| /external/libaom/libaom/av1/common/x86/ |
| D | highbd_jnt_convolve_sse4.c | 50 const __m128i offset_const = _mm_set1_epi32(offset); in av1_highbd_dist_wtd_convolve_y_sse4_1() local
111 __m128i res_unsigned_lo_0 = _mm_add_epi32(res_a_round0, offset_const); in av1_highbd_dist_wtd_convolve_y_sse4_1()
112 __m128i res_unsigned_lo_1 = _mm_add_epi32(res_a_round1, offset_const); in av1_highbd_dist_wtd_convolve_y_sse4_1()
132 highbd_convolve_rounding_sse2(&comp_avg_res_0, &offset_const, in av1_highbd_dist_wtd_convolve_y_sse4_1()
135 highbd_convolve_rounding_sse2(&comp_avg_res_1, &offset_const, in av1_highbd_dist_wtd_convolve_y_sse4_1()
175 __m128i res_unsigned_hi_0 = _mm_add_epi32(res_b_round0, offset_const); in av1_highbd_dist_wtd_convolve_y_sse4_1()
176 __m128i res_unsigned_hi_1 = _mm_add_epi32(res_b_round1, offset_const); in av1_highbd_dist_wtd_convolve_y_sse4_1()
203 highbd_convolve_rounding_sse2(&comp_avg_res_lo_0, &offset_const, in av1_highbd_dist_wtd_convolve_y_sse4_1()
206 highbd_convolve_rounding_sse2(&comp_avg_res_lo_1, &offset_const, in av1_highbd_dist_wtd_convolve_y_sse4_1()
209 highbd_convolve_rounding_sse2(&comp_avg_res_hi_0, &offset_const, in av1_highbd_dist_wtd_convolve_y_sse4_1()
[all …]
|
| D | highbd_jnt_convolve_avx2.c | 52 const __m256i offset_const = _mm256_set1_epi32(offset); in av1_highbd_dist_wtd_convolve_2d_copy_avx2() local
79 _mm256_add_epi32(res_32b_lo, offset_const); in av1_highbd_dist_wtd_convolve_2d_copy_avx2()
87 _mm256_add_epi32(res_32b_hi, offset_const); in av1_highbd_dist_wtd_convolve_2d_copy_avx2()
94 &comp_avg_res_lo, &offset_const, &rounding_const, rounding_shift); in av1_highbd_dist_wtd_convolve_2d_copy_avx2()
96 &comp_avg_res_hi, &offset_const, &rounding_const, rounding_shift); in av1_highbd_dist_wtd_convolve_2d_copy_avx2()
138 _mm256_add_epi32(res_32b, offset_const); in av1_highbd_dist_wtd_convolve_2d_copy_avx2()
145 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); in av1_highbd_dist_wtd_convolve_2d_copy_avx2()
183 _mm256_add_epi32(res_32b_lo, offset_const); in av1_highbd_dist_wtd_convolve_2d_copy_avx2()
191 _mm256_add_epi32(res_32b_hi, offset_const); in av1_highbd_dist_wtd_convolve_2d_copy_avx2()
198 highbd_convolve_rounding(&comp_avg_res_lo, &offset_const, in av1_highbd_dist_wtd_convolve_2d_copy_avx2()
[all …]
|
| D | highbd_convolve_2d_sse4.c | 51 const __m128i offset_const = _mm_set1_epi32(offset); in av1_highbd_dist_wtd_convolve_2d_copy_sse4_1() local
76 _mm_add_epi32(res_32b_lo, offset_const); in av1_highbd_dist_wtd_convolve_2d_copy_sse4_1()
84 _mm_add_epi32(res_32b_hi, offset_const); in av1_highbd_dist_wtd_convolve_2d_copy_sse4_1()
91 &comp_avg_res_lo, &offset_const, &rounding_const, rounding_shift); in av1_highbd_dist_wtd_convolve_2d_copy_sse4_1()
93 &comp_avg_res_hi, &offset_const, &rounding_const, rounding_shift); in av1_highbd_dist_wtd_convolve_2d_copy_sse4_1()
130 const __m128i res_unsigned_lo = _mm_add_epi32(res_32b, offset_const); in av1_highbd_dist_wtd_convolve_2d_copy_sse4_1()
134 _mm_add_epi32(res_32b_hi, offset_const); in av1_highbd_dist_wtd_convolve_2d_copy_sse4_1()
142 &comp_avg_res_lo, &offset_const, &rounding_const, rounding_shift); in av1_highbd_dist_wtd_convolve_2d_copy_sse4_1()
144 &comp_avg_res_hi, &offset_const, &rounding_const, rounding_shift); in av1_highbd_dist_wtd_convolve_2d_copy_sse4_1()
197 const __m128i offset_const = _mm_set1_epi32(offset); in av1_highbd_dist_wtd_convolve_2d_sse4_1() local
[all …]
|
| D | jnt_convolve_avx2.c | 55 const __m256i offset_const = _mm256_set1_epi16(offset); in av1_dist_wtd_convolve_x_avx2() local
96 const __m256i res_unsigned = _mm256_add_epi16(res, offset_const); in av1_dist_wtd_convolve_x_avx2()
106 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); in av1_dist_wtd_convolve_x_avx2()
151 const __m256i res_unsigned = _mm256_add_epi16(res, offset_const); in av1_dist_wtd_convolve_x_avx2()
161 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); in av1_dist_wtd_convolve_x_avx2()
211 const __m256i offset_const = _mm256_set1_epi16(offset); in av1_dist_wtd_convolve_y_avx2() local
303 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); in av1_dist_wtd_convolve_y_avx2()
368 convolve_rounding(&comp_avg_res_lo, &offset_const, in av1_dist_wtd_convolve_y_avx2()
372 convolve_rounding(&comp_avg_res_hi, &offset_const, in av1_dist_wtd_convolve_y_avx2()
482 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); in av1_dist_wtd_convolve_y_avx2()
[all …]
|
| D | jnt_convolve_sse2.c | 44 const __m128i offset_const = _mm_set1_epi16(offset); in av1_dist_wtd_convolve_x_sse2() local
73 const __m128i res_unsigned = _mm_add_epi16(res_16b, offset_const); in av1_dist_wtd_convolve_x_sse2()
83 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); in av1_dist_wtd_convolve_x_sse2()
129 const __m128i res_unsigned = _mm_add_epi16(res_16b, offset_const); in av1_dist_wtd_convolve_x_sse2()
140 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); in av1_dist_wtd_convolve_x_sse2()
174 const __m128i offset_const = _mm_set1_epi16(offset); in av1_dist_wtd_convolve_y_sse2() local
221 __m128i res_unsigned = _mm_add_epi16(res_16b, offset_const); in av1_dist_wtd_convolve_y_sse2()
231 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); in av1_dist_wtd_convolve_y_sse2()
250 res_unsigned = _mm_add_epi16(res_16b, offset_const); in av1_dist_wtd_convolve_y_sse2()
260 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); in av1_dist_wtd_convolve_y_sse2()
[all …]
|
| D | convolve_2d_sse2.c | 387 const __m128i offset_const = _mm_set1_epi16(offset); in av1_dist_wtd_convolve_2d_copy_sse2() local
403 const __m128i res_unsigned_lo = _mm_add_epi16(res_lo, offset_const); in av1_dist_wtd_convolve_2d_copy_sse2()
406 const __m128i res_unsigned_hi = _mm_add_epi16(res_hi, offset_const); in av1_dist_wtd_convolve_2d_copy_sse2()
417 &comp_avg_res_lo, &offset_const, &rounding_const, rounding_shift); in av1_dist_wtd_convolve_2d_copy_sse2()
423 &comp_avg_res_hi, &offset_const, &rounding_const, rounding_shift); in av1_dist_wtd_convolve_2d_copy_sse2()
445 const __m128i res_unsigned = _mm_add_epi16(res, offset_const); in av1_dist_wtd_convolve_2d_copy_sse2()
454 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); in av1_dist_wtd_convolve_2d_copy_sse2()
|
| D | jnt_convolve_ssse3.c | 50 const __m128i offset_const = _mm_set1_epi16(offset); in av1_dist_wtd_convolve_2d_ssse3() local
205 const __m128i res_unsigned = _mm_add_epi16(res_16b, offset_const); in av1_dist_wtd_convolve_2d_ssse3()
216 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); in av1_dist_wtd_convolve_2d_ssse3()
|
| /external/libaom/libaom/aom_dsp/x86/ |
| D | convolve_sse2.h | 102 const __m128i *const offset_const, in convolve_rounding() argument
105 const __m128i res_signed = _mm_sub_epi16(*res_unsigned, *offset_const); in convolve_rounding()
112 const __m128i *const res_unsigned, const __m128i *const offset_const, in highbd_convolve_rounding_sse2() argument
114 const __m128i res_signed = _mm_sub_epi32(*res_unsigned, *offset_const); in highbd_convolve_rounding_sse2()
|
| D | convolve_avx2.h | 204 const __m256i res_unsigned = _mm256_add_epi16(res_16b, offset_const); \
213 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); \
232 const __m256i res_unsigned = _mm256_add_epi16(res_16b, offset_const); \
242 &comp_avg_res, &offset_const, &rounding_const, rounding_shift); \
426 const __m256i *const offset_const, in convolve_rounding() argument
429 const __m256i res_signed = _mm256_sub_epi16(*res_unsigned, *offset_const); in convolve_rounding()
454 const __m256i *const res_unsigned, const __m256i *const offset_const, in highbd_convolve_rounding() argument
456 const __m256i res_signed = _mm256_sub_epi32(*res_unsigned, *offset_const); in highbd_convolve_rounding()
|
| /external/libaom/libaom/av1/common/arm/ |
| D | convolve_neon.c | 125 const int32x4_t round_shift_vec, const int32x4_t offset_const, in convolve8_vert_4x4_s32() argument
140 sum0 = vaddq_s32(sum0, offset_const); in convolve8_vert_4x4_s32()
154 const int32x4_t round_shift_vec, const int32x4_t offset_const, in convolve8_vert_8x4_s32() argument
178 sum0 = vaddq_s32(sum0, offset_const); in convolve8_vert_8x4_s32()
179 sum1 = vaddq_s32(sum1, offset_const); in convolve8_vert_8x4_s32()
1293 const int32x4_t offset_const = vdupq_n_s32(1 << offset_bits); in av1_convolve_2d_sr_neon() local
1342 round_shift_vec, offset_const, in av1_convolve_2d_sr_neon()
1345 round_shift_vec, offset_const, in av1_convolve_2d_sr_neon()
1348 round_shift_vec, offset_const, in av1_convolve_2d_sr_neon()
1351 round_shift_vec, offset_const, in av1_convolve_2d_sr_neon()
[all …]
|
| D | convolve_neon.h | 206 const int32x4_t round_shift_vec, const int32x4_t offset_const) { in convolve8_4x4_s32() argument
220 sum0 = vaddq_s32(sum0, offset_const); in convolve8_4x4_s32()
|
| D | jnt_convolve_neon.c | 581 const int32x4_t offset_const = vdupq_n_s32(1 << offset); in dist_wtd_convolve_2d_vert_neon() local
638 round_shift_vec, offset_const); in dist_wtd_convolve_2d_vert_neon()
640 round_shift_vec, offset_const); in dist_wtd_convolve_2d_vert_neon()
642 round_shift_vec, offset_const); in dist_wtd_convolve_2d_vert_neon()
644 round_shift_vec, offset_const); in dist_wtd_convolve_2d_vert_neon()
683 round_shift_vec, offset_const); in dist_wtd_convolve_2d_vert_neon()
|