#include "libavcodec/ac3dec.h"
#include "dsputil_mmx.h"
#include "idct_xvid.h"
+#include "diracdsp_mmx.h"
//#undef NDEBUG
//#include <assert.h>
put_vp_no_rnd_pixels8_l2_mmx(dst+8, a+8, b+8, stride, h);
}
+#if CONFIG_DIRAC_DECODER
+#define DIRAC_PIXOP(OPNAME, EXT)\
+void ff_ ## OPNAME ## _dirac_pixels8_ ## EXT(uint8_t *dst, const uint8_t *src[5], int stride, int h)\
+{\
+ OPNAME ## _pixels8_ ## EXT(dst, src[0], stride, h);\
+}\
+void ff_ ## OPNAME ## _dirac_pixels16_ ## EXT(uint8_t *dst, const uint8_t *src[5], int stride, int h)\
+{\
+ OPNAME ## _pixels16_ ## EXT(dst, src[0], stride, h);\
+}\
+void ff_ ## OPNAME ## _dirac_pixels32_ ## EXT(uint8_t *dst, const uint8_t *src[5], int stride, int h)\
+{\
+ OPNAME ## _pixels16_ ## EXT(dst , src[0] , stride, h);\
+ OPNAME ## _pixels16_ ## EXT(dst+16, src[0]+16, stride, h);\
+}
+
+DIRAC_PIXOP(put, mmx)
+DIRAC_PIXOP(avg, mmx)
+DIRAC_PIXOP(avg, mmx2)
+
+void ff_put_dirac_pixels16_sse2(uint8_t *dst, const uint8_t *src[5], int stride, int h)
+{
+ put_pixels16_sse2(dst, src[0], stride, h);
+}
+void ff_avg_dirac_pixels16_sse2(uint8_t *dst, const uint8_t *src[5], int stride, int h)
+{
+ avg_pixels16_sse2(dst, src[0], stride, h);
+}
+void ff_put_dirac_pixels32_sse2(uint8_t *dst, const uint8_t *src[5], int stride, int h)
+{
+ put_pixels16_sse2(dst , src[0] , stride, h);
+ put_pixels16_sse2(dst+16, src[0]+16, stride, h);
+}
+void ff_avg_dirac_pixels32_sse2(uint8_t *dst, const uint8_t *src[5], int stride, int h)
+{
+ avg_pixels16_sse2(dst , src[0] , stride, h);
+ avg_pixels16_sse2(dst+16, src[0]+16, stride, h);
+}
+#endif
+
/* XXX: those functions should be suppressed ASAP when all IDCTs are
converted */
#if CONFIG_GPL
int32_t max, unsigned int len);
void ff_vector_clip_int32_sse2 (int32_t *dst, const int32_t *src, int32_t min,
int32_t max, unsigned int len);
-void ff_vector_clip_int32_sse2_int(int32_t *dst, const int32_t *src, int32_t min,
+void ff_vector_clip_int32_int_sse2(int32_t *dst, const int32_t *src, int32_t min,
int32_t max, unsigned int len);
-void ff_vector_clip_int32_sse41 (int32_t *dst, const int32_t *src, int32_t min,
+void ff_vector_clip_int32_sse4 (int32_t *dst, const int32_t *src, int32_t min,
int32_t max, unsigned int len);
+extern void ff_butterflies_float_interleave_sse(float *dst, const float *src0,
+ const float *src1, int len);
+extern void ff_butterflies_float_interleave_avx(float *dst, const float *src0,
+ const float *src1, int len);
+
void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx)
{
int mm_flags = av_get_cpu_flags();
}
#if HAVE_YASM
- if (!high_bit_depth) {
+ if (!high_bit_depth && CONFIG_H264CHROMA) {
c->put_h264_chroma_pixels_tab[0]= ff_put_h264_chroma_mc8_mmx_rnd;
c->put_h264_chroma_pixels_tab[1]= ff_put_h264_chroma_mc4_mmx;
}
SET_QPEL_FUNCS(avg_2tap_qpel, 1, 8, mmx2, );
#if HAVE_YASM
- if (!high_bit_depth) {
+ if (!high_bit_depth && CONFIG_H264CHROMA) {
c->avg_h264_chroma_pixels_tab[0]= ff_avg_h264_chroma_mc8_mmx2_rnd;
c->avg_h264_chroma_pixels_tab[1]= ff_avg_h264_chroma_mc4_mmx2;
c->avg_h264_chroma_pixels_tab[2]= ff_avg_h264_chroma_mc2_mmx2;
c->put_h264_chroma_pixels_tab[2]= ff_put_h264_chroma_mc2_mmx2;
}
- if (bit_depth == 10) {
+ if (bit_depth == 10 && CONFIG_H264CHROMA) {
c->put_h264_chroma_pixels_tab[2]= ff_put_h264_chroma_mc2_10_mmxext;
c->avg_h264_chroma_pixels_tab[2]= ff_avg_h264_chroma_mc2_10_mmxext;
c->put_h264_chroma_pixels_tab[1]= ff_put_h264_chroma_mc4_10_mmxext;
SET_QPEL_FUNCS(avg_2tap_qpel, 1, 8, 3dnow, );
#if HAVE_YASM
- if (!high_bit_depth) {
+ if (!high_bit_depth && CONFIG_H264CHROMA) {
c->avg_h264_chroma_pixels_tab[0]= ff_avg_h264_chroma_mc8_3dnow_rnd;
c->avg_h264_chroma_pixels_tab[1]= ff_avg_h264_chroma_mc4_3dnow;
}
H264_QPEL_FUNCS_10(2, 0, sse2_cache64)
H264_QPEL_FUNCS_10(3, 0, sse2_cache64)
- c->put_h264_chroma_pixels_tab[0]= ff_put_h264_chroma_mc8_10_sse2;
- c->avg_h264_chroma_pixels_tab[0]= ff_avg_h264_chroma_mc8_10_sse2;
+ if (CONFIG_H264CHROMA) {
+ c->put_h264_chroma_pixels_tab[0] = ff_put_h264_chroma_mc8_10_sse2;
+ c->avg_h264_chroma_pixels_tab[0] = ff_avg_h264_chroma_mc8_10_sse2;
+ }
}
#endif
}
H264_QPEL_FUNCS_10(2, 0, ssse3_cache64)
H264_QPEL_FUNCS_10(3, 0, ssse3_cache64)
}
- if (!high_bit_depth) {
+ if (!high_bit_depth && CONFIG_H264CHROMA) {
c->put_h264_chroma_pixels_tab[0]= ff_put_h264_chroma_mc8_ssse3_rnd;
c->avg_h264_chroma_pixels_tab[0]= ff_avg_h264_chroma_mc8_ssse3_rnd;
c->put_h264_chroma_pixels_tab[1]= ff_put_h264_chroma_mc4_ssse3;
c->vector_clipf = vector_clipf_sse;
#if HAVE_YASM
c->scalarproduct_float = ff_scalarproduct_float_sse;
+ c->butterflies_float_interleave = ff_butterflies_float_interleave_sse;
+
+ if (!high_bit_depth)
+ c->emulated_edge_mc = emulated_edge_mc_sse;
+ c->gmc = gmc_sse;
#endif
}
if (HAVE_AMD3DNOW && (mm_flags & AV_CPU_FLAG_3DNOW))
c->scalarproduct_int16 = ff_scalarproduct_int16_sse2;
c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_sse2;
if (mm_flags & AV_CPU_FLAG_ATOM) {
- c->vector_clip_int32 = ff_vector_clip_int32_sse2_int;
+ c->vector_clip_int32 = ff_vector_clip_int32_int_sse2;
} else {
c->vector_clip_int32 = ff_vector_clip_int32_sse2;
}
c->apply_window_int16 = ff_apply_window_int16_sse2;
}
}
-
- if (!high_bit_depth)
- c->emulated_edge_mc = emulated_edge_mc_sse;
- c->gmc= gmc_sse;
#endif
}
if (mm_flags & AV_CPU_FLAG_SSSE3) {
if (mm_flags & AV_CPU_FLAG_SSE4 && HAVE_SSE) {
#if HAVE_YASM
- c->vector_clip_int32 = ff_vector_clip_int32_sse41;
+ c->vector_clip_int32 = ff_vector_clip_int32_sse4;
#endif
}
H264_QPEL_FUNCS_10(2, 0, sse2)
H264_QPEL_FUNCS_10(3, 0, sse2)
- c->put_h264_chroma_pixels_tab[0]= ff_put_h264_chroma_mc8_10_avx;
- c->avg_h264_chroma_pixels_tab[0]= ff_avg_h264_chroma_mc8_10_avx;
+ if (CONFIG_H264CHROMA) {
+ c->put_h264_chroma_pixels_tab[0] = ff_put_h264_chroma_mc8_10_avx;
+ c->avg_h264_chroma_pixels_tab[0] = ff_avg_h264_chroma_mc8_10_avx;
+ }
}
+ c->butterflies_float_interleave = ff_butterflies_float_interleave_avx;
}
#endif
}