/external/libaom/libaom/aom_dsp/x86/ |
D | avg_intrin_avx2.c | 30 __m256i b1 = _mm256_sub_epi16(a0, a1); in hadamard_col8x2_avx2() 32 __m256i b3 = _mm256_sub_epi16(a2, a3); in hadamard_col8x2_avx2() 34 __m256i b5 = _mm256_sub_epi16(a4, a5); in hadamard_col8x2_avx2() 36 __m256i b7 = _mm256_sub_epi16(a6, a7); in hadamard_col8x2_avx2() 40 a2 = _mm256_sub_epi16(b0, b2); in hadamard_col8x2_avx2() 41 a3 = _mm256_sub_epi16(b1, b3); in hadamard_col8x2_avx2() 44 a6 = _mm256_sub_epi16(b4, b6); in hadamard_col8x2_avx2() 45 a7 = _mm256_sub_epi16(b5, b7); in hadamard_col8x2_avx2() 52 b2 = _mm256_sub_epi16(a0, a4); in hadamard_col8x2_avx2() 53 b6 = _mm256_sub_epi16(a1, a5); in hadamard_col8x2_avx2() [all …]
|
D | sse_avx2.c | 29 const __m256i v_d00_w = _mm256_sub_epi16(v_a00_w, v_b00_w); in sse_w32_avx2() 30 const __m256i v_d01_w = _mm256_sub_epi16(v_a01_w, v_b01_w); in sse_w32_avx2() 85 const __m256i v_d_w = _mm256_sub_epi16(v_a_w, v_b_w); in sse_w4x4_avx2() 96 const __m256i v_d_w = _mm256_sub_epi16(v_a_w, v_b_w); in sse_w8x2_avx2() 138 const __m256i v_asub = _mm256_sub_epi16(v_al, v_bl); in aom_sse_avx2() 139 const __m256i v_bsub = _mm256_sub_epi16(v_au, v_bu); in aom_sse_avx2() 221 const __m256i v_d_w = _mm256_sub_epi16(v_a_w, v_b_w); in highbd_sse_w16_avx2() 240 const __m256i v_d_w = _mm256_sub_epi16(v_a_w, v_b_w); in highbd_sse_w4x4_avx2() 249 const __m256i v_d_w = _mm256_sub_epi16(v_a_w, v_b_w); in highbd_sse_w8x2_avx2()
|
D | subtract_avx2.c | 23 const __m256i d_0 = _mm256_sub_epi16(s_0, p_0); in subtract32_avx2() 24 const __m256i d_1 = _mm256_sub_epi16(s_1, p_1); in subtract32_avx2() 37 const __m256i d_0 = _mm256_sub_epi16(s_0, p_0); in subtract_block_16xn_avx2()
|
D | masked_sad_intrin_avx2.c | 224 const __m256i m_inv = _mm256_sub_epi16(mask_max, m); in highbd_masked_sad8xh_avx2() 244 const __m256i diff = _mm256_abs_epi16(_mm256_sub_epi16(pred, src)); in highbd_masked_sad8xh_avx2() 281 const __m256i m_inv = _mm256_sub_epi16(mask_max, m); in highbd_masked_sad16xh_avx2() 301 const __m256i diff = _mm256_abs_epi16(_mm256_sub_epi16(pred, src)); in highbd_masked_sad16xh_avx2()
|
D | highbd_variance_avx2.c | 35 const __m256i v_diff = _mm256_sub_epi16(v_p_a, v_p_b); in aom_highbd_calc8x8var_avx2() 65 const __m256i v_diff = _mm256_sub_epi16(v_p_a, v_p_b); in aom_highbd_calc16x16var_avx2()
|
D | intrapred_avx2.c | 715 _mm256_sub_epi16(_mm256_add_epi16(*top, *left), *topleft); in paeth_pred() 717 __m256i pl = _mm256_abs_epi16(_mm256_sub_epi16(base, *left)); in paeth_pred() 718 __m256i pt = _mm256_abs_epi16(_mm256_sub_epi16(base, *top)); in paeth_pred() 719 __m256i ptl = _mm256_abs_epi16(_mm256_sub_epi16(base, *topleft)); in paeth_pred() 1100 diff = _mm256_sub_epi16(a1, a0); // a[x+1] - a[x] in highbd_dr_prediction_z1_4xN_internal_avx2() 1353 diff = _mm256_sub_epi16(a1, a0); // a[x+1] - a[x] in highbd_dr_prediction_z1_8xN_internal_avx2() 1503 diff = _mm256_sub_epi16(a1, a0); // a[x+1] - a[x] in highbd_dr_prediction_z1_16xN_internal_avx2() 1681 diff = _mm256_sub_epi16(a1, a0); // a[x+1] - a[x] in highbd_dr_prediction_z1_32xN_internal_avx2() 1876 diff = _mm256_sub_epi16(a1, a0); // a[x+1] - a[x] in highbd_dr_prediction_z1_64xN_avx2() 2193 diff = _mm256_sub_epi16(a1_x, a0_x); // a[x+1] - a[x] in highbd_dr_prediction_z2_Nx4_avx2() [all …]
|
D | blend_a64_mask_avx2.c | 32 const __m256i max_minus_m0 = _mm256_sub_epi16(*v_maxval, *m0); in blend_a64_d16_mask_w16_avx2() 53 const __m256i max_minus_m0 = _mm256_sub_epi16(*v_maxval, *m0); in blend_a64_d16_mask_w32_avx2() 54 const __m256i max_minus_m1 = _mm256_sub_epi16(*v_maxval, *m1); in blend_a64_d16_mask_w32_avx2() 922 const __m256i mask1 = _mm256_sub_epi16(*mask_max, *mask0); in highbd_blend_a64_d16_mask_w4_avx2() 1039 const __m256i mask1a = _mm256_sub_epi16(*mask_max, *mask0a); in highbd_blend_a64_d16_mask_w8_avx2() 1040 const __m256i mask1b = _mm256_sub_epi16(*mask_max, *mask0b); in highbd_blend_a64_d16_mask_w8_avx2() 1175 const __m256i mask1a = _mm256_sub_epi16(*mask_max, *mask0a); in highbd_blend_a64_d16_mask_w16_avx2() 1176 const __m256i mask1b = _mm256_sub_epi16(*mask_max, *mask0b); in highbd_blend_a64_d16_mask_w16_avx2()
|
D | variance_impl_avx2.c | 75 exp_src_lo = _mm256_sub_epi16(exp_src_lo, exp_dst_lo); \ 76 exp_src_hi = _mm256_sub_epi16(exp_src_hi, exp_dst_hi); \
|
D | adaptive_quantize_avx2.c | 27 *zbin = _mm256_sub_epi16(*zbin, _mm256_set1_epi16(1)); in load_b_values_avx2()
|
D | convolve_avx2.h | 429 const __m256i res_signed = _mm256_sub_epi16(*res_unsigned, *offset_const); in convolve_rounding()
|
/external/libvpx/libvpx/vpx_dsp/x86/ |
D | avg_intrin_avx2.c | 232 __m256i b1 = _mm256_sub_epi16(a0, a1); in hadamard_col8x2_avx2() 234 __m256i b3 = _mm256_sub_epi16(a2, a3); in hadamard_col8x2_avx2() 236 __m256i b5 = _mm256_sub_epi16(a4, a5); in hadamard_col8x2_avx2() 238 __m256i b7 = _mm256_sub_epi16(a6, a7); in hadamard_col8x2_avx2() 242 a2 = _mm256_sub_epi16(b0, b2); in hadamard_col8x2_avx2() 243 a3 = _mm256_sub_epi16(b1, b3); in hadamard_col8x2_avx2() 246 a6 = _mm256_sub_epi16(b4, b6); in hadamard_col8x2_avx2() 247 a7 = _mm256_sub_epi16(b5, b7); in hadamard_col8x2_avx2() 254 b2 = _mm256_sub_epi16(a0, a4); in hadamard_col8x2_avx2() 255 b6 = _mm256_sub_epi16(a1, a5); in hadamard_col8x2_avx2() [all …]
|
D | fwd_dct32x32_impl_avx2.h | 135 step1b[-3] = _mm256_sub_epi16(ina3, inb3); in FDCT32x32_2D_AVX2() 136 step1b[-2] = _mm256_sub_epi16(ina2, inb2); in FDCT32x32_2D_AVX2() 137 step1b[-1] = _mm256_sub_epi16(ina1, inb1); in FDCT32x32_2D_AVX2() 138 step1b[-0] = _mm256_sub_epi16(ina0, inb0); in FDCT32x32_2D_AVX2() 171 step1b[-3] = _mm256_sub_epi16(ina3, inb3); in FDCT32x32_2D_AVX2() 172 step1b[-2] = _mm256_sub_epi16(ina2, inb2); in FDCT32x32_2D_AVX2() 173 step1b[-1] = _mm256_sub_epi16(ina1, inb1); in FDCT32x32_2D_AVX2() 174 step1b[-0] = _mm256_sub_epi16(ina0, inb0); in FDCT32x32_2D_AVX2() 207 step1b[-3] = _mm256_sub_epi16(ina3, inb3); in FDCT32x32_2D_AVX2() 208 step1b[-2] = _mm256_sub_epi16(ina2, inb2); in FDCT32x32_2D_AVX2() [all …]
|
D | loopfilter_avx2.c | 655 pixelFilter_q = _mm256_sub_epi16(pixelFilter_p, p256_6); in vpx_lpf_horizontal_16_dual_avx2() 657 pixelFilter_p = _mm256_sub_epi16(pixelFilter_p, q256_6); in vpx_lpf_horizontal_16_dual_avx2() 671 pixetFilter_q2q1q0 = _mm256_sub_epi16(pixetFilter_p2p1p0, p256_2); in vpx_lpf_horizontal_16_dual_avx2() 673 pixetFilter_p2p1p0 = _mm256_sub_epi16(pixetFilter_p2p1p0, q256_2); in vpx_lpf_horizontal_16_dual_avx2() 699 pixelFilter_p = _mm256_sub_epi16(pixelFilter_p, q256_5); in vpx_lpf_horizontal_16_dual_avx2() 701 pixelFilter_q = _mm256_sub_epi16(pixelFilter_q, p256_5); in vpx_lpf_horizontal_16_dual_avx2() 715 pixetFilter_p2p1p0 = _mm256_sub_epi16(pixetFilter_p2p1p0, q256_1); in vpx_lpf_horizontal_16_dual_avx2() 717 pixetFilter_q2q1q0 = _mm256_sub_epi16(pixetFilter_q2q1q0, p256_1); in vpx_lpf_horizontal_16_dual_avx2() 739 pixelFilter_p = _mm256_sub_epi16(pixelFilter_p, q256_4); in vpx_lpf_horizontal_16_dual_avx2() 741 pixelFilter_q = _mm256_sub_epi16(pixelFilter_q, p256_4); in vpx_lpf_horizontal_16_dual_avx2() [all …]
|
D | variance_avx2.c | 193 exp_src_lo = _mm256_sub_epi16(exp_src_lo, exp_dst_lo); \ 194 exp_src_hi = _mm256_sub_epi16(exp_src_hi, exp_dst_hi); \
|
/external/libvpx/libvpx/vp9/encoder/x86/ |
D | vp9_error_avx2.c | 35 dqcoeff_256 = _mm256_sub_epi16(dqcoeff_256, coeff_256); in vp9_block_error_avx2() 63 dqcoeff_0 = _mm256_sub_epi16(dqcoeff_0, coeff_0); in vp9_block_error_avx2() 64 dqcoeff_1 = _mm256_sub_epi16(dqcoeff_1, coeff_1); in vp9_block_error_avx2() 123 const __m256i diff = _mm256_sub_epi16(_dqcoeff, _coeff); in vp9_block_error_fp_avx2() 137 const __m256i diff = _mm256_sub_epi16(_dqcoeff, _coeff); in vp9_block_error_fp_avx2()
|
D | vp9_quantize_avx2.c | 45 const __m256i iscan_plus_one = _mm256_sub_epi16(iscan, nzero_coeff0); in scan_eob_256()
|
/external/libaom/libaom/av1/encoder/x86/ |
D | error_intrin_avx2.c | 45 const __m256i diff = _mm256_sub_epi16(_dqcoeff, _coeff); in av1_block_error_lp_avx2() 59 const __m256i diff = _mm256_sub_epi16(_dqcoeff, _coeff); in av1_block_error_lp_avx2() 104 dqcoeff_reg = _mm256_sub_epi16(dqcoeff_reg, coeff_reg); in av1_block_error_avx2()
|
D | av1_quantize_avx2.c | 126 __m256i cur_eob = _mm256_sub_epi16(iscan, nzero_coeff); in quantize() 142 const __m256i iscan_plus_one = _mm256_sub_epi16(iscan, nzero_coeff0); in scan_eob_256() 317 __m256i cur_eob = _mm256_sub_epi16(iscan, nzero_coeff); in quantize_32x32() 396 __m256i cur_eob = _mm256_sub_epi16(iscan, nzero_coeff); in quantize_64x64()
|
D | pickrst_avx2.c | 528 const __m256i flt0_0_sub_u = _mm256_sub_epi16(flt0_16b, u0); in av1_lowbd_pixel_proj_error_avx2() 529 const __m256i flt1_0_sub_u = _mm256_sub_epi16(flt1_16b, u0); in av1_lowbd_pixel_proj_error_avx2() 538 const __m256i e0 = _mm256_sub_epi16( in av1_lowbd_pixel_proj_error_avx2() 583 const __m256i e0 = _mm256_sub_epi16( in av1_lowbd_pixel_proj_error_avx2() 610 const __m256i diff0 = _mm256_sub_epi16(d0, s0); in av1_lowbd_pixel_proj_error_avx2() 934 const __m256i e0 = _mm256_sub_epi16(_mm256_add_epi16(vr, d0), s0); in av1_highbd_pixel_proj_error_avx2() 1011 const __m256i e0 = _mm256_sub_epi16(_mm256_add_epi16(vr, d0), s0); in av1_highbd_pixel_proj_error_avx2() 1050 const __m256i diffl = _mm256_sub_epi16(d0l, s0l); in av1_highbd_pixel_proj_error_avx2() 1051 const __m256i diffh = _mm256_sub_epi16(d0h, s0h); in av1_highbd_pixel_proj_error_avx2()
|
D | temporal_filter_avx2.c | 49 vdiff1 = _mm256_sub_epi16(vf1, vf2); in get_squared_error_16x16_avx2()
|
/external/libaom/libaom/av1/common/x86/ |
D | reconinter_avx2.c | 24 const __m256i diff = _mm256_abs_epi16(_mm256_sub_epi16(s0, s1)); in calc_mask_avx2() 168 const __m256i diff_const_16 = _mm256_sub_epi16(*clip_diff, diff_clamp); in calc_mask_d16_inv_avx2() 542 _mm256_abs_epi16(_mm256_sub_epi16(s0, s1)), DIFF_FACTOR_LOG2); in av1_build_compound_diffwtd_mask_highbd_avx2() 546 m = _mm256_sub_epi16(yAOM_BLEND_A64_MAX_ALPHA, m); in av1_build_compound_diffwtd_mask_highbd_avx2() 562 _mm256_abs_epi16(_mm256_sub_epi16(s0, s1)), DIFF_FACTOR_LOG2); in av1_build_compound_diffwtd_mask_highbd_avx2() 584 _mm256_abs_epi16(_mm256_sub_epi16(s0, s1)), xshift); in av1_build_compound_diffwtd_mask_highbd_avx2() 588 m = _mm256_sub_epi16(yAOM_BLEND_A64_MAX_ALPHA, m); in av1_build_compound_diffwtd_mask_highbd_avx2() 604 _mm256_abs_epi16(_mm256_sub_epi16(s0, s1)), xshift); in av1_build_compound_diffwtd_mask_highbd_avx2()
|
D | cfl_avx2.c | 449 _mm256_sub_epi16(_mm256_loadu_si256(src), avg_epi16)); in subtract_average_avx2() 452 dst + 1, _mm256_sub_epi16(_mm256_loadu_si256(src + 1), avg_epi16)); in subtract_average_avx2()
|
D | warp_plane_avx2.c | 1056 _mm256_add_epi16(_mm256_sub_epi16(dst_1_16, ref_1_16), dup_255); in av1_calc_frame_error_avx2() 1058 _mm256_add_epi16(_mm256_sub_epi16(dst_2_16, ref_2_16), dup_255); in av1_calc_frame_error_avx2() 1060 _mm256_add_epi16(_mm256_sub_epi16(dst_3_16, ref_3_16), dup_255); in av1_calc_frame_error_avx2() 1062 _mm256_add_epi16(_mm256_sub_epi16(dst_4_16, ref_4_16), dup_255); in av1_calc_frame_error_avx2()
|
/external/libaom/libaom/aom_dsp/simd/ |
D | v256_intrinsics_x86.h | 119 SIMD_INLINE v256 v256_sub_16(v256 a, v256 b) { return _mm256_sub_epi16(a, b); } in v256_sub_16() 468 v256 l = _mm256_sub_epi16(_mm256_unpacklo_epi8(a, _mm256_setzero_si256()), in v256_ssd_u8() 470 v256 h = _mm256_sub_epi16(_mm256_unpackhi_epi8(a, _mm256_setzero_si256()), in v256_ssd_u8() 533 return _mm256_sub_epi16( in v256_rdavg_u16()
|
/external/skia/src/opts/ |
D | SkBlitRow_opts.h | 46 __m256i scale_x2 = _mm256_sub_epi16(_mm256_set1_epi16(256), in SkPMSrcOver_AVX2()
|