/external/libjpeg-turbo/simd/arm/ |
D | jidctint-neon.c | 201 int16x4_t row0 = vld1_s16(coef_block + 0 * DCTSIZE); in jsimd_idct_islow_neon() 202 int16x4_t row1 = vld1_s16(coef_block + 1 * DCTSIZE); in jsimd_idct_islow_neon() 203 int16x4_t row2 = vld1_s16(coef_block + 2 * DCTSIZE); in jsimd_idct_islow_neon() 204 int16x4_t row3 = vld1_s16(coef_block + 3 * DCTSIZE); in jsimd_idct_islow_neon() 205 int16x4_t row4 = vld1_s16(coef_block + 4 * DCTSIZE); in jsimd_idct_islow_neon() 206 int16x4_t row5 = vld1_s16(coef_block + 5 * DCTSIZE); in jsimd_idct_islow_neon() 207 int16x4_t row6 = vld1_s16(coef_block + 6 * DCTSIZE); in jsimd_idct_islow_neon() 208 int16x4_t row7 = vld1_s16(coef_block + 7 * DCTSIZE); in jsimd_idct_islow_neon() 211 int16x4_t quant_row0 = vld1_s16(quantptr + 0 * DCTSIZE); in jsimd_idct_islow_neon() 212 int16x4_t quant_row1 = vld1_s16(quantptr + 1 * DCTSIZE); in jsimd_idct_islow_neon() [all …]
|
D | jidctred-neon.c | 103 const int16x4_t consts = vld1_s16(jsimd_idct_2x2_neon_consts); in jsimd_idct_2x2_neon() 230 const int16x4_t consts1 = vld1_s16(jsimd_idct_4x4_neon_consts); in jsimd_idct_4x4_neon() 231 const int16x4_t consts2 = vld1_s16(jsimd_idct_4x4_neon_consts + 4); in jsimd_idct_4x4_neon() 232 const int16x4_t consts3 = vld1_s16(jsimd_idct_4x4_neon_consts + 8); in jsimd_idct_4x4_neon() 254 int16x4_t quant_row1 = vld1_s16(quantptr + 1 * DCTSIZE + 4); in jsimd_idct_4x4_neon() 255 int16x4_t quant_row2 = vld1_s16(quantptr + 2 * DCTSIZE + 4); in jsimd_idct_4x4_neon() 256 int16x4_t quant_row3 = vld1_s16(quantptr + 3 * DCTSIZE + 4); in jsimd_idct_4x4_neon() 257 int16x4_t quant_row5 = vld1_s16(quantptr + 5 * DCTSIZE + 4); in jsimd_idct_4x4_neon() 258 int16x4_t quant_row6 = vld1_s16(quantptr + 6 * DCTSIZE + 4); in jsimd_idct_4x4_neon() 259 int16x4_t quant_row7 = vld1_s16(quantptr + 7 * DCTSIZE + 4); in jsimd_idct_4x4_neon() [all …]
|
D | jidctfst-neon.c | 95 const int16x4_t consts = vld1_s16(jsimd_idct_ifast_neon_consts); in jsimd_idct_ifast_neon() 118 int16x4_t quant_row1 = vld1_s16(quantptr + 1 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 119 int16x4_t quant_row2 = vld1_s16(quantptr + 2 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 120 int16x4_t quant_row3 = vld1_s16(quantptr + 3 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 121 int16x4_t quant_row4 = vld1_s16(quantptr + 4 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 122 int16x4_t quant_row5 = vld1_s16(quantptr + 5 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 123 int16x4_t quant_row6 = vld1_s16(quantptr + 6 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 124 int16x4_t quant_row7 = vld1_s16(quantptr + 7 * DCTSIZE + 4); in jsimd_idct_ifast_neon() 193 int16x4_t quant_row1 = vld1_s16(quantptr + 1 * DCTSIZE); in jsimd_idct_ifast_neon() 194 int16x4_t quant_row2 = vld1_s16(quantptr + 2 * DCTSIZE); in jsimd_idct_ifast_neon() [all …]
|
/external/libaom/aom_dsp/arm/ |
D | sum_squares_neon.c | 21 int16x4_t s0 = vld1_s16(src + 0 * stride); in aom_sum_squares_2d_i16_4x4_neon() 22 int16x4_t s1 = vld1_s16(src + 1 * stride); in aom_sum_squares_2d_i16_4x4_neon() 23 int16x4_t s2 = vld1_s16(src + 2 * stride); in aom_sum_squares_2d_i16_4x4_neon() 24 int16x4_t s3 = vld1_s16(src + 3 * stride); in aom_sum_squares_2d_i16_4x4_neon() 40 int16x4_t s0 = vld1_s16(src + 0 * stride); in aom_sum_squares_2d_i16_4xn_neon() 41 int16x4_t s1 = vld1_s16(src + 1 * stride); in aom_sum_squares_2d_i16_4xn_neon() 42 int16x4_t s2 = vld1_s16(src + 2 * stride); in aom_sum_squares_2d_i16_4xn_neon() 43 int16x4_t s3 = vld1_s16(src + 3 * stride); in aom_sum_squares_2d_i16_4xn_neon() 114 int16x4_t s0 = vld1_s16(src + 0 * stride); in aom_sum_sse_2d_i16_4x4_neon() 115 int16x4_t s1 = vld1_s16(src + 1 * stride); in aom_sum_sse_2d_i16_4x4_neon() [all …]
|
D | highbd_quantize_neon.c | 143 const int16x4_t v_quant = vld1_s16(quant_ptr); in highbd_quantize_b_neon() 144 const int16x4_t v_dequant = vld1_s16(dequant_ptr); in highbd_quantize_b_neon() 147 const int16x4_t v_round_no_scale = vld1_s16(round_ptr); in highbd_quantize_b_neon() 152 const int16x4_t v_quant_shift = vld1_s16(quant_shift_ptr); in highbd_quantize_b_neon() 153 const int16x4_t v_zbin_no_scale = vld1_s16(zbin_ptr); in highbd_quantize_b_neon() 277 const int16x4_t v_quant = vld1_s16(quant_ptr); in highbd_quantize_b_adaptive_neon() 278 const int16x4_t v_dequant = vld1_s16(dequant_ptr); in highbd_quantize_b_adaptive_neon() 281 const int16x4_t v_round_no_scale = vld1_s16(round_ptr); in highbd_quantize_b_adaptive_neon() 286 const int16x4_t v_quant_shift = vld1_s16(quant_shift_ptr); in highbd_quantize_b_adaptive_neon() 287 const int16x4_t v_zbin_no_scale = vld1_s16(zbin_ptr); in highbd_quantize_b_adaptive_neon()
|
D | mem_neon.h | 169 *s0 = vld1_s16(s); in load_s16_4x8() 171 *s1 = vld1_s16(s); in load_s16_4x8() 173 *s2 = vld1_s16(s); in load_s16_4x8() 175 *s3 = vld1_s16(s); in load_s16_4x8() 177 *s4 = vld1_s16(s); in load_s16_4x8() 179 *s5 = vld1_s16(s); in load_s16_4x8() 181 *s6 = vld1_s16(s); in load_s16_4x8() 183 *s7 = vld1_s16(s); in load_s16_4x8() 189 *s0 = vld1_s16(s); in load_s16_4x4() 191 *s1 = vld1_s16(s); in load_s16_4x4() [all …]
|
D | fwd_txfm_neon.c | 23 *input_0 = vshl_n_s16(vld1_s16(input + 0 * stride), 4); in aom_fdct4x4_helper() 24 *input_1 = vshl_n_s16(vld1_s16(input + 1 * stride), 4); in aom_fdct4x4_helper() 25 *input_2 = vshl_n_s16(vld1_s16(input + 2 * stride), 4); in aom_fdct4x4_helper() 26 *input_3 = vshl_n_s16(vld1_s16(input + 3 * stride), 4); in aom_fdct4x4_helper()
|
/external/libvpx/vpx_dsp/arm/ |
D | fdct4x4_neon.c | 27 in[0] = vshl_n_s16(vld1_s16(input + 0 * stride), 4); in vpx_fdct4x4_neon() 28 in[1] = vshl_n_s16(vld1_s16(input + 1 * stride), 4); in vpx_fdct4x4_neon() 29 in[2] = vshl_n_s16(vld1_s16(input + 2 * stride), 4); in vpx_fdct4x4_neon() 30 in[3] = vshl_n_s16(vld1_s16(input + 3 * stride), 4); in vpx_fdct4x4_neon() 60 in[0] = vshll_n_s16(vld1_s16(input + 0 * stride), 4); in vpx_highbd_fdct4x4_neon() 61 in[1] = vshll_n_s16(vld1_s16(input + 1 * stride), 4); in vpx_highbd_fdct4x4_neon() 62 in[2] = vshll_n_s16(vld1_s16(input + 2 * stride), 4); in vpx_highbd_fdct4x4_neon() 63 in[3] = vshll_n_s16(vld1_s16(input + 3 * stride), 4); in vpx_highbd_fdct4x4_neon()
|
D | highbd_quantize_neon.c | 111 int32x4_t zbin = vmovl_s16(vld1_s16(zbin_ptr)); in vpx_highbd_quantize_b_neon() 112 int32x4_t round = vmovl_s16(vld1_s16(round_ptr)); in vpx_highbd_quantize_b_neon() 115 int32x4_t quant = vshlq_n_s32(vmovl_s16(vld1_s16(quant_ptr)), 15); in vpx_highbd_quantize_b_neon() 116 int32x4_t quant_shift = vshlq_n_s32(vmovl_s16(vld1_s16(quant_shift_ptr)), 15); in vpx_highbd_quantize_b_neon() 117 int32x4_t dequant = vmovl_s16(vld1_s16(dequant_ptr)); in vpx_highbd_quantize_b_neon() 240 int32x4_t zbin = vrshrq_n_s32(vmovl_s16(vld1_s16(zbin_ptr)), 1); in vpx_highbd_quantize_b_32x32_neon() 241 int32x4_t round = vrshrq_n_s32(vmovl_s16(vld1_s16(round_ptr)), 1); in vpx_highbd_quantize_b_32x32_neon() 244 int32x4_t quant = vshlq_n_s32(vmovl_s16(vld1_s16(quant_ptr)), 15); in vpx_highbd_quantize_b_32x32_neon() 245 int32x4_t quant_shift = vshlq_n_s32(vmovl_s16(vld1_s16(quant_shift_ptr)), 16); in vpx_highbd_quantize_b_32x32_neon() 246 int32x4_t dequant = vmovl_s16(vld1_s16(dequant_ptr)); in vpx_highbd_quantize_b_32x32_neon()
|
D | sum_squares_neon.c | 24 s[0] = vld1_s16(src + 0 * stride); in vpx_sum_squares_2d_i16_neon() 25 s[1] = vld1_s16(src + 1 * stride); in vpx_sum_squares_2d_i16_neon() 26 s[2] = vld1_s16(src + 2 * stride); in vpx_sum_squares_2d_i16_neon() 27 s[3] = vld1_s16(src + 3 * stride); in vpx_sum_squares_2d_i16_neon()
|
D | fdct_partial_neon.c | 23 a0 = vld1_s16(input); in vpx_fdct4x4_1_neon() 25 a1 = vld1_s16(input); in vpx_fdct4x4_1_neon() 27 a2 = vld1_s16(input); in vpx_fdct4x4_1_neon() 29 a3 = vld1_s16(input); in vpx_fdct4x4_1_neon()
|
D | idct16x16_add_neon.c | 600 ind[0] = vld1_s16(input); in vpx_idct16x16_10_add_half1d_pass2() 602 ind[1] = vld1_s16(input); in vpx_idct16x16_10_add_half1d_pass2() 604 ind[2] = vld1_s16(input); in vpx_idct16x16_10_add_half1d_pass2() 606 ind[3] = vld1_s16(input); in vpx_idct16x16_10_add_half1d_pass2() 608 ind[4] = vld1_s16(input); in vpx_idct16x16_10_add_half1d_pass2() 610 ind[5] = vld1_s16(input); in vpx_idct16x16_10_add_half1d_pass2() 612 ind[6] = vld1_s16(input); in vpx_idct16x16_10_add_half1d_pass2() 614 ind[7] = vld1_s16(input); in vpx_idct16x16_10_add_half1d_pass2()
|
/external/libhevc/common/arm/ |
D | ihevc_weighted_pred_neon_intr.c | 149 pi2_src_val1 = vld1_s16((int16_t *)pi2_src); in ihevc_weighted_pred_uni_neonintr() 152 pi2_src_val2 = vld1_s16((int16_t *)pi2_src_tmp); in ihevc_weighted_pred_uni_neonintr() 296 pi2_src_val1 = vld1_s16((int16_t *)pi2_src); in ihevc_weighted_pred_chroma_uni_neonintr() 299 pi2_src_val2 = vld1_s16((int16_t *)pi2_src_tmp); in ihevc_weighted_pred_chroma_uni_neonintr() 449 pi2_src1_val1 = vld1_s16((int16_t *)pi2_src1); in ihevc_weighted_pred_bi_neonintr() 453 pi2_src2_val1 = vld1_s16((int16_t *)pi2_src2); in ihevc_weighted_pred_bi_neonintr() 457 pi2_src1_val2 = vld1_s16((int16_t *)pi2_src_tmp1); in ihevc_weighted_pred_bi_neonintr() 460 pi2_src2_val2 = vld1_s16((int16_t *)pi2_src_tmp2); in ihevc_weighted_pred_bi_neonintr() 635 pi2_src1_val1 = vld1_s16((int16_t *)pi2_src1); in ihevc_weighted_pred_chroma_bi_neonintr() 639 pi2_src2_val1 = vld1_s16((int16_t *)pi2_src2); in ihevc_weighted_pred_chroma_bi_neonintr() [all …]
|
D | ihevc_quant_iquant_ssd_neon_intr.c | 161 s0 = vld1_s16(pi2_coeffs + j); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 162 s1 = vld1_s16(pi2_coeffs + j + (src_strd)); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 163 s2 = vld1_s16(pi2_coeffs + j + (2 * src_strd)); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 164 s3 = vld1_s16(pi2_coeffs + j + (3 * src_strd)); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 447 s0 = vld1_s16(pi2_coeffs + j); in ihevc_q_iq_ssd_flat_scale_mat_var_rnd_fact_neon() 448 s1 = vld1_s16(pi2_coeffs + j + (src_strd)); in ihevc_q_iq_ssd_flat_scale_mat_var_rnd_fact_neon() 449 s2 = vld1_s16(pi2_coeffs + j + (2 * src_strd)); in ihevc_q_iq_ssd_flat_scale_mat_var_rnd_fact_neon() 450 s3 = vld1_s16(pi2_coeffs + j + (3 * src_strd)); in ihevc_q_iq_ssd_flat_scale_mat_var_rnd_fact_neon()
|
/external/libaom/av1/encoder/arm/neon/ |
D | hybrid_fwd_txfm_neon.c | 32 a1 = vld1_s16(&input[0]); in av1_fwht4x4_neon() 33 b1 = vld1_s16(&input[1 * stride]); in av1_fwht4x4_neon() 34 c1 = vld1_s16(&input[2 * stride]); in av1_fwht4x4_neon() 35 d1 = vld1_s16(&input[3 * stride]); in av1_fwht4x4_neon()
|
D | av1_highbd_quantize_neon.c | 95 const int16x4_t v_quant = vld1_s16(quant_ptr); in av1_highbd_quantize_fp_neon() 96 const int16x4_t v_dequant = vld1_s16(dequant_ptr); in av1_highbd_quantize_fp_neon() 99 const int16x4_t v_round_no_scale = vld1_s16(round_ptr); in av1_highbd_quantize_fp_neon()
|
/external/libhevc/encoder/arm/ |
D | ihevce_scan_coeffs_neon.c | 193 quant0 = vld1_s16(pi2_temp_quant_coeff + 0 * trans_size); in ihevce_scan_coeffs_neon() 194 quant1 = vld1_s16(pi2_temp_quant_coeff + 1 * trans_size); in ihevce_scan_coeffs_neon() 195 quant2 = vld1_s16(pi2_temp_quant_coeff + 2 * trans_size); in ihevce_scan_coeffs_neon() 196 quant3 = vld1_s16(pi2_temp_quant_coeff + 3 * trans_size); in ihevce_scan_coeffs_neon() 385 quant0 = vld1_s16(pi2_temp_quant_coeff + 0 * trans_size); in ihevce_scan_coeffs_neon() 386 quant1 = vld1_s16(pi2_temp_quant_coeff + 1 * trans_size); in ihevce_scan_coeffs_neon() 387 quant2 = vld1_s16(pi2_temp_quant_coeff + 2 * trans_size); in ihevce_scan_coeffs_neon() 388 quant3 = vld1_s16(pi2_temp_quant_coeff + 3 * trans_size); in ihevce_scan_coeffs_neon()
|
/external/libvpx/vp8/common/arm/neon/ |
D | shortidct4x4llm_neon.c | 34 d2 = vld1_s16(input); in vp8_short_idct4x4llm_neon() 35 d3 = vld1_s16(input + 4); in vp8_short_idct4x4llm_neon() 36 d4 = vld1_s16(input + 8); in vp8_short_idct4x4llm_neon() 37 d5 = vld1_s16(input + 12); in vp8_short_idct4x4llm_neon()
|
/external/libavc/common/arm/svc/ |
D | isvc_intra_sampling_neon.c | 249 i4_samp_horz_16x4_0 = vld1_s16(pi2_tmp); in isvc_interpolate_base_luma_dyadic_neon() 250 i4_samp_horz_16x4_1 = vld1_s16(pi2_tmp + 1); in isvc_interpolate_base_luma_dyadic_neon() 251 i4_samp_horz_16x4_2 = vld1_s16(pi2_tmp + 2); in isvc_interpolate_base_luma_dyadic_neon() 252 i4_samp_horz_16x4_3 = vld1_s16(pi2_tmp + 3); in isvc_interpolate_base_luma_dyadic_neon() 253 i4_samp_horz_16x4_4 = vld1_s16(pi2_tmp + 4); in isvc_interpolate_base_luma_dyadic_neon() 254 i4_samp_horz_16x4_5 = vld1_s16(pi2_tmp + 5); in isvc_interpolate_base_luma_dyadic_neon() 255 i4_samp_horz_16x4_6 = vld1_s16(pi2_tmp + 6); in isvc_interpolate_base_luma_dyadic_neon() 256 i4_samp_horz_16x4_7 = vld1_s16(pi2_tmp + 7); in isvc_interpolate_base_luma_dyadic_neon() 257 i4_samp_horz_16x4_8 = vld1_s16(pi2_tmp + 8); in isvc_interpolate_base_luma_dyadic_neon()
|
/external/libvpx/vp8/encoder/arm/neon/ |
D | vp8_shortwalsh4x4_neon.c | 36 d0s16 = vld1_s16(input); in vp8_short_walsh4x4_neon() 38 d1s16 = vld1_s16(input); in vp8_short_walsh4x4_neon() 40 d2s16 = vld1_s16(input); in vp8_short_walsh4x4_neon() 42 d3s16 = vld1_s16(input); in vp8_short_walsh4x4_neon()
|
D | shortfdct_neon.c | 33 d0s16 = vld1_s16(input); in vp8_short_fdct4x4_neon() 35 d1s16 = vld1_s16(input); in vp8_short_fdct4x4_neon() 37 d2s16 = vld1_s16(input); in vp8_short_fdct4x4_neon() 39 d3s16 = vld1_s16(input); in vp8_short_fdct4x4_neon()
|
/external/libopus/celt/arm/ |
D | pitch_neon_intr.c | 54 const int16x4_t x_s16x4 = vld1_s16(&x[i]); in celt_inner_prod_neon() 55 const int16x4_t y_s16x4 = vld1_s16(&y[i]); in celt_inner_prod_neon() 97 const int16x4_t x_s16x4 = vld1_s16(&x[i]); in dual_inner_prod_neon() 98 const int16x4_t y01_s16x4 = vld1_s16(&y01[i]); in dual_inner_prod_neon() 99 const int16x4_t y02_s16x4 = vld1_s16(&y02[i]); in dual_inner_prod_neon()
|
/external/libavc/encoder/arm/svc/ |
D | isvce_residual_pred_neon.c | 375 i4_horz_samp_16x4_r0_1 = vld1_s16(pi2_refarray_buffer); in isvce_luma_residual_sampler_2x_neon() 376 i4_horz_samp_16x4_r0_2 = vld1_s16(pi2_refarray_buffer + 4); in isvce_luma_residual_sampler_2x_neon() 378 i4_horz_samp_16x4_r1_1 = vld1_s16(pi2_refarray_buffer + 8); in isvce_luma_residual_sampler_2x_neon() 379 i4_horz_samp_16x4_r1_2 = vld1_s16(pi2_refarray_buffer + 12); in isvce_luma_residual_sampler_2x_neon() 381 i4_horz_samp_16x4_r2_1 = vld1_s16(pi2_refarray_buffer + 16); in isvce_luma_residual_sampler_2x_neon() 382 i4_horz_samp_16x4_r2_2 = vld1_s16(pi2_refarray_buffer + 20); in isvce_luma_residual_sampler_2x_neon() 384 i4_horz_samp_16x4_r3_1 = vld1_s16(pi2_refarray_buffer + 24); in isvce_luma_residual_sampler_2x_neon() 385 i4_horz_samp_16x4_r3_2 = vld1_s16(pi2_refarray_buffer + 28); in isvce_luma_residual_sampler_2x_neon()
|
/external/libvpx/vp9/encoder/arm/neon/ |
D | vp9_quantize_neon.c | 268 const int16x4_t v_quant = vld1_s16(quant_ptr); in vp9_highbd_quantize_fp_neon() 269 const int16x4_t v_dequant = vld1_s16(dequant_ptr); in vp9_highbd_quantize_fp_neon() 270 const int16x4_t v_round = vld1_s16(round_ptr); in vp9_highbd_quantize_fp_neon() 356 const int16x4_t v_quant = vld1_s16(quant_ptr); in vp9_highbd_quantize_fp_32x32_neon() 357 const int16x4_t v_dequant = vld1_s16(dequant_ptr); in vp9_highbd_quantize_fp_32x32_neon() 360 vqrdmulh_n_s16(vld1_s16(round_ptr), (int16_t)(1 << 14)); in vp9_highbd_quantize_fp_32x32_neon()
|
/external/webp/sharpyuv/ |
D | sharpyuv_neon.c | 122 const int16x4_t a0 = vld1_s16(A + i + 0); in SharpYuvFilterRow32_NEON() 123 const int16x4_t a1 = vld1_s16(A + i + 1); in SharpYuvFilterRow32_NEON() 124 const int16x4_t b0 = vld1_s16(B + i + 0); in SharpYuvFilterRow32_NEON() 125 const int16x4_t b1 = vld1_s16(B + i + 1); in SharpYuvFilterRow32_NEON()
|