/external/libvpx/libvpx/vp8/common/arm/neon/ |
D | bilinearpredict_neon.c | 84 d0 = vmull_u8(vreinterpret_u8_u32(c0.val[0]), filter0); in vp8_bilinear_predict4x4_neon() 85 d1 = vmull_u8(vreinterpret_u8_u32(c1.val[0]), filter0); in vp8_bilinear_predict4x4_neon() 86 d2 = vmull_u8(a4, filter0); in vp8_bilinear_predict4x4_neon() 105 uint16x8_t b0 = vmull_u8(e0, filter0); in vp8_bilinear_predict4x4_neon() 106 uint16x8_t b1 = vmull_u8(e1, filter0); 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() [all …]
|
D | sixtappredict_neon.c | 125 c0 = vmull_u8(b0, filter0); in yonly4x4() 126 c1 = vmull_u8(b2, filter0); in yonly4x4() 127 c2 = vmull_u8(b5, filter5); in yonly4x4() 128 c3 = vmull_u8(b7, filter5); in yonly4x4() 226 d0123 = vmull_u8(vreinterpret_u8_u32(s01_f5_q.val[0]), filter5); in vp8_sixtap_predict4x4_neon() 227 d4567 = vmull_u8(vreinterpret_u8_u32(s23_f5_q.val[0]), filter5); in vp8_sixtap_predict4x4_neon() 260 d0123_a = vmull_u8(vreinterpret_u8_u32(s01_f3_q.val[0]), filter3); in vp8_sixtap_predict4x4_neon() 261 d4567_a = vmull_u8(vreinterpret_u8_u32(s23_f3_q.val[0]), filter3); in vp8_sixtap_predict4x4_neon() 302 d0123 = vmull_u8(vreinterpret_u8_u32(s01_f5_q.val[0]), filter5); in vp8_sixtap_predict4x4_neon() 303 d4567 = vmull_u8(vreinterpret_u8_u32(s23_f5_q.val[0]), filter5); in vp8_sixtap_predict4x4_neon() [all …]
|
/external/libhevc/common/arm/ |
D | ihevc_intra_pred_chroma_mode_27_to_33.s | 179 vmull.u8 q5,d8,d30 @(i row)vmull_u8(ref_main_idx, dup_const_32_fract) 182 vmlal.u8 q5,d9,d31 @(i row)vmull_u8(ref_main_idx_1, dup_const_fract) 190 vmull.u8 q7,d12,d28 @(ii)vmull_u8(ref_main_idx, dup_const_32_fract) 194 vmlal.u8 q7,d13,d29 @(ii)vmull_u8(ref_main_idx_1, dup_const_fract) 205 vmull.u8 q9,d16,d26 @(iii)vmull_u8(ref_main_idx, dup_const_32_fract) 208 vmlal.u8 q9,d17,d27 @(iii)vmull_u8(ref_main_idx_1, dup_const_fract) 222 vmull.u8 q11,d20,d24 @(iv)vmull_u8(ref_main_idx, dup_const_32_fract) 226 vmlal.u8 q11,d21,d25 @(iv)vmull_u8(ref_main_idx_1, dup_const_fract) 240 vmull.u8 q5,d8,d30 @(v)vmull_u8(ref_main_idx, dup_const_32_fract) 243 vmlal.u8 q5,d9,d31 @(v)vmull_u8(ref_main_idx_1, dup_const_fract) [all …]
|
D | ihevc_intra_pred_luma_mode_27_to_33.s | 181 vmull.u8 q5,d8,d30 @(i row)vmull_u8(ref_main_idx, dup_const_32_fract) 184 vmlal.u8 q5,d9,d31 @(i row)vmull_u8(ref_main_idx_1, dup_const_fract) 192 vmull.u8 q7,d12,d28 @(ii)vmull_u8(ref_main_idx, dup_const_32_fract) 196 vmlal.u8 q7,d13,d29 @(ii)vmull_u8(ref_main_idx_1, dup_const_fract) 207 vmull.u8 q9,d16,d26 @(iii)vmull_u8(ref_main_idx, dup_const_32_fract) 210 vmlal.u8 q9,d17,d27 @(iii)vmull_u8(ref_main_idx_1, dup_const_fract) 223 vmull.u8 q11,d20,d24 @(iv)vmull_u8(ref_main_idx, dup_const_32_fract) 227 vmlal.u8 q11,d21,d25 @(iv)vmull_u8(ref_main_idx_1, dup_const_fract) 241 vmull.u8 q5,d8,d30 @(v)vmull_u8(ref_main_idx, dup_const_32_fract) 244 vmlal.u8 q5,d9,d31 @(v)vmull_u8(ref_main_idx_1, dup_const_fract) [all …]
|
D | ihevc_intra_pred_filters_luma_mode_19_to_25.s | 291 vmull.u8 q5,d8,d30 @(i row)vmull_u8(ref_main_idx, dup_const_32_fract) 294 vmlal.u8 q5,d9,d31 @(i row)vmull_u8(ref_main_idx_1, dup_const_fract) 301 vmull.u8 q7,d12,d28 @(ii)vmull_u8(ref_main_idx, dup_const_32_fract) 305 vmlal.u8 q7,d13,d29 @(ii)vmull_u8(ref_main_idx_1, dup_const_fract) 316 vmull.u8 q9,d16,d26 @(iii)vmull_u8(ref_main_idx, dup_const_32_fract) 319 vmlal.u8 q9,d17,d27 @(iii)vmull_u8(ref_main_idx_1, dup_const_fract) 331 vmull.u8 q11,d20,d24 @(iv)vmull_u8(ref_main_idx, dup_const_32_fract) 335 vmlal.u8 q11,d21,d25 @(iv)vmull_u8(ref_main_idx_1, dup_const_fract) 348 vmull.u8 q5,d8,d30 @(v)vmull_u8(ref_main_idx, dup_const_32_fract) 351 vmlal.u8 q5,d9,d31 @(v)vmull_u8(ref_main_idx_1, dup_const_fract) [all …]
|
D | ihevc_intra_pred_filters_chroma_mode_19_to_25.s | 288 vmull.u8 q5,d8,d30 @(i row)vmull_u8(ref_main_idx, dup_const_32_fract) 291 vmlal.u8 q5,d9,d31 @(i row)vmull_u8(ref_main_idx_1, dup_const_fract) 298 vmull.u8 q7,d12,d28 @(ii)vmull_u8(ref_main_idx, dup_const_32_fract) 302 vmlal.u8 q7,d13,d29 @(ii)vmull_u8(ref_main_idx_1, dup_const_fract) 313 vmull.u8 q9,d16,d26 @(iii)vmull_u8(ref_main_idx, dup_const_32_fract) 316 vmlal.u8 q9,d17,d27 @(iii)vmull_u8(ref_main_idx_1, dup_const_fract) 329 vmull.u8 q11,d20,d24 @(iv)vmull_u8(ref_main_idx, dup_const_32_fract) 333 vmlal.u8 q11,d21,d25 @(iv)vmull_u8(ref_main_idx_1, dup_const_fract) 346 vmull.u8 q5,d8,d30 @(v)vmull_u8(ref_main_idx, dup_const_32_fract) 349 vmlal.u8 q5,d9,d31 @(v)vmull_u8(ref_main_idx_1, dup_const_fract) [all …]
|
D | ihevc_inter_pred_chroma_horz.s | 175 vmull.u8 q15,d2,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 212 vmull.u8 q11,d10,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 238 vmull.u8 q10,d11,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 250 vmull.u8 q15,d2,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 297 vmull.u8 q11,d10,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 311 vmull.u8 q10,d11,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 319 vmull.u8 q15,d2,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 349 vmull.u8 q11,d10,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 355 vmull.u8 q10,d11,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 405 vmull.u8 q4,d1,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ [all …]
|
D | ihevc_inter_pred_chroma_horz_w16out.s | 194 vmull.u8 q15,d2,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 225 vmull.u8 q11,d10,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 253 vmull.u8 q10,d11,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 266 vmull.u8 q15,d2,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 306 vmull.u8 q11,d10,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 320 vmull.u8 q10,d11,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 328 vmull.u8 q15,d2,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 351 vmull.u8 q11,d10,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 357 vmull.u8 q10,d11,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ 402 vmull.u8 q4,d1,d25 @mul_res = vmull_u8(src[0_3], coeffabs_3)@ [all …]
|
D | ihevc_intra_pred_filters_neon_intr.c | 742 prod_t1 = vmull_u8(const_nt_1_col_t, pu1_ref_two_nt_1_row_dup); in ihevc_intra_pred_luma_planar_neonintr() 746 prod_t2 = vmull_u8(const_col_1_t, pu1_ref_three_nt_1_dup); in ihevc_intra_pred_luma_planar_neonintr() 822 prod_t1 = vmull_u8(const_nt_1_col_t, pu1_ref_two_nt_1_row_dup); in ihevc_intra_pred_luma_planar_neonintr() 829 prod_t2 = vmull_u8(const_col_1_t, pu1_ref_three_nt_1_dup); in ihevc_intra_pred_luma_planar_neonintr() 2147 mul_res1 = vmull_u8(ref_main_idx, dup_const_32_fract); in ihevc_intra_pred_luma_mode_3_to_9_neonintr() 2148 mul_res2 = vmull_u8(ref_main_idx_1, dup_const_fract); in ihevc_intra_pred_luma_mode_3_to_9_neonintr() 2232 mul_res1 = vmull_u8(vreinterpret_u8_u32(pu1_ref_val1), dup_32_fract); in ihevc_intra_pred_luma_mode_3_to_9_neonintr() 2233 mul_res2 = vmull_u8(vreinterpret_u8_u32(pu1_ref_val2), dup_fract); in ihevc_intra_pred_luma_mode_3_to_9_neonintr() 2427 mul_res1 = vmull_u8(ref_main_idx, dup_const_32_fract); in ihevc_intra_pred_luma_mode_11_to_17_neonintr() 2428 mul_res2 = vmull_u8(ref_main_idx_1, dup_const_fract); in ihevc_intra_pred_luma_mode_11_to_17_neonintr() [all …]
|
D | ihevc_inter_pred_filters_luma_vert.s | 164 vmull.u8 q4,d1,d23 @mul_res1 = vmull_u8(src_tmp2, coeffabs_1)@ 182 vmull.u8 q5,d2,d23 @mul_res2 = vmull_u8(src_tmp3, coeffabs_1)@ 247 vmull.u8 q4,d1,d23 @mul_res1 = vmull_u8(src_tmp2, coeffabs_1)@ 273 vmull.u8 q5,d2,d23 @mul_res2 = vmull_u8(src_tmp3, coeffabs_1)@ 353 vmull.u8 q4,d1,d23 @mul_res1 = vmull_u8(src_tmp2, coeffabs_1)@ 366 vmull.u8 q5,d2,d23 @mul_res2 = vmull_u8(src_tmp3, coeffabs_1)@ 440 …vmull.u8 q0,d5,d23 @mul_res1 = vmull_u8(vreinterpret_u8_u32(src_tmp2), coeffa… 453 …vmull.u8 q1,d7,d25 @mul_res2 = vmull_u8(vreinterpret_u8_u32(src_tmp4), coeffa… 618 vmull.u8 q4,d1,d23 @mul_res1 = vmull_u8(src_tmp2, coeffabs_1)@ 636 vmull.u8 q5,d2,d23 @mul_res2 = vmull_u8(src_tmp3, coeffabs_1)@ [all …]
|
D | ihevc_inter_pred_filters_luma_vert_w16inp.s | 153 vmull.s16 q4,d1,d23 @mul_res1 = vmull_u8(src_tmp2, coeffabs_1)@ 169 vmull.s16 q5,d2,d23 @mul_res2 = vmull_u8(src_tmp3, coeffabs_1)@ 225 vmull.s16 q4,d1,d23 @mul_res1 = vmull_u8(src_tmp2, coeffabs_1)@ 241 vmull.s16 q5,d2,d23 @mul_res2 = vmull_u8(src_tmp3, coeffabs_1)@ 310 vmull.s16 q4,d1,d23 @mul_res1 = vmull_u8(src_tmp2, coeffabs_1)@ 324 vmull.s16 q5,d2,d23 @mul_res2 = vmull_u8(src_tmp3, coeffabs_1)@
|
/external/libhevc/encoder/arm/ |
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() 76 sqabs_high = vmull_u8(vget_high_u8(abs), vget_high_u8(abs)); in ihevce_4x4_ssd_computer_neon() 102 sqabs = vmull_u8(abs, abs); in ihevce_1x8_ssd_computer_neon() 127 sqabs_low = vmull_u8(vget_low_u8(abs), vget_low_u8(abs)); in ihevce_1x16_ssd_computer_neon() 128 sqabs_high = vmull_u8(vget_high_u8(abs), vget_high_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() 160 sqabs_1 = vmull_u8(vget_high_u8(abs_0), vget_high_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() 162 sqabs_3 = vmull_u8(vget_high_u8(abs_1), vget_high_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() [all …]
|
D | ihevce_ssd_and_sad_calculator_neon.c | 71 const uint16x8_t sq_abs_l = vmull_u8(abs_l, abs_l); in ihevce_ssd_and_sad_calculator_neon() 72 const uint16x8_t sq_abs_h = vmull_u8(abs_h, abs_h); in ihevce_ssd_and_sad_calculator_neon() 194 sqabs_l = vmull_u8(abs_l, abs_l); in ihevce_ssd_and_sad_calculator_neon() 195 sqabs_h = vmull_u8(abs_h, abs_h); in ihevce_ssd_and_sad_calculator_neon() 205 sqabs_l = vmull_u8(abs_l, abs_l); in ihevce_ssd_and_sad_calculator_neon() 206 sqabs_h = vmull_u8(abs_h, abs_h); in ihevce_ssd_and_sad_calculator_neon() 257 sqabs_l = vmull_u8(abs_l, abs_l); in ihevce_ssd_and_sad_calculator_neon() 258 sqabs_h = vmull_u8(abs_h, abs_h); in ihevce_ssd_and_sad_calculator_neon() 268 sqabs_l = vmull_u8(abs_l, abs_l); in ihevce_ssd_and_sad_calculator_neon() 269 sqabs_h = vmull_u8(abs_h, abs_h); in ihevce_ssd_and_sad_calculator_neon() [all …]
|
D | ihevce_scale_by_2_neon.c | 89 p = vreinterpretq_s16_u16(vmull_u8(c, wt_0)); // a[0] * 66 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() 180 p = vreinterpretq_s16_u16(vmull_u8(vget_high_u8(src[c]), wt_0)); in ihevce_vert_scale_neon_w16()
|
/external/libgav1/libgav1/src/dsp/arm/ |
D | intrapred_smooth_neon.cc | 98 const uint16x8_t weighted_bl = vmull_u8(scaled_weights_y, bottom_left_v); in Smooth4Or8xN_NEON() 100 const uint16x8_t weighted_top = vmull_u8(weights_y_v, top_v); in Smooth4Or8xN_NEON() 101 const uint16x8_t weighted_left = vmull_u8(weights_x_v, left_v); in Smooth4Or8xN_NEON() 102 const uint16x8_t weighted_tr = vmull_u8(scaled_weights_x, top_right_v); in Smooth4Or8xN_NEON() 123 const uint16x8_t weighted_top_low = vmull_u8(weights_y, vget_low_u8(top)); in CalculateWeightsAndPred() 124 const uint16x8_t weighted_left_low = vmull_u8(vget_low_u8(weights_x), left); in CalculateWeightsAndPred() 126 vmull_u8(vget_low_u8(scaled_weights_x), top_right); in CalculateWeightsAndPred() 135 const uint16x8_t weighted_top_high = vmull_u8(weights_y, vget_high_u8(top)); in CalculateWeightsAndPred() 136 const uint16x8_t weighted_left_high = vmull_u8(vget_high_u8(weights_x), left); in CalculateWeightsAndPred() 138 vmull_u8(vget_high_u8(scaled_weights_x), top_right); in CalculateWeightsAndPred() [all …]
|
D | obmc_neon.cc | 44 const uint16x8_t weighted_pred = vmull_u8(pred_mask, pred_val); in WriteObmcLine4() 80 const uint16x8_t weighted_pred = vmull_u8(pred_mask, pred_val); in OverlapBlend2xH_NEON() 129 const uint16x8_t weighted_pred = vmull_u8(pred_mask, pred_val); in OverlapBlendFromLeft8xH_NEON() 177 vmull_u8(vget_low_u8(pred_mask), vget_low_u8(pred_val)); in OverlapBlendFromLeft_NEON() 183 vmull_u8(vget_high_u8(pred_mask), vget_high_u8(pred_val)); in OverlapBlendFromLeft_NEON() 299 const uint16x8_t weighted_pred = vmull_u8(pred_mask, pred_val); in OverlapBlendFromTop8xH_NEON() 349 vmull_u8(pred_mask, vget_low_u8(pred_val)); in OverlapBlendFromTop_NEON() 355 vmull_u8(pred_mask, vget_high_u8(pred_val)); in OverlapBlendFromTop_NEON()
|
D | loop_restoration_neon.cc | 915 sq[0] = vmull_u8(s[0], s[0]); in BoxSum() 921 sq[1] = vmull_u8(s[1], s[1]); in BoxSum() 952 sq[0] = vmull_u8(s[0], s[0]); in BoxSum() 958 sq[1] = vmull_u8(s[1], s[1]); in BoxSum() 1153 sq[0][0] = vmull_u8(vget_low_u8(s[0][0]), vget_low_u8(s[0][0])); in BoxFilterPreProcess5Lo() 1154 sq[1][0] = vmull_u8(vget_low_u8(s[1][0]), vget_low_u8(s[1][0])); in BoxFilterPreProcess5Lo() 1155 sq[0][1] = vmull_u8(vget_high_u8(s[0][0]), vget_high_u8(s[0][0])); in BoxFilterPreProcess5Lo() 1156 sq[1][1] = vmull_u8(vget_high_u8(s[1][0]), vget_high_u8(s[1][0])); in BoxFilterPreProcess5Lo() 1188 sq[0][2] = vmull_u8(vget_low_u8(s[0][1]), vget_low_u8(s[0][1])); in BoxFilterPreProcess5() 1189 sq[1][2] = vmull_u8(vget_low_u8(s[1][1]), vget_low_u8(s[1][1])); in BoxFilterPreProcess5() [all …]
|
/external/libaom/libaom/av1/common/arm/ |
D | blend_a64_hmask_neon.c | 51 res_low = vmull_u8(vget_low_u8(m_q), vget_low_u8(tmp0_q)); in aom_blend_a64_hmask_neon() 54 res_high = vmull_u8(vget_high_u8(m_q), vget_high_u8(tmp0_q)); in aom_blend_a64_hmask_neon() 78 res = vmull_u8(m, tmp0); in aom_blend_a64_hmask_neon() 97 res = vmull_u8(m, tmp0); in aom_blend_a64_hmask_neon() 121 res = vmull_u8(m, tmp0); in aom_blend_a64_hmask_neon()
|
D | blend_a64_vmask_neon.c | 49 res_low = vmull_u8(m, vget_low_u8(tmp0_q)); in aom_blend_a64_vmask_neon() 51 res_high = vmull_u8(m, vget_high_u8(tmp0_q)); in aom_blend_a64_vmask_neon() 72 res = vmull_u8(m, tmp0); in aom_blend_a64_vmask_neon() 96 res = vmull_u8(m, tmp0); in aom_blend_a64_vmask_neon() 128 res = vmull_u8(m, tmp0); in aom_blend_a64_vmask_neon()
|
/external/libvpx/libvpx/vpx_dsp/arm/ |
D | subpel_variance_neon.c | 39 const uint16x8_t a = vmull_u8(src_0, f0); in var_filter_block2d_bil_w4() 61 const uint16x8_t a = vmull_u8(src_0, f0); in var_filter_block2d_bil_w8() 85 const uint16x8_t a = vmull_u8(vget_low_u8(src_0), f0); in var_filter_block2d_bil_w16() 88 const uint16x8_t c = vmull_u8(vget_high_u8(src_0), f0); in var_filter_block2d_bil_w16()
|
/external/libaom/libaom/aom_dsp/arm/ |
D | subpel_variance_neon.c | 36 const uint16x8_t a = vmull_u8(src_0, f0); in var_filter_block2d_bil_w8() 60 const uint16x8_t a = vmull_u8(vget_low_u8(src_0), f0); in var_filter_block2d_bil_w16() 63 const uint16x8_t c = vmull_u8(vget_high_u8(src_0), f0); in var_filter_block2d_bil_w16()
|
/external/skqp/src/opts/ |
D | Sk4px_NEON.h | 16 return Sk16h(vmull_u8(vget_low_u8 (this->fVec), vget_low_u8 (other.fVec)), in mulWiden() 17 vmull_u8(vget_high_u8(this->fVec), vget_high_u8(other.fVec))); in mulWiden()
|
/external/skia/src/opts/ |
D | Sk4px_NEON.h | 16 return Sk16h(vmull_u8(vget_low_u8 (this->fVec), vget_low_u8 (other.fVec)), in mulWiden() 17 vmull_u8(vget_high_u8(this->fVec), vget_high_u8(other.fVec))); in mulWiden()
|
/external/webp/src/dsp/ |
D | alpha_processing_neon.c | 27 const uint16x8_t r1 = vmull_u8((V).val[1], alpha); \ 28 const uint16x8_t g1 = vmull_u8((V).val[2], alpha); \ 29 const uint16x8_t b1 = vmull_u8((V).val[(OTHER)], alpha); \
|
/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() 51 sum2 = vmull_u8( vget_low_u8(t1), vget_low_u8(d0)); in NEON_accum32()
|