Home
last modified time | relevance | path

Searched refs:_mm256_sub_epi16 (Results 1 – 25 of 26) sorted by relevance

12

/external/libaom/libaom/aom_dsp/x86/
Davg_intrin_avx2.c30 __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 …]
Dsad_highbd_avx2.c56 __m256i u0 = _mm256_sub_epi16(s0, r0); in aom_highbd_sad16x8_avx2()
57 __m256i u1 = _mm256_sub_epi16(s1, r1); in aom_highbd_sad16x8_avx2()
58 __m256i u2 = _mm256_sub_epi16(s2, r2); in aom_highbd_sad16x8_avx2()
59 __m256i u3 = _mm256_sub_epi16(s3, r3); in aom_highbd_sad16x8_avx2()
85 u0 = _mm256_sub_epi16(s0, r0); in aom_highbd_sad16x8_avx2()
86 u1 = _mm256_sub_epi16(s1, r1); in aom_highbd_sad16x8_avx2()
87 u2 = _mm256_sub_epi16(s2, r2); in aom_highbd_sad16x8_avx2()
88 u3 = _mm256_sub_epi16(s3, r3); in aom_highbd_sad16x8_avx2()
134 u0 = _mm256_sub_epi16(s0, r0); in aom_highbd_sad16x16_avx2()
135 u1 = _mm256_sub_epi16(s1, r1); in aom_highbd_sad16x16_avx2()
[all …]
Dsse_avx2.c29 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()
83 const __m256i v_d_w = _mm256_sub_epi16(v_a_w, v_b_w); in sse_w4x4_avx2()
94 const __m256i v_d_w = _mm256_sub_epi16(v_a_w, v_b_w); in sse_w8x2_avx2()
136 const __m256i v_asub = _mm256_sub_epi16(v_al, v_bl); in aom_sse_avx2()
137 const __m256i v_bsub = _mm256_sub_epi16(v_au, v_bu); in aom_sse_avx2()
218 const __m256i v_d_w = _mm256_sub_epi16(v_a_w, v_b_w); in highbd_sse_w16_avx2()
237 const __m256i v_d_w = _mm256_sub_epi16(v_a_w, v_b_w); in highbd_sse_w4x4_avx2()
246 const __m256i v_d_w = _mm256_sub_epi16(v_a_w, v_b_w); in highbd_sse_w8x2_avx2()
Dsubtract_avx2.c23 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 aom_subtract_block_16xn_avx2()
Dmasked_sad_intrin_avx2.c224 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()
Dhighbd_variance_avx2.c35 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()
Dblend_a64_mask_avx2.c32 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()
921 const __m256i mask1 = _mm256_sub_epi16(*mask_max, *mask0); in highbd_blend_a64_d16_mask_w4_avx2()
1038 const __m256i mask1a = _mm256_sub_epi16(*mask_max, *mask0a); in highbd_blend_a64_d16_mask_w8_avx2()
1039 const __m256i mask1b = _mm256_sub_epi16(*mask_max, *mask0b); in highbd_blend_a64_d16_mask_w8_avx2()
1174 const __m256i mask1a = _mm256_sub_epi16(*mask_max, *mask0a); in highbd_blend_a64_d16_mask_w16_avx2()
1175 const __m256i mask1b = _mm256_sub_epi16(*mask_max, *mask0b); in highbd_blend_a64_d16_mask_w16_avx2()
Dvariance_impl_avx2.c75 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); \
Dintrapred_avx2.c668 _mm256_sub_epi16(_mm256_add_epi16(*top, *left), *topleft); in paeth_pred()
670 __m256i pl = _mm256_abs_epi16(_mm256_sub_epi16(base, *left)); in paeth_pred()
671 __m256i pt = _mm256_abs_epi16(_mm256_sub_epi16(base, *top)); in paeth_pred()
672 __m256i ptl = _mm256_abs_epi16(_mm256_sub_epi16(base, *topleft)); in paeth_pred()
2069 diff = _mm256_sub_epi16(a1_x, a0_x); // a[x+1] - a[x] in highbd_dr_prediction_z2_Nx8_avx2()
2391 diff = _mm256_sub_epi16(a1_x, a0_x); // a[x+1] - a[x] in highbd_dr_prediction_z2_HxW_avx2()
2411 y_c256 = _mm256_sub_epi16(r6, mul16); in highbd_dr_prediction_z2_HxW_avx2()
2435 diff = _mm256_sub_epi16(a1_y, a0_y); // a[x+1] - a[x] in highbd_dr_prediction_z2_HxW_avx2()
3074 diff = _mm256_sub_epi16(a1, a0); // a[x+1] - a[x] in dr_prediction_z1_4xN_internal_avx2()
3801 diff = _mm256_sub_epi16(a1_x, a0_x); // a[x+1] - a[x] in dr_prediction_z2_Nx8_avx2()
[all …]
Dconvolve_avx2.h404 const __m256i res_signed = _mm256_sub_epi16(*res_unsigned, *offset_const); in convolve_rounding()
Dvariance_avx2.c417 const __m256i a_inv = _mm256_sub_epi16(alpha_max, a); in highbd_comp_mask_pred_line_avx2()
/external/libvpx/libvpx/vpx_dsp/x86/
Davg_intrin_avx2.c232 __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 …]
Dfwd_dct32x32_impl_avx2.h135 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 …]
Dloopfilter_avx2.c655 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 …]
Dvariance_avx2.c193 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/
Dvp9_error_avx2.c35 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()
Dvp9_quantize_avx2.c45 const __m256i iscan_plus_one = _mm256_sub_epi16(iscan, nzero_coeff0); in scan_eob_256()
/external/libaom/libaom/av1/common/x86/
Dreconinter_avx2.c24 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()
Dcfl_avx2.c445 _mm256_sub_epi16(_mm256_loadu_si256(src), avg_epi16)); in subtract_average_avx2()
448 dst + 1, _mm256_sub_epi16(_mm256_loadu_si256(src + 1), avg_epi16)); in subtract_average_avx2()
/external/libaom/libaom/av1/encoder/x86/
Dpickrst_avx2.c520 const __m256i flt0_0_sub_u = _mm256_sub_epi16(flt0_16b, u0); in av1_lowbd_pixel_proj_error_avx2()
521 const __m256i flt1_0_sub_u = _mm256_sub_epi16(flt1_16b, u0); in av1_lowbd_pixel_proj_error_avx2()
530 const __m256i e0 = _mm256_sub_epi16( in av1_lowbd_pixel_proj_error_avx2()
575 const __m256i e0 = _mm256_sub_epi16( in av1_lowbd_pixel_proj_error_avx2()
602 const __m256i diff0 = _mm256_sub_epi16(d0, s0); in av1_lowbd_pixel_proj_error_avx2()
694 const __m256i e0 = _mm256_sub_epi16(_mm256_add_epi16(vr, d0), s0); in av1_highbd_pixel_proj_error_avx2()
771 const __m256i e0 = _mm256_sub_epi16(_mm256_add_epi16(vr, d0), s0); in av1_highbd_pixel_proj_error_avx2()
810 const __m256i diffl = _mm256_sub_epi16(d0l, s0l); in av1_highbd_pixel_proj_error_avx2()
811 const __m256i diffh = _mm256_sub_epi16(d0h, s0h); in av1_highbd_pixel_proj_error_avx2()
Dav1_quantize_avx2.c126 __m256i cur_eob = _mm256_sub_epi16(iscan, nzero_coeff); in quantize()
202 __m256i cur_eob = _mm256_sub_epi16(iscan, nzero_coeff); in quantize_32x32()
281 __m256i cur_eob = _mm256_sub_epi16(iscan, nzero_coeff); in quantize_64x64()
Derror_intrin_avx2.c51 dqcoeff_reg = _mm256_sub_epi16(dqcoeff_reg, coeff_reg); in av1_block_error_avx2()
/external/libaom/libaom/aom_dsp/simd/
Dv256_intrinsics_x86.h119 SIMD_INLINE v256 v256_sub_16(v256 a, v256 b) { return _mm256_sub_epi16(a, b); } in v256_sub_16()
470 v256 l = _mm256_sub_epi16(_mm256_unpacklo_epi8(a, _mm256_setzero_si256()), in v256_ssd_u8()
472 v256 h = _mm256_sub_epi16(_mm256_unpackhi_epi8(a, _mm256_setzero_si256()), in v256_ssd_u8()
535 return _mm256_sub_epi16( in v256_rdavg_u16()
/external/clang/test/CodeGen/
Davx2-builtins.c1135 return _mm256_sub_epi16(a, b); in test_mm256_sub_epi16()
/external/clang/lib/Headers/
Davx2intrin.h737 _mm256_sub_epi16(__m256i __a, __m256i __b) in _mm256_sub_epi16() function

12