diff --git a/openair1/PHY/LTE_UE_TRANSPORT/dlsch_demodulation.c b/openair1/PHY/LTE_UE_TRANSPORT/dlsch_demodulation.c index 81347085539ef01ace695a7eb85e891fee79f0a9..54ef07cf89d510fa6673e283f4e13723b05f1bc4 100644 --- a/openair1/PHY/LTE_UE_TRANSPORT/dlsch_demodulation.c +++ b/openair1/PHY/LTE_UE_TRANSPORT/dlsch_demodulation.c @@ -546,6 +546,7 @@ int rx_pdsch(PHY_VARS_UE *ue, #if UE_TIMING_TRACE start_meas(&ue->generic_stat_bis[ue->current_thread_id[subframe]][slot]); #endif + // Now channel compensation if (dlsch0_harq->mimo_mode<LARGE_CDD) { dlsch_channel_compensation(pdsch_vars[eNB_id]->rxdataF_ext, @@ -564,7 +565,7 @@ int rx_pdsch(PHY_VARS_UE *ue, if (symbol == 5) { LOG_M("rxF_comp_d.m","rxF_c_d",&pdsch_vars[eNB_id]->rxdataF_comp0[0][symbol*frame_parms->N_RB_DL*12],frame_parms->N_RB_DL*12,1,1); } - + if ((rx_type==rx_IC_single_stream) && (eNB_id_i<ue->n_connected_eNB)) { dlsch_channel_compensation(pdsch_vars[eNB_id_i]->rxdataF_ext, @@ -1706,6 +1707,204 @@ void dlsch_channel_compensation(int **rxdataF_ext, #endif } +void dlsch_channel_compensation_core(int **rxdataF_ext, + int **dl_ch_estimates_ext, + int **dl_ch_mag, + int **dl_ch_magb, + int **rxdataF_comp, + int **rho, + unsigned char n_tx, + unsigned char n_rx, + unsigned char mod_order, + unsigned char output_shift, + int length, + int start_point) + +{ + + unsigned short ii; + int length_mod8 = 0; + int length2; + __m128i *dl_ch128,*dl_ch_mag128,*dl_ch_mag128b, *dl_ch128_2, *rxdataF128,*rxdataF_comp128,*rho128; + __m128i mmtmpD0,mmtmpD1,mmtmpD2,mmtmpD3,QAM_amp128,QAM_amp128b; + int aatx = 0, aarx = 0; + + for (aatx=0; aatx<n_tx; aatx++) { + + if (mod_order == 4) { + QAM_amp128 = _mm_set1_epi16(QAM16_n1); // 2/sqrt(10) + QAM_amp128b = _mm_setzero_si128(); + } else if (mod_order == 6) { + QAM_amp128 = _mm_set1_epi16(QAM64_n1); // + QAM_amp128b = _mm_set1_epi16(QAM64_n2); + } + + for (aarx=0; aarx<n_rx; aarx++) { + + dl_ch128 = (__m128i *)&dl_ch_estimates_ext[aatx*n_rx + aarx][start_point]; + dl_ch_mag128 = (__m128i *)&dl_ch_mag[aatx*n_rx + aarx][start_point]; + dl_ch_mag128b = (__m128i *)&dl_ch_magb[aatx*n_rx + aarx][start_point]; + rxdataF128 = (__m128i *)&rxdataF_ext[aarx][start_point]; + rxdataF_comp128 = (__m128i *)&rxdataF_comp[aatx*n_rx + aarx][start_point]; + + length_mod8 = length&7; + if (length_mod8 == 0){ + length2 = length>>3; + + for (ii=0; ii<length2; ++ii) { + if (mod_order>2) { + // get channel amplitude if not QPSK + + mmtmpD0 = _mm_madd_epi16(dl_ch128[0],dl_ch128[0]); + mmtmpD0 = _mm_srai_epi32(mmtmpD0,output_shift); + + mmtmpD1 = _mm_madd_epi16(dl_ch128[1],dl_ch128[1]); + mmtmpD1 = _mm_srai_epi32(mmtmpD1,output_shift); + + mmtmpD0 = _mm_packs_epi32(mmtmpD0,mmtmpD1); + + // store channel magnitude here in a new field of dlsch + + dl_ch_mag128[0] = _mm_unpacklo_epi16(mmtmpD0,mmtmpD0); + dl_ch_mag128b[0] = dl_ch_mag128[0]; + dl_ch_mag128[0] = _mm_mulhi_epi16(dl_ch_mag128[0],QAM_amp128); + dl_ch_mag128[0] = _mm_slli_epi16(dl_ch_mag128[0],1); + //print_ints("Re(ch):",(int16_t*)&mmtmpD0); + //print_shorts("QAM_amp:",(int16_t*)&QAM_amp128); + //print_shorts("mag:",(int16_t*)&dl_ch_mag128[0]); + dl_ch_mag128[1] = _mm_unpackhi_epi16(mmtmpD0,mmtmpD0); + dl_ch_mag128b[1] = dl_ch_mag128[1]; + dl_ch_mag128[1] = _mm_mulhi_epi16(dl_ch_mag128[1],QAM_amp128); + dl_ch_mag128[1] = _mm_slli_epi16(dl_ch_mag128[1],1); + + dl_ch_mag128b[0] = _mm_mulhi_epi16(dl_ch_mag128b[0],QAM_amp128b); + dl_ch_mag128b[0] = _mm_slli_epi16(dl_ch_mag128b[0],1); + + dl_ch_mag128b[1] = _mm_mulhi_epi16(dl_ch_mag128b[1],QAM_amp128b); + dl_ch_mag128b[1] = _mm_slli_epi16(dl_ch_mag128b[1],1); + + } + + // multiply by conjugated channel + mmtmpD0 = _mm_madd_epi16(dl_ch128[0],rxdataF128[0]); + + // mmtmpD0 contains real part of 4 consecutive outputs (32-bit) + mmtmpD1 = _mm_shufflelo_epi16(dl_ch128[0],_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_shufflehi_epi16(mmtmpD1,_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_sign_epi16(mmtmpD1,*(__m128i*)&conjugate[0]); + // print_ints("im",&mmtmpD1); + mmtmpD1 = _mm_madd_epi16(mmtmpD1,rxdataF128[0]); + // mmtmpD1 contains imag part of 4 consecutive outputs (32-bit) + mmtmpD0 = _mm_srai_epi32(mmtmpD0,output_shift); + // print_ints("re(shift)",&mmtmpD0); + mmtmpD1 = _mm_srai_epi32(mmtmpD1,output_shift); + // print_ints("im(shift)",&mmtmpD1); + mmtmpD2 = _mm_unpacklo_epi32(mmtmpD0,mmtmpD1); + mmtmpD3 = _mm_unpackhi_epi32(mmtmpD0,mmtmpD1); + // print_ints("c0",&mmtmpD2); + // print_ints("c1",&mmtmpD3); + rxdataF_comp128[0] = _mm_packs_epi32(mmtmpD2,mmtmpD3); + // print_shorts("rx:",rxdataF128); + // print_shorts("ch:",dl_ch128); + // print_shorts("pack:",rxdataF_comp128); + + // multiply by conjugated channel + mmtmpD0 = _mm_madd_epi16(dl_ch128[1],rxdataF128[1]); + // mmtmpD0 contains real part of 4 consecutive outputs (32-bit) + mmtmpD1 = _mm_shufflelo_epi16(dl_ch128[1],_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_shufflehi_epi16(mmtmpD1,_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_sign_epi16(mmtmpD1,*(__m128i*)conjugate); + mmtmpD1 = _mm_madd_epi16(mmtmpD1,rxdataF128[1]); + // mmtmpD1 contains imag part of 4 consecutive outputs (32-bit) + mmtmpD0 = _mm_srai_epi32(mmtmpD0,output_shift); + mmtmpD1 = _mm_srai_epi32(mmtmpD1,output_shift); + mmtmpD2 = _mm_unpacklo_epi32(mmtmpD0,mmtmpD1); + mmtmpD3 = _mm_unpackhi_epi32(mmtmpD0,mmtmpD1); + + rxdataF_comp128[1] = _mm_packs_epi32(mmtmpD2,mmtmpD3); + // print_shorts("rx:",rxdataF128+1); + // print_shorts("ch:",dl_ch128+1); + //print_shorts("pack:",rxdataF_comp128+1); + + dl_ch128+=2; + dl_ch_mag128+=2; + dl_ch_mag128b+=2; + rxdataF128+=2; + rxdataF_comp128+=2; + } + }else { + printf ("Channel Compensation: Received number of subcarriers is not multiple of 8, \n" + "need to adapt the code!\n"); + } + } + } + +/*This part of code makes sense only for processing in 2x2 blocks*/ + if (rho) { + + + for (aarx=0; aarx<n_rx; aarx++) { + rho128 = (__m128i *)&rho[aarx][start_point]; + dl_ch128 = (__m128i *)&dl_ch_estimates_ext[aarx][start_point]; + dl_ch128_2 = (__m128i *)&dl_ch_estimates_ext[2+aarx][start_point]; + + if (length_mod8 == 0){ + length2 = length>>3; + + for (ii=0; ii<length2; ++ii) { + // multiply by conjugated channel + mmtmpD0 = _mm_madd_epi16(dl_ch128[0],dl_ch128_2[0]); + // print_ints("re",&mmtmpD0); + + // mmtmpD0 contains real part of 4 consecutive outputs (32-bit) + mmtmpD1 = _mm_shufflelo_epi16(dl_ch128[0],_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_shufflehi_epi16(mmtmpD1,_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_sign_epi16(mmtmpD1,*(__m128i*)&conjugate[0]); + // print_ints("im",&mmtmpD1); + mmtmpD1 = _mm_madd_epi16(mmtmpD1,dl_ch128_2[0]); + // mmtmpD1 contains imag part of 4 consecutive outputs (32-bit) + mmtmpD0 = _mm_srai_epi32(mmtmpD0,output_shift); + // print_ints("re(shift)",&mmtmpD0); + mmtmpD1 = _mm_srai_epi32(mmtmpD1,output_shift); + // print_ints("im(shift)",&mmtmpD1); + mmtmpD2 = _mm_unpacklo_epi32(mmtmpD0,mmtmpD1); + mmtmpD3 = _mm_unpackhi_epi32(mmtmpD0,mmtmpD1); + // print_ints("c0",&mmtmpD2); + // print_ints("c1",&mmtmpD3); + rho128[0] = _mm_packs_epi32(mmtmpD2,mmtmpD3); + + //print_shorts("rx:",dl_ch128_2); + //print_shorts("ch:",dl_ch128); + //print_shorts("pack:",rho128); + + // multiply by conjugated channel + mmtmpD0 = _mm_madd_epi16(dl_ch128[1],dl_ch128_2[1]); + // mmtmpD0 contains real part of 4 consecutive outputs (32-bit) + mmtmpD1 = _mm_shufflelo_epi16(dl_ch128[1],_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_shufflehi_epi16(mmtmpD1,_MM_SHUFFLE(2,3,0,1)); + mmtmpD1 = _mm_sign_epi16(mmtmpD1,*(__m128i*)conjugate); + mmtmpD1 = _mm_madd_epi16(mmtmpD1,dl_ch128_2[1]); + // mmtmpD1 contains imag part of 4 consecutive outputs (32-bit) + mmtmpD0 = _mm_srai_epi32(mmtmpD0,output_shift); + mmtmpD1 = _mm_srai_epi32(mmtmpD1,output_shift); + mmtmpD2 = _mm_unpacklo_epi32(mmtmpD0,mmtmpD1); + mmtmpD3 = _mm_unpackhi_epi32(mmtmpD0,mmtmpD1); + + rho128[1] =_mm_packs_epi32(mmtmpD2,mmtmpD3); + dl_ch128+=2; + dl_ch128_2+=2; + rho128+=2; + } + }else { + printf ("Channel Compensation: Received number of subcarriers is not multiple of 8, \n" + "need to adapt the code!\n"); + } + } + } + _mm_empty(); + _m_empty(); +} + #if defined(__x86_64__) || defined(__i386__) void prec2A_TM56_128(unsigned char pmi,__m128i *ch0,__m128i *ch1) @@ -3095,127 +3294,6 @@ void dlsch_dual_stream_correlation(LTE_DL_FRAME_PARMS *frame_parms, } -/*void dlsch_dual_stream_correlationTM34(LTE_DL_FRAME_PARMS *frame_parms, - unsigned char symbol, - unsigned short nb_rb, - int **dl_ch_estimates_ext, - int **dl_ch_estimates_ext_i, - int **dl_ch_rho_ext, - unsigned char output_shift0, - unsigned char output_shift1) -{ - -#if defined(__x86_64__)||defined(__i386__) - - unsigned short rb; - __m128i *dl_ch128,*dl_ch128i,*dl_ch_rho128,mmtmpD0,mmtmpD1,mmtmpD2,mmtmpD3; - unsigned char aarx,symbol_mod,pilots=0; - int output_shift; - - // printf("dlsch_dual_stream_correlation: symbol %d\n",symbol); - - symbol_mod = (symbol>=(7-frame_parms->Ncp)) ? symbol-(7-frame_parms->Ncp) : symbol; - - if ((symbol_mod == 0) || (symbol_mod == (4-frame_parms->Ncp))) { - pilots=1; - } - - // printf("Dual stream correlation (%p)\n",dl_ch_estimates_ext_i); - - for (aarx=0; aarx<frame_parms->nb_antennas_rx; aarx++) { - - if (aarx==0) { - output_shift=output_shift0; - } - else { - output_shift=output_shift1; - } - - //printf ("antenna %d", aarx); - dl_ch128 = (__m128i *)&dl_ch_estimates_ext[aarx][symbol*frame_parms->N_RB_DL*12]; - - if (dl_ch_estimates_ext_i == NULL) // TM3/4 - dl_ch128i = (__m128i *)&dl_ch_estimates_ext[2+aarx][symbol*frame_parms->N_RB_DL*12]; - else - dl_ch128i = (__m128i *)&dl_ch_estimates_ext_i[aarx][symbol*frame_parms->N_RB_DL*12]; - - dl_ch_rho128 = (__m128i *)&dl_ch_rho_ext[aarx][symbol*frame_parms->N_RB_DL*12]; - - - for (rb=0; rb<nb_rb; rb++) { - // multiply by conjugated channel - mmtmpD0 = _mm_madd_epi16(dl_ch128[0],dl_ch128i[0]); - // print_ints("re",&mmtmpD0); - // mmtmpD0 contains real part of 4 consecutive outputs (32-bit) - mmtmpD1 = _mm_shufflelo_epi16(dl_ch128[0],_MM_SHUFFLE(2,3,0,1)); - mmtmpD1 = _mm_shufflehi_epi16(mmtmpD1,_MM_SHUFFLE(2,3,0,1)); - mmtmpD1 = _mm_sign_epi16(mmtmpD1,*(__m128i*)&conjugate[0]); - mmtmpD1 = _mm_madd_epi16(mmtmpD1,dl_ch128i[0]); - // print_ints("im",&mmtmpD1); - // mmtmpD1 contains imag part of 4 consecutive outputs (32-bit) - mmtmpD0 = _mm_srai_epi32(mmtmpD0,output_shift); - // print_ints("re(shift)",&mmtmpD0); - mmtmpD1 = _mm_srai_epi32(mmtmpD1,output_shift); - // print_ints("im(shift)",&mmtmpD1); - mmtmpD2 = _mm_unpacklo_epi32(mmtmpD0,mmtmpD1); - mmtmpD3 = _mm_unpackhi_epi32(mmtmpD0,mmtmpD1); - // print_ints("c0",&mmtmpD2); - // print_ints("c1",&mmtmpD3); - dl_ch_rho128[0] = _mm_packs_epi32(mmtmpD2,mmtmpD3); - // print_shorts("rho 0:",dl_ch_rho128); - // multiply by conjugated channel - mmtmpD0 = _mm_madd_epi16(dl_ch128[1],dl_ch128i[1]); - // mmtmpD0 contains real part of 4 consecutive outputs (32-bit) - mmtmpD1 = _mm_shufflelo_epi16(dl_ch128[1],_MM_SHUFFLE(2,3,0,1)); - mmtmpD1 = _mm_shufflehi_epi16(mmtmpD1,_MM_SHUFFLE(2,3,0,1)); - mmtmpD1 = _mm_sign_epi16(mmtmpD1,*(__m128i*)conjugate); - mmtmpD1 = _mm_madd_epi16(mmtmpD1,dl_ch128i[1]); - // mmtmpD1 contains imag part of 4 consecutive outputs (32-bit) - mmtmpD0 = _mm_srai_epi32(mmtmpD0,output_shift); - mmtmpD1 = _mm_srai_epi32(mmtmpD1,output_shift); - mmtmpD2 = _mm_unpacklo_epi32(mmtmpD0,mmtmpD1); - mmtmpD3 = _mm_unpackhi_epi32(mmtmpD0,mmtmpD1); - dl_ch_rho128[1] =_mm_packs_epi32(mmtmpD2,mmtmpD3); - - - if (pilots==0) { - - // multiply by conjugated channel - mmtmpD0 = _mm_madd_epi16(dl_ch128[2],dl_ch128i[2]); - // mmtmpD0 contains real part of 4 consecutive outputs (32-bit) - mmtmpD1 = _mm_shufflelo_epi16(dl_ch128[2],_MM_SHUFFLE(2,3,0,1)); - mmtmpD1 = _mm_shufflehi_epi16(mmtmpD1,_MM_SHUFFLE(2,3,0,1)); - mmtmpD1 = _mm_sign_epi16(mmtmpD1,*(__m128i*)conjugate); - mmtmpD1 = _mm_madd_epi16(mmtmpD1,dl_ch128i[2]); - // mmtmpD1 contains imag part of 4 consecutive outputs (32-bit) - mmtmpD0 = _mm_srai_epi32(mmtmpD0,output_shift); - mmtmpD1 = _mm_srai_epi32(mmtmpD1,output_shift); - mmtmpD2 = _mm_unpacklo_epi32(mmtmpD0,mmtmpD1); - mmtmpD3 = _mm_unpackhi_epi32(mmtmpD0,mmtmpD1); - dl_ch_rho128[2] = _mm_packs_epi32(mmtmpD2,mmtmpD3); - - dl_ch128+=3; - dl_ch128i+=3; - dl_ch_rho128+=3; - } else { - - dl_ch128+=2; - dl_ch128i+=2; - dl_ch_rho128+=2; - } - } - - } - - _mm_empty(); - _m_empty(); - -#elif defined(__arm__) - -#endif -} -*/ - void dlsch_detection_mrc(LTE_DL_FRAME_PARMS *frame_parms, int **rxdataF_comp, int **rxdataF_comp_i, @@ -3356,7 +3434,6 @@ void dlsch_detection_mrc(LTE_DL_FRAME_PARMS *frame_parms, #endif } - void dlsch_detection_mrc_TM34(LTE_DL_FRAME_PARMS *frame_parms, LTE_UE_PDSCH *pdsch_vars, int harq_pid, @@ -3435,8 +3512,6 @@ void dlsch_detection_mrc_TM34(LTE_DL_FRAME_PARMS *frame_parms, _m_empty(); } - - void dlsch_scale_channel(int **dl_ch_estimates_ext, LTE_DL_FRAME_PARMS *frame_parms, LTE_UE_DLSCH_t **dlsch_ue, @@ -3462,7 +3537,7 @@ void dlsch_scale_channel(int **dl_ch_estimates_ext, // Determine scaling amplitude based the symbol ch_amp = ((pilots) ? (dlsch_ue[0]->sqrt_rho_b) : (dlsch_ue[0]->sqrt_rho_a)); - + LOG_D(PHY,"Scaling PDSCH Chest in OFDM symbol %d by %d, pilots %d nb_rb %d NCP %d symbol %d\n",symbol_mod,ch_amp,pilots,nb_rb,frame_parms->Ncp,symbol); // printf("Scaling PDSCH Chest in OFDM symbol %d by %d\n",symbol_mod,ch_amp); @@ -3633,6 +3708,133 @@ void dlsch_channel_level(int **dl_ch_estimates_ext, #endif } +void dlsch_channel_level_core(int **dl_ch_estimates_ext, + int32_t *avg, + int n_tx, + int n_rx, + int length, + int start_point) +{ + +#if defined(__x86_64__)||defined(__i386__) + + short ii; + int aatx,aarx; + int length_mod4; + int length2; + __m128i *dl_ch128, avg128D; + + int16_t x = factor2(length); + int16_t y = (length)>>x; + + for (aatx=0; aatx<n_tx; aatx++) + for (aarx=0; aarx<n_rx; aarx++) { + //clear average level + //printf("aatx = %d, aarx = %d, aatx*frame_parms->nb_antennas_rx + aarx] = %d \n", aatx, aarx, aatx*frame_parms->nb_antennas_rx + aarx); + + avg128D = _mm_setzero_si128(); + // 5 is always a symbol with no pilots for both normal and extended prefix + + dl_ch128=(__m128i *)&dl_ch_estimates_ext[aatx*n_rx + aarx][start_point]; + + length_mod4=length&3; + + if (length_mod4 == 0){ + + length2 = length>>2; + + for (ii=0;ii<length2;ii++) { + //printf("rb %d : ",rb); + avg128D = _mm_add_epi32(avg128D,_mm_srai_epi16(_mm_madd_epi16(dl_ch128[0],dl_ch128[0]),x)); + avg128D = _mm_add_epi32(avg128D,_mm_srai_epi16(_mm_madd_epi16(dl_ch128[1],dl_ch128[1]),x)); + + //avg128D = _mm_add_epi32(avg128D,_mm_madd_epi16(dl_ch128[0],_mm_srai_epi16(_mm_mulhi_epi16(dl_ch128[0], coeff128),15))); + //avg128D = _mm_add_epi32(avg128D,_mm_madd_epi16(dl_ch128[1],_mm_srai_epi16(_mm_mulhi_epi16(dl_ch128[1], coeff128),15))); + + if (ii == 0){ + //print_shorts("dl_ch128",&dl_ch128[0]); + //print_shorts("dl_ch128",&dl_ch128[1]); + } + + dl_ch128+=2; + } + }else { + printf ("Channel level: Received number of subcarriers is not multiple of 4, \n" + "need to adapt the code!\n"); + } + + + avg[aatx*n_rx + aarx] =(((int32_t*)&avg128D)[0] + + ((int32_t*)&avg128D)[1] + + ((int32_t*)&avg128D)[2] + + ((int32_t*)&avg128D)[3])/y; + //printf("Channel level [%d]: %d\n",aatx*n_rx + aarx, avg[aatx*n_rx + aarx]); + } + + _mm_empty(); + _m_empty(); + +#elif defined(__arm__) + + short rb; + unsigned char aatx,aarx,nre=12,symbol_mod; + int32x4_t avg128D; + int16x4_t *dl_ch128; + + symbol_mod = (symbol>=(7-frame_parms->Ncp)) ? symbol-(7-frame_parms->Ncp) : symbol; + + for (aatx=0; aatx<frame_parms->nb_antenna_ports_eNB; aatx++) + for (aarx=0; aarx<frame_parms->nb_antennas_rx; aarx++) { + //clear average level + avg128D = vdupq_n_s32(0); + // 5 is always a symbol with no pilots for both normal and extended prefix + + dl_ch128=(int16x4_t *)&dl_ch_estimates_ext[aatx*frame_parms->nb_antennas_rx + aarx][symbol*frame_parms->N_RB_DL*12]; + + for (rb=0; rb<nb_rb; rb++) { + // printf("rb %d : ",rb); + // print_shorts("ch",&dl_ch128[0]); + avg128D = vqaddq_s32(avg128D,vmull_s16(dl_ch128[0],dl_ch128[0])); + avg128D = vqaddq_s32(avg128D,vmull_s16(dl_ch128[1],dl_ch128[1])); + avg128D = vqaddq_s32(avg128D,vmull_s16(dl_ch128[2],dl_ch128[2])); + avg128D = vqaddq_s32(avg128D,vmull_s16(dl_ch128[3],dl_ch128[3])); + + if (((symbol_mod == 0) || (symbol_mod == (frame_parms->Ncp-1)))&&(frame_parms->nb_antenna_ports_eNB!=1)) { + dl_ch128+=4; + } else { + avg128D = vqaddq_s32(avg128D,vmull_s16(dl_ch128[4],dl_ch128[4])); + avg128D = vqaddq_s32(avg128D,vmull_s16(dl_ch128[5],dl_ch128[5])); + dl_ch128+=6; + } + + /* + if (rb==0) { + print_shorts("dl_ch128",&dl_ch128[0]); + print_shorts("dl_ch128",&dl_ch128[1]); + print_shorts("dl_ch128",&dl_ch128[2]); + } + */ + } + + + if (((symbol_mod == 0) || (symbol_mod == (frame_parms->Ncp-1)))&&(frame_parms->nb_antenna_ports_eNB!=1)) + nre=8; + else if (((symbol_mod == 0) || (symbol_mod == (frame_parms->Ncp-1)))&&(frame_parms->nb_antenna_ports_eNB==1)) + nre=10; + else + nre=12; + + avg[aatx*frame_parms->nb_antennas_rx + aarx] = (((int32_t*)&avg128D)[0] + + ((int32_t*)&avg128D)[1] + + ((int32_t*)&avg128D)[2] + + ((int32_t*)&avg128D)[3])/(nb_rb*nre); + + //printf("Channel level : %d\n",avg[aatx*(frame_parms->nb_antennas_rx-1) + aarx]); + } +#endif + +} + //compute average channel_level of effective (precoded) channel //compute average channel_level of effective (precoded) channel diff --git a/openair1/PHY/LTE_UE_TRANSPORT/transport_proto_ue.h b/openair1/PHY/LTE_UE_TRANSPORT/transport_proto_ue.h index ade621faa2a3fb3b386e4df3e8a1782549f56219..178ed53dcc4441b9a06d65ebd5e616c109301d9a 100644 --- a/openair1/PHY/LTE_UE_TRANSPORT/transport_proto_ue.h +++ b/openair1/PHY/LTE_UE_TRANSPORT/transport_proto_ue.h @@ -816,6 +816,19 @@ void dlsch_channel_compensation(int32_t **rxdataF_ext, uint8_t output_shift, PHY_MEASUREMENTS *phy_measurements); +void dlsch_channel_compensation_core(int **rxdataF_ext, + int **dl_ch_estimates_ext, + int **dl_ch_mag, + int **dl_ch_magb, + int **rxdataF_comp, + int **rho, + unsigned char n_tx, + unsigned char n_rx, + unsigned char mod_order, + unsigned char output_shift, + int length, + int start_point); + void dlsch_dual_stream_correlation(LTE_DL_FRAME_PARMS *frame_parms, unsigned char symbol, unsigned short nb_rb, @@ -929,6 +942,13 @@ void dlsch_channel_level(int32_t **dl_ch_estimates_ext, uint8_t pilots_flag, uint16_t nb_rb); +void dlsch_channel_level_core(int32_t **dl_ch_estimates_ext, + int32_t *avg, + int n_tx, + int n_rx, + int length, + int start_point); + void dlsch_channel_level_TM34(int **dl_ch_estimates_ext, LTE_DL_FRAME_PARMS *frame_parms, @@ -1672,8 +1692,8 @@ double computeRhoB_UE(PDSCH_CONFIG_DEDICATED *pdsch_config_dedicated, LTE_UE_DLSCH_t *dlsch_ue); */ -uint8_t get_prach_prb_offset(LTE_DL_FRAME_PARMS *frame_parms, - uint8_t prach_ConfigIndex, +uint8_t get_prach_prb_offset(LTE_DL_FRAME_PARMS *frame_parms, + uint8_t prach_ConfigIndex, uint8_t n_ra_prboffset, uint8_t tdd_mapindex, uint16_t Nf);