/external/libjpeg-turbo/simd/arm/ |
D | jfdctint-neon.c | 198 z3_l = vaddq_s32(z3_l, z5_l); in jsimd_fdct_islow_neon() 199 z3_h = vaddq_s32(z3_h, z5_h); in jsimd_fdct_islow_neon() 200 z4_l = vaddq_s32(z4_l, z5_l); in jsimd_fdct_islow_neon() 201 z4_h = vaddq_s32(z4_h, z5_h); in jsimd_fdct_islow_neon() 203 tmp4_l = vaddq_s32(tmp4_l, z1_l); in jsimd_fdct_islow_neon() 204 tmp4_h = vaddq_s32(tmp4_h, z1_h); in jsimd_fdct_islow_neon() 205 tmp4_l = vaddq_s32(tmp4_l, z3_l); in jsimd_fdct_islow_neon() 206 tmp4_h = vaddq_s32(tmp4_h, z3_h); in jsimd_fdct_islow_neon() 210 tmp5_l = vaddq_s32(tmp5_l, z2_l); in jsimd_fdct_islow_neon() 211 tmp5_h = vaddq_s32(tmp5_h, z2_h); in jsimd_fdct_islow_neon() [all …]
|
D | jidctint-neon.c | 382 int32x4_t tmp10 = vaddq_s32(tmp0, tmp3); in jsimd_idct_islow_pass1_regular() 384 int32x4_t tmp11 = vaddq_s32(tmp1, tmp2); in jsimd_idct_islow_pass1_regular() 438 tmp0 = vaddq_s32(tmp0, z3); in jsimd_idct_islow_pass1_regular() 439 tmp1 = vaddq_s32(tmp1, z4); in jsimd_idct_islow_pass1_regular() 440 tmp2 = vaddq_s32(tmp2, z3); in jsimd_idct_islow_pass1_regular() 441 tmp3 = vaddq_s32(tmp3, z4); in jsimd_idct_islow_pass1_regular() 445 vrshrn_n_s32(vaddq_s32(tmp10, tmp3), DESCALE_P1), in jsimd_idct_islow_pass1_regular() 446 vrshrn_n_s32(vaddq_s32(tmp11, tmp2), DESCALE_P1), in jsimd_idct_islow_pass1_regular() 447 vrshrn_n_s32(vaddq_s32(tmp12, tmp1), DESCALE_P1), in jsimd_idct_islow_pass1_regular() 448 vrshrn_n_s32(vaddq_s32(tmp13, tmp0), DESCALE_P1) in jsimd_idct_islow_pass1_regular() [all …]
|
D | jidctred-neon.c | 124 row0 = vcombine_s16(vrshrn_n_s32(vaddq_s32(tmp10_l, tmp0_l), CONST_BITS), in jsimd_idct_2x2_neon() 125 vrshrn_n_s32(vaddq_s32(tmp10_h, tmp0_h), CONST_BITS)); in jsimd_idct_2x2_neon() 270 int32x4_t tmp10 = vaddq_s32(tmp0, tmp2); in jsimd_idct_4x4_neon() 290 row0 = vcombine_s16(dcval, vrshrn_n_s32(vaddq_s32(tmp10, tmp2), in jsimd_idct_4x4_neon() 294 row1 = vcombine_s16(dcval, vrshrn_n_s32(vaddq_s32(tmp12, tmp0), in jsimd_idct_4x4_neon() 323 int32x4_t tmp10 = vaddq_s32(tmp0, tmp2); in jsimd_idct_4x4_neon() 343 row0 = vcombine_s16(vrshrn_n_s32(vaddq_s32(tmp10, tmp2), in jsimd_idct_4x4_neon() 347 row1 = vcombine_s16(vrshrn_n_s32(vaddq_s32(tmp12, tmp0), in jsimd_idct_4x4_neon() 372 int32x4_t tmp10_l = vaddq_s32(tmp0_l, tmp2_l); in jsimd_idct_4x4_neon() 373 int32x4_t tmp10_h = vaddq_s32(tmp0_h, tmp2_h); in jsimd_idct_4x4_neon() [all …]
|
/external/libhevc/common/arm/ |
D | ihevc_resi_trans_neon.c | 141 e_03 = vaddq_s32(temp1, temp4); in ihevc_resi_trans_4x4_neon() 142 e_12 = vaddq_s32(temp3, temp2); in ihevc_resi_trans_4x4_neon() 146 e_0_a_e_1 = vaddq_s32(e_03, e_12); in ihevc_resi_trans_4x4_neon() 297 src1_4x32b = vaddq_s32(src1_4x32b, add_val); in ihevc_resi_trans_4x4_ttype1_neon() 298 src0_4x32b = vaddq_s32(src0_4x32b, add_val); in ihevc_resi_trans_4x4_ttype1_neon() 299 src2_4x32b = vaddq_s32(src2_4x32b, add_val); in ihevc_resi_trans_4x4_ttype1_neon() 300 src3_4x32b = vaddq_s32(src3_4x32b, add_val); in ihevc_resi_trans_4x4_ttype1_neon() 316 c0_4x32b = vaddq_s32(src0_4x32b, src3_4x32b); /* r0+r3 */ in ihevc_resi_trans_4x4_ttype1_neon() 317 c1_4x32b = vaddq_s32(src1_4x32b, src3_4x32b); /* r1+r3 */ in ihevc_resi_trans_4x4_ttype1_neon() 320 src1_4x32b = vaddq_s32(src0_4x32b, src1_4x32b); /* r0+r1 */ in ihevc_resi_trans_4x4_ttype1_neon() [all …]
|
D | ihevc_quant_iquant_ssd_neon_intr.c | 191 qtmp_0 = vaddq_s32(qtmp_0, add_q); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 192 qtmp_1 = vaddq_s32(qtmp_1, add_q); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 193 qtmp_2 = vaddq_s32(qtmp_2, add_q); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 194 qtmp_3 = vaddq_s32(qtmp_3, add_q); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 250 iqtmp_0 = vaddq_s32(iqtmp_0, add_iq); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 251 iqtmp_1 = vaddq_s32(iqtmp_1, add_iq); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 252 iqtmp_2 = vaddq_s32(iqtmp_2, add_iq); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 253 iqtmp_3 = vaddq_s32(iqtmp_3, add_iq); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 301 r0 = vaddq_s32(r0, r1); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() 302 r2 = vaddq_s32(r2, r3); in ihevc_quant_iquant_ssd_flat_scale_mat_neon() [all …]
|
D | ihevc_weighted_pred_neon_intr.c | 155 i4_tmp1_t = vaddq_s32(i4_tmp1_t, tmp_lvl_shift_t); in ihevc_weighted_pred_uni_neonintr() 159 i4_tmp2_t = vaddq_s32(i4_tmp2_t, tmp_lvl_shift_t); in ihevc_weighted_pred_uni_neonintr() 302 i4_tmp1_t = vaddq_s32(i4_tmp1_t, tmp_lvl_shift_t.val[0]); in ihevc_weighted_pred_chroma_uni_neonintr() 306 i4_tmp2_t = vaddq_s32(i4_tmp2_t, tmp_lvl_shift_t.val[0]); in ihevc_weighted_pred_chroma_uni_neonintr() 461 i4_tmp1_t1 = vaddq_s32(i4_tmp1_t1, i4_tmp1_t2); in ihevc_weighted_pred_bi_neonintr() 464 i4_tmp1_t1 = vaddq_s32(i4_tmp1_t1, tmp_lvl_shift_t); in ihevc_weighted_pred_bi_neonintr() 469 i4_tmp2_t1 = vaddq_s32(i4_tmp2_t1, i4_tmp2_t2); in ihevc_weighted_pred_bi_neonintr() 472 i4_tmp2_t1 = vaddq_s32(i4_tmp2_t1, tmp_lvl_shift_t); in ihevc_weighted_pred_bi_neonintr() 647 i4_tmp1_t1 = vaddq_s32(i4_tmp1_t1, i4_tmp1_t2); in ihevc_weighted_pred_chroma_bi_neonintr() 650 i4_tmp1_t1 = vaddq_s32(i4_tmp1_t1, tmp_lvl_shift_t.val[0]); in ihevc_weighted_pred_chroma_bi_neonintr() [all …]
|
D | ihevc_resi_trans_neon_32x32.c | 171 sum_val = vaddq_s32(sum_val,vpaddlq_s16(abs)); in ihevc_resi_trans_32x32_neon() 344 a[12].val[0] = vaddq_s32(a[12].val[0], a[4].val[0]); //q4 in ihevc_resi_trans_32x32_neon() 345 a[12].val[1] = vaddq_s32(a[12].val[1], a[4].val[1]); //q8 in ihevc_resi_trans_32x32_neon() 346 a[12].val[1] = vaddq_s32(a[12].val[1], a[12].val[0]); //q8 in ihevc_resi_trans_32x32_neon() 358 a[28].val[0] = vaddq_s32(a[28].val[0], a[20].val[0]); //q15 in ihevc_resi_trans_32x32_neon() 359 a[28].val[1] = vaddq_s32(a[28].val[1], a[20].val[1]); //q5 in ihevc_resi_trans_32x32_neon() 360 a[28].val[1] = vaddq_s32(a[28].val[1], a[28].val[0]); //q15 in ihevc_resi_trans_32x32_neon() 439 a[2].val[0] = vaddq_s32(a[2].val[0], a[6].val[0]); //q2 in ihevc_resi_trans_32x32_neon() 440 a[2].val[1] = vaddq_s32(a[2].val[1], a[6].val[1]); //q9 in ihevc_resi_trans_32x32_neon() 441 a[2].val[1] = vaddq_s32(a[2].val[1], a[2].val[0]); //q9 in ihevc_resi_trans_32x32_neon() [all …]
|
D | ihevc_weighted_pred_bi_default.s | 195 vqadd.s16 d18,d18,d0 @vaddq_s32(i4_tmp1_t1, tmp_lvl_shift_t) 197 vqadd.s16 d20,d8,d9 @vaddq_s32(i4_tmp2_t1, i4_tmp2_t2) 198 vqadd.s16 d19,d20,d0 @vaddq_s32(i4_tmp2_t1, tmp_lvl_shift_t) 203 vqadd.s16 d30,d30,d0 @vaddq_s32(i4_tmp1_t1, tmp_lvl_shift_t) iii iteration 206 vqadd.s16 d18,d24,d25 @vaddq_s32(i4_tmp2_t1, i4_tmp2_t2) iv iteration 211 …e pu1_dst iii iteration @vaddq_s32(i4_tmp2_t1, tmp_… 245 vqadd.s16 d18,d18,d0 @vaddq_s32(i4_tmp1_t1, tmp_lvl_shift_t) 247 vqadd.s16 d20,d8,d9 @vaddq_s32(i4_tmp2_t1, i4_tmp2_t2) 248 vqadd.s16 d19,d20,d0 @vaddq_s32(i4_tmp2_t1, tmp_lvl_shift_t) 285 vqadd.s16 q12,q12,q0 @vaddq_s32(i4_tmp1_t1, tmp_lvl_shift_t) [all …]
|
D | ihevc_weighted_pred_bi.s | 208 vadd.s32 q2,q2,q4 @vaddq_s32(i4_tmp1_t1, i4_tmp1_t2) 214 vadd.s32 q2,q2,q15 @vaddq_s32(i4_tmp1_t1, tmp_lvl_shift_t) 222 vadd.s32 q5,q5,q6 @vaddq_s32(i4_tmp2_t1, i4_tmp2_t2) ii iteration 227 vadd.s32 q5,q5,q15 @vaddq_s32(i4_tmp2_t1, tmp_lvl_shift_t) ii iteration 229 vadd.s32 q7,q7,q8 @vaddq_s32(i4_tmp1_t1, i4_tmp1_t2) iii iteration 234 vadd.s32 q7,q7,q15 @vaddq_s32(i4_tmp1_t1, tmp_lvl_shift_t) iii iteration 242 vadd.s32 q9,q9,q10 @vaddq_s32(i4_tmp2_t1, i4_tmp2_t2) iv iteration 245 vadd.s32 q9,q9,q15 @vaddq_s32(i4_tmp2_t1, tmp_lvl_shift_t) iv iteration
|
/external/libaom/libaom/av1/common/arm/ |
D | selfguided_neon.c | 283 r12 = vaddq_s32(d1, d2); in boxsum2_square_sum_calc() 284 r34 = vaddq_s32(d3, d4); in boxsum2_square_sum_calc() 285 r67 = vaddq_s32(d6, d7); in boxsum2_square_sum_calc() 286 r89 = vaddq_s32(d8, d9); in boxsum2_square_sum_calc() 287 r1011 = vaddq_s32(d10, d11); in boxsum2_square_sum_calc() 288 r345 = vaddq_s32(r34, d5); in boxsum2_square_sum_calc() 289 r6789 = vaddq_s32(r67, r89); in boxsum2_square_sum_calc() 291 *r0 = vaddq_s32(r12, r345); in boxsum2_square_sum_calc() 292 *r1 = vaddq_s32(r67, r345); in boxsum2_square_sum_calc() 293 *r2 = vaddq_s32(d5, r6789); in boxsum2_square_sum_calc() [all …]
|
D | convolve_neon.h | 98 sum_0 = vaddq_s32(sum_0, s3_0); in wiener_convolve8_horiz_8x8() 99 sum_1 = vaddq_s32(sum_1, s3_1); in wiener_convolve8_horiz_8x8() 102 sum_0 = vaddq_s32(sum_0, round_vec_0); in wiener_convolve8_horiz_8x8() 103 sum_1 = vaddq_s32(sum_1, round_vec_0); in wiener_convolve8_horiz_8x8() 147 sum_0 = vaddq_s32(sum_0, s3_0); in wiener_convolve8_horiz_4x8() 149 sum_0 = vaddq_s32(sum_0, round_vec_0); in wiener_convolve8_horiz_4x8() 220 sum0 = vaddq_s32(sum0, offset_const); in convolve8_4x4_s32()
|
/external/libvpx/libvpx/vpx_dsp/arm/ |
D | highbd_idct16x16_add_neon.c | 465 out[0].val[0] = vaddq_s32(step2[0].val[0], step2[15].val[0]); in highbd_idct16x16_add_stage7_dual() 466 out[0].val[1] = vaddq_s32(step2[0].val[1], step2[15].val[1]); in highbd_idct16x16_add_stage7_dual() 467 out[1].val[0] = vaddq_s32(step2[1].val[0], step2[14].val[0]); in highbd_idct16x16_add_stage7_dual() 468 out[1].val[1] = vaddq_s32(step2[1].val[1], step2[14].val[1]); in highbd_idct16x16_add_stage7_dual() 469 out[2].val[0] = vaddq_s32(step2[2].val[0], step2[13].val[0]); in highbd_idct16x16_add_stage7_dual() 470 out[2].val[1] = vaddq_s32(step2[2].val[1], step2[13].val[1]); in highbd_idct16x16_add_stage7_dual() 471 out[3].val[0] = vaddq_s32(step2[3].val[0], step2[12].val[0]); in highbd_idct16x16_add_stage7_dual() 472 out[3].val[1] = vaddq_s32(step2[3].val[1], step2[12].val[1]); in highbd_idct16x16_add_stage7_dual() 473 out[4].val[0] = vaddq_s32(step2[4].val[0], step2[11].val[0]); in highbd_idct16x16_add_stage7_dual() 474 out[4].val[1] = vaddq_s32(step2[4].val[1], step2[11].val[1]); in highbd_idct16x16_add_stage7_dual() [all …]
|
D | highbd_idct_neon.h | 43 b0 = vaddq_s32(a[0], a[2]); in idct4x4_16_kernel_bd10() 55 a[0] = vaddq_s32(b0, b3); in idct4x4_16_kernel_bd10() 56 a[1] = vaddq_s32(b1, b2); in idct4x4_16_kernel_bd10() 67 b0 = vaddq_s32(a[0], a[2]); in idct4x4_16_kernel_bd12() 93 a[0] = vaddq_s32(b0, b3); in idct4x4_16_kernel_bd12() 94 a[1] = vaddq_s32(b1, b2); in idct4x4_16_kernel_bd12() 206 step2[4] = vaddq_s32(step1[4], step1[5]); in idct8x8_64_half1d_bd10() 209 step2[7] = vaddq_s32(step1[7], step1[6]); in idct8x8_64_half1d_bd10() 212 step1[0] = vaddq_s32(step2[0], step2[3]); in idct8x8_64_half1d_bd10() 213 step1[1] = vaddq_s32(step2[1], step2[2]); in idct8x8_64_half1d_bd10() [all …]
|
D | highbd_idct8x8_add_neon.c | 98 step2[4] = vaddq_s32(step1[4], step1[5]); in idct8x8_12_half1d_bd10() 101 step2[7] = vaddq_s32(step1[7], step1[6]); in idct8x8_12_half1d_bd10() 104 step1[0] = vaddq_s32(step2[1], step2[3]); in idct8x8_12_half1d_bd10() 105 step1[1] = vaddq_s32(step2[1], step2[2]); in idct8x8_12_half1d_bd10() 116 *io0 = vaddq_s32(step1[0], step2[7]); in idct8x8_12_half1d_bd10() 117 *io1 = vaddq_s32(step1[1], step1[6]); in idct8x8_12_half1d_bd10() 118 *io2 = vaddq_s32(step1[2], step1[5]); in idct8x8_12_half1d_bd10() 119 *io3 = vaddq_s32(step1[3], step2[4]); in idct8x8_12_half1d_bd10() 187 step2[4] = vaddq_s32(step1[4], step1[5]); in idct8x8_12_half1d_bd12() 190 step2[7] = vaddq_s32(step1[7], step1[6]); in idct8x8_12_half1d_bd12() [all …]
|
D | variance_neon.c | 71 vaddq_s32(sse_lo_s32, sse_hi_s32))), in variance_neon_w4x4() 120 vaddq_s32(sse_lo_s32, sse_hi_s32))), in variance_neon_w16() 162 vaddq_s32(sse_lo_s32, sse_hi_s32))), in variance_neon_w8x2() 319 q7s32 = vaddq_s32(q7s32, q8s32); in vpx_mse16x16_neon() 320 q9s32 = vaddq_s32(q9s32, q10s32); in vpx_mse16x16_neon() 321 q10s32 = vaddq_s32(q7s32, q9s32); in vpx_mse16x16_neon() 372 q7s32 = vaddq_s32(q7s32, q8s32); in vpx_get4x4sse_cs_neon() 373 q9s32 = vaddq_s32(q9s32, q10s32); in vpx_get4x4sse_cs_neon() 374 q9s32 = vaddq_s32(q7s32, q9s32); in vpx_get4x4sse_cs_neon()
|
/external/libvpx/libvpx/vp9/common/arm/neon/ |
D | vp9_iht_neon.h | 45 s[0] = vaddq_s32(s[0], s[3]); in iadst4() 46 s[0] = vaddq_s32(s[0], s[5]); in iadst4() 52 output[0] = vaddq_s32(s[0], s[3]); in iadst4() 53 output[1] = vaddq_s32(s[1], s[3]); in iadst4() 55 output[3] = vaddq_s32(s[0], s[1]); in iadst4() 69 t0[0] = vaddq_s32(x0_lo, x1_lo); in iadst_half_butterfly_neon() 70 t0[1] = vaddq_s32(x0_hi, x1_hi); in iadst_half_butterfly_neon() 87 t0[0] = vaddq_s32(x0_lo, x1_lo); in iadst_half_butterfly_neg_neon() 88 t0[1] = vaddq_s32(x0_hi, x1_hi); in iadst_half_butterfly_neg_neon() 105 t0[0] = vaddq_s32(x0_lo, x1_lo); in iadst_half_butterfly_pos_neon() [all …]
|
/external/webrtc/modules/audio_coding/codecs/isac/fix/source/ |
D | lattice_neon.c | 63 tmp2a = vaddq_s32(tmp0a, ptr2va); in WebRtcIsacfix_FilterMaLoopNeon() 64 tmp2b = vaddq_s32(tmp0b, ptr2vb); in WebRtcIsacfix_FilterMaLoopNeon() 94 ptr1va = vaddq_s32(tmp1a, tmp3a); in WebRtcIsacfix_FilterMaLoopNeon() 95 ptr1vb = vaddq_s32(tmp1b, tmp3b); in WebRtcIsacfix_FilterMaLoopNeon() 115 tmp2a = vaddq_s32(tmp0a, ptr2va); in WebRtcIsacfix_FilterMaLoopNeon() 135 ptr1va = vaddq_s32(tmp1a, tmp3a); in WebRtcIsacfix_FilterMaLoopNeon()
|
/external/libvpx/libvpx/vp8/encoder/arm/neon/ |
D | vp8_shortwalsh4x4_neon.c | 84 q0s32 = vaddq_s32(q8s32, q9s32); in vp8_short_walsh4x4_neon() 85 q1s32 = vaddq_s32(q11s32, q10s32); in vp8_short_walsh4x4_neon() 104 q8s32 = vaddq_s32(q0s32, q15s32); in vp8_short_walsh4x4_neon() 105 q9s32 = vaddq_s32(q1s32, q15s32); in vp8_short_walsh4x4_neon() 106 q10s32 = vaddq_s32(q2s32, q15s32); in vp8_short_walsh4x4_neon() 107 q11s32 = vaddq_s32(q3s32, q15s32); in vp8_short_walsh4x4_neon()
|
/external/tensorflow/tensorflow/lite/kernels/internal/optimized/integer_ops/ |
D | mean.h | 85 temp_sum.val[0] = vaddq_s32(temp_sum.val[0], input_low_low); in MeanImpl() 86 temp_sum.val[1] = vaddq_s32(temp_sum.val[1], input_high_low); in MeanImpl() 87 temp_sum.val[2] = vaddq_s32(temp_sum.val[2], input_low_high); in MeanImpl() 88 temp_sum.val[3] = vaddq_s32(temp_sum.val[3], input_high_high); in MeanImpl() 95 temp_sum.val[0] = vaddq_s32(temp_sum.val[0], bias_dup); in MeanImpl() 96 temp_sum.val[1] = vaddq_s32(temp_sum.val[1], bias_dup); in MeanImpl() 97 temp_sum.val[2] = vaddq_s32(temp_sum.val[2], bias_dup); in MeanImpl() 98 temp_sum.val[3] = vaddq_s32(temp_sum.val[3], bias_dup); in MeanImpl()
|
/external/libxaac/decoder/armv7/ |
D | ixheaacd_fft32x32_ld.s | 197 VADD.S32 q3, q9, q6 @c_data4_r=vaddq_s32(b_data4_r,b_data6_i)@ 206 VADD.S32 q9, q4, q5 @c_data6_i=vaddq_s32(b_data4_i,b_data6_r)@ 209 VADD.S32 q4, q8, q1 @c_data0_r=vaddq_s32(b_data0_r,b_data2_r)@ 215 VADD.S32 q8, q0, q2 @c_data0_i=vaddq_s32(b_data0_i,b_data2_i)@ 263 VADD.S32 q11, q7, q15 @c_data1_r=vaddq_s32(b_data1_r,b_data3_r)@ 265 VADD.S32 q7, q14, q10 @c_data1_i=vaddq_s32(b_data1_i,b_data3_i)@ 268 VADD.S32 q14, q3, q12 @c_data5_r=vaddq_s32(b_data5_r,b_data5_i)@ 270 VADD.S32 q3, q13, q1 @c_data7_r=vaddq_s32(b_data7_r,b_data7_i)@ 273 VADD.S32 q1 , q14, q12 @b_data5_r=vaddq_s32(c_data7_i,c_data5_r)@ 278 VADD.S32 q14, q3, q10 @b_data7_r=vaddq_s32(c_data5_i,c_data7_r)@ [all …]
|
/external/libaom/libaom/aom_dsp/arm/ |
D | variance_neon.c | 46 *sse = (unsigned int)horizontal_add_s32x4(vaddq_s32(v_sse_lo, v_sse_hi)); in variance_neon_w8() 211 q10s32 = vaddq_s32(q10s32, q9s32); in aom_variance16x8_neon() 273 q10s32 = vaddq_s32(q10s32, q9s32); in aom_variance8x16_neon() 341 q7s32 = vaddq_s32(q7s32, q8s32); in aom_mse16x16_neon() 342 q9s32 = vaddq_s32(q9s32, q10s32); in aom_mse16x16_neon() 343 q10s32 = vaddq_s32(q7s32, q9s32); in aom_mse16x16_neon() 393 q7s32 = vaddq_s32(q7s32, q8s32); in aom_get4x4sse_cs_neon() 394 q9s32 = vaddq_s32(q9s32, q10s32); in aom_get4x4sse_cs_neon() 395 q9s32 = vaddq_s32(q7s32, q9s32); in aom_get4x4sse_cs_neon()
|
/external/libopus/silk/arm/ |
D | NSQ_neon.c | 55 int32x4_t c0 = vaddq_s32(b0, b1); in silk_noise_shape_quantizer_short_prediction_neon() 56 int32x4_t c1 = vaddq_s32(b2, b3); in silk_noise_shape_quantizer_short_prediction_neon() 58 int32x4_t d = vaddq_s32(c0, c1); in silk_noise_shape_quantizer_short_prediction_neon()
|
/external/XNNPACK/src/f32-sigmoid/gen/ |
D | neonfma-rr1-lut64-p2-div-x24.c | 117 …const float32x4_t vs0123 = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl0123), ve0123)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24() 118 …const float32x4_t vs4567 = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl4567), ve4567)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24() 119 …const float32x4_t vs89AB = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl89AB), ve89AB)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24() 120 …const float32x4_t vsCDEF = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vlCDEF), veCDEF)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24() 121 …const float32x4_t vsGHIJ = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vlGHIJ), veGHIJ)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24() 122 …const float32x4_t vsKLMN = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vlKLMN), veKLMN)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24() 218 const float32x4_t vs = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl), ve)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24() 252 const float32x4_t vs = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl), ve)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut64_p2_div_x24()
|
D | neonfma-rr1-lut2048-p1-div-x24.c | 116 …const float32x4_t vs0123 = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl0123), ve0123)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x24() 117 …const float32x4_t vs4567 = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl4567), ve4567)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x24() 118 …const float32x4_t vs89AB = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl89AB), ve89AB)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x24() 119 …const float32x4_t vsCDEF = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vlCDEF), veCDEF)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x24() 120 …const float32x4_t vsGHIJ = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vlGHIJ), veGHIJ)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x24() 121 …const float32x4_t vsKLMN = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vlKLMN), veKLMN)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x24() 210 const float32x4_t vs = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl), ve)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x24() 243 const float32x4_t vs = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl), ve)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x24()
|
D | neonfma-rr1-lut2048-p1-div-x20.c | 104 …const float32x4_t vs0123 = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl0123), ve0123)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x20() 105 …const float32x4_t vs4567 = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl4567), ve4567)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x20() 106 …const float32x4_t vs89AB = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl89AB), ve89AB)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x20() 107 …const float32x4_t vsCDEF = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vlCDEF), veCDEF)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x20() 108 …const float32x4_t vsGHIJ = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vlGHIJ), veGHIJ)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x20() 187 const float32x4_t vs = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl), ve)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x20() 220 const float32x4_t vs = vreinterpretq_f32_s32(vaddq_s32(vreinterpretq_s32_f32(vl), ve)); in xnn_f32_sigmoid_ukernel__neonfma_rr1_lut2048_p1_div_x20()
|