| /external/libaom/libaom/av1/common/x86/ |
| D | convolve_sse2.c | 86 const __m128i round_const = _mm_set1_epi32((1 << FILTER_BITS) >> 1); in av1_convolve_y_sr_sse2() local
130 res_round = _mm_sra_epi32(_mm_add_epi32(res, round_const), round_shift); in av1_convolve_y_sr_sse2()
143 res_round = _mm_sra_epi32(_mm_add_epi32(res, round_const), round_shift); in av1_convolve_y_sr_sse2()
203 _mm_sra_epi32(_mm_add_epi32(res_lo, round_const), round_shift); in av1_convolve_y_sr_sse2()
205 _mm_sra_epi32(_mm_add_epi32(res_hi, round_const), round_shift); in av1_convolve_y_sr_sse2()
217 _mm_sra_epi32(_mm_add_epi32(res_lo, round_const), round_shift); in av1_convolve_y_sr_sse2()
219 _mm_sra_epi32(_mm_add_epi32(res_hi, round_const), round_shift); in av1_convolve_y_sr_sse2()
250 const __m128i round_const = _mm_set1_epi32((1 << bits) >> 1); in av1_convolve_x_sr_sse2() local
280 _mm_sra_epi32(_mm_add_epi32(res_lo_round, round_const), round_shift); in av1_convolve_x_sr_sse2()
323 res_lo_round = _mm_sra_epi32(_mm_add_epi32(res_lo_round, round_const), in av1_convolve_x_sr_sse2()
[all …]
|
| D | jnt_convolve_sse2.c | 32 const __m128i round_const = _mm_set1_epi32((1 << conv_params->round_0) >> 1); in av1_dist_wtd_convolve_x_sse2() local
69 _mm_sra_epi32(_mm_add_epi32(res_lo, round_const), round_shift); in av1_dist_wtd_convolve_x_sse2()
122 _mm_sra_epi32(_mm_add_epi32(res_lo, round_const), round_shift); in av1_dist_wtd_convolve_x_sse2()
124 _mm_sra_epi32(_mm_add_epi32(res_hi, round_const), round_shift); in av1_dist_wtd_convolve_x_sse2()
178 const __m128i round_const = _mm_set1_epi32((1 << conv_params->round_1) >> 1); in av1_dist_wtd_convolve_y_sse2() local
218 _mm_sra_epi32(_mm_add_epi32(res_shift, round_const), round_shift); in av1_dist_wtd_convolve_y_sse2()
247 _mm_sra_epi32(_mm_add_epi32(res_shift, round_const), round_shift); in av1_dist_wtd_convolve_y_sse2()
320 res_lo_shift = _mm_sra_epi32(_mm_add_epi32(res_lo_shift, round_const), in av1_dist_wtd_convolve_y_sse2()
322 res_hi_shift = _mm_sra_epi32(_mm_add_epi32(res_hi_shift, round_const), in av1_dist_wtd_convolve_y_sse2()
350 res_lo_shift = _mm_sra_epi32(_mm_add_epi32(res_lo_shift, round_const), in av1_dist_wtd_convolve_y_sse2()
[all …]
|
| D | highbd_wiener_convolve_ssse3.c | 65 const __m128i round_const = _mm_set1_epi32( in av1_highbd_wiener_convolve_add_src_ssse3() local
86 res_even = _mm_srai_epi32(_mm_add_epi32(res_even, round_const), in av1_highbd_wiener_convolve_add_src_ssse3()
101 res_odd = _mm_srai_epi32(_mm_add_epi32(res_odd, round_const), in av1_highbd_wiener_convolve_add_src_ssse3()
133 const __m128i round_const = in av1_highbd_wiener_convolve_add_src_ssse3() local
189 _mm_add_epi32(res_lo, round_const), conv_params->round_1); in av1_highbd_wiener_convolve_add_src_ssse3()
191 _mm_add_epi32(res_hi, round_const), conv_params->round_1); in av1_highbd_wiener_convolve_add_src_ssse3()
|
| D | wiener_convolve_sse2.c | 64 const __m128i round_const = _mm_set1_epi32( in av1_wiener_convolve_add_src_sse2() local
84 res_even = _mm_srai_epi32(_mm_add_epi32(res_even, round_const), in av1_wiener_convolve_add_src_sse2()
99 res_odd = _mm_srai_epi32(_mm_add_epi32(res_odd, round_const), in av1_wiener_convolve_add_src_sse2()
131 const __m128i round_const = in av1_wiener_convolve_add_src_sse2() local
187 _mm_add_epi32(res_lo, round_const), conv_params->round_1); in av1_wiener_convolve_add_src_sse2()
189 _mm_add_epi32(res_hi, round_const), conv_params->round_1); in av1_wiener_convolve_add_src_sse2()
|
| D | highbd_wiener_convolve_avx2.c | 87 const __m256i round_const = _mm256_set1_epi32( in av1_highbd_wiener_convolve_add_src_avx2() local
119 _mm256_add_epi32(res_even_sum, round_const), conv_params->round_0); in av1_highbd_wiener_convolve_add_src_avx2()
124 _mm256_add_epi32(res_odd_sum, round_const), conv_params->round_0); in av1_highbd_wiener_convolve_add_src_avx2()
170 const __m256i round_const = in av1_highbd_wiener_convolve_add_src_avx2() local
229 _mm256_add_epi32(res_lo, round_const), conv_params->round_1); in av1_highbd_wiener_convolve_add_src_avx2()
231 _mm256_add_epi32(res_hi, round_const), conv_params->round_1); in av1_highbd_wiener_convolve_add_src_avx2()
|
| D | jnt_convolve_ssse3.c | 75 const __m128i round_const = _mm_set1_epi32( in av1_dist_wtd_convolve_2d_ssse3() local
99 _mm_sra_epi32(_mm_add_epi32(res_even, round_const), round_shift); in av1_dist_wtd_convolve_2d_ssse3()
114 _mm_sra_epi32(_mm_add_epi32(res_odd, round_const), round_shift); in av1_dist_wtd_convolve_2d_ssse3()
143 const __m128i round_const = _mm_set1_epi32( in av1_dist_wtd_convolve_2d_ssse3() local
200 _mm_sra_epi32(_mm_add_epi32(res_lo, round_const), round_shift); in av1_dist_wtd_convolve_2d_ssse3()
202 _mm_sra_epi32(_mm_add_epi32(res_hi, round_const), round_shift); in av1_dist_wtd_convolve_2d_ssse3()
|
| D | warp_plane_avx2.c | 80 const __m256i *round_const, in filter_src_pixels_avx2() argument
95 _mm256_add_epi16(_mm256_add_epi16(res_even, res_odd), *round_const); in filter_src_pixels_avx2()
257 const __m256i *round_const, in horizontal_filter_avx2() argument
261 filter_src_pixels_avx2(src, horz_out, coeff, shuffle_src, round_const, shift, in horizontal_filter_avx2()
302 const __m256i *round_const, const __m128i *shift, in warp_horizontal_filter_avx2() argument
319 round_const, shift); in warp_horizontal_filter_avx2()
328 filter_src_pixels_avx2(src_01, horz_out, coeff, shuffle_src, round_const, in warp_horizontal_filter_avx2()
335 const __m256i *round_const, const __m128i *shift, in warp_horizontal_filter_alpha0_avx2() argument
353 filter_src_pixels_avx2(src_01, horz_out, coeff, shuffle_src, round_const, in warp_horizontal_filter_alpha0_avx2()
363 filter_src_pixels_avx2(src_01, horz_out, coeff, shuffle_src, round_const, in warp_horizontal_filter_alpha0_avx2()
[all …]
|
| D | jnt_convolve_avx2.c | 63 const __m256i round_const = in av1_dist_wtd_convolve_x_avx2() local
93 res = _mm256_sra_epi16(_mm256_add_epi16(res, round_const), round_shift); in av1_dist_wtd_convolve_x_avx2()
147 res = _mm256_sra_epi16(_mm256_add_epi16(res, round_const), round_shift); in av1_dist_wtd_convolve_x_avx2()
202 const __m256i round_const = in av1_dist_wtd_convolve_y_avx2() local
280 _mm256_add_epi32(res_lo_0_shift, round_const), round_shift); in av1_dist_wtd_convolve_y_avx2()
286 _mm256_add_epi32(res_lo_1_shift, round_const), round_shift); in av1_dist_wtd_convolve_y_avx2()
338 _mm256_add_epi32(res_hi_0_shift, round_const), round_shift); in av1_dist_wtd_convolve_y_avx2()
344 _mm256_add_epi32(res_hi_1_shift, round_const), round_shift); in av1_dist_wtd_convolve_y_avx2()
459 _mm256_add_epi32(res_lo_0_shift, round_const), round_shift); in av1_dist_wtd_convolve_y_avx2()
465 _mm256_add_epi32(res_lo_1_shift, round_const), round_shift); in av1_dist_wtd_convolve_y_avx2()
[all …]
|
| D | highbd_convolve_2d_sse4.c | 228 const __m128i round_const = _mm_set1_epi32( in av1_highbd_dist_wtd_convolve_2d_sse4_1() local
251 _mm_sra_epi32(_mm_add_epi32(res_even, round_const), round_shift); in av1_highbd_dist_wtd_convolve_2d_sse4_1()
266 _mm_sra_epi32(_mm_add_epi32(res_odd, round_const), round_shift); in av1_highbd_dist_wtd_convolve_2d_sse4_1()
295 const __m128i round_const = _mm_set1_epi32( in av1_highbd_dist_wtd_convolve_2d_sse4_1() local
352 _mm_sra_epi32(_mm_add_epi32(res_lo, round_const), round_shift); in av1_highbd_dist_wtd_convolve_2d_sse4_1()
383 _mm_sra_epi32(_mm_add_epi32(res_hi, round_const), round_shift); in av1_highbd_dist_wtd_convolve_2d_sse4_1()
|
| D | convolve_2d_sse2.c | 65 const __m128i round_const = _mm_set1_epi32( in av1_convolve_2d_sr_sse2() local
87 _mm_sra_epi32(_mm_add_epi32(res_even, round_const), round_shift); in av1_convolve_2d_sr_sse2()
102 _mm_sra_epi32(_mm_add_epi32(res_odd, round_const), round_shift); in av1_convolve_2d_sr_sse2()
135 const __m128i round_const = _mm_set1_epi32( in av1_convolve_2d_sr_sse2() local
196 res_lo_round = _mm_sra_epi32(_mm_add_epi32(res_lo_round, round_const), in av1_convolve_2d_sr_sse2()
198 res_hi_round = _mm_sra_epi32(_mm_add_epi32(res_hi_round, round_const), in av1_convolve_2d_sr_sse2()
|
| D | convolve_avx2.c | 274 const __m256i round_const = _mm256_set1_epi16((1 << bits) >> 1); in av1_convolve_x_sr_avx2() local
313 res_16b = _mm256_sra_epi16(_mm256_add_epi16(res_16b, round_const), in av1_convolve_x_sr_avx2()
351 res_16b = _mm256_sra_epi16(_mm256_add_epi16(res_16b, round_const), in av1_convolve_x_sr_avx2()
386 res_16b = _mm256_sra_epi16(_mm256_add_epi16(res_16b, round_const), in av1_convolve_x_sr_avx2()
423 res_16b = _mm256_sra_epi16(_mm256_add_epi16(res_16b, round_const), in av1_convolve_x_sr_avx2()
|
| D | reconinter_sse4.c | 105 const __m128i round_const = _mm_set1_epi16((1 << round) >> 1); in av1_build_compound_diffwtd_mask_d16_sse4_1() local
130 _mm_srli_epi16(_mm_adds_epu16(diff, round_const), round); in av1_build_compound_diffwtd_mask_d16_sse4_1()
|
| D | highbd_warp_plane_sse4.c | 121 const __m128i round_const = _mm_set1_epi32((1 << offset_bits_horiz) + in highbd_filter_src_pixels() local
134 res_even = _mm_sra_epi32(_mm_add_epi32(res_even, round_const), in highbd_filter_src_pixels()
148 res_odd = _mm_sra_epi32(_mm_add_epi32(res_odd, round_const), in highbd_filter_src_pixels()
602 const __m128i round_const = in av1_highbd_warp_affine_sse4_1() local
607 _mm_add_epi32(res_lo, round_const), reduce_bits_vert); in av1_highbd_warp_affine_sse4_1()
609 _mm_add_epi32(res_hi, round_const), reduce_bits_vert); in av1_highbd_warp_affine_sse4_1()
|
| D | reconinter_avx2.c | 140 const __m256i *round_const, in calc_mask_d16_avx2() argument
147 _mm256_srli_epi16(_mm256_adds_epu16(diff, *round_const), round); in calc_mask_d16_avx2()
156 const __m256i *round_const, in calc_mask_d16_inv_avx2() argument
164 _mm256_srli_epi16(_mm256_adds_epu16(diff, *round_const), round); in calc_mask_d16_inv_avx2()
|
| D | warp_plane_sse4.c | 265 const __m128i round_const = _mm_set1_epi16((1 << offset_bits_horiz) + in filter_src_pixels() local
279 _mm_add_epi16(_mm_add_epi16(res_even, res_odd), round_const); in filter_src_pixels()
|
| /external/libaom/libaom/aom_dsp/x86/ |
| D | convolve_sse2.h | 103 const __m128i *const round_const, in convolve_rounding() argument
107 _mm_srai_epi16(_mm_add_epi16(res_signed, *round_const), round_shift); in convolve_rounding()
113 const __m128i *const round_const, const int round_shift) { in highbd_convolve_rounding_sse2() argument
116 _mm_srai_epi32(_mm_add_epi32(res_signed, *round_const), round_shift); in highbd_convolve_rounding_sse2()
|
| D | masked_sad_intrin_ssse3.c | 296 const __m128i round_const = in highbd_masked_sad_ssse3() local
313 pred_l = _mm_srai_epi32(_mm_add_epi32(pred_l, round_const), in highbd_masked_sad_ssse3()
319 pred_r = _mm_srai_epi32(_mm_add_epi32(pred_r, round_const), in highbd_masked_sad_ssse3()
355 const __m128i round_const = in aom_highbd_masked_sad4xh_ssse3() local
380 pred_l = _mm_srai_epi32(_mm_add_epi32(pred_l, round_const), in aom_highbd_masked_sad4xh_ssse3()
386 pred_r = _mm_srai_epi32(_mm_add_epi32(pred_r, round_const), in aom_highbd_masked_sad4xh_ssse3()
|
| D | masked_sad_intrin_avx2.c | 212 const __m256i round_const = in highbd_masked_sad8xh_avx2() local
229 pred_l = _mm256_srai_epi32(_mm256_add_epi32(pred_l, round_const), in highbd_masked_sad8xh_avx2()
235 pred_r = _mm256_srai_epi32(_mm256_add_epi32(pred_r, round_const), in highbd_masked_sad8xh_avx2()
269 const __m256i round_const = in highbd_masked_sad16xh_avx2() local
286 pred_l = _mm256_srai_epi32(_mm256_add_epi32(pred_l, round_const), in highbd_masked_sad16xh_avx2()
292 pred_r = _mm256_srai_epi32(_mm256_add_epi32(pred_r, round_const), in highbd_masked_sad16xh_avx2()
|
| D | convolve_avx2.h | 427 const __m256i *const round_const, in convolve_rounding() argument
431 _mm256_add_epi16(res_signed, *round_const), round_shift); in convolve_rounding()
455 const __m256i *const round_const, const int round_shift) { in highbd_convolve_rounding() argument
458 _mm256_add_epi32(res_signed, *round_const), round_shift); in highbd_convolve_rounding()
|
| D | variance_avx2.c | 424 const __m256i round_const = in highbd_comp_mask_pred_line_avx2() local
432 _mm256_add_epi32(pred_lo, round_const), AOM_BLEND_A64_ROUND_BITS); in highbd_comp_mask_pred_line_avx2()
438 _mm256_add_epi32(pred_hi, round_const), AOM_BLEND_A64_ROUND_BITS); in highbd_comp_mask_pred_line_avx2()
|
| D | masked_variance_intrin_ssse3.c | 906 const __m128i round_const = in highbd_masked_variance() local
923 pred_l = _mm_srai_epi32(_mm_add_epi32(pred_l, round_const), in highbd_masked_variance()
929 pred_r = _mm_srai_epi32(_mm_add_epi32(pred_r, round_const), in highbd_masked_variance()
978 const __m128i round_const = in highbd_masked_variance4xh() local
998 pred_l = _mm_srai_epi32(_mm_add_epi32(pred_l, round_const), in highbd_masked_variance4xh()
1004 pred_r = _mm_srai_epi32(_mm_add_epi32(pred_r, round_const), in highbd_masked_variance4xh()
|
| D | variance_sse2.c | 655 const __m128i round_const = in highbd_comp_mask_pred_line_sse2() local
662 const __m128i pred_l = _mm_srai_epi32(_mm_add_epi32(pred_lo, round_const), in highbd_comp_mask_pred_line_sse2()
668 const __m128i pred_h = _mm_srai_epi32(_mm_add_epi32(pred_hi, round_const), in highbd_comp_mask_pred_line_sse2()
|
| /external/libaom/libaom/av1/common/arm/ |
| D | convolve_neon.h | 29 const int32_t round_const = (1 << (bd + round1_bits - 1)); in wiener_convolve8_vert_4x8() local
32 const int32x4_t round_vec = vdupq_n_s32(round_const); in wiener_convolve8_vert_4x8()
|
| /external/vixl/src/aarch64/ |
| D | logic-aarch64.cc | 3952 int round_const = round ? (1 << (esize - 2)) : 0; in sqrdmulh() local
3958 product += round_const; in sqrdmulh()
4030 int round_const = round ? (1 << (esize - 2)) : 0; in sqrdmlash() local
4041 accum += round_const; in sqrdmlash()
|
| /external/arm-neon-tests/ |
| D | ref-rvct-neon-nofp16.txt | 2113 VRSHL/VRSHLQ (checking round_const overflow: shift by -1) output:
2137 VRSHL/VRSHLQ (checking round_const overflow: shift by -3) output:
|