diff options
Diffstat (limited to 'vp9/common/vp9_rtcd_defs.sh')
| -rw-r--r-- | vp9/common/vp9_rtcd_defs.sh | 482 |
1 files changed, 203 insertions, 279 deletions
diff --git a/vp9/common/vp9_rtcd_defs.sh b/vp9/common/vp9_rtcd_defs.sh index 43bc3cb1f..f281e08e9 100644 --- a/vp9/common/vp9_rtcd_defs.sh +++ b/vp9/common/vp9_rtcd_defs.sh @@ -5,14 +5,13 @@ cat <<EOF */ #include "vpx/vpx_integer.h" +#include "vp9/common/vp9_enums.h" struct loop_filter_info; -struct blockd; struct macroblockd; struct loop_filter_info; /* Encoder forward decls */ -struct block; struct macroblock; struct vp9_variance_vtable; @@ -26,32 +25,26 @@ forward_decls vp9_common_forward_decls # # Dequant # -prototype void vp9_dequant_idct_add_y_block_8x8 "int16_t *q, const int16_t *dq, uint8_t *pre, uint8_t *dst, int stride, struct macroblockd *xd" -specialize vp9_dequant_idct_add_y_block_8x8 +prototype void vp9_idct_add_y_block_8x8 "int16_t *q, uint8_t *dst, int stride, struct macroblockd *xd" +specialize vp9_idct_add_y_block_8x8 -prototype void vp9_dequant_idct_add_uv_block_8x8 "int16_t *q, const int16_t *dq, uint8_t *pre, uint8_t *dstu, uint8_t *dstv, int stride, struct macroblockd *xd" -specialize vp9_dequant_idct_add_uv_block_8x8 +prototype void vp9_idct_add_16x16 "int16_t *input, uint8_t *dest, int stride, int eob" +specialize vp9_idct_add_16x16 -prototype void vp9_dequant_idct_add_16x16 "int16_t *input, const int16_t *dq, uint8_t *pred, uint8_t *dest, int pitch, int stride, int eob" -specialize vp9_dequant_idct_add_16x16 +prototype void vp9_idct_add_8x8 "int16_t *input, uint8_t *dest, int stride, int eob" +specialize vp9_idct_add_8x8 -prototype void vp9_dequant_idct_add_8x8 "int16_t *input, const int16_t *dq, uint8_t *pred, uint8_t *dest, int pitch, int stride, int eob" -specialize vp9_dequant_idct_add_8x8 +prototype void vp9_idct_add "int16_t *input, uint8_t *dest, int stride, int eob" +specialize vp9_idct_add -prototype void vp9_dequant_idct_add "int16_t *input, const int16_t *dq, uint8_t *pred, uint8_t *dest, int pitch, int stride, int eob" -specialize vp9_dequant_idct_add +prototype void vp9_idct_add_y_block "int16_t *q, uint8_t *dst, int stride, struct macroblockd *xd" +specialize vp9_idct_add_y_block -prototype void vp9_dequant_idct_add_y_block "int16_t *q, const int16_t *dq, uint8_t *pre, uint8_t *dst, int stride, struct macroblockd *xd" -specialize vp9_dequant_idct_add_y_block +prototype void vp9_idct_add_uv_block "int16_t *q, uint8_t *dst, int stride, uint16_t *eobs" +specialize vp9_idct_add_uv_block -prototype void vp9_dequant_idct_add_uv_block "int16_t *q, const int16_t *dq, uint8_t *pre, uint8_t *dstu, uint8_t *dstv, int stride, struct macroblockd *xd" -specialize vp9_dequant_idct_add_uv_block - -prototype void vp9_dequant_idct_add_32x32 "int16_t *q, const int16_t *dq, uint8_t *pre, uint8_t *dst, int pitch, int stride, int eob" -specialize vp9_dequant_idct_add_32x32 - -prototype void vp9_dequant_idct_add_uv_block_16x16 "int16_t *q, const int16_t *dq, uint8_t *dstu, uint8_t *dstv, int stride, struct macroblockd *xd" -specialize vp9_dequant_idct_add_uv_block_16x16 +prototype void vp9_idct_add_32x32 "int16_t *q, uint8_t *dst, int stride, int eob" +specialize vp9_idct_add_32x32 # # RECON @@ -67,151 +60,49 @@ vp9_copy_mem8x8_dspr2=vp9_copy_mem8x8_dspr2 prototype void vp9_copy_mem8x4 "const uint8_t *src, int src_pitch, uint8_t *dst, int dst_pitch" specialize vp9_copy_mem8x4 mmx -prototype void vp9_recon_b "uint8_t *pred_ptr, int16_t *diff_ptr, uint8_t *dst_ptr, int stride" -specialize vp9_recon_b - -prototype void vp9_recon_uv_b "uint8_t *pred_ptr, int16_t *diff_ptr, uint8_t *dst_ptr, int stride" -specialize vp9_recon_uv_b - -prototype void vp9_recon2b "uint8_t *pred_ptr, int16_t *diff_ptr, uint8_t *dst_ptr, int stride" -specialize vp9_recon2b sse2 - -prototype void vp9_recon4b "uint8_t *pred_ptr, int16_t *diff_ptr, uint8_t *dst_ptr, int stride" -specialize vp9_recon4b sse2 - -prototype void vp9_recon_mb "struct macroblockd *x" -specialize vp9_recon_mb - -prototype void vp9_recon_mby "struct macroblockd *x" -specialize vp9_recon_mby - -prototype void vp9_recon_mby_s "struct macroblockd *x, uint8_t *dst" -specialize vp9_recon_mby_s - -prototype void vp9_recon_mbuv_s "struct macroblockd *x, uint8_t *udst, uint8_t *vdst" -specialize void vp9_recon_mbuv_s - -prototype void vp9_recon_sby_s "struct macroblockd *x, uint8_t *dst" -specialize vp9_recon_sby_s - -prototype void vp9_recon_sbuv_s "struct macroblockd *x, uint8_t *udst, uint8_t *vdst" -specialize void vp9_recon_sbuv_s - -prototype void vp9_recon_sb64y_s "struct macroblockd *x, uint8_t *dst" -specialize vp9_recon_sb64y_s - -prototype void vp9_recon_sb64uv_s "struct macroblockd *x, uint8_t *udst, uint8_t *vdst" -specialize void vp9_recon_sb64uv_s - -prototype void vp9_build_intra_predictors_mby_s "struct macroblockd *x" -specialize vp9_build_intra_predictors_mby_s +prototype void vp9_build_intra_predictors "uint8_t *src, int src_stride, uint8_t *pred, int y_stride, int mode, int bw, int bh, int up_available, int left_available, int right_available" +specialize void vp9_build_intra_predictors -prototype void vp9_build_intra_predictors_sby_s "struct macroblockd *x" -specialize vp9_build_intra_predictors_sby_s; +prototype void vp9_build_intra_predictors_sby_s "struct macroblockd *x, enum BLOCK_SIZE_TYPE bsize" +specialize vp9_build_intra_predictors_sby_s -prototype void vp9_build_intra_predictors_sbuv_s "struct macroblockd *x" -specialize vp9_build_intra_predictors_sbuv_s; +prototype void vp9_build_intra_predictors_sbuv_s "struct macroblockd *x, enum BLOCK_SIZE_TYPE bsize" +specialize vp9_build_intra_predictors_sbuv_s -prototype void vp9_build_intra_predictors_mby "struct macroblockd *x" -specialize vp9_build_intra_predictors_mby; - -prototype void vp9_build_intra_predictors_mby_s "struct macroblockd *x" -specialize vp9_build_intra_predictors_mby_s; - -prototype void vp9_build_intra_predictors_mbuv "struct macroblockd *x" -specialize vp9_build_intra_predictors_mbuv; - -prototype void vp9_build_intra_predictors_mbuv_s "struct macroblockd *x" -specialize vp9_build_intra_predictors_mbuv_s; - -prototype void vp9_build_intra_predictors_sb64y_s "struct macroblockd *x" -specialize vp9_build_intra_predictors_sb64y_s; - -prototype void vp9_build_intra_predictors_sb64uv_s "struct macroblockd *x" -specialize vp9_build_intra_predictors_sb64uv_s; - -prototype void vp9_intra4x4_predict "struct macroblockd *xd, struct blockd *x, int b_mode, uint8_t *predictor" +prototype void vp9_intra4x4_predict "struct macroblockd *xd, int block, enum BLOCK_SIZE_TYPE bsize, int b_mode, uint8_t *predictor, int pre_stride" specialize vp9_intra4x4_predict; -prototype void vp9_intra8x8_predict "struct macroblockd *xd, struct blockd *x, int b_mode, uint8_t *predictor" -specialize vp9_intra8x8_predict; - -prototype void vp9_intra_uv4x4_predict "struct macroblockd *xd, struct blockd *x, int b_mode, uint8_t *predictor" -specialize vp9_intra_uv4x4_predict; - if [ "$CONFIG_VP9_DECODER" = "yes" ]; then -prototype void vp9_add_residual_4x4 "const int16_t *diff, const uint8_t *pred, int pitch, uint8_t *dest, int stride" -specialize vp9_add_residual_4x4 sse2 - -prototype void vp9_add_residual_8x8 "const int16_t *diff, const uint8_t *pred, int pitch, uint8_t *dest, int stride" -specialize vp9_add_residual_8x8 sse2 - -prototype void vp9_add_residual_16x16 "const int16_t *diff, const uint8_t *pred, int pitch, uint8_t *dest, int stride" -specialize vp9_add_residual_16x16 sse2 - -prototype void vp9_add_residual_32x32 "const int16_t *diff, const uint8_t *pred, int pitch, uint8_t *dest, int stride" -specialize vp9_add_residual_32x32 sse2 - -prototype void vp9_add_constant_residual_8x8 "const int16_t diff, const uint8_t *pred, int pitch, uint8_t *dest, int stride" +prototype void vp9_add_constant_residual_8x8 "const int16_t diff, uint8_t *dest, int stride" specialize vp9_add_constant_residual_8x8 sse2 -prototype void vp9_add_constant_residual_16x16 "const int16_t diff, const uint8_t *pred, int pitch, uint8_t *dest, int stride" +prototype void vp9_add_constant_residual_16x16 "const int16_t diff, uint8_t *dest, int stride" specialize vp9_add_constant_residual_16x16 sse2 -prototype void vp9_add_constant_residual_32x32 "const int16_t diff, const uint8_t *pred, int pitch, uint8_t *dest, int stride" +prototype void vp9_add_constant_residual_32x32 "const int16_t diff, uint8_t *dest, int stride" specialize vp9_add_constant_residual_32x32 sse2 fi # # Loopfilter # -prototype void vp9_loop_filter_mbv "uint8_t *y, uint8_t *u, uint8_t *v, int ystride, int uv_stride, struct loop_filter_info *lfi" -specialize vp9_loop_filter_mbv sse2 - -prototype void vp9_loop_filter_bv "uint8_t *y, uint8_t *u, uint8_t *v, int ystride, int uv_stride, struct loop_filter_info *lfi" -specialize vp9_loop_filter_bv sse2 +prototype void vp9_mb_lpf_vertical_edge_w "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count" +specialize vp9_mb_lpf_vertical_edge_w -prototype void vp9_loop_filter_bv8x8 "uint8_t *y, uint8_t *u, uint8_t *v, int ystride, int uv_stride, struct loop_filter_info *lfi" -specialize vp9_loop_filter_bv8x8 sse2 +prototype void vp9_mbloop_filter_vertical_edge "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count" +specialize vp9_mbloop_filter_vertical_edge -prototype void vp9_loop_filter_mbh "uint8_t *y, uint8_t *u, uint8_t *v, int ystride, int uv_stride, struct loop_filter_info *lfi" -specialize vp9_loop_filter_mbh sse2 +prototype void vp9_loop_filter_vertical_edge "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count" +specialize vp9_loop_filter_vertical_edge -prototype void vp9_loop_filter_bh "uint8_t *y, uint8_t *u, uint8_t *v, int ystride, int uv_stride, struct loop_filter_info *lfi" -specialize vp9_loop_filter_bh sse2 +prototype void vp9_mb_lpf_horizontal_edge_w "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count" +specialize vp9_mb_lpf_horizontal_edge_w -prototype void vp9_loop_filter_bh8x8 "uint8_t *y, uint8_t *u, uint8_t *v, int ystride, int uv_stride, struct loop_filter_info *lfi" -specialize vp9_loop_filter_bh8x8 sse2 +prototype void vp9_mbloop_filter_horizontal_edge "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count" +specialize vp9_mbloop_filter_horizontal_edge -prototype void vp9_loop_filter_simple_mbv "uint8_t *y, int ystride, const uint8_t *blimit" -specialize vp9_loop_filter_simple_mbv mmx sse2 -vp9_loop_filter_simple_mbv_c=vp9_loop_filter_simple_vertical_edge_c -vp9_loop_filter_simple_mbv_mmx=vp9_loop_filter_simple_vertical_edge_mmx -vp9_loop_filter_simple_mbv_sse2=vp9_loop_filter_simple_vertical_edge_sse2 - -prototype void vp9_loop_filter_simple_mbh "uint8_t *y, int ystride, const uint8_t *blimit" -specialize vp9_loop_filter_simple_mbh mmx sse2 -vp9_loop_filter_simple_mbh_c=vp9_loop_filter_simple_horizontal_edge_c -vp9_loop_filter_simple_mbh_mmx=vp9_loop_filter_simple_horizontal_edge_mmx -vp9_loop_filter_simple_mbh_sse2=vp9_loop_filter_simple_horizontal_edge_sse2 - -prototype void vp9_loop_filter_simple_bv "uint8_t *y, int ystride, const uint8_t *blimit" -specialize vp9_loop_filter_simple_bv mmx sse2 -vp9_loop_filter_simple_bv_c=vp9_loop_filter_bvs_c -vp9_loop_filter_simple_bv_mmx=vp9_loop_filter_bvs_mmx -vp9_loop_filter_simple_bv_sse2=vp9_loop_filter_bvs_sse2 - -prototype void vp9_loop_filter_simple_bh "uint8_t *y, int ystride, const uint8_t *blimit" -specialize vp9_loop_filter_simple_bh mmx sse2 -vp9_loop_filter_simple_bh_c=vp9_loop_filter_bhs_c -vp9_loop_filter_simple_bh_mmx=vp9_loop_filter_bhs_mmx -vp9_loop_filter_simple_bh_sse2=vp9_loop_filter_bhs_sse2 - -prototype void vp9_lpf_mbh_w "unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr, int y_stride, int uv_stride, struct loop_filter_info *lfi" -specialize vp9_lpf_mbh_w sse2 - -prototype void vp9_lpf_mbv_w "unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr, int y_stride, int uv_stride, struct loop_filter_info *lfi" -specialize vp9_lpf_mbv_w sse2 +prototype void vp9_loop_filter_horizontal_edge "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count" +specialize vp9_loop_filter_horizontal_edge # # post proc @@ -225,7 +116,7 @@ prototype void vp9_mbpost_proc_across_ip "uint8_t *src, int pitch, int rows, int specialize vp9_mbpost_proc_across_ip sse2 vp9_mbpost_proc_across_ip_sse2=vp9_mbpost_proc_across_ip_xmm -prototype void vp9_post_proc_down_and_across "uint8_t *src_ptr, uint8_t *dst_ptr, int src_pixels_per_line, int dst_pixels_per_line, int rows, int cols, int flimit" +prototype void vp9_post_proc_down_and_across "const uint8_t *src_ptr, uint8_t *dst_ptr, int src_pixels_per_line, int dst_pixels_per_line, int rows, int cols, int flimit" specialize vp9_post_proc_down_and_across mmx sse2 vp9_post_proc_down_and_across_sse2=vp9_post_proc_down_and_across_xmm @@ -244,18 +135,6 @@ prototype void vp9_blend_b "uint8_t *y, uint8_t *u, uint8_t *v, int y1, int u1, specialize vp9_blend_b # -# sad 16x3, 3x16 -# -prototype unsigned int vp9_sad16x3 "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride" -specialize vp9_sad16x3 sse2 - -prototype unsigned int vp9_sad3x16 "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride" -specialize vp9_sad3x16 sse2 - -prototype unsigned int vp9_sub_pixel_variance16x2 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" -specialize vp9_sub_pixel_variance16x2 sse2 - -# # Sub Pixel Filters # prototype void vp9_convolve8 "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" @@ -276,122 +155,63 @@ specialize vp9_convolve8_avg_horiz ssse3 prototype void vp9_convolve8_avg_vert "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" specialize vp9_convolve8_avg_vert ssse3 -#if CONFIG_IMPLICIT_COMPOUNDINTER_WEIGHT -prototype void vp9_convolve8_1by8 "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_1by8 - -prototype void vp9_convolve8_qtr "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_qtr - -prototype void vp9_convolve8_3by8 "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_3by8 - -prototype void vp9_convolve8_5by8 "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_5by8 - -prototype void vp9_convolve8_3qtr "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_3qtr - -prototype void vp9_convolve8_7by8 "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_7by8 - -prototype void vp9_convolve8_1by8_horiz "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_1by8_horiz - -prototype void vp9_convolve8_qtr_horiz "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_qtr_horiz - -prototype void vp9_convolve8_3by8_horiz "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_3by8_horiz - -prototype void vp9_convolve8_5by8_horiz "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_5by8_horiz - -prototype void vp9_convolve8_3qtr_horiz "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_3qtr_horiz - -prototype void vp9_convolve8_7by8_horiz "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_7by8_horiz - -prototype void vp9_convolve8_1by8_vert "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_1by8_vert - -prototype void vp9_convolve8_qtr_vert "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_qtr_vert - -prototype void vp9_convolve8_3by8_vert "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_3by8_vert - -prototype void vp9_convolve8_5by8_vert "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_5by8_vert - -prototype void vp9_convolve8_3qtr_vert "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_3qtr_vert - -prototype void vp9_convolve8_7by8_vert "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h" -specialize vp9_convolve8_7by8_vert -#endif - # # dct # -prototype void vp9_short_idct4x4_1 "int16_t *input, int16_t *output, int pitch" -specialize vp9_short_idct4x4_1 +prototype void vp9_short_idct4x4_1_add "int16_t *input, uint8_t *dest, int dest_stride" +specialize vp9_short_idct4x4_1_add -prototype void vp9_short_idct4x4 "int16_t *input, int16_t *output, int pitch" -specialize vp9_short_idct4x4 sse2 +prototype void vp9_short_idct4x4_add "int16_t *input, uint8_t *dest, int dest_stride" +specialize vp9_short_idct4x4_add sse2 -prototype void vp9_short_idct8x8 "int16_t *input, int16_t *output, int pitch" -specialize vp9_short_idct8x8 sse2 +prototype void vp9_short_idct8x8_add "int16_t *input, uint8_t *dest, int dest_stride" +specialize vp9_short_idct8x8_add sse2 -prototype void vp9_short_idct10_8x8 "int16_t *input, int16_t *output, int pitch" -specialize vp9_short_idct10_8x8 sse2 +prototype void vp9_short_idct10_8x8_add "int16_t *input, uint8_t *dest, int dest_stride" +specialize vp9_short_idct10_8x8_add sse2 prototype void vp9_short_idct1_8x8 "int16_t *input, int16_t *output" specialize vp9_short_idct1_8x8 -prototype void vp9_short_idct16x16 "int16_t *input, int16_t *output, int pitch" -specialize vp9_short_idct16x16 sse2 +prototype void vp9_short_idct16x16_add "int16_t *input, uint8_t *dest, int dest_stride" +specialize vp9_short_idct16x16_add sse2 -prototype void vp9_short_idct10_16x16 "int16_t *input, int16_t *output, int pitch" -specialize vp9_short_idct10_16x16 sse2 +prototype void vp9_short_idct10_16x16_add "int16_t *input, uint8_t *dest, int dest_stride" +specialize vp9_short_idct10_16x16_add sse2 prototype void vp9_short_idct1_16x16 "int16_t *input, int16_t *output" specialize vp9_short_idct1_16x16 - -prototype void vp9_short_idct32x32 "int16_t *input, int16_t *output, int pitch" -specialize vp9_short_idct32x32 sse2 +prototype void vp9_short_idct32x32_add "int16_t *input, uint8_t *dest, int dest_stride" +specialize vp9_short_idct32x32_add sse2 prototype void vp9_short_idct1_32x32 "int16_t *input, int16_t *output" specialize vp9_short_idct1_32x32 -prototype void vp9_short_idct10_32x32 "int16_t *input, int16_t *output, int pitch" -specialize vp9_short_idct10_32x32 +prototype void vp9_short_idct10_32x32_add "int16_t *input, uint8_t *dest, int dest_stride" +specialize vp9_short_idct10_32x32_add -prototype void vp9_short_iht8x8 "int16_t *input, int16_t *output, int pitch, int tx_type" -specialize vp9_short_iht8x8 +prototype void vp9_short_iht4x4_add "int16_t *input, uint8_t *dest, int dest_stride, int tx_type" +specialize vp9_short_iht4x4_add -prototype void vp9_short_iht4x4 "int16_t *input, int16_t *output, int pitch, int tx_type" -specialize vp9_short_iht4x4 +prototype void vp9_short_iht8x8_add "int16_t *input, uint8_t *dest, int dest_stride, int tx_type" +specialize vp9_short_iht8x8_add -prototype void vp9_short_iht16x16 "int16_t *input, int16_t *output, int pitch, int tx_type" -specialize vp9_short_iht16x16 +prototype void vp9_short_iht16x16_add "int16_t *input, uint8_t *output, int pitch, int tx_type" +specialize vp9_short_iht16x16_add prototype void vp9_idct4_1d "int16_t *input, int16_t *output" specialize vp9_idct4_1d sse2 - # dct and add prototype void vp9_dc_only_idct_add "int input_dc, uint8_t *pred_ptr, uint8_t *dst_ptr, int pitch, int stride" specialize vp9_dc_only_idct_add sse2 -prototype void vp9_short_iwalsh4x4_1 "int16_t *input, int16_t *output, int pitch" -specialize vp9_short_iwalsh4x4_1 -prototype void vp9_short_iwalsh4x4 "int16_t *input, int16_t *output, int pitch" -specialize vp9_short_iwalsh4x4 -prototype void vp9_dc_only_inv_walsh_add "int input_dc, uint8_t *pred_ptr, uint8_t *dst_ptr, int pitch, int stride" -specialize vp9_dc_only_inv_walsh_add +prototype void vp9_short_iwalsh4x4_1_add "int16_t *input, uint8_t *dest, int dest_stride" +specialize vp9_short_iwalsh4x4_1_add + +prototype void vp9_short_iwalsh4x4_add "int16_t *input, uint8_t *dest, int dest_stride" +specialize vp9_short_iwalsh4x4_add prototype unsigned int vp9_sad32x3 "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, int max_sad" specialize vp9_sad32x3 @@ -408,66 +228,148 @@ if [ "$CONFIG_VP9_ENCODER" = "yes" ]; then # variance [ $arch = "x86_64" ] && mmx_x86_64=mmx && sse2_x86_64=sse2 +prototype unsigned int vp9_variance32x16 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" +specialize vp9_variance32x16 sse2 + +prototype unsigned int vp9_variance16x32 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" +specialize vp9_variance16x32 sse2 + +prototype unsigned int vp9_variance64x32 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" +specialize vp9_variance64x32 sse2 + +prototype unsigned int vp9_variance32x64 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" +specialize vp9_variance32x64 sse2 + prototype unsigned int vp9_variance32x32 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" -specialize vp9_variance32x32 +specialize vp9_variance32x32 sse2 prototype unsigned int vp9_variance64x64 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" -specialize vp9_variance64x64 +specialize vp9_variance64x64 sse2 prototype unsigned int vp9_variance16x16 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_variance16x16 mmx sse2 -vp9_variance16x16_sse2=vp9_variance16x16_wmt -vp9_variance16x16_mmx=vp9_variance16x16_mmx prototype unsigned int vp9_variance16x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_variance16x8 mmx sse2 -vp9_variance16x8_sse2=vp9_variance16x8_wmt -vp9_variance16x8_mmx=vp9_variance16x8_mmx prototype unsigned int vp9_variance8x16 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_variance8x16 mmx sse2 -vp9_variance8x16_sse2=vp9_variance8x16_wmt -vp9_variance8x16_mmx=vp9_variance8x16_mmx prototype unsigned int vp9_variance8x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_variance8x8 mmx sse2 -vp9_variance8x8_sse2=vp9_variance8x8_wmt -vp9_variance8x8_mmx=vp9_variance8x8_mmx + +prototype void vp9_get_sse_sum_8x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, int *sum" +specialize vp9_get_sse_sum_8x8 sse2 +vp9_get_sse_sum_8x8_sse2=vp9_get8x8var_sse2 + +prototype unsigned int vp9_variance8x4 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" +specialize vp9_variance8x4 sse2 + +prototype unsigned int vp9_variance4x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" +specialize vp9_variance4x8 sse2 prototype unsigned int vp9_variance4x4 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_variance4x4 mmx sse2 -vp9_variance4x4_sse2=vp9_variance4x4_wmt -vp9_variance4x4_mmx=vp9_variance4x4_mmx prototype unsigned int vp9_sub_pixel_variance64x64 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_sub_pixel_variance64x64 sse2 +prototype unsigned int vp9_sub_pixel_avg_variance64x64 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance64x64 + +prototype unsigned int vp9_sub_pixel_variance32x64 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" +specialize vp9_sub_pixel_variance32x64 + +prototype unsigned int vp9_sub_pixel_avg_variance32x64 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance32x64 + +prototype unsigned int vp9_sub_pixel_variance64x32 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" +specialize vp9_sub_pixel_variance64x32 + +prototype unsigned int vp9_sub_pixel_avg_variance64x32 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance64x32 + +prototype unsigned int vp9_sub_pixel_variance32x16 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" +specialize vp9_sub_pixel_variance32x16 + +prototype unsigned int vp9_sub_pixel_avg_variance32x16 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance32x16 + +prototype unsigned int vp9_sub_pixel_variance16x32 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" +specialize vp9_sub_pixel_variance16x32 + +prototype unsigned int vp9_sub_pixel_avg_variance16x32 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance16x32 + prototype unsigned int vp9_sub_pixel_variance32x32 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_sub_pixel_variance32x32 sse2 +prototype unsigned int vp9_sub_pixel_avg_variance32x32 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance32x32 + prototype unsigned int vp9_sub_pixel_variance16x16 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_sub_pixel_variance16x16 sse2 mmx ssse3 +prototype unsigned int vp9_sub_pixel_avg_variance16x16 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance16x16 + prototype unsigned int vp9_sub_pixel_variance8x16 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_sub_pixel_variance8x16 sse2 mmx vp9_sub_pixel_variance8x16_sse2=vp9_sub_pixel_variance8x16_wmt +prototype unsigned int vp9_sub_pixel_avg_variance8x16 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance8x16 + prototype unsigned int vp9_sub_pixel_variance16x8 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_sub_pixel_variance16x8 sse2 mmx ssse3 vp9_sub_pixel_variance16x8_sse2=vp9_sub_pixel_variance16x8_ssse3; vp9_sub_pixel_variance16x8_sse2=vp9_sub_pixel_variance16x8_wmt +prototype unsigned int vp9_sub_pixel_avg_variance16x8 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance16x8 + prototype unsigned int vp9_sub_pixel_variance8x8 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_sub_pixel_variance8x8 sse2 mmx vp9_sub_pixel_variance8x8_sse2=vp9_sub_pixel_variance8x8_wmt +prototype unsigned int vp9_sub_pixel_avg_variance8x8 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance8x8 + +# TODO(jingning): need to convert 8x4/4x8 functions into mmx/sse form +prototype unsigned int vp9_sub_pixel_variance8x4 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" +specialize vp9_sub_pixel_variance8x4 + +prototype unsigned int vp9_sub_pixel_avg_variance8x4 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance8x4 + +prototype unsigned int vp9_sub_pixel_variance4x8 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" +specialize vp9_sub_pixel_variance4x8 + +prototype unsigned int vp9_sub_pixel_avg_variance4x8 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance4x8 + prototype unsigned int vp9_sub_pixel_variance4x4 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_sub_pixel_variance4x4 sse2 mmx vp9_sub_pixel_variance4x4_sse2=vp9_sub_pixel_variance4x4_wmt +prototype unsigned int vp9_sub_pixel_avg_variance4x4 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred" +specialize vp9_sub_pixel_avg_variance4x4 + prototype unsigned int vp9_sad64x64 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad" specialize vp9_sad64x64 sse2 +prototype unsigned int vp9_sad32x64 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad" +specialize vp9_sad32x64 sse2 + +prototype unsigned int vp9_sad64x32 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad" +specialize vp9_sad64x32 sse2 + +prototype unsigned int vp9_sad32x16 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad" +specialize vp9_sad32x16 sse2 + +prototype unsigned int vp9_sad16x32 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad" +specialize vp9_sad16x32 sse2 + prototype unsigned int vp9_sad32x32 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad" specialize vp9_sad32x32 sse2 @@ -483,6 +385,13 @@ specialize vp9_sad8x16 mmx sse2 prototype unsigned int vp9_sad8x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad" specialize vp9_sad8x8 mmx sse2 +# TODO(jingning): need to covert these functions into mmx/sse2 form +prototype unsigned int vp9_sad8x4 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad" +specialize vp9_sad8x4 + +prototype unsigned int vp9_sad4x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad" +specialize vp9_sad4x8 + prototype unsigned int vp9_sad4x4 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad" specialize vp9_sad4x4 mmx sse @@ -555,12 +464,30 @@ specialize vp9_sad8x16x8 sse4 prototype void vp9_sad8x8x8 "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, uint32_t *sad_array" specialize vp9_sad8x8x8 sse4 +prototype void vp9_sad8x4x8 "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, uint32_t *sad_array" +specialize vp9_sad8x4x8 + +prototype void vp9_sad4x8x8 "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, uint32_t *sad_array" +specialize vp9_sad4x8x8 + prototype void vp9_sad4x4x8 "const uint8_t *src_ptr, int src_stride, const uint8_t *ref_ptr, int ref_stride, uint32_t *sad_array" specialize vp9_sad4x4x8 sse4 prototype void vp9_sad64x64x4d "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array" specialize vp9_sad64x64x4d sse2 +prototype void vp9_sad32x64x4d "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array" +specialize vp9_sad32x64x4d sse2 + +prototype void vp9_sad64x32x4d "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array" +specialize vp9_sad64x32x4d sse2 + +prototype void vp9_sad32x16x4d "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array" +specialize vp9_sad32x16x4d sse2 + +prototype void vp9_sad16x32x4d "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array" +specialize vp9_sad16x32x4d sse2 + prototype void vp9_sad32x32x4d "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array" specialize vp9_sad32x32x4d sse2 @@ -576,6 +503,13 @@ specialize vp9_sad8x16x4d sse2 prototype void vp9_sad8x8x4d "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array" specialize vp9_sad8x8x4d sse2 +# TODO(jingning): need to convert these 4x8/8x4 functions into sse2 form +prototype void vp9_sad8x4x4d "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array" +specialize vp9_sad8x4x4d + +prototype void vp9_sad4x8x4d "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array" +specialize vp9_sad4x8x4d + prototype void vp9_sad4x4x4d "const uint8_t *src_ptr, int src_stride, const uint8_t* const ref_ptr[], int ref_stride, unsigned int *sad_array" specialize vp9_sad4x4x4d sse prototype unsigned int vp9_sub_pixel_mse16x16 "const uint8_t *src_ptr, int src_pixels_per_line, int xoffset, int yoffset, const uint8_t *dst_ptr, int dst_pixels_per_line, unsigned int *sse" @@ -585,6 +519,15 @@ prototype unsigned int vp9_mse16x16 "const uint8_t *src_ptr, int source_stride, specialize vp9_mse16x16 mmx sse2 vp9_mse16x16_sse2=vp9_mse16x16_wmt +prototype unsigned int vp9_mse8x16 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int recon_stride, unsigned int *sse" +specialize vp9_mse8x16 + +prototype unsigned int vp9_mse16x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int recon_stride, unsigned int *sse" +specialize vp9_mse16x8 + +prototype unsigned int vp9_mse8x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int recon_stride, unsigned int *sse" +specialize vp9_mse8x8 + prototype unsigned int vp9_sub_pixel_mse64x64 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse" specialize vp9_sub_pixel_mse64x64 @@ -594,30 +537,11 @@ specialize vp9_sub_pixel_mse32x32 prototype unsigned int vp9_get_mb_ss "const int16_t *" specialize vp9_get_mb_ss mmx sse2 # ENCODEMB INVOKE -prototype int vp9_mbblock_error "struct macroblock *mb" -specialize vp9_mbblock_error mmx sse2 -vp9_mbblock_error_sse2=vp9_mbblock_error_xmm prototype int vp9_block_error "int16_t *coeff, int16_t *dqcoeff, int block_size" specialize vp9_block_error mmx sse2 vp9_block_error_sse2=vp9_block_error_xmm -prototype void vp9_subtract_b "struct block *be, struct blockd *bd, int pitch" -specialize vp9_subtract_b mmx sse2 - -prototype int vp9_mbuverror "struct macroblock *mb" -specialize vp9_mbuverror mmx sse2 -vp9_mbuverror_sse2=vp9_mbuverror_xmm - -prototype void vp9_subtract_b "struct block *be, struct blockd *bd, int pitch" -specialize vp9_subtract_b mmx sse2 - -prototype void vp9_subtract_mby "int16_t *diff, uint8_t *src, uint8_t *pred, int stride" -specialize vp9_subtract_mby mmx sse2 - -prototype void vp9_subtract_mbuv "int16_t *diff, uint8_t *usrc, uint8_t *vsrc, uint8_t *pred, int stride" -specialize vp9_subtract_mbuv mmx sse2 - # # Structured Similarity (SSIM) # @@ -665,16 +589,16 @@ specialize vp9_short_walsh8x4 # # Motion search # -prototype int vp9_full_search_sad "struct macroblock *x, struct block *b, struct blockd *d, union int_mv *ref_mv, int sad_per_bit, int distance, struct vp9_variance_vtable *fn_ptr, DEC_MVCOSTS, union int_mv *center_mv" +prototype int vp9_full_search_sad "struct macroblock *x, union int_mv *ref_mv, int sad_per_bit, int distance, struct vp9_variance_vtable *fn_ptr, DEC_MVCOSTS, union int_mv *center_mv, int n" specialize vp9_full_search_sad sse3 sse4_1 vp9_full_search_sad_sse3=vp9_full_search_sadx3 vp9_full_search_sad_sse4_1=vp9_full_search_sadx8 -prototype int vp9_refining_search_sad "struct macroblock *x, struct block *b, struct blockd *d, union int_mv *ref_mv, int sad_per_bit, int distance, struct vp9_variance_vtable *fn_ptr, DEC_MVCOSTS, union int_mv *center_mv" +prototype int vp9_refining_search_sad "struct macroblock *x, union int_mv *ref_mv, int sad_per_bit, int distance, struct vp9_variance_vtable *fn_ptr, DEC_MVCOSTS, union int_mv *center_mv" specialize vp9_refining_search_sad sse3 vp9_refining_search_sad_sse3=vp9_refining_search_sadx4 -prototype int vp9_diamond_search_sad "struct macroblock *x, struct block *b, struct blockd *d, union int_mv *ref_mv, union int_mv *best_mv, int search_param, int sad_per_bit, int *num00, struct vp9_variance_vtable *fn_ptr, DEC_MVCOSTS, union int_mv *center_mv" +prototype int vp9_diamond_search_sad "struct macroblock *x, union int_mv *ref_mv, union int_mv *best_mv, int search_param, int sad_per_bit, int *num00, struct vp9_variance_vtable *fn_ptr, DEC_MVCOSTS, union int_mv *center_mv" specialize vp9_diamond_search_sad sse3 vp9_diamond_search_sad_sse3=vp9_diamond_search_sadx4 |