/external/libhevc/common/arm/ |
D | ihevc_resi_trans_neon_32x32.c | 135 vget_high_s16(diff_16[2][0]), vget_low_s16(diff_16[2][0])); in ihevc_resi_trans_32x32_neon() 139 vget_high_s16(diff_16[3][0]), vget_low_s16(diff_16[3][0])); in ihevc_resi_trans_32x32_neon() 160 vget_high_s16(diff_16[2][1]), vget_low_s16(diff_16[2][1])); in ihevc_resi_trans_32x32_neon() 164 vget_high_s16(diff_16[3][1]), vget_low_s16(diff_16[3][1])); in ihevc_resi_trans_32x32_neon() 237 e0_1 = vcombine_s16(vget_high_s16(e0_1), vget_low_s16(e0_1)); in ihevc_resi_trans_32x32_neon() 242 e1_1 = vcombine_s16(vget_high_s16(e1_1), vget_low_s16(e1_1)); in ihevc_resi_trans_32x32_neon() 254 vrev64_s16(vget_high_s16(ee0)), vrev64_s16(vget_high_s16(ee1))); in ihevc_resi_trans_32x32_neon() 264 vreinterpret_s32_s16(vget_high_s16(eee))); in ihevc_resi_trans_32x32_neon() 272 vtrn_s16(vget_low_s16(eeee), vget_high_s16(eeee)); in ihevc_resi_trans_32x32_neon() 286 a[0].val[0], vget_high_s16(g_ai2_ihevc_trans_32_01_8), eeee_01); in ihevc_resi_trans_32x32_neon() [all …]
|
D | ihevc_resi_trans_neon.c | 112 diff_1_high = vget_high_s16(diff_1); in ihevc_resi_trans_4x4_neon() 114 diff_2_high = vget_high_s16(diff_2); in ihevc_resi_trans_4x4_neon() 124 e_1 = vget_high_s16(e_01); in ihevc_resi_trans_4x4_neon() 125 o_0 = vget_high_s16(o_32); in ihevc_resi_trans_4x4_neon() 258 src1_4x16b = vget_high_s16(src_reg0); in ihevc_resi_trans_4x4_ttype1_neon() 260 src3_4x16b = vget_high_s16(src_reg1); in ihevc_resi_trans_4x4_ttype1_neon() 483 a2.val[1] = vmull_n_s16(vget_high_s16(eo0), 83); in ihevc_resi_trans_8x8_neon() 485 a6.val[1] = vmull_n_s16(vget_high_s16(eo0), 36); in ihevc_resi_trans_8x8_neon() 488 a6.val[1] = vmlsl_n_s16(a6.val[1], vget_high_s16(eo1), 83); in ihevc_resi_trans_8x8_neon() 490 a2.val[1] = vmlal_n_s16(a2.val[1], vget_high_s16(eo1), 36); in ihevc_resi_trans_8x8_neon() [all …]
|
/external/libvpx/libvpx/vpx_dsp/arm/ |
D | sum_squares_neon.c | 63 s0 = vmlal_s16(s0, vget_high_s16(s[0]), vget_high_s16(s[0])); in vpx_sum_squares_2d_i16_neon() 64 s0 = vmlal_s16(s0, vget_high_s16(s[1]), vget_high_s16(s[1])); in vpx_sum_squares_2d_i16_neon() 65 s0 = vmlal_s16(s0, vget_high_s16(s[2]), vget_high_s16(s[2])); in vpx_sum_squares_2d_i16_neon() 66 s0 = vmlal_s16(s0, vget_high_s16(s[3]), vget_high_s16(s[3])); in vpx_sum_squares_2d_i16_neon() 67 s0 = vmlal_s16(s0, vget_high_s16(s[4]), vget_high_s16(s[4])); in vpx_sum_squares_2d_i16_neon() 68 s0 = vmlal_s16(s0, vget_high_s16(s[5]), vget_high_s16(s[5])); in vpx_sum_squares_2d_i16_neon() 69 s0 = vmlal_s16(s0, vget_high_s16(s[6]), vget_high_s16(s[6])); in vpx_sum_squares_2d_i16_neon() 70 s0 = vmlal_s16(s0, vget_high_s16(s[7]), vget_high_s16(s[7])); in vpx_sum_squares_2d_i16_neon()
|
D | fwd_txfm_neon.c | 49 int32x4_t v_t0_hi = vaddl_s16(vget_high_s16(v_x0), vget_high_s16(v_x1)); in vpx_fdct8x8_neon() 51 int32x4_t v_t1_hi = vsubl_s16(vget_high_s16(v_x0), vget_high_s16(v_x1)); in vpx_fdct8x8_neon() 53 int32x4_t v_t2_hi = vmull_n_s16(vget_high_s16(v_x2), cospi_24_64); in vpx_fdct8x8_neon() 55 int32x4_t v_t3_hi = vmull_n_s16(vget_high_s16(v_x3), cospi_24_64); in vpx_fdct8x8_neon() 57 v_t2_hi = vmlal_n_s16(v_t2_hi, vget_high_s16(v_x3), cospi_8_64); in vpx_fdct8x8_neon() 59 v_t3_hi = vmlsl_n_s16(v_t3_hi, vget_high_s16(v_x2), cospi_8_64); in vpx_fdct8x8_neon() 82 v_t0_hi = vmull_n_s16(vget_high_s16(v_x0), cospi_16_64); in vpx_fdct8x8_neon() 84 v_t1_hi = vmull_n_s16(vget_high_s16(v_x1), cospi_16_64); in vpx_fdct8x8_neon() 100 v_t0_hi = vmull_n_s16(vget_high_s16(v_x3), cospi_4_64); in vpx_fdct8x8_neon() 102 v_t0_hi = vmlal_n_s16(v_t0_hi, vget_high_s16(v_x0), cospi_28_64); in vpx_fdct8x8_neon() [all …]
|
D | idct_neon.h | 129 t[1] = vaddl_s16(vget_high_s16(a), vget_high_s16(b)); in add_multiply_shift_and_narrow_s16() 141 t[1] = vsubl_s16(vget_high_s16(a), vget_high_s16(b)); in sub_multiply_shift_and_narrow_s16() 154 t[1] = vmull_n_s16(vget_high_s16(a), a_const); in multiply_accumulate_shift_and_narrow_s16() 156 t[1] = vmlal_n_s16(t[1], vget_high_s16(b), b_const); in multiply_accumulate_shift_and_narrow_s16() 298 b[1] = vget_high_s16(a[0]); in idct4x4_16_kernel_bd8() 300 b[3] = vget_high_s16(a[1]); in idct4x4_16_kernel_bd8() 402 t32[3] = vmull_lane_s16(vget_high_s16(step2[6]), cospis0, 2); in idct8x8_12_pass2_bd8() 404 t32[1] = vmlsl_lane_s16(t32[3], vget_high_s16(step2[5]), cospis0, 2); in idct8x8_12_pass2_bd8() 406 t32[3] = vmlal_lane_s16(t32[3], vget_high_s16(step2[5]), cospis0, 2); in idct8x8_12_pass2_bd8() 431 input1h = vget_high_s16(io[1]); in idct8x8_64_1d_bd8_kernel() [all …]
|
D | variance_neon.c | 57 sse_lo_s32 = vmlal_s16(sse_lo_s32, vget_high_s16(diff_lo_s16), in variance_neon_w4x4() 58 vget_high_s16(diff_lo_s16)); in variance_neon_w4x4() 62 sse_hi_s32 = vmlal_s16(sse_hi_s32, vget_high_s16(diff_hi_s16), in variance_neon_w4x4() 63 vget_high_s16(diff_hi_s16)); in variance_neon_w4x4() 106 sse_lo_s32 = vmlal_s16(sse_lo_s32, vget_high_s16(diff_lo_s16), in variance_neon_w16() 107 vget_high_s16(diff_lo_s16)); in variance_neon_w16() 111 sse_hi_s32 = vmlal_s16(sse_hi_s32, vget_high_s16(diff_hi_s16), in variance_neon_w16() 112 vget_high_s16(diff_hi_s16)); in variance_neon_w16() 151 sse_hi_s32 = vmlal_s16(sse_hi_s32, vget_high_s16(diff_0_s16), in variance_neon_w8x2() 152 vget_high_s16(diff_0_s16)); in variance_neon_w8x2() [all …]
|
D | deblock_neon.c | 361 xy_high = vmull_s16(vget_high_s16(x), vget_high_s16(y)); in vpx_mbpost_proc_across_ip_neon() 374 accumulate_sum_sumsq(vget_high_s16(x), xy_high, &sum_high, &sumsq_high); in vpx_mbpost_proc_across_ip_neon() 431 sumsq_high = vmull_s16(vget_high_s16(sum), vget_high_s16(sum_tmp)); in vpx_mbpost_proc_down_neon() 440 sumsq_high = vmlal_s16(sumsq_high, vget_high_s16(b), vget_high_s16(b)); in vpx_mbpost_proc_down_neon() 458 xy_high = vmull_s16(vget_high_s16(x), vget_high_s16(y)); in vpx_mbpost_proc_down_neon() 465 mask = combine_mask(vget_low_s16(sum), vget_high_s16(sum), sumsq_low, in vpx_mbpost_proc_down_neon()
|
/external/libvpx/libvpx/vp8/common/arm/neon/ |
D | iwalsh_neon.c | 26 d4s16 = vadd_s16(vget_low_s16(q0s16), vget_high_s16(q1s16)); in vp8_short_inv_walsh4x4_neon() 27 d6s16 = vadd_s16(vget_high_s16(q0s16), vget_low_s16(q1s16)); in vp8_short_inv_walsh4x4_neon() 28 d5s16 = vsub_s16(vget_low_s16(q0s16), vget_high_s16(q1s16)); in vp8_short_inv_walsh4x4_neon() 29 d7s16 = vsub_s16(vget_high_s16(q0s16), vget_low_s16(q1s16)); in vp8_short_inv_walsh4x4_neon() 39 v2tmp3 = vtrn_s32(vreinterpret_s32_s16(vget_high_s16(q0s16)), in vp8_short_inv_walsh4x4_neon() 40 vreinterpret_s32_s16(vget_high_s16(q1s16))); in vp8_short_inv_walsh4x4_neon() 68 vst1_lane_s16(mb_dqcoeff, vget_high_s16(q0s16), 0); in vp8_short_inv_walsh4x4_neon() 72 vst1_lane_s16(mb_dqcoeff, vget_high_s16(q1s16), 0); in vp8_short_inv_walsh4x4_neon() 77 vst1_lane_s16(mb_dqcoeff, vget_high_s16(q0s16), 1); in vp8_short_inv_walsh4x4_neon() 81 vst1_lane_s16(mb_dqcoeff, vget_high_s16(q1s16), 1); in vp8_short_inv_walsh4x4_neon() [all …]
|
D | shortidct4x4llm_neon.c | 46 d12 = vqadd_s16(vget_low_s16(q1s16), vget_high_s16(q1s16)); // a1 in vp8_short_idct4x4llm_neon() 47 d13 = vqsub_s16(vget_low_s16(q1s16), vget_high_s16(q1s16)); // b1 in vp8_short_idct4x4llm_neon() 53 d10 = vqsub_s16(vget_low_s16(q3s16), vget_high_s16(q4s16)); // c1 in vp8_short_idct4x4llm_neon() 54 d11 = vqadd_s16(vget_high_s16(q3s16), vget_low_s16(q4s16)); // d1 in vp8_short_idct4x4llm_neon() 75 d12 = vqadd_s16(vget_low_s16(q1s16), vget_high_s16(q1s16)); // a1 in vp8_short_idct4x4llm_neon() 76 d13 = vqsub_s16(vget_low_s16(q1s16), vget_high_s16(q1s16)); // b1 in vp8_short_idct4x4llm_neon() 82 d10 = vqsub_s16(vget_low_s16(q3s16), vget_high_s16(q4s16)); // c1 in vp8_short_idct4x4llm_neon() 83 d11 = vqadd_s16(vget_high_s16(q3s16), vget_low_s16(q4s16)); // d1 in vp8_short_idct4x4llm_neon()
|
D | dequant_idct_neon.c | 64 q2 = vcombine_s16(vget_high_s16(q1), vget_high_s16(q2)); in vp8_dequant_idct_add_neon() 73 d10 = vqsub_s16(vget_low_s16(q3), vget_high_s16(q4)); in vp8_dequant_idct_add_neon() 74 d11 = vqadd_s16(vget_high_s16(q3), vget_low_s16(q4)); in vp8_dequant_idct_add_neon() 101 d10 = vqsub_s16(vget_low_s16(q3), vget_high_s16(q4)); in vp8_dequant_idct_add_neon() 102 d11 = vqadd_s16(vget_high_s16(q3), vget_low_s16(q4)); in vp8_dequant_idct_add_neon()
|
/external/libvpx/libvpx/vp9/common/arm/neon/ |
D | vp9_iht_neon.h | 32 x[2] = vget_high_s16(io[0]); in iadst4() 33 x[3] = vget_high_s16(io[1]); in iadst4() 64 const int32x4_t x0_hi = vmull_lane_s16(vget_high_s16(x[0]), c, 0); in iadst_half_butterfly_neon() 66 const int32x4_t x1_hi = vmull_lane_s16(vget_high_s16(x[1]), c, 0); in iadst_half_butterfly_neon() 82 const int32x4_t x0_hi = vmull_lane_s16(vget_high_s16(*x0), c, 1); in iadst_half_butterfly_neg_neon() 84 const int32x4_t x1_hi = vmull_lane_s16(vget_high_s16(*x1), c, 1); in iadst_half_butterfly_neg_neon() 100 const int32x4_t x0_hi = vmull_lane_s16(vget_high_s16(*x0), c, 0); in iadst_half_butterfly_pos_neon() 102 const int32x4_t x1_hi = vmull_lane_s16(vget_high_s16(*x1), c, 0); in iadst_half_butterfly_pos_neon() 119 s0[1] = vmull_lane_s16(vget_high_s16(in0), c, 0); in iadst_butterfly_lane_0_1_neon() 121 s1[1] = vmull_lane_s16(vget_high_s16(in0), c, 1); in iadst_butterfly_lane_0_1_neon() [all …]
|
D | vp9_iht4x4_add_neon.c | 38 a[1] = vcombine_s16(vget_high_s16(a[1]), vget_low_s16(a[1])); in vp9_iht4x4_16_add_neon() 41 a[1] = vcombine_s16(vget_high_s16(a[1]), vget_low_s16(a[1])); in vp9_iht4x4_16_add_neon() 46 a[1] = vcombine_s16(vget_high_s16(a[1]), vget_low_s16(a[1])); in vp9_iht4x4_16_add_neon() 55 a[1] = vcombine_s16(vget_high_s16(a[1]), vget_low_s16(a[1])); in vp9_iht4x4_16_add_neon()
|
/external/libaom/libaom/aom_dsp/arm/ |
D | fwd_txfm_neon.c | 46 int32x4_t v_t0_hi = vaddl_s16(vget_high_s16(v_x0), vget_high_s16(v_x1)); in aom_fdct8x8_neon() 48 int32x4_t v_t1_hi = vsubl_s16(vget_high_s16(v_x0), vget_high_s16(v_x1)); in aom_fdct8x8_neon() 50 int32x4_t v_t2_hi = vmull_n_s16(vget_high_s16(v_x2), (int16_t)cospi_24_64); in aom_fdct8x8_neon() 52 int32x4_t v_t3_hi = vmull_n_s16(vget_high_s16(v_x3), (int16_t)cospi_24_64); in aom_fdct8x8_neon() 54 v_t2_hi = vmlal_n_s16(v_t2_hi, vget_high_s16(v_x3), (int16_t)cospi_8_64); in aom_fdct8x8_neon() 56 v_t3_hi = vmlsl_n_s16(v_t3_hi, vget_high_s16(v_x2), (int16_t)cospi_8_64); in aom_fdct8x8_neon() 79 v_t0_hi = vmull_n_s16(vget_high_s16(v_x0), (int16_t)cospi_16_64); in aom_fdct8x8_neon() 81 v_t1_hi = vmull_n_s16(vget_high_s16(v_x1), (int16_t)cospi_16_64); in aom_fdct8x8_neon() 97 v_t0_hi = vmull_n_s16(vget_high_s16(v_x3), (int16_t)cospi_4_64); in aom_fdct8x8_neon() 99 v_t0_hi = vmlal_n_s16(v_t0_hi, vget_high_s16(v_x0), (int16_t)cospi_28_64); in aom_fdct8x8_neon() [all …]
|
/external/libaom/libaom/av1/common/arm/ |
D | warp_plane_neon.c | 217 *res = vpadd_s16(vget_low_s16(pix_0), vget_high_s16(pix_0)); in convolve() 285 b0 = vtrn_s32(vreinterpret_s32_s16(vget_high_s16(f0)), in horizontal_filter_neon() 286 vreinterpret_s32_s16(vget_high_s16(f2))); in horizontal_filter_neon() 287 b1 = vtrn_s32(vreinterpret_s32_s16(vget_high_s16(f4)), in horizontal_filter_neon() 288 vreinterpret_s32_s16(vget_high_s16(f6))); in horizontal_filter_neon() 291 b0 = vtrn_s32(vreinterpret_s32_s16(vget_high_s16(f1)), in horizontal_filter_neon() 292 vreinterpret_s32_s16(vget_high_s16(f3))); in horizontal_filter_neon() 293 b1 = vtrn_s32(vreinterpret_s32_s16(vget_high_s16(f5)), in horizontal_filter_neon() 294 vreinterpret_s32_s16(vget_high_s16(f7))); in horizontal_filter_neon() 385 fltr_0 = vget_high_s16(vreinterpretq_s16_s32(d0.val[0])); in vertical_filter_neon() [all …]
|
D | convolve_neon.h | 43 sum1 = vmull_n_s16(vget_high_s16(ss0), filter_y[0]); in wiener_convolve8_vert_4x8() 44 sum1 = vmlal_n_s16(sum1, vget_high_s16(ss1), filter_y[1]); in wiener_convolve8_vert_4x8() 45 sum1 = vmlal_n_s16(sum1, vget_high_s16(ss2), filter_y[2]); in wiener_convolve8_vert_4x8() 46 sum1 = vmlal_n_s16(sum1, vget_high_s16(s3), filter_y[3]); in wiener_convolve8_vert_4x8() 90 sum_1 = vmovl_s16(vget_high_s16(sum)); in wiener_convolve8_horiz_8x8() 97 s3_1 = vmull_n_s16(vget_high_s16(s3), filter_x[3]); in wiener_convolve8_horiz_8x8()
|
D | selfguided_neon.c | 365 vget_high_s16(t1), vget_high_s16(t2), vget_high_s16(t3), in boxsum2() 366 vget_high_s16(t4), vget_high_s16(t5), vget_high_s16(t6), in boxsum2() 367 vget_high_s16(t7), vget_high_s16(t8), vget_high_s16(t9), in boxsum2() 368 vget_high_s16(t10), vget_high_s16(t11), &r12345, &r34567, &r56789, in boxsum2() 724 r2 = vmull_s16(vget_high_s16(s2), vget_high_s16(s2)); in boxsum1() 725 r3 = vmull_s16(vget_high_s16(s3), vget_high_s16(s3)); in boxsum1() 726 r4_high = vmull_s16(vget_high_s16(s4), vget_high_s16(s4)); in boxsum1() 768 r5 = vmull_s16(vget_high_s16(s5), vget_high_s16(s5)); in boxsum1() 769 r6 = vmull_s16(vget_high_s16(s6), vget_high_s16(s6)); in boxsum1() 770 r7 = vmull_s16(vget_high_s16(s7), vget_high_s16(s7)); in boxsum1() [all …]
|
/external/tensorflow/tensorflow/lite/kernels/internal/optimized/ |
D | depthwiseconv_uint8.h | 69 acc[1].val[i] = vmlal_s16(acc[1].val[i], vget_high_s16(filter[i]), 70 vget_high_s16(input_dup2.val[i])); 117 vmlal_s16(acc[1], vget_high_s16(filter), vget_high_s16(input[0])); 120 vmlal_s16(acc[3], vget_high_s16(filter), vget_high_s16(input[1])); 141 acc[1] = vmlal_s16(acc[1], vget_high_s16(filter), vget_high_s16(input)); 180 acc[2 * i + 1] = vmlal_s16(acc[2 * i + 1], vget_high_s16(filter), 181 vget_high_s16(input_dup2.val[i])); 210 acc[1] = vmlal_s16(acc[1], vget_high_s16(filter), input_dup2.val[1]); 253 acc[1] = vmlal_lane_s16(acc[1], vget_high_s16(filter[0]), input, 0); 255 acc[3] = vmlal_lane_s16(acc[3], vget_high_s16(filter[1]), input, 1); [all …]
|
/external/webrtc/webrtc/common_audio/signal_processing/ |
D | downsample_fast_neon.c | 65 int16x4_t in16x4_2 = vget_high_s16(in16x8x2.val[0]); in WebRtcSpl_DownsampleFastNeon() 66 int16x4_t in16x4_3 = vget_high_s16(in16x8x2.val[1]); in WebRtcSpl_DownsampleFastNeon() 84 int16x4_t in16x4_1 = vget_high_s16(in16x8x2.val[0]); in WebRtcSpl_DownsampleFastNeon() 96 int16x4_t in16x4_1 = vget_high_s16(in16x8x2.val[0]); in WebRtcSpl_DownsampleFastNeon() 136 int16x4_t in16x4_1 = vget_high_s16(in16x8x4.val[0]); in WebRtcSpl_DownsampleFastNeon() 137 int16x4_t in16x4_3 = vget_high_s16(in16x8x4.val[1]); in WebRtcSpl_DownsampleFastNeon() 138 int16x4_t in16x4_5 = vget_high_s16(in16x8x4.val[2]); in WebRtcSpl_DownsampleFastNeon() 139 int16x4_t in16x4_7 = vget_high_s16(in16x8x4.val[3]); in WebRtcSpl_DownsampleFastNeon() 159 int16x4_t in16x4_1 = vget_high_s16(in16x8x4.val[0]); in WebRtcSpl_DownsampleFastNeon()
|
/external/webp/src/dsp/ |
D | enc_neon.c | 101 vcombine_s16(vget_high_s16(rows->val[0]), vget_high_s16(rows->val[1])); in TransformPass_NEON() 112 const int16x4_t c = vqsub_s16(vget_low_s16(C1), vget_high_s16(C0)); in TransformPass_NEON() 113 const int16x4_t d = vqadd_s16(vget_low_s16(C0), vget_high_s16(C1)); in TransformPass_NEON() 118 const int16x8_t E1 = vcombine_s16(vget_high_s16(E_tmp), vget_low_s16(E_tmp)); in TransformPass_NEON() 305 const int16x4_t D1 = vget_high_s16(D0D1); in FTransform_NEON() 307 const int16x4_t D3 = vget_high_s16(D2D3); in FTransform_NEON() 317 vget_high_s16(a0a1_2)); in FTransform_NEON() 319 vget_high_s16(a0a1_2)); in FTransform_NEON() 321 const int32x4_t a2_2217 = vmull_n_s16(vget_high_s16(a3a2), 2217); in FTransform_NEON() 323 const int32x4_t a3_m_a2 = vmlsl_n_s16(a3_2217, vget_high_s16(a3a2), 5352); in FTransform_NEON() [all …]
|
/external/webrtc/webrtc/modules/audio_coding/codecs/isac/fix/source/ |
D | transform_neon.c | 60 int32x4_t tmp2 = vmull_s16(vget_high_s16(tmpr), vget_high_s16(inre1)); in ComplexMulAndFindMaxNeon() 61 int32x4_t tmp3 = vmull_s16(vget_high_s16(tmpr), vget_high_s16(inre2)); in ComplexMulAndFindMaxNeon() 62 tmp2 = vmlal_s16(tmp2, vget_high_s16(tmpi), vget_high_s16(inre2)); in ComplexMulAndFindMaxNeon() 63 tmp3 = vmlsl_s16(tmp3, vget_high_s16(tmpi), vget_high_s16(inre1)); in ComplexMulAndFindMaxNeon() 384 int32x4_t outre1_1 = vmull_s16(vget_high_s16(inre16x8), in PostShiftAndDivideAndDemodulateNeon() 385 vget_high_s16(scale)); in PostShiftAndDivideAndDemodulateNeon() 386 int32x4_t outre2_1 = vmull_s16(vget_high_s16(inim16x8), in PostShiftAndDivideAndDemodulateNeon() 387 vget_high_s16(scale)); in PostShiftAndDivideAndDemodulateNeon() 402 int32x4_t tmpr_1 = vmovl_s16(vget_high_s16(tmpr)); in PostShiftAndDivideAndDemodulateNeon() 403 int32x4_t tmpi_1 = vmovl_s16(vget_high_s16(tmpi)); in PostShiftAndDivideAndDemodulateNeon()
|
/external/libaom/libaom/av1/encoder/arm/neon/ |
D | quantize_neon.c | 61 vmull_s16(vget_high_s16(v_tmp), vget_high_s16(v_quant)); in av1_quantize_fp_neon() 86 vmull_s16(vget_high_s16(v_tmp), vget_high_s16(v_quant)); in av1_quantize_fp_neon() 101 vget_low_s16(v_eobmax_76543210), vget_high_s16(v_eobmax_76543210)); in av1_quantize_fp_neon()
|
/external/tensorflow/tensorflow/core/kernels/ |
D | quantized_mul_op.cc | 79 vget_high_s16(full_input_minus_offset_high_16x8); in ScalarMultiply() 83 vget_high_s16(full_input_minus_offset_low_16x8); in ScalarMultiply() 156 const int16x4_t x_high_high_16x4 = vget_high_s16(x_minus_offset_high_16x8); in VectorMultiply() 158 const int16x4_t x_low_high_16x4 = vget_high_s16(x_minus_offset_low_16x8); in VectorMultiply() 160 const int16x4_t y_high_high_16x4 = vget_high_s16(y_minus_offset_high_16x8); in VectorMultiply() 162 const int16x4_t y_low_high_16x4 = vget_high_s16(y_minus_offset_low_16x8); in VectorMultiply() 242 vget_high_s16(x_minus_offset_high_16x8); in VectorTensorMultiply() 244 const int16x4_t x_low_high_16x4 = vget_high_s16(x_minus_offset_low_16x8); in VectorTensorMultiply() 247 vget_high_s16(y_minus_offset_high_16x8); in VectorTensorMultiply() 249 const int16x4_t y_low_high_16x4 = vget_high_s16(y_minus_offset_low_16x8); in VectorTensorMultiply()
|
/external/libvpx/libvpx/vp9/encoder/arm/neon/ |
D | vp9_quantize_neon.c | 62 vmull_s16(vget_high_s16(v_tmp), vget_high_s16(v_quant)); in vp9_quantize_fp_neon() 87 vmull_s16(vget_high_s16(v_tmp), vget_high_s16(v_quant)); in vp9_quantize_fp_neon() 105 vget_high_s16(v_eobmax_76543210)); in vp9_quantize_fp_neon() 168 dqcoeff_1 = vmull_n_s16(vget_high_s16(qcoeff), dequant_ptr[1]); in vp9_quantize_fp_32x32_neon() 213 dqcoeff_1 = vmull_n_s16(vget_high_s16(qcoeff), dequant_ptr[1]); in vp9_quantize_fp_32x32_neon()
|
/external/webrtc/webrtc/modules/audio_processing/ns/ |
D | nsx_core_neon.c | 255 tmp32x4 = vmull_s16(vget_high_s16(tmp16x8_0), Q3_16x4); in WebRtcNsx_NoiseEstimationNeon() 382 int32x4_t tmp_r_1 = vmull_s16(vget_high_s16(real), in WebRtcNsx_PrepareSpectrumNeon() 383 vget_high_s16(ns_filter)); in WebRtcNsx_PrepareSpectrumNeon() 384 int32x4_t tmp_i_1 = vmull_s16(vget_high_s16(imag), in WebRtcNsx_PrepareSpectrumNeon() 385 vget_high_s16(ns_filter)); in WebRtcNsx_PrepareSpectrumNeon() 468 int32x4_t tmp32a_0_high = vmull_s16(vget_high_s16(real_0), in WebRtcNsx_SynthesisUpdateNeon() 469 vget_high_s16(window_0)); in WebRtcNsx_SynthesisUpdateNeon() 473 int32x4_t tmp32a_1_high = vmull_s16(vget_high_s16(real_1), in WebRtcNsx_SynthesisUpdateNeon() 474 vget_high_s16(window_1)); in WebRtcNsx_SynthesisUpdateNeon() 585 int32x4_t tmp32_high = vmull_s16(vget_high_s16(window), in WebRtcNsx_AnalysisUpdateNeon() [all …]
|
/external/gemmlowp/internal/ |
D | output_neon.h | 278 vst1_s16(dst->data(row, col + 1), vget_high_s16(src.buf.reg[0])); 280 vst1_s16(dst->data(row, col + 3), vget_high_s16(src.buf.reg[1])); 283 vtrn_s16(vget_low_s16(src.buf.reg[0]), vget_high_s16(src.buf.reg[0])); 285 vtrn_s16(vget_low_s16(src.buf.reg[1]), vget_high_s16(src.buf.reg[1])); 292 vget_high_s16(vreinterpretq_s16_s32(t.val[0]))); 296 vget_high_s16(vreinterpretq_s16_s32(t.val[1]))); 361 vget_high_s16(vreinterpretq_s16_s32(u0.val[0]))); 363 vget_high_s16(vreinterpretq_s16_s32(u1.val[0]))); 365 vget_high_s16(vreinterpretq_s16_s32(u0.val[1]))); 367 vget_high_s16(vreinterpretq_s16_s32(u1.val[1]))); [all …]
|