/external/libaom/aom_dsp/arm/ |
D | highbd_intrapred_neon.c | 53 sum_q = vaddl_u16(vld1_u16(above), vld1_u16(left)); in highbd_dc_predictor() 111 const uint16x4_t row = vld1_u16(above); in vertical4xh_neon() 226 top = vcombine_u16(vld1_u16(top_row), vdup_n_u16(0)); in highbd_paeth_4or8_x_h_neon() 395 const uint16x4_t top_v = vld1_u16(top_row); in highbd_smooth_4xh_neon() 397 const uint16x4_t weights_x_v = vld1_u16(smooth_weights_u16); in highbd_smooth_4xh_neon() 453 const uint16x4x2_t top_vals = { { vld1_u16(top_row), in highbd_smooth_8xh_neon() 454 vld1_u16(top_row + 4) } }; in highbd_smooth_8xh_neon() 456 const uint16x4x2_t weights_x = { { vld1_u16(smooth_weights_u16 + 4), in highbd_smooth_8xh_neon() 457 vld1_u16(smooth_weights_u16 + 8) } }; in highbd_smooth_8xh_neon() 509 vld1_u16(smooth_weights_u16 + (W)-4 + x); \ [all …]
|
D | highbd_loopfilter_neon.c | 234 const uint16x4_t src[4] = { vld1_u16(dst_p1), vld1_u16(dst_p0), in aom_highbd_lpf_horizontal_4_neon() 235 vld1_u16(dst_q0), vld1_u16(dst_q1) }; in aom_highbd_lpf_horizontal_4_neon() 299 uint16x4_t src[4] = { vld1_u16(dst_p1), vld1_u16(dst_p0), vld1_u16(dst_q0), in aom_highbd_lpf_vertical_4_neon() 300 vld1_u16(dst_q1) }; in aom_highbd_lpf_vertical_4_neon() 421 const uint16x4_t src[6] = { vld1_u16(dst_p2), vld1_u16(dst_p1), in aom_highbd_lpf_horizontal_6_neon() 422 vld1_u16(dst_p0), vld1_u16(dst_q0), in aom_highbd_lpf_horizontal_6_neon() 423 vld1_u16(dst_q1), vld1_u16(dst_q2) }; in aom_highbd_lpf_horizontal_6_neon() 666 const uint16x4_t src[8] = { vld1_u16(dst_p3), vld1_u16(dst_p2), in aom_highbd_lpf_horizontal_8_neon() 667 vld1_u16(dst_p1), vld1_u16(dst_p0), in aom_highbd_lpf_horizontal_8_neon() 668 vld1_u16(dst_q0), vld1_u16(dst_q1), in aom_highbd_lpf_horizontal_8_neon() [all …]
|
D | highbd_variance_neon.c | 86 const uint16x4_t v_a_r0 = vld1_u16(&src[0]); in aom_highbd_calc4x4var_neon() 87 const uint16x4_t v_b_r0 = vld1_u16(&ref[0]); in aom_highbd_calc4x4var_neon() 88 const uint16x4_t v_a_r1 = vld1_u16(&src[src_stride]); in aom_highbd_calc4x4var_neon() 89 const uint16x4_t v_b_r1 = vld1_u16(&ref[ref_stride]); in aom_highbd_calc4x4var_neon()
|
D | subtract_neon.c | 151 const uint16x4_t v_src_r0 = vld1_u16(&src[0]); in aom_highbd_subtract_block_neon() 152 const uint16x4_t v_src_r1 = vld1_u16(&src[src_stride]); in aom_highbd_subtract_block_neon() 153 const uint16x4_t v_pred_r0 = vld1_u16(&pred[0]); in aom_highbd_subtract_block_neon() 154 const uint16x4_t v_pred_r1 = vld1_u16(&pred[pred_stride]); in aom_highbd_subtract_block_neon()
|
D | avg_neon.c | 198 const uint16x4_t r0 = vld1_u16(src); in aom_highbd_avg_4x4_neon() 201 r1 = vld1_u16(src); in aom_highbd_avg_4x4_neon() 203 r2 = vld1_u16(src); in aom_highbd_avg_4x4_neon() 205 r3 = vld1_u16(src); in aom_highbd_avg_4x4_neon()
|
/external/libgav1/src/dsp/arm/ |
D | loop_filter_10bit_neon.cc | 231 const uint16x4_t src[4] = {vld1_u16(dst_p1), vld1_u16(dst_p0), in Horizontal4_NEON() 232 vld1_u16(dst_q0), vld1_u16(dst_q1)}; in Horizontal4_NEON() 287 uint16x4_t src[4] = {vld1_u16(dst_p1), vld1_u16(dst_p0), vld1_u16(dst_q0), in Vertical4_NEON() 288 vld1_u16(dst_q1)}; in Vertical4_NEON() 399 const uint16x4_t src[6] = {vld1_u16(dst_p2), vld1_u16(dst_p1), in Horizontal6_NEON() 400 vld1_u16(dst_p0), vld1_u16(dst_q0), in Horizontal6_NEON() 401 vld1_u16(dst_q1), vld1_u16(dst_q2)}; in Horizontal6_NEON() 626 vld1_u16(dst_p3), vld1_u16(dst_p2), vld1_u16(dst_p1), vld1_u16(dst_p0), in Horizontal8_NEON() 627 vld1_u16(dst_q0), vld1_u16(dst_q1), vld1_u16(dst_q2), vld1_u16(dst_q3)}; in Horizontal8_NEON() 898 vld1_u16(dst_p6), vld1_u16(dst_p5), vld1_u16(dst_p4), vld1_u16(dst_p3), in Horizontal14_NEON() [all …]
|
D | intrapred_smooth_neon.cc | 634 const uint16x4_t top_v = vld1_u16(top); in Smooth4xH_NEON() 636 const uint16x4_t weights_x_v = vld1_u16(kSmoothWeights); in Smooth4xH_NEON() 696 const uint16x4x2_t top_vals = {vld1_u16(top), vld1_u16(top + 4)}; in Smooth8xH_NEON() 698 const uint16x4x2_t weights_x = {vld1_u16(kSmoothWeights + 4), in Smooth8xH_NEON() 699 vld1_u16(kSmoothWeights + 8)}; in Smooth8xH_NEON() 737 const uint16x4_t weights_x_low = vld1_u16(kSmoothWeights + width - 4 + x); in SmoothWxH_NEON() 739 const uint16x4_t weights_x_high = vld1_u16(kSmoothWeights + width + x); in SmoothWxH_NEON() 750 const uint16x4x2_t top_vals = {vld1_u16(top + x), vld1_u16(top + x + 4)}; in SmoothWxH_NEON() 756 const uint16x4x2_t weights_x = {vld1_u16(kSmoothWeights + width - 4 + x), in SmoothWxH_NEON() 757 vld1_u16(kSmoothWeights + width + x)}; in SmoothWxH_NEON() [all …]
|
D | distance_weighted_blend_neon.cc | 227 x.val[0] = vld1_u16(ptr); in LoadU16x4_x2() 228 x.val[1] = vld1_u16(ptr + 4); in LoadU16x4_x2() 237 x.val[0] = vld1_u16(ptr); in LoadU16x4_x4() 238 x.val[1] = vld1_u16(ptr + 4); in LoadU16x4_x4() 239 x.val[2] = vld1_u16(ptr + 8); in LoadU16x4_x4() 240 x.val[3] = vld1_u16(ptr + 12); in LoadU16x4_x4()
|
D | cdef_neon.cc | 456 output[0] = vcombine_u16(vld1_u16(src + y_0 * stride + x_0), in LoadDirection4() 457 vld1_u16(src + y_0 * stride + stride + x_0)); in LoadDirection4() 458 output[1] = vcombine_u16(vld1_u16(src - y_0 * stride - x_0), in LoadDirection4() 459 vld1_u16(src - y_0 * stride + stride - x_0)); in LoadDirection4() 460 output[2] = vcombine_u16(vld1_u16(src + y_1 * stride + x_1), in LoadDirection4() 461 vld1_u16(src + y_1 * stride + stride + x_1)); in LoadDirection4() 462 output[3] = vcombine_u16(vld1_u16(src - y_1 * stride - x_1), in LoadDirection4() 463 vld1_u16(src - y_1 * stride + stride - x_1)); in LoadDirection4() 619 pixel = vcombine_u16(vld1_u16(src), vld1_u16(src + src_stride)); in CdefFilter_NEON()
|
D | common_neon_test.cc | 116 input_4x4[0] = vld1_u16(src_block_[0]); in TEST_F() 117 input_4x4[1] = vld1_u16(src_block_[1]); in TEST_F() 118 input_4x4[2] = vld1_u16(src_block_[2]); in TEST_F() 119 input_4x4[3] = vld1_u16(src_block_[3]); in TEST_F()
|
/external/libvpx/vpx_dsp/arm/ |
D | highbd_vpx_convolve8_neon.c | 404 d01 = vcombine_u16(vld1_u16(dst + 0 * dst_stride), in vpx_highbd_convolve8_avg_horiz_neon() 405 vld1_u16(dst + 2 * dst_stride)); in vpx_highbd_convolve8_avg_horiz_neon() 406 d23 = vcombine_u16(vld1_u16(dst + 1 * dst_stride), in vpx_highbd_convolve8_avg_horiz_neon() 407 vld1_u16(dst + 3 * dst_stride)); in vpx_highbd_convolve8_avg_horiz_neon() 466 d0 = vcombine_u16(vld1_u16(dst + 0 * dst_stride), in vpx_highbd_convolve8_avg_horiz_neon() 467 vld1_u16(dst + 4 * dst_stride)); in vpx_highbd_convolve8_avg_horiz_neon() 468 d1 = vcombine_u16(vld1_u16(dst + 1 * dst_stride), in vpx_highbd_convolve8_avg_horiz_neon() 469 vld1_u16(dst + 5 * dst_stride)); in vpx_highbd_convolve8_avg_horiz_neon() 470 d2 = vcombine_u16(vld1_u16(dst + 2 * dst_stride), in vpx_highbd_convolve8_avg_horiz_neon() 471 vld1_u16(dst + 6 * dst_stride)); in vpx_highbd_convolve8_avg_horiz_neon() [all …]
|
D | highbd_sad_neon.c | 31 const uint16x4_t src_u16 = vld1_u16(src16_ptr + j); in highbd_sad4_neon() 32 const uint16x4_t ref_u16 = vld1_u16(ref16_ptr + j); in highbd_sad4_neon() 77 const uint16x4_t a_u16 = vld1_u16(src16_ptr + j); in highbd_sad4_avg_neon() 78 const uint16x4_t b_u16 = vld1_u16(ref16_ptr + j); in highbd_sad4_avg_neon() 79 const uint16x4_t c_u16 = vld1_u16(pred_ptr + j); in highbd_sad4_avg_neon()
|
D | highbd_vpx_convolve_avg_neon.c | 32 s0 = vld1_u16(src); in vpx_highbd_convolve_avg_neon() 33 d0 = vld1_u16(dst); in vpx_highbd_convolve_avg_neon() 35 s1 = vld1_u16(src); in vpx_highbd_convolve_avg_neon() 36 d1 = vld1_u16(dst + dst_stride); in vpx_highbd_convolve_avg_neon()
|
D | highbd_variance_neon.c | 61 const int16x4_t src_s16 = vreinterpret_s16_u16(vld1_u16(&src_ptr[j])); in highbd_variance16() 62 const int16x4_t ref_s16 = vreinterpret_s16_u16(vld1_u16(&ref_ptr[j])); in highbd_variance16() 256 const uint16x4_t src1_u16 = vld1_u16(&src_ptr[j]); in highbd_var_filter_block2d_bil_first_pass() 257 const uint16x4_t src2_u16 = vld1_u16(&src_ptr[j + pixel_step]); in highbd_var_filter_block2d_bil_first_pass() 305 const uint16x4_t src1_u16 = vld1_u16(&src_ptr[j]); in highbd_var_filter_block2d_bil_second_pass() 306 const uint16x4_t src2_u16 = vld1_u16(&src_ptr[j + pixel_step]); in highbd_var_filter_block2d_bil_second_pass() 456 const uint16x4_t pred_u16 = vld1_u16(&pred[j]); in vpx_highbd_comp_avg_pred_neon() 457 const uint16x4_t ref_u16 = vld1_u16(&ref[j]); in vpx_highbd_comp_avg_pred_neon()
|
/external/XNNPACK/src/x16-transposec/gen/ |
D | 4x4-reuse-multi-zip-neon.c | 54 const uint16x4_t v2_0 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_multi_zip_neon() 55 const uint16x4_t v2_1 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_multi_zip_neon() 56 const uint16x4_t v2_2 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_multi_zip_neon() 57 const uint16x4_t v2_3 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_multi_zip_neon() 72 const uint16x4_t v2_0 = vld1_u16(i0); in xnn_x16_transposec_ukernel__4x4_reuse_multi_zip_neon() 77 const uint16x4_t v2_1 = vld1_u16(i1); in xnn_x16_transposec_ukernel__4x4_reuse_multi_zip_neon() 82 const uint16x4_t v2_2 = vld1_u16(i2); in xnn_x16_transposec_ukernel__4x4_reuse_multi_zip_neon()
|
D | 4x4-reuse-dec-zip-neon.c | 46 const uint16x4_t v2_0 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_dec_zip_neon() 47 const uint16x4_t v2_1 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_dec_zip_neon() 48 const uint16x4_t v2_2 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_dec_zip_neon() 49 const uint16x4_t v2_3 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_dec_zip_neon() 75 const uint16x4_t v2_0 = vld1_u16(i0); in xnn_x16_transposec_ukernel__4x4_reuse_dec_zip_neon() 80 const uint16x4_t v2_1 = vld1_u16(i1); in xnn_x16_transposec_ukernel__4x4_reuse_dec_zip_neon() 85 const uint16x4_t v2_2 = vld1_u16(i2); in xnn_x16_transposec_ukernel__4x4_reuse_dec_zip_neon()
|
D | 4x4-multi-multi-zip-neon.c | 58 const uint16x4_t v2_0 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_multi_zip_neon() 59 const uint16x4_t v2_1 = vld1_u16(i1); i1 = (uint16_t*) ((uintptr_t) i1 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_multi_zip_neon() 60 const uint16x4_t v2_2 = vld1_u16(i2); i2 = (uint16_t*) ((uintptr_t) i2 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_multi_zip_neon() 61 const uint16x4_t v2_3 = vld1_u16(i3); i3 = (uint16_t*) ((uintptr_t) i3 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_multi_zip_neon() 76 const uint16x4_t v2_0 = vld1_u16(i0); in xnn_x16_transposec_ukernel__4x4_multi_multi_zip_neon() 80 const uint16x4_t v2_1 = vld1_u16(i1); in xnn_x16_transposec_ukernel__4x4_multi_multi_zip_neon() 84 const uint16x4_t v2_2 = vld1_u16(i2); in xnn_x16_transposec_ukernel__4x4_multi_multi_zip_neon()
|
D | 4x4-multi-switch-zip-neon.c | 49 const uint16x4_t v2_0 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_switch_zip_neon() 50 const uint16x4_t v2_1 = vld1_u16(i1); i1 = (uint16_t*) ((uintptr_t) i1 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_switch_zip_neon() 51 const uint16x4_t v2_2 = vld1_u16(i2); i2 = (uint16_t*) ((uintptr_t) i2 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_switch_zip_neon() 52 const uint16x4_t v2_3 = vld1_u16(i3); i3 = (uint16_t*) ((uintptr_t) i3 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_switch_zip_neon() 77 const uint16x4_t v2_0 = vld1_u16(i0); in xnn_x16_transposec_ukernel__4x4_multi_switch_zip_neon() 81 const uint16x4_t v2_1 = vld1_u16(i1); in xnn_x16_transposec_ukernel__4x4_multi_switch_zip_neon() 85 const uint16x4_t v2_2 = vld1_u16(i2); in xnn_x16_transposec_ukernel__4x4_multi_switch_zip_neon()
|
D | 4x4-reuse-switch-zip-neon.c | 45 const uint16x4_t v2_0 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_switch_zip_neon() 46 const uint16x4_t v2_1 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_switch_zip_neon() 47 const uint16x4_t v2_2 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_switch_zip_neon() 48 const uint16x4_t v2_3 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_switch_zip_neon() 73 const uint16x4_t v2_0 = vld1_u16(i0); in xnn_x16_transposec_ukernel__4x4_reuse_switch_zip_neon() 78 const uint16x4_t v2_1 = vld1_u16(i1); in xnn_x16_transposec_ukernel__4x4_reuse_switch_zip_neon() 83 const uint16x4_t v2_2 = vld1_u16(i2); in xnn_x16_transposec_ukernel__4x4_reuse_switch_zip_neon()
|
D | 4x4-multi-dec-zip-neon.c | 50 const uint16x4_t v2_0 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_dec_zip_neon() 51 const uint16x4_t v2_1 = vld1_u16(i1); i1 = (uint16_t*) ((uintptr_t) i1 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_dec_zip_neon() 52 const uint16x4_t v2_2 = vld1_u16(i2); i2 = (uint16_t*) ((uintptr_t) i2 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_dec_zip_neon() 53 const uint16x4_t v2_3 = vld1_u16(i3); i3 = (uint16_t*) ((uintptr_t) i3 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_dec_zip_neon() 79 const uint16x4_t v2_0 = vld1_u16(i0); in xnn_x16_transposec_ukernel__4x4_multi_dec_zip_neon() 83 const uint16x4_t v2_1 = vld1_u16(i1); in xnn_x16_transposec_ukernel__4x4_multi_dec_zip_neon() 87 const uint16x4_t v2_2 = vld1_u16(i2); in xnn_x16_transposec_ukernel__4x4_multi_dec_zip_neon()
|
D | 4x4-reuse-mov-zip-neon.c | 46 const uint16x4_t v2_0 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_mov_zip_neon() 47 const uint16x4_t v2_1 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_mov_zip_neon() 48 const uint16x4_t v2_2 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_mov_zip_neon() 49 const uint16x4_t v2_3 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_stride); in xnn_x16_transposec_ukernel__4x4_reuse_mov_zip_neon() 78 const uint16x4_t v2_0 = vld1_u16(i0); in xnn_x16_transposec_ukernel__4x4_reuse_mov_zip_neon() 83 const uint16x4_t v2_1 = vld1_u16(i1); in xnn_x16_transposec_ukernel__4x4_reuse_mov_zip_neon() 88 const uint16x4_t v2_2 = vld1_u16(i2); in xnn_x16_transposec_ukernel__4x4_reuse_mov_zip_neon()
|
D | 4x4-multi-mov-zip-neon.c | 50 const uint16x4_t v2_0 = vld1_u16(i0); i0 = (uint16_t*) ((uintptr_t) i0 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_mov_zip_neon() 51 const uint16x4_t v2_1 = vld1_u16(i1); i1 = (uint16_t*) ((uintptr_t) i1 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_mov_zip_neon() 52 const uint16x4_t v2_2 = vld1_u16(i2); i2 = (uint16_t*) ((uintptr_t) i2 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_mov_zip_neon() 53 const uint16x4_t v2_3 = vld1_u16(i3); i3 = (uint16_t*) ((uintptr_t) i3 + input_offset); in xnn_x16_transposec_ukernel__4x4_multi_mov_zip_neon() 82 const uint16x4_t v2_0 = vld1_u16(i0); in xnn_x16_transposec_ukernel__4x4_multi_mov_zip_neon() 86 const uint16x4_t v2_1 = vld1_u16(i1); in xnn_x16_transposec_ukernel__4x4_multi_mov_zip_neon() 90 const uint16x4_t v2_2 = vld1_u16(i2); in xnn_x16_transposec_ukernel__4x4_multi_mov_zip_neon()
|
/external/libavc/common/arm/svc/ |
D | isvc_resi_trans_quant_neon.c | 87 threshold0_16x4 = vld1_u16(pu2_threshold_matrix); in isvc_resi_trans_quant_4x4_neon() 88 threshold1_16x4 = vld1_u16(pu2_threshold_matrix + 4); in isvc_resi_trans_quant_4x4_neon() 89 threshold2_16x4 = vld1_u16(pu2_threshold_matrix + 8); in isvc_resi_trans_quant_4x4_neon() 90 threshold3_16x4 = vld1_u16(pu2_threshold_matrix + 12); in isvc_resi_trans_quant_4x4_neon() 92 scale_mat0_16x4 = vld1_u16(pu2_scale_matrix); in isvc_resi_trans_quant_4x4_neon() 93 scale_mat1_16x4 = vld1_u16(pu2_scale_matrix + 4); in isvc_resi_trans_quant_4x4_neon() 94 scale_mat2_16x4 = vld1_u16(pu2_scale_matrix + 8); in isvc_resi_trans_quant_4x4_neon() 95 scale_mat3_16x4 = vld1_u16(pu2_scale_matrix + 12); in isvc_resi_trans_quant_4x4_neon() 325 threshold0_16x4 = vld1_u16(pu2_threshold_matrix); in isvc_resi_trans_quant_4x4_with_residual_sub_neon() 326 threshold1_16x4 = vld1_u16(pu2_threshold_matrix + 4); in isvc_resi_trans_quant_4x4_with_residual_sub_neon() [all …]
|
/external/libaom/av1/common/arm/ |
D | reconinter_neon.c | 71 tmp0 = vcombine_u16(vld1_u16(src0_1 + (0 * src0_stride)), in av1_build_compound_diffwtd_mask_d16_neon() 72 vld1_u16(src0_1 + (1 * src0_stride))); in av1_build_compound_diffwtd_mask_d16_neon() 73 tmp1 = vcombine_u16(vld1_u16(src1_1 + (0 * src1_stride)), in av1_build_compound_diffwtd_mask_d16_neon() 74 vld1_u16(src1_1 + (1 * src1_stride))); in av1_build_compound_diffwtd_mask_d16_neon()
|
D | cfl_neon.c | 150 const uint16x4_t top = vld1_u16(input); in cfl_luma_subsampling_420_hbd_neon() 151 const uint16x4_t bot = vld1_u16(input + input_stride); in cfl_luma_subsampling_420_hbd_neon() 196 const uint16x4_t top = vld1_u16(input); in cfl_luma_subsampling_422_hbd_neon() 230 const uint16x4_t top = vld1_u16(input); in cfl_luma_subsampling_444_hbd_neon() 277 vadd_u16(vld1_u16(sum_buf), vld1_u16(sum_buf + CFL_BUF_LINE)); in CFL_GET_SUBSAMPLE_FUNCTION() 278 const uint16x4_t a1 = vadd_u16(vld1_u16(sum_buf + 2 * CFL_BUF_LINE), in CFL_GET_SUBSAMPLE_FUNCTION() 279 vld1_u16(sum_buf + 3 * CFL_BUF_LINE)); in CFL_GET_SUBSAMPLE_FUNCTION() 345 vst1_s16(dst, vsub_s16(vreinterpret_s16_u16(vld1_u16(src)), avg_16x4)); in CFL_GET_SUBSAMPLE_FUNCTION()
|