DECLARE_ALIGNED(16, const uint64_t, ff_pdw_80000000)[2] =
{0x8000000080000000ULL, 0x8000000080000000ULL};
-DECLARE_ALIGNED(8, const uint64_t, ff_pw_1 ) = 0x0001000100010001ULL;
+DECLARE_ALIGNED(16, const xmm_reg, ff_pw_1 ) = {0x0001000100010001ULL, 0x0001000100010001ULL};
+DECLARE_ALIGNED(16, const xmm_reg, ff_pw_2 ) = {0x0002000200020002ULL, 0x0002000200020002ULL};
DECLARE_ALIGNED(16, const xmm_reg, ff_pw_3 ) = {0x0003000300030003ULL, 0x0003000300030003ULL};
DECLARE_ALIGNED(16, const xmm_reg, ff_pw_4 ) = {0x0004000400040004ULL, 0x0004000400040004ULL};
DECLARE_ALIGNED(16, const xmm_reg, ff_pw_5 ) = {0x0005000500050005ULL, 0x0005000500050005ULL};
dst[i+0] += src[i+0];
}
-#if HAVE_7REGS && HAVE_TEN_OPERANDS
+#if HAVE_7REGS
static void add_hfyu_median_prediction_cmov(uint8_t *dst, const uint8_t *top, const uint8_t *diff, int w, int *left, int *left_top) {
x86_reg w2 = -w;
x86_reg x;
/* draw the edges of width 'w' of an image of size width, height
this mmx version can only handle w==8 || w==16 */
-static void draw_edges_mmx(uint8_t *buf, int wrap, int width, int height, int w, int sides)
+static void draw_edges_mmx(uint8_t *buf, int wrap, int width, int height, int w, int h, int sides)
{
uint8_t *ptr, *last_line;
int i;
/* top and bottom (and hopefully also the corners) */
if (sides&EDGE_TOP) {
- for(i = 0; i < w; i += 4) {
+ for(i = 0; i < h; i += 4) {
ptr= buf - (i + 1) * wrap - w;
__asm__ volatile(
"1: \n\t"
int dxx, int dxy, int dyx, int dyy, int shift, int r, int width, int height)
{
gmc(dst, src, stride, h, ox, oy, dxx, dxy, dyx, dyy, shift, r,
- width, height, &ff_emulated_edge_mc);
+ width, height, &ff_emulated_edge_mc_8);
}
#endif
void ff_avg_h264_chroma_mc4_ssse3 (uint8_t *dst, uint8_t *src,
int stride, int h, int x, int y);
+#define CHROMA_MC(OP, NUM, DEPTH, OPT) \
+void ff_ ## OP ## _h264_chroma_mc ## NUM ## _ ## DEPTH ## _ ## OPT \
+ (uint8_t *dst, uint8_t *src,\
+ int stride, int h, int x, int y);
+
+CHROMA_MC(put, 2, 10, mmxext)
+CHROMA_MC(avg, 2, 10, mmxext)
+CHROMA_MC(put, 4, 10, mmxext)
+CHROMA_MC(avg, 4, 10, mmxext)
+CHROMA_MC(put, 8, 10, sse2)
+CHROMA_MC(avg, 8, 10, sse2)
+CHROMA_MC(put, 8, 10, avx)
+CHROMA_MC(avg, 8, 10, avx)
/* CAVS specific */
void ff_put_cavs_qpel8_mc00_mmx2(uint8_t *dst, uint8_t *src, int stride) {
float ff_scalarproduct_float_sse(const float *v1, const float *v2, int order);
+void ff_vector_clip_int32_mmx (int32_t *dst, const int32_t *src, int32_t min,
+ 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,
+ int32_t max, unsigned int len);
+void ff_vector_clip_int32_sse41 (int32_t *dst, const int32_t *src, int32_t min,
+ int32_t max, unsigned int len);
+
void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx)
{
int mm_flags = av_get_cpu_flags();
- const int h264_high_depth = avctx->codec_id == CODEC_ID_H264 && avctx->bits_per_raw_sample > 8;
+ const int high_bit_depth = avctx->codec_id == CODEC_ID_H264 && avctx->bits_per_raw_sample > 8;
+ const int bit_depth = avctx->bits_per_raw_sample;
if (avctx->dsp_mask) {
if (avctx->dsp_mask & AV_CPU_FLAG_FORCE)
c->put_pixels_clamped = ff_put_pixels_clamped_mmx;
c->put_signed_pixels_clamped = ff_put_signed_pixels_clamped_mmx;
c->add_pixels_clamped = ff_add_pixels_clamped_mmx;
- if (!h264_high_depth) {
+ if (!high_bit_depth) {
c->clear_block = clear_block_mmx;
c->clear_blocks = clear_blocks_mmx;
if ((mm_flags & AV_CPU_FLAG_SSE) &&
c->PFX ## _pixels_tab[IDX][2] = PFX ## _pixels ## SIZE ## _y2_ ## CPU; \
c->PFX ## _pixels_tab[IDX][3] = PFX ## _pixels ## SIZE ## _xy2_ ## CPU
- if (!h264_high_depth) {
+ if (!high_bit_depth) {
SET_HPEL_FUNCS(put, 0, 16, mmx);
SET_HPEL_FUNCS(put_no_rnd, 0, 16, mmx);
SET_HPEL_FUNCS(avg, 0, 16, mmx);
c->gmc= gmc_mmx;
#endif
#if ARCH_X86_32 && HAVE_YASM
- if (!h264_high_depth)
+ if (!high_bit_depth)
c->emulated_edge_mc = emulated_edge_mc_mmx;
#endif
c->add_bytes= add_bytes_mmx;
- if (!h264_high_depth)
+ if (!high_bit_depth)
c->draw_edges = draw_edges_mmx;
if (CONFIG_H263_DECODER || CONFIG_H263_ENCODER) {
}
#if HAVE_YASM
- if (!h264_high_depth) {
+ if (!high_bit_depth) {
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;
}
c->put_rv40_chroma_pixels_tab[0]= ff_put_rv40_chroma_mc8_mmx;
c->put_rv40_chroma_pixels_tab[1]= ff_put_rv40_chroma_mc4_mmx;
+
+ c->vector_clip_int32 = ff_vector_clip_int32_mmx;
#endif
if (mm_flags & AV_CPU_FLAG_MMX2) {
c->prefetch = prefetch_mmx2;
- if (!h264_high_depth) {
+ if (!high_bit_depth) {
c->put_pixels_tab[0][1] = put_pixels16_x2_mmx2;
c->put_pixels_tab[0][2] = put_pixels16_y2_mmx2;
}
if(!(avctx->flags & CODEC_FLAG_BITEXACT)){
- if (!h264_high_depth) {
+ if (!high_bit_depth) {
c->put_no_rnd_pixels_tab[0][1] = put_no_rnd_pixels16_x2_mmx2;
c->put_no_rnd_pixels_tab[0][2] = put_no_rnd_pixels16_y2_mmx2;
c->put_no_rnd_pixels_tab[1][1] = put_no_rnd_pixels8_x2_mmx2;
c->put_no_rnd_pixels_tab[1][2] = put_no_rnd_pixels8_y2_exact_mmx2;
}
-#define SET_QPEL_FUNCS(PFX, IDX, SIZE, CPU) \
- c->PFX ## _pixels_tab[IDX][ 0] = PFX ## SIZE ## _mc00_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][ 1] = PFX ## SIZE ## _mc10_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][ 2] = PFX ## SIZE ## _mc20_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][ 3] = PFX ## SIZE ## _mc30_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][ 4] = PFX ## SIZE ## _mc01_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][ 5] = PFX ## SIZE ## _mc11_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][ 6] = PFX ## SIZE ## _mc21_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][ 7] = PFX ## SIZE ## _mc31_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][ 8] = PFX ## SIZE ## _mc02_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][ 9] = PFX ## SIZE ## _mc12_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][10] = PFX ## SIZE ## _mc22_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][11] = PFX ## SIZE ## _mc32_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][12] = PFX ## SIZE ## _mc03_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][13] = PFX ## SIZE ## _mc13_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][14] = PFX ## SIZE ## _mc23_ ## CPU; \
- c->PFX ## _pixels_tab[IDX][15] = PFX ## SIZE ## _mc33_ ## CPU
-
- SET_QPEL_FUNCS(put_qpel, 0, 16, mmx2);
- SET_QPEL_FUNCS(put_qpel, 1, 8, mmx2);
- SET_QPEL_FUNCS(put_no_rnd_qpel, 0, 16, mmx2);
- SET_QPEL_FUNCS(put_no_rnd_qpel, 1, 8, mmx2);
- SET_QPEL_FUNCS(avg_qpel, 0, 16, mmx2);
- SET_QPEL_FUNCS(avg_qpel, 1, 8, mmx2);
-
- if (!h264_high_depth) {
- SET_QPEL_FUNCS(put_h264_qpel, 0, 16, mmx2);
- SET_QPEL_FUNCS(put_h264_qpel, 1, 8, mmx2);
- SET_QPEL_FUNCS(put_h264_qpel, 2, 4, mmx2);
- SET_QPEL_FUNCS(avg_h264_qpel, 0, 16, mmx2);
- SET_QPEL_FUNCS(avg_h264_qpel, 1, 8, mmx2);
- SET_QPEL_FUNCS(avg_h264_qpel, 2, 4, mmx2);
+#define SET_QPEL_FUNCS(PFX, IDX, SIZE, CPU, PREFIX) \
+ c->PFX ## _pixels_tab[IDX][ 0] = PREFIX ## PFX ## SIZE ## _mc00_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][ 1] = PREFIX ## PFX ## SIZE ## _mc10_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][ 2] = PREFIX ## PFX ## SIZE ## _mc20_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][ 3] = PREFIX ## PFX ## SIZE ## _mc30_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][ 4] = PREFIX ## PFX ## SIZE ## _mc01_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][ 5] = PREFIX ## PFX ## SIZE ## _mc11_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][ 6] = PREFIX ## PFX ## SIZE ## _mc21_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][ 7] = PREFIX ## PFX ## SIZE ## _mc31_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][ 8] = PREFIX ## PFX ## SIZE ## _mc02_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][ 9] = PREFIX ## PFX ## SIZE ## _mc12_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][10] = PREFIX ## PFX ## SIZE ## _mc22_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][11] = PREFIX ## PFX ## SIZE ## _mc32_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][12] = PREFIX ## PFX ## SIZE ## _mc03_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][13] = PREFIX ## PFX ## SIZE ## _mc13_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][14] = PREFIX ## PFX ## SIZE ## _mc23_ ## CPU; \
+ c->PFX ## _pixels_tab[IDX][15] = PREFIX ## PFX ## SIZE ## _mc33_ ## CPU
+
+ SET_QPEL_FUNCS(put_qpel, 0, 16, mmx2, );
+ SET_QPEL_FUNCS(put_qpel, 1, 8, mmx2, );
+ SET_QPEL_FUNCS(put_no_rnd_qpel, 0, 16, mmx2, );
+ SET_QPEL_FUNCS(put_no_rnd_qpel, 1, 8, mmx2, );
+ SET_QPEL_FUNCS(avg_qpel, 0, 16, mmx2, );
+ SET_QPEL_FUNCS(avg_qpel, 1, 8, mmx2, );
+
+ if (!high_bit_depth) {
+ SET_QPEL_FUNCS(put_h264_qpel, 0, 16, mmx2, );
+ SET_QPEL_FUNCS(put_h264_qpel, 1, 8, mmx2, );
+ SET_QPEL_FUNCS(put_h264_qpel, 2, 4, mmx2, );
+ SET_QPEL_FUNCS(avg_h264_qpel, 0, 16, mmx2, );
+ SET_QPEL_FUNCS(avg_h264_qpel, 1, 8, mmx2, );
+ SET_QPEL_FUNCS(avg_h264_qpel, 2, 4, mmx2, );
+ }
+ else if (bit_depth == 10) {
+#if HAVE_YASM
+#if !ARCH_X86_64
+ SET_QPEL_FUNCS(avg_h264_qpel, 0, 16, 10_mmxext, ff_);
+ SET_QPEL_FUNCS(put_h264_qpel, 0, 16, 10_mmxext, ff_);
+ SET_QPEL_FUNCS(put_h264_qpel, 1, 8, 10_mmxext, ff_);
+ SET_QPEL_FUNCS(avg_h264_qpel, 1, 8, 10_mmxext, ff_);
+#endif
+ SET_QPEL_FUNCS(put_h264_qpel, 2, 4, 10_mmxext, ff_);
+ SET_QPEL_FUNCS(avg_h264_qpel, 2, 4, 10_mmxext, ff_);
+#endif
}
- SET_QPEL_FUNCS(put_2tap_qpel, 0, 16, mmx2);
- SET_QPEL_FUNCS(put_2tap_qpel, 1, 8, mmx2);
- SET_QPEL_FUNCS(avg_2tap_qpel, 0, 16, mmx2);
- SET_QPEL_FUNCS(avg_2tap_qpel, 1, 8, mmx2);
+ SET_QPEL_FUNCS(put_2tap_qpel, 0, 16, mmx2, );
+ SET_QPEL_FUNCS(put_2tap_qpel, 1, 8, mmx2, );
+ SET_QPEL_FUNCS(avg_2tap_qpel, 0, 16, mmx2, );
+ SET_QPEL_FUNCS(avg_2tap_qpel, 1, 8, mmx2, );
#if HAVE_YASM
c->avg_rv40_chroma_pixels_tab[0]= ff_avg_rv40_chroma_mc8_mmx2;
c->avg_rv40_chroma_pixels_tab[1]= ff_avg_rv40_chroma_mc4_mmx2;
- if (!h264_high_depth) {
+ if (!high_bit_depth) {
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) {
+ 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;
+ c->avg_h264_chroma_pixels_tab[1]= ff_avg_h264_chroma_mc4_10_mmxext;
+ }
c->add_hfyu_median_prediction = ff_add_hfyu_median_prediction_mmx2;
#endif
-#if HAVE_7REGS && HAVE_TEN_OPERANDS
+#if HAVE_7REGS
if( mm_flags&AV_CPU_FLAG_3DNOW )
c->add_hfyu_median_prediction = add_hfyu_median_prediction_cmov;
#endif
} else if (mm_flags & AV_CPU_FLAG_3DNOW) {
c->prefetch = prefetch_3dnow;
- if (!h264_high_depth) {
+ if (!high_bit_depth) {
c->put_pixels_tab[0][1] = put_pixels16_x2_3dnow;
c->put_pixels_tab[0][2] = put_pixels16_y2_3dnow;
c->put_no_rnd_pixels_tab[1][2] = put_no_rnd_pixels8_y2_exact_3dnow;
}
- SET_QPEL_FUNCS(put_qpel, 0, 16, 3dnow);
- SET_QPEL_FUNCS(put_qpel, 1, 8, 3dnow);
- SET_QPEL_FUNCS(put_no_rnd_qpel, 0, 16, 3dnow);
- SET_QPEL_FUNCS(put_no_rnd_qpel, 1, 8, 3dnow);
- SET_QPEL_FUNCS(avg_qpel, 0, 16, 3dnow);
- SET_QPEL_FUNCS(avg_qpel, 1, 8, 3dnow);
-
- if (!h264_high_depth) {
- SET_QPEL_FUNCS(put_h264_qpel, 0, 16, 3dnow);
- SET_QPEL_FUNCS(put_h264_qpel, 1, 8, 3dnow);
- SET_QPEL_FUNCS(put_h264_qpel, 2, 4, 3dnow);
- SET_QPEL_FUNCS(avg_h264_qpel, 0, 16, 3dnow);
- SET_QPEL_FUNCS(avg_h264_qpel, 1, 8, 3dnow);
- SET_QPEL_FUNCS(avg_h264_qpel, 2, 4, 3dnow);
+ SET_QPEL_FUNCS(put_qpel, 0, 16, 3dnow, );
+ SET_QPEL_FUNCS(put_qpel, 1, 8, 3dnow, );
+ SET_QPEL_FUNCS(put_no_rnd_qpel, 0, 16, 3dnow, );
+ SET_QPEL_FUNCS(put_no_rnd_qpel, 1, 8, 3dnow, );
+ SET_QPEL_FUNCS(avg_qpel, 0, 16, 3dnow, );
+ SET_QPEL_FUNCS(avg_qpel, 1, 8, 3dnow, );
+
+ if (!high_bit_depth) {
+ SET_QPEL_FUNCS(put_h264_qpel, 0, 16, 3dnow, );
+ SET_QPEL_FUNCS(put_h264_qpel, 1, 8, 3dnow, );
+ SET_QPEL_FUNCS(put_h264_qpel, 2, 4, 3dnow, );
+ SET_QPEL_FUNCS(avg_h264_qpel, 0, 16, 3dnow, );
+ SET_QPEL_FUNCS(avg_h264_qpel, 1, 8, 3dnow, );
+ SET_QPEL_FUNCS(avg_h264_qpel, 2, 4, 3dnow, );
}
- SET_QPEL_FUNCS(put_2tap_qpel, 0, 16, 3dnow);
- SET_QPEL_FUNCS(put_2tap_qpel, 1, 8, 3dnow);
- SET_QPEL_FUNCS(avg_2tap_qpel, 0, 16, 3dnow);
- SET_QPEL_FUNCS(avg_2tap_qpel, 1, 8, 3dnow);
+ SET_QPEL_FUNCS(put_2tap_qpel, 0, 16, 3dnow, );
+ SET_QPEL_FUNCS(put_2tap_qpel, 1, 8, 3dnow, );
+ SET_QPEL_FUNCS(avg_2tap_qpel, 0, 16, 3dnow, );
+ SET_QPEL_FUNCS(avg_2tap_qpel, 1, 8, 3dnow, );
#if HAVE_YASM
- if (!h264_high_depth) {
+ if (!high_bit_depth) {
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;
}
c->avg_h264_qpel_pixels_tab[1][x+y*4] = avg_h264_qpel8_mc##x##y##_##CPU;
if((mm_flags & AV_CPU_FLAG_SSE2) && !(mm_flags & AV_CPU_FLAG_3DNOW)){
// these functions are slower than mmx on AMD, but faster on Intel
- if (!h264_high_depth) {
+ if (!high_bit_depth) {
c->put_pixels_tab[0][0] = put_pixels16_sse2;
c->put_no_rnd_pixels_tab[0][0] = put_pixels16_sse2;
c->avg_pixels_tab[0][0] = avg_pixels16_sse2;
}
}
if(mm_flags & AV_CPU_FLAG_SSE2){
- if (!h264_high_depth) {
+ if (!high_bit_depth) {
H264_QPEL_FUNCS(0, 1, sse2);
H264_QPEL_FUNCS(0, 2, sse2);
H264_QPEL_FUNCS(0, 3, sse2);
H264_QPEL_FUNCS(3, 2, sse2);
H264_QPEL_FUNCS(3, 3, sse2);
}
+#if HAVE_YASM
+#define H264_QPEL_FUNCS_10(x, y, CPU)\
+ c->put_h264_qpel_pixels_tab[0][x+y*4] = ff_put_h264_qpel16_mc##x##y##_10_##CPU;\
+ c->put_h264_qpel_pixels_tab[1][x+y*4] = ff_put_h264_qpel8_mc##x##y##_10_##CPU;\
+ c->avg_h264_qpel_pixels_tab[0][x+y*4] = ff_avg_h264_qpel16_mc##x##y##_10_##CPU;\
+ c->avg_h264_qpel_pixels_tab[1][x+y*4] = ff_avg_h264_qpel8_mc##x##y##_10_##CPU;
+ if (bit_depth == 10) {
+ SET_QPEL_FUNCS(put_h264_qpel, 0, 16, 10_sse2, ff_);
+ SET_QPEL_FUNCS(put_h264_qpel, 1, 8, 10_sse2, ff_);
+ SET_QPEL_FUNCS(avg_h264_qpel, 0, 16, 10_sse2, ff_);
+ SET_QPEL_FUNCS(avg_h264_qpel, 1, 8, 10_sse2, ff_);
+ H264_QPEL_FUNCS_10(1, 0, sse2_cache64)
+ 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;
+ }
+#endif
}
#if HAVE_SSSE3
if(mm_flags & AV_CPU_FLAG_SSSE3){
- if (!h264_high_depth) {
+ if (!high_bit_depth) {
H264_QPEL_FUNCS(1, 0, ssse3);
H264_QPEL_FUNCS(1, 1, ssse3);
H264_QPEL_FUNCS(1, 2, ssse3);
H264_QPEL_FUNCS(3, 3, ssse3);
}
#if HAVE_YASM
- if (!h264_high_depth) {
+ else if (bit_depth == 10) {
+ H264_QPEL_FUNCS_10(1, 0, ssse3_cache64)
+ H264_QPEL_FUNCS_10(2, 0, ssse3_cache64)
+ H264_QPEL_FUNCS_10(3, 0, ssse3_cache64)
+ }
+ if (!high_bit_depth) {
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;
#if HAVE_YASM
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;
+ } else {
+ c->vector_clip_int32 = ff_vector_clip_int32_sse2;
+ }
if (avctx->flags & CODEC_FLAG_BITEXACT) {
c->apply_window_int16 = ff_apply_window_int16_sse2_ba;
} else {
}
}
- if (!h264_high_depth)
+ if (!high_bit_depth)
c->emulated_edge_mc = emulated_edge_mc_sse;
c->gmc= gmc_sse;
#endif
}
#endif
}
+
+ if (mm_flags & AV_CPU_FLAG_SSE4 && HAVE_SSE) {
+#if HAVE_YASM
+ c->vector_clip_int32 = ff_vector_clip_int32_sse41;
+#endif
+ }
+
+#if HAVE_AVX && HAVE_YASM
+ if (mm_flags & AV_CPU_FLAG_AVX) {
+ if (bit_depth == 10) {
+ //AVX implies !cache64.
+ //TODO: Port cache(32|64) detection from x264.
+ H264_QPEL_FUNCS_10(1, 0, sse2)
+ 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;
+ }
+ }
+#endif
}
if (CONFIG_ENCODERS)