1 common_forward_decls() { 2 cat <<EOF 3 #include "vp8/common/blockd.h" 4 5 struct blockd; 6 struct macroblockd; 7 struct loop_filter_info; 8 9 /* Encoder forward decls */ 10 struct block; 11 struct macroblock; 12 struct variance_vtable; 13 union int_mv; 14 struct yv12_buffer_config; 15 EOF 16 } 17 forward_decls common_forward_decls 18 19 # 20 # Dequant 21 # 22 prototype void vp8_dequantize_b "struct blockd*, short *dqc" 23 specialize vp8_dequantize_b mmx media neon 24 vp8_dequantize_b_media=vp8_dequantize_b_v6 25 26 prototype void vp8_dequant_idct_add "short *input, short *dq, unsigned char *output, int stride" 27 specialize vp8_dequant_idct_add mmx media neon dspr2 28 vp8_dequant_idct_add_media=vp8_dequant_idct_add_v6 29 vp8_dequant_idct_add_dspr2=vp8_dequant_idct_add_dspr2 30 31 prototype void vp8_dequant_idct_add_y_block "short *q, short *dq, unsigned char *dst, int stride, char *eobs" 32 specialize vp8_dequant_idct_add_y_block mmx sse2 media neon dspr2 33 vp8_dequant_idct_add_y_block_media=vp8_dequant_idct_add_y_block_v6 34 vp8_dequant_idct_add_y_block_dspr2=vp8_dequant_idct_add_y_block_dspr2 35 36 prototype void vp8_dequant_idct_add_uv_block "short *q, short *dq, unsigned char *dst_u, unsigned char *dst_v, int stride, char *eobs" 37 specialize vp8_dequant_idct_add_uv_block mmx sse2 media neon dspr2 38 vp8_dequant_idct_add_uv_block_media=vp8_dequant_idct_add_uv_block_v6 39 vp8_dequant_idct_add_y_block_dspr2=vp8_dequant_idct_add_y_block_dspr2 40 41 # 42 # Loopfilter 43 # 44 prototype void vp8_loop_filter_mbv "unsigned char *y, unsigned char *u, unsigned char *v, int ystride, int uv_stride, struct loop_filter_info *lfi" 45 specialize vp8_loop_filter_mbv mmx sse2 media neon dspr2 46 vp8_loop_filter_mbv_media=vp8_loop_filter_mbv_armv6 47 vp8_loop_filter_mbv_dspr2=vp8_loop_filter_mbv_dspr2 48 49 prototype void vp8_loop_filter_bv "unsigned char *y, unsigned char *u, unsigned char *v, int ystride, int uv_stride, struct loop_filter_info *lfi" 50 specialize vp8_loop_filter_bv mmx sse2 media neon dspr2 51 vp8_loop_filter_bv_media=vp8_loop_filter_bv_armv6 52 vp8_loop_filter_bv_dspr2=vp8_loop_filter_bv_dspr2 53 54 prototype void vp8_loop_filter_mbh "unsigned char *y, unsigned char *u, unsigned char *v, int ystride, int uv_stride, struct loop_filter_info *lfi" 55 specialize vp8_loop_filter_mbh mmx sse2 media neon dspr2 56 vp8_loop_filter_mbh_media=vp8_loop_filter_mbh_armv6 57 vp8_loop_filter_mbh_dspr2=vp8_loop_filter_mbh_dspr2 58 59 prototype void vp8_loop_filter_bh "unsigned char *y, unsigned char *u, unsigned char *v, int ystride, int uv_stride, struct loop_filter_info *lfi" 60 specialize vp8_loop_filter_bh mmx sse2 media neon dspr2 61 vp8_loop_filter_bh_media=vp8_loop_filter_bh_armv6 62 vp8_loop_filter_bh_dspr2=vp8_loop_filter_bh_dspr2 63 64 65 prototype void vp8_loop_filter_simple_mbv "unsigned char *y, int ystride, const unsigned char *blimit" 66 specialize vp8_loop_filter_simple_mbv mmx sse2 media neon 67 vp8_loop_filter_simple_mbv_c=vp8_loop_filter_simple_vertical_edge_c 68 vp8_loop_filter_simple_mbv_mmx=vp8_loop_filter_simple_vertical_edge_mmx 69 vp8_loop_filter_simple_mbv_sse2=vp8_loop_filter_simple_vertical_edge_sse2 70 vp8_loop_filter_simple_mbv_media=vp8_loop_filter_simple_vertical_edge_armv6 71 vp8_loop_filter_simple_mbv_neon=vp8_loop_filter_mbvs_neon 72 73 prototype void vp8_loop_filter_simple_mbh "unsigned char *y, int ystride, const unsigned char *blimit" 74 specialize vp8_loop_filter_simple_mbh mmx sse2 media neon 75 vp8_loop_filter_simple_mbh_c=vp8_loop_filter_simple_horizontal_edge_c 76 vp8_loop_filter_simple_mbh_mmx=vp8_loop_filter_simple_horizontal_edge_mmx 77 vp8_loop_filter_simple_mbh_sse2=vp8_loop_filter_simple_horizontal_edge_sse2 78 vp8_loop_filter_simple_mbh_media=vp8_loop_filter_simple_horizontal_edge_armv6 79 vp8_loop_filter_simple_mbh_neon=vp8_loop_filter_mbhs_neon 80 81 prototype void vp8_loop_filter_simple_bv "unsigned char *y, int ystride, const unsigned char *blimit" 82 specialize vp8_loop_filter_simple_bv mmx sse2 media neon 83 vp8_loop_filter_simple_bv_c=vp8_loop_filter_bvs_c 84 vp8_loop_filter_simple_bv_mmx=vp8_loop_filter_bvs_mmx 85 vp8_loop_filter_simple_bv_sse2=vp8_loop_filter_bvs_sse2 86 vp8_loop_filter_simple_bv_media=vp8_loop_filter_bvs_armv6 87 vp8_loop_filter_simple_bv_neon=vp8_loop_filter_bvs_neon 88 89 prototype void vp8_loop_filter_simple_bh "unsigned char *y, int ystride, const unsigned char *blimit" 90 specialize vp8_loop_filter_simple_bh mmx sse2 media neon 91 vp8_loop_filter_simple_bh_c=vp8_loop_filter_bhs_c 92 vp8_loop_filter_simple_bh_mmx=vp8_loop_filter_bhs_mmx 93 vp8_loop_filter_simple_bh_sse2=vp8_loop_filter_bhs_sse2 94 vp8_loop_filter_simple_bh_media=vp8_loop_filter_bhs_armv6 95 vp8_loop_filter_simple_bh_neon=vp8_loop_filter_bhs_neon 96 97 # 98 # IDCT 99 # 100 #idct16 101 prototype void vp8_short_idct4x4llm "short *input, unsigned char *pred, int pitch, unsigned char *dst, int dst_stride" 102 specialize vp8_short_idct4x4llm mmx media neon dspr2 103 vp8_short_idct4x4llm_media=vp8_short_idct4x4llm_v6_dual 104 vp8_short_idct4x4llm_dspr2=vp8_short_idct4x4llm_dspr2 105 106 #iwalsh1 107 prototype void vp8_short_inv_walsh4x4_1 "short *input, short *output" 108 specialize vp8_short_inv_walsh4x4_1 dspr2 109 vp8_short_inv_walsh4x4_1_dspr2=vp8_short_inv_walsh4x4_1_dspr2 110 # no asm yet 111 112 #iwalsh16 113 prototype void vp8_short_inv_walsh4x4 "short *input, short *output" 114 specialize vp8_short_inv_walsh4x4 mmx sse2 media neon dspr2 115 vp8_short_inv_walsh4x4_media=vp8_short_inv_walsh4x4_v6 116 vp8_short_inv_walsh4x4_dspr2=vp8_short_inv_walsh4x4_dspr2 117 118 #idct1_scalar_add 119 prototype void vp8_dc_only_idct_add "short input, unsigned char *pred, int pred_stride, unsigned char *dst, int dst_stride" 120 specialize vp8_dc_only_idct_add mmx media neon dspr2 121 vp8_dc_only_idct_add_media=vp8_dc_only_idct_add_v6 122 vp8_dc_only_idct_add_dspr2=vp8_dc_only_idct_add_dspr2 123 124 # 125 # RECON 126 # 127 prototype void vp8_copy_mem16x16 "unsigned char *src, int src_pitch, unsigned char *dst, int dst_pitch" 128 specialize vp8_copy_mem16x16 mmx sse2 media neon dspr2 129 vp8_copy_mem16x16_media=vp8_copy_mem16x16_v6 130 vp8_copy_mem16x16_dspr2=vp8_copy_mem16x16_dspr2 131 132 prototype void vp8_copy_mem8x8 "unsigned char *src, int src_pitch, unsigned char *dst, int dst_pitch" 133 specialize vp8_copy_mem8x8 mmx media neon dspr2 134 vp8_copy_mem8x8_media=vp8_copy_mem8x8_v6 135 vp8_copy_mem8x8_dspr2=vp8_copy_mem8x8_dspr2 136 137 prototype void vp8_copy_mem8x4 "unsigned char *src, int src_pitch, unsigned char *dst, int dst_pitch" 138 specialize vp8_copy_mem8x4 mmx media neon dspr2 139 vp8_copy_mem8x4_media=vp8_copy_mem8x4_v6 140 vp8_copy_mem8x4_dspr2=vp8_copy_mem8x4_dspr2 141 142 prototype void vp8_build_intra_predictors_mby_s "struct macroblockd *x, unsigned char * yabove_row, unsigned char * yleft, int left_stride, unsigned char * ypred_ptr, int y_stride" 143 specialize vp8_build_intra_predictors_mby_s sse2 ssse3 144 #TODO: fix assembly for neon 145 146 prototype void vp8_build_intra_predictors_mbuv_s "struct macroblockd *x, unsigned char * uabove_row, unsigned char * vabove_row, unsigned char *uleft, unsigned char *vleft, int left_stride, unsigned char * upred_ptr, unsigned char * vpred_ptr, int pred_stride" 147 specialize vp8_build_intra_predictors_mbuv_s sse2 ssse3 148 149 prototype void vp8_intra4x4_predict "unsigned char *Above, unsigned char *yleft, int left_stride, B_PREDICTION_MODE b_mode, unsigned char *dst, int dst_stride, unsigned char top_left" 150 specialize vp8_intra4x4_predict media 151 vp8_intra4x4_predict_media=vp8_intra4x4_predict_armv6 152 153 # 154 # Postproc 155 # 156 if [ "$CONFIG_POSTPROC" = "yes" ]; then 157 prototype void vp8_mbpost_proc_down "unsigned char *dst, int pitch, int rows, int cols,int flimit" 158 specialize vp8_mbpost_proc_down mmx sse2 159 vp8_mbpost_proc_down_sse2=vp8_mbpost_proc_down_xmm 160 161 prototype void vp8_mbpost_proc_across_ip "unsigned char *dst, int pitch, int rows, int cols,int flimit" 162 specialize vp8_mbpost_proc_across_ip sse2 163 vp8_mbpost_proc_across_ip_sse2=vp8_mbpost_proc_across_ip_xmm 164 165 prototype void vp8_post_proc_down_and_across_mb_row "unsigned char *src, unsigned char *dst, int src_pitch, int dst_pitch, int cols, unsigned char *flimits, int size" 166 specialize vp8_post_proc_down_and_across_mb_row sse2 167 168 prototype void vp8_plane_add_noise "unsigned char *s, char *noise, char blackclamp[16], char whiteclamp[16], char bothclamp[16], unsigned int w, unsigned int h, int pitch" 169 specialize vp8_plane_add_noise mmx sse2 170 vp8_plane_add_noise_sse2=vp8_plane_add_noise_wmt 171 172 prototype void vp8_blend_mb_inner "unsigned char *y, unsigned char *u, unsigned char *v, int y1, int u1, int v1, int alpha, int stride" 173 # no asm yet 174 175 prototype void vp8_blend_mb_outer "unsigned char *y, unsigned char *u, unsigned char *v, int y1, int u1, int v1, int alpha, int stride" 176 # no asm yet 177 178 prototype void vp8_blend_b "unsigned char *y, unsigned char *u, unsigned char *v, int y1, int u1, int v1, int alpha, int stride" 179 # no asm yet 180 181 prototype void vp8_filter_by_weight16x16 "unsigned char *src, int src_stride, unsigned char *dst, int dst_stride, int src_weight" 182 specialize vp8_filter_by_weight16x16 sse2 183 184 prototype void vp8_filter_by_weight8x8 "unsigned char *src, int src_stride, unsigned char *dst, int dst_stride, int src_weight" 185 specialize vp8_filter_by_weight8x8 sse2 186 187 prototype void vp8_filter_by_weight4x4 "unsigned char *src, int src_stride, unsigned char *dst, int dst_stride, int src_weight" 188 # no asm yet 189 fi 190 191 # 192 # Subpixel 193 # 194 prototype void vp8_sixtap_predict16x16 "unsigned char *src, int src_pitch, int xofst, int yofst, unsigned char *dst, int dst_pitch" 195 specialize vp8_sixtap_predict16x16 mmx sse2 ssse3 media neon dspr2 196 vp8_sixtap_predict16x16_media=vp8_sixtap_predict16x16_armv6 197 vp8_sixtap_predict16x16_dspr2=vp8_sixtap_predict16x16_dspr2 198 199 prototype void vp8_sixtap_predict8x8 "unsigned char *src, int src_pitch, int xofst, int yofst, unsigned char *dst, int dst_pitch" 200 specialize vp8_sixtap_predict8x8 mmx sse2 ssse3 media neon dspr2 201 vp8_sixtap_predict8x8_media=vp8_sixtap_predict8x8_armv6 202 vp8_sixtap_predict8x8_dspr2=vp8_sixtap_predict8x8_dspr2 203 204 prototype void vp8_sixtap_predict8x4 "unsigned char *src, int src_pitch, int xofst, int yofst, unsigned char *dst, int dst_pitch" 205 specialize vp8_sixtap_predict8x4 mmx sse2 ssse3 media neon dspr2 206 vp8_sixtap_predict8x4_media=vp8_sixtap_predict8x4_armv6 207 vp8_sixtap_predict8x4_dspr2=vp8_sixtap_predict8x4_dspr2 208 209 prototype void vp8_sixtap_predict4x4 "unsigned char *src, int src_pitch, int xofst, int yofst, unsigned char *dst, int dst_pitch" 210 specialize vp8_sixtap_predict4x4 mmx ssse3 media neon dspr2 211 vp8_sixtap_predict4x4_media=vp8_sixtap_predict4x4_armv6 212 vp8_sixtap_predict4x4_dspr2=vp8_sixtap_predict4x4_dspr2 213 214 prototype void vp8_bilinear_predict16x16 "unsigned char *src, int src_pitch, int xofst, int yofst, unsigned char *dst, int dst_pitch" 215 specialize vp8_bilinear_predict16x16 mmx sse2 ssse3 media neon 216 vp8_bilinear_predict16x16_media=vp8_bilinear_predict16x16_armv6 217 218 prototype void vp8_bilinear_predict8x8 "unsigned char *src, int src_pitch, int xofst, int yofst, unsigned char *dst, int dst_pitch" 219 specialize vp8_bilinear_predict8x8 mmx sse2 ssse3 media neon 220 vp8_bilinear_predict8x8_media=vp8_bilinear_predict8x8_armv6 221 222 prototype void vp8_bilinear_predict8x4 "unsigned char *src, int src_pitch, int xofst, int yofst, unsigned char *dst, int dst_pitch" 223 specialize vp8_bilinear_predict8x4 mmx media neon 224 vp8_bilinear_predict8x4_media=vp8_bilinear_predict8x4_armv6 225 226 prototype void vp8_bilinear_predict4x4 "unsigned char *src, int src_pitch, int xofst, int yofst, unsigned char *dst, int dst_pitch" 227 specialize vp8_bilinear_predict4x4 mmx media neon 228 vp8_bilinear_predict4x4_media=vp8_bilinear_predict4x4_armv6 229 230 # 231 # Whole-pixel Variance 232 # 233 prototype unsigned int vp8_variance4x4 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sse" 234 specialize vp8_variance4x4 mmx sse2 235 vp8_variance4x4_sse2=vp8_variance4x4_wmt 236 237 prototype unsigned int vp8_variance8x8 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sse" 238 specialize vp8_variance8x8 mmx sse2 media neon 239 vp8_variance8x8_sse2=vp8_variance8x8_wmt 240 vp8_variance8x8_media=vp8_variance8x8_armv6 241 242 prototype unsigned int vp8_variance8x16 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sse" 243 specialize vp8_variance8x16 mmx sse2 neon 244 vp8_variance8x16_sse2=vp8_variance8x16_wmt 245 246 prototype unsigned int vp8_variance16x8 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sse" 247 specialize vp8_variance16x8 mmx sse2 neon 248 vp8_variance16x8_sse2=vp8_variance16x8_wmt 249 250 prototype unsigned int vp8_variance16x16 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sse" 251 specialize vp8_variance16x16 mmx sse2 media neon 252 vp8_variance16x16_sse2=vp8_variance16x16_wmt 253 vp8_variance16x16_media=vp8_variance16x16_armv6 254 255 # 256 # Sub-pixel Variance 257 # 258 prototype unsigned int vp8_sub_pixel_variance4x4 "const unsigned char *src_ptr, int source_stride, int xoffset, int yoffset, const unsigned char *ref_ptr, int Refstride, unsigned int *sse" 259 specialize vp8_sub_pixel_variance4x4 mmx sse2 260 vp8_sub_pixel_variance4x4_sse2=vp8_sub_pixel_variance4x4_wmt 261 262 prototype unsigned int vp8_sub_pixel_variance8x8 "const unsigned char *src_ptr, int source_stride, int xoffset, int yoffset, const unsigned char *ref_ptr, int Refstride, unsigned int *sse" 263 specialize vp8_sub_pixel_variance8x8 mmx sse2 media neon 264 vp8_sub_pixel_variance8x8_sse2=vp8_sub_pixel_variance8x8_wmt 265 vp8_sub_pixel_variance8x8_media=vp8_sub_pixel_variance8x8_armv6 266 267 prototype unsigned int vp8_sub_pixel_variance8x16 "const unsigned char *src_ptr, int source_stride, int xoffset, int yoffset, const unsigned char *ref_ptr, int Refstride, unsigned int *sse" 268 specialize vp8_sub_pixel_variance8x16 mmx sse2 269 vp8_sub_pixel_variance8x16_sse2=vp8_sub_pixel_variance8x16_wmt 270 271 prototype unsigned int vp8_sub_pixel_variance16x8 "const unsigned char *src_ptr, int source_stride, int xoffset, int yoffset, const unsigned char *ref_ptr, int Refstride, unsigned int *sse" 272 specialize vp8_sub_pixel_variance16x8 mmx sse2 ssse3 273 vp8_sub_pixel_variance16x8_sse2=vp8_sub_pixel_variance16x8_wmt 274 275 prototype unsigned int vp8_sub_pixel_variance16x16 "const unsigned char *src_ptr, int source_stride, int xoffset, int yoffset, const unsigned char *ref_ptr, int Refstride, unsigned int *sse" 276 specialize vp8_sub_pixel_variance16x16 mmx sse2 ssse3 media neon 277 vp8_sub_pixel_variance16x16_sse2=vp8_sub_pixel_variance16x16_wmt 278 vp8_sub_pixel_variance16x16_media=vp8_sub_pixel_variance16x16_armv6 279 280 prototype unsigned int vp8_variance_halfpixvar16x16_h "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sse" 281 specialize vp8_variance_halfpixvar16x16_h mmx sse2 media neon 282 vp8_variance_halfpixvar16x16_h_sse2=vp8_variance_halfpixvar16x16_h_wmt 283 vp8_variance_halfpixvar16x16_h_media=vp8_variance_halfpixvar16x16_h_armv6 284 285 prototype unsigned int vp8_variance_halfpixvar16x16_v "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sse" 286 specialize vp8_variance_halfpixvar16x16_v mmx sse2 media neon 287 vp8_variance_halfpixvar16x16_v_sse2=vp8_variance_halfpixvar16x16_v_wmt 288 vp8_variance_halfpixvar16x16_v_media=vp8_variance_halfpixvar16x16_v_armv6 289 290 prototype unsigned int vp8_variance_halfpixvar16x16_hv "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sse" 291 specialize vp8_variance_halfpixvar16x16_hv mmx sse2 media neon 292 vp8_variance_halfpixvar16x16_hv_sse2=vp8_variance_halfpixvar16x16_hv_wmt 293 vp8_variance_halfpixvar16x16_hv_media=vp8_variance_halfpixvar16x16_hv_armv6 294 295 # 296 # Single block SAD 297 # 298 prototype unsigned int vp8_sad4x4 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int max_sad" 299 specialize vp8_sad4x4 mmx sse2 neon 300 vp8_sad4x4_sse2=vp8_sad4x4_wmt 301 302 prototype unsigned int vp8_sad8x8 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int max_sad" 303 specialize vp8_sad8x8 mmx sse2 neon 304 vp8_sad8x8_sse2=vp8_sad8x8_wmt 305 306 prototype unsigned int vp8_sad8x16 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int max_sad" 307 specialize vp8_sad8x16 mmx sse2 neon 308 vp8_sad8x16_sse2=vp8_sad8x16_wmt 309 310 prototype unsigned int vp8_sad16x8 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int max_sad" 311 specialize vp8_sad16x8 mmx sse2 neon 312 vp8_sad16x8_sse2=vp8_sad16x8_wmt 313 314 prototype unsigned int vp8_sad16x16 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int max_sad" 315 specialize vp8_sad16x16 mmx sse2 sse3 media neon 316 vp8_sad16x16_sse2=vp8_sad16x16_wmt 317 vp8_sad16x16_media=vp8_sad16x16_armv6 318 319 # 320 # Multi-block SAD, comparing a reference to N blocks 1 pixel apart horizontally 321 # 322 prototype void vp8_sad4x4x3 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sad_array" 323 specialize vp8_sad4x4x3 sse3 324 325 prototype void vp8_sad8x8x3 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sad_array" 326 specialize vp8_sad8x8x3 sse3 327 328 prototype void vp8_sad8x16x3 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sad_array" 329 specialize vp8_sad8x16x3 sse3 330 331 prototype void vp8_sad16x8x3 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sad_array" 332 specialize vp8_sad16x8x3 sse3 ssse3 333 334 prototype void vp8_sad16x16x3 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sad_array" 335 specialize vp8_sad16x16x3 sse3 ssse3 336 337 # Note the only difference in the following prototypes is that they return into 338 # an array of short 339 prototype void vp8_sad4x4x8 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned short *sad_array" 340 specialize vp8_sad4x4x8 sse4_1 341 vp8_sad4x4x8_sse4_1=vp8_sad4x4x8_sse4 342 343 prototype void vp8_sad8x8x8 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned short *sad_array" 344 specialize vp8_sad8x8x8 sse4_1 345 vp8_sad8x8x8_sse4_1=vp8_sad8x8x8_sse4 346 347 prototype void vp8_sad8x16x8 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned short *sad_array" 348 specialize vp8_sad8x16x8 sse4_1 349 vp8_sad8x16x8_sse4_1=vp8_sad8x16x8_sse4 350 351 prototype void vp8_sad16x8x8 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned short *sad_array" 352 specialize vp8_sad16x8x8 sse4_1 353 vp8_sad16x8x8_sse4_1=vp8_sad16x8x8_sse4 354 355 prototype void vp8_sad16x16x8 "const unsigned char *src_ptr, int src_stride, const unsigned char *ref_ptr, int ref_stride, unsigned short *sad_array" 356 specialize vp8_sad16x16x8 sse4_1 357 vp8_sad16x16x8_sse4_1=vp8_sad16x16x8_sse4 358 359 # 360 # Multi-block SAD, comparing a reference to N independent blocks 361 # 362 prototype void vp8_sad4x4x4d "const unsigned char *src_ptr, int src_stride, const unsigned char * const ref_ptr[], int ref_stride, unsigned int *sad_array" 363 specialize vp8_sad4x4x4d sse3 364 365 prototype void vp8_sad8x8x4d "const unsigned char *src_ptr, int src_stride, const unsigned char * const ref_ptr[], int ref_stride, unsigned int *sad_array" 366 specialize vp8_sad8x8x4d sse3 367 368 prototype void vp8_sad8x16x4d "const unsigned char *src_ptr, int src_stride, const unsigned char * const ref_ptr[], int ref_stride, unsigned int *sad_array" 369 specialize vp8_sad8x16x4d sse3 370 371 prototype void vp8_sad16x8x4d "const unsigned char *src_ptr, int src_stride, const unsigned char * const ref_ptr[], int ref_stride, unsigned int *sad_array" 372 specialize vp8_sad16x8x4d sse3 373 374 prototype void vp8_sad16x16x4d "const unsigned char *src_ptr, int src_stride, const unsigned char * const ref_ptr[], int ref_stride, unsigned int *sad_array" 375 specialize vp8_sad16x16x4d sse3 376 377 # 378 # Encoder functions below this point. 379 # 380 if [ "$CONFIG_VP8_ENCODER" = "yes" ]; then 381 382 # 383 # Sum of squares (vector) 384 # 385 prototype unsigned int vp8_get_mb_ss "const short *" 386 specialize vp8_get_mb_ss mmx sse2 387 388 # 389 # SSE (Sum Squared Error) 390 # 391 prototype unsigned int vp8_sub_pixel_mse16x16 "const unsigned char *src_ptr, int source_stride, int xoffset, int yoffset, const unsigned char *ref_ptr, int Refstride, unsigned int *sse" 392 specialize vp8_sub_pixel_mse16x16 mmx sse2 393 vp8_sub_pixel_mse16x16_sse2=vp8_sub_pixel_mse16x16_wmt 394 395 prototype unsigned int vp8_mse16x16 "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, unsigned int *sse" 396 specialize vp8_mse16x16 mmx sse2 media neon 397 vp8_mse16x16_sse2=vp8_mse16x16_wmt 398 vp8_mse16x16_media=vp8_mse16x16_armv6 399 400 prototype unsigned int vp8_get4x4sse_cs "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride" 401 specialize vp8_get4x4sse_cs mmx neon 402 403 # 404 # Block copy 405 # 406 case $arch in 407 x86*) 408 prototype void vp8_copy32xn "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int ref_stride, int n" 409 specialize vp8_copy32xn sse2 sse3 410 ;; 411 esac 412 413 # 414 # Structured Similarity (SSIM) 415 # 416 if [ "$CONFIG_INTERNAL_STATS" = "yes" ]; then 417 [ $arch = "x86_64" ] && sse2_on_x86_64=sse2 418 419 prototype void vp8_ssim_parms_8x8 "unsigned char *s, int sp, unsigned char *r, int rp, unsigned long *sum_s, unsigned long *sum_r, unsigned long *sum_sq_s, unsigned long *sum_sq_r, unsigned long *sum_sxr" 420 specialize vp8_ssim_parms_8x8 $sse2_on_x86_64 421 422 prototype void vp8_ssim_parms_16x16 "unsigned char *s, int sp, unsigned char *r, int rp, unsigned long *sum_s, unsigned long *sum_r, unsigned long *sum_sq_s, unsigned long *sum_sq_r, unsigned long *sum_sxr" 423 specialize vp8_ssim_parms_16x16 $sse2_on_x86_64 424 fi 425 426 # 427 # Forward DCT 428 # 429 prototype void vp8_short_fdct4x4 "short *input, short *output, int pitch" 430 specialize vp8_short_fdct4x4 mmx sse2 media neon 431 vp8_short_fdct4x4_media=vp8_short_fdct4x4_armv6 432 433 prototype void vp8_short_fdct8x4 "short *input, short *output, int pitch" 434 specialize vp8_short_fdct8x4 mmx sse2 media neon 435 vp8_short_fdct8x4_media=vp8_short_fdct8x4_armv6 436 437 prototype void vp8_short_walsh4x4 "short *input, short *output, int pitch" 438 specialize vp8_short_walsh4x4 sse2 media neon 439 vp8_short_walsh4x4_media=vp8_short_walsh4x4_armv6 440 441 # 442 # Quantizer 443 # 444 prototype void vp8_regular_quantize_b "struct block *, struct blockd *" 445 specialize vp8_regular_quantize_b sse2 sse4_1 446 vp8_regular_quantize_b_sse4_1=vp8_regular_quantize_b_sse4 447 448 prototype void vp8_fast_quantize_b "struct block *, struct blockd *" 449 specialize vp8_fast_quantize_b sse2 ssse3 media neon 450 vp8_fast_quantize_b_media=vp8_fast_quantize_b_armv6 451 452 prototype void vp8_regular_quantize_b_pair "struct block *b1, struct block *b2, struct blockd *d1, struct blockd *d2" 453 # no asm yet 454 455 prototype void vp8_fast_quantize_b_pair "struct block *b1, struct block *b2, struct blockd *d1, struct blockd *d2" 456 specialize vp8_fast_quantize_b_pair neon 457 458 prototype void vp8_quantize_mb "struct macroblock *" 459 specialize vp8_quantize_mb neon 460 461 prototype void vp8_quantize_mby "struct macroblock *" 462 specialize vp8_quantize_mby neon 463 464 prototype void vp8_quantize_mbuv "struct macroblock *" 465 specialize vp8_quantize_mbuv neon 466 467 # 468 # Block subtraction 469 # 470 prototype int vp8_block_error "short *coeff, short *dqcoeff" 471 specialize vp8_block_error mmx sse2 472 vp8_block_error_sse2=vp8_block_error_xmm 473 474 prototype int vp8_mbblock_error "struct macroblock *mb, int dc" 475 specialize vp8_mbblock_error mmx sse2 476 vp8_mbblock_error_sse2=vp8_mbblock_error_xmm 477 478 prototype int vp8_mbuverror "struct macroblock *mb" 479 specialize vp8_mbuverror mmx sse2 480 vp8_mbuverror_sse2=vp8_mbuverror_xmm 481 482 prototype void vp8_subtract_b "struct block *be, struct blockd *bd, int pitch" 483 specialize vp8_subtract_b mmx sse2 media neon 484 vp8_subtract_b_media=vp8_subtract_b_armv6 485 486 prototype void vp8_subtract_mby "short *diff, unsigned char *src, int src_stride, unsigned char *pred, int pred_stride" 487 specialize vp8_subtract_mby mmx sse2 media neon 488 vp8_subtract_mby_media=vp8_subtract_mby_armv6 489 490 prototype void vp8_subtract_mbuv "short *diff, unsigned char *usrc, unsigned char *vsrc, int src_stride, unsigned char *upred, unsigned char *vpred, int pred_stride" 491 specialize vp8_subtract_mbuv mmx sse2 media neon 492 vp8_subtract_mbuv_media=vp8_subtract_mbuv_armv6 493 494 # 495 # Motion search 496 # 497 prototype int vp8_full_search_sad "struct macroblock *x, struct block *b, struct blockd *d, union int_mv *ref_mv, int sad_per_bit, int distance, struct variance_vtable *fn_ptr, int *mvcost[2], union int_mv *center_mv" 498 specialize vp8_full_search_sad sse3 sse4_1 499 vp8_full_search_sad_sse3=vp8_full_search_sadx3 500 vp8_full_search_sad_sse4_1=vp8_full_search_sadx8 501 502 prototype int vp8_refining_search_sad "struct macroblock *x, struct block *b, struct blockd *d, union int_mv *ref_mv, int sad_per_bit, int distance, struct variance_vtable *fn_ptr, int *mvcost[2], union int_mv *center_mv" 503 specialize vp8_refining_search_sad sse3 504 vp8_refining_search_sad_sse3=vp8_refining_search_sadx4 505 506 prototype int vp8_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 variance_vtable *fn_ptr, int *mvcost[2], union int_mv *center_mv" 507 vp8_diamond_search_sad_sse3=vp8_diamond_search_sadx4 508 509 # 510 # Alt-ref Noise Reduction (ARNR) 511 # 512 if [ "$CONFIG_REALTIME_ONLY" != "yes" ]; then 513 prototype void vp8_temporal_filter_apply "unsigned char *frame1, unsigned int stride, unsigned char *frame2, unsigned int block_size, int strength, int filter_weight, unsigned int *accumulator, unsigned short *count" 514 specialize vp8_temporal_filter_apply sse2 515 fi 516 517 # 518 # Pick Loopfilter 519 # 520 prototype void vp8_yv12_copy_partial_frame "struct yv12_buffer_config *src_ybc, struct yv12_buffer_config *dst_ybc" 521 specialize vp8_yv12_copy_partial_frame neon 522 523 # 524 # Denoiser filter 525 # 526 if [ "$CONFIG_TEMPORAL_DENOISING" = "yes" ]; then 527 prototype int vp8_denoiser_filter "struct yv12_buffer_config* mc_running_avg, struct yv12_buffer_config* running_avg, struct macroblock* signal, unsigned int motion_magnitude2, int y_offset, int uv_offset" 528 specialize vp8_denoiser_filter sse2 529 fi 530 531 # End of encoder only functions 532 fi 533 534 # Scaler functions 535 if [ "CONFIG_SPATIAL_RESAMPLING" != "yes" ]; then 536 prototype void vp8_horizontal_line_4_5_scale "const unsigned char *source, unsigned int source_width, unsigned char *dest, unsigned int dest_width" 537 prototype void vp8_vertical_band_4_5_scale "unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 538 prototype void vp8_last_vertical_band_4_5_scale "unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 539 prototype void vp8_horizontal_line_2_3_scale "const unsigned char *source, unsigned int source_width, unsigned char *dest, unsigned int dest_width" 540 prototype void vp8_vertical_band_2_3_scale "unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 541 prototype void vp8_last_vertical_band_2_3_scale "unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 542 prototype void vp8_horizontal_line_3_5_scale "const unsigned char *source, unsigned int source_width, unsigned char *dest, unsigned int dest_width" 543 prototype void vp8_vertical_band_3_5_scale "unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 544 prototype void vp8_last_vertical_band_3_5_scale "unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 545 prototype void vp8_horizontal_line_3_4_scale "const unsigned char *source, unsigned int source_width, unsigned char *dest, unsigned int dest_width" 546 prototype void vp8_vertical_band_3_4_scale "unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 547 prototype void vp8_last_vertical_band_3_4_scale "unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 548 prototype void vp8_horizontal_line_1_2_scale "const unsigned char *source, unsigned int source_width, unsigned char *dest, unsigned int dest_width" 549 prototype void vp8_vertical_band_1_2_scale "unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 550 prototype void vp8_last_vertical_band_1_2_scale "unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 551 prototype void vp8_horizontal_line_5_4_scale "const unsigned char *source, unsigned int source_width, unsigned char *dest, unsigned int dest_width" 552 prototype void vp8_vertical_band_5_4_scale "unsigned char *source, unsigned int src_pitch, unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 553 prototype void vp8_horizontal_line_5_3_scale "const unsigned char *source, unsigned int source_width, unsigned char *dest, unsigned int dest_width" 554 prototype void vp8_vertical_band_5_3_scale "unsigned char *source, unsigned int src_pitch, unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 555 prototype void vp8_horizontal_line_2_1_scale "const unsigned char *source, unsigned int source_width, unsigned char *dest, unsigned int dest_width" 556 prototype void vp8_vertical_band_2_1_scale "unsigned char *source, unsigned int src_pitch, unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 557 prototype void vp8_vertical_band_2_1_scale_i "unsigned char *source, unsigned int src_pitch, unsigned char *dest, unsigned int dest_pitch, unsigned int dest_width" 558 fi 559 560 prototype void vp8_yv12_extend_frame_borders "struct yv12_buffer_config *ybf" 561 specialize vp8_yv12_extend_frame_borders neon 562 563 prototype void vp8_yv12_copy_frame "struct yv12_buffer_config *src_ybc, struct yv12_buffer_config *dst_ybc" 564 specialize vp8_yv12_copy_frame neon 565 566 prototype void vp8_yv12_copy_y "struct yv12_buffer_config *src_ybc, struct yv12_buffer_config *dst_ybc" 567 specialize vp8_yv12_copy_y neon 568 569
