/external/libvpx/libvpx/vpx_dsp/arm/ |
D | sad_neon.c | 24 uint16x8_t abs = vabdl_u8(vget_low_u8(src_u8), vget_low_u8(ref_u8)); in vpx_sad4x4_neon() 36 uint16x8_t abs = vabdl_u8(vget_low_u8(src_u8), vget_low_u8(avg)); in vpx_sad4x4_avg_neon() 50 abs = vabal_u8(abs, vget_low_u8(src_u8), vget_low_u8(ref_u8)); in vpx_sad4x8_neon() 70 abs = vabal_u8(abs, vget_low_u8(src_u8), vget_low_u8(avg)); in vpx_sad4x8_avg_neon() 143 abs = vabal_u8(abs, vget_low_u8(a_u8), vget_low_u8(b_u8)); in sad16x() 164 abs = vabal_u8(abs, vget_low_u8(a_u8), vget_low_u8(avg)); in sad16x_avg() 203 abs = vabal_u8(abs, vget_low_u8(a_lo), vget_low_u8(b_lo)); in sad32x() 205 abs = vabal_u8(abs, vget_low_u8(a_hi), vget_low_u8(b_hi)); in sad32x() 230 abs = vabal_u8(abs, vget_low_u8(a_lo), vget_low_u8(avg_lo)); in sad32x_avg() 232 abs = vabal_u8(abs, vget_low_u8(a_hi), vget_low_u8(avg_hi)); in sad32x_avg() [all …]
|
D | subtract_neon.c | 32 const uint16x8_t d0 = vsubl_u8(vget_low_u8(s0), vget_low_u8(p0)); in vpx_subtract_block_neon() 34 const uint16x8_t d2 = vsubl_u8(vget_low_u8(s1), vget_low_u8(p1)); in vpx_subtract_block_neon() 49 const uint16x8_t d0 = vsubl_u8(vget_low_u8(s), vget_low_u8(p)); in vpx_subtract_block_neon()
|
D | avg_neon.c | 24 const uint16x8_t c = vaddl_u8(vget_low_u8(b), vget_high_u8(b)); in vpx_avg_4x4_neon() 96 vec_sum_lo = vaddw_u8(vec_sum_lo, vget_low_u8(vec_row1)); in vpx_int_pro_row_neon() 99 vec_sum_lo = vaddw_u8(vec_sum_lo, vget_low_u8(vec_row2)); in vpx_int_pro_row_neon() 102 vec_sum_lo = vaddw_u8(vec_sum_lo, vget_low_u8(vec_row3)); in vpx_int_pro_row_neon() 105 vec_sum_lo = vaddw_u8(vec_sum_lo, vget_low_u8(vec_row4)); in vpx_int_pro_row_neon() 108 vec_sum_lo = vaddw_u8(vec_sum_lo, vget_low_u8(vec_row5)); in vpx_int_pro_row_neon() 111 vec_sum_lo = vaddw_u8(vec_sum_lo, vget_low_u8(vec_row6)); in vpx_int_pro_row_neon() 114 vec_sum_lo = vaddw_u8(vec_sum_lo, vget_low_u8(vec_row7)); in vpx_int_pro_row_neon() 117 vec_sum_lo = vaddw_u8(vec_sum_lo, vget_low_u8(vec_row8)); in vpx_int_pro_row_neon() 137 vec_sum = vaddw_u8(vec_sum, vget_low_u8(vec_row)); in vpx_int_pro_col_neon() [all …]
|
D | vpx_scaled_convolve8_neon.c | 244 s[0] = vget_low_u8(ss[0]); in scaledconvolve_vert_w16() 245 s[1] = vget_low_u8(ss[1]); in scaledconvolve_vert_w16() 246 s[2] = vget_low_u8(ss[2]); in scaledconvolve_vert_w16() 247 s[3] = vget_low_u8(ss[3]); in scaledconvolve_vert_w16() 248 s[4] = vget_low_u8(ss[4]); in scaledconvolve_vert_w16() 249 s[5] = vget_low_u8(ss[5]); in scaledconvolve_vert_w16() 250 s[6] = vget_low_u8(ss[6]); in scaledconvolve_vert_w16() 251 s[7] = vget_low_u8(ss[7]); in scaledconvolve_vert_w16()
|
D | loopfilter_neon.c | 187 *sum0 = vsubw_u8(*sum0, vget_low_u8(sub0)); in filter_update_16() 189 *sum0 = vsubw_u8(*sum0, vget_low_u8(sub1)); in filter_update_16() 191 *sum0 = vaddw_u8(*sum0, vget_low_u8(add0)); in filter_update_16() 193 *sum0 = vaddw_u8(*sum0, vget_low_u8(add1)); in filter_update_16() 261 sum0 = vaddl_u8(vget_low_u8(p3), vget_low_u8(p3)); // 2*p3 in calc_7_tap_filter_16() 263 sum0 = vaddw_u8(sum0, vget_low_u8(p3)); // 3*p3 in calc_7_tap_filter_16() 265 sum0 = vaddw_u8(sum0, vget_low_u8(p2)); // 3*p3+p2 in calc_7_tap_filter_16() 267 sum0 = vaddw_u8(sum0, vget_low_u8(p2)); // 3*p3+2*p2 in calc_7_tap_filter_16() 269 sum0 = vaddw_u8(sum0, vget_low_u8(p1)); // 3*p3+2*p2+p1 in calc_7_tap_filter_16() 271 sum0 = vaddw_u8(sum0, vget_low_u8(p0)); // 3*p3+2*p2+p1+p0 in calc_7_tap_filter_16() [all …]
|
/external/libaom/libaom/aom_dsp/arm/ |
D | sad4d_neon.c | 46 *vec_sum_ref_lo = vabal_u8(*vec_sum_ref_lo, vget_low_u8(vec_src_00), in sad_neon_64() 47 vget_low_u8(vec_ref_00)); in sad_neon_64() 50 *vec_sum_ref_lo = vabal_u8(*vec_sum_ref_lo, vget_low_u8(vec_src_16), in sad_neon_64() 51 vget_low_u8(vec_ref_16)); in sad_neon_64() 54 *vec_sum_ref_lo = vabal_u8(*vec_sum_ref_lo, vget_low_u8(vec_src_32), in sad_neon_64() 55 vget_low_u8(vec_ref_32)); in sad_neon_64() 58 *vec_sum_ref_lo = vabal_u8(*vec_sum_ref_lo, vget_low_u8(vec_src_48), in sad_neon_64() 59 vget_low_u8(vec_ref_48)); in sad_neon_64() 73 *vec_sum_ref_lo = vabal_u8(*vec_sum_ref_lo, vget_low_u8(vec_src_00), in sad_neon_32() 74 vget_low_u8(vec_ref_00)); in sad_neon_32() [all …]
|
D | sad_neon.c | 90 q12 = vabdl_u8(vget_low_u8(q0), vget_low_u8(q4)); in aom_sad16x8_neon() 98 q12 = vabal_u8(q12, vget_low_u8(q0), vget_low_u8(q4)); in aom_sad16x8_neon() 147 vec_accum_lo = vabal_u8(vec_accum_lo, vget_low_u8(vec_src_00), in aom_sad64x64_neon() 148 vget_low_u8(vec_ref_00)); in aom_sad64x64_neon() 151 vec_accum_lo = vabal_u8(vec_accum_lo, vget_low_u8(vec_src_16), in aom_sad64x64_neon() 152 vget_low_u8(vec_ref_16)); in aom_sad64x64_neon() 155 vec_accum_lo = vabal_u8(vec_accum_lo, vget_low_u8(vec_src_32), in aom_sad64x64_neon() 156 vget_low_u8(vec_ref_32)); in aom_sad64x64_neon() 159 vec_accum_lo = vabal_u8(vec_accum_lo, vget_low_u8(vec_src_48), in aom_sad64x64_neon() 160 vget_low_u8(vec_ref_48)); in aom_sad64x64_neon() [all …]
|
D | subtract_neon.c | 32 vsubl_u8(vget_low_u8(v_src_00), vget_low_u8(v_pred_00)); in aom_subtract_block_neon() 36 vsubl_u8(vget_low_u8(v_src_16), vget_low_u8(v_pred_16)); in aom_subtract_block_neon() 53 vsubl_u8(vget_low_u8(v_src), vget_low_u8(v_pred)); in aom_subtract_block_neon()
|
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() 325 vpaddl_u8(vget_low_u8(t0)), vpaddl_u8(vget_high_u8(t0)))); in aom_lowbd_blend_a64_d16_mask_neon() 327 vpaddl_u8(vget_low_u8(t1)), vpaddl_u8(vget_high_u8(t1)))); in aom_lowbd_blend_a64_d16_mask_neon() 329 vpaddl_u8(vget_low_u8(t2)), vpaddl_u8(vget_high_u8(t2)))); in aom_lowbd_blend_a64_d16_mask_neon() 331 vpaddl_u8(vget_low_u8(t3)), vpaddl_u8(vget_high_u8(t3)))); in aom_lowbd_blend_a64_d16_mask_neon()
|
/external/libjpeg-turbo/simd/arm/ |
D | jdsample-neon.c | 88 vmlal_u8(vmovl_u8(vget_low_u8(s1)), vget_low_u8(s0), three_u8); in jsimd_h2v1_fancy_upsample_neon() 92 vmlal_u8(vmovl_u8(vget_low_u8(s0)), vget_low_u8(s1), three_u8); in jsimd_h2v1_fancy_upsample_neon() 127 vmlal_u8(vmovl_u8(vget_low_u8(s1)), vget_low_u8(s0), three_u8); in jsimd_h2v1_fancy_upsample_neon() 131 vmlal_u8(vmovl_u8(vget_low_u8(s0)), vget_low_u8(s1), three_u8); in jsimd_h2v1_fancy_upsample_neon() 247 uint16x8_t s0colsum0_l = vmlal_u8(vmovl_u8(vget_low_u8(s0A)), in jsimd_h2v2_fancy_upsample_neon() 248 vget_low_u8(s0B), three_u8); in jsimd_h2v2_fancy_upsample_neon() 251 uint16x8_t s0colsum1_l = vmlal_u8(vmovl_u8(vget_low_u8(s0C)), in jsimd_h2v2_fancy_upsample_neon() 252 vget_low_u8(s0B), three_u8); in jsimd_h2v2_fancy_upsample_neon() 259 uint16x8_t s1colsum0_l = vmlal_u8(vmovl_u8(vget_low_u8(s1A)), in jsimd_h2v2_fancy_upsample_neon() 260 vget_low_u8(s1B), three_u8); in jsimd_h2v2_fancy_upsample_neon() [all …]
|
D | jcsample-neon.c | 112 uint8x8x2_t table = { { vget_low_u8(pixels), vget_high_u8(pixels) } }; in jsimd_h2v1_downsample_neon() 113 pixels = vcombine_u8(vtbl2_u8(table, vget_low_u8(expand_mask)), in jsimd_h2v1_downsample_neon() 176 { { vget_low_u8(pixels_r0), vget_high_u8(pixels_r0) } }; in jsimd_h2v2_downsample_neon() 178 { { vget_low_u8(pixels_r1), vget_high_u8(pixels_r1) } }; in jsimd_h2v2_downsample_neon() 179 pixels_r0 = vcombine_u8(vtbl2_u8(table_r0, vget_low_u8(expand_mask)), in jsimd_h2v2_downsample_neon() 181 pixels_r1 = vcombine_u8(vtbl2_u8(table_r1, vget_low_u8(expand_mask)), in jsimd_h2v2_downsample_neon()
|
/external/libvpx/libvpx/vp8/common/arm/neon/ |
D | sixtappredict_neon.c | 50 const uint32x2x2_t a_shuf = vzip_u32(vreinterpret_u32_u8(vget_low_u8(a)), in filter_add_accumulate() 52 const uint32x2x2_t b_shuf = vzip_u32(vreinterpret_u32_u8(vget_low_u8(b)), in filter_add_accumulate() 61 const uint32x2x2_t a_shuf = vzip_u32(vreinterpret_u32_u8(vget_low_u8(a)), in filter_sub_accumulate() 63 const uint32x2x2_t b_shuf = vzip_u32(vreinterpret_u32_u8(vget_low_u8(b)), in filter_sub_accumulate() 216 s0_f5 = vext_u8(vget_low_u8(s0), vget_high_u8(s0), 5); in vp8_sixtap_predict4x4_neon() 217 s1_f5 = vext_u8(vget_low_u8(s1), vget_high_u8(s1), 5); in vp8_sixtap_predict4x4_neon() 218 s2_f5 = vext_u8(vget_low_u8(s2), vget_high_u8(s2), 5); in vp8_sixtap_predict4x4_neon() 219 s3_f5 = vext_u8(vget_low_u8(s3), vget_high_u8(s3), 5); in vp8_sixtap_predict4x4_neon() 221 s01_f0 = vcombine_u8(vget_low_u8(s0), vget_low_u8(s1)); in vp8_sixtap_predict4x4_neon() 222 s23_f0 = vcombine_u8(vget_low_u8(s2), vget_low_u8(s3)); in vp8_sixtap_predict4x4_neon() [all …]
|
D | bilinearpredict_neon.c | 75 c0 = vzip_u32(vreinterpret_u32_u8(vget_low_u8(a01)), in vp8_bilinear_predict4x4_neon() 77 c1 = vzip_u32(vreinterpret_u32_u8(vget_low_u8(a23)), in vp8_bilinear_predict4x4_neon() 79 c2 = vzip_u32(vreinterpret_u32_u8(vget_low_u8(b01)), in vp8_bilinear_predict4x4_neon() 81 c3 = vzip_u32(vreinterpret_u32_u8(vget_low_u8(b23)), in vp8_bilinear_predict4x4_neon() 155 q6u16 = vmull_u8(vget_low_u8(q1u8), d0u8); in vp8_bilinear_predict8x4_neon() 156 q7u16 = vmull_u8(vget_low_u8(q2u8), d0u8); in vp8_bilinear_predict8x4_neon() 157 q8u16 = vmull_u8(vget_low_u8(q3u8), d0u8); in vp8_bilinear_predict8x4_neon() 158 q9u16 = vmull_u8(vget_low_u8(q4u8), d0u8); in vp8_bilinear_predict8x4_neon() 159 q10u16 = vmull_u8(vget_low_u8(q5u8), d0u8); in vp8_bilinear_predict8x4_neon() 161 d3u8 = vext_u8(vget_low_u8(q1u8), vget_high_u8(q1u8), 1); in vp8_bilinear_predict8x4_neon() [all …]
|
D | mbloopfilter_neon.c | 264 vst1_u8(u, vget_low_u8(q4)); in vp8_mbloop_filter_horizontal_edge_uv_neon() 268 vst1_u8(u, vget_low_u8(q5)); in vp8_mbloop_filter_horizontal_edge_uv_neon() 272 vst1_u8(u, vget_low_u8(q6)); in vp8_mbloop_filter_horizontal_edge_uv_neon() 276 vst1_u8(u, vget_low_u8(q7)); in vp8_mbloop_filter_horizontal_edge_uv_neon() 280 vst1_u8(u, vget_low_u8(q8)); in vp8_mbloop_filter_horizontal_edge_uv_neon() 284 vst1_u8(u, vget_low_u8(q9)); in vp8_mbloop_filter_horizontal_edge_uv_neon() 418 vst1_u8(s1, vget_low_u8(q3)); in vp8_mbloop_filter_vertical_edge_y_neon() 422 vst1_u8(s1, vget_low_u8(q4)); in vp8_mbloop_filter_vertical_edge_y_neon() 426 vst1_u8(s1, vget_low_u8(q5)); in vp8_mbloop_filter_vertical_edge_y_neon() 430 vst1_u8(s1, vget_low_u8(q6)); in vp8_mbloop_filter_vertical_edge_y_neon() [all …]
|
/external/libhevc/encoder/arm/ |
D | ihevce_sad_compute_neon.c | 61 uint16x8_t abs = vabdl_u8(vget_low_u8(src_u8), vget_low_u8(ref_u8)); in ihevce_4x4_sad_computer_neon() 116 abs_0 = vabal_u8(abs_0, vget_low_u8(src), vget_low_u8(pred)); in ihevce_16xn_sad_computer_neon() 148 abs_0 = vabal_u8(abs_0, vget_low_u8(src_0), vget_low_u8(pred_0)); in ihevce_32xn_sad_computer_neon() 150 abs_1 = vabal_u8(abs_1, vget_low_u8(src_1), vget_low_u8(pred_1)); in ihevce_32xn_sad_computer_neon() 186 abs_0 = vabal_u8(abs_0, vget_low_u8(src_0), vget_low_u8(pred_0)); in ihevce_64xn_sad_computer_neon() 188 abs_0 = vabal_u8(abs_0, vget_low_u8(src_1), vget_low_u8(pred_1)); in ihevce_64xn_sad_computer_neon() 190 abs_1 = vabal_u8(abs_1, vget_low_u8(src_2), vget_low_u8(pred_2)); in ihevce_64xn_sad_computer_neon() 192 abs_1 = vabal_u8(abs_1, vget_low_u8(src_3), vget_low_u8(pred_3)); in ihevce_64xn_sad_computer_neon()
|
D | ihevce_ssd_calculator_neon.c | 75 sqabs_low = vmull_u8(vget_low_u8(abs), vget_low_u8(abs)); in ihevce_4x4_ssd_computer_neon() 127 sqabs_low = vmull_u8(vget_low_u8(abs), vget_low_u8(abs)); in ihevce_1x16_ssd_computer_neon() 159 sqabs_0 = vmull_u8(vget_low_u8(abs_0), vget_low_u8(abs_0)); in ihevce_1x32_ssd_computer_neon() 161 sqabs_2 = vmull_u8(vget_low_u8(abs_1), vget_low_u8(abs_1)); in ihevce_1x32_ssd_computer_neon() 211 sqabs_0 = vmull_u8(vget_low_u8(abs_0), vget_low_u8(abs_0)); in ihevce_1x64_ssd_computer_neon() 213 sqabs_2 = vmull_u8(vget_low_u8(abs_1), vget_low_u8(abs_1)); in ihevce_1x64_ssd_computer_neon() 215 sqabs_4 = vmull_u8(vget_low_u8(abs_2), vget_low_u8(abs_2)); in ihevce_1x64_ssd_computer_neon() 217 sqabs_6 = vmull_u8(vget_low_u8(abs_3), vget_low_u8(abs_3)); in ihevce_1x64_ssd_computer_neon()
|
D | ihevce_ssd_and_sad_calculator_neon.c | 69 const uint8x8_t abs_l = vabd_u8(vget_low_u8(src_u8), vget_low_u8(ref_u8)); in ihevce_ssd_and_sad_calculator_neon() 141 abs_l = vabdl_u8(vget_low_u8(src), vget_low_u8(pred)); in ihevce_ssd_and_sad_calculator_neon() 190 abs_l = vabd_u8(vget_low_u8(src_0), vget_low_u8(pred_0)); in ihevce_ssd_and_sad_calculator_neon() 201 abs_l = vabd_u8(vget_low_u8(src_1), vget_low_u8(pred_1)); in ihevce_ssd_and_sad_calculator_neon() 253 abs_l = vabd_u8(vget_low_u8(src_0), vget_low_u8(pred_0)); in ihevce_ssd_and_sad_calculator_neon() 264 abs_l = vabd_u8(vget_low_u8(src_1), vget_low_u8(pred_1)); in ihevce_ssd_and_sad_calculator_neon() 275 abs_l = vabd_u8(vget_low_u8(src_2), vget_low_u8(pred_2)); in ihevce_ssd_and_sad_calculator_neon() 286 abs_l = vabd_u8(vget_low_u8(src_3), vget_low_u8(pred_3)); in ihevce_ssd_and_sad_calculator_neon()
|
D | ihevce_scale_by_2_neon.c | 84 c = vext_u8(vget_low_u8(src.val[1]), vget_high_u8(src.val[1]), 1); in ihevce_horz_scale_neon_w16() 85 l0 = vext_u8(vget_low_u8(src.val[0]), vget_high_u8(src.val[0]), 1); in ihevce_horz_scale_neon_w16() 86 r0 = vext_u8(vget_low_u8(src.val[0]), vget_high_u8(src.val[0]), 2); in ihevce_horz_scale_neon_w16() 87 r3 = vext_u8(vget_low_u8(src.val[0]), vget_high_u8(src.val[0]), 3); 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() 162 p = vreinterpretq_s16_u16(vmull_u8(vget_low_u8(src[c]), wt_0)); in ihevce_vert_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()
|
D | ihevce_common_utils_neon.c | 124 a2 = vreinterpretq_s16_u16(vmovl_u8(vget_low_u8(a0))); in ihevce_wt_avg_2d_16x1_neon() 126 a4 = vreinterpretq_s16_u16(vmovl_u8(vget_low_u8(a1))); in ihevce_wt_avg_2d_16x1_neon() 250 a0 = vmovl_u8(vget_low_u8(src0_u8)); in ihevce_wt_avg_2d_4xn_neon() 252 a2 = vmovl_u8(vget_low_u8(src1_u8)); in ihevce_wt_avg_2d_4xn_neon() 541 vsubl_u8(vget_low_u8(src_buf_8x16), vget_low_u8(recon_buf_8x16))); in ihevce_get_luma_eo_sao_params_neon() 547 vsubl_u8(vget_low_u8(recon_buf_8x16), vget_low_u8(recon_buf0_8x16))); in ihevce_get_luma_eo_sao_params_neon() 553 vsubl_u8(vget_low_u8(recon_buf_8x16), vget_low_u8(recon_buf1_8x16))); in ihevce_get_luma_eo_sao_params_neon() 726 vsubl_u8(vget_low_u8(src_buf_8x16), vget_low_u8(recon_buf_8x16))); in ihevce_get_luma_eo_sao_params_neon() 731 vsubl_u8(vget_low_u8(recon_buf_8x16), vget_low_u8(recon_buf0_8x16))); in ihevce_get_luma_eo_sao_params_neon() 737 vsubl_u8(vget_low_u8(recon_buf_8x16), vget_low_u8(recon_buf1_8x16))); in ihevce_get_luma_eo_sao_params_neon() [all …]
|
/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() 80 sum_lo = vmlal_u8(sum_lo, vget_low_u8(src_1), krn1); in IntraEdgeFilter_NEON() 102 uint16x8_t sum_lo = vaddl_u8(vget_low_u8(src_0), vget_low_u8(src_2)); in IntraEdgeFilter_NEON() 104 sum_lo = vmlal_u8(sum_lo, vget_low_u8(src_1), krn1); 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() 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() 255 const uint8x8_t src1 = vget_low_u8(src); in IntraEdgeUpsampler_NEON() 271 const uint8x8_t result_lo = Upsample(vget_low_u8(src0), vget_low_u8(src1), in IntraEdgeUpsampler_NEON() [all …]
|
D | super_res_neon.cc | 74 res = vmlal_u8(res, src[4], vget_low_u8(f[2])); in SuperRes() 75 res = vmlal_u8(res, src[6], vget_low_u8(f[3])); in SuperRes() 76 uint16x8_t temp = vmull_u8(src[0], vget_low_u8(f[0])); in SuperRes() 77 temp = vmlal_u8(temp, src[2], vget_low_u8(f[1])); in SuperRes() 115 sr[0] = vget_low_u8(s[0]); in SuperRes_NEON() 116 sr[1] = vget_low_u8(s[1]); in SuperRes_NEON() 117 sr[2] = vget_low_u8(s[2]); in SuperRes_NEON() 118 sr[3] = vget_low_u8(s[3]); in SuperRes_NEON() 119 sr[4] = vget_low_u8(s[4]); in SuperRes_NEON() 120 sr[5] = vget_low_u8(s[5]); in SuperRes_NEON() [all …]
|
/external/rust/crates/libz-sys/src/zlib-ng/arch/arm/ |
D | adler32_neon.c | 36 sum2 = vmull_u8( vget_low_u8(t0), vget_low_u8(d0)); in NEON_accum32() 38 sum2 = vmlal_u8(sum2, vget_low_u8(t1), vget_low_u8(d1)); in NEON_accum32() 51 sum2 = vmull_u8( vget_low_u8(t1), vget_low_u8(d0)); in NEON_accum32()
|
/external/XNNPACK/src/u8-clamp/ |
D | neon-x64.c | 45 vout = vmin_u8(vout, vget_low_u8(voutput_max)); in xnn_u8_clamp_ukernel__neon_x64() 46 vout = vmax_u8(vout, vget_low_u8(voutput_min)); in xnn_u8_clamp_ukernel__neon_x64() 51 vout = vmin_u8(vout, vget_low_u8(voutput_max)); in xnn_u8_clamp_ukernel__neon_x64() 52 vout = vmax_u8(vout, vget_low_u8(voutput_min)); in xnn_u8_clamp_ukernel__neon_x64()
|
/external/XNNPACK/src/qu8-vadd/ |
D | minmax-neon.c | 39 const int16x8_t vxa0 = vreinterpretq_s16_u16(vsubl_u8(vget_low_u8(va01), va_zero_point)); in xnn_qu8_vadd_minmax_ukernel__neon() 40 const int16x8_t vxb0 = vreinterpretq_s16_u16(vsubl_u8(vget_low_u8(vb01), vb_zero_point)); in xnn_qu8_vadd_minmax_ukernel__neon() 43 const int16x8_t vxa2 = vreinterpretq_s16_u16(vsubl_u8(vget_low_u8(va23), va_zero_point)); in xnn_qu8_vadd_minmax_ukernel__neon() 44 const int16x8_t vxb2 = vreinterpretq_s16_u16(vsubl_u8(vget_low_u8(vb23), vb_zero_point)); in xnn_qu8_vadd_minmax_ukernel__neon() 109 const int16x8_t vxa0 = vreinterpretq_s16_u16(vsubl_u8(vget_low_u8(va01), va_zero_point)); in xnn_qu8_vadd_minmax_ukernel__neon() 110 const int16x8_t vxb0 = vreinterpretq_s16_u16(vsubl_u8(vget_low_u8(vb01), vb_zero_point)); in xnn_qu8_vadd_minmax_ukernel__neon() 188 vy = vmax_u8(vy, vget_low_u8(vy_min)); in xnn_qu8_vadd_minmax_ukernel__neon() 189 vy = vmin_u8(vy, vget_low_u8(vy_max)); in xnn_qu8_vadd_minmax_ukernel__neon() 231 vy = vmax_u8(vy, vget_low_u8(vy_min)); in xnn_qu8_vadd_minmax_ukernel__neon() 232 vy = vmin_u8(vy, vget_low_u8(vy_max)); in xnn_qu8_vadd_minmax_ukernel__neon()
|
/external/skqp/src/opts/ |
D | Sk4px_NEON.h | 11 return Sk16h(vmovl_u8(vget_low_u8 (this->fVec)), in widen() 16 return Sk16h(vmull_u8(vget_low_u8 (this->fVec), vget_low_u8 (other.fVec)), in mulWiden()
|