| /external/libgav1/libgav1/src/dsp/arm/ |
| D | super_res_neon.cc | 217 uint32x4_t res_hi = vmull_u16(vget_high_u16(src[1]), vget_high_u16(f[1])); in SuperRes() local
218 res_hi = vmlal_u16(res_hi, vget_high_u16(src[3]), vget_high_u16(f[3])); in SuperRes()
219 res_hi = vmlal_u16(res_hi, vget_high_u16(src[4]), vget_high_u16(f[4])); in SuperRes()
220 res_hi = vmlal_u16(res_hi, vget_high_u16(src[6]), vget_high_u16(f[6])); in SuperRes()
227 res_hi = vqsubq_u32(res_hi, temp_hi); in SuperRes()
230 vqrshrn_n_u32(res_hi, kFilterBits)); in SuperRes()
|
| D | distance_weighted_blend_neon.cc | 116 const int16x8_t res_hi = in DistanceWeightedBlendLarge_NEON() local
120 vcombine_u8(vqmovun_s16(res_lo), vqmovun_s16(res_hi)); in DistanceWeightedBlendLarge_NEON()
206 const int32x4_t res_hi = vsubq_s32(vreinterpretq_s32_u32(blended_hi), offset); in ComputeWeightedAverage8() local
213 vmin_u16(vqrshrun_n_s32(res_hi, kInterPostRoundBit + 4), bd_max); in ComputeWeightedAverage8()
|
| D | average_blend_neon.cc | 157 const uint16x4_t res_hi = vqrshrun_n_s32(offset_hi, kInterPostRoundBit + 1); in AverageBlend8Row() local
158 return vminq_u16(vcombine_u16(res_lo, res_hi), v_bitdepth); in AverageBlend8Row()
|
| /external/libaom/libaom/av1/common/x86/ |
| D | highbd_warp_plane_sse4.c | 534 __m128i res_hi = _mm_unpackhi_epi32(res_even, res_odd); in av1_highbd_warp_affine_sse4_1() local
572 res_hi = _mm_add_epi32(res_hi, res_add_const); in av1_highbd_warp_affine_sse4_1()
573 res_hi = in av1_highbd_warp_affine_sse4_1()
574 _mm_sra_epi32(_mm_add_epi32(res_hi, reduce_bits_vert_const), in av1_highbd_warp_affine_sse4_1()
582 res_hi = _mm_add_epi32(_mm_mullo_epi32(p4_32, wt0), in av1_highbd_warp_affine_sse4_1()
583 _mm_mullo_epi32(res_hi, wt1)); in av1_highbd_warp_affine_sse4_1()
584 res_hi = _mm_srai_epi32(res_hi, DIST_PRECISION_BITS); in av1_highbd_warp_affine_sse4_1()
586 res_hi = _mm_srai_epi32(_mm_add_epi32(p4_32, res_hi), 1); in av1_highbd_warp_affine_sse4_1()
589 __m128i res32_hi = _mm_add_epi32(res_hi, res_sub_const); in av1_highbd_warp_affine_sse4_1()
596 res_hi = _mm_packus_epi32(res_hi, res_hi); in av1_highbd_warp_affine_sse4_1()
[all …]
|
| D | warp_plane_sse4.c | 544 __m128i *res_lo, __m128i *res_hi, in filter_src_pixels_vertical() argument
578 *res_hi = _mm_unpackhi_epi32(res_even, res_odd); in filter_src_pixels_vertical()
582 __m128i *res_lo, __m128i *res_hi, const __m128i *res_add_const, in store_vertical_filter_output() argument
588 __m128i res_hi_1 = *res_hi; in store_vertical_filter_output()
694 __m128i res_hi; in warp_vertical_filter() local
695 filter_src_pixels_vertical(tmp, coeffs, &res_lo, &res_hi, k); in warp_vertical_filter()
697 store_vertical_filter_output(&res_lo, &res_hi, res_add_const, &wt, in warp_vertical_filter()
722 __m128i res_hi; in warp_vertical_filter_gamma0() local
723 filter_src_pixels_vertical(tmp, coeffs, &res_lo, &res_hi, k); in warp_vertical_filter_gamma0()
725 store_vertical_filter_output(&res_lo, &res_hi, res_add_const, &wt, in warp_vertical_filter_gamma0()
[all …]
|
| D | convolve_sse2.c | 167 __m128i s[8], src6, res_lo, res_hi; in av1_convolve_y_sr_sse2() local
200 res_hi = convolve_hi_y(s, coeffs); // Filter high index pixels 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()
214 res_hi = convolve_hi_y(s + 1, coeffs); // Filter high index pixels 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()
320 const __m128i res_hi = _mm_unpackhi_epi32(res_even, res_odd); in av1_convolve_x_sr_sse2() local
326 _mm_sra_epi32(_mm_add_epi32(res_hi, round_0_const), round_0_shift); in av1_convolve_x_sr_sse2()
|
| D | jnt_convolve_sse2.c | 120 const __m128i res_hi = _mm_unpackhi_epi32(res_even, res_odd); in av1_dist_wtd_convolve_x_sse2() local
124 _mm_sra_epi32(_mm_add_epi32(res_hi, round_const), round_shift); in av1_dist_wtd_convolve_x_sse2()
285 __m128i s[8], src6, res_lo, res_hi, res_lo_shift, res_hi_shift; in av1_dist_wtd_convolve_y_sse2() local
317 res_hi = convolve_hi_y(s, coeffs); // Filter high index pixels in av1_dist_wtd_convolve_y_sse2()
319 res_hi_shift = _mm_sll_epi32(res_hi, left_shift); in av1_dist_wtd_convolve_y_sse2()
347 res_hi = convolve_hi_y(s + 1, coeffs); // Filter high index pixels in av1_dist_wtd_convolve_y_sse2()
349 res_hi_shift = _mm_sll_epi32(res_hi, left_shift); in av1_dist_wtd_convolve_y_sse2()
581 const __m128i res_hi = _mm_unpackhi_epi32(res_even, res_odd); in av1_dist_wtd_convolve_2d_sse2() local
586 _mm_sra_epi32(_mm_add_epi32(res_hi, round_const), round_shift); in av1_dist_wtd_convolve_2d_sse2()
|
| D | jnt_convolve_avx2.c | 330 __m256i res_hi = convolve_lowbd_4tap(s + 3, coeffs + 1); in av1_dist_wtd_convolve_y_avx2() local
332 res_hi = _mm256_add_epi16(res_hi, offset_const_1); in av1_dist_wtd_convolve_y_avx2()
334 const __m256i res_hi_0_32b = _mm256_unpacklo_epi16(res_hi, zero); in av1_dist_wtd_convolve_y_avx2()
340 const __m256i res_hi_1_32b = _mm256_unpackhi_epi16(res_hi, zero); in av1_dist_wtd_convolve_y_avx2()
509 __m256i res_hi = convolve_lowbd(s + 4, coeffs); in av1_dist_wtd_convolve_y_avx2() local
511 res_hi = _mm256_add_epi16(res_hi, offset_const_1); in av1_dist_wtd_convolve_y_avx2()
513 const __m256i res_hi_0_32b = _mm256_unpacklo_epi16(res_hi, zero); in av1_dist_wtd_convolve_y_avx2()
519 const __m256i res_hi_1_32b = _mm256_unpackhi_epi16(res_hi, zero); in av1_dist_wtd_convolve_y_avx2()
|
| D | warp_plane_avx2.c | 628 __m256i *res_hi, int row) { in filter_src_pixels_vertical_avx2() argument
655 *res_hi = _mm256_unpackhi_epi32(res_even, res_odd); in filter_src_pixels_vertical_avx2()
659 const __m256i *res_lo, const __m256i *res_hi, const __m256i *res_add_const, in store_vertical_filter_output_avx2() argument
665 __m256i res_hi_1 = *res_hi; in store_vertical_filter_output_avx2()
810 __m256i res_lo, res_hi; in warp_vertical_filter_avx2() local
811 filter_src_pixels_vertical_avx2(horz_out, src, coeffs, &res_lo, &res_hi, in warp_vertical_filter_avx2()
813 store_vertical_filter_output_avx2(&res_lo, &res_hi, res_add_const, wt, in warp_vertical_filter_avx2()
860 __m256i res_lo, res_hi; in warp_vertical_filter_gamma0_avx2() local
861 filter_src_pixels_vertical_avx2(horz_out, src, coeffs, &res_lo, &res_hi, in warp_vertical_filter_gamma0_avx2()
863 store_vertical_filter_output_avx2(&res_lo, &res_hi, res_add_const, wt, in warp_vertical_filter_gamma0_avx2()
[all …]
|
| D | convolve_avx2.c | 103 const __m256i res_hi = convolve_lowbd_4tap(s + 3, coeffs + 1); in av1_convolve_y_sr_avx2() local
108 _mm256_add_epi16(res_hi, right_shift_const), right_shift); in av1_convolve_y_sr_avx2()
216 const __m256i res_hi = convolve_lowbd(s + 4, coeffs); in av1_convolve_y_sr_avx2() local
221 _mm256_add_epi16(res_hi, right_shift_const), right_shift); in av1_convolve_y_sr_avx2()
|
| D | convolve_2d_sse2.c | 189 const __m128i res_hi = _mm_unpackhi_epi32(res_even, res_odd); in av1_convolve_2d_sr_sse2() local
194 _mm_sra_epi32(_mm_add_epi32(res_hi, sum_round), sum_shift); in av1_convolve_2d_sr_sse2()
405 const __m128i res_hi = _mm_sll_epi16(d16_hi, left_shift); in av1_dist_wtd_convolve_2d_copy_sse2() local
406 const __m128i res_unsigned_hi = _mm_add_epi16(res_hi, offset_const); in av1_dist_wtd_convolve_2d_copy_sse2()
|
| D | highbd_wiener_convolve_ssse3.c | 186 const __m128i res_hi = _mm_unpackhi_epi32(res_even, res_odd); in av1_highbd_wiener_convolve_add_src_ssse3() local
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 | 184 const __m128i res_hi = _mm_unpackhi_epi32(res_even, res_odd); in av1_wiener_convolve_add_src_sse2() local
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 | 226 const __m256i res_hi = _mm256_unpackhi_epi32(res_even, res_odd); in av1_highbd_wiener_convolve_add_src_avx2() local
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 | 197 const __m128i res_hi = _mm_unpackhi_epi32(res_even, res_odd); in av1_dist_wtd_convolve_2d_ssse3() local
202 _mm_sra_epi32(_mm_add_epi32(res_hi, round_const), round_shift); in av1_dist_wtd_convolve_2d_ssse3()
|
| D | highbd_convolve_2d_sse4.c | 349 const __m128i res_hi = _mm_unpackhi_epi32(res_even, res_odd); in av1_highbd_dist_wtd_convolve_2d_sse4_1() local
383 _mm_sra_epi32(_mm_add_epi32(res_hi, round_const), round_shift); in av1_highbd_dist_wtd_convolve_2d_sse4_1()
|
| /external/libaom/libaom/av1/common/arm/ |
| D | warp_plane_neon.c | 488 int32x4_t res_lo, res_hi; in av1_warp_affine_neon() local
627 vertical_filter_neon(v_src, &res_lo, &res_hi, sy, gamma); in av1_warp_affine_neon()
630 res_hi = vaddq_s32(res_hi, add_const_vert); in av1_warp_affine_neon()
667 res_hi = vrshlq_s32(res_hi, shift_vert); in av1_warp_affine_neon()
675 res_hi = vmulq_s32(res_hi, bwd); in av1_warp_affine_neon()
677 tmp32_hi = vaddq_s32(tmp32_hi, res_hi); in av1_warp_affine_neon()
680 tmp32_hi = vaddq_s32(tmp32_hi, res_hi); in av1_warp_affine_neon()
691 uint16x4_t res_u16_high = vqmovun_s32(res_hi); in av1_warp_affine_neon()
697 res_hi = vrshlq_s32(res_hi, shift_vert); in av1_warp_affine_neon()
699 result_final = vcombine_s16(vmovn_s32(res_lo), vmovn_s32(res_hi)); in av1_warp_affine_neon()
|
| /external/mesa3d/src/compiler/nir/ |
| D | nir_lower_int64.c | 305 nir_ssa_def *res_hi = nir_iadd(b, carry, nir_iadd(b, x_hi, y_hi)); in lower_iadd64() local
307 return nir_pack_64_2x32_split(b, res_lo, res_hi); in lower_iadd64()
320 nir_ssa_def *res_hi = nir_iadd(b, nir_isub(b, x_hi, y_hi), borrow); in lower_isub64() local
322 return nir_pack_64_2x32_split(b, res_lo, res_hi); in lower_isub64()
404 nir_ssa_def *res_hi = sign_extend ? nir_imul_high(b, x, y) in lower_mul_2x32_64() local
407 return nir_pack_64_2x32_split(b, nir_imul(b, x, y), res_hi); in lower_mul_2x32_64()
419 nir_ssa_def *res_hi = nir_iadd(b, nir_unpack_64_2x32_split_y(b, mul_lo), in lower_imul64() local
424 res_hi); in lower_imul64()
491 nir_ssa_def *res_hi = nir_ishr_imm(b, x_hi, 31); in lower_isign64() local
492 nir_ssa_def *res_lo = nir_ior(b, res_hi, nir_b2i32(b, is_non_zero)); in lower_isign64()
[all …]
|
| /external/libaom/libaom/aom_dsp/x86/ |
| D | blend_mask_sse4.h | 54 __m128i res_hi = _mm_madd_epi16(_mm_unpackhi_epi16(s0, s1), in blend_a64_d16_mask_w8_sse41() local
57 res_hi = _mm_srai_epi32(_mm_sub_epi32(res_hi, *v_round_offset), shift); in blend_a64_d16_mask_w8_sse41()
58 const __m128i res_e = _mm_packs_epi32(res_lo, res_hi); in blend_a64_d16_mask_w8_sse41()
|
| D | convolve_sse2.h | 91 const __m128i res_hi = _mm_srai_epi32(wt_res_hi, DIST_PRECISION_BITS); in comp_avg() local
93 res = _mm_packs_epi32(res_lo, res_hi); in comp_avg()
|
| D | aom_subpixel_8t_intrin_sse2.c | 63 __m128i res_hi = _mm_unpackhi_epi32(srcRegFilt32b1_1, srcRegFilt32b1_2); in aom_filter_block1d16_h4_sse2() local
64 srcRegFilt32b1_1 = _mm_packs_epi32(res_lo, res_hi); in aom_filter_block1d16_h4_sse2()
87 res_hi = _mm_unpackhi_epi32(srcRegFilt32b2_1, srcRegFilt32b2_2); in aom_filter_block1d16_h4_sse2()
88 srcRegFilt32b2_1 = _mm_packs_epi32(res_lo, res_hi); in aom_filter_block1d16_h4_sse2()
301 __m128i res_hi = _mm_unpackhi_epi32(srcRegFilt32b1_1, srcRegFilt32b1_2); in aom_filter_block1d8_h4_sse2() local
302 srcRegFilt32b1_1 = _mm_packs_epi32(res_lo, res_hi); in aom_filter_block1d8_h4_sse2()
|
| D | convolve_avx2.h | 415 const __m256i res_hi = _mm256_srai_epi32(wt_res_hi, DIST_PRECISION_BITS); in comp_avg() local
417 res = _mm256_packs_epi32(res_lo, res_hi); in comp_avg()
|
| /external/libgav1/libgav1/src/dsp/x86/ |
| D | distance_weighted_blend_sse4.cc | 145 const __m128i res_hi = in DistanceWeightedBlendLarge_SSE4_1() local
148 StoreUnaligned16(dst + x, _mm_packus_epi16(res_lo, res_hi)); in DistanceWeightedBlendLarge_SSE4_1()
353 const __m128i res_hi = in DistanceWeightedBlendLarge_SSE4_1() local
358 StoreUnaligned16(dst + x, res_hi); in DistanceWeightedBlendLarge_SSE4_1()
|
| /external/mesa3d/src/gallium/auxiliary/gallivm/ |
| D | lp_bld_arit.h | 91 LLVMValueRef *res_hi);
97 LLVMValueRef *res_hi);
|
| /external/ruy/ruy/ |
| D | kernel_avx.cc | 236 __m128i res_hi = _mm_srl_epi64(a, hi_count);
240 _mm_castpd_si128(_mm_permute_pd(_mm_castsi128_pd(res_hi), 1));
262 __m128i res_hi = _mm_sll_epi64(a, hi_count);
266 _mm_castpd_si128(_mm_permute_pd(_mm_castsi128_pd(res_hi), 1));
373 __m128i res_hi = _mm_unpackhi_epi64(tmp_lo, tmp_hi);
374 return _mm256_set_m128i(res_hi, res_lo);
|