_mmx; |
393 |
|
rgba_to_yv12 = rgba_to_yv12_mmx; |
394 |
yuyv_to_yv12 = yuyv_to_yv12_mmx; |
yuyv_to_yv12 = yuyv_to_yv12_mmx; |
395 |
uyvy_to_yv12 = uyvy_to_yv12_mmx; |
uyvy_to_yv12 = uyvy_to_yv12_mmx; |
396 |
|
|
412 |
dev16 = dev16_mmx; |
dev16 = dev16_mmx; |
413 |
sad16v = sad16v_mmx; |
sad16v = sad16v_mmx; |
414 |
sse8_16bit = sse8_16bit_mmx; |
sse8_16bit = sse8_16bit_mmx; |
415 |
|
sse8_8bit = sse8_8bit_mmx; |
416 |
} |
} |
417 |
|
|
418 |
/* these 3dnow functions are faster than mmx, but slower than xmm. */ |
/* these 3dnow functions are faster than mmx, but slower than xmm. */ |
426 |
|
|
427 |
yuyv_to_yv12 = yuyv_to_yv12_3dn; |
yuyv_to_yv12 = yuyv_to_yv12_3dn; |
428 |
uyvy_to_yv12 = uyvy_to_yv12_3dn; |
uyvy_to_yv12 = uyvy_to_yv12_3dn; |
429 |
|
|
430 |
} |
} |
431 |
|
|
432 |
|
|
441 |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_xmm; |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_xmm; |
442 |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm; |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm; |
443 |
|
|
444 |
/* reduced resolution */ |
interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_xmm; |
445 |
copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_xmm; |
interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_xmm; |
446 |
add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_xmm; |
interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_xmm; |
447 |
|
|
448 |
|
interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_xmm; |
449 |
|
interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_xmm; |
450 |
|
interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_xmm; |
451 |
|
interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_xmm; |
452 |
|
|
453 |
/* Quantization */ |
/* Quantization */ |
|
quant_mpeg_intra = quant_mpeg_intra_xmm; |
|
454 |
quant_mpeg_inter = quant_mpeg_inter_xmm; |
quant_mpeg_inter = quant_mpeg_inter_xmm; |
455 |
|
|
456 |
dequant_h263_intra = dequant_h263_intra_xmm; |
dequant_h263_intra = dequant_h263_intra_xmm; |
458 |
|
|
459 |
/* Buffer transfer */ |
/* Buffer transfer */ |
460 |
transfer_8to16sub2 = transfer_8to16sub2_xmm; |
transfer_8to16sub2 = transfer_8to16sub2_xmm; |
461 |
|
transfer_8to16sub2ro = transfer_8to16sub2ro_xmm; |
462 |
|
|
463 |
/* Colorspace transformation */ |
/* Colorspace transformation */ |
464 |
yv12_to_yv12 = yv12_to_yv12_xmm; |
/* yv12_to_yv12 = yv12_to_yv12_xmm; */ /* appears to be slow on many machines */ |
465 |
yuyv_to_yv12 = yuyv_to_yv12_xmm; |
yuyv_to_yv12 = yuyv_to_yv12_xmm; |
466 |
uyvy_to_yv12 = uyvy_to_yv12_xmm; |
uyvy_to_yv12 = uyvy_to_yv12_xmm; |
467 |
|
|
480 |
interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn; |
interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn; |
481 |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn; |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn; |
482 |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn; |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn; |
483 |
|
|
484 |
|
interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_3dn; |
485 |
|
interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_3dn; |
486 |
|
interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_3dn; |
487 |
} |
} |
488 |
|
|
489 |
if ((cpu_flags & XVID_CPU_3DNOWEXT)) { |
if ((cpu_flags & XVID_CPU_3DNOWEXT)) { |
495 |
transfer_8to16subro = transfer_8to16subro_3dne; |
transfer_8to16subro = transfer_8to16subro_3dne; |
496 |
transfer_16to8add = transfer_16to8add_3dne; |
transfer_16to8add = transfer_16to8add_3dne; |
497 |
transfer8x8_copy = transfer8x8_copy_3dne; |
transfer8x8_copy = transfer8x8_copy_3dne; |
498 |
|
transfer8x4_copy = transfer8x4_copy_3dne; |
499 |
|
|
500 |
if ((cpu_flags & XVID_CPU_MMXEXT)) { |
if ((cpu_flags & XVID_CPU_MMXEXT)) { |
501 |
/* Inverse DCT */ |
/* Inverse DCT */ |
509 |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne; |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne; |
510 |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne; |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne; |
511 |
|
|
512 |
|
interpolate8x4_halfpel_h = interpolate8x4_halfpel_h_3dne; |
513 |
|
interpolate8x4_halfpel_v = interpolate8x4_halfpel_v_3dne; |
514 |
|
interpolate8x4_halfpel_hv = interpolate8x4_halfpel_hv_3dne; |
515 |
|
|
516 |
/* Quantization */ |
/* Quantization */ |
517 |
quant_h263_intra = quant_h263_intra_3dne; /* cmov only */ |
quant_h263_intra = quant_h263_intra_3dne; /* cmov only */ |
518 |
quant_h263_inter = quant_h263_inter_3dne; |
quant_h263_inter = quant_h263_inter_3dne; |
522 |
dequant_h263_inter = dequant_h263_inter_3dne; |
dequant_h263_inter = dequant_h263_inter_3dne; |
523 |
|
|
524 |
/* ME functions */ |
/* ME functions */ |
|
calc_cbp = calc_cbp_3dne; |
|
|
|
|
525 |
sad16 = sad16_3dne; |
sad16 = sad16_3dne; |
526 |
sad8 = sad8_3dne; |
sad8 = sad8_3dne; |
527 |
sad16bi = sad16bi_3dne; |
sad16bi = sad16bi_3dne; |
544 |
sad16 = sad16_sse2; |
sad16 = sad16_sse2; |
545 |
dev16 = dev16_sse2; |
dev16 = dev16_sse2; |
546 |
|
|
547 |
/* DCT operators |
/* DCT operators */ |
|
* no iDCT because it's not "Walken matching" */ |
|
548 |
fdct = fdct_sse2_skal; |
fdct = fdct_sse2_skal; |
549 |
|
idct = idct_sse2_skal; /* Is now IEEE1180 and Walken compliant. */ |
550 |
|
|
551 |
|
/* postprocessing */ |
552 |
|
image_brightness = image_brightness_sse2; |
553 |
|
|
554 |
|
} |
555 |
|
|
556 |
|
if ((cpu_flags & XVID_CPU_SSE3)) { |
557 |
|
|
558 |
|
/* SAD operators */ |
559 |
|
sad16 = sad16_sse3; |
560 |
|
dev16 = dev16_sse3; |
561 |
} |
} |
562 |
|
|
563 |
#endif /* ARCH_IS_IA32 */ |
#endif /* ARCH_IS_IA32 */ |
564 |
|
|
565 |
#if defined(ARCH_IS_IA64) |
#if defined(ARCH_IS_IA64) |
618 |
interpolate8x8_avg2 = interpolate8x8_avg2_altivec_c; |
interpolate8x8_avg2 = interpolate8x8_avg2_altivec_c; |
619 |
interpolate8x8_avg4 = interpolate8x8_avg4_altivec_c; |
interpolate8x8_avg4 = interpolate8x8_avg4_altivec_c; |
620 |
|
|
621 |
interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_altivec_c; |
interpolate8x8_halfpel_add = interpolate8x8_halfpel_add_altivec_c; |
622 |
|
interpolate8x8_halfpel_h_add = interpolate8x8_halfpel_h_add_altivec_c; |
623 |
|
interpolate8x8_halfpel_v_add = interpolate8x8_halfpel_v_add_altivec_c; |
624 |
|
interpolate8x8_halfpel_hv_add = interpolate8x8_halfpel_hv_add_altivec_c; |
625 |
|
|
626 |
/* Colorspace conversion */ |
/* Colorspace conversion */ |
627 |
bgra_to_yv12 = bgra_to_yv12_altivec_c; |
bgra_to_yv12 = bgra_to_yv12_altivec_c; |
640 |
quant_h263_inter = quant_h263_inter_altivec_c; |
quant_h263_inter = quant_h263_inter_altivec_c; |
641 |
dequant_h263_intra = dequant_h263_intra_altivec_c; |
dequant_h263_intra = dequant_h263_intra_altivec_c; |
642 |
dequant_h263_inter = dequant_h263_inter_altivec_c; |
dequant_h263_inter = dequant_h263_inter_altivec_c; |
643 |
|
|
644 |
|
dequant_mpeg_intra = dequant_mpeg_intra_altivec_c; |
645 |
|
dequant_mpeg_inter = dequant_mpeg_inter_altivec_c; |
646 |
|
|
647 |
|
/* Qpel stuff */ |
648 |
|
xvid_QP_Funcs = &xvid_QP_Funcs_Altivec_C; |
649 |
|
xvid_QP_Add_Funcs = &xvid_QP_Add_Funcs_Altivec_C; |
650 |
} |
} |
651 |
#endif |
#endif |
652 |
|
|
665 |
return XVID_ERR_VERSION; |
return XVID_ERR_VERSION; |
666 |
|
|
667 |
info->actual_version = XVID_VERSION; |
info->actual_version = XVID_VERSION; |
668 |
info->build = "xvid-1.0.0"; |
info->build = "xvid-1.3.0-dev"; |
669 |
info->cpu_flags = detect_cpu_flags(); |
info->cpu_flags = detect_cpu_flags(); |
670 |
|
info->num_threads = 0; |
671 |
|
|
672 |
|
#if defined(_WIN32) |
673 |
|
|
674 |
|
{ |
675 |
|
SYSTEM_INFO siSysInfo; |
676 |
|
GetSystemInfo(&siSysInfo); |
677 |
|
info->num_threads = siSysInfo.dwNumberOfProcessors; /* number of _logical_ cores */ |
678 |
|
} |
679 |
|
|
|
#if defined(_SMP) && defined(WIN32) |
|
|
info->num_threads = pthread_num_processors_np();; |
|
680 |
#else |
#else |
681 |
info->num_threads = 0; |
|
682 |
|
info->num_threads = sysconf(_SC_NPROCESSORS_CONF); |
683 |
|
|
684 |
#endif |
#endif |
685 |
|
|
686 |
return 0; |
return 0; |