/external/libaom/libaom/av1/common/arm/ |
D | wiener_convolve_neon.c | 118 res0 = vreinterpretq_s16_u16(vaddl_u8(t0, t6)); in av1_wiener_convolve_add_src_neon() 119 res1 = vreinterpretq_s16_u16(vaddl_u8(t1, t5)); in av1_wiener_convolve_add_src_neon() 120 res2 = vreinterpretq_s16_u16(vaddl_u8(t2, t4)); in av1_wiener_convolve_add_src_neon() 125 res0 = vreinterpretq_s16_u16(vaddl_u8(t1, t7)); in av1_wiener_convolve_add_src_neon() 126 res1 = vreinterpretq_s16_u16(vaddl_u8(t2, t6)); in av1_wiener_convolve_add_src_neon() 127 res2 = vreinterpretq_s16_u16(vaddl_u8(t3, t5)); in av1_wiener_convolve_add_src_neon() 132 res0 = vreinterpretq_s16_u16(vaddl_u8(t2, t8)); in av1_wiener_convolve_add_src_neon() 133 res1 = vreinterpretq_s16_u16(vaddl_u8(t3, t7)); in av1_wiener_convolve_add_src_neon() 134 res2 = vreinterpretq_s16_u16(vaddl_u8(t4, t6)); in av1_wiener_convolve_add_src_neon() 139 res0 = vreinterpretq_s16_u16(vaddl_u8(t3, t9)); in av1_wiener_convolve_add_src_neon() [all …]
|
D | cfl_neon.c | 65 const uint16x8_t top_0 = vaddl_u8(top.val[0], top.val[1]); in cfl_luma_subsampling_420_lbd_neon() 67 const uint16x8_t bot_0 = vaddl_u8(bot.val[0], bot.val[1]); in cfl_luma_subsampling_420_lbd_neon() 69 const uint16x8_t top_1 = vaddl_u8(top.val[2], top.val[3]); in cfl_luma_subsampling_420_lbd_neon() 71 const uint16x8_t bot_1 = vaddl_u8(bot.val[2], bot.val[3]); in cfl_luma_subsampling_420_lbd_neon() 100 sum.val[0] = vshlq_n_u16(vaddl_u8(top.val[0], top.val[1]), 2); in cfl_luma_subsampling_422_lbd_neon() 101 sum.val[1] = vshlq_n_u16(vaddl_u8(top.val[2], top.val[3]), 2); in cfl_luma_subsampling_422_lbd_neon()
|
/external/libaom/libaom/aom_dsp/arm/ |
D | blend_a64_mask_neon.c | 234 vreinterpretq_s16_u16(vaddl_u8(vget_low_u8(t0), vget_low_u8(t1))); in aom_lowbd_blend_a64_d16_mask_neon() 236 vreinterpretq_s16_u16(vaddl_u8(vget_low_u8(t2), vget_low_u8(t3))); in aom_lowbd_blend_a64_d16_mask_neon() 238 vreinterpretq_s16_u16(vaddl_u8(vget_low_u8(t4), vget_low_u8(t5))); in aom_lowbd_blend_a64_d16_mask_neon() 240 vreinterpretq_s16_u16(vaddl_u8(vget_low_u8(t6), vget_low_u8(t7))); in aom_lowbd_blend_a64_d16_mask_neon() 243 vaddl_u8(vget_high_u8(t0), vget_high_u8(t1))); in aom_lowbd_blend_a64_d16_mask_neon() 245 vaddl_u8(vget_high_u8(t2), vget_high_u8(t3))); in aom_lowbd_blend_a64_d16_mask_neon() 247 vaddl_u8(vget_high_u8(t4), vget_high_u8(t5))); in aom_lowbd_blend_a64_d16_mask_neon() 249 vaddl_u8(vget_high_u8(t6), vget_high_u8(t7))); in aom_lowbd_blend_a64_d16_mask_neon() 291 mask0 = vreinterpretq_s16_u16(vaddl_u8(mask0_l, mask1_l)); in aom_lowbd_blend_a64_d16_mask_neon() 292 mask1 = vreinterpretq_s16_u16(vaddl_u8(mask2_l, mask3_l)); in aom_lowbd_blend_a64_d16_mask_neon() [all …]
|
D | avg_neon.c | 21 const uint16x8_t c = vaddl_u8(vget_low_u8(b), vget_high_u8(b)); in aom_avg_4x4_neon() 38 sum = vaddl_u8(b, c); in aom_avg_8x8_neon()
|
D | loopfilter_neon.c | 231 out = vaddl_u8(*p3q3, *p2q2); in lpf_14_neon() 275 p6q6_2 = vaddl_u8(*p6q6, *p6q6); in lpf_14_neon() 286 qp_sum = vaddl_u8(q2p2, q1p1); in lpf_14_neon() 410 out = vaddl_u8(*p3q3, *p2q2); in lpf_8_neon() 524 out = vaddl_u8(*p0q0, *p1q1); in lpf_6_neon()
|
/external/XNNPACK/src/qu8-avgpool/ |
D | 9p8x-minmax-neon-c8.c | 105 const uint16x8_t vsum018 = vaddw_u8(vaddl_u8(vi0, vi1), vi8); in xnn_qu8_avgpool_minmax_ukernel_9p8x__neon_c8() 106 const uint16x8_t vsum23 = vaddl_u8(vi2, vi3); in xnn_qu8_avgpool_minmax_ukernel_9p8x__neon_c8() 107 const uint16x8_t vsum45 = vaddl_u8(vi4, vi5); in xnn_qu8_avgpool_minmax_ukernel_9p8x__neon_c8() 108 const uint16x8_t vsum67 = vaddl_u8(vi6, vi7); in xnn_qu8_avgpool_minmax_ukernel_9p8x__neon_c8() 178 const uint16x8_t vsum01 = vaddl_u8(vi0, vi1); in xnn_qu8_avgpool_minmax_ukernel_9p8x__neon_c8() 179 const uint16x8_t vsum23 = vaddl_u8(vi2, vi3); in xnn_qu8_avgpool_minmax_ukernel_9p8x__neon_c8() 180 const uint16x8_t vsum45 = vaddl_u8(vi4, vi5); in xnn_qu8_avgpool_minmax_ukernel_9p8x__neon_c8() 181 const uint16x8_t vsum67 = vaddl_u8(vi6, vi7); in xnn_qu8_avgpool_minmax_ukernel_9p8x__neon_c8() 273 const int16x8_t vsum01 = vreinterpretq_s16_u16(vaddl_u8(vi0, vi1)); in xnn_qu8_avgpool_minmax_ukernel_9p8x__neon_c8() 274 const int16x8_t vsum23 = vreinterpretq_s16_u16(vaddl_u8(vi2, vi3)); in xnn_qu8_avgpool_minmax_ukernel_9p8x__neon_c8() [all …]
|
D | 9x-minmax-neon-c8.c | 129 const uint16x8_t vsum018 = vaddw_u8(vaddl_u8(vi0, vi1), vi8); in xnn_qu8_avgpool_minmax_ukernel_9x__neon_c8() 130 const uint16x8_t vsum23 = vaddl_u8(vi2, vi3); in xnn_qu8_avgpool_minmax_ukernel_9x__neon_c8() 131 const uint16x8_t vsum45 = vaddl_u8(vi4, vi5); in xnn_qu8_avgpool_minmax_ukernel_9x__neon_c8() 132 const uint16x8_t vsum67 = vaddl_u8(vi6, vi7); in xnn_qu8_avgpool_minmax_ukernel_9x__neon_c8() 202 const uint16x8_t vsum018 = vaddw_u8(vaddl_u8(vi0, vi1), vi8); in xnn_qu8_avgpool_minmax_ukernel_9x__neon_c8() 203 const uint16x8_t vsum23 = vaddl_u8(vi2, vi3); in xnn_qu8_avgpool_minmax_ukernel_9x__neon_c8() 204 const uint16x8_t vsum45 = vaddl_u8(vi4, vi5); in xnn_qu8_avgpool_minmax_ukernel_9x__neon_c8() 205 const uint16x8_t vsum67 = vaddl_u8(vi6, vi7); in xnn_qu8_avgpool_minmax_ukernel_9x__neon_c8()
|
/external/libgav1/libgav1/src/dsp/arm/ |
D | intra_edge_neon.cc | 78 uint16x8_t sum_lo = vaddl_u8(vget_low_u8(src_0), vget_low_u8(src_2)); in IntraEdgeFilter_NEON() 81 uint16x8_t sum_hi = vaddl_u8(vget_high_u8(src_0), vget_high_u8(src_2)); in IntraEdgeFilter_NEON() 102 uint16x8_t sum_lo = vaddl_u8(vget_low_u8(src_0), vget_low_u8(src_2)); in IntraEdgeFilter_NEON() 105 uint16x8_t sum_hi = vaddl_u8(vget_high_u8(src_0), vget_high_u8(src_2)); in IntraEdgeFilter_NEON() 159 vshlq_n_u16(vaddl_u8(vget_low_u8(src_0), vget_low_u8(src_4)), 1); in IntraEdgeFilter_NEON() 161 vaddl_u8(vget_low_u8(src_1), vget_low_u8(src_2)), vget_low_u8(src_3)); in IntraEdgeFilter_NEON() 165 vshlq_n_u16(vaddl_u8(vget_high_u8(src_0), vget_high_u8(src_4)), 1); in IntraEdgeFilter_NEON() 167 vaddw_u8(vaddl_u8(vget_high_u8(src_1), vget_high_u8(src_2)), in IntraEdgeFilter_NEON() 189 vshlq_n_u16(vaddl_u8(vget_low_u8(src_0), vget_low_u8(src_4)), 1); in IntraEdgeFilter_NEON() 191 vaddl_u8(vget_low_u8(src_1), vget_low_u8(src_2)), vget_low_u8(src_3)); in IntraEdgeFilter_NEON() [all …]
|
D | loop_filter_neon.cc | 293 const uint16x8_t p2q2_double = vaddl_u8(p2q2, p2q2); in Filter6() 300 sum = vaddq_u16(vaddl_u8(p1q1, p1q1), sum); in Filter6() 306 sum = vaddq_u16(vaddl_u8(p0q0, p0q0), sum); in Filter6() 322 sum = vaddq_u16(vaddl_u8(q0p0, q1p1), sum); in Filter6() 530 uint16x8_t sum = vaddw_u8(vaddl_u8(p3q3, p3q3), p3q3); in Filter8() 536 sum = vaddq_u16(vaddl_u8(p2q2, p2q2), sum); in Filter8() 542 sum = vaddq_u16(vaddl_u8(p1q1, p0q0), sum); in Filter8() 556 sum = vsubq_u16(sum, vaddl_u8(p3q3, p2q2)); in Filter8() 558 sum = vaddq_u16(vaddl_u8(p1q1, q1p1), sum); in Filter8() 565 sum = vsubq_u16(sum, vaddl_u8(p3q3, p1q1)); in Filter8() [all …]
|
D | convolve_neon.cc | 2843 row[0] = vaddl_u8(vld1_u8(src), vld1_u8(src + 1)); in IntraBlockCopy2D() 2846 row[1] = vaddl_u8(vld1_u8(src), vld1_u8(src + 1)); in IntraBlockCopy2D() 2849 row[2] = vaddl_u8(vld1_u8(src), vld1_u8(src + 1)); in IntraBlockCopy2D() 2851 row[3] = vaddl_u8(vld1_u8(src), vld1_u8(src + 1)); in IntraBlockCopy2D() 2854 row[4] = vaddl_u8(vld1_u8(src), vld1_u8(src + 1)); in IntraBlockCopy2D() 2856 row[5] = vaddl_u8(vld1_u8(src), vld1_u8(src + 1)); in IntraBlockCopy2D() 2858 row[6] = vaddl_u8(vld1_u8(src), vld1_u8(src + 1)); in IntraBlockCopy2D() 2860 row[7] = vaddl_u8(vld1_u8(src), vld1_u8(src + 1)); in IntraBlockCopy2D() 2863 row[8] = vaddl_u8(vld1_u8(src), vld1_u8(src + 1)); in IntraBlockCopy2D() 2865 row[9] = vaddl_u8(vld1_u8(src), vld1_u8(src + 1)); in IntraBlockCopy2D() [all …]
|
D | cdef_neon.cc | 93 *partial_lo = vaddl_u8(v_src[0], vext_u8(v_zero, v_src[1], 7)); in AddPartial_D0_D4() 100 vaddl_u8(vext_u8(v_src[1], v_zero, 7), vext_u8(v_src[2], v_zero, 6)); in AddPartial_D0_D4() 203 v_pair_add[0] = vaddl_u8(v_src[0], v_src[1]); in AddPartial_D5_D7() 204 v_pair_add[1] = vaddl_u8(v_src[2], v_src[3]); in AddPartial_D5_D7() 205 v_pair_add[2] = vaddl_u8(v_src[4], v_src[5]); in AddPartial_D5_D7() 206 v_pair_add[3] = vaddl_u8(v_src[6], v_src[7]); in AddPartial_D5_D7() 288 partial_lo[6] = vaddl_u8(v_src[0], v_src[1]); in AddPartial()
|
D | loop_restoration_neon.cc | 83 const int16x8_t ss = vreinterpretq_s16_u16(vaddl_u8(s0, s1)); in WienerHorizontal2() 103 const int16x8_t s_0_2 = vreinterpretq_s16_u16(vaddl_u8(s[0], s[2])); in WienerHorizontalSum() 673 const uint16x8_t sum = vaddl_u8(src[0], src[1]); in Sum3W_16() 678 const uint16x8_t sum = vaddl_u8(vget_low_u8(src[0]), vget_low_u8(src[1])); in Sum3WLo16() 683 const uint16x8_t sum = vaddl_u8(vget_high_u8(src[0]), vget_high_u8(src[1])); in Sum3WHi16() 688 const uint16x8_t sum01 = vaddl_u8(vget_low_u8(src[0]), vget_low_u8(src[1])); in Sum5WLo16() 689 const uint16x8_t sum23 = vaddl_u8(vget_low_u8(src[2]), vget_low_u8(src[3])); in Sum5WLo16() 695 const uint16x8_t sum01 = vaddl_u8(vget_high_u8(src[0]), vget_high_u8(src[1])); in Sum5WHi16() 696 const uint16x8_t sum23 = vaddl_u8(vget_high_u8(src[2]), vget_high_u8(src[3])); in Sum5WHi16() 771 const uint16x8_t sum01 = vaddl_u8(s[0], s[1]); in Sum5Horizontal() [all …]
|
D | intrapred_neon.cc | 343 vabdq_u16(vaddl_u8(top, left), top_left_x2); in Paeth4Or8xN_NEON() 412 vaddl_u8(vget_low_u8(top[num]), vget_low_u8(left)), top_left_x2); \ 414 vaddl_u8(vget_high_u8(top[num]), vget_low_u8(left)), top_left_x2)
|
/external/XNNPACK/src/qu8-gavgpool/ |
D | 7p7x-minmax-neon-c8.c | 52 const uint16x8_t vsum01 = vaddl_u8(vi0, vi1); in xnn_qu8_gavgpool_minmax_ukernel_7p7x__neon_c8() 53 const uint16x8_t vsum23 = vaddl_u8(vi2, vi3); in xnn_qu8_gavgpool_minmax_ukernel_7p7x__neon_c8() 54 const uint16x8_t vsum45 = vaddl_u8(vi4, vi5); in xnn_qu8_gavgpool_minmax_ukernel_7p7x__neon_c8() 89 const uint16x8_t vsum01 = vaddl_u8(vi0, vi1); in xnn_qu8_gavgpool_minmax_ukernel_7p7x__neon_c8() 90 const uint16x8_t vsum23 = vaddl_u8(vi2, vi3); in xnn_qu8_gavgpool_minmax_ukernel_7p7x__neon_c8() 91 const uint16x8_t vsum45 = vaddl_u8(vi4, vi5); in xnn_qu8_gavgpool_minmax_ukernel_7p7x__neon_c8() 151 const uint16x8_t vsum01 = vaddl_u8(vi0, vi1); in xnn_qu8_gavgpool_minmax_ukernel_7p7x__neon_c8() 152 const uint16x8_t vsum23 = vaddl_u8(vi2, vi3); in xnn_qu8_gavgpool_minmax_ukernel_7p7x__neon_c8() 153 const uint16x8_t vsum45 = vaddl_u8(vi4, vi5); in xnn_qu8_gavgpool_minmax_ukernel_7p7x__neon_c8() 223 const uint16x8_t vsum01 = vaddl_u8(vi0, vi1); in xnn_qu8_gavgpool_minmax_ukernel_7p7x__neon_c8() [all …]
|
D | 7x-minmax-neon-c8.c | 75 const uint16x8_t vsum01 = vaddl_u8(vi0, vi1); in xnn_qu8_gavgpool_minmax_ukernel_7x__neon_c8() 76 const uint16x8_t vsum23 = vaddl_u8(vi2, vi3); in xnn_qu8_gavgpool_minmax_ukernel_7x__neon_c8() 77 const uint16x8_t vsum45 = vaddl_u8(vi4, vi5); in xnn_qu8_gavgpool_minmax_ukernel_7x__neon_c8() 145 const uint16x8_t vsum01 = vaddl_u8(vi0, vi1); in xnn_qu8_gavgpool_minmax_ukernel_7x__neon_c8() 146 const uint16x8_t vsum23 = vaddl_u8(vi2, vi3); in xnn_qu8_gavgpool_minmax_ukernel_7x__neon_c8() 147 const uint16x8_t vsum45 = vaddl_u8(vi4, vi5); in xnn_qu8_gavgpool_minmax_ukernel_7x__neon_c8()
|
/external/libhevc/encoder/arm/ |
D | ihevce_scale_by_2_neon.c | 90 q = vreinterpretq_s16_u16(vaddl_u8(l0, r0)); in ihevce_horz_scale_neon_w16() 92 r = vreinterpretq_s16_u16(vaddl_u8(r3, vget_low_u8(src.val[0]))); in ihevce_horz_scale_neon_w16() 164 q = vreinterpretq_s16_u16(vaddl_u8(vget_low_u8(src[t1]), vget_low_u8(src[b1]))); in ihevce_vert_scale_neon_w16() 167 r = vreinterpretq_s16_u16(vaddl_u8(vget_low_u8(src[t2]), vget_low_u8(src[b2]))); in ihevce_vert_scale_neon_w16() 182 q = vreinterpretq_s16_u16(vaddl_u8(vget_high_u8(src[t1]), vget_high_u8(src[b1]))); in ihevce_vert_scale_neon_w16() 185 r = vreinterpretq_s16_u16(vaddl_u8(vget_high_u8(src[t2]), vget_high_u8(src[b2]))); in ihevce_vert_scale_neon_w16()
|
D | ihevce_ssd_and_sad_calculator_neon.c | 78 abs_sum = vaddl_u8(abs_l, abs_h); in ihevce_ssd_and_sad_calculator_neon() 193 abs_sum_l = vaddl_u8(abs_l, abs_h); in ihevce_ssd_and_sad_calculator_neon() 204 abs_sum_h = vaddl_u8(abs_l, abs_h); in ihevce_ssd_and_sad_calculator_neon() 256 abs_sum_0 = vaddl_u8(abs_l, abs_h); in ihevce_ssd_and_sad_calculator_neon() 267 abs_sum_1 = vaddl_u8(abs_l, abs_h); in ihevce_ssd_and_sad_calculator_neon() 278 abs_sum_2 = vaddl_u8(abs_l, abs_h); in ihevce_ssd_and_sad_calculator_neon() 289 abs_sum_3 = vaddl_u8(abs_l, abs_h); in ihevce_ssd_and_sad_calculator_neon()
|
/external/libvpx/libvpx/vpx_dsp/arm/ |
D | avg_neon.c | 24 const uint16x8_t c = vaddl_u8(vget_low_u8(b), vget_high_u8(b)); in vpx_avg_4x4_neon() 38 sum = vaddl_u8(b, c); in vpx_avg_8x8_neon()
|
D | deblock_neon.c | 354 y = vreinterpretq_s16_u16(vaddl_u8(right_context, left_context)); in vpx_mbpost_proc_across_ip_neon() 451 y = vreinterpretq_s16_u16(vaddl_u8(below_context, above_context[0])); in vpx_mbpost_proc_down_neon()
|
/external/webp/src/dsp/ |
D | upsampling_neon.c | 37 const uint16x8_t ad = vaddl_u8(a, d); \ 38 const uint16x8_t bc = vaddl_u8(b, c); \
|
D | dec_neon.c | 1290 const uint16x8_t s0 = vaddl_u8(L0, L1); in DC4_NEON() 1291 const uint16x8_t s1 = vaddl_u8(L2, L3); in DC4_NEON() 1451 const uint16x8_t s0 = vaddl_u8(L0, L1); in DC8_NEON() 1452 const uint16x8_t s1 = vaddl_u8(L2, L3); in DC8_NEON() 1453 const uint16x8_t s2 = vaddl_u8(L4, L5); in DC8_NEON() 1454 const uint16x8_t s3 = vaddl_u8(L6, L7); in DC8_NEON() 1538 const uint16x8_t s0 = vaddl_u8(L0, L1); in DC16_NEON() 1539 const uint16x8_t s1 = vaddl_u8(L2, L3); in DC16_NEON() 1540 const uint16x8_t s2 = vaddl_u8(L4, L5); in DC16_NEON() 1541 const uint16x8_t s3 = vaddl_u8(L6, L7); in DC16_NEON()
|
D | filters_neon.c | 157 const int16x8_t C = vreinterpretq_s16_u16(vaddl_u8(A, B)); in GradientPredictDirect_NEON()
|
D | enc_neon.c | 603 const int16x8_t q_a0 = vreinterpretq_s16_u16(vaddl_u8(q4_in.val[0], in DistoVerticalPass_NEON() 605 const int16x8_t q_a1 = vreinterpretq_s16_u16(vaddl_u8(q4_in.val[1], in DistoVerticalPass_NEON()
|
/external/libpng/arm/ |
D | filter_neon_intrinsics.c | 266 p1 = vaddl_u8(a, b); /* a + b */ in paeth() 267 pc = vaddl_u8(c, c); /* c * 2 */ in paeth()
|
/external/pdfium/third_party/libpng16/arm/ |
D | filter_neon_intrinsics.c | 266 p1 = vaddl_u8(a, b); /* a + b */ in paeth() 267 pc = vaddl_u8(c, c); /* c * 2 */ in paeth()
|