/external/libaom/libaom/av1/encoder/x86/ |
D | encodetxb_avx2.c | 45 const __m256i abs01 = _mm256_abs_epi16(_mm256_packs_epi32(c0, c1)); in av1_txb_init_levels_avx2() 60 const __m256i coeffAB = _mm256_packs_epi32(coeffA, coeffB); in av1_txb_init_levels_avx2() 61 const __m256i coeffCD = _mm256_packs_epi32(coeffC, coeffD); in av1_txb_init_levels_avx2() 87 const __m256i coeffAB = _mm256_packs_epi32(coeffA, coeffB); in av1_txb_init_levels_avx2() 88 const __m256i coeffCD = _mm256_packs_epi32(coeffC, coeffD); in av1_txb_init_levels_avx2() 108 const __m256i coeffAB = _mm256_packs_epi32(coeffA, coeffB); in av1_txb_init_levels_avx2() 109 const __m256i coeffCD = _mm256_packs_epi32(coeffC, coeffD); in av1_txb_init_levels_avx2()
|
D | wedge_utils_avx2.c | 59 const __m256i v_t0_w = _mm256_packs_epi32(v_t0l_d, v_t0h_d); in av1_wedge_sse_from_residuals_avx2() 200 const __m256i v_r0_w = _mm256_packs_epi32(v_r0l_w, v_r0h_w); in av1_wedge_compute_delta_squares_avx2() 201 const __m256i v_r1_w = _mm256_packs_epi32(v_r1l_w, v_r1h_w); in av1_wedge_compute_delta_squares_avx2() 202 const __m256i v_r2_w = _mm256_packs_epi32(v_r2l_w, v_r2h_w); in av1_wedge_compute_delta_squares_avx2() 203 const __m256i v_r3_w = _mm256_packs_epi32(v_r3l_w, v_r3h_w); in av1_wedge_compute_delta_squares_avx2()
|
D | pickrst_avx2.c | 520 _mm256_packs_epi32(yy_loadu_256(flt0 + j), in av1_lowbd_pixel_proj_error_avx2() 524 _mm256_packs_epi32(yy_loadu_256(flt1 + j), in av1_lowbd_pixel_proj_error_avx2() 539 _mm256_add_epi16(_mm256_packs_epi32(vr0, vr1), d0), s0); in av1_lowbd_pixel_proj_error_avx2() 572 _mm256_packs_epi32(yy_loadu_256(flt + j), in av1_lowbd_pixel_proj_error_avx2() 584 _mm256_add_epi16(_mm256_packs_epi32(vr0, vr1), d0), s0); in av1_lowbd_pixel_proj_error_avx2() 929 _mm256_permute4x64_epi64(_mm256_packs_epi32(vrl, vrh), 0xd8); in av1_highbd_pixel_proj_error_avx2() 1006 _mm256_permute4x64_epi64(_mm256_packs_epi32(vrl, vrh), 0xd8); in av1_highbd_pixel_proj_error_avx2()
|
D | error_intrin_avx2.c | 25 const __m256i y = _mm256_packs_epi32(x0, x1); in read_coeff()
|
/external/libvpx/libvpx/vpx_dsp/x86/ |
D | fwd_dct32x32_impl_avx2.h | 431 step2[20] = _mm256_packs_epi32(s2_20_6, s2_20_7); in FDCT32x32_2D_AVX2() 432 step2[21] = _mm256_packs_epi32(s2_21_6, s2_21_7); in FDCT32x32_2D_AVX2() 433 step2[22] = _mm256_packs_epi32(s2_22_6, s2_22_7); in FDCT32x32_2D_AVX2() 434 step2[23] = _mm256_packs_epi32(s2_23_6, s2_23_7); in FDCT32x32_2D_AVX2() 435 step2[24] = _mm256_packs_epi32(s2_24_6, s2_24_7); in FDCT32x32_2D_AVX2() 436 step2[25] = _mm256_packs_epi32(s2_25_6, s2_25_7); in FDCT32x32_2D_AVX2() 437 step2[26] = _mm256_packs_epi32(s2_26_6, s2_26_7); in FDCT32x32_2D_AVX2() 438 step2[27] = _mm256_packs_epi32(s2_27_6, s2_27_7); in FDCT32x32_2D_AVX2() 632 step3[10] = _mm256_packs_epi32(s3_10_6, s3_10_7); in FDCT32x32_2D_AVX2() 633 step3[11] = _mm256_packs_epi32(s3_11_6, s3_11_7); in FDCT32x32_2D_AVX2() [all …]
|
D | bitdepth_conversion_avx2.h | 25 return _mm256_packs_epi32(a_low, a_high); in load_tran_low()
|
/external/tensorflow/third_party/eigen3/unsupported/Eigen/CXX11/src/FixedPoint/ |
D | TypeCastingAVX512.h | 44 _mm256_packs_epi32(_mm512_castsi512_si256(a_int), 48 _mm256_packs_epi32(_mm512_extracti32x8_epi32(a_int, 1), 75 _mm256_packs_epi32(_mm512_castsi512_si256(a_int), 79 _mm256_packs_epi32(_mm512_castsi512_si256(c_int), 83 _mm256_packs_epi32(_mm512_extracti32x8_epi32(a_int, 1), 87 _mm256_packs_epi32(_mm512_extracti32x8_epi32(c_int, 1),
|
D | TypeCastingAVX2.h | 38 __m256i converted = _mm256_packs_epi16(_mm256_packs_epi32(a.m_val, b.m_val), 39 _mm256_packs_epi32(c.m_val, d.m_val)); 59 __m256i converted = _mm256_packs_epi16(_mm256_packs_epi32(a_conv, b_conv), 60 _mm256_packs_epi32(c_conv, d_conv));
|
/external/libaom/libaom/av1/common/x86/ |
D | highbd_convolve_2d_avx2.c | 98 __m256i res_even1 = _mm256_packs_epi32(res_even, res_even); in av1_highbd_convolve_2d_sr_avx2() 99 __m256i res_odd1 = _mm256_packs_epi32(res_odd, res_odd); in av1_highbd_convolve_2d_sr_avx2() 149 __m256i res_16bit = _mm256_packs_epi32(res_a_round, res_b_round); in av1_highbd_convolve_2d_sr_avx2() 158 res_a_round = _mm256_packs_epi32(res_a_round, res_a_round); in av1_highbd_convolve_2d_sr_avx2() 167 res_a_round = _mm256_packs_epi32(res_a_round, res_a_round); in av1_highbd_convolve_2d_sr_avx2()
|
D | highbd_wiener_convolve_avx2.c | 130 const __m256i res = _mm256_packs_epi32(res_even, res_odd); in av1_highbd_wiener_convolve_add_src_avx2() 236 _mm256_packs_epi32(res_lo_round, res_hi_round); in av1_highbd_wiener_convolve_add_src_avx2()
|
D | jnt_convolve_avx2.c | 289 _mm256_packs_epi32(res_lo_0_round, res_lo_1_round); in av1_dist_wtd_convolve_y_avx2() 347 _mm256_packs_epi32(res_hi_0_round, res_hi_1_round); in av1_dist_wtd_convolve_y_avx2() 468 _mm256_packs_epi32(res_lo_0_round, res_lo_1_round); in av1_dist_wtd_convolve_y_avx2() 526 _mm256_packs_epi32(res_hi_0_round, res_hi_1_round); in av1_dist_wtd_convolve_y_avx2() 717 const __m256i res_16b = _mm256_packs_epi32(res_a_round, res_b_round); in av1_dist_wtd_convolve_2d_avx2() 747 const __m256i res_16b = _mm256_packs_epi32(res_a_round, res_a_round); in av1_dist_wtd_convolve_2d_avx2()
|
/external/libaom/libaom/aom_dsp/x86/ |
D | bitdepth_conversion_avx2.h | 21 return _mm256_packs_epi32(a_low, a_high); in load_tran_low()
|
D | highbd_adaptive_quantize_avx2.c | 57 _mm256_permute4x64_epi64(_mm256_packs_epi32(cmp_mask0, cmp_mask1), 0xd8); in highbd_update_mask0_avx2() 154 cmp_mask = _mm256_permute4x64_epi64(_mm256_packs_epi32(temp0, temp1), 0xd8); in aom_highbd_quantize_b_adaptive_avx2() 195 cmp_mask = _mm256_permute4x64_epi64(_mm256_packs_epi32(temp0, temp1), 0xd8); in aom_highbd_quantize_b_adaptive_avx2() 336 cmp_mask = _mm256_permute4x64_epi64(_mm256_packs_epi32(temp0, temp1), 0xd8); in aom_highbd_quantize_b_32x32_adaptive_avx2() 379 cmp_mask = _mm256_permute4x64_epi64(_mm256_packs_epi32(temp0, temp1), 0xd8); in aom_highbd_quantize_b_32x32_adaptive_avx2()
|
D | highbd_convolve_avx2.c | 223 __m256i res_16bit = _mm256_packs_epi32(res_a_round, res_b_round); in av1_highbd_convolve_y_sr_avx2() 232 res_a_round = _mm256_packs_epi32(res_a_round, res_a_round); in av1_highbd_convolve_y_sr_avx2() 241 res_a_round = _mm256_packs_epi32(res_a_round, res_a_round); in av1_highbd_convolve_y_sr_avx2() 334 __m256i res_even1 = _mm256_packs_epi32(res_even, res_even); in av1_highbd_convolve_x_sr_avx2() 335 __m256i res_odd1 = _mm256_packs_epi32(res_odd, res_odd); in av1_highbd_convolve_x_sr_avx2() 602 res = _mm256_packs_epi32(res, res); in aom_highbd_filter_block1d4_h4_avx2() 628 res = _mm256_packs_epi32(res, res); in aom_highbd_filter_block1d4_h4_avx2() 687 __m256i res = _mm256_packs_epi32(res_even, res_odd); in aom_highbd_filter_block1d8_h4_avx2() 716 res = _mm256_packs_epi32(res, res); in aom_highbd_filter_block1d8_h4_avx2() 1114 res_16bit = _mm256_packs_epi32(res_16bit, res_16bit); in aom_highbd_filter_block1d4_v4_avx2() [all …]
|
D | convolve_avx2.h | 126 const __m256i res_16bit = _mm256_packs_epi32(res_a_round, res_b_round); \ 203 const __m256i res_16b = _mm256_packs_epi32(res_a_round, res_b_round); \ 231 const __m256i res_16b = _mm256_packs_epi32(res_a_round, res_a_round); \ 417 res = _mm256_packs_epi32(res_lo, res_hi); in comp_avg()
|
D | txfm_common_avx2.h | 48 *in0 = _mm256_packs_epi32(c0, c1); in btf_16_w16_avx2() 49 *in1 = _mm256_packs_epi32(d0, d1); in btf_16_w16_avx2() 105 const __m256i b = _mm256_packs_epi32(a_low, *(const __m256i *)(a + 8)); in load_32bit_to_16bit_w16_avx2()
|
D | adaptive_quantize_avx2.c | 42 return _mm256_packs_epi32(coeff1, coeff2); in load_coefficients_avx2() 69 _mm256_permute4x64_epi64(_mm256_packs_epi32(cmp_mask0, cmp_mask1), 0xd8); in update_mask0_avx2()
|
/external/XNNPACK/src/qs8-vaddc/gen/ |
D | minmax-avx2-mul32-ld64-x32.c | 58 …__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), v… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x32() 59 …__m256i voutGHIJOPQRKLMNSTUV = _mm256_adds_epi16(_mm256_packs_epi32(vaccGHIJKLMN, vaccOPQRSTUV), v… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x32()
|
D | minmax-avx2-mul32-ld64-x16.c | 50 …__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), v… in xnn_qs8_vaddc_minmax_ukernel__avx2_mul32_ld64_x16()
|
/external/XNNPACK/src/qs8-vadd/gen/ |
D | minmax-avx2-mul32-ld64-x32.c | 67 …__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), v… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32() 68 …__m256i voutGHIJOPQRKLMNSTUV = _mm256_adds_epi16(_mm256_packs_epi32(vaccGHIJKLMN, vaccOPQRSTUV), v… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x32()
|
D | minmax-avx2-mul32-ld64-x16.c | 55 …__m256i vout012389AB4567CDEF = _mm256_adds_epi16(_mm256_packs_epi32(vacc01234567, vacc89ABCDEF), v… in xnn_qs8_vadd_minmax_ukernel__avx2_mul32_ld64_x16()
|
/external/XNNPACK/src/qs8-dwconv/ |
D | unipass-avx512skx-mul32.c.in | 109 …ABC[C+8:C+12]}${ABC[C+4:C+8]}${ABC[C+12:C+16]} = _mm256_adds_epi16(_mm256_packs_epi32(_mm512_casts… 111 …ABC[C+8:C+12]}${ABC[C+4:C+8]}${ABC[C+12:C+16]} = _mm256_adds_epi16(_mm256_packs_epi32(_mm512_casts… 186 …{ABC[0:4]}${ABC[8:12]}${ABC[4:8]}${ABC[12:16]} = _mm256_adds_epi16(_mm256_packs_epi32(_mm512_casts… 188 …{ABC[0:4]}${ABC[8:12]}${ABC[4:8]}${ABC[12:16]} = _mm256_adds_epi16(_mm256_packs_epi32(_mm512_casts…
|
/external/XNNPACK/src/qs8-igemm/gen/ |
D | 3x8c8-minmax-avx2.c | 197 …__m256i vacc01x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc1x01234567), vo… in xnn_qs8_igemm_minmax_ukernel_3x8c8__avx2() 198 …__m256i vacc22x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc2x01234567, vacc2x01234567), vo… in xnn_qs8_igemm_minmax_ukernel_3x8c8__avx2()
|
/external/XNNPACK/src/qs8-gemm/gen/ |
D | 3x8c8-minmax-avx2.c | 180 …__m256i vacc01x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc1x01234567), vo… in xnn_qs8_gemm_minmax_ukernel_3x8c8__avx2() 181 …__m256i vacc22x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc2x01234567, vacc2x01234567), vo… in xnn_qs8_gemm_minmax_ukernel_3x8c8__avx2()
|
D | 3x8c8-xw-minmax-avx2.c | 176 …__m256i vacc01x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc0x01234567, vacc1x01234567), vo… in xnn_qs8_gemm_xw_minmax_ukernel_3x8c8__avx2() 177 …__m256i vacc22x01234567 = _mm256_adds_epi16(_mm256_packs_epi32(vacc2x01234567, vacc2x01234567), vo… in xnn_qs8_gemm_xw_minmax_ukernel_3x8c8__avx2()
|