diff --git a/cmake_targets/autotests/test_case_list.xml b/cmake_targets/autotests/test_case_list.xml index 50da90b5ec..3d15361430 100755 --- a/cmake_targets/autotests/test_case_list.xml +++ b/cmake_targets/autotests/test_case_list.xml @@ -159,9 +159,9 @@ -n100 -R106 -b106 -s5 -n100 -R217 -b217 -s5 -n100 -R273 -b273 -s5 - -n100 -s1 -t25 - -n100 -s1 -t33 - -n100 -s1 -t50 + -n100 -s1 -S2 -t25 + -n100 -s1 -S2 -t33 + -n100 -s5 -S7 -t50 -n100 -m0 -e0 -R25 -b25 -i 2 1 0 test1 test2 test3 test4 test5 test6 test7 PDSCH test OK @@ -193,15 +193,17 @@ (Test3: 106 MCS-TABLE 256 QAM MCS Index 26), (Test4: MCS 0, low SNR performance), (Test5: 4x4 MIMO, 1 Layer), - (Test6: 4x4 MIMO, 2 Layers) + (Test6: 4x4 MIMO, 2 Layers), + (Test7: 4x4 MIMO, 4 Layers) nr_dlsim -n100 -e27 -s30 - -n100 -e16 -s10 + -n100 -e16 -s11 -S13 -n100 -q1 -e26 -s30 -n100 -e0 -t95 -S-1.0 -i 2 1 0 -n10 -s20 -U 3 0 0 2 -gR -x1 -y4 -z4 - -n10 -s20 -U 3 0 0 2 -gR -x2 -y4 -z4 - test1 test2 test3 test4 test5 test6 + -n10 -s20 -U 3 0 0 2 -gR -x2 -y4 -z4 + -n10 -s20 -U 3 0 0 2 -x4 -y4 -z4 + test1 test2 test3 test4 test5 test6 test7 PDSCH test OK segmentation fault|assertion|exiting|fatal 3 @@ -218,9 +220,9 @@ -n100 -s5 -T 2 2 2 -n100 -s5 -T 2 1 2 -n100 -s5 -T 2 0 4 - -n100 -s2 -U 2 0 1 - -n100 -s2 -U 2 0 2 - -n100 -s2 -U 2 1 3 + -n100 -s5 -S7 -U 2 0 1 + -n100 -s5 -S7 -U 2 0 2 + -n100 -s5 -S7 -U 2 1 3 test1 test2 test3 test4 test5 test6 PDSCH test OK segmentation fault|assertion|exiting|fatal diff --git a/nfapi/open-nFAPI/nfapi/public_inc/fapi_nr_ue_interface.h b/nfapi/open-nFAPI/nfapi/public_inc/fapi_nr_ue_interface.h index 0ddc0986bd..94bf1b02dc 100644 --- a/nfapi/open-nFAPI/nfapi/public_inc/fapi_nr_ue_interface.h +++ b/nfapi/open-nFAPI/nfapi/public_inc/fapi_nr_ue_interface.h @@ -87,6 +87,7 @@ typedef struct { typedef struct { uint16_t rnti; uint8_t dci_format; + uint8_t coreset_type; int ss_type; // n_CCE index of first CCE for PDCCH reception int n_CCE; diff --git a/openair1/PHY/NR_ESTIMATION/nr_ul_channel_estimation.c b/openair1/PHY/NR_ESTIMATION/nr_ul_channel_estimation.c index 53ca4bf762..ebc225d82b 100644 --- a/openair1/PHY/NR_ESTIMATION/nr_ul_channel_estimation.c +++ b/openair1/PHY/NR_ESTIMATION/nr_ul_channel_estimation.c @@ -249,69 +249,21 @@ int nr_pusch_channel_estimation(PHY_VARS_gNB *gNB, } #endif - } else if (pusch_pdu->dmrs_config_type == pusch_dmrs_type2 && chest_freq == 0) { //pusch_dmrs_type2 |p_r,p_l,d,d,d,d,p_r,p_l,d,d,d,d| - LOG_D(PHY,"PUSCH estimation DMRS type 2, Freq-domain interpolation"); - c16_t *pil = pilot; - int re_offset = k0; - // Treat first DMRS specially (left edge) - *ul_ch=c16mulShift(*pil, - rxdataF[soffset+nushift+re_offset], - 15); - pil++; - ul_ch++; - re_offset = (re_offset + 1)%symbolSize; - ch_offset++; - - // TO verify: used after the loop, likely a piece of code is missing for ch_r - c16_t ch_r; - for (int re_cnt = 1; re_cnt < (nb_rb_pusch*NR_NB_SC_PER_RB) - 5; re_cnt += 6) { - c16_t ch_l=c16mulShift(*pil, - rxdataF[soffset+nushift+re_offset], - 15); - *ul_ch = ch_l; + } else if (pusch_pdu->dmrs_config_type == pusch_dmrs_type2 && chest_freq == 0) { // pusch_dmrs_type2 |p_r,p_l,d,d,d,d,p_r,p_l,d,d,d,d| + LOG_D(PHY, "PUSCH estimation DMRS type 2, Freq-domain interpolation\n"); + c16_t *pil = pilot; + c16_t *rx = &rxdataF[soffset + nushift]; + for (int n = 0; n < nb_rb_pusch * NR_NB_SC_PER_RB; n += 6) { + c16_t ch0 = c16mulShift(*pil, rx[(k0 + n) % symbolSize], 15); pil++; - ul_ch++; - ch_offset++; - multadd_real_four_symbols_vector_complex_scalar(filt8_ml2, - &ch_l, - ul_ch); - *max_ch = max(*max_ch, max(abs(ch_l.r), abs(ch_l.i))); - - re_offset = (re_offset+5)%symbolSize; - ch_r=c16mulShift(*pil, - rxdataF[soffset+nushift+re_offset], - 15); - multadd_real_four_symbols_vector_complex_scalar(filt8_mr2, - &ch_r, - ul_ch); - *max_ch = max(*max_ch, max(abs(ch_r.r), abs(ch_r.i))); - - //for (int re_idx = 0; re_idx < 8; re_idx += 2) - //printf("ul_ch = %d + j*%d\n", ul_ch[re_idx], ul_ch[re_idx+1]); - ul_ch += 4; - ch_offset += 4; - *ul_ch = ch_r; + c16_t ch1 = c16mulShift(*pil, rx[(k0 + n + 1) % symbolSize], 15); pil++; - ul_ch++; - ch_offset++; - re_offset = (re_offset + 1)%symbolSize; + c16_t ch = c16addShift(ch0, ch1, 1); + *max_ch = max(*max_ch, max(abs(ch.r), abs(ch.i))); + multadd_real_four_symbols_vector_complex_scalar(filt8_rep4, &ch, &ul_ch[n]); + ul_ch[n + 4] = ch; + ul_ch[n + 5] = ch; } - - // Treat last pilot specially (right edge) - c16_t ch_l=c16mulShift(*pil, - rxdataF[soffset+nushift+re_offset], - 15); - *ul_ch = ch_l; - ul_ch++; - ch_offset++; - multadd_real_four_symbols_vector_complex_scalar(filt8_rr1, - &ch_l, - ul_ch); - multadd_real_four_symbols_vector_complex_scalar(filt8_rr2, - &ch_r, - ul_ch); - __m128i *ul_ch_128 = (__m128i *)&ul_ch_estimates[p*gNB->frame_parms.nb_antennas_rx+aarx][ch_offset]; - ul_ch_128[0] = _mm_slli_epi16 (ul_ch_128[0], 2); } else if (pusch_pdu->dmrs_config_type == pusch_dmrs_type1) { // this is case without frequency-domain linear interpolation, just take average of LS channel estimates of 6 DMRS REs and use a common value for the whole PRB diff --git a/openair1/PHY/NR_UE_ESTIMATION/filt16a_32.c b/openair1/PHY/NR_UE_ESTIMATION/filt16a_32.c index d2a25b1258..52ede9018e 100644 --- a/openair1/PHY/NR_UE_ESTIMATION/filt16a_32.c +++ b/openair1/PHY/NR_UE_ESTIMATION/filt16a_32.c @@ -342,3 +342,26 @@ short filt16_ul_middle[16] = {2048, 2048, 2048, 2048, 2048, 2048, 2048, 2048, short filt16_ul_last[16] = {4096, 4096, 4096, 4096, 8192, 8192, 8192, 8192, 0, 0, 0, 0, 0, 0, 0, 0}; + +short filt8_rep4[8] = {16384, 16384, 16384, 16384, 0, 0, 0, 0}; + +// DL +// DMRS_Type1 +short filt16_dl_first[16] = {12228, 12228, 12228, 12228, 8192, 8192, 8192, 8192, + 4096, 4096, 4096, 4096, 0, 0, 0, 0}; + +short filt16_dl_middle[16] = {2048, 2048, 2048, 2048, 2048, 2048, 2048, 2048, + 2048, 2048, 2048, 2048, 2048, 2048, 2048, 2048}; + +short filt16_dl_last[16] = {4096, 4096, 4096, 4096, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0}; + +// DMRS_Type2 +short filt16_dl_first_type2[16] = {16384, 16384, 16384, 8192, 8192, 8192, 8192, 8192, + 8192, 0, 0, 0, 0, 0, 0}; + +short filt16_dl_middle_type2[16] = {8192, 8192, 8192, 8192, 8192, 8192, 8192, 8192, + 8192, 8192, 8192, 8192, 0, 0, 0, 0}; + +short filt16_dl_last_type2[16] = {8192, 8192, 8192, 8192, 8192, 8192, 16384, 16384, + 16384, 0, 0, 0, 0, 0, 0, 0}; diff --git a/openair1/PHY/NR_UE_ESTIMATION/filt16a_32.h b/openair1/PHY/NR_UE_ESTIMATION/filt16a_32.h index 8a3b0a9870..b7bd6aec18 100644 --- a/openair1/PHY/NR_UE_ESTIMATION/filt16a_32.h +++ b/openair1/PHY/NR_UE_ESTIMATION/filt16a_32.h @@ -224,4 +224,16 @@ extern short filt16_ul_p0[16]; extern short filt16_ul_p1p2[16]; extern short filt16_ul_middle[16]; extern short filt16_ul_last[16]; +extern short filt8_rep4[8]; + +/*DL*/ +// DL DMRS_Type1 +extern short filt16_dl_first[16]; +extern short filt16_dl_middle[16]; +extern short filt16_dl_last[16]; + +// DL DMRS_Type2 +extern short filt16_dl_first_type2[16]; +extern short filt16_dl_middle_type2[16]; +extern short filt16_dl_last_type2[16]; #endif diff --git a/openair1/PHY/NR_UE_ESTIMATION/nr_dl_channel_estimation.c b/openair1/PHY/NR_UE_ESTIMATION/nr_dl_channel_estimation.c index 6b988393d3..814de26b88 100644 --- a/openair1/PHY/NR_UE_ESTIMATION/nr_dl_channel_estimation.c +++ b/openair1/PHY/NR_UE_ESTIMATION/nr_dl_channel_estimation.c @@ -1230,6 +1230,360 @@ void nr_pdcch_channel_estimation(PHY_VARS_NR_UE *ue, } } +void NFAPI_NR_DMRS_TYPE1_linear_interp(NR_DL_FRAME_PARMS *frame_parms, + c16_t *rxF, + c16_t *pil, + c16_t *dl_ch, + unsigned short bwp_start_subcarrier, + unsigned short nb_rb_pdsch, + int8_t delta) +{ + int re_offset = bwp_start_subcarrier % frame_parms->ofdm_symbol_size; + c16_t *pil0 = pil; + c16_t *dl_ch0 = dl_ch; + c16_t ch = {0}; + + int16_t *fdcl = NULL; + int16_t *fdcr = NULL; + int16_t *fdclh = NULL; + int16_t *fdcrh = NULL; + switch (delta) { + case 0: // port 0,1 + fdcl = filt8_dcl0; // left DC interpolation Filter (even RB) + fdcr = filt8_dcr0; // right DC interpolation Filter (even RB) + fdclh = filt8_dcl0_h; // left DC interpolation Filter (odd RB) + fdcrh = filt8_dcr0_h; // right DC interpolation Filter (odd RB) + break; + + case 1: // port2,3 + fdcl = filt8_dcl1; + fdcr = filt8_dcr1; + fdclh = filt8_dcl1_h; + fdcrh = filt8_dcr1_h; + break; + + default: + AssertFatal(1 == 0, "pdsch_channel_estimation: Invalid delta %i\n", delta); + break; + } + + for (int pilot_cnt = 0; pilot_cnt < 6 * nb_rb_pdsch; pilot_cnt++) { + if (pilot_cnt % 2 == 0) { + ch = c16mulShift(*pil, rxF[re_offset], 15); + pil++; + re_offset = (re_offset + 2) % frame_parms->ofdm_symbol_size; + ch = c16maddShift(*pil, rxF[re_offset], ch, 15); + ch = c16Shift(ch, 1); + pil++; + re_offset = (re_offset + 2) % frame_parms->ofdm_symbol_size; + } + +#ifdef DEBUG_PDSCH + printf("pilot %3u: pil -> (%6d,%6d), rxF -> (%4d,%4d), ch -> (%4d,%4d) \n", pilot_cnt, pil->r, pil->i, rxF[re_offset].r, rxF[re_offset].i, ch.r, ch.i); +#endif + + if (pilot_cnt == 0) { // Treat first pilot + c16multaddVectRealComplex(filt16_dl_first, &ch, dl_ch, 16); + } else if (pilot_cnt == 6 * nb_rb_pdsch - 1) { // Treat last pilot + c16multaddVectRealComplex(filt16_dl_last, &ch, dl_ch, 16); + } else { // Treat middle pilots + c16multaddVectRealComplex(filt16_dl_middle, &ch, dl_ch, 16); + if (pilot_cnt % 2 == 0) { + dl_ch += 4; + } + } + } + + // check if PRB crosses DC and improve estimates around DC + if ((bwp_start_subcarrier < frame_parms->ofdm_symbol_size) && (bwp_start_subcarrier + nb_rb_pdsch * 12 >= frame_parms->ofdm_symbol_size)) { + dl_ch = dl_ch0; + uint16_t idxDC = 2 * (frame_parms->ofdm_symbol_size - bwp_start_subcarrier); + uint16_t idxPil = idxDC / 4; + re_offset = bwp_start_subcarrier % frame_parms->ofdm_symbol_size; + pil = pil0; + pil += (idxPil - 1); + dl_ch += (idxDC / 2 - 2); + dl_ch = memset(dl_ch, 0, sizeof(c16_t) * 5); + re_offset = (re_offset + idxDC / 2 - 2) % frame_parms->ofdm_symbol_size; + ch = c16mulShift(*pil, rxF[re_offset], 15); + pil++; + re_offset = (re_offset + 2) % frame_parms->ofdm_symbol_size; + ch = c16maddShift(*pil, rxF[re_offset], ch, 15); + ch = c16Shift(ch, 1); + + // for proper allignment of SIMD vectors + if ((frame_parms->N_RB_DL & 1) == 0) { + c16multaddVectRealComplex(fdcl, &ch, dl_ch - 2, 8); + pil++; + re_offset = (re_offset + 2) % frame_parms->ofdm_symbol_size; + ch = c16mulShift(*pil, rxF[re_offset], 15); + pil++; + re_offset = (re_offset + 2) % frame_parms->ofdm_symbol_size; + ch = c16maddShift(*pil, rxF[re_offset], ch, 15); + ch = c16Shift(ch, 1); + c16multaddVectRealComplex(fdcr, &ch, dl_ch - 2, 8); + } else { + c16multaddVectRealComplex(fdclh, &ch, dl_ch, 8); + pil++; + re_offset = (re_offset + 2) % frame_parms->ofdm_symbol_size; + ch = c16mulShift(*pil, rxF[re_offset], 15); + pil++; + re_offset = (re_offset + 2) % frame_parms->ofdm_symbol_size; + ch = c16maddShift(*pil, rxF[re_offset], ch, 15); + ch = c16Shift(ch, 1); + c16multaddVectRealComplex(fdcrh, &ch, dl_ch, 8); + } + } +} + +void NFAPI_NR_DMRS_TYPE1_average_prb(NR_DL_FRAME_PARMS *frame_parms, + c16_t *rxF, + c16_t *pil, + c16_t *dl_ch, + unsigned short bwp_start_subcarrier, + unsigned short nb_rb_pdsch) +{ + int re_offset = bwp_start_subcarrier % frame_parms->ofdm_symbol_size; + c16_t ch = {0}; + int P_average = 6; + + c32_t ch32 = {0}; + for (int p_av = 0; p_av < P_average; p_av++) { + ch32 = c32x16maddShift(*pil, rxF[re_offset], ch32, 15); + pil++; + re_offset = (re_offset + 2) % frame_parms->ofdm_symbol_size; + } + ch = c16x32div(ch32, P_average); + +#if NO_INTERP + for (int i = 0; i < 2 * P_average; i++) { + dl_ch[i] = ch; + } + dl_ch += 2 * P_average; +#else + c16multaddVectRealComplex(filt8_avlip0, &ch, dl_ch, 8); + dl_ch += 16; + c16multaddVectRealComplex(filt8_avlip1, &ch, dl_ch, 8); + dl_ch += 16; + c16multaddVectRealComplex(filt8_avlip2, &ch, dl_ch, 8); + dl_ch -= 24; +#endif + + for (int pilot_cnt = P_average; pilot_cnt < 6 * (nb_rb_pdsch - 1); pilot_cnt += P_average) { + ch32.r = 0; + ch32.i = 0; + for (int p_av = 0; p_av < P_average; p_av++) { + ch32 = c32x16maddShift(*pil, rxF[re_offset], ch32, 15); + pil++; + re_offset = (re_offset + 2) % frame_parms->ofdm_symbol_size; + } + ch = c16x32div(ch32, P_average); + +#if NO_INTERP + for (int i = 0; i < 2 * P_average; i++) { + dl_ch[i] = ch; + } + dl_ch += 2 * P_average; +#else + dl_ch[3].r += (ch.r * 1365) >> 15; // 1/12*16384 + dl_ch[3].i += (ch.i * 1365) >> 15; // 1/12*16384 + dl_ch += 4; + c16multaddVectRealComplex(filt8_avlip3, &ch, dl_ch, 8); + dl_ch += 8; + c16multaddVectRealComplex(filt8_avlip4, &ch, dl_ch, 8); + dl_ch += 8; + c16multaddVectRealComplex(filt8_avlip5, &ch, dl_ch, 8); + dl_ch -= 8; +#endif + } + + ch32.r = 0; + ch32.i = 0; + for (int p_av = 0; p_av < P_average; p_av++) { + ch32 = c32x16maddShift(*pil, rxF[re_offset], ch32, 15); + pil++; + re_offset = (re_offset + 2) % frame_parms->ofdm_symbol_size; + } + ch = c16x32div(ch32, P_average); + +#if NO_INTERP + for (int i = 0; i < 2 * P_average; i++) { + dl_ch[i] = ch; + } + dl_ch += 2 * P_average; +#else + dl_ch[3].r += (ch.r * 1365) >> 15; // 1/12*16384 + dl_ch[3].i += (ch.i * 1365) >> 15; // 1/12*16384 + dl_ch += 4; + c16multaddVectRealComplex(filt8_avlip3, &ch, dl_ch, 8); + dl_ch += 8; + c16multaddVectRealComplex(filt8_avlip6, &ch, dl_ch, 8); +#endif +} + +void NFAPI_NR_DMRS_TYPE2_linear_interp(NR_DL_FRAME_PARMS *frame_parms, + c16_t *rxF, + c16_t *pil, + c16_t *dl_ch, + unsigned short bwp_start_subcarrier, + unsigned short nb_rb_pdsch, + int8_t delta, + unsigned short p) +{ + int re_offset = bwp_start_subcarrier % frame_parms->ofdm_symbol_size; + c16_t *dl_ch0 = dl_ch; + c16_t ch = {0}; + c16_t ch_l = {0}; + c16_t ch_r = {0}; + + for (int pilot_cnt = 0; pilot_cnt < 4 * nb_rb_pdsch; pilot_cnt += 2) { + ch_l = c16mulShift(*pil, rxF[re_offset], 15); +#ifdef DEBUG_PDSCH + printf("pilot %3u: pil -> (%6d,%6d), rxF -> (%4d,%4d), ch -> (%4d,%4d) \n", pilot_cnt, pil->r, pil->i, rxF[re_offset].r, rxF[re_offset].i, ch_l.r, ch_l.i); +#endif + pil++; + re_offset = (re_offset + 1) % frame_parms->ofdm_symbol_size; + ch_r = c16mulShift(*pil, rxF[re_offset], 15); +#ifdef DEBUG_PDSCH + printf("pilot %3u: pil -> (%6d,%6d), rxF -> (%4d,%4d), ch -> (%4d,%4d) \n", pilot_cnt, pil->r, pil->i, rxF[re_offset].r, rxF[re_offset].i, ch_r.r, ch_r.i); +#endif + ch = c16addShift(ch_l, ch_r, 1); + if (pilot_cnt == 1) { + multadd_real_vector_complex_scalar(filt16_dl_first_type2, (int16_t *)&ch, (int16_t *)dl_ch, 16); + dl_ch += 3; + } else if (pilot_cnt == (4 * nb_rb_pdsch - 1)) { + multadd_real_vector_complex_scalar(filt16_dl_last_type2, (int16_t *)&ch, (int16_t *)dl_ch, 16); + } else { + multadd_real_vector_complex_scalar(filt16_dl_middle_type2, (int16_t *)&ch, (int16_t *)dl_ch, 16); + dl_ch += 6; + } + pil++; + re_offset = (re_offset + 5) % frame_parms->ofdm_symbol_size; + } + + // check if PRB crosses DC and improve estimates around DC + if ((bwp_start_subcarrier < frame_parms->ofdm_symbol_size) && (bwp_start_subcarrier + nb_rb_pdsch * 12 >= frame_parms->ofdm_symbol_size) && (p < 2)) { + dl_ch = dl_ch0; + uint16_t idxDC = 2 * (frame_parms->ofdm_symbol_size - bwp_start_subcarrier); + dl_ch += (idxDC / 2 - 4); + dl_ch = memset(dl_ch, 0, sizeof(c16_t) * 10); + + dl_ch--; + ch_r = *dl_ch; + dl_ch += 11; + ch_l = *dl_ch; + + // for proper allignment of SIMD vectors + if ((frame_parms->N_RB_DL & 1) == 0) { + dl_ch -= 10; + // Interpolate fdcrl1 with ch_r + c16multaddVectRealComplex(filt8_dcrl1, &ch_r, dl_ch, 8); + // Interpolate fdclh1 with ch_l + c16multaddVectRealComplex(filt8_dclh1, &ch_l, dl_ch, 8); + dl_ch += 8; + // Interpolate fdcrh1 with ch_r + c16multaddVectRealComplex(filt8_dcrh1, &ch_r, dl_ch, 8); + // Interpolate fdcll1 with ch_l + c16multaddVectRealComplex(filt8_dcll1, &ch_l, dl_ch, 8); + } else { + dl_ch -= 14; + // Interpolate fdcrl1 with ch_r + c16multaddVectRealComplex(filt8_dcrl2, &ch_r, dl_ch, 8); + // Interpolate fdclh1 with ch_l + c16multaddVectRealComplex(filt8_dclh2, &ch_l, dl_ch, 8); + dl_ch += 8; + // Interpolate fdcrh1 with ch_r + c16multaddVectRealComplex(filt8_dcrh2, &ch_r, dl_ch, 8); + // Interpolate fdcll1 with ch_l + c16multaddVectRealComplex(filt8_dcll2, &ch_l, dl_ch, 8); + } + } +} + +void NFAPI_NR_DMRS_TYPE2_average_prb(NR_DL_FRAME_PARMS *frame_parms, + c16_t *rxF, + c16_t *pil, + c16_t *dl_ch, + unsigned short bwp_start_subcarrier, + unsigned short nb_rb_pdsch) +{ + int re_offset = bwp_start_subcarrier % frame_parms->ofdm_symbol_size; + c16_t ch = {0}; + int P_average = 4; + + c32_t ch32 = {0}; + for (int p_av = 0; p_av < P_average; p_av++) { + ch32 = c32x16maddShift(*pil, rxF[re_offset], ch32, 15); + pil++; + re_offset = (re_offset + 1) % frame_parms->ofdm_symbol_size; + } + ch = c16x32div(ch32, P_average); + +#if NO_INTERP + for (int i = 0; i < 3 * P_average; i++) { + dl_ch[i] = ch; + } + dl_ch += 3 * P_average; +#else + c16multaddVectRealComplex(filt8_avlip0, &ch, dl_ch, 8); + dl_ch += 8; + c16multaddVectRealComplex(filt8_avlip1, &ch, dl_ch, 8); + dl_ch += 8; + c16multaddVectRealComplex(filt8_avlip2, &ch, dl_ch, 8); + dl_ch -= 12; +#endif + + for (int pilot_cnt = P_average; pilot_cnt < 4 * (nb_rb_pdsch - 1); pilot_cnt += P_average) { + ch32.r = 0; + ch32.i = 0; + for (int p_av = 0; p_av < P_average; p_av++) { + ch32 = c32x16maddShift(*pil, rxF[re_offset], ch32, 15); + pil++; + re_offset = (re_offset + 5) % frame_parms->ofdm_symbol_size; + } + ch = c16x32div(ch32, P_average); + +#if NO_INTERP + for (int i = 0; i < 3 * P_average; i++) { + dl_ch[i] = ch; + } + dl_ch += 3 * P_average; +#else + dl_ch[3].r += (ch.r * 1365) >> 15; // 1/12*16384 + dl_ch[3].i += (ch.i * 1365) >> 15; // 1/12*16384 + dl_ch += 4; + c16multaddVectRealComplex(filt8_avlip3, &ch, dl_ch, 8); + dl_ch += 8; + c16multaddVectRealComplex(filt8_avlip4, &ch, dl_ch, 8); + dl_ch += 8; + c16multaddVectRealComplex(filt8_avlip5, &ch, dl_ch, 8); + dl_ch -= 8; +#endif + } + + ch32.r = 0; + ch32.i = 0; + for (int p_av = 0; p_av < P_average; p_av++) { + ch32 = c32x16maddShift(*pil, rxF[re_offset], ch32, 15); + pil++; + re_offset = (re_offset + 5) % frame_parms->ofdm_symbol_size; + } + ch = c16x32div(ch32, P_average); + +#if NO_INTERP + for (int i = 0; i < 3 * P_average; i++) { + dl_ch[i] = ch; + } + dl_ch += 3 * P_average; +#else + dl_ch[3].r += (ch.r * 1365) >> 15; // 1/12*16384 + dl_ch[3].i += (ch.i * 1365) >> 15; // 1/12*16384 + dl_ch += 4; + c16multaddVectRealComplex(filt8_avlip3, &ch, dl_ch, 8); + dl_ch += 8; + c16multaddVectRealComplex(filt8_avlip6, &ch, dl_ch, 8); +#endif +} int nr_pdsch_channel_estimation(PHY_VARS_NR_UE *ue, UE_nr_rxtx_proc_t *proc, bool is_SI, @@ -1243,35 +1597,23 @@ int nr_pdsch_channel_estimation(PHY_VARS_NR_UE *ue, unsigned short nb_rb_pdsch, uint32_t pdsch_est_size, int32_t dl_ch_estimates[][pdsch_est_size], - c16_t rxdataF[][ue->frame_parms.samples_per_slot_wCP]) + int rxdataFsize, + c16_t rxdataF[][rxdataFsize]) { int gNB_id = proc->gNB_id; int Ns = proc->nr_slot_rx; - int pilot[3280] __attribute__((aligned(16))); - unsigned char aarx; - unsigned short k; - unsigned int pilot_cnt; - int16_t ch_l[2],ch_r[2],ch[2],*pil,*rxF,*dl_ch; - int16_t *fl=NULL,*fm=NULL,*fr=NULL,*fml=NULL,*fmr=NULL,*fmm=NULL,*fdcl=NULL,*fdcr=NULL,*fdclh=NULL,*fdcrh=NULL, *frl=NULL, *frr=NULL; - int ch_offset,symbol_offset; - - uint8_t nushift; - ch_offset = ue->frame_parms.ofdm_symbol_size*symbol; - - symbol_offset = ue->frame_parms.ofdm_symbol_size*symbol; - - k = bwp_start_subcarrier; - int re_offset = k; + const int ch_offset = ue->frame_parms.ofdm_symbol_size * symbol; + const int symbol_offset = ue->frame_parms.ofdm_symbol_size * symbol; #ifdef DEBUG_PDSCH - printf("PDSCH Channel Estimation : gNB_id %d ch_offset %d, symbol_offset %d OFDM size %d, Ncp=%d, Ns=%d, k=%d symbol %d\n", + printf("PDSCH Channel Estimation : gNB_id %d ch_offset %d, symbol_offset %d OFDM size %d, Ncp=%d, Ns=%d, bwp_start_subcarrier=%d symbol %d\n", gNB_id, ch_offset, symbol_offset, ue->frame_parms.ofdm_symbol_size, ue->frame_parms.Ncp, Ns, - k, + bwp_start_subcarrier, symbol); #endif @@ -1283,95 +1625,26 @@ int nr_pdsch_channel_estimation(PHY_VARS_NR_UE *ue, int8_t delta = get_delta(p, config_type); // checking if re-initialization of scrambling IDs is needed - if (scrambling_id != ue->scramblingID_dlsch[nscid]){ + if (scrambling_id != ue->scramblingID_dlsch[nscid]) { ue->scramblingID_dlsch[nscid] = scrambling_id; nr_gold_pdsch(ue, nscid, scrambling_id); } - nr_pdsch_dmrs_rx(ue, Ns, ue->nr_gold_pdsch[gNB_id][Ns][symbol][0], &pilot[0], 1000+p, 0, nb_rb_pdsch+rb_offset, config_type); + c16_t pilot[3280] __attribute__((aligned(16))); + nr_pdsch_dmrs_rx(ue, Ns, ue->nr_gold_pdsch[gNB_id][Ns][symbol][0], (int32_t *)pilot, 1000 + p, 0, nb_rb_pdsch + rb_offset, config_type); - if (config_type == NFAPI_NR_DMRS_TYPE1){ - nushift = (p>>1)&1; - if (p<4) ue->frame_parms.nushift = nushift; - switch (delta) { - - case 0://port 0,1 - fl = filt8_l0;//left interpolation Filter for DMRS config. 1 - fm = filt8_m0;//left middle interpolation Filter - fr = filt8_r0;//right interpolation Filter - fmm = filt8_mm0;;//middle middle interpolation Filter - fml = filt8_m0;//left middle interpolation Filter - fmr = filt8_mr0;//middle right interpolation Filter - fdcl = filt8_dcl0;//left DC interpolation Filter (even RB) - fdcr = filt8_dcr0;//right DC interpolation Filter (even RB) - fdclh = filt8_dcl0_h;//left DC interpolation Filter (odd RB) - fdcrh = filt8_dcr0_h;//right DC interpolation Filter (odd RB) - frl = NULL; - frr = NULL; - break; - - case 1://port2,3 - fl = filt8_l1; - fm = filt8_m1; - fr = filt8_r1; - fmm = filt8_mm1; - fml = filt8_ml1; - fmr = filt8_m1; - fdcl = filt8_dcl1; - fdcr = filt8_dcr1; - fdclh = filt8_dcl1_h; - fdcrh = filt8_dcr1_h; - frl = NULL; - frr = NULL; - break; - - default: - LOG_E(PHY,"pdsch_channel_estimation: nushift=%d -> ERROR\n",nushift); - return -1; - break; - } - } else {//NFAPI_NR_DMRS_TYPE2 + uint8_t nushift = 0; + if (config_type == NFAPI_NR_DMRS_TYPE1) { + nushift = (p >> 1) & 1; + if (p < 4) + ue->frame_parms.nushift = nushift; + } else { // NFAPI_NR_DMRS_TYPE2 nushift = delta; - if (p<6) ue->frame_parms.nushift = nushift; - switch (delta) { - case 0://port 0,1 - fl = filt8_l2;//left interpolation Filter should be fml - fr = filt8_r2;//right interpolation Filter should be fmr - fm = filt8_l2; - fmm = filt8_r2; - fml = filt8_ml2; - fmr = filt8_mr2; - frl = filt8_rl2; - frr = filt8_rm2; - fdcl = filt8_dcl1; - fdcr = filt8_dcr1; - fdclh = filt8_dcl1_h; - fdcrh = filt8_dcr1_h; - break; - - case 2://port2,3 - fl = filt8_l3; - fm = filt8_m2; - fr = filt8_r3; - fmm = filt8_mm2; - fml = filt8_l2; - fmr = filt8_r2; - frl = filt8_rl3; - frr = filt8_rr3; - fdcl = NULL; - fdcr = NULL; - fdclh = NULL; - fdcrh = NULL; - break; - - default: - LOG_E(PHY,"pdsch_channel_estimation: nushift=%d -> ERROR\n",nushift); - return -1; - break; - } + if (p < 6) + ue->frame_parms.nushift = nushift; } - for (aarx = 0; aarx < ue->frame_parms.nb_antennas_rx; aarx++) { + for (int aarx = 0; aarx < ue->frame_parms.nb_antennas_rx; aarx++) { #ifdef DEBUG_PDSCH printf("\n============================================\n"); @@ -1379,643 +1652,57 @@ int nr_pdsch_channel_estimation(PHY_VARS_NR_UE *ue, printf("============================================\n"); #endif - pil = (int16_t *)&pilot[rb_offset * ((config_type == NFAPI_NR_DMRS_TYPE1) ? 6 : 4)]; - k = k % ue->frame_parms.ofdm_symbol_size; - re_offset = k; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset + re_offset + nushift)]; - dl_ch = (int16_t *)&dl_ch_estimates[p * ue->frame_parms.nb_antennas_rx + aarx][ch_offset]; - memset(dl_ch, 0, 4 * (ue->frame_parms.ofdm_symbol_size)); + c16_t *rxF = &rxdataF[aarx][symbol_offset + nushift]; + c16_t *dl_ch = (c16_t *)&dl_ch_estimates[p * ue->frame_parms.nb_antennas_rx + aarx][ch_offset]; + memset(dl_ch, 0, sizeof(*dl_ch) * ue->frame_parms.ofdm_symbol_size); if (config_type == NFAPI_NR_DMRS_TYPE1 && ue->chest_freq == 0) { + NFAPI_NR_DMRS_TYPE1_linear_interp(&ue->frame_parms, + rxF, + &pilot[6 * rb_offset], + dl_ch, + bwp_start_subcarrier, + nb_rb_pdsch, + delta); - for (pilot_cnt = 0; pilot_cnt < 6 * nb_rb_pdsch; pilot_cnt++) { + } else if (config_type == NFAPI_NR_DMRS_TYPE2 && ue->chest_freq == 0) { + NFAPI_NR_DMRS_TYPE2_linear_interp(&ue->frame_parms, + rxF, + &pilot[4 * rb_offset], + dl_ch, + bwp_start_subcarrier, + nb_rb_pdsch, + delta, + p); - if (pilot_cnt % 2 == 0) { - ch[0] = (int16_t)(((int32_t)pil[0] * rxF[0] - (int32_t)pil[1] * rxF[1]) >> 15); - ch[1] = (int16_t)(((int32_t)pil[0] * rxF[1] + (int32_t)pil[1] * rxF[0]) >> 15); - pil += 2; - re_offset = (re_offset + 2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset + nushift + re_offset)]; - ch[0] += (int16_t)(((int32_t)pil[0] * rxF[0] - (int32_t)pil[1] * rxF[1]) >> 15); - ch[1] += (int16_t)(((int32_t)pil[0] * rxF[1] + (int32_t)pil[1] * rxF[0]) >> 15); - ch[0] >>= 1; - ch[1] >>= 1; - pil += 2; - re_offset = (re_offset + 2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset + nushift + re_offset)]; - } + } else if (config_type == NFAPI_NR_DMRS_TYPE1) { + NFAPI_NR_DMRS_TYPE1_average_prb(&ue->frame_parms, + rxF, + &pilot[6 * rb_offset], + dl_ch, + bwp_start_subcarrier, + nb_rb_pdsch); -#ifdef DEBUG_PDSCH - printf("pilot %3u: pil -> (%6d,%6d), rxF -> (%4d,%4d), ch -> (%4d,%4d) \n", pilot_cnt, pil[0], pil[1], rxF[0], rxF[1], ch[0], ch[1]); -#endif - - if (pilot_cnt == 0) { - multadd_real_vector_complex_scalar(fl, ch, dl_ch, 8); - } else if (pilot_cnt == 1) { - multadd_real_vector_complex_scalar(fml, ch, dl_ch, 8); - } else if (pilot_cnt == 6 * nb_rb_pdsch - 2) { - multadd_real_vector_complex_scalar(fmr, ch, dl_ch, 8); - dl_ch += 8; - } else if (pilot_cnt == 6 * nb_rb_pdsch - 1) { - multadd_real_vector_complex_scalar(fr, ch, dl_ch, 8); - } else if (pilot_cnt % 2 == 0) { - multadd_real_vector_complex_scalar(fmm, ch, dl_ch, 8); - dl_ch += 8; - } else { - multadd_real_vector_complex_scalar(fm, ch, dl_ch, 8); - } - } - - // check if PRB crosses DC and improve estimates around DC - if ((bwp_start_subcarrier < ue->frame_parms.ofdm_symbol_size) && (bwp_start_subcarrier+nb_rb_pdsch*12 >= ue->frame_parms.ofdm_symbol_size)) { - dl_ch = (int16_t *)&dl_ch_estimates[aarx][ch_offset]; - uint16_t idxDC = 2*(ue->frame_parms.ofdm_symbol_size - bwp_start_subcarrier); - uint16_t idxPil = idxDC/2; - re_offset = k; - pil = (int16_t *)&pilot[rb_offset*((config_type == NFAPI_NR_DMRS_TYPE1) ? 6:4)]; - pil += (idxPil-2); - dl_ch += (idxDC-4); - dl_ch = memset(dl_ch, 0, sizeof(int16_t)*10); - re_offset = (re_offset+idxDC/2-2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); - ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); - pil += 2; - re_offset = (re_offset + 2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset + nushift + re_offset)]; - ch[0] += (int16_t)(((int32_t)pil[0] * rxF[0] - (int32_t)pil[1] * rxF[1]) >> 15); - ch[1] += (int16_t)(((int32_t)pil[0] * rxF[1] + (int32_t)pil[1] * rxF[0]) >> 15); - ch[0] >>= 1; - ch[1] >>= 1; - - // for proper allignment of SIMD vectors - if((ue->frame_parms.N_RB_DL&1) == 0) { - - multadd_real_vector_complex_scalar(fdcl, - ch, - dl_ch-4, - 8); - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); - ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); - pil += 2; - re_offset = (re_offset + 2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset + nushift + re_offset)]; - ch[0] += (int16_t)(((int32_t)pil[0] * rxF[0] - (int32_t)pil[1] * rxF[1]) >> 15); - ch[1] += (int16_t)(((int32_t)pil[0] * rxF[1] + (int32_t)pil[1] * rxF[0]) >> 15); - ch[0] >>= 1; - ch[1] >>= 1; - - multadd_real_vector_complex_scalar(fdcr, - ch, - dl_ch-4, - 8); - } else { - - multadd_real_vector_complex_scalar(fdclh, - ch, - dl_ch, - 8); - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - ch[0] = (int16_t)(((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15); - ch[1] = (int16_t)(((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15); - pil += 2; - re_offset = (re_offset + 2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset + nushift + re_offset)]; - ch[0] += (int16_t)(((int32_t)pil[0] * rxF[0] - (int32_t)pil[1] * rxF[1]) >> 15); - ch[1] += (int16_t)(((int32_t)pil[0] * rxF[1] + (int32_t)pil[1] * rxF[0]) >> 15); - ch[0] >>= 1; - ch[1] >>= 1; - - multadd_real_vector_complex_scalar(fdcrh, - ch, - dl_ch, - 8); - } - } - } else if (config_type == NFAPI_NR_DMRS_TYPE2 && ue->chest_freq == 0){ //pdsch_dmrs_type2 |dmrs_r,dmrs_l,0,0,0,0,dmrs_r,dmrs_l,0,0,0,0| - - for (pilot_cnt = 0; pilot_cnt < 4 * nb_rb_pdsch; pilot_cnt++) { - if (pilot_cnt % 2 == 0) { - ch_l[0] = (int16_t)(((int32_t)pil[0] * rxF[0] - (int32_t)pil[1] * rxF[1]) >> 15); - ch_l[1] = (int16_t)(((int32_t)pil[0] * rxF[1] + (int32_t)pil[1] * rxF[0]) >> 15); - -#ifdef DEBUG_PDSCH - printf("pilot %3u: pil -> (%6d,%6d), rxF -> (%4d,%4d), ch -> (%4d,%4d) \n", pilot_cnt, pil[0], pil[1], rxF[0], rxF[1], ch_l[0], ch_l[1]); -#endif - - } else { - ch_r[0] = (int16_t)(((int32_t)pil[0] * rxF[0] - (int32_t)pil[1] * rxF[1]) >> 15); - ch_r[1] = (int16_t)(((int32_t)pil[0] * rxF[1] + (int32_t)pil[1] * rxF[0]) >> 15); - -#ifdef DEBUG_PDSCH - printf("pilot %3u: pil -> (%6d,%6d), rxF -> (%4d,%4d), ch -> (%4d,%4d) \n", pilot_cnt, pil[0], pil[1], rxF[0], rxF[1], ch_r[0], ch_r[1]); -#endif - ch[0] = (ch_l[0] + ch_r[0]) >> 1; - ch[1] = (ch_l[1] + ch_r[1]) >> 1; - - if (pilot_cnt == 3) { - multadd_real_vector_complex_scalar(fr, ch, dl_ch, 8); - dl_ch += 12; - } else if (pilot_cnt % 4 == 1) { - multadd_real_vector_complex_scalar(fmr, ch, dl_ch, 8); - dl_ch += 8; - } else if (pilot_cnt > 3) { - multadd_real_vector_complex_scalar(fmm, ch, dl_ch, 8); - dl_ch += 12; - } - - dl_ch[0 + (nushift << 1)] = ch[0]; - dl_ch[1 + (nushift << 1)] = ch[1]; - dl_ch[2 + (nushift << 1)] = ch[0]; - dl_ch[3 + (nushift << 1)] = ch[1]; - - if (pilot_cnt == 1) { - multadd_real_vector_complex_scalar(fl, ch, dl_ch, 8); - } else if (pilot_cnt % 4 == 1) { - multadd_real_vector_complex_scalar(fm, ch, dl_ch, 8); - - } else if (pilot_cnt == (4 * nb_rb_pdsch - 1)) { - dl_ch += 4; - // Treat last 2 pilots specially (right edge) - multadd_real_vector_complex_scalar(frl, dl_ch - 2 + (nushift << 1), dl_ch, 8); - multadd_real_vector_complex_scalar(frr, dl_ch - 14 + (nushift << 1), dl_ch, 8); - - } else { - dl_ch += 4; - multadd_real_vector_complex_scalar(fml, ch, dl_ch, 8); - } - } - - pil += 2; - int re_offset_add = pilot_cnt % 2 == 0 ? 1 : 5; - re_offset = (re_offset + re_offset_add) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset + nushift + re_offset)]; - } - - // check if PRB crosses DC and improve estimates around DC - if ((bwp_start_subcarrier < ue->frame_parms.ofdm_symbol_size) && (bwp_start_subcarrier+nb_rb_pdsch*12 >= ue->frame_parms.ofdm_symbol_size) && (p<2)) { - - dl_ch = (int16_t *)&dl_ch_estimates[p*ue->frame_parms.nb_antennas_rx+aarx][ch_offset]; - uint16_t idxDC = 2*(ue->frame_parms.ofdm_symbol_size - bwp_start_subcarrier); - dl_ch += (idxDC-8); - dl_ch = memset(dl_ch, 0, sizeof(int16_t)*20); - - dl_ch -= 2; - - ch_r[0] = dl_ch[0]; - ch_r[1]= dl_ch[1] ; - dl_ch += 22; - ch_l[0] = dl_ch[0]; - ch_l[1]= dl_ch[1] ; - - // for proper allignment of SIMD vectors - if((ue->frame_parms.N_RB_DL&1) == 0) { - dl_ch -= 20; - //Interpolate fdcrl1 with ch_r - multadd_real_vector_complex_scalar(filt8_dcrl1, - ch_r, - dl_ch, - 8); - //Interpolate fdclh1 with ch_l - multadd_real_vector_complex_scalar(filt8_dclh1, - ch_l, - dl_ch, - 8); - dl_ch += 16; - //Interpolate fdcrh1 with ch_r - multadd_real_vector_complex_scalar(filt8_dcrh1, - ch_r, - dl_ch, - 8); - //Interpolate fdcll1 with ch_l - multadd_real_vector_complex_scalar(filt8_dcll1, - ch_l, - dl_ch, - 8); - } else { - dl_ch -= 28; - //Interpolate fdcrl1 with ch_r - multadd_real_vector_complex_scalar(filt8_dcrl2, - ch_r, - dl_ch, - 8); - //Interpolate fdclh1 with ch_l - multadd_real_vector_complex_scalar(filt8_dclh2, - ch_l, - dl_ch, - 8); - dl_ch += 16; - //Interpolate fdcrh1 with ch_r - multadd_real_vector_complex_scalar(filt8_dcrh2, - ch_r, - dl_ch, - 8); - //Interpolate fdcll1 with ch_l - multadd_real_vector_complex_scalar(filt8_dcll2, - ch_l, - dl_ch, - 8); - } - } + } else { + NFAPI_NR_DMRS_TYPE2_average_prb(&ue->frame_parms, + rxF, + &pilot[4 * rb_offset], + dl_ch, + bwp_start_subcarrier, + nb_rb_pdsch); } - else if (config_type == NFAPI_NR_DMRS_TYPE1) { // this is case without frequency-domain linear interpolation, just take average of LS channel estimates of 6 DMRS REs and use a common value for the whole PRB - int32_t ch_0, ch_1; - - ch_0 = ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 = ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch[0] = ch_0 / 6; - ch[1] = ch_1 / 6; - -#if NO_INTERP - for (int i=0;i<12;i++) ((int32_t*)dl_ch)[i] = *(int32_t*)ch; - dl_ch+=24; -#else - multadd_real_vector_complex_scalar(filt8_avlip0, - ch, - dl_ch, - 8); - - dl_ch += 16; - multadd_real_vector_complex_scalar(filt8_avlip1, - ch, - dl_ch, - 8); - - dl_ch += 16; - multadd_real_vector_complex_scalar(filt8_avlip2, - ch, - dl_ch, - 8); - dl_ch -= 24; -#endif - - for (pilot_cnt=6; pilot_cnt<6*(nb_rb_pdsch-1); pilot_cnt += 6) { - - ch_0 = ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 = ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch[0] = ch_0 / 6; - ch[1] = ch_1 / 6; - -#if NO_INTERP - for (int i=0;i<12;i++) ((int32_t*)dl_ch)[i] = *(int32_t*)ch; - dl_ch+=24; -#else - dl_ch[6] += (ch[0] * 1365)>>15; // 1/12*16384 - dl_ch[7] += (ch[1] * 1365)>>15; // 1/12*16384 - - dl_ch += 8; - multadd_real_vector_complex_scalar(filt8_avlip3, - ch, - dl_ch, - 8); - - dl_ch += 16; - multadd_real_vector_complex_scalar(filt8_avlip4, - ch, - dl_ch, - 8); - - dl_ch += 16; - multadd_real_vector_complex_scalar(filt8_avlip5, - ch, - dl_ch, - 8); - dl_ch -= 16; -#endif - } - ch_0 = ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 = ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+2) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch[0] = ch_0 / 6; - ch[1] = ch_1 / 6; - -#if NO_INTERP - for (int i=0;i<12;i++) ((int32_t*)dl_ch)[i] = *(int32_t*)ch; - dl_ch+=24; -#else - dl_ch[6] += (ch[0] * 1365)>>15; // 1/12*16384 - dl_ch[7] += (ch[1] * 1365)>>15; // 1/12*16384 - - dl_ch += 8; - multadd_real_vector_complex_scalar(filt8_avlip3, - ch, - dl_ch, - 8); - - dl_ch += 16; - multadd_real_vector_complex_scalar(filt8_avlip6, - ch, - dl_ch, - 8); -#endif - } - else { // this is case without frequency-domain linear interpolation, just take average of LS channel estimates of 4 DMRS REs and use a common value for the whole PRB - int32_t ch_0, ch_1; - - ch_0 = ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 = ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+1) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+5) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+1) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+5) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch[0] = ch_0 / 4; - ch[1] = ch_1 / 4; - -#if NO_INTERP - for (int i=0;i<12;i++) ((int32_t*)dl_ch)[i] = *(int32_t*)ch; - dl_ch+=24; -#else - multadd_real_vector_complex_scalar(filt8_avlip0, - ch, - dl_ch, - 8); - - dl_ch += 16; - multadd_real_vector_complex_scalar(filt8_avlip1, - ch, - dl_ch, - 8); - - dl_ch += 16; - multadd_real_vector_complex_scalar(filt8_avlip2, - ch, - dl_ch, - 8); - dl_ch -= 24; -#endif - - for (pilot_cnt=4; pilot_cnt<4*(nb_rb_pdsch-1); pilot_cnt += 4) { - int32_t ch_0, ch_1; - - ch_0 = ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 = ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+1) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+5) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+1) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+5) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch[0] = ch_0 / 4; - ch[1] = ch_1 / 4; - -#if NO_INTERP - for (int i=0;i<12;i++) ((int32_t*)dl_ch)[i] = *(int32_t*)ch; - dl_ch+=24; -#else - dl_ch[6] += (ch[0] * 1365)>>15; // 1/12*16384 - dl_ch[7] += (ch[1] * 1365)>>15; // 1/12*16384 - - dl_ch += 8; - multadd_real_vector_complex_scalar(filt8_avlip3, - ch, - dl_ch, - 8); - - dl_ch += 16; - multadd_real_vector_complex_scalar(filt8_avlip4, - ch, - dl_ch, - 8); - - dl_ch += 16; - multadd_real_vector_complex_scalar(filt8_avlip5, - ch, - dl_ch, - 8); - dl_ch -= 16; -#endif - } - - ch_0 = ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 = ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+1) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+5) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+1) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch_0 += ((int32_t)pil[0]*rxF[0] - (int32_t)pil[1]*rxF[1])>>15; - ch_1 += ((int32_t)pil[0]*rxF[1] + (int32_t)pil[1]*rxF[0])>>15; - - pil += 2; - re_offset = (re_offset+5) % ue->frame_parms.ofdm_symbol_size; - rxF = (int16_t *)&rxdataF[aarx][(symbol_offset+nushift+re_offset)]; - - ch[0] = ch_0 / 4; - ch[1] = ch_1 / 4; - -#if NO_INTERP - for (int i=0;i<12;i++) ((int32_t*)dl_ch)[i] = *(int32_t*)ch; - dl_ch+=24; -#else - dl_ch[6] += (ch[0] * 1365)>>15; // 1/12*16384 - dl_ch[7] += (ch[1] * 1365)>>15; // 1/12*16384 - - dl_ch += 8; - multadd_real_vector_complex_scalar(filt8_avlip3, - ch, - dl_ch, - 8); - - dl_ch += 16; - multadd_real_vector_complex_scalar(filt8_avlip6, - ch, - dl_ch, - 8); -#endif - } #ifdef DEBUG_PDSCH - dl_ch = (int16_t *)&dl_ch_estimates[p*ue->frame_parms.nb_antennas_rx+aarx][ch_offset]; - for(uint16_t idxP=0; idxPframe_parms.nb_antennas_rx + aarx][ch_offset]; + for (uint16_t idxP = 0; idxP < ceil((float)nb_rb_pdsch * 12 / 8); idxP++) { + for (uint8_t idxI = 0; idxI < 8; idxI++) { + printf("%4d\t%4d\t", dl_ch[idxP * 8 + idxI].r, dl_ch[idxP * 8 + idxI].i); } - printf("%2d\n",idxP); + printf("%2d\n", idxP); } #endif } - return(0); + return 0; } /******************************************************************* @@ -2045,10 +1732,11 @@ int nr_pdsch_channel_estimation(PHY_VARS_NR_UE *ue, * 3) Compensate signal with PTRS estimation for slot *********************************************************************/ void nr_pdsch_ptrs_processing(PHY_VARS_NR_UE *ue, + int nbRx, c16_t ptrs_phase_per_slot[][14], int32_t ptrs_re_per_slot[][14], - uint32_t rx_size, - int32_t rxdataF_comp[][rx_size], + uint32_t rx_size_symbol, + int32_t rxdataF_comp[][nbRx][rx_size_symbol * NR_SYMBOLS_PER_SLOT], NR_DL_FRAME_PARMS *frame_parms, NR_DL_UE_HARQ_t *dlsch0_harq, NR_DL_UE_HARQ_t *dlsch1_harq, @@ -2139,8 +1827,9 @@ void nr_pdsch_ptrs_processing(PHY_VARS_NR_UE *ue, nr_ptrs_cpe_estimation(*K_ptrs,*ptrsReOffset,*dmrsConfigType,*nb_rb, rnti, nr_slot_rx, - symbol,frame_parms->ofdm_symbol_size, - (int16_t*)&rxdataF_comp[aarx][(symbol * nb_re_pdsch)], + symbol, + frame_parms->ofdm_symbol_size, + (int16_t *)(rxdataF_comp[0][aarx] + symbol * nb_re_pdsch), ue->nr_gold_pdsch[gNB_id][nr_slot_rx][symbol][0], (int16_t*)&phase_per_symbol[symbol], &ptrs_re_symbol[symbol]); @@ -2161,9 +1850,7 @@ void nr_pdsch_ptrs_processing(PHY_VARS_NR_UE *ue, } #ifdef DEBUG_DL_PTRS LOG_M("ptrsEst.m","est",ptrs_phase_per_slot[aarx],frame_parms->symbols_per_slot,1,1 ); - LOG_M("rxdataF_bf_ptrs_comp.m","bf_ptrs_cmp", - &rxdataF_comp[aarx][(*startSymbIndex) * NR_NB_SC_PER_RB * (*nb_rb) ], - (*nb_rb) * NR_NB_SC_PER_RB * (*nbSymb),1,1); + LOG_M("rxdataF_bf_ptrs_comp.m", "bf_ptrs_cmp", rxdataF_comp[0][aarx] + (*startSymbIndex) * NR_NB_SC_PER_RB * (*nb_rb), (*nb_rb) * NR_NB_SC_PER_RB * (*nbSymb), 1, 1); #endif /*------------------------------------------------------------------------------------------------------- */ /* 3) Compensated DMRS based estimated signal with PTRS estimation */ @@ -2175,10 +1862,11 @@ void nr_pdsch_ptrs_processing(PHY_VARS_NR_UE *ue, #ifdef DEBUG_DL_PTRS printf("[PHY][DL][PTRS]: Rotate Symbol %2d with %d + j* %d\n", i, phase_per_symbol[i].r,phase_per_symbol[i].i); #endif - rotate_cpx_vector((c16_t*)&rxdataF_comp[aarx][(i * (*nb_rb) * NR_NB_SC_PER_RB)], + rotate_cpx_vector((c16_t *)&rxdataF_comp[0][aarx][(i * (*nb_rb) * NR_NB_SC_PER_RB)], &phase_per_symbol[i], - (c16_t*)&rxdataF_comp[aarx][(i * (*nb_rb) * NR_NB_SC_PER_RB)], - ((*nb_rb) * NR_NB_SC_PER_RB), 15); + (c16_t *)&rxdataF_comp[0][aarx][(i * (*nb_rb) * NR_NB_SC_PER_RB)], + ((*nb_rb) * NR_NB_SC_PER_RB), + 15); }// if not DMRS Symbol }// symbol loop }// last symbol check diff --git a/openair1/PHY/NR_UE_ESTIMATION/nr_estimation.h b/openair1/PHY/NR_UE_ESTIMATION/nr_estimation.h index b779882c86..11dad31602 100644 --- a/openair1/PHY/NR_UE_ESTIMATION/nr_estimation.h +++ b/openair1/PHY/NR_UE_ESTIMATION/nr_estimation.h @@ -91,7 +91,8 @@ int nr_pdsch_channel_estimation(PHY_VARS_NR_UE *ue, unsigned short nb_rb_pdsch, uint32_t pdsch_est_size, int32_t dl_ch_estimates[][pdsch_est_size], - c16_t rxdataF[][ue->frame_parms.samples_per_slot_wCP]); + int rxdataFsize, + c16_t rxdataF[][rxdataFsize]); void nr_adjust_synch_ue(NR_DL_FRAME_PARMS *frame_parms, PHY_VARS_NR_UE *ue, @@ -123,10 +124,11 @@ void phy_adjust_gain_nr(PHY_VARS_NR_UE *ue, uint8_t gNB_id); void nr_pdsch_ptrs_processing(PHY_VARS_NR_UE *ue, + int nbRx, c16_t ptrs_phase_per_slot[][14], int32_t ptrs_re_per_slot[][14], - uint32_t rx_size, - int32_t rxdataF_comp[][rx_size], + uint32_t rx_size_symbol, + int32_t rxdataF_comp[][nbRx][rx_size_symbol * NR_SYMBOLS_PER_SLOT], NR_DL_FRAME_PARMS *frame_parms, NR_DL_UE_HARQ_t *dlsch0_harq, NR_DL_UE_HARQ_t *dlsch1_harq, diff --git a/openair1/PHY/NR_UE_TRANSPORT/csi_rx.c b/openair1/PHY/NR_UE_TRANSPORT/csi_rx.c index 08aa58d5f9..4eaa36f1bc 100644 --- a/openair1/PHY/NR_UE_TRANSPORT/csi_rx.c +++ b/openair1/PHY/NR_UE_TRANSPORT/csi_rx.c @@ -518,9 +518,8 @@ int nr_csi_rs_ri_estimation(const PHY_VARS_NR_UE *ue, // construct Hh x H elements if(ant_rx_conjch == ant_rx_ch) { - nr_a_sum_b((__m128i *)&csi_rs_estimated_A_MF[port_tx_conjch][port_tx_ch][k_offset], - (__m128i *)&csi_rs_estimated_conjch_ch[ant_rx_conjch][port_tx_conjch][ant_rx_ch][port_tx_ch][k_offset], - 1); + nr_a_sum_b( + (c16_t *)&csi_rs_estimated_A_MF[port_tx_conjch][port_tx_ch][k_offset], (c16_t *)&csi_rs_estimated_conjch_ch[ant_rx_conjch][port_tx_conjch][ant_rx_ch][port_tx_ch][k_offset], 1); } } } diff --git a/openair1/PHY/NR_UE_TRANSPORT/dci_nr.c b/openair1/PHY/NR_UE_TRANSPORT/dci_nr.c index 2809a21cd4..d4ffb66b65 100644 --- a/openair1/PHY/NR_UE_TRANSPORT/dci_nr.c +++ b/openair1/PHY/NR_UE_TRANSPORT/dci_nr.c @@ -935,6 +935,7 @@ uint8_t nr_dci_decoding_procedure(PHY_VARS_NR_UE *ue, dci_ind->dci_list[dci_ind->number_of_dcis].N_CCE = L; dci_ind->dci_list[dci_ind->number_of_dcis].dci_format = rel15->dci_format_options[k]; dci_ind->dci_list[dci_ind->number_of_dcis].ss_type = rel15->dci_type_options[k]; + dci_ind->dci_list[dci_ind->number_of_dcis].coreset_type = rel15->coreset.CoreSetType; int n_rb, rb_offset; get_coreset_rballoc(rel15->coreset.frequency_domain_resource, &n_rb, &rb_offset); dci_ind->dci_list[dci_ind->number_of_dcis].cset_start = rel15->BWPStart + rb_offset; diff --git a/openair1/PHY/NR_UE_TRANSPORT/nr_dlsch_demodulation.c b/openair1/PHY/NR_UE_TRANSPORT/nr_dlsch_demodulation.c index af1980f820..d568172153 100644 --- a/openair1/PHY/NR_UE_TRANSPORT/nr_dlsch_demodulation.c +++ b/openair1/PHY/NR_UE_TRANSPORT/nr_dlsch_demodulation.c @@ -40,6 +40,7 @@ #include "PHY/NR_REFSIG/dmrs_nr.h" #include "common/utils/nr/nr_common.h" #include +#include "openair1/PHY/TOOLS/phy_scope_interface.h" /* dynamic shift for LLR computation for TM3/4 * set as command line argument, see lte-softmodem.c @@ -81,15 +82,15 @@ unsigned char offset_mumimo_llr_drange[29][3]={{8,8,8},{7,7,7},{7,7,7},{7,7,7},{ #define print_shorts(s,x) printf("%s = [%d+j*%d, %d+j*%d, %d+j*%d, %d+j*%d]\n",s,(x)[0],(x)[1],(x)[2],(x)[3],(x)[4],(x)[5],(x)[6],(x)[7]) /* compute H_h_H matrix inversion up to 4x4 matrices */ -uint8_t nr_zero_forcing_rx(uint32_t rx_size, - int32_t rxdataF_comp[][rx_size], - int32_t dl_ch_mag[][rx_size], - int32_t dl_ch_magb[][rx_size], - int32_t dl_ch_magr[][rx_size], - int32_t dl_ch_estimates_ext[][rx_size], - unsigned short nb_rb, +uint8_t nr_zero_forcing_rx(uint32_t rx_size_symbol, unsigned char n_rx, unsigned char n_tx,//number of layer + int32_t rxdataF_comp[][n_rx][rx_size_symbol * NR_SYMBOLS_PER_SLOT], + int32_t dl_ch_mag[][n_rx][rx_size_symbol], + int32_t dl_ch_magb[][n_rx][rx_size_symbol], + int32_t dl_ch_magr[][n_rx][rx_size_symbol], + int32_t dl_ch_estimates_ext[][rx_size_symbol], + unsigned short nb_rb, unsigned char mod_order, int shift, unsigned char symbol, @@ -105,13 +106,14 @@ static void nr_dlsch_layer_demapping(int16_t *llr_cw[2], int16_t *llr_layers[NR_MAX_NB_LAYERS]); /* compute LLR */ -static int nr_dlsch_llr(uint32_t rx_size, +static int nr_dlsch_llr(uint32_t rx_size_symbol, + int nbRx, int16_t *layer_llr[NR_MAX_NB_LAYERS], NR_DL_FRAME_PARMS *frame_parms, - int32_t rxdataF_comp[][rx_size], - int32_t dl_ch_mag[][rx_size], - int32_t dl_ch_magb[][rx_size], - int32_t dl_ch_magr[][rx_size], + int32_t rxdataF_comp[][nbRx][rx_size_symbol * NR_SYMBOLS_PER_SLOT], + int32_t dl_ch_mag[rx_size_symbol], + int32_t dl_ch_magb[rx_size_symbol], + int32_t dl_ch_magr[rx_size_symbol], NR_DL_UE_HARQ_t *dlsch0_harq, NR_DL_UE_HARQ_t *dlsch1_harq, unsigned char harq_pid, @@ -124,6 +126,122 @@ static int nr_dlsch_llr(uint32_t rx_size, uint8_t nr_slot_rx, NR_UE_DLSCH_t dlsch[2], uint32_t llr_offset[NR_SYMBOLS_PER_SLOT]); +/** \fn dlsch_extract_rbs(int32_t **rxdataF, + int32_t **dl_ch_estimates, + int32_t **rxdataF_ext, + int32_t **dl_ch_estimates_ext, + unsigned char symbol + uint8_t pilots, + uint8_t config_type, + unsigned short start_rb, + unsigned short nb_rb_pdsch, + uint8_t n_dmrs_cdm_groups, + uint8_t Nl, + NR_DL_FRAME_PARMS *frame_parms, + uint16_t dlDmrsSymbPos) + \brief This function extracts the received resource blocks, both channel estimates and data symbols, + for the current allocation and for multiple layer antenna gNB transmission. + @param rxdataF Raw FFT output of received signal + @param dl_ch_estimates Channel estimates of current slot + @param rxdataF_ext FFT output for RBs in this allocation + @param dl_ch_estimates_ext Channel estimates for RBs in this allocation + @param Nl nb of antenna layers + @param symbol Symbol to extract + @param n_dmrs_cdm_groups + @param frame_parms Pointer to frame descriptor +*/ +static void nr_dlsch_extract_rbs(uint32_t rxdataF_sz, + c16_t rxdataF[][rxdataF_sz], + uint32_t rx_size_symbol, + uint32_t pdsch_est_size, + int32_t dl_ch_estimates[][pdsch_est_size], + c16_t rxdataF_ext[][rx_size_symbol], + int32_t dl_ch_estimates_ext[][rx_size_symbol], + unsigned char symbol, + uint8_t pilots, + uint8_t config_type, + unsigned short start_rb, + unsigned short nb_rb_pdsch, + uint8_t n_dmrs_cdm_groups, + uint8_t Nl, + NR_DL_FRAME_PARMS *frame_parms, + uint16_t dlDmrsSymbPos, + int chest_time_type); + +static void nr_dlsch_channel_level_median(uint32_t rx_size_symbol, int32_t dl_ch_estimates_ext[][rx_size_symbol], int32_t *median, int n_tx, int n_rx, int length); + +/** \brief This function performs channel compensation (matched filtering) on the received RBs for this allocation. In addition, it computes the squared-magnitude of the channel with weightings for + 16QAM/64QAM detection as well as dual-stream detection (cross-correlation) + @param rxdataF_ext Frequency-domain received signal in RBs to be demodulated + @param dl_ch_estimates_ext Frequency-domain channel estimates in RBs to be demodulated + @param dl_ch_mag First Channel magnitudes (16QAM/64QAM) + @param dl_ch_magb Second weighted Channel magnitudes (64QAM) + @param rxdataF_comp Compensated received waveform + @param rho Cross-correlation between two spatial channels on each RX antenna + @param frame_parms Pointer to frame descriptor + @param symbol Symbol on which to operate + @param first_symbol_flag set to 1 on first DLSCH symbol + @param mod_order Modulation order of allocation + @param nb_rb Number of RBs in allocation + @param output_shift Rescaling for compensated output (should be energy-normalizing) + @param phy_measurements Pointer to UE PHY measurements +*/ + +void nr_dlsch_channel_compensation(uint32_t rx_size_symbol, + int nbRx, + c16_t rxdataF_ext[][rx_size_symbol], + int32_t dl_ch_estimates_ext[][rx_size_symbol], + int32_t dl_ch_mag[][nbRx][rx_size_symbol], + int32_t dl_ch_magb[][nbRx][rx_size_symbol], + int32_t dl_ch_magr[][nbRx][rx_size_symbol], + int32_t rxdataF_comp[][nbRx][rx_size_symbol * NR_SYMBOLS_PER_SLOT], + int ***rho, + NR_DL_FRAME_PARMS *frame_parms, + uint8_t nb_aatx, + unsigned char symbol, + int length, + uint8_t first_symbol_flag, + unsigned char mod_order, + unsigned short nb_rb, + unsigned char output_shift, + PHY_NR_MEASUREMENTS *measurements); + +/** \brief This function computes the average channel level over all allocated RBs and antennas (TX/RX) in order to compute output shift for compensated signal + @param dl_ch_estimates_ext Channel estimates in allocated RBs + @param frame_parms Pointer to frame descriptor + @param avg Pointer to average signal strength + @param pilots_flag Flag to indicate pilots in symbol + @param nb_rb Number of allocated RBs +*/ +static void nr_dlsch_channel_level(uint32_t rx_size_symbol, + int32_t dl_ch_estimates_ext[][rx_size_symbol], + NR_DL_FRAME_PARMS *frame_parms, + uint8_t n_tx, + int32_t *avg, + uint8_t symbol, + uint32_t len, + unsigned short nb_rb); + +void nr_dlsch_scale_channel(uint32_t rx_size_symbol, + int32_t dl_ch_estimates_ext[][rx_size_symbol], + NR_DL_FRAME_PARMS *frame_parms, + uint8_t n_tx, + uint8_t n_rx, + uint8_t symbol, + uint8_t pilots, + uint32_t len, + unsigned short nb_rb); +void nr_dlsch_detection_mrc(uint32_t rx_size_symbol, + short n_tx, + short n_rx, + int32_t rxdataF_comp[][n_rx][rx_size_symbol * NR_SYMBOLS_PER_SLOT], + int ***rho, + int32_t dl_ch_mag[][n_rx][rx_size_symbol], + int32_t dl_ch_magb[][n_rx][rx_size_symbol], + int32_t dl_ch_magr[][n_rx][rx_size_symbol], + unsigned char symbol, + unsigned short nb_rb, + int length); /* Main Function */ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, @@ -135,21 +253,28 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, uint32_t pdsch_est_size, int32_t dl_ch_estimates[][pdsch_est_size], int16_t *llr[2], - c16_t ptrs_phase_per_slot[][NR_SYMBOLS_PER_SLOT], - int32_t ptrs_re_per_slot[][NR_SYMBOLS_PER_SLOT], uint32_t dl_valid_re[NR_SYMBOLS_PER_SLOT], - uint32_t rx_size, - int32_t dl_ch_estimates_ext[][rx_size], - int32_t rxdataF_ext[][rx_size], - int32_t rxdataF_comp[][rx_size], - int32_t dl_ch_mag[][rx_size], - int32_t dl_ch_magb[][rx_size], - int32_t dl_ch_magr[][rx_size], c16_t rxdataF[][ue->frame_parms.samples_per_slot_wCP], uint32_t llr_offset[NR_SYMBOLS_PER_SLOT], - int32_t *log2_maxh) + int32_t *log2_maxh, + int rx_size_symbol, + int nbRx, + int32_t rxdataF_comp[][nbRx][rx_size_symbol], + c16_t ptrs_phase_per_slot[][NR_SYMBOLS_PER_SLOT], + int32_t ptrs_re_per_slot[][NR_SYMBOLS_PER_SLOT]) { + const int matrixSz = ue->frame_parms.nb_antennas_rx * ue->frame_parms.nb_antennas_tx; + __attribute__((aligned(32))) int32_t dl_ch_estimates_ext[matrixSz][rx_size_symbol]; + memset(dl_ch_estimates_ext, 0, sizeof(dl_ch_estimates_ext)); + __attribute__((aligned(32))) int32_t dl_ch_mag[ue->frame_parms.nb_antennas_tx][ue->frame_parms.nb_antennas_rx][rx_size_symbol]; + memset(dl_ch_mag, 0, sizeof(dl_ch_mag)); + + __attribute__((aligned(32))) int32_t dl_ch_magb[ue->frame_parms.nb_antennas_tx][nbRx][rx_size_symbol]; + memset(dl_ch_magb, 0, sizeof(dl_ch_magb)); + + __attribute__((aligned(32))) int32_t dl_ch_magr[ue->frame_parms.nb_antennas_tx][nbRx][rx_size_symbol]; + memset(dl_ch_magr, 0, sizeof(dl_ch_magr)); NR_UE_COMMON *common_vars = &ue->common_vars; NR_DL_FRAME_PARMS *frame_parms = &ue->frame_parms; PHY_NR_MEASUREMENTS *measurements = &ue->measurements; @@ -279,64 +404,59 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, //---------------------------------------------------------- //--------------------- RBs extraction --------------------- //---------------------------------------------------------- - start_meas(&ue->generic_stat_bis[slot]); - nr_dlsch_extract_rbs(ue->frame_parms.samples_per_slot_wCP, - rxdataF, - rx_size, - pdsch_est_size, - dl_ch_estimates, - rxdataF_ext, - dl_ch_estimates_ext, - symbol, - pilots, - config_type, - start_rb + dlsch[0].dlsch_config.BWPStart, - nb_rb_pdsch, - dlsch[0].dlsch_config.n_dmrs_cdm_groups, - dlsch[0].Nl, - frame_parms, - dlsch[0].dlsch_config.dlDmrsSymbPos, - ue->chest_time); - stop_meas(&ue->generic_stat_bis[slot]); - if (cpumeas(CPUMEAS_GETSTATE)) - LOG_D(PHY, "[AbsSFN %u.%d] Slot%d Symbol %d: Pilot/Data extraction %5.2f \n", - frame,nr_slot_rx,slot,symbol,ue->generic_stat_bis[slot].p_time/(cpuf*1000.0)); + const int n_rx = frame_parms->nb_antennas_rx; + const int nl = dlsch[0].Nl; - int nl = dlsch[0].Nl; - int n_rx = frame_parms->nb_antennas_rx; - nb_re_pdsch = (pilots==1)? ((config_type==NFAPI_NR_DMRS_TYPE1)?nb_rb_pdsch*(12-6*dlsch[0].dlsch_config.n_dmrs_cdm_groups): nb_rb_pdsch*(12-4*dlsch[0].dlsch_config.n_dmrs_cdm_groups)) : (nb_rb_pdsch*12); + start_meas(&ue->generic_stat_bis[slot]); + { + __attribute__((aligned(32))) c16_t rxdataF_ext[nbRx][rx_size_symbol]; + memset(rxdataF_ext, 0, sizeof(rxdataF_ext)); + + nr_dlsch_extract_rbs(ue->frame_parms.samples_per_slot_wCP, + rxdataF, + rx_size_symbol, + pdsch_est_size, + dl_ch_estimates, + rxdataF_ext, + dl_ch_estimates_ext, + symbol, + pilots, + config_type, + start_rb + dlsch[0].dlsch_config.BWPStart, + nb_rb_pdsch, + dlsch[0].dlsch_config.n_dmrs_cdm_groups, + dlsch[0].Nl, + frame_parms, + dlsch[0].dlsch_config.dlDmrsSymbPos, + ue->chest_time); + stop_meas(&ue->generic_stat_bis[slot]); + if (ue->phy_sim_pdsch_rxdataF_ext) + for (unsigned char aarx = 0; aarx < frame_parms->nb_antennas_rx; aarx++) { + int offset = ((void *)rxdataF_ext[aarx] - (void *)rxdataF_ext) + symbol * rx_size_symbol; + memcpy(ue->phy_sim_pdsch_rxdataF_ext + offset, rxdataF_ext, rx_size_symbol * sizeof(c16_t)); + } + if (cpumeas(CPUMEAS_GETSTATE)) + LOG_D(PHY, "[AbsSFN %u.%d] Slot%d Symbol %d: Pilot/Data extraction %5.2f \n", frame, nr_slot_rx, slot, symbol, ue->generic_stat_bis[slot].p_time / (cpuf * 1000.0)); + + nb_re_pdsch = (pilots == 1) + ? ((config_type == NFAPI_NR_DMRS_TYPE1) ? nb_rb_pdsch * (12 - 6 * dlsch[0].dlsch_config.n_dmrs_cdm_groups) : nb_rb_pdsch * (12 - 4 * dlsch[0].dlsch_config.n_dmrs_cdm_groups)) + : (nb_rb_pdsch * 12); //---------------------------------------------------------- //--------------------- Channel Scaling -------------------- //---------------------------------------------------------- start_meas(&ue->generic_stat_bis[slot]); - nr_dlsch_scale_channel(rx_size, - dl_ch_estimates_ext, - frame_parms, - nl, - n_rx, - symbol, - pilots, - nb_re_pdsch, - nb_rb_pdsch); + nr_dlsch_scale_channel(rx_size_symbol, dl_ch_estimates_ext, frame_parms, nl, n_rx, symbol, pilots, nb_re_pdsch, nb_rb_pdsch); stop_meas(&ue->generic_stat_bis[slot]); if (cpumeas(CPUMEAS_GETSTATE)) - LOG_D(PHY, "[AbsSFN %u.%d] Slot%d Symbol %d: Channel Scale %5.2f \n", - frame,nr_slot_rx,slot,symbol,ue->generic_stat_bis[slot].p_time/(cpuf*1000.0)); + LOG_D(PHY, "[AbsSFN %u.%d] Slot%d Symbol %d: Channel Scale %5.2f \n", frame, nr_slot_rx, slot, symbol, ue->generic_stat_bis[slot].p_time / (cpuf * 1000.0)); //---------------------------------------------------------- //--------------------- Channel Level Calc. ---------------- //---------------------------------------------------------- start_meas(&ue->generic_stat_bis[slot]); if (first_symbol_flag==1) { - nr_dlsch_channel_level(rx_size, - dl_ch_estimates_ext, - frame_parms, - nl, - avg, - symbol, - nb_re_pdsch, - nb_rb_pdsch); + nr_dlsch_channel_level(rx_size_symbol, dl_ch_estimates_ext, frame_parms, nl, avg, symbol, nb_re_pdsch, nb_rb_pdsch); avgs = 0; for (aatx=0;aatx 1) { - nr_dlsch_channel_level_median(rx_size, - dl_ch_estimates_ext, - median, - nl, - n_rx, - nb_re_pdsch, - symbol*nb_rb_pdsch*12); + nr_dlsch_channel_level_median(rx_size_symbol, dl_ch_estimates_ext, median, nl, n_rx, nb_re_pdsch); for (aatx = 0; aatx < nl; aatx++) { for (aarx = 0; aarx < n_rx; aarx++) { avgs = cmax(avgs, median[aatx*n_rx + aarx]); @@ -361,12 +475,7 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, } *log2_maxh = (log2_approx(avgs)/2) + 1; //LOG_I(PHY, "avgs Power per SC is %d lg2_maxh %d\n", avgs, log2_maxh); - LOG_D(PHY,"[DLSCH] AbsSubframe %d.%d log2_maxh = %d (%d,%d)\n", - frame%1024, - nr_slot_rx, - *log2_maxh, - avg[0], - avgs); + LOG_D(PHY, "[DLSCH] AbsSubframe %d.%d log2_maxh = %d (%d,%d)\n", frame % 1024, nr_slot_rx, *log2_maxh, avg[0], avgs); } stop_meas(&ue->generic_stat_bis[slot]); @@ -385,7 +494,8 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, //---------------------------------------------------------- // Disable correlation measurement for optimizing UE start_meas(&ue->generic_stat_bis[slot]); - nr_dlsch_channel_compensation(rx_size, + nr_dlsch_channel_compensation(rx_size_symbol, + nbRx, rxdataF_ext, dl_ch_estimates_ext, dl_ch_mag, @@ -402,38 +512,26 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, nb_rb_pdsch, *log2_maxh, measurements); // log2_maxh+I0_shift - stop_meas(&ue->generic_stat_bis[slot]); + stop_meas(&ue->generic_stat_bis[slot]); + } if (cpumeas(CPUMEAS_GETSTATE)) - LOG_D(PHY, "[AbsSFN %u.%d] Slot%d Symbol %d log2_maxh %d channel_level %d: Channel Comp %5.2f \n", frame, nr_slot_rx, slot, symbol, *log2_maxh, proc->channel_level, ue->generic_stat_bis[slot].p_time/(cpuf*1000.0)); + LOG_D(PHY, + "[AbsSFN %u.%d] Slot%d Symbol %d log2_maxh %d channel_level %d: Channel Comp %5.2f \n", + frame, + nr_slot_rx, + slot, + symbol, + *log2_maxh, + proc->channel_level, + ue->generic_stat_bis[slot].p_time / (cpuf * 1000.0)); start_meas(&ue->generic_stat_bis[slot]); if (n_rx > 1) { - nr_dlsch_detection_mrc(rx_size, - rxdataF_comp, - NULL, - dl_ch_mag, - dl_ch_magb, - dl_ch_magr, - nl, - n_rx, - symbol, - nb_rb_pdsch, - nb_re_pdsch); + nr_dlsch_detection_mrc(rx_size_symbol, nl, n_rx, rxdataF_comp, NULL, dl_ch_mag, dl_ch_magb, dl_ch_magr, symbol, nb_rb_pdsch, nb_re_pdsch); if (nl >= 2)//Apply zero forcing for 2, 3, and 4 Tx layers - nr_zero_forcing_rx(rx_size, - rxdataF_comp, - dl_ch_mag, - dl_ch_magb, - dl_ch_magr, - dl_ch_estimates_ext, - nb_rb_pdsch, - n_rx, - nl, - dlsch[0].dlsch_config.qamModOrder, - *log2_maxh, - symbol, - nb_re_pdsch); + nr_zero_forcing_rx( + rx_size_symbol, n_rx, nl, rxdataF_comp, dl_ch_mag, dl_ch_magb, dl_ch_magr, dl_ch_estimates_ext, nb_rb_pdsch, dlsch[0].dlsch_config.qamModOrder, *log2_maxh, symbol, nb_re_pdsch); } stop_meas(&ue->generic_stat_bis[slot]); @@ -452,19 +550,8 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, /* Check for PTRS bitmap and process it respectively */ if((pduBitmap & 0x1) && (dlsch[0].rnti_type == _C_RNTI_)) { - nr_pdsch_ptrs_processing(ue, - ptrs_phase_per_slot, - ptrs_re_per_slot, - rx_size, - rxdataF_comp, - frame_parms, - dlsch0_harq, - dlsch1_harq, - gNB_id, - nr_slot_rx, - symbol, - (nb_rb_pdsch*12), - dlsch[0].rnti,dlsch); + nr_pdsch_ptrs_processing( + ue, nbRx, ptrs_phase_per_slot, ptrs_re_per_slot, rx_size_symbol, rxdataF_comp, frame_parms, dlsch0_harq, dlsch1_harq, gNB_id, nr_slot_rx, symbol, (nb_rb_pdsch * 12), dlsch[0].rnti, dlsch); dl_valid_re[symbol-1] -= ptrs_re_per_slot[0][symbol]; } @@ -499,13 +586,14 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, first_symbol_flag=0; } /* Calculate LLR's for each symbol */ - nr_dlsch_llr(rx_size, + nr_dlsch_llr(rx_size_symbol, + nbRx, layer_llr, frame_parms, rxdataF_comp, - dl_ch_mag, - dl_ch_magb, - dl_ch_magr, + dl_ch_mag[0][0], + dl_ch_magb[0][0], + dl_ch_magr[0][0], dlsch0_harq, dlsch1_harq, harq_pid, @@ -514,7 +602,7 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, nb_rb_pdsch, codeword_TB0, codeword_TB1, - dl_valid_re[i-1], + dl_valid_re[i - 1], nr_slot_rx, dlsch, llr_offset); @@ -570,13 +658,28 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, LOG_D(PHY, "[AbsSFN %u.%d] Slot%d Symbol %d: LLR Computation %5.2f \n", frame, nr_slot_rx, slot, symbol, ue->generic_stat_bis[slot].p_time / (cpuf * 1000.0)); #if T_TRACER - T(T_UE_PHY_PDSCH_IQ, T_INT(gNB_id), T_INT(ue->Mod_id), T_INT(frame%1024), - T_INT(nr_slot_rx), T_INT(nb_rb_pdsch), - T_INT(frame_parms->N_RB_UL), T_INT(frame_parms->symbols_per_slot), - T_BUFFER(&rxdataF_comp[gNB_id][0], 2 * /* ulsch[UE_id]->harq_processes[harq_pid]->nb_rb */ frame_parms->N_RB_UL *12*frame_parms->symbols_per_slot*2)); + T(T_UE_PHY_PDSCH_IQ, + T_INT(gNB_id), + T_INT(ue->Mod_id), + T_INT(frame % 1024), + T_INT(nr_slot_rx), + T_INT(nb_rb_pdsch), + T_INT(frame_parms->N_RB_UL), + T_INT(frame_parms->symbols_per_slot), + T_BUFFER(&rxdataF_comp[gNB_id][0], 2 * /* ulsch[UE_id]->harq_processes[harq_pid]->nb_rb */ frame_parms->N_RB_UL * 12 * 2)); #endif - return(0); + UEscopeCopy(ue, pdschRxdataF_comp, rxdataF_comp, sizeof(c16_t), nbRx, rx_size_symbol); + + if (ue->phy_sim_pdsch_rxdataF_comp) + for (int a = 0; a < nbRx; a++) { + int offset = (void *)rxdataF_comp[0][a] - (void *)rxdataF_comp[0] + symbol * rx_size_symbol * sizeof(c16_t); + memcpy(ue->phy_sim_pdsch_rxdataF_comp + offset, rxdataF_comp[0][a] + symbol * rx_size_symbol, sizeof(c16_t) * rx_size_symbol); + memcpy(ue->phy_sim_pdsch_dl_ch_estimates + offset, dl_ch_estimates, pdsch_est_size * sizeof(c16_t)); + } + if (ue->phy_sim_pdsch_dl_ch_estimates_ext) + memcpy(ue->phy_sim_pdsch_dl_ch_estimates_ext + symbol * rx_size_symbol, dl_ch_estimates_ext, sizeof(dl_ch_estimates_ext)); + return (0); } void nr_dlsch_deinterleaving(uint8_t symbol, @@ -630,13 +733,14 @@ void nr_dlsch_deinterleaving(uint8_t symbol, // Pre-processing for LLR computation //============================================================================================== -void nr_dlsch_channel_compensation(uint32_t rx_size, - int32_t rxdataF_ext[][rx_size], - int32_t dl_ch_estimates_ext[][rx_size], - int32_t dl_ch_mag[][rx_size], - int32_t dl_ch_magb[][rx_size], - int32_t dl_ch_magr[][rx_size], - int32_t rxdataF_comp[][rx_size], +void nr_dlsch_channel_compensation(uint32_t rx_size_symbol, + int nbRx, + c16_t rxdataF_ext[][rx_size_symbol], + int32_t dl_ch_estimates_ext[][rx_size_symbol], + int32_t dl_ch_mag[][nbRx][rx_size_symbol], + int32_t dl_ch_magb[][nbRx][rx_size_symbol], + int32_t dl_ch_magr[][nbRx][rx_size_symbol], + int32_t rxdataF_comp[][nbRx][rx_size_symbol * NR_SYMBOLS_PER_SLOT], int ***rho, NR_DL_FRAME_PARMS *frame_parms, uint8_t nb_aatx, @@ -676,12 +780,12 @@ void nr_dlsch_channel_compensation(uint32_t rx_size, for (aarx=0; aarxnb_antennas_rx; aarx++) { - dl_ch128 = (__m128i *)&dl_ch_estimates_ext[(aatx*frame_parms->nb_antennas_rx)+aarx][symbol*nb_rb*12]; - dl_ch_mag128 = (__m128i *)&dl_ch_mag[(aatx*frame_parms->nb_antennas_rx)+aarx][symbol*nb_rb*12]; - dl_ch_mag128b = (__m128i *)&dl_ch_magb[(aatx*frame_parms->nb_antennas_rx)+aarx][symbol*nb_rb*12]; - dl_ch_mag128r = (__m128i *)&dl_ch_magr[(aatx*frame_parms->nb_antennas_rx)+aarx][symbol*nb_rb*12]; - rxdataF128 = (__m128i *)&rxdataF_ext[aarx][symbol*nb_rb*12]; - rxdataF_comp128 = (__m128i *)&rxdataF_comp[(aatx*frame_parms->nb_antennas_rx)+aarx][symbol*nb_rb*12]; + dl_ch128 = (__m128i *)dl_ch_estimates_ext[(aatx*frame_parms->nb_antennas_rx)+aarx]; + dl_ch_mag128 = (__m128i *)dl_ch_mag[aatx][aarx]; + dl_ch_mag128b = (__m128i *)dl_ch_magb[aatx][aarx]; + dl_ch_mag128r = (__m128i *)dl_ch_magr[aatx][aarx]; + rxdataF128 = (__m128i *)rxdataF_ext[aarx]; + rxdataF_comp128 = (__m128i *)(rxdataF_comp[aatx][aarx] + symbol * nb_rb * 12); for (rb=0; rb2) { @@ -839,8 +943,8 @@ void nr_dlsch_channel_compensation(uint32_t rx_size, //avg_rho_re[aarx][aatx*nb_aatx+atx] = 0; //avg_rho_im[aarx][aatx*nb_aatx+atx] = 0; rho128 = (__m128i *)&rho[aarx][aatx*nb_aatx+atx][symbol*nb_rb*12]; - dl_ch128 = (__m128i *)&dl_ch_estimates_ext[aatx*frame_parms->nb_antennas_rx+aarx][symbol*nb_rb*12]; - dl_ch128_2 = (__m128i *)&dl_ch_estimates_ext[atx*frame_parms->nb_antennas_rx+aarx][symbol*nb_rb*12]; + dl_ch128 = (__m128i *)dl_ch_estimates_ext[aatx*frame_parms->nb_antennas_rx+aarx]; + dl_ch128_2 = (__m128i *)dl_ch_estimates_ext[atx*frame_parms->nb_antennas_rx+aarx]; for (rb=0; rbnb_antennas_rx; aarx++) { - dl_ch128 = (int16x4_t*)&dl_ch_estimates_ext[(aatx<<1)+aarx][symbol*frame_parms->N_RB_DL*12]; - dl_ch_mag128 = (int16x8_t*)&dl_ch_mag[(aatx<<1)+aarx][symbol*frame_parms->N_RB_DL*12]; - dl_ch_mag128b = (int16x8_t*)&dl_ch_magb[(aatx<<1)+aarx][symbol*frame_parms->N_RB_DL*12]; - rxdataF128 = (int16x4_t*)&rxdataF_ext[aarx][symbol*frame_parms->N_RB_DL*12]; - rxdataF_comp128 = (int16x4x2_t*)&rxdataF_comp[(aatx<<1)+aarx][symbol*frame_parms->N_RB_DL*12]; + dl_ch128 = (int16x4_t*)dl_ch_estimates_ext[(aatx<<1)+aarx]; + dl_ch_mag128 = (int16x8_t *)dl_ch_mag[aatx][aarx]; + dl_ch_mag128b = (int16x8_t *)dl_ch_magb[aatx][aarx]; + rxdataF128 = (int16x4_t *)rxdataF_ext[aarx]; + rxdataF_comp128 = (int16x4x2_t *)(rxdataF_comp[aatx][aarx] + symbol * nb_rb * 12); for (rb=0; rb2) { @@ -1076,8 +1180,8 @@ void nr_dlsch_channel_compensation(uint32_t rx_size, if (rho) { for (aarx=0; aarxnb_antennas_rx; aarx++) { rho128 = (int16x4x2_t*)&rho[aarx][symbol*frame_parms->N_RB_DL*12]; - dl_ch128 = (int16x4_t*)&dl_ch_estimates_ext[aarx][symbol*frame_parms->N_RB_DL*12]; - dl_ch128_2 = (int16x4_t*)&dl_ch_estimates_ext[2+aarx][symbol*frame_parms->N_RB_DL*12]; + dl_ch128 = (int16x4_t*)dl_ch_estimates_ext[aarx]; + dl_ch128_2 = (int16x4_t*)dl_ch_estimates_ext[2+aarx]; for (rb=0; rb 1 ? n_rx : 2; - - dl_ch128 = (__m128i *)&dl_ch_estimates_ext[aatx*x + aarx][start_point]; - dl_ch_mag128 = (__m128i *)&dl_ch_mag[aatx*x + aarx][start_point]; - dl_ch_mag128b = (__m128i *)&dl_ch_magb[aatx*x + aarx][start_point]; - rxdataF128 = (__m128i *)&rxdataF_ext[aarx][start_point]; - rxdataF_comp128 = (__m128i *)&rxdataF_comp[aatx*x + aarx][start_point]; - - length_mod8 = length&7; - if (length_mod8 == 0){ - length2 = length>>3; - - for (ii=0; ii2) { - // 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>3; - - for (ii=0; iinb_antennas_rx)+aarx][symbol*nb_rb*12]; + dl_ch128=(__m128i *)dl_ch_estimates_ext[(aatx*frame_parms->nb_antennas_rx)+aarx]; for (rb=0;rbN_RB_DL*12]; + dl_ch128=(int16x4_t *)dl_ch_estimates_ext[(aatx<<1)+aarx]; for (rb=0; rb>2;//length = number of REs, hence length2=nb_REs*(32/128) in SIMD loop @@ -1565,7 +1459,7 @@ void nr_dlsch_channel_level_median(uint32_t rx_size, min = 0; norm128D = vdupq_n_s32(0); - dl_ch128=(int16x4_t *)&dl_ch_estimates_ext[aatx*n_rx + aarx][start_point]; + dl_ch128=(int16x4_t *)dl_ch_estimates_ext[aatx*n_rx + aarx]; length_mod8=length&3; length2 = length>>2; @@ -1598,23 +1492,23 @@ void nr_dlsch_channel_level_median(uint32_t rx_size, // Extraction functions //============================================================================================== -void nr_dlsch_extract_rbs(uint32_t rxdataF_sz, - c16_t rxdataF[][rxdataF_sz], - const uint32_t rx_size, - const uint32_t pdsch_est_size, - int32_t dl_ch_estimates[][pdsch_est_size], - int32_t rxdataF_ext[][rx_size], - int32_t dl_ch_estimates_ext[][rx_size], - unsigned char symbol, - uint8_t pilots, - uint8_t config_type, - unsigned short start_rb, - unsigned short nb_rb_pdsch, - uint8_t n_dmrs_cdm_groups, - uint8_t Nl, - NR_DL_FRAME_PARMS *frame_parms, - uint16_t dlDmrsSymbPos, - int chest_time_type) +static void nr_dlsch_extract_rbs(uint32_t rxdataF_sz, + c16_t rxdataF[][rxdataF_sz], + uint32_t rx_size_symbol, + uint32_t pdsch_est_size, + int32_t dl_ch_estimates[][pdsch_est_size], + c16_t rxdataF_ext[][rx_size_symbol], + int32_t dl_ch_estimates_ext[][rx_size_symbol], + unsigned char symbol, + uint8_t pilots, + uint8_t config_type, + unsigned short start_rb, + unsigned short nb_rb_pdsch, + uint8_t n_dmrs_cdm_groups, + uint8_t Nl, + NR_DL_FRAME_PARMS *frame_parms, + uint16_t dlDmrsSymbPos, + int chest_time_type) { if (config_type == NFAPI_NR_DMRS_TYPE1) { AssertFatal(n_dmrs_cdm_groups == 1 || n_dmrs_cdm_groups == 2, @@ -1633,14 +1527,13 @@ void nr_dlsch_extract_rbs(uint32_t rxdataF_sz, validDmrsEst = get_next_dmrs_symbol_in_slot(dlDmrsSymbPos,0,14); // get first dmrs symbol index for (unsigned char aarx = 0; aarx < frame_parms->nb_antennas_rx; aarx++) { - - int32_t *rxF_ext = &rxdataF_ext[aarx][symbol * nb_rb_pdsch * NR_NB_SC_PER_RB]; - int32_t *rxF = (int32_t *)&rxdataF[aarx][symbol * frame_parms->ofdm_symbol_size]; + c16_t *rxF_ext = rxdataF_ext[aarx]; + c16_t *rxF = &rxdataF[aarx][symbol * frame_parms->ofdm_symbol_size]; for (unsigned char aatx = 0; aatx < Nl; aatx++) { int32_t *dl_ch0 = &dl_ch_estimates[(aatx*frame_parms->nb_antennas_rx)+aarx][validDmrsEst * frame_parms->ofdm_symbol_size]; - int32_t *dl_ch0_ext = &dl_ch_estimates_ext[(aatx*frame_parms->nb_antennas_rx)+aarx][symbol * nb_rb_pdsch * NR_NB_SC_PER_RB]; + int32_t *dl_ch0_ext = dl_ch_estimates_ext[(aatx*frame_parms->nb_antennas_rx)+aarx]; if (pilots == 0) { //data symbol only if (aatx == 0) { @@ -1720,17 +1613,18 @@ void nr_dlsch_extract_rbs(uint32_t rxdataF_sz, } } -void nr_dlsch_detection_mrc(uint32_t rx_size, - int32_t rxdataF_comp[][rx_size], - int ***rho, - int32_t dl_ch_mag[][rx_size], - int32_t dl_ch_magb[][rx_size], - int32_t dl_ch_magr[][rx_size], +void nr_dlsch_detection_mrc(uint32_t rx_size_symbol, short n_tx, short n_rx, + int32_t rxdataF_comp[][n_rx][rx_size_symbol * NR_SYMBOLS_PER_SLOT], + int ***rho, + int32_t dl_ch_mag[][n_rx][rx_size_symbol], + int32_t dl_ch_magb[][n_rx][rx_size_symbol], + int32_t dl_ch_magr[][n_rx][rx_size_symbol], unsigned char symbol, unsigned short nb_rb, - int length) { + int length) +{ #if defined(__x86_64__)||defined(__i386__) unsigned char aatx, aarx; int i; @@ -1740,15 +1634,15 @@ void nr_dlsch_detection_mrc(uint32_t rx_size, if (n_rx>1) { for (aatx=0; aatx 0, ""); if(size==1) { - nr_element_sign(a44[0],//a - ad_bc,//b + nr_element_sign(a44[0][0], // a + ad_bc, // b nb_rb, sign); } else { - + int16_t k, rr[size - 1], cc[size - 1]; + c16_t outtemp[12 * nb_rb] __attribute__((aligned(32))); + c16_t outtemp1[12 * nb_rb] __attribute__((aligned(32))); + c16_t *sub_matrix[size - 1][size - 1]; for (int rtx=0;rtx 0); if(size==1) { - return((double complex)a44_cpx[0]*sign); + return (a44_cpx[0][0] * sign); }else { + double complex sub_matrix[size - 1][size - 1]; + int16_t k, rr[size - 1], cc[size - 1]; outtemp1 = 0; for (int rtx=0;rtx 1); int16_t k,rr[size-1],cc[size-1]; if(flag) {//fixed point SIMD calc. //Allocate the submatrix elements - int32_t **sub_matrix; - sub_matrix = (int32_t **)malloc16_clear( (size-1)*(size-1)*sizeof(int32_t *) ); - for (int rtx=0;rtx<(size-1);rtx++) {//row - for (int ctx=0;ctx<(size-1);ctx++) {//column - sub_matrix[ctx*(size-1)+rtx] = (int32_t *)malloc16_clear( 12*nb_rb*sizeof(int32_t) ); - } - } + c16_t *sub_matrix[size - 1][size - 1]; //Compute Matrix determinant - nr_determin(a44,// - ad_bc,//determinant - size,//size + nr_determin(size, + a44, // + ad_bc, // determinant nb_rb, +1, shift0); @@ -2049,39 +1908,37 @@ uint8_t nr_matrix_inverse(int32_t **a44,//Input matrix//conjH_H_elements[0] //fill out the sub matrix corresponds to this element for (int ridx=0;ridx<(size-1);ridx++) for (int cidx=0;cidx<(size-1);cidx++) - sub_matrix[cidx*(size-1)+ridx]= (int32_t *)&a44[cc[cidx]*size+rr[ridx]][0]; + // To verify + sub_matrix[cidx][ridx]=a44[cc[cidx]][rr[ridx]]; - nr_determin(sub_matrix, - inv_H_h_H[rtx*size+ctx],//out transpose - size-1,//size + nr_determin(size - 1, // size + sub_matrix, + inv_H_h_H[rtx][ctx], // out transpose nb_rb, - ((rtx&1)==1?-1:1)*((ctx&1)==1?-1:1), + ((rtx & 1) == 1 ? -1 : 1) * ((ctx & 1) == 1 ? -1 : 1), shift0); - //printf("H_h_H(r%d,c%d)=%d+j%d --> inv_H_h_H(%d,%d) = %d+j%d \n",rtx,ctx,((short *)a44[ctx*size+rtx])[0],((short *)a44[ctx*size+rtx])[1],ctx,rtx,((short *)inv_H_h_H[rtx*size+ctx])[0],((short *)inv_H_h_H[rtx*size+ctx])[1]); } } - _mm_empty(); - _m_empty(); } else {//floating point calc. //Allocate the submatrix elements - double complex sub_matrix_cpx[(size-1)*(size-1)]; + double complex sub_matrix_cpx[size - 1][size - 1]; //Convert the IQ samples (in Q15 format) to float complex - double complex a44_cpx[size*size]; - double complex inv_H_h_H_cpx[size*size]; + double complex a44_cpx[size][size]; + double complex inv_H_h_H_cpx[size][size]; double complex determin_cpx; for (int i=0; i<12*nb_rb; i++) { //Convert Q15 to floating point for (int rtx=0;rtx inv_H_h_H(%d,%d) = %lf+j%lf \n",rtx,ctx,creal(a44_cpx[ctx*size+rtx]),cimag(a44_cpx[ctx*size+rtx]),ctx,rtx,creal(inv_H_h_H_cpx[rtx*size+ctx]),cimag(inv_H_h_H_cpx[rtx*size+ctx])); - if (creal(inv_H_h_H_cpx[rtx*size+ctx])>0) - ((short *) inv_H_h_H[rtx*size+ctx])[i<<1] = (short) ((creal(inv_H_h_H_cpx[rtx*size+ctx])*(1<<(shift0-1)))+0.5);//Convert to Q 18 + if (creal(inv_H_h_H_cpx[rtx][ctx]) > 0) + inv_H_h_H[rtx][ctx][i].r = (short)((creal(inv_H_h_H_cpx[rtx][ctx]) * (1 << (shift0 - 1))) + 0.5); // Convert to Q 18 else - ((short *) inv_H_h_H[rtx*size+ctx])[i<<1] = (short) ((creal(inv_H_h_H_cpx[rtx*size+ctx])*(1<<(shift0-1)))-0.5);// + inv_H_h_H[rtx][ctx][i].r = (short)((creal(inv_H_h_H_cpx[rtx][ctx]) * (1 << (shift0 - 1))) - 0.5); // - if (cimag(inv_H_h_H_cpx[rtx*size+ctx])>0) - ((short *) inv_H_h_H[rtx*size+ctx])[(i<<1)+1] = (short) ((cimag(inv_H_h_H_cpx[rtx*size+ctx])*(1<<(shift0-1)))+0.5);// + if (cimag(inv_H_h_H_cpx[rtx][ctx]) > 0) + inv_H_h_H[rtx][ctx][i].i = (short)((cimag(inv_H_h_H_cpx[rtx][ctx]) * (1 << (shift0 - 1))) + 0.5); // else - ((short *) inv_H_h_H[rtx*size+ctx])[(i<<1)+1] = (short) ((cimag(inv_H_h_H_cpx[rtx*size+ctx])*(1<<(shift0-1)))-0.5);// + inv_H_h_H[rtx][ctx][i].i = (short)((cimag(inv_H_h_H_cpx[rtx][ctx]) * (1 << (shift0 - 1))) - 0.5); // //if (i<4) printf("inv_H_h_H_FP(%d,%d)= %d+j%d \n",ctx,rtx, ((short *) inv_H_h_H[rtx*size+ctx])[i<<1],((short *) inv_H_h_H[rtx*size+ctx])[(i<<1)+1]); } @@ -2183,110 +2040,95 @@ void nr_conjch0_mult_ch1(int *ch0, * * * */ -uint8_t nr_zero_forcing_rx(uint32_t rx_size, - int32_t rxdataF_comp[][rx_size], - int32_t dl_ch_mag[][rx_size], - int32_t dl_ch_magb[][rx_size], - int32_t dl_ch_magr[][rx_size], - int32_t dl_ch_estimates_ext[][rx_size], - unsigned short nb_rb, +uint8_t nr_zero_forcing_rx(uint32_t rx_size_symbol, unsigned char n_rx, unsigned char n_tx,//number of layer + int32_t rxdataF_comp[][n_rx][rx_size_symbol * NR_SYMBOLS_PER_SLOT], + int32_t dl_ch_mag[][n_rx][rx_size_symbol], + int32_t dl_ch_magb[][n_rx][rx_size_symbol], + int32_t dl_ch_magr[][n_rx][rx_size_symbol], + int32_t dl_ch_estimates_ext[][rx_size_symbol], + unsigned short nb_rb, unsigned char mod_order, int shift, unsigned char symbol, int length) { int *ch0r, *ch0c; - int32_t *** conjH_H_elements; uint32_t nb_rb_0 = length/12 + ((length%12)?1:0); - int32_t determ_fin[12*nb_rb_0] __attribute__((aligned(32))); + c16_t determ_fin[12 * nb_rb_0] __attribute__((aligned(32))); ///Allocate H^*H matrix elements and sub elements - conjH_H_elements = (int32_t ***)malloc16_clear( n_rx*sizeof(int32_t **) ); - for (int aarx=0;aarx6 -#define interference_abs_64qam_epi16(psi,int_ch_mag,int_two_ch_mag,int_three_ch_mag,a,c1,c3,c5,c7) tmp_result = _mm_cmplt_epi16(psi,int_two_ch_mag); tmp_result3 = _mm_xor_si128(tmp_result,(*(__m128i*)&nr_ones[0])); tmp_result2 = _mm_cmplt_epi16(psi,int_ch_mag); tmp_result = _mm_xor_si128(tmp_result,tmp_result2); tmp_result4 = _mm_cmpgt_epi16(psi,int_three_ch_mag); tmp_result3 = _mm_xor_si128(tmp_result3,tmp_result4); tmp_result = _mm_and_si128(tmp_result,c3); tmp_result2 = _mm_and_si128(tmp_result2,c1); tmp_result3 = _mm_and_si128(tmp_result3,c5); tmp_result4 = _mm_and_si128(tmp_result4,c7); tmp_result = _mm_or_si128(tmp_result,tmp_result2); tmp_result3 = _mm_or_si128(tmp_result3,tmp_result4); a = _mm_or_si128(tmp_result,tmp_result3); - -// calculates a_sq = int_ch_mag*(a_r^2 + a_i^2)*scale_factor -#define square_a_epi16(a_r,a_i,int_ch_mag,scale_factor,a_sq) tmp_result = _mm_mulhi_epi16(a_r,a_r); tmp_result = _mm_slli_epi16(tmp_result,1); tmp_result = _mm_mulhi_epi16(tmp_result,scale_factor); tmp_result = _mm_slli_epi16(tmp_result,1); tmp_result = _mm_mulhi_epi16(tmp_result,int_ch_mag); tmp_result = _mm_slli_epi16(tmp_result,1); tmp_result2 = _mm_mulhi_epi16(a_i,a_i); tmp_result2 = _mm_slli_epi16(tmp_result2,1); tmp_result2 = _mm_mulhi_epi16(tmp_result2,scale_factor); tmp_result2 = _mm_slli_epi16(tmp_result2,1); tmp_result2 = _mm_mulhi_epi16(tmp_result2,int_ch_mag); tmp_result2 = _mm_slli_epi16(tmp_result2,1); a_sq = _mm_adds_epi16(tmp_result,tmp_result2); - -// calculates a_sq = int_ch_mag*(a_r^2 + a_i^2)*scale_factor for 64-QAM -#define square_a_64qam_epi16(a_r,a_i,int_ch_mag,scale_factor,a_sq) tmp_result = _mm_mulhi_epi16(a_r,a_r); tmp_result = _mm_slli_epi16(tmp_result,1); tmp_result = _mm_mulhi_epi16(tmp_result,scale_factor); tmp_result = _mm_slli_epi16(tmp_result,3); tmp_result = _mm_mulhi_epi16(tmp_result,int_ch_mag); tmp_result = _mm_slli_epi16(tmp_result,1); tmp_result2 = _mm_mulhi_epi16(a_i,a_i); tmp_result2 = _mm_slli_epi16(tmp_result2,1); tmp_result2 = _mm_mulhi_epi16(tmp_result2,scale_factor); tmp_result2 = _mm_slli_epi16(tmp_result2,3); tmp_result2 = _mm_mulhi_epi16(tmp_result2,int_ch_mag); tmp_result2 = _mm_slli_epi16(tmp_result2,1); a_sq = _mm_adds_epi16(tmp_result,tmp_result2); - -#elif defined(__arm__) || defined(__aarch64__) - +__m128i rho_rmi __attribute__((aligned(16))); #endif //============================================================================================== @@ -703,9 +129,9 @@ void nr_dlsch_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, #endif #if defined(__x86_64__) || defined(__i386__) - ch_mag = (__m128i*)&dl_ch_mag[(symbol*nb_rb*12)]; + ch_mag = (__m128i *)dl_ch_mag; #elif defined(__arm__) || defined(__aarch64__) - ch_mag = (int16x8_t*)&dl_ch_mag[(symbol*nb_rb*12)]; + ch_mag = (int16x8_t *)dl_ch_mag; #endif @@ -719,6 +145,7 @@ void nr_dlsch_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, for (i=0; i>2; // length in quad words (4 REs) len2+=((len_mod4==0)?0:1); for (i=0; i>2; i+=2) { // in each iteration, we take 8 complex samples #if defined(__x86_64__) || defined(__i386__) - xmm0 = rho01_128i[i]; // 4 symbols - xmm1 = rho01_128i[i+1]; + __m128i xmm0 = rho01_128i[i]; // 4 symbols + __m128i xmm1 = rho01_128i[i + 1]; // put (rho_r + rho_i)/2sqrt2 in rho_rpi // put (rho_r - rho_i)/2sqrt2 in rho_rmi @@ -1113,10 +502,10 @@ void nr_qpsk_qpsk(short *stream0_in, xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) - xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) - rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) - rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) + __m128i xmm2 = _mm_unpacklo_epi64(xmm0, xmm1); // Re(rho) + __m128i xmm3 = _mm_unpackhi_epi64(xmm0, xmm1); // Im(rho) + __m128i rho_rpi = _mm_adds_epi16(xmm2, xmm3); // rho = Re(rho) + Im(rho) + __m128i rho_rmi = _mm_subs_epi16(xmm2, xmm3); // rho* = Re(rho) - Im(rho) // divide by sqrt(8), no shift needed ONE_OVER_SQRT_8 = Q1.16 rho_rpi = _mm_mulhi_epi16(rho_rpi,ONE_OVER_SQRT_8); @@ -1140,11 +529,11 @@ void nr_qpsk_qpsk(short *stream0_in, xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] - y0i = _mm_unpackhi_epi64(xmm0,xmm1); + __m128i y0r = _mm_unpacklo_epi64(xmm0, xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] + __m128i y0i = _mm_unpackhi_epi64(xmm0, xmm1); - y0r_over2 = _mm_srai_epi16(y0r,1); // divide by 2 - y0i_over2 = _mm_srai_epi16(y0i,1); // divide by 2 + __m128i y0r_over2 = _mm_srai_epi16(y0r, 1); // divide by 2 + __m128i y0i_over2 = _mm_srai_epi16(y0i, 1); // divide by 2 #elif defined(__arm__) || defined(__aarch64__) @@ -1162,11 +551,11 @@ void nr_qpsk_qpsk(short *stream0_in, xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] - y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] + __m128i y1r = _mm_unpacklo_epi64(xmm0, xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] + __m128i y1i = _mm_unpackhi_epi64(xmm0, xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] - y1r_over2 = _mm_srai_epi16(y1r,1); // divide by 2 - y1i_over2 = _mm_srai_epi16(y1i,1); // divide by 2 + __m128i y1r_over2 = _mm_srai_epi16(y1r, 1); // divide by 2 + __m128i y1i_over2 = _mm_srai_epi16(y1i, 1); // divide by 2 // Compute the terms for the LLR of first bit @@ -1178,7 +567,7 @@ void nr_qpsk_qpsk(short *stream0_in, xmm2 = _mm_adds_epi16(A,y0i_over2); // = |y1r/2 - rho/sqrt(8)| + y0i/2 xmm3 = _mm_subs_epi16(y1i_over2,rho_rmi); B = _mm_abs_epi16(xmm3); // B = |y1i/2 - rho*/sqrt(8)| - logmax_num_re0 = _mm_adds_epi16(B,xmm2); // = |y1r/2 - rho/sqrt(8)|+|y1i/2 - rho*/sqrt(8)| + y0i/2 + __m128i logmax_num_re0 = _mm_adds_epi16(B, xmm2); // = |y1r/2 - rho/sqrt(8)|+|y1i/2 - rho*/sqrt(8)| + y0i/2 // 2 term for numerator of LLR xmm3 = _mm_subs_epi16(y1r_over2,rho_rmi); @@ -1195,7 +584,7 @@ void nr_qpsk_qpsk(short *stream0_in, xmm2 = _mm_adds_epi16(E,y0i_over2); // = |y1r/2 + rho*/4| + y0i/2 xmm3 = _mm_subs_epi16(y1i_over2,rho_rpi); F = _mm_abs_epi16(xmm3); // F = |y1i/2 - rho/4| - logmax_den_re0 = _mm_adds_epi16(F,xmm2); // = |y1r/2 + rho*/4| + |y1i/2 - rho/4| + y0i/2 + __m128i logmax_den_re0 = _mm_adds_epi16(F, xmm2); // = |y1r/2 + rho*/4| + |y1i/2 - rho/4| + y0i/2 // 2 term for denominator of LLR xmm3 = _mm_adds_epi16(y1r_over2,rho_rpi); @@ -1210,7 +599,7 @@ void nr_qpsk_qpsk(short *stream0_in, // 1 term for nominator of LLR xmm2 = _mm_adds_epi16(A,y0r_over2); - logmax_num_im0 = _mm_adds_epi16(B,xmm2); // = |y1r/2 - rho/4| + |y1i/2 - rho*/4| + y0r/2 + __m128i logmax_num_im0 = _mm_adds_epi16(B, xmm2); // = |y1r/2 - rho/4| + |y1i/2 - rho*/4| + y0r/2 // 2 term for nominator of LLR xmm2 = _mm_subs_epi16(E,y0r_over2); @@ -1220,7 +609,7 @@ void nr_qpsk_qpsk(short *stream0_in, // 1 term for denominator of LLR xmm2 = _mm_adds_epi16(C,y0r_over2); - logmax_den_im0 = _mm_adds_epi16(D,xmm2); // = |y1r/2 - rho*/4| + |y1i/2 + rho/4| - y0r/2 + __m128i logmax_den_im0 = _mm_adds_epi16(D, xmm2); // = |y1r/2 - rho*/4| + |y1i/2 + rho/4| - y0r/2 xmm2 = _mm_subs_epi16(G,y0r_over2); xmm2 = _mm_adds_epi16(xmm2,H); // = |y1r/2 + rho/4| + |y1i/2 + rho*/4| - y0r/2 @@ -1250,7233 +639,3 @@ void nr_qpsk_qpsk(short *stream0_in, _m_empty(); #endif } - -int nr_dlsch_qpsk_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, - int32_t **rxdataF_comp, - int32_t **rxdataF_comp_i, - int32_t **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) - int32_t **rho_i, - int16_t *dlsch_llr, - uint8_t symbol, - uint8_t first_symbol_flag, - uint16_t nb_rb, - uint16_t pbch_pss_sss_adjust, - int16_t **llr16p) -{ - - int16_t *rxF=(int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *rxF_i=(int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *ch_mag_i = (int16_t*)&dl_ch_mag_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *rho=(int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *llr16; - int len; - uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; - - if (first_symbol_flag == 1) { - llr16 = (int16_t*)dlsch_llr; - } else { - llr16 = (int16_t*)(*llr16p); - } - - AssertFatal(llr16!=NULL,"nr_dlsch_qpsk_qpsk_llr: llr is null, symbol %d\n",symbol); - - - if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { - // if symbol has pilots - if (frame_parms->nb_antenna_ports_gNB!=1) - // in 2 antenna ports we have 8 REs per symbol per RB - len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); - else - // for 1 antenna port we have 10 REs per symbol per RB - len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); - } else { - // symbol has no pilots - len = (nb_rb*12) - pbch_pss_sss_adjust; - } - - nr_qpsk_qam16((short *)rxF, - (short *)rxF_i, - (short *)ch_mag_i, - (short *)llr16, - (short *)rho, - len); - - llr16 += (len<<1); - *llr16p = (short *)llr16; - - return(0); -} - -/* -#if defined(__x86_64__) || defined(__i386__) -__m128i ONE_OVER_SQRT_2 __attribute__((aligned(16))); -__m128i ONE_OVER_SQRT_10 __attribute__((aligned(16))); -__m128i THREE_OVER_SQRT_10 __attribute__((aligned(16))); -__m128i ONE_OVER_SQRT_10_Q15 __attribute__((aligned(16))); -__m128i SQRT_10_OVER_FOUR __attribute__((aligned(16))); -__m128i ch_mag_int; -#endif -*/ -void nr_qpsk_qam16(int16_t *stream0_in, - int16_t *stream1_in, - int16_t *ch_mag_i, - int16_t *stream0_out, - int16_t *rho01, - int32_t length - ) -{ - /* - This function computes the LLRs of stream 0 (s_0) in presence of the interfering stream 1 (s_1) assuming that both symbols are QPSK. It can be used for both MU-MIMO interference-aware receiver or for SU-MIMO receivers. - - Parameters: - stream0_in = Matched filter output y0' = (h0*g0)*y0 - stream1_in = Matched filter output y1' = (h0*g1)*y0 - stream0_out = LLRs - rho01 = Correlation between the two effective channels \rho_{10} = (h1*g1)*(h0*g0) - length = number of resource elements - */ - -#if defined(__x86_64__) || defined(__i386__) - __m128i *rho01_128i = (__m128i *)rho01; - __m128i *stream0_128i_in = (__m128i *)stream0_in; - __m128i *stream1_128i_in = (__m128i *)stream1_in; - __m128i *stream0_128i_out = (__m128i *)stream0_out; - __m128i *ch_mag_128i_i = (__m128i *)ch_mag_i; - __m128i ONE_OVER_SQRT_2 = _mm_set1_epi16(23170); // round(1/sqrt(2)*2^15) - __m128i ONE_OVER_SQRT_10_Q15 = _mm_set1_epi16(10362); // round(1/sqrt(10)*2^15) - __m128i THREE_OVER_SQRT_10 = _mm_set1_epi16(31086); // round(3/sqrt(10)*2^15) - __m128i SQRT_10_OVER_FOUR = _mm_set1_epi16(25905); // round(sqrt(10)/4*2^15) - __m128i ch_mag_int __attribute__((aligned(16))); -#elif defined(__arm__) || defined(__aarch64__) - int16x8_t *rho01_128i = (int16x8_t *)rho01; - int16x8_t *stream0_128i_in = (int16x8_t *)stream0_in; - int16x8_t *stream1_128i_in = (int16x8_t *)stream1_in; - int16x8_t *stream0_128i_out = (int16x8_t *)stream0_out; - int16x8_t *ch_mag_128i_i = (int16x8_t *)ch_mag_i; - int16x8_t ONE_OVER_SQRT_2 = vdupq_n_s16(23170); // round(1/sqrt(2)*2^15) - int16x8_t ONE_OVER_SQRT_10_Q15 = vdupq_n_s16(10362); // round(1/sqrt(10)*2^15) - int16x8_t THREE_OVER_SQRT_10 = vdupq_n_s16(31086); // round(3/sqrt(10)*2^15) - int16x8_t SQRT_10_OVER_FOUR = vdupq_n_s16(25905); // round(sqrt(10)/4*2^15) - int16x8_t ch_mag_int __attribute__((aligned(16))); -#endif - -#ifdef DEBUG_LLR - print_shorts2("rho01_128i:\n",rho01_128i); -#endif - - int i; - - - for (i=0; i>2; i+=2) { - // in each iteration, we take 8 complex samples - -#if defined(__x86_64__) || defined(__i386__) - - xmm0 = rho01_128i[i]; // 4 symbols - xmm1 = rho01_128i[i+1]; - - // put (rho_r + rho_i)/2sqrt2 in rho_rpi - // put (rho_r - rho_i)/2sqrt2 in rho_rmi - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) - xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) - rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) - rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) - - // divide by sqrt(2) - rho_rpi = _mm_mulhi_epi16(rho_rpi, ONE_OVER_SQRT_2); - rho_rmi = _mm_mulhi_epi16(rho_rmi, ONE_OVER_SQRT_2); - rho_rpi = _mm_slli_epi16(rho_rpi,1); - rho_rmi = _mm_slli_epi16(rho_rmi,1); - - // Compute LLR for first bit of stream 0 - - // Compute real and imaginary parts of MF output for stream 0 - xmm0 = stream0_128i_in[i]; - xmm1 = stream0_128i_in[i+1]; - - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] - y0i = _mm_unpackhi_epi64(xmm0,xmm1); - - // divide by sqrt(2) - y0r_over2 = _mm_mulhi_epi16(y0r, ONE_OVER_SQRT_2); - y0i_over2 = _mm_mulhi_epi16(y0i, ONE_OVER_SQRT_2); - y0r_over2 = _mm_slli_epi16(y0r,1); - y0i_over2 = _mm_slli_epi16(y0i,1); - - y0_p_1_1 = _mm_adds_epi16(y0r_over2, y0i_over2); - y0_m_1_1 = _mm_subs_epi16(y0r_over2, y0i_over2); - - // Compute real and imaginary parts of MF output for stream 1 - xmm0 = stream1_128i_in[i]; - xmm1 = stream1_128i_in[i+1]; - - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] - y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] - - xmm0 = _mm_setzero_si128(); // ZERO - - // compute psi - xmm3 = _mm_subs_epi16(y1r,rho_rpi); - psi_r_p1_p1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_subs_epi16(y1i,rho_rmi); - psi_i_p1_p1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_subs_epi16(y1r,rho_rmi); - psi_r_p1_m1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_adds_epi16(y1i,rho_rpi); - psi_i_p1_m1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_adds_epi16(y1r,rho_rmi); - psi_r_m1_p1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_subs_epi16(y1i,rho_rpi); - psi_i_m1_p1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_adds_epi16(y1r,rho_rpi); - psi_r_m1_m1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_adds_epi16(y1i,rho_rmi); - psi_i_m1_m1 = _mm_abs_epi16(xmm3); - - // Rearrange interfering channel magnitudes - xmm2 = ch_mag_128i_i[i]; - xmm3 = ch_mag_128i_i[i+1]; - - xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - - ch_mag_int = _mm_unpacklo_epi64(xmm2,xmm3); - - // calculate optimal interference amplitudes - interference_abs_epi16(psi_r_p1_p1 , ch_mag_int, a_r_p1_p1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_p1 , ch_mag_int, a_i_p1_p1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p1_m1 , ch_mag_int, a_r_p1_m1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_m1 , ch_mag_int, a_i_p1_m1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_p1 , ch_mag_int, a_r_m1_p1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_p1 , ch_mag_int, a_i_m1_p1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_m1 , ch_mag_int, a_r_m1_m1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_m1 , ch_mag_int, a_i_m1_m1 , ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - - // prodsum - prodsum_psi_a_epi16(psi_r_p1_p1, a_r_p1_p1, psi_i_p1_p1, a_i_p1_p1, psi_a_p1_p1); - prodsum_psi_a_epi16(psi_r_p1_m1, a_r_p1_m1, psi_i_p1_m1, a_i_p1_m1, psi_a_p1_m1); - prodsum_psi_a_epi16(psi_r_m1_p1, a_r_m1_p1, psi_i_m1_p1, a_i_m1_p1, psi_a_m1_p1); - prodsum_psi_a_epi16(psi_r_m1_m1, a_r_m1_m1, psi_i_m1_m1, a_i_m1_m1, psi_a_m1_m1); - - // squares - square_a_epi16(a_r_p1_p1, a_i_p1_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_p1); - square_a_epi16(a_r_p1_m1, a_i_p1_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_m1); - square_a_epi16(a_r_m1_p1, a_i_m1_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_p1); - square_a_epi16(a_r_m1_m1, a_i_m1_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_m1); - - // Computing Metrics - xmm0 = _mm_subs_epi16(psi_a_p1_p1, a_sq_p1_p1); - bit_met_p1_p1 = _mm_adds_epi16(xmm0, y0_p_1_1); - - xmm0 = _mm_subs_epi16(psi_a_p1_m1, a_sq_p1_m1); - bit_met_p1_m1 = _mm_adds_epi16(xmm0, y0_m_1_1); - - xmm0 = _mm_subs_epi16(psi_a_m1_p1, a_sq_m1_p1); - bit_met_m1_p1 = _mm_subs_epi16(xmm0, y0_m_1_1); - - xmm0 = _mm_subs_epi16(psi_a_m1_m1, a_sq_m1_m1); - bit_met_m1_m1 = _mm_subs_epi16(xmm0, y0_p_1_1); - - // MSB - logmax_num_re0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_m1); // bit=0 - logmax_den_re0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_m1); // bit=1 - - y0r = _mm_subs_epi16(logmax_num_re0,logmax_den_re0); - - // LSB - logmax_num_im0 = _mm_max_epi16(bit_met_p1_p1,bit_met_m1_p1); // bit=0 - logmax_den_im0 = _mm_max_epi16(bit_met_p1_m1,bit_met_m1_m1); // bit=1 - - y0i = _mm_subs_epi16(logmax_num_im0,logmax_den_im0); - - stream0_128i_out[i] = _mm_unpacklo_epi16(y0r,y0i); // = [L1(1), L2(1), L1(2), L2(2)] - - if (i<((length>>1) - 1)) // false if only 2 REs remain - stream0_128i_out[i+1] = _mm_unpackhi_epi16(y0r,y0i); - -#elif defined(__arm__) || defined(__aarch64__) - -#endif - } - -#if defined(__x86_64__) || defined(__i386__) - _mm_empty(); - _m_empty(); -#endif -} - -int nr_dlsch_qpsk_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, - int32_t **rxdataF_comp, - int32_t **rxdataF_comp_i, - int32_t **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) - int32_t **rho_i, - int16_t *dlsch_llr, - uint8_t symbol, - uint8_t first_symbol_flag, - uint16_t nb_rb, - uint16_t pbch_pss_sss_adjust, - int16_t **llr16p) -{ - - int16_t *rxF=(int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *rxF_i=(int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *ch_mag_i = (int16_t*)&dl_ch_mag_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *rho=(int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *llr16; - int len; - uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; - - - if (first_symbol_flag == 1) { - llr16 = (int16_t*)dlsch_llr; - } else { - llr16 = (int16_t*)(*llr16p); - } - - AssertFatal(llr16!=NULL,"nr_dlsch_qpsk_qam64_llr: llr is null, symbol %d\n",symbol); - - if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { - // if symbol has pilots - if (frame_parms->nb_antenna_ports_gNB!=1) - // in 2 antenna ports we have 8 REs per symbol per RB - len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); - else - // for 1 antenna port we have 10 REs per symbol per RB - len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); - } else { - // symbol has no pilots - len = (nb_rb*12) - pbch_pss_sss_adjust; - } - - nr_qpsk_qam64((short *)rxF, - (short *)rxF_i, - (short *)ch_mag_i, - (short *)llr16, - (short *)rho, - len); - - llr16 += (len<<1); - *llr16p = (short *)llr16; - - return(0); -} -/* -__m128i ONE_OVER_SQRT_2_42 __attribute__((aligned(16))); -__m128i THREE_OVER_SQRT_2_42 __attribute__((aligned(16))); -__m128i FIVE_OVER_SQRT_2_42 __attribute__((aligned(16))); -__m128i SEVEN_OVER_SQRT_2_42 __attribute__((aligned(16))); - -__m128i ch_mag_int_with_sigma2 __attribute__((aligned(16))); -__m128i two_ch_mag_int_with_sigma2 __attribute__((aligned(16))); -__m128i three_ch_mag_int_with_sigma2 __attribute__((aligned(16))); -__m128i SQRT_42_OVER_FOUR __attribute__((aligned(16))); -*/ -void nr_qpsk_qam64(short *stream0_in, - short *stream1_in, - short *ch_mag_i, - short *stream0_out, - short *rho01, - int length - ) -{ - - /* - This function computes the LLRs of stream 0 (s_0) in presence of the interfering stream 1 (s_1) assuming that both symbols are QPSK. It can be used for both MU-MIMO interference-aware receiver or for SU-MIMO receivers. - - Parameters: - stream0_in = Matched filter output y0' = (h0*g0)*y0 - stream1_in = Matched filter output y1' = (h0*g1)*y0 - stream0_out = LLRs - rho01 = Correlation between the two effective channels \rho_{10} = (h1*g1)*(h0*g0) - length = number of resource elements - */ - -#if defined(__x86_64__) || defined(__i386__) - __m128i *rho01_128i = (__m128i *)rho01; - __m128i *stream0_128i_in = (__m128i *)stream0_in; - __m128i *stream1_128i_in = (__m128i *)stream1_in; - __m128i *stream0_128i_out = (__m128i *)stream0_out; - __m128i *ch_mag_128i_i = (__m128i *)ch_mag_i; - __m128i ONE_OVER_SQRT_2 = _mm_set1_epi16(23170); // round(1/sqrt(2)*2^15) - __m128i ONE_OVER_SQRT_2_42 = _mm_set1_epi16(3575); // round(1/sqrt(2*42)*2^15) - __m128i THREE_OVER_SQRT_2_42 = _mm_set1_epi16(10726); // round(3/sqrt(2*42)*2^15) - __m128i FIVE_OVER_SQRT_2_42 = _mm_set1_epi16(17876); // round(5/sqrt(2*42)*2^15) - __m128i SEVEN_OVER_SQRT_2_42 = _mm_set1_epi16(25027); // round(7/sqrt(2*42)*2^15) - __m128i SQRT_42_OVER_FOUR = _mm_set1_epi16(13272); // round(sqrt(42)/4*2^13), Q3.1 - __m128i ch_mag_int; - __m128i ch_mag_int_with_sigma2; - __m128i two_ch_mag_int_with_sigma2; - __m128i three_ch_mag_int_with_sigma2; -#elif defined(__arm__) || defined(__aarch64__) - -#endif - -#ifdef DEBUG_LLR - print_shorts2("rho01_128i:\n",rho01_128i); -#endif - - int i; - - - for (i=0; i>2; i+=2) { - // in each iteration, we take 8 complex samples - -#if defined(__x86_64__) || defined(__i386__) - - xmm0 = rho01_128i[i]; // 4 symbols - xmm1 = rho01_128i[i+1]; - - // put (rho_r + rho_i)/sqrt2 in rho_rpi - // put (rho_r - rho_i)/sqrt2 in rho_rmi - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) - xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) - rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) - rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) - - // divide by sqrt(2) - rho_rpi = _mm_mulhi_epi16(rho_rpi, ONE_OVER_SQRT_2); - rho_rmi = _mm_mulhi_epi16(rho_rmi, ONE_OVER_SQRT_2); - rho_rpi = _mm_slli_epi16(rho_rpi,1); - rho_rmi = _mm_slli_epi16(rho_rmi,1); - - // Compute LLR for first bit of stream 0 - - // Compute real and imaginary parts of MF output for stream 0 - xmm0 = stream0_128i_in[i]; - xmm1 = stream0_128i_in[i+1]; - - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] - y0i = _mm_unpackhi_epi64(xmm0,xmm1); - - // divide by sqrt(2) - y0r_over2 = _mm_mulhi_epi16(y0r, ONE_OVER_SQRT_2); - y0i_over2 = _mm_mulhi_epi16(y0i, ONE_OVER_SQRT_2); - y0r_over2 = _mm_slli_epi16(y0r,1); - y0i_over2 = _mm_slli_epi16(y0i,1); - - y0_p_1_1 = _mm_adds_epi16(y0r_over2, y0i_over2); - y0_m_1_1 = _mm_subs_epi16(y0r_over2, y0i_over2); - - // Compute real and imaginary parts of MF output for stream 1 - xmm0 = stream1_128i_in[i]; - xmm1 = stream1_128i_in[i+1]; - - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] - y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] - - xmm0 = _mm_setzero_si128(); // ZERO - - // compute psi - xmm3 = _mm_subs_epi16(y1r,rho_rpi); - psi_r_p1_p1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_subs_epi16(y1i,rho_rmi); - psi_i_p1_p1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_subs_epi16(y1r,rho_rmi); - psi_r_p1_m1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_adds_epi16(y1i,rho_rpi); - psi_i_p1_m1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_adds_epi16(y1r,rho_rmi); - psi_r_m1_p1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_subs_epi16(y1i,rho_rpi); - psi_i_m1_p1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_adds_epi16(y1r,rho_rpi); - psi_r_m1_m1 = _mm_abs_epi16(xmm3); - xmm3 = _mm_adds_epi16(y1i,rho_rmi); - psi_i_m1_m1 = _mm_abs_epi16(xmm3); - - // Rearrange interfering channel magnitudes - xmm2 = ch_mag_128i_i[i]; - xmm3 = ch_mag_128i_i[i+1]; - - xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - - ch_mag_int = _mm_unpacklo_epi64(xmm2,xmm3); - ch_mag_int_with_sigma2 = _mm_srai_epi16(ch_mag_int, 1); // *2 - two_ch_mag_int_with_sigma2 = ch_mag_int; // *4 - three_ch_mag_int_with_sigma2 = _mm_adds_epi16(ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2); // *6 - - interference_abs_64qam_epi16(psi_r_p1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - - // prodsum - prodsum_psi_a_epi16(psi_r_p1_p1, a_r_p1_p1, psi_i_p1_p1, a_i_p1_p1, psi_a_p1_p1); - prodsum_psi_a_epi16(psi_r_p1_m1, a_r_p1_m1, psi_i_p1_m1, a_i_p1_m1, psi_a_p1_m1); - prodsum_psi_a_epi16(psi_r_m1_p1, a_r_m1_p1, psi_i_m1_p1, a_i_m1_p1, psi_a_m1_p1); - prodsum_psi_a_epi16(psi_r_m1_m1, a_r_m1_m1, psi_i_m1_m1, a_i_m1_m1, psi_a_m1_m1); - - // Multiply by sqrt(2) - psi_a_p1_p1 = _mm_mulhi_epi16(psi_a_p1_p1, ONE_OVER_SQRT_2); - psi_a_p1_p1 = _mm_slli_epi16(psi_a_p1_p1, 2); - psi_a_p1_m1 = _mm_mulhi_epi16(psi_a_p1_m1, ONE_OVER_SQRT_2); - psi_a_p1_m1 = _mm_slli_epi16(psi_a_p1_m1, 2); - psi_a_m1_p1 = _mm_mulhi_epi16(psi_a_m1_p1, ONE_OVER_SQRT_2); - psi_a_m1_p1 = _mm_slli_epi16(psi_a_m1_p1, 2); - psi_a_m1_m1 = _mm_mulhi_epi16(psi_a_m1_m1, ONE_OVER_SQRT_2); - psi_a_m1_m1 = _mm_slli_epi16(psi_a_m1_m1, 2); - - square_a_64qam_epi16(a_r_p1_p1, a_i_p1_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_p1); - square_a_64qam_epi16(a_r_p1_m1, a_i_p1_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_m1); - square_a_64qam_epi16(a_r_m1_p1, a_i_m1_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_p1); - square_a_64qam_epi16(a_r_m1_m1, a_i_m1_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_m1); - - // Computing Metrics - xmm0 = _mm_subs_epi16(psi_a_p1_p1, a_sq_p1_p1); - bit_met_p1_p1 = _mm_adds_epi16(xmm0, y0_p_1_1); - - xmm0 = _mm_subs_epi16(psi_a_p1_m1, a_sq_p1_m1); - bit_met_p1_m1 = _mm_adds_epi16(xmm0, y0_m_1_1); - - xmm0 = _mm_subs_epi16(psi_a_m1_p1, a_sq_m1_p1); - bit_met_m1_p1 = _mm_subs_epi16(xmm0, y0_m_1_1); - - xmm0 = _mm_subs_epi16(psi_a_m1_m1, a_sq_m1_m1); - bit_met_m1_m1 = _mm_subs_epi16(xmm0, y0_p_1_1); - - // MSB - logmax_num_re0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_m1); // bit=0 - logmax_den_re0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_m1); // bit=1 - - y0r = _mm_subs_epi16(logmax_num_re0,logmax_den_re0); - - // LSB - logmax_num_im0 = _mm_max_epi16(bit_met_p1_p1,bit_met_m1_p1); // bit=0 - logmax_den_im0 = _mm_max_epi16(bit_met_p1_m1,bit_met_m1_m1); // bit=1 - - y0i = _mm_subs_epi16(logmax_num_im0,logmax_den_im0); - - stream0_128i_out[i] = _mm_unpacklo_epi16(y0r,y0i); // = [L1(1), L2(1), L1(2), L2(2)] - - if (i<((length>>1) - 1)) // false if only 2 REs remain - stream0_128i_out[i+1] = _mm_unpackhi_epi16(y0r,y0i); - -#elif defined(__arm__) || defined(__aarch64__) - -#endif - } - -#if defined(__x86_64__) || defined(__i386__) - _mm_empty(); - _m_empty(); -#endif -} - - -//---------------------------------------------------------------------------------------------- -// 16-QAM -//---------------------------------------------------------------------------------------------- - -/* -__m128i ONE_OVER_TWO_SQRT_10 __attribute__((aligned(16))); -__m128i NINE_OVER_TWO_SQRT_10 __attribute__((aligned(16))); - -__m128i y0r_over_sqrt10 __attribute__ ((aligned(16))); -__m128i y0i_over_sqrt10 __attribute__ ((aligned(16))); -__m128i y0r_three_over_sqrt10 __attribute__ ((aligned(16))); -__m128i y0i_three_over_sqrt10 __attribute__ ((aligned(16))); - -__m128i ch_mag_des __attribute__((aligned(16))); -__m128i ch_mag_over_10 __attribute__ ((aligned(16))); -__m128i ch_mag_over_2 __attribute__ ((aligned(16))); -__m128i ch_mag_9_over_10 __attribute__ ((aligned(16))); -*/ - -void nr_qam16_qpsk(short *stream0_in, - short *stream1_in, - short *ch_mag, - short *stream0_out, - short *rho01, - int length - ) -{ - - /* - Author: Sebastian Wagner - Date: 2012-06-04 - - Input: - stream0_in: MF filter for 1st stream, i.e., y0=h0'*y - stream!_in: MF filter for 2nd stream, i.e., y1=h1'*y - ch_mag: 2*h0/sqrt(00), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc - ch_mag_i: 2*h1/sqrt(00), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc - rho01: Channel cross correlation, i.e., h1'*h0 - - Output: - stream0_out: output LLRs for 1st stream - */ - -#if defined(__x86_64__) || defined(__i386__) - __m128i *rho01_128i = (__m128i *)rho01; - __m128i *stream0_128i_in = (__m128i *)stream0_in; - __m128i *stream1_128i_in = (__m128i *)stream1_in; - __m128i *stream0_128i_out = (__m128i *)stream0_out; - __m128i *ch_mag_128i = (__m128i *)ch_mag; - __m128i ONE_OVER_SQRT_2 = _mm_set1_epi16(23170); // round(1/sqrt(2)*2^15) - __m128i ONE_OVER_SQRT_10 = _mm_set1_epi16(20724); // round(1/sqrt(10)*2^16) - __m128i THREE_OVER_SQRT_10 = _mm_set1_epi16(31086); // round(3/sqrt(10)*2^15) - __m128i SQRT_10_OVER_FOUR = _mm_set1_epi16(25905); // round(sqrt(10)/4*2^15) - __m128i ONE_OVER_TWO_SQRT_10 = _mm_set1_epi16(10362); // round(1/2/sqrt(10)*2^16) - __m128i NINE_OVER_TWO_SQRT_10 = _mm_set1_epi16(23315); // round(9/2/sqrt(10)*2^14) - __m128i y0r_over_sqrt10; - __m128i y0i_over_sqrt10; - __m128i y0r_three_over_sqrt10; - __m128i y0i_three_over_sqrt10; - - __m128i ch_mag_des; - __m128i ch_mag_over_10; - __m128i ch_mag_over_2; - __m128i ch_mag_9_over_10; -#elif defined(__arm__) || defined(__aarch64__) - -#endif - - int i; - - - for (i=0; i>2; i+=2) { - // In one iteration, we deal with 8 REs - -#if defined(__x86_64__) || defined(__i386__) - // Get rho - xmm0 = rho01_128i[i]; - xmm1 = rho01_128i[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) - xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) - rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) - rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) - - // Compute the different rhos - rho_rpi_1_1 = _mm_mulhi_epi16(rho_rpi,ONE_OVER_SQRT_10); - rho_rmi_1_1 = _mm_mulhi_epi16(rho_rmi,ONE_OVER_SQRT_10); - rho_rpi_3_3 = _mm_mulhi_epi16(rho_rpi,THREE_OVER_SQRT_10); - rho_rmi_3_3 = _mm_mulhi_epi16(rho_rmi,THREE_OVER_SQRT_10); - rho_rpi_3_3 = _mm_slli_epi16(rho_rpi_3_3,1); - rho_rmi_3_3 = _mm_slli_epi16(rho_rmi_3_3,1); - - xmm4 = _mm_mulhi_epi16(xmm2,ONE_OVER_SQRT_10); // Re(rho) - xmm5 = _mm_mulhi_epi16(xmm3,THREE_OVER_SQRT_10); // Im(rho) - xmm5 = _mm_slli_epi16(xmm5,1); - - rho_rpi_1_3 = _mm_adds_epi16(xmm4,xmm5); - rho_rmi_1_3 = _mm_subs_epi16(xmm4,xmm5); - - xmm6 = _mm_mulhi_epi16(xmm2,THREE_OVER_SQRT_10); // Re(rho) - xmm7 = _mm_mulhi_epi16(xmm3,ONE_OVER_SQRT_10); // Im(rho) - xmm6 = _mm_slli_epi16(xmm6,1); - - rho_rpi_3_1 = _mm_adds_epi16(xmm6,xmm7); - rho_rmi_3_1 = _mm_subs_epi16(xmm6,xmm7); - - // Rearrange interfering MF output - xmm0 = stream1_128i_in[i]; - xmm1 = stream1_128i_in[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] - y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] - - xmm0 = _mm_setzero_si128(); // ZERO - xmm2 = _mm_subs_epi16(rho_rpi_1_1,y1r); // = [Re(rho)+ Im(rho)]/sqrt(10) - y1r - psi_r_p1_p1 = _mm_abs_epi16(xmm2); // = |[Re(rho)+ Im(rho)]/sqrt(10) - y1r| - - xmm2= _mm_subs_epi16(rho_rmi_1_1,y1r); - psi_r_p1_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_1_1,y1i); - psi_i_p1_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_1_3,y1r); - psi_r_p1_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_1_3,y1r); - psi_r_p1_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_3_1,y1i); - psi_i_p1_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_3_1,y1r); - psi_r_p3_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_3_1,y1r); - psi_r_p3_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_1_3,y1i); - psi_i_p3_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_3_3,y1r); - psi_r_p3_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_3_3,y1r); - psi_r_p3_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_3_3,y1i); - psi_i_p3_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_1_1,y1i); - psi_i_m1_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_3_1,y1i); - psi_i_m1_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_1_3,y1i); - psi_i_m3_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_3_3,y1i); - psi_i_m3_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_1_1,y1i); - psi_i_p1_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_3_1,y1i); - psi_i_p1_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_1_3,y1i); - psi_i_p3_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_3_3,y1i); - psi_i_p3_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_1_1,y1r); - psi_r_m1_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_1_3,y1r); - psi_r_m1_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_3_1,y1r); - psi_r_m3_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_3_3,y1r); - psi_r_m3_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1r,rho_rmi_1_1); - psi_r_m1_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1r,rho_rmi_1_3); - psi_r_m1_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1i,rho_rmi_1_1); - psi_i_m1_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1i,rho_rmi_3_1); - psi_i_m1_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1r,rho_rmi_3_1); - psi_r_m3_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1r,rho_rmi_3_3); - psi_r_m3_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1i,rho_rmi_1_3); - psi_i_m3_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1i,rho_rmi_3_3); - psi_i_m3_m3 = _mm_abs_epi16(xmm2); - - // Rearrange desired MF output - xmm0 = stream0_128i_in[i]; - xmm1 = stream0_128i_in[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] - y0i = _mm_unpackhi_epi64(xmm0,xmm1); - - // Rearrange desired channel magnitudes - xmm2 = ch_mag_128i[i]; // = [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2)]*(2/sqrt(10)) - xmm3 = ch_mag_128i[i+1]; // = [|h|^2(3),|h|^2(3),|h|^2(4),|h|^2(4)]*(2/sqrt(10)) - xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - - ch_mag_des = _mm_unpacklo_epi64(xmm2,xmm3); // = [|h|^2(1),|h|^2(2),|h|^2(3),|h|^2(4)]*(2/sqrt(10)) - - // Scale MF output of desired signal - y0r_over_sqrt10 = _mm_mulhi_epi16(y0r,ONE_OVER_SQRT_10); - y0i_over_sqrt10 = _mm_mulhi_epi16(y0i,ONE_OVER_SQRT_10); - y0r_three_over_sqrt10 = _mm_mulhi_epi16(y0r,THREE_OVER_SQRT_10); - y0i_three_over_sqrt10 = _mm_mulhi_epi16(y0i,THREE_OVER_SQRT_10); - y0r_three_over_sqrt10 = _mm_slli_epi16(y0r_three_over_sqrt10,1); - y0i_three_over_sqrt10 = _mm_slli_epi16(y0i_three_over_sqrt10,1); - - // Compute necessary combination of required terms - y0_p_1_1 = _mm_adds_epi16(y0r_over_sqrt10,y0i_over_sqrt10); - y0_m_1_1 = _mm_subs_epi16(y0r_over_sqrt10,y0i_over_sqrt10); - - y0_p_1_3 = _mm_adds_epi16(y0r_over_sqrt10,y0i_three_over_sqrt10); - y0_m_1_3 = _mm_subs_epi16(y0r_over_sqrt10,y0i_three_over_sqrt10); - - y0_p_3_1 = _mm_adds_epi16(y0r_three_over_sqrt10,y0i_over_sqrt10); - y0_m_3_1 = _mm_subs_epi16(y0r_three_over_sqrt10,y0i_over_sqrt10); - - y0_p_3_3 = _mm_adds_epi16(y0r_three_over_sqrt10,y0i_three_over_sqrt10); - y0_m_3_3 = _mm_subs_epi16(y0r_three_over_sqrt10,y0i_three_over_sqrt10); - - // Add psi - psi_a_p1_p1 = _mm_adds_epi16(psi_r_p1_p1 ,psi_i_p1_p1); - psi_a_p1_p3 = _mm_adds_epi16(psi_r_p1_p3 ,psi_i_p1_p3); - psi_a_p3_p1 = _mm_adds_epi16(psi_r_p3_p1 ,psi_i_p3_p1); - psi_a_p3_p3 = _mm_adds_epi16(psi_r_p3_p3 ,psi_i_p3_p3); - psi_a_p1_m1 = _mm_adds_epi16(psi_r_p1_m1 ,psi_i_p1_m1); - psi_a_p1_m3 = _mm_adds_epi16(psi_r_p1_m3 ,psi_i_p1_m3); - psi_a_p3_m1 = _mm_adds_epi16(psi_r_p3_m1 ,psi_i_p3_m1); - psi_a_p3_m3 = _mm_adds_epi16(psi_r_p3_m3 ,psi_i_p3_m3); - psi_a_m1_p1 = _mm_adds_epi16(psi_r_m1_p1 ,psi_i_m1_p1); - psi_a_m1_p3 = _mm_adds_epi16(psi_r_m1_p3 ,psi_i_m1_p3); - psi_a_m3_p1 = _mm_adds_epi16(psi_r_m3_p1 ,psi_i_m3_p1); - psi_a_m3_p3 = _mm_adds_epi16(psi_r_m3_p3 ,psi_i_m3_p3); - psi_a_m1_m1 = _mm_adds_epi16(psi_r_m1_m1 ,psi_i_m1_m1); - psi_a_m1_m3 = _mm_adds_epi16(psi_r_m1_m3 ,psi_i_m1_m3); - psi_a_m3_m1 = _mm_adds_epi16(psi_r_m3_m1 ,psi_i_m3_m1); - psi_a_m3_m3 = _mm_adds_epi16(psi_r_m3_m3 ,psi_i_m3_m3); - - // scale by sqrt(2) - psi_a_p1_p1 = _mm_mulhi_epi16(psi_a_p1_p1,ONE_OVER_SQRT_2); - psi_a_p1_p1 = _mm_slli_epi16(psi_a_p1_p1,1); - psi_a_p1_p3 = _mm_mulhi_epi16(psi_a_p1_p3,ONE_OVER_SQRT_2); - psi_a_p1_p3 = _mm_slli_epi16(psi_a_p1_p3,1); - psi_a_p3_p1 = _mm_mulhi_epi16(psi_a_p3_p1,ONE_OVER_SQRT_2); - psi_a_p3_p1 = _mm_slli_epi16(psi_a_p3_p1,1); - psi_a_p3_p3 = _mm_mulhi_epi16(psi_a_p3_p3,ONE_OVER_SQRT_2); - psi_a_p3_p3 = _mm_slli_epi16(psi_a_p3_p3,1); - - psi_a_p1_m1 = _mm_mulhi_epi16(psi_a_p1_m1,ONE_OVER_SQRT_2); - psi_a_p1_m1 = _mm_slli_epi16(psi_a_p1_m1,1); - psi_a_p1_m3 = _mm_mulhi_epi16(psi_a_p1_m3,ONE_OVER_SQRT_2); - psi_a_p1_m3 = _mm_slli_epi16(psi_a_p1_m3,1); - psi_a_p3_m1 = _mm_mulhi_epi16(psi_a_p3_m1,ONE_OVER_SQRT_2); - psi_a_p3_m1 = _mm_slli_epi16(psi_a_p3_m1,1); - psi_a_p3_m3 = _mm_mulhi_epi16(psi_a_p3_m3,ONE_OVER_SQRT_2); - psi_a_p3_m3 = _mm_slli_epi16(psi_a_p3_m3,1); - - psi_a_m1_p1 = _mm_mulhi_epi16(psi_a_m1_p1,ONE_OVER_SQRT_2); - psi_a_m1_p1 = _mm_slli_epi16(psi_a_m1_p1,1); - psi_a_m1_p3 = _mm_mulhi_epi16(psi_a_m1_p3,ONE_OVER_SQRT_2); - psi_a_m1_p3 = _mm_slli_epi16(psi_a_m1_p3,1); - psi_a_m3_p1 = _mm_mulhi_epi16(psi_a_m3_p1,ONE_OVER_SQRT_2); - psi_a_m3_p1 = _mm_slli_epi16(psi_a_m3_p1,1); - psi_a_m3_p3 = _mm_mulhi_epi16(psi_a_m3_p3,ONE_OVER_SQRT_2); - psi_a_m3_p3 = _mm_slli_epi16(psi_a_m3_p3,1); - - psi_a_m1_m1 = _mm_mulhi_epi16(psi_a_m1_m1,ONE_OVER_SQRT_2); - psi_a_m1_m1 = _mm_slli_epi16(psi_a_m1_m1,1); - psi_a_m1_m3 = _mm_mulhi_epi16(psi_a_m1_m3,ONE_OVER_SQRT_2); - psi_a_m1_m3 = _mm_slli_epi16(psi_a_m1_m3,1); - psi_a_m3_m1 = _mm_mulhi_epi16(psi_a_m3_m1,ONE_OVER_SQRT_2); - psi_a_m3_m1 = _mm_slli_epi16(psi_a_m3_m1,1); - psi_a_m3_m3 = _mm_mulhi_epi16(psi_a_m3_m3,ONE_OVER_SQRT_2); - psi_a_m3_m3 = _mm_slli_epi16(psi_a_m3_m3,1); - - // Computing different multiples of channel norms - ch_mag_over_10=_mm_mulhi_epi16(ch_mag_des, ONE_OVER_TWO_SQRT_10); - ch_mag_over_2=_mm_mulhi_epi16(ch_mag_des, SQRT_10_OVER_FOUR); - ch_mag_over_2=_mm_slli_epi16(ch_mag_over_2, 1); - ch_mag_9_over_10=_mm_mulhi_epi16(ch_mag_des, NINE_OVER_TWO_SQRT_10); - ch_mag_9_over_10=_mm_slli_epi16(ch_mag_9_over_10, 2); - - // Computing Metrics - xmm1 = _mm_adds_epi16(psi_a_p1_p1, y0_p_1_1); - bit_met_p1_p1= _mm_subs_epi16(xmm1, ch_mag_over_10); - - xmm1 = _mm_adds_epi16(psi_a_p1_p3, y0_p_1_3); - bit_met_p1_p3= _mm_subs_epi16(xmm1, ch_mag_over_2); - - xmm1 = _mm_adds_epi16(psi_a_p1_m1, y0_m_1_1); - bit_met_p1_m1= _mm_subs_epi16(xmm1, ch_mag_over_10); - - xmm1 = _mm_adds_epi16(psi_a_p1_m3, y0_m_1_3); - bit_met_p1_m3= _mm_subs_epi16(xmm1, ch_mag_over_2); - - xmm1 = _mm_adds_epi16(psi_a_p3_p1, y0_p_3_1); - bit_met_p3_p1= _mm_subs_epi16(xmm1, ch_mag_over_2); - - xmm1 = _mm_adds_epi16(psi_a_p3_p3, y0_p_3_3); - bit_met_p3_p3= _mm_subs_epi16(xmm1, ch_mag_9_over_10); - - xmm1 = _mm_adds_epi16(psi_a_p3_m1, y0_m_3_1); - bit_met_p3_m1= _mm_subs_epi16(xmm1, ch_mag_over_2); - - xmm1 = _mm_adds_epi16(psi_a_p3_m3, y0_m_3_3); - bit_met_p3_m3= _mm_subs_epi16(xmm1, ch_mag_9_over_10); - - xmm1 = _mm_subs_epi16(psi_a_m1_p1, y0_m_1_1); - bit_met_m1_p1= _mm_subs_epi16(xmm1, ch_mag_over_10); - - xmm1 = _mm_subs_epi16(psi_a_m1_p3, y0_m_1_3); - bit_met_m1_p3= _mm_subs_epi16(xmm1, ch_mag_over_2); - - xmm1 = _mm_subs_epi16(psi_a_m1_m1, y0_p_1_1); - bit_met_m1_m1= _mm_subs_epi16(xmm1, ch_mag_over_10); - - xmm1 = _mm_subs_epi16(psi_a_m1_m3, y0_p_1_3); - bit_met_m1_m3= _mm_subs_epi16(xmm1, ch_mag_over_2); - - xmm1 = _mm_subs_epi16(psi_a_m3_p1, y0_m_3_1); - bit_met_m3_p1= _mm_subs_epi16(xmm1, ch_mag_over_2); - - xmm1 = _mm_subs_epi16(psi_a_m3_p3, y0_m_3_3); - bit_met_m3_p3= _mm_subs_epi16(xmm1, ch_mag_9_over_10); - - xmm1 = _mm_subs_epi16(psi_a_m3_m1, y0_p_3_1); - bit_met_m3_m1= _mm_subs_epi16(xmm1, ch_mag_over_2); - - xmm1 = _mm_subs_epi16(psi_a_m3_m3, y0_p_3_3); - bit_met_m3_m3= _mm_subs_epi16(xmm1, ch_mag_9_over_10); - - // LLR of the first bit - // Bit = 1 - xmm0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_p3); - xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m1_m3); - xmm2 = _mm_max_epi16(bit_met_m3_p1,bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m3_m1,bit_met_m3_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_num_re0= _mm_max_epi16(xmm4,xmm5); - - // Bit = 0 - xmm0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_p3); - xmm1 = _mm_max_epi16(bit_met_p1_m1,bit_met_p1_m3); - xmm2 = _mm_max_epi16(bit_met_p3_p1,bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_p3_m1,bit_met_p3_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4,xmm5); - - // LLR of first bit [L1(1), L1(2), L1(3), L1(4), L1(5), L1(6), L1(7), L1(8)] - y0r = _mm_subs_epi16(logmax_den_re0,logmax_num_re0); - - // LLR of the second bit - // Bit = 1 - xmm0 = _mm_max_epi16(bit_met_p1_m1,bit_met_p3_m1); - xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m3_m1); - xmm2 = _mm_max_epi16(bit_met_p1_m3,bit_met_p3_m3); - xmm3 = _mm_max_epi16(bit_met_m1_m3,bit_met_m3_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_num_re1 = _mm_max_epi16(xmm4,xmm5); - - // Bit = 0 - xmm0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p3_p1); - xmm1 = _mm_max_epi16(bit_met_m1_p1,bit_met_m3_p1); - xmm2 = _mm_max_epi16(bit_met_p1_p3,bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_m1_p3,bit_met_m3_p3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_den_re1 = _mm_max_epi16(xmm4,xmm5); - - // LLR of second bit [L2(1), L2(2), L2(3), L2(4)] - y1r = _mm_subs_epi16(logmax_den_re1,logmax_num_re1); - - // LLR of the third bit - // Bit = 1 - xmm0 = _mm_max_epi16(bit_met_m3_p1,bit_met_m3_p3); - xmm1 = _mm_max_epi16(bit_met_m3_m1,bit_met_m3_m3); - xmm2 = _mm_max_epi16(bit_met_p3_p1,bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_p3_m1,bit_met_p3_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_num_im0 = _mm_max_epi16(xmm4,xmm5); - - // Bit = 0 - xmm0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_p3); - xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m1_m3); - xmm2 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_p3); - xmm3 = _mm_max_epi16(bit_met_p1_m1,bit_met_p1_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_den_im0 = _mm_max_epi16(xmm4,xmm5); - - // LLR of third bit [L3(1), L3(2), L3(3), L3(4)] - y0i = _mm_subs_epi16(logmax_den_im0,logmax_num_im0); - - // LLR of the fourth bit - // Bit = 1 - xmm0 = _mm_max_epi16(bit_met_p1_m3,bit_met_p3_m3); - xmm1 = _mm_max_epi16(bit_met_m1_m3,bit_met_m3_m3); - xmm2 = _mm_max_epi16(bit_met_p1_p3,bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_m1_p3,bit_met_m3_p3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_num_im1 = _mm_max_epi16(xmm4,xmm5); - - // Bit = 0 - xmm0 = _mm_max_epi16(bit_met_p1_m1,bit_met_p3_m1); - xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m3_m1); - xmm2 = _mm_max_epi16(bit_met_p1_p1,bit_met_p3_p1); - xmm3 = _mm_max_epi16(bit_met_m1_p1,bit_met_m3_p1); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_den_im1 = _mm_max_epi16(xmm4,xmm5); - - // LLR of fourth bit [L4(1), L4(2), L4(3), L4(4)] - y1i = _mm_subs_epi16(logmax_den_im1,logmax_num_im1); - - // Pack LLRs in output - // [L1(1), L2(1), L1(2), L2(2), L1(3), L2(3), L1(4), L2(4)] - xmm0 = _mm_unpacklo_epi16(y0r,y1r); - // [L1(5), L2(5), L1(6), L2(6), L1(7), L2(7), L1(8), L2(8)] - xmm1 = _mm_unpackhi_epi16(y0r,y1r); - // [L3(1), L4(1), L3(2), L4(2), L3(3), L4(3), L3(4), L4(4)] - xmm2 = _mm_unpacklo_epi16(y0i,y1i); - // [L3(5), L4(5), L3(6), L4(6), L3(7), L4(7), L3(8), L4(8)] - xmm3 = _mm_unpackhi_epi16(y0i,y1i); - - stream0_128i_out[2*i+0] = _mm_unpacklo_epi32(xmm0,xmm2); // 8LLRs, 2REs - stream0_128i_out[2*i+1] = _mm_unpackhi_epi32(xmm0,xmm2); - stream0_128i_out[2*i+2] = _mm_unpacklo_epi32(xmm1,xmm3); - stream0_128i_out[2*i+3] = _mm_unpackhi_epi32(xmm1,xmm3); - -#elif defined(__arm__) || defined(__aarch64__) - -#endif - } - -#if defined(__x86_64__) || defined(__i386__) - _mm_empty(); - _m_empty(); -#endif - -} - -int nr_dlsch_16qam_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, - int32_t **rxdataF_comp, - int32_t **rxdataF_comp_i, - int32_t **dl_ch_mag, //|h_0|^2*(2/sqrt{10}) - int32_t **rho_i, - int16_t *dlsch_llr, - uint8_t symbol, - uint8_t first_symbol_flag, - uint16_t nb_rb, - uint16_t pbch_pss_sss_adjust, - int16_t **llr16p) -{ - - int16_t *rxF = (int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *rxF_i = (int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *ch_mag = (int16_t*)&dl_ch_mag[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *rho = (int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *llr16; - int len; - uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; - - // first symbol has different structure due to more pilots - if (first_symbol_flag == 1) { - llr16 = (int16_t*)dlsch_llr; - } else { - llr16 = (int16_t*)(*llr16p); - } - - AssertFatal(llr16!=NULL,"nr_dlsch_16qam_qpsk_llr: llr is null, symbol %d\n",symbol); - - if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { - // if symbol has pilots - if (frame_parms->nb_antenna_ports_gNB!=1) - // in 2 antenna ports we have 8 REs per symbol per RB - len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); - else - // for 1 antenna port we have 10 REs per symbol per RB - len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); - } else { - // symbol has no pilots - len = (nb_rb*12) - pbch_pss_sss_adjust; - } - - // printf("symbol %d: qam16_llr, len %d (llr16 %p)\n",symbol,len,llr16); - - nr_qam16_qpsk((short *)rxF, - (short *)rxF_i, - (short *)ch_mag, - (short *)llr16, - (short *)rho, - len); - - llr16 += (len<<2); - *llr16p = (short *)llr16; - - return(0); -} - -void nr_qam16_qam16(short *stream0_in, - short *stream1_in, - short *ch_mag, - short *ch_mag_i, - short *stream0_out, - short *rho01, - int length - ) -{ - - /* - Author: Sebastian Wagner - Date: 2012-06-04 - - Input: - stream0_in: MF filter for 1st stream, i.e., y0=h0'*y - stream!_in: MF filter for 2nd stream, i.e., y1=h1'*y - ch_mag: 2*h0/sqrt(00), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc - ch_mag_i: 2*h1/sqrt(00), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc - rho01: Channel cross correlation, i.e., h1'*h0 - - Output: - stream0_out: output LLRs for 1st stream - */ -#if defined(__x86_64__) || defined(__i386__) - __m128i *rho01_128i = (__m128i *)rho01; - __m128i *stream0_128i_in = (__m128i *)stream0_in; - __m128i *stream1_128i_in = (__m128i *)stream1_in; - __m128i *stream0_128i_out = (__m128i *)stream0_out; - __m128i *ch_mag_128i = (__m128i *)ch_mag; - __m128i *ch_mag_128i_i = (__m128i *)ch_mag_i; - - - - __m128i ONE_OVER_SQRT_10 = _mm_set1_epi16(20724); // round(1/sqrt(10)*2^16) - __m128i ONE_OVER_SQRT_10_Q15 = _mm_set1_epi16(10362); // round(1/sqrt(10)*2^15) - __m128i THREE_OVER_SQRT_10 = _mm_set1_epi16(31086); // round(3/sqrt(10)*2^15) - __m128i SQRT_10_OVER_FOUR = _mm_set1_epi16(25905); // round(sqrt(10)/4*2^15) - __m128i ONE_OVER_TWO_SQRT_10 = _mm_set1_epi16(10362); // round(1/2/sqrt(10)*2^16) - __m128i NINE_OVER_TWO_SQRT_10 = _mm_set1_epi16(23315); // round(9/2/sqrt(10)*2^14) - __m128i ch_mag_des,ch_mag_int; - __m128i y0r_over_sqrt10; - __m128i y0i_over_sqrt10; - __m128i y0r_three_over_sqrt10; - __m128i y0i_three_over_sqrt10; - __m128i ch_mag_over_10; - __m128i ch_mag_over_2; - __m128i ch_mag_9_over_10; -#elif defined(__arm__) || defined(__aarch64__) - -#endif - - int i; - - for (i=0; i>2; i+=2) { - // In one iteration, we deal with 8 REs - -#if defined(__x86_64__) || defined(__i386__) - // Get rho - xmm0 = rho01_128i[i]; - xmm1 = rho01_128i[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) - xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) - rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) - rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) - - // Compute the different rhos - rho_rpi_1_1 = _mm_mulhi_epi16(rho_rpi,ONE_OVER_SQRT_10); - rho_rmi_1_1 = _mm_mulhi_epi16(rho_rmi,ONE_OVER_SQRT_10); - rho_rpi_3_3 = _mm_mulhi_epi16(rho_rpi,THREE_OVER_SQRT_10); - rho_rmi_3_3 = _mm_mulhi_epi16(rho_rmi,THREE_OVER_SQRT_10); - rho_rpi_3_3 = _mm_slli_epi16(rho_rpi_3_3,1); - rho_rmi_3_3 = _mm_slli_epi16(rho_rmi_3_3,1); - - xmm4 = _mm_mulhi_epi16(xmm2,ONE_OVER_SQRT_10); // Re(rho) - xmm5 = _mm_mulhi_epi16(xmm3,THREE_OVER_SQRT_10); // Im(rho) - xmm5 = _mm_slli_epi16(xmm5,1); - - rho_rpi_1_3 = _mm_adds_epi16(xmm4,xmm5); - rho_rmi_1_3 = _mm_subs_epi16(xmm4,xmm5); - - xmm6 = _mm_mulhi_epi16(xmm2,THREE_OVER_SQRT_10); // Re(rho) - xmm7 = _mm_mulhi_epi16(xmm3,ONE_OVER_SQRT_10); // Im(rho) - xmm6 = _mm_slli_epi16(xmm6,1); - - rho_rpi_3_1 = _mm_adds_epi16(xmm6,xmm7); - rho_rmi_3_1 = _mm_subs_epi16(xmm6,xmm7); - - // Rearrange interfering MF output - xmm0 = stream1_128i_in[i]; - xmm1 = stream1_128i_in[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] - y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] - - xmm0 = _mm_setzero_si128(); // ZERO - xmm2 = _mm_subs_epi16(rho_rpi_1_1,y1r); // = [Re(rho)+ Im(rho)]/sqrt(10) - y1r - psi_r_p1_p1 = _mm_abs_epi16(xmm2); // = |[Re(rho)+ Im(rho)]/sqrt(10) - y1r| - - xmm2= _mm_subs_epi16(rho_rmi_1_1,y1r); - psi_r_p1_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_1_1,y1i); - psi_i_p1_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_1_3,y1r); - psi_r_p1_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_1_3,y1r); - psi_r_p1_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_3_1,y1i); - psi_i_p1_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_3_1,y1r); - psi_r_p3_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_3_1,y1r); - psi_r_p3_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_1_3,y1i); - psi_i_p3_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_3_3,y1r); - psi_r_p3_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_3_3,y1r); - psi_r_p3_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_3_3,y1i); - psi_i_p3_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_1_1,y1i); - psi_i_m1_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_3_1,y1i); - psi_i_m1_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_1_3,y1i); - psi_i_m3_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_3_3,y1i); - psi_i_m3_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_1_1,y1i); - psi_i_p1_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_3_1,y1i); - psi_i_p1_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_1_3,y1i); - psi_i_p3_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_3_3,y1i); - psi_i_p3_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_1_1,y1r); - psi_r_m1_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_1_3,y1r); - psi_r_m1_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_3_1,y1r); - psi_r_m3_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_3_3,y1r); - psi_r_m3_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1r,rho_rmi_1_1); - psi_r_m1_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1r,rho_rmi_1_3); - psi_r_m1_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1i,rho_rmi_1_1); - psi_i_m1_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1i,rho_rmi_3_1); - psi_i_m1_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1r,rho_rmi_3_1); - psi_r_m3_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1r,rho_rmi_3_3); - psi_r_m3_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1i,rho_rmi_1_3); - psi_i_m3_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1i,rho_rmi_3_3); - psi_i_m3_m3 = _mm_abs_epi16(xmm2); - - // Rearrange desired MF output - xmm0 = stream0_128i_in[i]; - xmm1 = stream0_128i_in[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] - y0i = _mm_unpackhi_epi64(xmm0,xmm1); - - // Rearrange desired channel magnitudes - xmm2 = ch_mag_128i[i]; // = [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2)]*(2/sqrt(10)) - xmm3 = ch_mag_128i[i+1]; // = [|h|^2(3),|h|^2(3),|h|^2(4),|h|^2(4)]*(2/sqrt(10)) - xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - - ch_mag_des = _mm_unpacklo_epi64(xmm2,xmm3); // = [|h|^2(1),|h|^2(2),|h|^2(3),|h|^2(4)]*(2/sqrt(10)) - - // Rearrange interfering channel magnitudes - xmm2 = ch_mag_128i_i[i]; - xmm3 = ch_mag_128i_i[i+1]; - - xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - - ch_mag_int = _mm_unpacklo_epi64(xmm2,xmm3); - - // Scale MF output of desired signal - y0r_over_sqrt10 = _mm_mulhi_epi16(y0r,ONE_OVER_SQRT_10); - y0i_over_sqrt10 = _mm_mulhi_epi16(y0i,ONE_OVER_SQRT_10); - y0r_three_over_sqrt10 = _mm_mulhi_epi16(y0r,THREE_OVER_SQRT_10); - y0i_three_over_sqrt10 = _mm_mulhi_epi16(y0i,THREE_OVER_SQRT_10); - y0r_three_over_sqrt10 = _mm_slli_epi16(y0r_three_over_sqrt10,1); - y0i_three_over_sqrt10 = _mm_slli_epi16(y0i_three_over_sqrt10,1); - - // Compute necessary combination of required terms - y0_p_1_1 = _mm_adds_epi16(y0r_over_sqrt10,y0i_over_sqrt10); - y0_m_1_1 = _mm_subs_epi16(y0r_over_sqrt10,y0i_over_sqrt10); - - y0_p_1_3 = _mm_adds_epi16(y0r_over_sqrt10,y0i_three_over_sqrt10); - y0_m_1_3 = _mm_subs_epi16(y0r_over_sqrt10,y0i_three_over_sqrt10); - - y0_p_3_1 = _mm_adds_epi16(y0r_three_over_sqrt10,y0i_over_sqrt10); - y0_m_3_1 = _mm_subs_epi16(y0r_three_over_sqrt10,y0i_over_sqrt10); - - y0_p_3_3 = _mm_adds_epi16(y0r_three_over_sqrt10,y0i_three_over_sqrt10); - y0_m_3_3 = _mm_subs_epi16(y0r_three_over_sqrt10,y0i_three_over_sqrt10); - - // Compute optimal interfering symbol magnitude - interference_abs_epi16(psi_r_p1_p1 ,ch_mag_int,a_r_p1_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_p1 ,ch_mag_int,a_i_p1_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p1_p3 ,ch_mag_int,a_r_p1_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_p3 ,ch_mag_int,a_i_p1_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p1_m1 ,ch_mag_int,a_r_p1_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_m1 ,ch_mag_int,a_i_p1_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p1_m3 ,ch_mag_int,a_r_p1_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_m3 ,ch_mag_int,a_i_p1_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p3_p1 ,ch_mag_int,a_r_p3_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p3_p1 ,ch_mag_int,a_i_p3_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p3_p3 ,ch_mag_int,a_r_p3_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p3_p3 ,ch_mag_int,a_i_p3_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p3_m1 ,ch_mag_int,a_r_p3_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p3_m1 ,ch_mag_int,a_i_p3_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p3_m3 ,ch_mag_int,a_r_p3_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p3_m3 ,ch_mag_int,a_i_p3_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_p1 ,ch_mag_int,a_r_m1_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_p1 ,ch_mag_int,a_i_m1_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_p3 ,ch_mag_int,a_r_m1_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_p3 ,ch_mag_int,a_i_m1_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_m1 ,ch_mag_int,a_r_m1_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_m1 ,ch_mag_int,a_i_m1_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_m3 ,ch_mag_int,a_r_m1_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_m3 ,ch_mag_int,a_i_m1_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m3_p1 ,ch_mag_int,a_r_m3_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m3_p1 ,ch_mag_int,a_i_m3_p1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m3_p3 ,ch_mag_int,a_r_m3_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m3_p3 ,ch_mag_int,a_i_m3_p3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m3_m1 ,ch_mag_int,a_r_m3_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m3_m1 ,ch_mag_int,a_i_m3_m1,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m3_m3 ,ch_mag_int,a_r_m3_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m3_m3 ,ch_mag_int,a_i_m3_m3,ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - - // Calculation of groups of two terms in the bit metric involving product of psi and interference magnitude - prodsum_psi_a_epi16(psi_r_p1_p1,a_r_p1_p1,psi_i_p1_p1,a_i_p1_p1,psi_a_p1_p1); - prodsum_psi_a_epi16(psi_r_p1_p3,a_r_p1_p3,psi_i_p1_p3,a_i_p1_p3,psi_a_p1_p3); - prodsum_psi_a_epi16(psi_r_p3_p1,a_r_p3_p1,psi_i_p3_p1,a_i_p3_p1,psi_a_p3_p1); - prodsum_psi_a_epi16(psi_r_p3_p3,a_r_p3_p3,psi_i_p3_p3,a_i_p3_p3,psi_a_p3_p3); - prodsum_psi_a_epi16(psi_r_p1_m1,a_r_p1_m1,psi_i_p1_m1,a_i_p1_m1,psi_a_p1_m1); - prodsum_psi_a_epi16(psi_r_p1_m3,a_r_p1_m3,psi_i_p1_m3,a_i_p1_m3,psi_a_p1_m3); - prodsum_psi_a_epi16(psi_r_p3_m1,a_r_p3_m1,psi_i_p3_m1,a_i_p3_m1,psi_a_p3_m1); - prodsum_psi_a_epi16(psi_r_p3_m3,a_r_p3_m3,psi_i_p3_m3,a_i_p3_m3,psi_a_p3_m3); - prodsum_psi_a_epi16(psi_r_m1_p1,a_r_m1_p1,psi_i_m1_p1,a_i_m1_p1,psi_a_m1_p1); - prodsum_psi_a_epi16(psi_r_m1_p3,a_r_m1_p3,psi_i_m1_p3,a_i_m1_p3,psi_a_m1_p3); - prodsum_psi_a_epi16(psi_r_m3_p1,a_r_m3_p1,psi_i_m3_p1,a_i_m3_p1,psi_a_m3_p1); - prodsum_psi_a_epi16(psi_r_m3_p3,a_r_m3_p3,psi_i_m3_p3,a_i_m3_p3,psi_a_m3_p3); - prodsum_psi_a_epi16(psi_r_m1_m1,a_r_m1_m1,psi_i_m1_m1,a_i_m1_m1,psi_a_m1_m1); - prodsum_psi_a_epi16(psi_r_m1_m3,a_r_m1_m3,psi_i_m1_m3,a_i_m1_m3,psi_a_m1_m3); - prodsum_psi_a_epi16(psi_r_m3_m1,a_r_m3_m1,psi_i_m3_m1,a_i_m3_m1,psi_a_m3_m1); - prodsum_psi_a_epi16(psi_r_m3_m3,a_r_m3_m3,psi_i_m3_m3,a_i_m3_m3,psi_a_m3_m3); - - - // squared interference magnitude times int. ch. power - square_a_epi16(a_r_p1_p1,a_i_p1_p1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p1_p1); - square_a_epi16(a_r_p1_p3,a_i_p1_p3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p1_p3); - square_a_epi16(a_r_p3_p1,a_i_p3_p1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p3_p1); - square_a_epi16(a_r_p3_p3,a_i_p3_p3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p3_p3); - square_a_epi16(a_r_p1_m1,a_i_p1_m1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p1_m1); - square_a_epi16(a_r_p1_m3,a_i_p1_m3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p1_m3); - square_a_epi16(a_r_p3_m1,a_i_p3_m1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p3_m1); - square_a_epi16(a_r_p3_m3,a_i_p3_m3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_p3_m3); - square_a_epi16(a_r_m1_p1,a_i_m1_p1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m1_p1); - square_a_epi16(a_r_m1_p3,a_i_m1_p3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m1_p3); - square_a_epi16(a_r_m3_p1,a_i_m3_p1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m3_p1); - square_a_epi16(a_r_m3_p3,a_i_m3_p3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m3_p3); - square_a_epi16(a_r_m1_m1,a_i_m1_m1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m1_m1); - square_a_epi16(a_r_m1_m3,a_i_m1_m3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m1_m3); - square_a_epi16(a_r_m3_m1,a_i_m3_m1,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m3_m1); - square_a_epi16(a_r_m3_m3,a_i_m3_m3,ch_mag_int,SQRT_10_OVER_FOUR,a_sq_m3_m3); - - // Computing different multiples of channel norms - ch_mag_over_10=_mm_mulhi_epi16(ch_mag_des, ONE_OVER_TWO_SQRT_10); - ch_mag_over_2=_mm_mulhi_epi16(ch_mag_des, SQRT_10_OVER_FOUR); - ch_mag_over_2=_mm_slli_epi16(ch_mag_over_2, 1); - ch_mag_9_over_10=_mm_mulhi_epi16(ch_mag_des, NINE_OVER_TWO_SQRT_10); - ch_mag_9_over_10=_mm_slli_epi16(ch_mag_9_over_10, 2); - - // Computing Metrics - xmm0 = _mm_subs_epi16(psi_a_p1_p1,a_sq_p1_p1); - xmm1 = _mm_adds_epi16(xmm0,y0_p_1_1); - bit_met_p1_p1= _mm_subs_epi16(xmm1,ch_mag_over_10); - - xmm0 = _mm_subs_epi16(psi_a_p1_p3,a_sq_p1_p3); - xmm1 = _mm_adds_epi16(xmm0,y0_p_1_3); - bit_met_p1_p3= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_p1_m1,a_sq_p1_m1); - xmm1 = _mm_adds_epi16(xmm0,y0_m_1_1); - bit_met_p1_m1= _mm_subs_epi16(xmm1,ch_mag_over_10); - - xmm0 = _mm_subs_epi16(psi_a_p1_m3,a_sq_p1_m3); - xmm1 = _mm_adds_epi16(xmm0,y0_m_1_3); - bit_met_p1_m3= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_p3_p1,a_sq_p3_p1); - xmm1 = _mm_adds_epi16(xmm0,y0_p_3_1); - bit_met_p3_p1= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_p3_p3,a_sq_p3_p3); - xmm1 = _mm_adds_epi16(xmm0,y0_p_3_3); - bit_met_p3_p3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); - - xmm0 = _mm_subs_epi16(psi_a_p3_m1,a_sq_p3_m1); - xmm1 = _mm_adds_epi16(xmm0,y0_m_3_1); - bit_met_p3_m1= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_p3_m3,a_sq_p3_m3); - xmm1 = _mm_adds_epi16(xmm0,y0_m_3_3); - bit_met_p3_m3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); - - xmm0 = _mm_subs_epi16(psi_a_m1_p1,a_sq_m1_p1); - xmm1 = _mm_subs_epi16(xmm0,y0_m_1_1); - bit_met_m1_p1= _mm_subs_epi16(xmm1,ch_mag_over_10); - - xmm0 = _mm_subs_epi16(psi_a_m1_p3,a_sq_m1_p3); - xmm1 = _mm_subs_epi16(xmm0,y0_m_1_3); - bit_met_m1_p3= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_m1_m1,a_sq_m1_m1); - xmm1 = _mm_subs_epi16(xmm0,y0_p_1_1); - bit_met_m1_m1= _mm_subs_epi16(xmm1,ch_mag_over_10); - - xmm0 = _mm_subs_epi16(psi_a_m1_m3,a_sq_m1_m3); - xmm1 = _mm_subs_epi16(xmm0,y0_p_1_3); - bit_met_m1_m3= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_m3_p1,a_sq_m3_p1); - xmm1 = _mm_subs_epi16(xmm0,y0_m_3_1); - bit_met_m3_p1= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_m3_p3,a_sq_m3_p3); - xmm1 = _mm_subs_epi16(xmm0,y0_m_3_3); - bit_met_m3_p3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); - - xmm0 = _mm_subs_epi16(psi_a_m3_m1,a_sq_m3_m1); - xmm1 = _mm_subs_epi16(xmm0,y0_p_3_1); - bit_met_m3_m1= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_m3_m3,a_sq_m3_m3); - xmm1 = _mm_subs_epi16(xmm0,y0_p_3_3); - bit_met_m3_m3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); - - // LLR of the first bit - // Bit = 1 - xmm0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_p3); - xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m1_m3); - xmm2 = _mm_max_epi16(bit_met_m3_p1,bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m3_m1,bit_met_m3_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_num_re0= _mm_max_epi16(xmm4,xmm5); - - // Bit = 0 - xmm0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_p3); - xmm1 = _mm_max_epi16(bit_met_p1_m1,bit_met_p1_m3); - xmm2 = _mm_max_epi16(bit_met_p3_p1,bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_p3_m1,bit_met_p3_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4,xmm5); - - // LLR of first bit [L1(1), L1(2), L1(3), L1(4), L1(5), L1(6), L1(7), L1(8)] - y0r = _mm_subs_epi16(logmax_den_re0,logmax_num_re0); - - // LLR of the second bit - // Bit = 1 - xmm0 = _mm_max_epi16(bit_met_p1_m1,bit_met_p3_m1); - xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m3_m1); - xmm2 = _mm_max_epi16(bit_met_p1_m3,bit_met_p3_m3); - xmm3 = _mm_max_epi16(bit_met_m1_m3,bit_met_m3_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_num_re1 = _mm_max_epi16(xmm4,xmm5); - - // Bit = 0 - xmm0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p3_p1); - xmm1 = _mm_max_epi16(bit_met_m1_p1,bit_met_m3_p1); - xmm2 = _mm_max_epi16(bit_met_p1_p3,bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_m1_p3,bit_met_m3_p3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_den_re1 = _mm_max_epi16(xmm4,xmm5); - - // LLR of second bit [L2(1), L2(2), L2(3), L2(4)] - y1r = _mm_subs_epi16(logmax_den_re1,logmax_num_re1); - - // LLR of the third bit - // Bit = 1 - xmm0 = _mm_max_epi16(bit_met_m3_p1,bit_met_m3_p3); - xmm1 = _mm_max_epi16(bit_met_m3_m1,bit_met_m3_m3); - xmm2 = _mm_max_epi16(bit_met_p3_p1,bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_p3_m1,bit_met_p3_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_num_im0 = _mm_max_epi16(xmm4,xmm5); - - // Bit = 0 - xmm0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_p3); - xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m1_m3); - xmm2 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_p3); - xmm3 = _mm_max_epi16(bit_met_p1_m1,bit_met_p1_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_den_im0 = _mm_max_epi16(xmm4,xmm5); - - // LLR of third bit [L3(1), L3(2), L3(3), L3(4)] - y0i = _mm_subs_epi16(logmax_den_im0,logmax_num_im0); - - // LLR of the fourth bit - // Bit = 1 - xmm0 = _mm_max_epi16(bit_met_p1_m3,bit_met_p3_m3); - xmm1 = _mm_max_epi16(bit_met_m1_m3,bit_met_m3_m3); - xmm2 = _mm_max_epi16(bit_met_p1_p3,bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_m1_p3,bit_met_m3_p3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_num_im1 = _mm_max_epi16(xmm4,xmm5); - - // Bit = 0 - xmm0 = _mm_max_epi16(bit_met_p1_m1,bit_met_p3_m1); - xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m3_m1); - xmm2 = _mm_max_epi16(bit_met_p1_p1,bit_met_p3_p1); - xmm3 = _mm_max_epi16(bit_met_m1_p1,bit_met_m3_p1); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_den_im1 = _mm_max_epi16(xmm4,xmm5); - - // LLR of fourth bit [L4(1), L4(2), L4(3), L4(4)] - y1i = _mm_subs_epi16(logmax_den_im1,logmax_num_im1); - - // Pack LLRs in output - // [L1(1), L2(1), L1(2), L2(2), L1(3), L2(3), L1(4), L2(4)] - xmm0 = _mm_unpacklo_epi16(y0r,y1r); - // [L1(5), L2(5), L1(6), L2(6), L1(7), L2(7), L1(8), L2(8)] - xmm1 = _mm_unpackhi_epi16(y0r,y1r); - // [L3(1), L4(1), L3(2), L4(2), L3(3), L4(3), L3(4), L4(4)] - xmm2 = _mm_unpacklo_epi16(y0i,y1i); - // [L3(5), L4(5), L3(6), L4(6), L3(7), L4(7), L3(8), L4(8)] - xmm3 = _mm_unpackhi_epi16(y0i,y1i); - - stream0_128i_out[2*i+0] = _mm_unpacklo_epi32(xmm0,xmm2); // 8LLRs, 2REs - stream0_128i_out[2*i+1] = _mm_unpackhi_epi32(xmm0,xmm2); - stream0_128i_out[2*i+2] = _mm_unpacklo_epi32(xmm1,xmm3); - stream0_128i_out[2*i+3] = _mm_unpackhi_epi32(xmm1,xmm3); -#elif defined(__arm__) || defined(__aarch64__) - -#endif - - } - -#if defined(__x86_64__) || defined(__i386__) - _mm_empty(); - _m_empty(); -#endif -} - -int nr_dlsch_16qam_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, - int32_t **rxdataF_comp, - int32_t **rxdataF_comp_i, - int32_t **dl_ch_mag, //|h_0|^2*(2/sqrt{10}) - int32_t **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) - int32_t **rho_i, - int16_t *dlsch_llr, - uint8_t symbol, - uint32_t len, - uint8_t first_symbol_flag, - uint16_t nb_rb, - uint16_t pbch_pss_sss_adjust, - int16_t **llr16p) -{ - - int16_t *rxF = (int16_t*)&rxdataF_comp[0][(symbol*nb_rb*12)]; - int16_t *rxF_i = (int16_t*)&rxdataF_comp_i[0][(symbol*nb_rb*12)]; - int16_t *ch_mag = (int16_t*)&dl_ch_mag[0][(symbol*nb_rb*12)]; - int16_t *ch_mag_i = (int16_t*)&dl_ch_mag_i[0][(symbol*nb_rb*12)]; - int16_t *rho = (int16_t*)&rho_i[0][(symbol*nb_rb*12)]; - int16_t *llr16; - - // first symbol has different structure due to more pilots - if (first_symbol_flag == 1) { - llr16 = (int16_t*)dlsch_llr; - } else { - llr16 = (int16_t*)(*llr16p); - } - - - AssertFatal(llr16!=NULL,"nr_dlsch_16qam_16qam_llr: llr is null, symbol %d\n",symbol); - - // printf("symbol %d: qam16_llr, len %d (llr16 %p)\n",symbol,len,llr16); - - nr_qam16_qam16((short *)rxF, - (short *)rxF_i, - (short *)ch_mag, - (short *)ch_mag_i, - (short *)llr16, - (short *)rho, - len); - - llr16 += (len<<2); - *llr16p = (short *)llr16; - - return(0); -} - -void nr_qam16_qam64(int16_t *stream0_in, - int16_t *stream1_in, - int16_t *ch_mag, - int16_t *ch_mag_i, - int16_t *stream0_out, - int16_t *rho01, - int32_t length - ) -{ - - /* - Author: Sebastian Wagner - Date: 2012-06-04 - - Input: - stream0_in: MF filter for 1st stream, i.e., y0=h0'*y - stream!_in: MF filter for 2nd stream, i.e., y1=h1'*y - ch_mag: 2*h0/sqrt(00), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc - ch_mag_i: 2*h1/sqrt(00), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc - rho01: Channel cross correlation, i.e., h1'*h0 - - Output: - stream0_out: output LLRs for 1st stream - */ - -#if defined(__x86_64__) || defined(__i386__) - __m128i *rho01_128i = (__m128i *)rho01; - __m128i *stream0_128i_in = (__m128i *)stream0_in; - __m128i *stream1_128i_in = (__m128i *)stream1_in; - __m128i *stream0_128i_out = (__m128i *)stream0_out; - __m128i *ch_mag_128i = (__m128i *)ch_mag; - __m128i *ch_mag_128i_i = (__m128i *)ch_mag_i; - - - __m128i ONE_OVER_SQRT_2 = _mm_set1_epi16(23170); // round(1/sqrt(2)*2^15) - __m128i ONE_OVER_SQRT_10 = _mm_set1_epi16(20724); // round(1/sqrt(10)*2^16) - __m128i THREE_OVER_SQRT_10 = _mm_set1_epi16(31086); // round(3/sqrt(10)*2^15) - __m128i SQRT_10_OVER_FOUR = _mm_set1_epi16(25905); // round(sqrt(10)/4*2^15) - __m128i ONE_OVER_TWO_SQRT_10 = _mm_set1_epi16(10362); // round(1/2/sqrt(10)*2^16) - __m128i NINE_OVER_TWO_SQRT_10 = _mm_set1_epi16(23315); // round(9/2/sqrt(10)*2^14) - __m128i ONE_OVER_SQRT_2_42 = _mm_set1_epi16(3575); // round(1/sqrt(2*42)*2^15) - __m128i THREE_OVER_SQRT_2_42 = _mm_set1_epi16(10726); // round(3/sqrt(2*42)*2^15) - __m128i FIVE_OVER_SQRT_2_42 = _mm_set1_epi16(17876); // round(5/sqrt(2*42)*2^15) - __m128i SEVEN_OVER_SQRT_2_42 = _mm_set1_epi16(25027); // round(7/sqrt(2*42)*2^15) - __m128i SQRT_42_OVER_FOUR = _mm_set1_epi16(13272); // round(sqrt(42)/4*2^13), Q3. - __m128i ch_mag_des,ch_mag_int; - __m128i y0r_over_sqrt10; - __m128i y0i_over_sqrt10; - __m128i y0r_three_over_sqrt10; - __m128i y0i_three_over_sqrt10; - __m128i ch_mag_over_10; - __m128i ch_mag_over_2; - __m128i ch_mag_9_over_10; - __m128i ch_mag_int_with_sigma2; - __m128i two_ch_mag_int_with_sigma2; - __m128i three_ch_mag_int_with_sigma2; - -#elif defined(__arm__) || defined(__aarch64__) - -#endif - int i; - - for (i=0; i>2; i+=2) { - // In one iteration, we deal with 8 REs - -#if defined(__x86_64__) || defined(__i386__) - // Get rho - xmm0 = rho01_128i[i]; - xmm1 = rho01_128i[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) - xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) - rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) - rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) - - // Compute the different rhos - rho_rpi_1_1 = _mm_mulhi_epi16(rho_rpi,ONE_OVER_SQRT_10); - rho_rmi_1_1 = _mm_mulhi_epi16(rho_rmi,ONE_OVER_SQRT_10); - rho_rpi_3_3 = _mm_mulhi_epi16(rho_rpi,THREE_OVER_SQRT_10); - rho_rmi_3_3 = _mm_mulhi_epi16(rho_rmi,THREE_OVER_SQRT_10); - rho_rpi_3_3 = _mm_slli_epi16(rho_rpi_3_3,1); - rho_rmi_3_3 = _mm_slli_epi16(rho_rmi_3_3,1); - - xmm4 = _mm_mulhi_epi16(xmm2,ONE_OVER_SQRT_10); // Re(rho) - xmm5 = _mm_mulhi_epi16(xmm3,THREE_OVER_SQRT_10); // Im(rho) - xmm5 = _mm_slli_epi16(xmm5,1); - - rho_rpi_1_3 = _mm_adds_epi16(xmm4,xmm5); - rho_rmi_1_3 = _mm_subs_epi16(xmm4,xmm5); - - xmm6 = _mm_mulhi_epi16(xmm2,THREE_OVER_SQRT_10); // Re(rho) - xmm7 = _mm_mulhi_epi16(xmm3,ONE_OVER_SQRT_10); // Im(rho) - xmm6 = _mm_slli_epi16(xmm6,1); - - rho_rpi_3_1 = _mm_adds_epi16(xmm6,xmm7); - rho_rmi_3_1 = _mm_subs_epi16(xmm6,xmm7); - - // Rearrange interfering MF output - xmm0 = stream1_128i_in[i]; - xmm1 = stream1_128i_in[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] - y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] - - xmm0 = _mm_setzero_si128(); // ZERO - xmm2 = _mm_subs_epi16(rho_rpi_1_1,y1r); // = [Re(rho)+ Im(rho)]/sqrt(10) - y1r - psi_r_p1_p1 = _mm_abs_epi16(xmm2); // = |[Re(rho)+ Im(rho)]/sqrt(10) - y1r| - - xmm2= _mm_subs_epi16(rho_rmi_1_1,y1r); - psi_r_p1_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_1_1,y1i); - psi_i_p1_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_1_3,y1r); - psi_r_p1_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_1_3,y1r); - psi_r_p1_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_3_1,y1i); - psi_i_p1_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_3_1,y1r); - psi_r_p3_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_3_1,y1r); - psi_r_p3_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_1_3,y1i); - psi_i_p3_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_3_3,y1r); - psi_r_p3_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_3_3,y1r); - psi_r_p3_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rmi_3_3,y1i); - psi_i_p3_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_1_1,y1i); - psi_i_m1_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_3_1,y1i); - psi_i_m1_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_1_3,y1i); - psi_i_m3_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_subs_epi16(rho_rpi_3_3,y1i); - psi_i_m3_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_1_1,y1i); - psi_i_p1_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_3_1,y1i); - psi_i_p1_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_1_3,y1i); - psi_i_p3_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_3_3,y1i); - psi_i_p3_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_1_1,y1r); - psi_r_m1_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_1_3,y1r); - psi_r_m1_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_3_1,y1r); - psi_r_m3_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(rho_rpi_3_3,y1r); - psi_r_m3_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1r,rho_rmi_1_1); - psi_r_m1_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1r,rho_rmi_1_3); - psi_r_m1_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1i,rho_rmi_1_1); - psi_i_m1_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1i,rho_rmi_3_1); - psi_i_m1_m3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1r,rho_rmi_3_1); - psi_r_m3_p1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1r,rho_rmi_3_3); - psi_r_m3_p3 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1i,rho_rmi_1_3); - psi_i_m3_m1 = _mm_abs_epi16(xmm2); - xmm2= _mm_adds_epi16(y1i,rho_rmi_3_3); - psi_i_m3_m3 = _mm_abs_epi16(xmm2); - - // Rearrange desired MF output - xmm0 = stream0_128i_in[i]; - xmm1 = stream0_128i_in[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] - y0i = _mm_unpackhi_epi64(xmm0,xmm1); - - // Rearrange desired channel magnitudes - xmm2 = ch_mag_128i[i]; // = [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2)]*(2/sqrt(10)) - xmm3 = ch_mag_128i[i+1]; // = [|h|^2(3),|h|^2(3),|h|^2(4),|h|^2(4)]*(2/sqrt(10)) - xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - - ch_mag_des = _mm_unpacklo_epi64(xmm2,xmm3); // = [|h|^2(1),|h|^2(2),|h|^2(3),|h|^2(4)]*(2/sqrt(10)) - - // Rearrange interfering channel magnitudes - xmm2 = ch_mag_128i_i[i]; - xmm3 = ch_mag_128i_i[i+1]; - - xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - - ch_mag_int = _mm_unpacklo_epi64(xmm2,xmm3); - - // Scale MF output of desired signal - y0r_over_sqrt10 = _mm_mulhi_epi16(y0r,ONE_OVER_SQRT_10); - y0i_over_sqrt10 = _mm_mulhi_epi16(y0i,ONE_OVER_SQRT_10); - y0r_three_over_sqrt10 = _mm_mulhi_epi16(y0r,THREE_OVER_SQRT_10); - y0i_three_over_sqrt10 = _mm_mulhi_epi16(y0i,THREE_OVER_SQRT_10); - y0r_three_over_sqrt10 = _mm_slli_epi16(y0r_three_over_sqrt10,1); - y0i_three_over_sqrt10 = _mm_slli_epi16(y0i_three_over_sqrt10,1); - - // Compute necessary combination of required terms - y0_p_1_1 = _mm_adds_epi16(y0r_over_sqrt10,y0i_over_sqrt10); - y0_m_1_1 = _mm_subs_epi16(y0r_over_sqrt10,y0i_over_sqrt10); - - y0_p_1_3 = _mm_adds_epi16(y0r_over_sqrt10,y0i_three_over_sqrt10); - y0_m_1_3 = _mm_subs_epi16(y0r_over_sqrt10,y0i_three_over_sqrt10); - - y0_p_3_1 = _mm_adds_epi16(y0r_three_over_sqrt10,y0i_over_sqrt10); - y0_m_3_1 = _mm_subs_epi16(y0r_three_over_sqrt10,y0i_over_sqrt10); - - y0_p_3_3 = _mm_adds_epi16(y0r_three_over_sqrt10,y0i_three_over_sqrt10); - y0_m_3_3 = _mm_subs_epi16(y0r_three_over_sqrt10,y0i_three_over_sqrt10); - - // Compute optimal interfering symbol magnitude - ch_mag_int_with_sigma2 = _mm_srai_epi16(ch_mag_int, 1); // *2 - two_ch_mag_int_with_sigma2 = ch_mag_int; // *4 - three_ch_mag_int_with_sigma2 = _mm_adds_epi16(ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2); // *6 - - interference_abs_64qam_epi16(psi_r_p1_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p1_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p1_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p1_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p3_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p3_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p3_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p3_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p3_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p3_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p3_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p3_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m3_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m3_p1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_p1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m3_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m3_p3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_p3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m3_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m3_m1 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_m1,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m3_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m3_m3 ,ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_m3,ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42,FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - - // Calculation of groups of two terms in the bit metric involving product of psi and interference magnitude - prodsum_psi_a_epi16(psi_r_p1_p1,a_r_p1_p1,psi_i_p1_p1,a_i_p1_p1,psi_a_p1_p1); - prodsum_psi_a_epi16(psi_r_p1_p3,a_r_p1_p3,psi_i_p1_p3,a_i_p1_p3,psi_a_p1_p3); - prodsum_psi_a_epi16(psi_r_p3_p1,a_r_p3_p1,psi_i_p3_p1,a_i_p3_p1,psi_a_p3_p1); - prodsum_psi_a_epi16(psi_r_p3_p3,a_r_p3_p3,psi_i_p3_p3,a_i_p3_p3,psi_a_p3_p3); - prodsum_psi_a_epi16(psi_r_p1_m1,a_r_p1_m1,psi_i_p1_m1,a_i_p1_m1,psi_a_p1_m1); - prodsum_psi_a_epi16(psi_r_p1_m3,a_r_p1_m3,psi_i_p1_m3,a_i_p1_m3,psi_a_p1_m3); - prodsum_psi_a_epi16(psi_r_p3_m1,a_r_p3_m1,psi_i_p3_m1,a_i_p3_m1,psi_a_p3_m1); - prodsum_psi_a_epi16(psi_r_p3_m3,a_r_p3_m3,psi_i_p3_m3,a_i_p3_m3,psi_a_p3_m3); - prodsum_psi_a_epi16(psi_r_m1_p1,a_r_m1_p1,psi_i_m1_p1,a_i_m1_p1,psi_a_m1_p1); - prodsum_psi_a_epi16(psi_r_m1_p3,a_r_m1_p3,psi_i_m1_p3,a_i_m1_p3,psi_a_m1_p3); - prodsum_psi_a_epi16(psi_r_m3_p1,a_r_m3_p1,psi_i_m3_p1,a_i_m3_p1,psi_a_m3_p1); - prodsum_psi_a_epi16(psi_r_m3_p3,a_r_m3_p3,psi_i_m3_p3,a_i_m3_p3,psi_a_m3_p3); - prodsum_psi_a_epi16(psi_r_m1_m1,a_r_m1_m1,psi_i_m1_m1,a_i_m1_m1,psi_a_m1_m1); - prodsum_psi_a_epi16(psi_r_m1_m3,a_r_m1_m3,psi_i_m1_m3,a_i_m1_m3,psi_a_m1_m3); - prodsum_psi_a_epi16(psi_r_m3_m1,a_r_m3_m1,psi_i_m3_m1,a_i_m3_m1,psi_a_m3_m1); - prodsum_psi_a_epi16(psi_r_m3_m3,a_r_m3_m3,psi_i_m3_m3,a_i_m3_m3,psi_a_m3_m3); - - // Multiply by sqrt(2) - psi_a_p1_p1 = _mm_mulhi_epi16(psi_a_p1_p1, ONE_OVER_SQRT_2); - psi_a_p1_p1 = _mm_slli_epi16(psi_a_p1_p1, 2); - psi_a_p1_p3 = _mm_mulhi_epi16(psi_a_p1_p3, ONE_OVER_SQRT_2); - psi_a_p1_p3 = _mm_slli_epi16(psi_a_p1_p3, 2); - psi_a_p3_p1 = _mm_mulhi_epi16(psi_a_p3_p1, ONE_OVER_SQRT_2); - psi_a_p3_p1 = _mm_slli_epi16(psi_a_p3_p1, 2); - psi_a_p3_p3 = _mm_mulhi_epi16(psi_a_p3_p3, ONE_OVER_SQRT_2); - psi_a_p3_p3 = _mm_slli_epi16(psi_a_p3_p3, 2); - psi_a_p1_m1 = _mm_mulhi_epi16(psi_a_p1_m1, ONE_OVER_SQRT_2); - psi_a_p1_m1 = _mm_slli_epi16(psi_a_p1_m1, 2); - psi_a_p1_m3 = _mm_mulhi_epi16(psi_a_p1_m3, ONE_OVER_SQRT_2); - psi_a_p1_m3 = _mm_slli_epi16(psi_a_p1_m3, 2); - psi_a_p3_m1 = _mm_mulhi_epi16(psi_a_p3_m1, ONE_OVER_SQRT_2); - psi_a_p3_m1 = _mm_slli_epi16(psi_a_p3_m1, 2); - psi_a_p3_m3 = _mm_mulhi_epi16(psi_a_p3_m3, ONE_OVER_SQRT_2); - psi_a_p3_m3 = _mm_slli_epi16(psi_a_p3_m3, 2); - psi_a_m1_p1 = _mm_mulhi_epi16(psi_a_m1_p1, ONE_OVER_SQRT_2); - psi_a_m1_p1 = _mm_slli_epi16(psi_a_m1_p1, 2); - psi_a_m1_p3 = _mm_mulhi_epi16(psi_a_m1_p3, ONE_OVER_SQRT_2); - psi_a_m1_p3 = _mm_slli_epi16(psi_a_m1_p3, 2); - psi_a_m3_p1 = _mm_mulhi_epi16(psi_a_m3_p1, ONE_OVER_SQRT_2); - psi_a_m3_p1 = _mm_slli_epi16(psi_a_m3_p1, 2); - psi_a_m3_p3 = _mm_mulhi_epi16(psi_a_m3_p3, ONE_OVER_SQRT_2); - psi_a_m3_p3 = _mm_slli_epi16(psi_a_m3_p3, 2); - psi_a_m1_m1 = _mm_mulhi_epi16(psi_a_m1_m1, ONE_OVER_SQRT_2); - psi_a_m1_m1 = _mm_slli_epi16(psi_a_m1_m1, 2); - psi_a_m1_m3 = _mm_mulhi_epi16(psi_a_m1_m3, ONE_OVER_SQRT_2); - psi_a_m1_m3 = _mm_slli_epi16(psi_a_m1_m3, 2); - psi_a_m3_m1 = _mm_mulhi_epi16(psi_a_m3_m1, ONE_OVER_SQRT_2); - psi_a_m3_m1 = _mm_slli_epi16(psi_a_m3_m1, 2); - psi_a_m3_m3 = _mm_mulhi_epi16(psi_a_m3_m3, ONE_OVER_SQRT_2); - psi_a_m3_m3 = _mm_slli_epi16(psi_a_m3_m3, 2); - - // squared interference magnitude times int. ch. power - square_a_64qam_epi16(a_r_p1_p1,a_i_p1_p1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p1_p1); - square_a_64qam_epi16(a_r_p1_p3,a_i_p1_p3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p1_p3); - square_a_64qam_epi16(a_r_p3_p1,a_i_p3_p1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p3_p1); - square_a_64qam_epi16(a_r_p3_p3,a_i_p3_p3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p3_p3); - square_a_64qam_epi16(a_r_p1_m1,a_i_p1_m1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p1_m1); - square_a_64qam_epi16(a_r_p1_m3,a_i_p1_m3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p1_m3); - square_a_64qam_epi16(a_r_p3_m1,a_i_p3_m1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p3_m1); - square_a_64qam_epi16(a_r_p3_m3,a_i_p3_m3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_p3_m3); - square_a_64qam_epi16(a_r_m1_p1,a_i_m1_p1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m1_p1); - square_a_64qam_epi16(a_r_m1_p3,a_i_m1_p3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m1_p3); - square_a_64qam_epi16(a_r_m3_p1,a_i_m3_p1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m3_p1); - square_a_64qam_epi16(a_r_m3_p3,a_i_m3_p3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m3_p3); - square_a_64qam_epi16(a_r_m1_m1,a_i_m1_m1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m1_m1); - square_a_64qam_epi16(a_r_m1_m3,a_i_m1_m3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m1_m3); - square_a_64qam_epi16(a_r_m3_m1,a_i_m3_m1,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m3_m1); - square_a_64qam_epi16(a_r_m3_m3,a_i_m3_m3,ch_mag_int,SQRT_42_OVER_FOUR,a_sq_m3_m3); - - // Computing different multiples of channel norms - ch_mag_over_10=_mm_mulhi_epi16(ch_mag_des, ONE_OVER_TWO_SQRT_10); - ch_mag_over_2=_mm_mulhi_epi16(ch_mag_des, SQRT_10_OVER_FOUR); - ch_mag_over_2=_mm_slli_epi16(ch_mag_over_2, 1); - ch_mag_9_over_10=_mm_mulhi_epi16(ch_mag_des, NINE_OVER_TWO_SQRT_10); - ch_mag_9_over_10=_mm_slli_epi16(ch_mag_9_over_10, 2); - - // Computing Metrics - xmm0 = _mm_subs_epi16(psi_a_p1_p1,a_sq_p1_p1); - xmm1 = _mm_adds_epi16(xmm0,y0_p_1_1); - bit_met_p1_p1= _mm_subs_epi16(xmm1,ch_mag_over_10); - - xmm0 = _mm_subs_epi16(psi_a_p1_p3,a_sq_p1_p3); - xmm1 = _mm_adds_epi16(xmm0,y0_p_1_3); - bit_met_p1_p3= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_p1_m1,a_sq_p1_m1); - xmm1 = _mm_adds_epi16(xmm0,y0_m_1_1); - bit_met_p1_m1= _mm_subs_epi16(xmm1,ch_mag_over_10); - - xmm0 = _mm_subs_epi16(psi_a_p1_m3,a_sq_p1_m3); - xmm1 = _mm_adds_epi16(xmm0,y0_m_1_3); - bit_met_p1_m3= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_p3_p1,a_sq_p3_p1); - xmm1 = _mm_adds_epi16(xmm0,y0_p_3_1); - bit_met_p3_p1= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_p3_p3,a_sq_p3_p3); - xmm1 = _mm_adds_epi16(xmm0,y0_p_3_3); - bit_met_p3_p3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); - - xmm0 = _mm_subs_epi16(psi_a_p3_m1,a_sq_p3_m1); - xmm1 = _mm_adds_epi16(xmm0,y0_m_3_1); - bit_met_p3_m1= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_p3_m3,a_sq_p3_m3); - xmm1 = _mm_adds_epi16(xmm0,y0_m_3_3); - bit_met_p3_m3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); - - xmm0 = _mm_subs_epi16(psi_a_m1_p1,a_sq_m1_p1); - xmm1 = _mm_subs_epi16(xmm0,y0_m_1_1); - bit_met_m1_p1= _mm_subs_epi16(xmm1,ch_mag_over_10); - - xmm0 = _mm_subs_epi16(psi_a_m1_p3,a_sq_m1_p3); - xmm1 = _mm_subs_epi16(xmm0,y0_m_1_3); - bit_met_m1_p3= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_m1_m1,a_sq_m1_m1); - xmm1 = _mm_subs_epi16(xmm0,y0_p_1_1); - bit_met_m1_m1= _mm_subs_epi16(xmm1,ch_mag_over_10); - - xmm0 = _mm_subs_epi16(psi_a_m1_m3,a_sq_m1_m3); - xmm1 = _mm_subs_epi16(xmm0,y0_p_1_3); - bit_met_m1_m3= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_m3_p1,a_sq_m3_p1); - xmm1 = _mm_subs_epi16(xmm0,y0_m_3_1); - bit_met_m3_p1= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_m3_p3,a_sq_m3_p3); - xmm1 = _mm_subs_epi16(xmm0,y0_m_3_3); - bit_met_m3_p3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); - - xmm0 = _mm_subs_epi16(psi_a_m3_m1,a_sq_m3_m1); - xmm1 = _mm_subs_epi16(xmm0,y0_p_3_1); - bit_met_m3_m1= _mm_subs_epi16(xmm1,ch_mag_over_2); - - xmm0 = _mm_subs_epi16(psi_a_m3_m3,a_sq_m3_m3); - xmm1 = _mm_subs_epi16(xmm0,y0_p_3_3); - bit_met_m3_m3= _mm_subs_epi16(xmm1,ch_mag_9_over_10); - - // LLR of the first bit - // Bit = 1 - xmm0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_p3); - xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m1_m3); - xmm2 = _mm_max_epi16(bit_met_m3_p1,bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m3_m1,bit_met_m3_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_num_re0= _mm_max_epi16(xmm4,xmm5); - - // Bit = 0 - xmm0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_p3); - xmm1 = _mm_max_epi16(bit_met_p1_m1,bit_met_p1_m3); - xmm2 = _mm_max_epi16(bit_met_p3_p1,bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_p3_m1,bit_met_p3_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4,xmm5); - - // LLR of first bit [L1(1), L1(2), L1(3), L1(4), L1(5), L1(6), L1(7), L1(8)] - y0r = _mm_subs_epi16(logmax_den_re0,logmax_num_re0); - - // LLR of the second bit - // Bit = 1 - xmm0 = _mm_max_epi16(bit_met_p1_m1,bit_met_p3_m1); - xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m3_m1); - xmm2 = _mm_max_epi16(bit_met_p1_m3,bit_met_p3_m3); - xmm3 = _mm_max_epi16(bit_met_m1_m3,bit_met_m3_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_num_re1 = _mm_max_epi16(xmm4,xmm5); - - // Bit = 0 - xmm0 = _mm_max_epi16(bit_met_p1_p1,bit_met_p3_p1); - xmm1 = _mm_max_epi16(bit_met_m1_p1,bit_met_m3_p1); - xmm2 = _mm_max_epi16(bit_met_p1_p3,bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_m1_p3,bit_met_m3_p3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_den_re1 = _mm_max_epi16(xmm4,xmm5); - - // LLR of second bit [L2(1), L2(2), L2(3), L2(4)] - y1r = _mm_subs_epi16(logmax_den_re1,logmax_num_re1); - - // LLR of the third bit - // Bit = 1 - xmm0 = _mm_max_epi16(bit_met_m3_p1,bit_met_m3_p3); - xmm1 = _mm_max_epi16(bit_met_m3_m1,bit_met_m3_m3); - xmm2 = _mm_max_epi16(bit_met_p3_p1,bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_p3_m1,bit_met_p3_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_num_im0 = _mm_max_epi16(xmm4,xmm5); - - // Bit = 0 - xmm0 = _mm_max_epi16(bit_met_m1_p1,bit_met_m1_p3); - xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m1_m3); - xmm2 = _mm_max_epi16(bit_met_p1_p1,bit_met_p1_p3); - xmm3 = _mm_max_epi16(bit_met_p1_m1,bit_met_p1_m3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_den_im0 = _mm_max_epi16(xmm4,xmm5); - - // LLR of third bit [L3(1), L3(2), L3(3), L3(4)] - y0i = _mm_subs_epi16(logmax_den_im0,logmax_num_im0); - - // LLR of the fourth bit - // Bit = 1 - xmm0 = _mm_max_epi16(bit_met_p1_m3,bit_met_p3_m3); - xmm1 = _mm_max_epi16(bit_met_m1_m3,bit_met_m3_m3); - xmm2 = _mm_max_epi16(bit_met_p1_p3,bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_m1_p3,bit_met_m3_p3); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_num_im1 = _mm_max_epi16(xmm4,xmm5); - - // Bit = 0 - xmm0 = _mm_max_epi16(bit_met_p1_m1,bit_met_p3_m1); - xmm1 = _mm_max_epi16(bit_met_m1_m1,bit_met_m3_m1); - xmm2 = _mm_max_epi16(bit_met_p1_p1,bit_met_p3_p1); - xmm3 = _mm_max_epi16(bit_met_m1_p1,bit_met_m3_p1); - xmm4 = _mm_max_epi16(xmm0,xmm1); - xmm5 = _mm_max_epi16(xmm2,xmm3); - logmax_den_im1 = _mm_max_epi16(xmm4,xmm5); - - // LLR of fourth bit [L4(1), L4(2), L4(3), L4(4)] - y1i = _mm_subs_epi16(logmax_den_im1,logmax_num_im1); - - // Pack LLRs in output - // [L1(1), L2(1), L1(2), L2(2), L1(3), L2(3), L1(4), L2(4)] - xmm0 = _mm_unpacklo_epi16(y0r,y1r); - // [L1(5), L2(5), L1(6), L2(6), L1(7), L2(7), L1(8), L2(8)] - xmm1 = _mm_unpackhi_epi16(y0r,y1r); - // [L3(1), L4(1), L3(2), L4(2), L3(3), L4(3), L3(4), L4(4)] - xmm2 = _mm_unpacklo_epi16(y0i,y1i); - // [L3(5), L4(5), L3(6), L4(6), L3(7), L4(7), L3(8), L4(8)] - xmm3 = _mm_unpackhi_epi16(y0i,y1i); - - stream0_128i_out[2*i+0] = _mm_unpacklo_epi32(xmm0,xmm2); // 8LLRs, 2REs - stream0_128i_out[2*i+1] = _mm_unpackhi_epi32(xmm0,xmm2); - stream0_128i_out[2*i+2] = _mm_unpacklo_epi32(xmm1,xmm3); - stream0_128i_out[2*i+3] = _mm_unpackhi_epi32(xmm1,xmm3); -#elif defined(__arm__) || defined(__aarch64__) - -#endif - - } - -#if defined(__x86_64__) || defined(__i386__) - _mm_empty(); - _m_empty(); -#endif -} - -int nr_dlsch_16qam_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, - int32_t **rxdataF_comp, - int32_t **rxdataF_comp_i, - int32_t **dl_ch_mag, //|h_0|^2*(2/sqrt{10}) - int32_t **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) - int32_t **rho_i, - int16_t *dlsch_llr, - uint8_t symbol, - uint8_t first_symbol_flag, - uint16_t nb_rb, - uint16_t pbch_pss_sss_adjust, - int16_t **llr16p) -{ - - int16_t *rxF = (int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *rxF_i = (int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *ch_mag = (int16_t*)&dl_ch_mag[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *ch_mag_i = (int16_t*)&dl_ch_mag_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *rho = (int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *llr16; - int len; - uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; - - // first symbol has different structure due to more pilots - if (first_symbol_flag == 1) { - llr16 = (int16_t*)dlsch_llr; - } else { - llr16 = (int16_t*)(*llr16p); - } - - - AssertFatal(llr16!=NULL,"nr_dlsch_16qam_64qam_llr:llr is null, symbol %d\n",symbol); - - - if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { - // if symbol has pilots - if (frame_parms->nb_antenna_ports_gNB!=1) - // in 2 antenna ports we have 8 REs per symbol per RB - len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); - else - // for 1 antenna port we have 10 REs per symbol per RB - len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); - } else { - // symbol has no pilots - len = (nb_rb*12) - pbch_pss_sss_adjust; - } - - // printf("symbol %d: qam16_llr, len %d (llr16 %p)\n",symbol,len,llr16); - - nr_qam16_qam64((short *)rxF, - (short *)rxF_i, - (short *)ch_mag, - (short *)ch_mag_i, - (short *)llr16, - (short *)rho, - len); - - llr16 += (len<<2); - *llr16p = (short *)llr16; - - return(0); -} - -//---------------------------------------------------------------------------------------------- -// 64-QAM -//---------------------------------------------------------------------------------------------- - -/* -__m128i ONE_OVER_SQRT_42 __attribute__((aligned(16))); -__m128i THREE_OVER_SQRT_42 __attribute__((aligned(16))); -__m128i FIVE_OVER_SQRT_42 __attribute__((aligned(16))); -__m128i SEVEN_OVER_SQRT_42 __attribute__((aligned(16))); - -__m128i FORTYNINE_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); -__m128i THIRTYSEVEN_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); -__m128i TWENTYNINE_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); -__m128i TWENTYFIVE_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); -__m128i SEVENTEEN_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); -__m128i NINE_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); -__m128i THIRTEEN_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); -__m128i FIVE_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); -__m128i ONE_OVER_FOUR_SQRT_42 __attribute__((aligned(16))); - -__m128i y0r_one_over_sqrt_21 __attribute__((aligned(16))); -__m128i y0r_three_over_sqrt_21 __attribute__((aligned(16))); -__m128i y0r_five_over_sqrt_21 __attribute__((aligned(16))); -__m128i y0r_seven_over_sqrt_21 __attribute__((aligned(16))); -__m128i y0i_one_over_sqrt_21 __attribute__((aligned(16))); -__m128i y0i_three_over_sqrt_21 __attribute__((aligned(16))); -__m128i y0i_five_over_sqrt_21 __attribute__((aligned(16))); -__m128i y0i_seven_over_sqrt_21 __attribute__((aligned(16))); - -__m128i ch_mag_98_over_42_with_sigma2 __attribute__((aligned(16))); -__m128i ch_mag_74_over_42_with_sigma2 __attribute__((aligned(16))); -__m128i ch_mag_58_over_42_with_sigma2 __attribute__((aligned(16))); -__m128i ch_mag_50_over_42_with_sigma2 __attribute__((aligned(16))); -__m128i ch_mag_34_over_42_with_sigma2 __attribute__((aligned(16))); -__m128i ch_mag_18_over_42_with_sigma2 __attribute__((aligned(16))); -__m128i ch_mag_26_over_42_with_sigma2 __attribute__((aligned(16))); -__m128i ch_mag_10_over_42_with_sigma2 __attribute__((aligned(16))); -__m128i ch_mag_2_over_42_with_sigma2 __attribute__((aligned(16))); - -*/ - -void nr_qam64_qpsk(int16_t *stream0_in, - int16_t *stream1_in, - int16_t *ch_mag, - int16_t *stream0_out, - int16_t *rho01, - int32_t length - ) -{ - - /* - Author: S. Wagner - Date: 31-07-12 - - Input: - stream0_in: MF filter for 1st stream, i.e., y0=h0'*y - stream1_in: MF filter for 2nd stream, i.e., y1=h1'*y - ch_mag: 4*h0/sqrt(42), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc - ch_mag_i: 4*h1/sqrt(42), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc - rho01: Channel cross correlation, i.e., h1'*h0 - - Output: - stream0_out: output LLRs for 1st stream - */ - -#if defined(__x86_64__) || defined(__i386__) - __m128i *rho01_128i = (__m128i *)rho01; - __m128i *stream0_128i_in = (__m128i *)stream0_in; - __m128i *stream1_128i_in = (__m128i *)stream1_in; - __m128i *ch_mag_128i = (__m128i *)ch_mag; - - - __m128i ONE_OVER_SQRT_42 = _mm_set1_epi16(10112); // round(1/sqrt(42)*2^16) - __m128i THREE_OVER_SQRT_42 = _mm_set1_epi16(30337); // round(3/sqrt(42)*2^16) - __m128i FIVE_OVER_SQRT_42 = _mm_set1_epi16(25281); // round(5/sqrt(42)*2^15) - __m128i SEVEN_OVER_SQRT_42 = _mm_set1_epi16(17697); // round(5/sqrt(42)*2^15) - __m128i ONE_OVER_SQRT_2 = _mm_set1_epi16(23170); // round(1/sqrt(2)*2^15) - __m128i FORTYNINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(30969); // round(49/(4*sqrt(42))*2^14), Q2.14 - __m128i THIRTYSEVEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(23385); // round(37/(4*sqrt(42))*2^14), Q2.14 - __m128i TWENTYFIVE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(31601); // round(25/(4*sqrt(42))*2^15) - __m128i TWENTYNINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(18329); // round(29/(4*sqrt(42))*2^15), Q2.14 - __m128i SEVENTEEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(21489); // round(17/(4*sqrt(42))*2^15) - __m128i NINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(11376); // round(9/(4*sqrt(42))*2^15) - __m128i THIRTEEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(16433); // round(13/(4*sqrt(42))*2^15) - __m128i FIVE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(6320); // round(5/(4*sqrt(42))*2^15) - __m128i ONE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(1264); // round(1/(4*sqrt(42))*2^15) - - - __m128i ch_mag_des; - __m128i ch_mag_98_over_42_with_sigma2; - __m128i ch_mag_74_over_42_with_sigma2; - __m128i ch_mag_58_over_42_with_sigma2; - __m128i ch_mag_50_over_42_with_sigma2; - __m128i ch_mag_34_over_42_with_sigma2; - __m128i ch_mag_18_over_42_with_sigma2; - __m128i ch_mag_26_over_42_with_sigma2; - __m128i ch_mag_10_over_42_with_sigma2; - __m128i ch_mag_2_over_42_with_sigma2; - __m128i y0r_one_over_sqrt_21; - __m128i y0r_three_over_sqrt_21; - __m128i y0r_five_over_sqrt_21; - __m128i y0r_seven_over_sqrt_21; - __m128i y0i_one_over_sqrt_21; - __m128i y0i_three_over_sqrt_21; - __m128i y0i_five_over_sqrt_21; - __m128i y0i_seven_over_sqrt_21; -#elif defined(__arm__) || defined(__aarch64__) - -#endif - - int i,j; - - for (i=0; i>2; i+=2) { - -#if defined(__x86_64) || defined(__i386__) - // Get rho - xmm0 = rho01_128i[i]; - xmm1 = rho01_128i[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) - xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) - rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) - rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) - - // Compute the different rhos - rho_rpi_1_1 = _mm_mulhi_epi16(rho_rpi, ONE_OVER_SQRT_42); - rho_rmi_1_1 = _mm_mulhi_epi16(rho_rmi, ONE_OVER_SQRT_42); - rho_rpi_3_3 = _mm_mulhi_epi16(rho_rpi, THREE_OVER_SQRT_42); - rho_rmi_3_3 = _mm_mulhi_epi16(rho_rmi, THREE_OVER_SQRT_42); - rho_rpi_5_5 = _mm_mulhi_epi16(rho_rpi, FIVE_OVER_SQRT_42); - rho_rmi_5_5 = _mm_mulhi_epi16(rho_rmi, FIVE_OVER_SQRT_42); - rho_rpi_7_7 = _mm_mulhi_epi16(rho_rpi, SEVEN_OVER_SQRT_42); - rho_rmi_7_7 = _mm_mulhi_epi16(rho_rmi, SEVEN_OVER_SQRT_42); - - rho_rpi_5_5 = _mm_slli_epi16(rho_rpi_5_5, 1); - rho_rmi_5_5 = _mm_slli_epi16(rho_rmi_5_5, 1); - rho_rpi_7_7 = _mm_slli_epi16(rho_rpi_7_7, 2); - rho_rmi_7_7 = _mm_slli_epi16(rho_rmi_7_7, 2); - - xmm4 = _mm_mulhi_epi16(xmm2, ONE_OVER_SQRT_42); - xmm5 = _mm_mulhi_epi16(xmm3, ONE_OVER_SQRT_42); - xmm6 = _mm_mulhi_epi16(xmm3, THREE_OVER_SQRT_42); - xmm7 = _mm_mulhi_epi16(xmm3, FIVE_OVER_SQRT_42); - xmm8 = _mm_mulhi_epi16(xmm3, SEVEN_OVER_SQRT_42); - xmm7 = _mm_slli_epi16(xmm7, 1); - xmm8 = _mm_slli_epi16(xmm8, 2); - - rho_rpi_1_3 = _mm_adds_epi16(xmm4, xmm6); - rho_rmi_1_3 = _mm_subs_epi16(xmm4, xmm6); - rho_rpi_1_5 = _mm_adds_epi16(xmm4, xmm7); - rho_rmi_1_5 = _mm_subs_epi16(xmm4, xmm7); - rho_rpi_1_7 = _mm_adds_epi16(xmm4, xmm8); - rho_rmi_1_7 = _mm_subs_epi16(xmm4, xmm8); - - xmm4 = _mm_mulhi_epi16(xmm2, THREE_OVER_SQRT_42); - rho_rpi_3_1 = _mm_adds_epi16(xmm4, xmm5); - rho_rmi_3_1 = _mm_subs_epi16(xmm4, xmm5); - rho_rpi_3_5 = _mm_adds_epi16(xmm4, xmm7); - rho_rmi_3_5 = _mm_subs_epi16(xmm4, xmm7); - rho_rpi_3_7 = _mm_adds_epi16(xmm4, xmm8); - rho_rmi_3_7 = _mm_subs_epi16(xmm4, xmm8); - - xmm4 = _mm_mulhi_epi16(xmm2, FIVE_OVER_SQRT_42); - xmm4 = _mm_slli_epi16(xmm4, 1); - rho_rpi_5_1 = _mm_adds_epi16(xmm4, xmm5); - rho_rmi_5_1 = _mm_subs_epi16(xmm4, xmm5); - rho_rpi_5_3 = _mm_adds_epi16(xmm4, xmm6); - rho_rmi_5_3 = _mm_subs_epi16(xmm4, xmm6); - rho_rpi_5_7 = _mm_adds_epi16(xmm4, xmm8); - rho_rmi_5_7 = _mm_subs_epi16(xmm4, xmm8); - - xmm4 = _mm_mulhi_epi16(xmm2, SEVEN_OVER_SQRT_42); - xmm4 = _mm_slli_epi16(xmm4, 2); - rho_rpi_7_1 = _mm_adds_epi16(xmm4, xmm5); - rho_rmi_7_1 = _mm_subs_epi16(xmm4, xmm5); - rho_rpi_7_3 = _mm_adds_epi16(xmm4, xmm6); - rho_rmi_7_3 = _mm_subs_epi16(xmm4, xmm6); - rho_rpi_7_5 = _mm_adds_epi16(xmm4, xmm7); - rho_rmi_7_5 = _mm_subs_epi16(xmm4, xmm7); - - // Rearrange interfering MF output - xmm0 = stream1_128i_in[i]; - xmm1 = stream1_128i_in[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] - y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] - - // Psi_r calculation from rho_rpi or rho_rmi - xmm0 = _mm_setzero_si128(); // ZERO for abs_pi16 - xmm2 = _mm_subs_epi16(rho_rpi_7_7, y1r); - psi_r_p7_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_5, y1r); - psi_r_p7_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_3, y1r); - psi_r_p7_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_1, y1r); - psi_r_p7_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_1, y1r); - psi_r_p7_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_3, y1r); - psi_r_p7_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_5, y1r); - psi_r_p7_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_7, y1r); - psi_r_p7_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_7, y1r); - psi_r_p5_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_5, y1r); - psi_r_p5_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_3, y1r); - psi_r_p5_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_1, y1r); - psi_r_p5_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_1, y1r); - psi_r_p5_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_3, y1r); - psi_r_p5_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_5, y1r); - psi_r_p5_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_7, y1r); - psi_r_p5_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_7, y1r); - psi_r_p3_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_5, y1r); - psi_r_p3_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_3, y1r); - psi_r_p3_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_1, y1r); - psi_r_p3_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_1, y1r); - psi_r_p3_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_3, y1r); - psi_r_p3_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_5, y1r); - psi_r_p3_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_7, y1r); - psi_r_p3_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_7, y1r); - psi_r_p1_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_5, y1r); - psi_r_p1_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_3, y1r); - psi_r_p1_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_1, y1r); - psi_r_p1_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_1, y1r); - psi_r_p1_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_3, y1r); - psi_r_p1_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_5, y1r); - psi_r_p1_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_7, y1r); - psi_r_p1_m7 = _mm_abs_epi16(xmm2); - - xmm2 = _mm_adds_epi16(rho_rmi_1_7, y1r); - psi_r_m1_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_5, y1r); - psi_r_m1_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_3, y1r); - psi_r_m1_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_1, y1r); - psi_r_m1_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_1, y1r); - psi_r_m1_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_3, y1r); - psi_r_m1_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_5, y1r); - psi_r_m1_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_7, y1r); - psi_r_m1_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_7, y1r); - psi_r_m3_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_5, y1r); - psi_r_m3_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_3, y1r); - psi_r_m3_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_1, y1r); - psi_r_m3_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_1, y1r); - psi_r_m3_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_3, y1r); - psi_r_m3_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_5, y1r); - psi_r_m3_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_7, y1r); - psi_r_m3_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_7, y1r); - psi_r_m5_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_5, y1r); - psi_r_m5_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_3, y1r); - psi_r_m5_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_1, y1r); - psi_r_m5_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_1, y1r); - psi_r_m5_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_3, y1r); - psi_r_m5_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_5, y1r); - psi_r_m5_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_7, y1r); - psi_r_m5_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_7, y1r); - psi_r_m7_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_5, y1r); - psi_r_m7_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_3, y1r); - psi_r_m7_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_1, y1r); - psi_r_m7_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_1, y1r); - psi_r_m7_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_3, y1r); - psi_r_m7_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_5, y1r); - psi_r_m7_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_7, y1r); - psi_r_m7_m7 = _mm_abs_epi16(xmm2); - - // Psi_i calculation from rho_rpi or rho_rmi - xmm2 = _mm_subs_epi16(rho_rmi_7_7, y1i); - psi_i_p7_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_7, y1i); - psi_i_p7_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_7, y1i); - psi_i_p7_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_7, y1i); - psi_i_p7_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_7, y1i); - psi_i_p7_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_7, y1i); - psi_i_p7_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_7, y1i); - psi_i_p7_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_7, y1i); - psi_i_p7_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_5, y1i); - psi_i_p5_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_5, y1i); - psi_i_p5_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_5, y1i); - psi_i_p5_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_5, y1i); - psi_i_p5_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_5, y1i); - psi_i_p5_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_5, y1i); - psi_i_p5_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_5, y1i); - psi_i_p5_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_5, y1i); - psi_i_p5_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_3, y1i); - psi_i_p3_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_3, y1i); - psi_i_p3_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_3, y1i); - psi_i_p3_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_3, y1i); - psi_i_p3_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_3, y1i); - psi_i_p3_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_3, y1i); - psi_i_p3_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_3, y1i); - psi_i_p3_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_3, y1i); - psi_i_p3_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_1, y1i); - psi_i_p1_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_1, y1i); - psi_i_p1_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_1, y1i); - psi_i_p1_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_1, y1i); - psi_i_p1_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_1, y1i); - psi_i_p1_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_1, y1i); - psi_i_p1_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_1, y1i); - psi_i_p1_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_1, y1i); - psi_i_p1_m7 = _mm_abs_epi16(xmm2); - - xmm2 = _mm_subs_epi16(rho_rpi_7_1, y1i); - psi_i_m1_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_1, y1i); - psi_i_m1_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_1, y1i); - psi_i_m1_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_1, y1i); - psi_i_m1_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_1, y1i); - psi_i_m1_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_1, y1i); - psi_i_m1_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_1, y1i); - psi_i_m1_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_1, y1i); - psi_i_m1_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_3, y1i); - psi_i_m3_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_3, y1i); - psi_i_m3_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_3, y1i); - psi_i_m3_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_3, y1i); - psi_i_m3_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_3, y1i); - psi_i_m3_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_3, y1i); - psi_i_m3_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_3, y1i); - psi_i_m3_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_3, y1i); - psi_i_m3_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_5, y1i); - psi_i_m5_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_5, y1i); - psi_i_m5_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_5, y1i); - psi_i_m5_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_5, y1i); - psi_i_m5_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_5, y1i); - psi_i_m5_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_5, y1i); - psi_i_m5_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_5, y1i); - psi_i_m5_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_5, y1i); - psi_i_m5_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_7, y1i); - psi_i_m7_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_7, y1i); - psi_i_m7_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_7, y1i); - psi_i_m7_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_7, y1i); - psi_i_m7_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_7, y1i); - psi_i_m7_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_7, y1i); - psi_i_m7_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_7, y1i); - psi_i_m7_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_7, y1i); - psi_i_m7_m7 = _mm_abs_epi16(xmm2); - - - // Rearrange desired MF output - xmm0 = stream0_128i_in[i]; - xmm1 = stream0_128i_in[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] - y0i = _mm_unpackhi_epi64(xmm0,xmm1); - - // Rearrange desired channel magnitudes - xmm2 = ch_mag_128i[i]; // = [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2)]*(2/sqrt(10)) - xmm3 = ch_mag_128i[i+1]; // = [|h|^2(3),|h|^2(3),|h|^2(4),|h|^2(4)]*(2/sqrt(10)) - xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - ch_mag_des = _mm_unpacklo_epi64(xmm2,xmm3); - - y0r_one_over_sqrt_21 = _mm_mulhi_epi16(y0r, ONE_OVER_SQRT_42); - y0r_three_over_sqrt_21 = _mm_mulhi_epi16(y0r, THREE_OVER_SQRT_42); - y0r_five_over_sqrt_21 = _mm_mulhi_epi16(y0r, FIVE_OVER_SQRT_42); - y0r_five_over_sqrt_21 = _mm_slli_epi16(y0r_five_over_sqrt_21, 1); - y0r_seven_over_sqrt_21 = _mm_mulhi_epi16(y0r, SEVEN_OVER_SQRT_42); - y0r_seven_over_sqrt_21 = _mm_slli_epi16(y0r_seven_over_sqrt_21, 2); // Q2.14 - - y0i_one_over_sqrt_21 = _mm_mulhi_epi16(y0i, ONE_OVER_SQRT_42); - y0i_three_over_sqrt_21 = _mm_mulhi_epi16(y0i, THREE_OVER_SQRT_42); - y0i_five_over_sqrt_21 = _mm_mulhi_epi16(y0i, FIVE_OVER_SQRT_42); - y0i_five_over_sqrt_21 = _mm_slli_epi16(y0i_five_over_sqrt_21, 1); - y0i_seven_over_sqrt_21 = _mm_mulhi_epi16(y0i, SEVEN_OVER_SQRT_42); - y0i_seven_over_sqrt_21 = _mm_slli_epi16(y0i_seven_over_sqrt_21, 2); // Q2.14 - - y0_p_7_1 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21); - y0_p_7_3 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21); - y0_p_7_5 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21); - y0_p_7_7 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_p_5_1 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21); - y0_p_5_3 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21); - y0_p_5_5 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21); - y0_p_5_7 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_p_3_1 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21); - y0_p_3_3 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21); - y0_p_3_5 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21); - y0_p_3_7 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_p_1_1 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21); - y0_p_1_3 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21); - y0_p_1_5 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21); - y0_p_1_7 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21); - - y0_m_1_1 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21); - y0_m_1_3 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21); - y0_m_1_5 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21); - y0_m_1_7 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_m_3_1 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21); - y0_m_3_3 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21); - y0_m_3_5 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21); - y0_m_3_7 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_m_5_1 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21); - y0_m_5_3 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21); - y0_m_5_5 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21); - y0_m_5_7 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_m_7_1 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21); - y0_m_7_3 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21); - y0_m_7_5 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21); - y0_m_7_7 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21); - - // divide by sqrt(2) - psi_r_p7_p7 = _mm_mulhi_epi16(psi_r_p7_p7, ONE_OVER_SQRT_2); - psi_r_p7_p7 = _mm_slli_epi16(psi_r_p7_p7, 1); - psi_r_p7_p5 = _mm_mulhi_epi16(psi_r_p7_p5, ONE_OVER_SQRT_2); - psi_r_p7_p5 = _mm_slli_epi16(psi_r_p7_p5, 1); - psi_r_p7_p3 = _mm_mulhi_epi16(psi_r_p7_p3, ONE_OVER_SQRT_2); - psi_r_p7_p3 = _mm_slli_epi16(psi_r_p7_p3, 1); - psi_r_p7_p1 = _mm_mulhi_epi16(psi_r_p7_p1, ONE_OVER_SQRT_2); - psi_r_p7_p1 = _mm_slli_epi16(psi_r_p7_p1, 1); - psi_r_p7_m1 = _mm_mulhi_epi16(psi_r_p7_m1, ONE_OVER_SQRT_2); - psi_r_p7_m1 = _mm_slli_epi16(psi_r_p7_m1, 1); - psi_r_p7_m3 = _mm_mulhi_epi16(psi_r_p7_m3, ONE_OVER_SQRT_2); - psi_r_p7_m3 = _mm_slli_epi16(psi_r_p7_m3, 1); - psi_r_p7_m5 = _mm_mulhi_epi16(psi_r_p7_m5, ONE_OVER_SQRT_2); - psi_r_p7_m5 = _mm_slli_epi16(psi_r_p7_m5, 1); - psi_r_p7_m7 = _mm_mulhi_epi16(psi_r_p7_m7, ONE_OVER_SQRT_2); - psi_r_p7_m7 = _mm_slli_epi16(psi_r_p7_m7, 1); - psi_r_p5_p7 = _mm_mulhi_epi16(psi_r_p5_p7, ONE_OVER_SQRT_2); - psi_r_p5_p7 = _mm_slli_epi16(psi_r_p5_p7, 1); - psi_r_p5_p5 = _mm_mulhi_epi16(psi_r_p5_p5, ONE_OVER_SQRT_2); - psi_r_p5_p5 = _mm_slli_epi16(psi_r_p5_p5, 1); - psi_r_p5_p3 = _mm_mulhi_epi16(psi_r_p5_p3, ONE_OVER_SQRT_2); - psi_r_p5_p3 = _mm_slli_epi16(psi_r_p5_p3, 1); - psi_r_p5_p1 = _mm_mulhi_epi16(psi_r_p5_p1, ONE_OVER_SQRT_2); - psi_r_p5_p1 = _mm_slli_epi16(psi_r_p5_p1, 1); - psi_r_p5_m1 = _mm_mulhi_epi16(psi_r_p5_m1, ONE_OVER_SQRT_2); - psi_r_p5_m1 = _mm_slli_epi16(psi_r_p5_m1, 1); - psi_r_p5_m3 = _mm_mulhi_epi16(psi_r_p5_m3, ONE_OVER_SQRT_2); - psi_r_p5_m3 = _mm_slli_epi16(psi_r_p5_m3, 1); - psi_r_p5_m5 = _mm_mulhi_epi16(psi_r_p5_m5, ONE_OVER_SQRT_2); - psi_r_p5_m5 = _mm_slli_epi16(psi_r_p5_m5, 1); - psi_r_p5_m7 = _mm_mulhi_epi16(psi_r_p5_m7, ONE_OVER_SQRT_2); - psi_r_p5_m7 = _mm_slli_epi16(psi_r_p5_m7, 1); - psi_r_p3_p7 = _mm_mulhi_epi16(psi_r_p3_p7, ONE_OVER_SQRT_2); - psi_r_p3_p7 = _mm_slli_epi16(psi_r_p3_p7, 1); - psi_r_p3_p5 = _mm_mulhi_epi16(psi_r_p3_p5, ONE_OVER_SQRT_2); - psi_r_p3_p5 = _mm_slli_epi16(psi_r_p3_p5, 1); - psi_r_p3_p3 = _mm_mulhi_epi16(psi_r_p3_p3, ONE_OVER_SQRT_2); - psi_r_p3_p3 = _mm_slli_epi16(psi_r_p3_p3, 1); - psi_r_p3_p1 = _mm_mulhi_epi16(psi_r_p3_p1, ONE_OVER_SQRT_2); - psi_r_p3_p1 = _mm_slli_epi16(psi_r_p3_p1, 1); - psi_r_p3_m1 = _mm_mulhi_epi16(psi_r_p3_m1, ONE_OVER_SQRT_2); - psi_r_p3_m1 = _mm_slli_epi16(psi_r_p3_m1, 1); - psi_r_p3_m3 = _mm_mulhi_epi16(psi_r_p3_m3, ONE_OVER_SQRT_2); - psi_r_p3_m3 = _mm_slli_epi16(psi_r_p3_m3, 1); - psi_r_p3_m5 = _mm_mulhi_epi16(psi_r_p3_m5, ONE_OVER_SQRT_2); - psi_r_p3_m5 = _mm_slli_epi16(psi_r_p3_m5, 1); - psi_r_p3_m7 = _mm_mulhi_epi16(psi_r_p3_m7, ONE_OVER_SQRT_2); - psi_r_p3_m7 = _mm_slli_epi16(psi_r_p3_m7, 1); - psi_r_p1_p7 = _mm_mulhi_epi16(psi_r_p1_p7, ONE_OVER_SQRT_2); - psi_r_p1_p7 = _mm_slli_epi16(psi_r_p1_p7, 1); - psi_r_p1_p5 = _mm_mulhi_epi16(psi_r_p1_p5, ONE_OVER_SQRT_2); - psi_r_p1_p5 = _mm_slli_epi16(psi_r_p1_p5, 1); - psi_r_p1_p3 = _mm_mulhi_epi16(psi_r_p1_p3, ONE_OVER_SQRT_2); - psi_r_p1_p3 = _mm_slli_epi16(psi_r_p1_p3, 1); - psi_r_p1_p1 = _mm_mulhi_epi16(psi_r_p1_p1, ONE_OVER_SQRT_2); - psi_r_p1_p1 = _mm_slli_epi16(psi_r_p1_p1, 1); - psi_r_p1_m1 = _mm_mulhi_epi16(psi_r_p1_m1, ONE_OVER_SQRT_2); - psi_r_p1_m1 = _mm_slli_epi16(psi_r_p1_m1, 1); - psi_r_p1_m3 = _mm_mulhi_epi16(psi_r_p1_m3, ONE_OVER_SQRT_2); - psi_r_p1_m3 = _mm_slli_epi16(psi_r_p1_m3, 1); - psi_r_p1_m5 = _mm_mulhi_epi16(psi_r_p1_m5, ONE_OVER_SQRT_2); - psi_r_p1_m5 = _mm_slli_epi16(psi_r_p1_m5, 1); - psi_r_p1_m7 = _mm_mulhi_epi16(psi_r_p1_m7, ONE_OVER_SQRT_2); - psi_r_p1_m7 = _mm_slli_epi16(psi_r_p1_m7, 1); - psi_r_m1_p7 = _mm_mulhi_epi16(psi_r_m1_p7, ONE_OVER_SQRT_2); - psi_r_m1_p7 = _mm_slli_epi16(psi_r_m1_p7, 1); - psi_r_m1_p5 = _mm_mulhi_epi16(psi_r_m1_p5, ONE_OVER_SQRT_2); - psi_r_m1_p5 = _mm_slli_epi16(psi_r_m1_p5, 1); - psi_r_m1_p3 = _mm_mulhi_epi16(psi_r_m1_p3, ONE_OVER_SQRT_2); - psi_r_m1_p3 = _mm_slli_epi16(psi_r_m1_p3, 1); - psi_r_m1_p1 = _mm_mulhi_epi16(psi_r_m1_p1, ONE_OVER_SQRT_2); - psi_r_m1_p1 = _mm_slli_epi16(psi_r_m1_p1, 1); - psi_r_m1_m1 = _mm_mulhi_epi16(psi_r_m1_m1, ONE_OVER_SQRT_2); - psi_r_m1_m1 = _mm_slli_epi16(psi_r_m1_m1, 1); - psi_r_m1_m3 = _mm_mulhi_epi16(psi_r_m1_m3, ONE_OVER_SQRT_2); - psi_r_m1_m3 = _mm_slli_epi16(psi_r_m1_m3, 1); - psi_r_m1_m5 = _mm_mulhi_epi16(psi_r_m1_m5, ONE_OVER_SQRT_2); - psi_r_m1_m5 = _mm_slli_epi16(psi_r_m1_m5, 1); - psi_r_m1_m7 = _mm_mulhi_epi16(psi_r_m1_m7, ONE_OVER_SQRT_2); - psi_r_m1_m7 = _mm_slli_epi16(psi_r_m1_m7, 1); - psi_r_m3_p7 = _mm_mulhi_epi16(psi_r_m3_p7, ONE_OVER_SQRT_2); - psi_r_m3_p7 = _mm_slli_epi16(psi_r_m3_p7, 1); - psi_r_m3_p5 = _mm_mulhi_epi16(psi_r_m3_p5, ONE_OVER_SQRT_2); - psi_r_m3_p5 = _mm_slli_epi16(psi_r_m3_p5, 1); - psi_r_m3_p3 = _mm_mulhi_epi16(psi_r_m3_p3, ONE_OVER_SQRT_2); - psi_r_m3_p3 = _mm_slli_epi16(psi_r_m3_p3, 1); - psi_r_m3_p1 = _mm_mulhi_epi16(psi_r_m3_p1, ONE_OVER_SQRT_2); - psi_r_m3_p1 = _mm_slli_epi16(psi_r_m3_p1, 1); - psi_r_m3_m1 = _mm_mulhi_epi16(psi_r_m3_m1, ONE_OVER_SQRT_2); - psi_r_m3_m1 = _mm_slli_epi16(psi_r_m3_m1, 1); - psi_r_m3_m3 = _mm_mulhi_epi16(psi_r_m3_m3, ONE_OVER_SQRT_2); - psi_r_m3_m3 = _mm_slli_epi16(psi_r_m3_m3, 1); - psi_r_m3_m5 = _mm_mulhi_epi16(psi_r_m3_m5, ONE_OVER_SQRT_2); - psi_r_m3_m5 = _mm_slli_epi16(psi_r_m3_m5, 1); - psi_r_m3_m7 = _mm_mulhi_epi16(psi_r_m3_m7, ONE_OVER_SQRT_2); - psi_r_m3_m7 = _mm_slli_epi16(psi_r_m3_m7, 1); - psi_r_m5_p7 = _mm_mulhi_epi16(psi_r_m5_p7, ONE_OVER_SQRT_2); - psi_r_m5_p7 = _mm_slli_epi16(psi_r_m5_p7, 1); - psi_r_m5_p5 = _mm_mulhi_epi16(psi_r_m5_p5, ONE_OVER_SQRT_2); - psi_r_m5_p5 = _mm_slli_epi16(psi_r_m5_p5, 1); - psi_r_m5_p3 = _mm_mulhi_epi16(psi_r_m5_p3, ONE_OVER_SQRT_2); - psi_r_m5_p3 = _mm_slli_epi16(psi_r_m5_p3, 1); - psi_r_m5_p1 = _mm_mulhi_epi16(psi_r_m5_p1, ONE_OVER_SQRT_2); - psi_r_m5_p1 = _mm_slli_epi16(psi_r_m5_p1, 1); - psi_r_m5_m1 = _mm_mulhi_epi16(psi_r_m5_m1, ONE_OVER_SQRT_2); - psi_r_m5_m1 = _mm_slli_epi16(psi_r_m5_m1, 1); - psi_r_m5_m3 = _mm_mulhi_epi16(psi_r_m5_m3, ONE_OVER_SQRT_2); - psi_r_m5_m3 = _mm_slli_epi16(psi_r_m5_m3, 1); - psi_r_m5_m5 = _mm_mulhi_epi16(psi_r_m5_m5, ONE_OVER_SQRT_2); - psi_r_m5_m5 = _mm_slli_epi16(psi_r_m5_m5, 1); - psi_r_m5_m7 = _mm_mulhi_epi16(psi_r_m5_m7, ONE_OVER_SQRT_2); - psi_r_m5_m7 = _mm_slli_epi16(psi_r_m5_m7, 1); - psi_r_m7_p7 = _mm_mulhi_epi16(psi_r_m7_p7, ONE_OVER_SQRT_2); - psi_r_m7_p7 = _mm_slli_epi16(psi_r_m7_p7, 1); - psi_r_m7_p5 = _mm_mulhi_epi16(psi_r_m7_p5, ONE_OVER_SQRT_2); - psi_r_m7_p5 = _mm_slli_epi16(psi_r_m7_p5, 1); - psi_r_m7_p3 = _mm_mulhi_epi16(psi_r_m7_p3, ONE_OVER_SQRT_2); - psi_r_m7_p3 = _mm_slli_epi16(psi_r_m7_p3, 1); - psi_r_m7_p1 = _mm_mulhi_epi16(psi_r_m7_p1, ONE_OVER_SQRT_2); - psi_r_m7_p1 = _mm_slli_epi16(psi_r_m7_p1, 1); - psi_r_m7_m1 = _mm_mulhi_epi16(psi_r_m7_m1, ONE_OVER_SQRT_2); - psi_r_m7_m1 = _mm_slli_epi16(psi_r_m7_m1, 1); - psi_r_m7_m3 = _mm_mulhi_epi16(psi_r_m7_m3, ONE_OVER_SQRT_2); - psi_r_m7_m3 = _mm_slli_epi16(psi_r_m7_m3, 1); - psi_r_m7_m5 = _mm_mulhi_epi16(psi_r_m7_m5, ONE_OVER_SQRT_2); - psi_r_m7_m5 = _mm_slli_epi16(psi_r_m7_m5, 1); - psi_r_m7_m7 = _mm_mulhi_epi16(psi_r_m7_m7, ONE_OVER_SQRT_2); - psi_r_m7_m7 = _mm_slli_epi16(psi_r_m7_m7, 1); - - psi_i_p7_p7 = _mm_mulhi_epi16(psi_i_p7_p7, ONE_OVER_SQRT_2); - psi_i_p7_p7 = _mm_slli_epi16(psi_i_p7_p7, 1); - psi_i_p7_p5 = _mm_mulhi_epi16(psi_i_p7_p5, ONE_OVER_SQRT_2); - psi_i_p7_p5 = _mm_slli_epi16(psi_i_p7_p5, 1); - psi_i_p7_p3 = _mm_mulhi_epi16(psi_i_p7_p3, ONE_OVER_SQRT_2); - psi_i_p7_p3 = _mm_slli_epi16(psi_i_p7_p3, 1); - psi_i_p7_p1 = _mm_mulhi_epi16(psi_i_p7_p1, ONE_OVER_SQRT_2); - psi_i_p7_p1 = _mm_slli_epi16(psi_i_p7_p1, 1); - psi_i_p7_m1 = _mm_mulhi_epi16(psi_i_p7_m1, ONE_OVER_SQRT_2); - psi_i_p7_m1 = _mm_slli_epi16(psi_i_p7_m1, 1); - psi_i_p7_m3 = _mm_mulhi_epi16(psi_i_p7_m3, ONE_OVER_SQRT_2); - psi_i_p7_m3 = _mm_slli_epi16(psi_i_p7_m3, 1); - psi_i_p7_m5 = _mm_mulhi_epi16(psi_i_p7_m5, ONE_OVER_SQRT_2); - psi_i_p7_m5 = _mm_slli_epi16(psi_i_p7_m5, 1); - psi_i_p7_m7 = _mm_mulhi_epi16(psi_i_p7_m7, ONE_OVER_SQRT_2); - psi_i_p7_m7 = _mm_slli_epi16(psi_i_p7_m7, 1); - psi_i_p5_p7 = _mm_mulhi_epi16(psi_i_p5_p7, ONE_OVER_SQRT_2); - psi_i_p5_p7 = _mm_slli_epi16(psi_i_p5_p7, 1); - psi_i_p5_p5 = _mm_mulhi_epi16(psi_i_p5_p5, ONE_OVER_SQRT_2); - psi_i_p5_p5 = _mm_slli_epi16(psi_i_p5_p5, 1); - psi_i_p5_p3 = _mm_mulhi_epi16(psi_i_p5_p3, ONE_OVER_SQRT_2); - psi_i_p5_p3 = _mm_slli_epi16(psi_i_p5_p3, 1); - psi_i_p5_p1 = _mm_mulhi_epi16(psi_i_p5_p1, ONE_OVER_SQRT_2); - psi_i_p5_p1 = _mm_slli_epi16(psi_i_p5_p1, 1); - psi_i_p5_m1 = _mm_mulhi_epi16(psi_i_p5_m1, ONE_OVER_SQRT_2); - psi_i_p5_m1 = _mm_slli_epi16(psi_i_p5_m1, 1); - psi_i_p5_m3 = _mm_mulhi_epi16(psi_i_p5_m3, ONE_OVER_SQRT_2); - psi_i_p5_m3 = _mm_slli_epi16(psi_i_p5_m3, 1); - psi_i_p5_m5 = _mm_mulhi_epi16(psi_i_p5_m5, ONE_OVER_SQRT_2); - psi_i_p5_m5 = _mm_slli_epi16(psi_i_p5_m5, 1); - psi_i_p5_m7 = _mm_mulhi_epi16(psi_i_p5_m7, ONE_OVER_SQRT_2); - psi_i_p5_m7 = _mm_slli_epi16(psi_i_p5_m7, 1); - psi_i_p3_p7 = _mm_mulhi_epi16(psi_i_p3_p7, ONE_OVER_SQRT_2); - psi_i_p3_p7 = _mm_slli_epi16(psi_i_p3_p7, 1); - psi_i_p3_p5 = _mm_mulhi_epi16(psi_i_p3_p5, ONE_OVER_SQRT_2); - psi_i_p3_p5 = _mm_slli_epi16(psi_i_p3_p5, 1); - psi_i_p3_p3 = _mm_mulhi_epi16(psi_i_p3_p3, ONE_OVER_SQRT_2); - psi_i_p3_p3 = _mm_slli_epi16(psi_i_p3_p3, 1); - psi_i_p3_p1 = _mm_mulhi_epi16(psi_i_p3_p1, ONE_OVER_SQRT_2); - psi_i_p3_p1 = _mm_slli_epi16(psi_i_p3_p1, 1); - psi_i_p3_m1 = _mm_mulhi_epi16(psi_i_p3_m1, ONE_OVER_SQRT_2); - psi_i_p3_m1 = _mm_slli_epi16(psi_i_p3_m1, 1); - psi_i_p3_m3 = _mm_mulhi_epi16(psi_i_p3_m3, ONE_OVER_SQRT_2); - psi_i_p3_m3 = _mm_slli_epi16(psi_i_p3_m3, 1); - psi_i_p3_m5 = _mm_mulhi_epi16(psi_i_p3_m5, ONE_OVER_SQRT_2); - psi_i_p3_m5 = _mm_slli_epi16(psi_i_p3_m5, 1); - psi_i_p3_m7 = _mm_mulhi_epi16(psi_i_p3_m7, ONE_OVER_SQRT_2); - psi_i_p3_m7 = _mm_slli_epi16(psi_i_p3_m7, 1); - psi_i_p1_p7 = _mm_mulhi_epi16(psi_i_p1_p7, ONE_OVER_SQRT_2); - psi_i_p1_p7 = _mm_slli_epi16(psi_i_p1_p7, 1); - psi_i_p1_p5 = _mm_mulhi_epi16(psi_i_p1_p5, ONE_OVER_SQRT_2); - psi_i_p1_p5 = _mm_slli_epi16(psi_i_p1_p5, 1); - psi_i_p1_p3 = _mm_mulhi_epi16(psi_i_p1_p3, ONE_OVER_SQRT_2); - psi_i_p1_p3 = _mm_slli_epi16(psi_i_p1_p3, 1); - psi_i_p1_p1 = _mm_mulhi_epi16(psi_i_p1_p1, ONE_OVER_SQRT_2); - psi_i_p1_p1 = _mm_slli_epi16(psi_i_p1_p1, 1); - psi_i_p1_m1 = _mm_mulhi_epi16(psi_i_p1_m1, ONE_OVER_SQRT_2); - psi_i_p1_m1 = _mm_slli_epi16(psi_i_p1_m1, 1); - psi_i_p1_m3 = _mm_mulhi_epi16(psi_i_p1_m3, ONE_OVER_SQRT_2); - psi_i_p1_m3 = _mm_slli_epi16(psi_i_p1_m3, 1); - psi_i_p1_m5 = _mm_mulhi_epi16(psi_i_p1_m5, ONE_OVER_SQRT_2); - psi_i_p1_m5 = _mm_slli_epi16(psi_i_p1_m5, 1); - psi_i_p1_m7 = _mm_mulhi_epi16(psi_i_p1_m7, ONE_OVER_SQRT_2); - psi_i_p1_m7 = _mm_slli_epi16(psi_i_p1_m7, 1); - psi_i_m1_p7 = _mm_mulhi_epi16(psi_i_m1_p7, ONE_OVER_SQRT_2); - psi_i_m1_p7 = _mm_slli_epi16(psi_i_m1_p7, 1); - psi_i_m1_p5 = _mm_mulhi_epi16(psi_i_m1_p5, ONE_OVER_SQRT_2); - psi_i_m1_p5 = _mm_slli_epi16(psi_i_m1_p5, 1); - psi_i_m1_p3 = _mm_mulhi_epi16(psi_i_m1_p3, ONE_OVER_SQRT_2); - psi_i_m1_p3 = _mm_slli_epi16(psi_i_m1_p3, 1); - psi_i_m1_p1 = _mm_mulhi_epi16(psi_i_m1_p1, ONE_OVER_SQRT_2); - psi_i_m1_p1 = _mm_slli_epi16(psi_i_m1_p1, 1); - psi_i_m1_m1 = _mm_mulhi_epi16(psi_i_m1_m1, ONE_OVER_SQRT_2); - psi_i_m1_m1 = _mm_slli_epi16(psi_i_m1_m1, 1); - psi_i_m1_m3 = _mm_mulhi_epi16(psi_i_m1_m3, ONE_OVER_SQRT_2); - psi_i_m1_m3 = _mm_slli_epi16(psi_i_m1_m3, 1); - psi_i_m1_m5 = _mm_mulhi_epi16(psi_i_m1_m5, ONE_OVER_SQRT_2); - psi_i_m1_m5 = _mm_slli_epi16(psi_i_m1_m5, 1); - psi_i_m1_m7 = _mm_mulhi_epi16(psi_i_m1_m7, ONE_OVER_SQRT_2); - psi_i_m1_m7 = _mm_slli_epi16(psi_i_m1_m7, 1); - psi_i_m3_p7 = _mm_mulhi_epi16(psi_i_m3_p7, ONE_OVER_SQRT_2); - psi_i_m3_p7 = _mm_slli_epi16(psi_i_m3_p7, 1); - psi_i_m3_p5 = _mm_mulhi_epi16(psi_i_m3_p5, ONE_OVER_SQRT_2); - psi_i_m3_p5 = _mm_slli_epi16(psi_i_m3_p5, 1); - psi_i_m3_p3 = _mm_mulhi_epi16(psi_i_m3_p3, ONE_OVER_SQRT_2); - psi_i_m3_p3 = _mm_slli_epi16(psi_i_m3_p3, 1); - psi_i_m3_p1 = _mm_mulhi_epi16(psi_i_m3_p1, ONE_OVER_SQRT_2); - psi_i_m3_p1 = _mm_slli_epi16(psi_i_m3_p1, 1); - psi_i_m3_m1 = _mm_mulhi_epi16(psi_i_m3_m1, ONE_OVER_SQRT_2); - psi_i_m3_m1 = _mm_slli_epi16(psi_i_m3_m1, 1); - psi_i_m3_m3 = _mm_mulhi_epi16(psi_i_m3_m3, ONE_OVER_SQRT_2); - psi_i_m3_m3 = _mm_slli_epi16(psi_i_m3_m3, 1); - psi_i_m3_m5 = _mm_mulhi_epi16(psi_i_m3_m5, ONE_OVER_SQRT_2); - psi_i_m3_m5 = _mm_slli_epi16(psi_i_m3_m5, 1); - psi_i_m3_m7 = _mm_mulhi_epi16(psi_i_m3_m7, ONE_OVER_SQRT_2); - psi_i_m3_m7 = _mm_slli_epi16(psi_i_m3_m7, 1); - psi_i_m5_p7 = _mm_mulhi_epi16(psi_i_m5_p7, ONE_OVER_SQRT_2); - psi_i_m5_p7 = _mm_slli_epi16(psi_i_m5_p7, 1); - psi_i_m5_p5 = _mm_mulhi_epi16(psi_i_m5_p5, ONE_OVER_SQRT_2); - psi_i_m5_p5 = _mm_slli_epi16(psi_i_m5_p5, 1); - psi_i_m5_p3 = _mm_mulhi_epi16(psi_i_m5_p3, ONE_OVER_SQRT_2); - psi_i_m5_p3 = _mm_slli_epi16(psi_i_m5_p3, 1); - psi_i_m5_p1 = _mm_mulhi_epi16(psi_i_m5_p1, ONE_OVER_SQRT_2); - psi_i_m5_p1 = _mm_slli_epi16(psi_i_m5_p1, 1); - psi_i_m5_m1 = _mm_mulhi_epi16(psi_i_m5_m1, ONE_OVER_SQRT_2); - psi_i_m5_m1 = _mm_slli_epi16(psi_i_m5_m1, 1); - psi_i_m5_m3 = _mm_mulhi_epi16(psi_i_m5_m3, ONE_OVER_SQRT_2); - psi_i_m5_m3 = _mm_slli_epi16(psi_i_m5_m3, 1); - psi_i_m5_m5 = _mm_mulhi_epi16(psi_i_m5_m5, ONE_OVER_SQRT_2); - psi_i_m5_m5 = _mm_slli_epi16(psi_i_m5_m5, 1); - psi_i_m5_m7 = _mm_mulhi_epi16(psi_i_m5_m7, ONE_OVER_SQRT_2); - psi_i_m5_m7 = _mm_slli_epi16(psi_i_m5_m7, 1); - psi_i_m7_p7 = _mm_mulhi_epi16(psi_i_m7_p7, ONE_OVER_SQRT_2); - psi_i_m7_p7 = _mm_slli_epi16(psi_i_m7_p7, 1); - psi_i_m7_p5 = _mm_mulhi_epi16(psi_i_m7_p5, ONE_OVER_SQRT_2); - psi_i_m7_p5 = _mm_slli_epi16(psi_i_m7_p5, 1); - psi_i_m7_p3 = _mm_mulhi_epi16(psi_i_m7_p3, ONE_OVER_SQRT_2); - psi_i_m7_p3 = _mm_slli_epi16(psi_i_m7_p3, 1); - psi_i_m7_p1 = _mm_mulhi_epi16(psi_i_m7_p1, ONE_OVER_SQRT_2); - psi_i_m7_p1 = _mm_slli_epi16(psi_i_m7_p1, 1); - psi_i_m7_m1 = _mm_mulhi_epi16(psi_i_m7_m1, ONE_OVER_SQRT_2); - psi_i_m7_m1 = _mm_slli_epi16(psi_i_m7_m1, 1); - psi_i_m7_m3 = _mm_mulhi_epi16(psi_i_m7_m3, ONE_OVER_SQRT_2); - psi_i_m7_m3 = _mm_slli_epi16(psi_i_m7_m3, 1); - psi_i_m7_m5 = _mm_mulhi_epi16(psi_i_m7_m5, ONE_OVER_SQRT_2); - psi_i_m7_m5 = _mm_slli_epi16(psi_i_m7_m5, 1); - psi_i_m7_m7 = _mm_mulhi_epi16(psi_i_m7_m7, ONE_OVER_SQRT_2); - psi_i_m7_m7 = _mm_slli_epi16(psi_i_m7_m7, 1); - - psi_a_p7_p7 = _mm_adds_epi16(psi_r_p7_p7, psi_i_p7_p7); - psi_a_p7_p5 = _mm_adds_epi16(psi_r_p7_p5, psi_i_p7_p5); - psi_a_p7_p3 = _mm_adds_epi16(psi_r_p7_p3, psi_i_p7_p3); - psi_a_p7_p1 = _mm_adds_epi16(psi_r_p7_p1, psi_i_p7_p1); - psi_a_p7_m1 = _mm_adds_epi16(psi_r_p7_m1, psi_i_p7_m1); - psi_a_p7_m3 = _mm_adds_epi16(psi_r_p7_m3, psi_i_p7_m3); - psi_a_p7_m5 = _mm_adds_epi16(psi_r_p7_m5, psi_i_p7_m5); - psi_a_p7_m7 = _mm_adds_epi16(psi_r_p7_m7, psi_i_p7_m7); - psi_a_p5_p7 = _mm_adds_epi16(psi_r_p5_p7, psi_i_p5_p7); - psi_a_p5_p5 = _mm_adds_epi16(psi_r_p5_p5, psi_i_p5_p5); - psi_a_p5_p3 = _mm_adds_epi16(psi_r_p5_p3, psi_i_p5_p3); - psi_a_p5_p1 = _mm_adds_epi16(psi_r_p5_p1, psi_i_p5_p1); - psi_a_p5_m1 = _mm_adds_epi16(psi_r_p5_m1, psi_i_p5_m1); - psi_a_p5_m3 = _mm_adds_epi16(psi_r_p5_m3, psi_i_p5_m3); - psi_a_p5_m5 = _mm_adds_epi16(psi_r_p5_m5, psi_i_p5_m5); - psi_a_p5_m7 = _mm_adds_epi16(psi_r_p5_m7, psi_i_p5_m7); - psi_a_p3_p7 = _mm_adds_epi16(psi_r_p3_p7, psi_i_p3_p7); - psi_a_p3_p5 = _mm_adds_epi16(psi_r_p3_p5, psi_i_p3_p5); - psi_a_p3_p3 = _mm_adds_epi16(psi_r_p3_p3, psi_i_p3_p3); - psi_a_p3_p1 = _mm_adds_epi16(psi_r_p3_p1, psi_i_p3_p1); - psi_a_p3_m1 = _mm_adds_epi16(psi_r_p3_m1, psi_i_p3_m1); - psi_a_p3_m3 = _mm_adds_epi16(psi_r_p3_m3, psi_i_p3_m3); - psi_a_p3_m5 = _mm_adds_epi16(psi_r_p3_m5, psi_i_p3_m5); - psi_a_p3_m7 = _mm_adds_epi16(psi_r_p3_m7, psi_i_p3_m7); - psi_a_p1_p7 = _mm_adds_epi16(psi_r_p1_p7, psi_i_p1_p7); - psi_a_p1_p5 = _mm_adds_epi16(psi_r_p1_p5, psi_i_p1_p5); - psi_a_p1_p3 = _mm_adds_epi16(psi_r_p1_p3, psi_i_p1_p3); - psi_a_p1_p1 = _mm_adds_epi16(psi_r_p1_p1, psi_i_p1_p1); - psi_a_p1_m1 = _mm_adds_epi16(psi_r_p1_m1, psi_i_p1_m1); - psi_a_p1_m3 = _mm_adds_epi16(psi_r_p1_m3, psi_i_p1_m3); - psi_a_p1_m5 = _mm_adds_epi16(psi_r_p1_m5, psi_i_p1_m5); - psi_a_p1_m7 = _mm_adds_epi16(psi_r_p1_m7, psi_i_p1_m7); - psi_a_m1_p7 = _mm_adds_epi16(psi_r_m1_p7, psi_i_m1_p7); - psi_a_m1_p5 = _mm_adds_epi16(psi_r_m1_p5, psi_i_m1_p5); - psi_a_m1_p3 = _mm_adds_epi16(psi_r_m1_p3, psi_i_m1_p3); - psi_a_m1_p1 = _mm_adds_epi16(psi_r_m1_p1, psi_i_m1_p1); - psi_a_m1_m1 = _mm_adds_epi16(psi_r_m1_m1, psi_i_m1_m1); - psi_a_m1_m3 = _mm_adds_epi16(psi_r_m1_m3, psi_i_m1_m3); - psi_a_m1_m5 = _mm_adds_epi16(psi_r_m1_m5, psi_i_m1_m5); - psi_a_m1_m7 = _mm_adds_epi16(psi_r_m1_m7, psi_i_m1_m7); - psi_a_m3_p7 = _mm_adds_epi16(psi_r_m3_p7, psi_i_m3_p7); - psi_a_m3_p5 = _mm_adds_epi16(psi_r_m3_p5, psi_i_m3_p5); - psi_a_m3_p3 = _mm_adds_epi16(psi_r_m3_p3, psi_i_m3_p3); - psi_a_m3_p1 = _mm_adds_epi16(psi_r_m3_p1, psi_i_m3_p1); - psi_a_m3_m1 = _mm_adds_epi16(psi_r_m3_m1, psi_i_m3_m1); - psi_a_m3_m3 = _mm_adds_epi16(psi_r_m3_m3, psi_i_m3_m3); - psi_a_m3_m5 = _mm_adds_epi16(psi_r_m3_m5, psi_i_m3_m5); - psi_a_m3_m7 = _mm_adds_epi16(psi_r_m3_m7, psi_i_m3_m7); - psi_a_m5_p7 = _mm_adds_epi16(psi_r_m5_p7, psi_i_m5_p7); - psi_a_m5_p5 = _mm_adds_epi16(psi_r_m5_p5, psi_i_m5_p5); - psi_a_m5_p3 = _mm_adds_epi16(psi_r_m5_p3, psi_i_m5_p3); - psi_a_m5_p1 = _mm_adds_epi16(psi_r_m5_p1, psi_i_m5_p1); - psi_a_m5_m1 = _mm_adds_epi16(psi_r_m5_m1, psi_i_m5_m1); - psi_a_m5_m3 = _mm_adds_epi16(psi_r_m5_m3, psi_i_m5_m3); - psi_a_m5_m5 = _mm_adds_epi16(psi_r_m5_m5, psi_i_m5_m5); - psi_a_m5_m7 = _mm_adds_epi16(psi_r_m5_m7, psi_i_m5_m7); - psi_a_m7_p7 = _mm_adds_epi16(psi_r_m7_p7, psi_i_m7_p7); - psi_a_m7_p5 = _mm_adds_epi16(psi_r_m7_p5, psi_i_m7_p5); - psi_a_m7_p3 = _mm_adds_epi16(psi_r_m7_p3, psi_i_m7_p3); - psi_a_m7_p1 = _mm_adds_epi16(psi_r_m7_p1, psi_i_m7_p1); - psi_a_m7_m1 = _mm_adds_epi16(psi_r_m7_m1, psi_i_m7_m1); - psi_a_m7_m3 = _mm_adds_epi16(psi_r_m7_m3, psi_i_m7_m3); - psi_a_m7_m5 = _mm_adds_epi16(psi_r_m7_m5, psi_i_m7_m5); - psi_a_m7_m7 = _mm_adds_epi16(psi_r_m7_m7, psi_i_m7_m7); - - // Computing different multiples of ||h0||^2 - // x=1, y=1 - ch_mag_2_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,ONE_OVER_FOUR_SQRT_42); - ch_mag_2_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_2_over_42_with_sigma2,1); - // x=1, y=3 - ch_mag_10_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,FIVE_OVER_FOUR_SQRT_42); - ch_mag_10_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_10_over_42_with_sigma2,1); - // x=1, x=5 - ch_mag_26_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,THIRTEEN_OVER_FOUR_SQRT_42); - ch_mag_26_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_26_over_42_with_sigma2,1); - // x=1, y=7 - ch_mag_50_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42); - ch_mag_50_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_50_over_42_with_sigma2,1); - // x=3, y=3 - ch_mag_18_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,NINE_OVER_FOUR_SQRT_42); - ch_mag_18_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_18_over_42_with_sigma2,1); - // x=3, y=5 - ch_mag_34_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,SEVENTEEN_OVER_FOUR_SQRT_42); - ch_mag_34_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_34_over_42_with_sigma2,1); - // x=3, y=7 - ch_mag_58_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYNINE_OVER_FOUR_SQRT_42); - ch_mag_58_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_58_over_42_with_sigma2,2); - // x=5, y=5 - ch_mag_50_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42); - ch_mag_50_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_50_over_42_with_sigma2,1); - // x=5, y=7 - ch_mag_74_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,THIRTYSEVEN_OVER_FOUR_SQRT_42); - ch_mag_74_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_74_over_42_with_sigma2,2); - // x=7, y=7 - ch_mag_98_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,FORTYNINE_OVER_FOUR_SQRT_42); - ch_mag_98_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_98_over_42_with_sigma2,2); - - // Computing Metrics - xmm1 = _mm_adds_epi16(psi_a_p7_p7, y0_p_7_7); - bit_met_p7_p7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p7_p5, y0_p_7_5); - bit_met_p7_p5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p7_p3, y0_p_7_3); - bit_met_p7_p3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p7_p1, y0_p_7_1); - bit_met_p7_p1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p7_m1, y0_m_7_1); - bit_met_p7_m1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p7_m3, y0_m_7_3); - bit_met_p7_m3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p7_m5, y0_m_7_5); - bit_met_p7_m5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p7_m7, y0_m_7_7); - bit_met_p7_m7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p5_p7, y0_p_5_7); - bit_met_p5_p7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p5_p5, y0_p_5_5); - bit_met_p5_p5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p5_p3, y0_p_5_3); - bit_met_p5_p3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p5_p1, y0_p_5_1); - bit_met_p5_p1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p5_m1, y0_m_5_1); - bit_met_p5_m1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p5_m3, y0_m_5_3); - bit_met_p5_m3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p5_m5, y0_m_5_5); - bit_met_p5_m5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p5_m7, y0_m_5_7); - bit_met_p5_m7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p3_p7, y0_p_3_7); - bit_met_p3_p7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p3_p5, y0_p_3_5); - bit_met_p3_p5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p3_p3, y0_p_3_3); - bit_met_p3_p3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p3_p1, y0_p_3_1); - bit_met_p3_p1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p3_m1, y0_m_3_1); - bit_met_p3_m1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p3_m3, y0_m_3_3); - bit_met_p3_m3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p3_m5, y0_m_3_5); - bit_met_p3_m5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p3_m7, y0_m_3_7); - bit_met_p3_m7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p1_p7, y0_p_1_7); - bit_met_p1_p7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p1_p5, y0_p_1_5); - bit_met_p1_p5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p1_p3, y0_p_1_3); - bit_met_p1_p3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p1_p1, y0_p_1_1); - bit_met_p1_p1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p1_m1, y0_m_1_1); - bit_met_p1_m1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p1_m3, y0_m_1_3); - bit_met_p1_m3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p1_m5, y0_m_1_5); - bit_met_p1_m5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm1 = _mm_adds_epi16(psi_a_p1_m7, y0_m_1_7); - bit_met_p1_m7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - - xmm1 = _mm_subs_epi16(psi_a_m1_p7, y0_m_1_7); - bit_met_m1_p7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m1_p5, y0_m_1_5); - bit_met_m1_p5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m1_p3, y0_m_1_3); - bit_met_m1_p3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m1_p1, y0_m_1_1); - bit_met_m1_p1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m1_m1, y0_p_1_1); - bit_met_m1_m1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m1_m3, y0_p_1_3); - bit_met_m1_m3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m1_m5, y0_p_1_5); - bit_met_m1_m5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m1_m7, y0_p_1_7); - bit_met_m1_m7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m3_p7, y0_m_3_7); - bit_met_m3_p7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m3_p5, y0_m_3_5); - bit_met_m3_p5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m3_p3, y0_m_3_3); - bit_met_m3_p3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m3_p1, y0_m_3_1); - bit_met_m3_p1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m3_m1, y0_p_3_1); - bit_met_m3_m1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m3_m3, y0_p_3_3); - bit_met_m3_m3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m3_m5, y0_p_3_5); - bit_met_m3_m5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m3_m7, y0_p_3_7); - bit_met_m3_m7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m5_p7, y0_m_5_7); - bit_met_m5_p7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m5_p5, y0_m_5_5); - bit_met_m5_p5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m5_p3, y0_m_5_3); - bit_met_m5_p3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m5_p1, y0_m_5_1); - bit_met_m5_p1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m5_m1, y0_p_5_1); - bit_met_m5_m1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m5_m3, y0_p_5_3); - bit_met_m5_m3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m5_m5, y0_p_5_5); - bit_met_m5_m5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m5_m7, y0_p_5_7); - bit_met_m5_m7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m7_p7, y0_m_7_7); - bit_met_m7_p7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m7_p5, y0_m_7_5); - bit_met_m7_p5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m7_p3, y0_m_7_3); - bit_met_m7_p3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m7_p1, y0_m_7_1); - bit_met_m7_p1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m7_m1, y0_p_7_1); - bit_met_m7_m1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m7_m3, y0_p_7_3); - bit_met_m7_m3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m7_m5, y0_p_7_5); - bit_met_m7_m5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm1 = _mm_subs_epi16(psi_a_m7_m7, y0_p_7_7); - bit_met_m7_m7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); - - // Detection for 1st bit (LTE mapping) - // bit = 1 - xmm0 = _mm_max_epi16(bit_met_m7_p7, bit_met_m7_p5); - xmm1 = _mm_max_epi16(bit_met_m7_p3, bit_met_m7_p1); - xmm2 = _mm_max_epi16(bit_met_m7_m1, bit_met_m7_m3); - xmm3 = _mm_max_epi16(bit_met_m7_m5, bit_met_m7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m5_p7, bit_met_m5_p5); - xmm1 = _mm_max_epi16(bit_met_m5_p3, bit_met_m5_p1); - xmm2 = _mm_max_epi16(bit_met_m5_m1, bit_met_m5_m3); - xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m5_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_m3_p7, bit_met_m3_p5); - xmm1 = _mm_max_epi16(bit_met_m3_p3, bit_met_m3_p1); - xmm2 = _mm_max_epi16(bit_met_m3_m1, bit_met_m3_m3); - xmm3 = _mm_max_epi16(bit_met_m3_m5, bit_met_m3_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_m1_p7, bit_met_m1_p5); - xmm1 = _mm_max_epi16(bit_met_m1_p3, bit_met_m1_p1); - xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m1_m3); - xmm3 = _mm_max_epi16(bit_met_m1_m5, bit_met_m1_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - // bit = 0 - xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p7_p5); - xmm1 = _mm_max_epi16(bit_met_p7_p3, bit_met_p7_p1); - xmm2 = _mm_max_epi16(bit_met_p7_m1, bit_met_p7_m3); - xmm3 = _mm_max_epi16(bit_met_p7_m5, bit_met_p7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p5_p7, bit_met_p5_p5); - xmm1 = _mm_max_epi16(bit_met_p5_p3, bit_met_p5_p1); - xmm2 = _mm_max_epi16(bit_met_p5_m1, bit_met_p5_m3); - xmm3 = _mm_max_epi16(bit_met_p5_m5, bit_met_p5_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p3_p7, bit_met_p3_p5); - xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p3_p1); - xmm2 = _mm_max_epi16(bit_met_p3_m1, bit_met_p3_m3); - xmm3 = _mm_max_epi16(bit_met_p3_m5, bit_met_p3_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p1_p7, bit_met_p1_p5); - xmm1 = _mm_max_epi16(bit_met_p1_p3, bit_met_p1_p1); - xmm2 = _mm_max_epi16(bit_met_p1_m1, bit_met_p1_m3); - xmm3 = _mm_max_epi16(bit_met_p1_m5, bit_met_p1_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y0r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - // Detection for 2nd bit (LTE mapping) - // bit = 1 - xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); - xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); - xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); - xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); - xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); - xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); - xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); - xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); - xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); - xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - // bit = 0 - xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); - xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); - xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m7_p1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); - xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); - xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); - xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); - xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m7_p5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); - xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); - xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); - xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y1r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - // Detection for 3rd bit (LTE mapping) - xmm0 = _mm_max_epi16(bit_met_m7_m7, bit_met_m7_m5); - xmm1 = _mm_max_epi16(bit_met_m7_m3, bit_met_m7_m1); - xmm2 = _mm_max_epi16(bit_met_m7_p1, bit_met_m7_p3); - xmm3 = _mm_max_epi16(bit_met_m7_p5, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m5_m7, bit_met_m5_m5); - xmm1 = _mm_max_epi16(bit_met_m5_m3, bit_met_m5_m1); - xmm2 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p5_m7, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p5_m3, bit_met_p5_m1); - xmm2 = _mm_max_epi16(bit_met_p5_p1, bit_met_p5_p3); - xmm3 = _mm_max_epi16(bit_met_p5_p5, bit_met_p5_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p7_m5); - xmm1 = _mm_max_epi16(bit_met_p7_m3, bit_met_p7_m1); - xmm2 = _mm_max_epi16(bit_met_p7_p1, bit_met_p7_p3); - xmm3 = _mm_max_epi16(bit_met_p7_p5, bit_met_p7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - xmm0 = _mm_max_epi16(bit_met_m3_m7, bit_met_m3_m5); - xmm1 = _mm_max_epi16(bit_met_m3_m3, bit_met_m3_m1); - xmm2 = _mm_max_epi16(bit_met_m3_p1, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m3_p5, bit_met_m3_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m1_m7, bit_met_m1_m5); - xmm1 = _mm_max_epi16(bit_met_m1_m3, bit_met_m1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m1_p3); - xmm3 = _mm_max_epi16(bit_met_m1_p5, bit_met_m1_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p1_m7, bit_met_p1_m5); - xmm1 = _mm_max_epi16(bit_met_p1_m3, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_p1_p1, bit_met_p1_p3); - xmm3 = _mm_max_epi16(bit_met_p1_p5, bit_met_p1_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p3_m7, bit_met_p3_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p3_m1); - xmm2 = _mm_max_epi16(bit_met_p3_p1, bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_p3_p5, bit_met_p3_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y2r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - // Detection for 4th bit (LTE mapping) - xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); - xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); - xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); - xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); - xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); - xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); - xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); - xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); - xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); - xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); - xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); - xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); - xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); - xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); - xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); - xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); - xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m7_p1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); - xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); - xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y0i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - - // Detection for 5th bit (LTE mapping) - xmm0 = _mm_max_epi16(bit_met_m7_m7, bit_met_m7_m5); - xmm1 = _mm_max_epi16(bit_met_m7_m3, bit_met_m7_m1); - xmm2 = _mm_max_epi16(bit_met_m7_p1, bit_met_m7_p3); - xmm3 = _mm_max_epi16(bit_met_m7_p5, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m1_m7, bit_met_m1_m5); - xmm1 = _mm_max_epi16(bit_met_m1_m3, bit_met_m1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m1_p3); - xmm3 = _mm_max_epi16(bit_met_m1_p5, bit_met_m1_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p1_m7, bit_met_p1_m5); - xmm1 = _mm_max_epi16(bit_met_p1_m3, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_p1_p1, bit_met_p1_p3); - xmm3 = _mm_max_epi16(bit_met_p1_p5, bit_met_p1_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p7_m5); - xmm1 = _mm_max_epi16(bit_met_p7_m3, bit_met_p7_m1); - xmm2 = _mm_max_epi16(bit_met_p7_p1, bit_met_p7_p3); - xmm3 = _mm_max_epi16(bit_met_p7_p5, bit_met_p7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - xmm0 = _mm_max_epi16(bit_met_m5_m7, bit_met_m5_m5); - xmm1 = _mm_max_epi16(bit_met_m5_m3, bit_met_m5_m1); - xmm2 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m3_m7, bit_met_m3_m5); - xmm1 = _mm_max_epi16(bit_met_m3_m3, bit_met_m3_m1); - xmm2 = _mm_max_epi16(bit_met_m3_p1, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m3_p5, bit_met_m3_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p3_m7, bit_met_p3_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p3_m1); - xmm2 = _mm_max_epi16(bit_met_p3_p1, bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_p3_p5, bit_met_p3_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p5_m7, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p5_m3, bit_met_p5_m1); - xmm2 = _mm_max_epi16(bit_met_p5_p1, bit_met_p5_p3); - xmm3 = _mm_max_epi16(bit_met_p5_p5, bit_met_p5_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y1i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - // Detection for 6th bit (LTE mapping) - xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); - xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); - xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); - xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); - xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); - xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); - xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); - xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); - xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); - xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); - xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); - xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); - xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); - xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); - xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); - xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); - xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); - xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); - xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m7_p5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y2i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - - // map to output stream, difficult to do in SIMD since we have 6 16bit LLRs - // RE 1 - j = 24*i; - stream0_out[j + 0] = ((short *)&y0r)[0]; - stream0_out[j + 1] = ((short *)&y1r)[0]; - stream0_out[j + 2] = ((short *)&y2r)[0]; - stream0_out[j + 3] = ((short *)&y0i)[0]; - stream0_out[j + 4] = ((short *)&y1i)[0]; - stream0_out[j + 5] = ((short *)&y2i)[0]; - // RE 2 - stream0_out[j + 6] = ((short *)&y0r)[1]; - stream0_out[j + 7] = ((short *)&y1r)[1]; - stream0_out[j + 8] = ((short *)&y2r)[1]; - stream0_out[j + 9] = ((short *)&y0i)[1]; - stream0_out[j + 10] = ((short *)&y1i)[1]; - stream0_out[j + 11] = ((short *)&y2i)[1]; - // RE 3 - stream0_out[j + 12] = ((short *)&y0r)[2]; - stream0_out[j + 13] = ((short *)&y1r)[2]; - stream0_out[j + 14] = ((short *)&y2r)[2]; - stream0_out[j + 15] = ((short *)&y0i)[2]; - stream0_out[j + 16] = ((short *)&y1i)[2]; - stream0_out[j + 17] = ((short *)&y2i)[2]; - // RE 4 - stream0_out[j + 18] = ((short *)&y0r)[3]; - stream0_out[j + 19] = ((short *)&y1r)[3]; - stream0_out[j + 20] = ((short *)&y2r)[3]; - stream0_out[j + 21] = ((short *)&y0i)[3]; - stream0_out[j + 22] = ((short *)&y1i)[3]; - stream0_out[j + 23] = ((short *)&y2i)[3]; - // RE 5 - stream0_out[j + 24] = ((short *)&y0r)[4]; - stream0_out[j + 25] = ((short *)&y1r)[4]; - stream0_out[j + 26] = ((short *)&y2r)[4]; - stream0_out[j + 27] = ((short *)&y0i)[4]; - stream0_out[j + 28] = ((short *)&y1i)[4]; - stream0_out[j + 29] = ((short *)&y2i)[4]; - // RE 6 - stream0_out[j + 30] = ((short *)&y0r)[5]; - stream0_out[j + 31] = ((short *)&y1r)[5]; - stream0_out[j + 32] = ((short *)&y2r)[5]; - stream0_out[j + 33] = ((short *)&y0i)[5]; - stream0_out[j + 34] = ((short *)&y1i)[5]; - stream0_out[j + 35] = ((short *)&y2i)[5]; - // RE 7 - stream0_out[j + 36] = ((short *)&y0r)[6]; - stream0_out[j + 37] = ((short *)&y1r)[6]; - stream0_out[j + 38] = ((short *)&y2r)[6]; - stream0_out[j + 39] = ((short *)&y0i)[6]; - stream0_out[j + 40] = ((short *)&y1i)[6]; - stream0_out[j + 41] = ((short *)&y2i)[6]; - // RE 8 - stream0_out[j + 42] = ((short *)&y0r)[7]; - stream0_out[j + 43] = ((short *)&y1r)[7]; - stream0_out[j + 44] = ((short *)&y2r)[7]; - stream0_out[j + 45] = ((short *)&y0i)[7]; - stream0_out[j + 46] = ((short *)&y1i)[7]; - stream0_out[j + 47] = ((short *)&y2i)[7]; -#elif defined(__arm__) || defined(__aarch64__) - -#endif - } - -#if defined(__x86_64__) || defined(__i386__) - _mm_empty(); - _m_empty(); -#endif -} - - -int nr_dlsch_64qam_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, - int32_t **rxdataF_comp, - int32_t **rxdataF_comp_i, - int32_t **dl_ch_mag, - int32_t **rho_i, - int16_t *dlsch_llr, - uint8_t symbol, - uint8_t first_symbol_flag, - uint16_t nb_rb, - uint16_t pbch_pss_sss_adjust, - int16_t **llr16p) -{ - - int16_t *rxF = (int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *rxF_i = (int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *ch_mag = (int16_t*)&dl_ch_mag[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *rho = (int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *llr16; - int len; - uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; - - //first symbol has different structure due to more pilots - if (first_symbol_flag == 1) { - llr16 = (int16_t*)dlsch_llr; - } else { - llr16 = (int16_t*)(*llr16p); - } - - AssertFatal(llr16!=NULL,"nr_dlsch_16qam_64qam_llr:llr is null, symbol %d\n",symbol); - - if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { - // if symbol has pilots - if (frame_parms->nb_antenna_ports_gNB!=1) - // in 2 antenna ports we have 8 REs per symbol per RB - len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); - else - // for 1 antenna port we have 10 REs per symbol per RB - len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); - } else { - // symbol has no pilots - len = (nb_rb*12) - pbch_pss_sss_adjust; - } - - nr_qam64_qpsk((short *)rxF, - (short *)rxF_i, - (short *)ch_mag, - (short *)llr16, - (short *)rho, - len); - - llr16 += (6*len); - *llr16p = (short *)llr16; - return(0); -} - - - -void nr_qam64_qam16(short *stream0_in, - short *stream1_in, - short *ch_mag, - short *ch_mag_i, - short *stream0_out, - short *rho01, - int length - ) -{ - - /* - Author: S. Wagner - Date: 31-07-12 - - Input: - stream0_in: MF filter for 1st stream, i.e., y0=h0'*y - stream1_in: MF filter for 2nd stream, i.e., y1=h1'*y - ch_mag: 4*h0/sqrt(42), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc - ch_mag_i: 4*h1/sqrt(42), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc - rho01: Channel cross correlation, i.e., h1'*h0 - - Output: - stream0_out: output LLRs for 1st stream - */ - -#if defined(__x86_64__) || defined(__i386__) - - __m128i *rho01_128i = (__m128i *)rho01; - __m128i *stream0_128i_in = (__m128i *)stream0_in; - __m128i *stream1_128i_in = (__m128i *)stream1_in; - __m128i *ch_mag_128i = (__m128i *)ch_mag; - __m128i *ch_mag_128i_i = (__m128i *)ch_mag_i; - - __m128i ONE_OVER_SQRT_42 = _mm_set1_epi16(10112); // round(1/sqrt(42)*2^16) - __m128i THREE_OVER_SQRT_42 = _mm_set1_epi16(30337); // round(3/sqrt(42)*2^16) - __m128i FIVE_OVER_SQRT_42 = _mm_set1_epi16(25281); // round(5/sqrt(42)*2^15) - __m128i SEVEN_OVER_SQRT_42 = _mm_set1_epi16(17697); // round(5/sqrt(42)*2^15) - __m128i FORTYNINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(30969); // round(49/(4*sqrt(42))*2^14), Q2.14 - __m128i THIRTYSEVEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(23385); // round(37/(4*sqrt(42))*2^14), Q2.14 - __m128i TWENTYFIVE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(31601); // round(25/(4*sqrt(42))*2^15) - __m128i TWENTYNINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(18329); // round(29/(4*sqrt(42))*2^15), Q2.14 - __m128i SEVENTEEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(21489); // round(17/(4*sqrt(42))*2^15) - __m128i NINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(11376); // round(9/(4*sqrt(42))*2^15) - __m128i THIRTEEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(16433); // round(13/(4*sqrt(42))*2^15) - __m128i FIVE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(6320); // round(5/(4*sqrt(42))*2^15) - __m128i ONE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(1264); // round(1/(4*sqrt(42))*2^15) - __m128i ONE_OVER_SQRT_10_Q15 = _mm_set1_epi16(10362); // round(1/sqrt(10)*2^15) - __m128i THREE_OVER_SQRT_10 = _mm_set1_epi16(31086); // round(3/sqrt(10)*2^15) - __m128i SQRT_10_OVER_FOUR = _mm_set1_epi16(25905); // round(sqrt(10)/4*2^15) - - - __m128i ch_mag_int; - __m128i ch_mag_des; - __m128i ch_mag_98_over_42_with_sigma2; - __m128i ch_mag_74_over_42_with_sigma2; - __m128i ch_mag_58_over_42_with_sigma2; - __m128i ch_mag_50_over_42_with_sigma2; - __m128i ch_mag_34_over_42_with_sigma2; - __m128i ch_mag_18_over_42_with_sigma2; - __m128i ch_mag_26_over_42_with_sigma2; - __m128i ch_mag_10_over_42_with_sigma2; - __m128i ch_mag_2_over_42_with_sigma2; - __m128i y0r_one_over_sqrt_21; - __m128i y0r_three_over_sqrt_21; - __m128i y0r_five_over_sqrt_21; - __m128i y0r_seven_over_sqrt_21; - __m128i y0i_one_over_sqrt_21; - __m128i y0i_three_over_sqrt_21; - __m128i y0i_five_over_sqrt_21; - __m128i y0i_seven_over_sqrt_21; - -#elif defined(__arm__) || defined(__aarch64__) - -#endif - int i,j; - - - - for (i=0; i>2; i+=2) { - -#if defined(__x86_64__) || defined(__i386__) - // Get rho - xmm0 = rho01_128i[i]; - xmm1 = rho01_128i[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) - xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) - rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) - rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) - - // Compute the different rhos - rho_rpi_1_1 = _mm_mulhi_epi16(rho_rpi, ONE_OVER_SQRT_42); - rho_rmi_1_1 = _mm_mulhi_epi16(rho_rmi, ONE_OVER_SQRT_42); - rho_rpi_3_3 = _mm_mulhi_epi16(rho_rpi, THREE_OVER_SQRT_42); - rho_rmi_3_3 = _mm_mulhi_epi16(rho_rmi, THREE_OVER_SQRT_42); - rho_rpi_5_5 = _mm_mulhi_epi16(rho_rpi, FIVE_OVER_SQRT_42); - rho_rmi_5_5 = _mm_mulhi_epi16(rho_rmi, FIVE_OVER_SQRT_42); - rho_rpi_7_7 = _mm_mulhi_epi16(rho_rpi, SEVEN_OVER_SQRT_42); - rho_rmi_7_7 = _mm_mulhi_epi16(rho_rmi, SEVEN_OVER_SQRT_42); - - rho_rpi_5_5 = _mm_slli_epi16(rho_rpi_5_5, 1); - rho_rmi_5_5 = _mm_slli_epi16(rho_rmi_5_5, 1); - rho_rpi_7_7 = _mm_slli_epi16(rho_rpi_7_7, 2); - rho_rmi_7_7 = _mm_slli_epi16(rho_rmi_7_7, 2); - - xmm4 = _mm_mulhi_epi16(xmm2, ONE_OVER_SQRT_42); - xmm5 = _mm_mulhi_epi16(xmm3, ONE_OVER_SQRT_42); - xmm6 = _mm_mulhi_epi16(xmm3, THREE_OVER_SQRT_42); - xmm7 = _mm_mulhi_epi16(xmm3, FIVE_OVER_SQRT_42); - xmm8 = _mm_mulhi_epi16(xmm3, SEVEN_OVER_SQRT_42); - xmm7 = _mm_slli_epi16(xmm7, 1); - xmm8 = _mm_slli_epi16(xmm8, 2); - - rho_rpi_1_3 = _mm_adds_epi16(xmm4, xmm6); - rho_rmi_1_3 = _mm_subs_epi16(xmm4, xmm6); - rho_rpi_1_5 = _mm_adds_epi16(xmm4, xmm7); - rho_rmi_1_5 = _mm_subs_epi16(xmm4, xmm7); - rho_rpi_1_7 = _mm_adds_epi16(xmm4, xmm8); - rho_rmi_1_7 = _mm_subs_epi16(xmm4, xmm8); - - xmm4 = _mm_mulhi_epi16(xmm2, THREE_OVER_SQRT_42); - rho_rpi_3_1 = _mm_adds_epi16(xmm4, xmm5); - rho_rmi_3_1 = _mm_subs_epi16(xmm4, xmm5); - rho_rpi_3_5 = _mm_adds_epi16(xmm4, xmm7); - rho_rmi_3_5 = _mm_subs_epi16(xmm4, xmm7); - rho_rpi_3_7 = _mm_adds_epi16(xmm4, xmm8); - rho_rmi_3_7 = _mm_subs_epi16(xmm4, xmm8); - - xmm4 = _mm_mulhi_epi16(xmm2, FIVE_OVER_SQRT_42); - xmm4 = _mm_slli_epi16(xmm4, 1); - rho_rpi_5_1 = _mm_adds_epi16(xmm4, xmm5); - rho_rmi_5_1 = _mm_subs_epi16(xmm4, xmm5); - rho_rpi_5_3 = _mm_adds_epi16(xmm4, xmm6); - rho_rmi_5_3 = _mm_subs_epi16(xmm4, xmm6); - rho_rpi_5_7 = _mm_adds_epi16(xmm4, xmm8); - rho_rmi_5_7 = _mm_subs_epi16(xmm4, xmm8); - - xmm4 = _mm_mulhi_epi16(xmm2, SEVEN_OVER_SQRT_42); - xmm4 = _mm_slli_epi16(xmm4, 2); - rho_rpi_7_1 = _mm_adds_epi16(xmm4, xmm5); - rho_rmi_7_1 = _mm_subs_epi16(xmm4, xmm5); - rho_rpi_7_3 = _mm_adds_epi16(xmm4, xmm6); - rho_rmi_7_3 = _mm_subs_epi16(xmm4, xmm6); - rho_rpi_7_5 = _mm_adds_epi16(xmm4, xmm7); - rho_rmi_7_5 = _mm_subs_epi16(xmm4, xmm7); - - // Rearrange interfering MF output - xmm0 = stream1_128i_in[i]; - xmm1 = stream1_128i_in[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] - y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] - - // Psi_r calculation from rho_rpi or rho_rmi - xmm0 = _mm_setzero_si128(); // ZERO for abs_pi16 - xmm2 = _mm_subs_epi16(rho_rpi_7_7, y1r); - psi_r_p7_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_5, y1r); - psi_r_p7_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_3, y1r); - psi_r_p7_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_1, y1r); - psi_r_p7_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_1, y1r); - psi_r_p7_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_3, y1r); - psi_r_p7_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_5, y1r); - psi_r_p7_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_7, y1r); - psi_r_p7_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_7, y1r); - psi_r_p5_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_5, y1r); - psi_r_p5_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_3, y1r); - psi_r_p5_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_1, y1r); - psi_r_p5_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_1, y1r); - psi_r_p5_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_3, y1r); - psi_r_p5_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_5, y1r); - psi_r_p5_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_7, y1r); - psi_r_p5_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_7, y1r); - psi_r_p3_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_5, y1r); - psi_r_p3_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_3, y1r); - psi_r_p3_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_1, y1r); - psi_r_p3_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_1, y1r); - psi_r_p3_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_3, y1r); - psi_r_p3_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_5, y1r); - psi_r_p3_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_7, y1r); - psi_r_p3_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_7, y1r); - psi_r_p1_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_5, y1r); - psi_r_p1_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_3, y1r); - psi_r_p1_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_1, y1r); - psi_r_p1_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_1, y1r); - psi_r_p1_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_3, y1r); - psi_r_p1_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_5, y1r); - psi_r_p1_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_7, y1r); - psi_r_p1_m7 = _mm_abs_epi16(xmm2); - - xmm2 = _mm_adds_epi16(rho_rmi_1_7, y1r); - psi_r_m1_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_5, y1r); - psi_r_m1_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_3, y1r); - psi_r_m1_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_1, y1r); - psi_r_m1_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_1, y1r); - psi_r_m1_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_3, y1r); - psi_r_m1_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_5, y1r); - psi_r_m1_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_7, y1r); - psi_r_m1_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_7, y1r); - psi_r_m3_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_5, y1r); - psi_r_m3_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_3, y1r); - psi_r_m3_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_1, y1r); - psi_r_m3_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_1, y1r); - psi_r_m3_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_3, y1r); - psi_r_m3_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_5, y1r); - psi_r_m3_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_7, y1r); - psi_r_m3_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_7, y1r); - psi_r_m5_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_5, y1r); - psi_r_m5_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_3, y1r); - psi_r_m5_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_1, y1r); - psi_r_m5_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_1, y1r); - psi_r_m5_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_3, y1r); - psi_r_m5_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_5, y1r); - psi_r_m5_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_7, y1r); - psi_r_m5_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_7, y1r); - psi_r_m7_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_5, y1r); - psi_r_m7_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_3, y1r); - psi_r_m7_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_1, y1r); - psi_r_m7_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_1, y1r); - psi_r_m7_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_3, y1r); - psi_r_m7_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_5, y1r); - psi_r_m7_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_7, y1r); - psi_r_m7_m7 = _mm_abs_epi16(xmm2); - - // Psi_i calculation from rho_rpi or rho_rmi - xmm2 = _mm_subs_epi16(rho_rmi_7_7, y1i); - psi_i_p7_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_7, y1i); - psi_i_p7_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_7, y1i); - psi_i_p7_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_7, y1i); - psi_i_p7_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_7, y1i); - psi_i_p7_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_7, y1i); - psi_i_p7_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_7, y1i); - psi_i_p7_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_7, y1i); - psi_i_p7_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_5, y1i); - psi_i_p5_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_5, y1i); - psi_i_p5_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_5, y1i); - psi_i_p5_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_5, y1i); - psi_i_p5_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_5, y1i); - psi_i_p5_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_5, y1i); - psi_i_p5_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_5, y1i); - psi_i_p5_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_5, y1i); - psi_i_p5_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_3, y1i); - psi_i_p3_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_3, y1i); - psi_i_p3_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_3, y1i); - psi_i_p3_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_3, y1i); - psi_i_p3_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_3, y1i); - psi_i_p3_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_3, y1i); - psi_i_p3_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_3, y1i); - psi_i_p3_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_3, y1i); - psi_i_p3_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_1, y1i); - psi_i_p1_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_1, y1i); - psi_i_p1_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_1, y1i); - psi_i_p1_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_1, y1i); - psi_i_p1_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_1, y1i); - psi_i_p1_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_1, y1i); - psi_i_p1_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_1, y1i); - psi_i_p1_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_1, y1i); - psi_i_p1_m7 = _mm_abs_epi16(xmm2); - - xmm2 = _mm_subs_epi16(rho_rpi_7_1, y1i); - psi_i_m1_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_1, y1i); - psi_i_m1_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_1, y1i); - psi_i_m1_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_1, y1i); - psi_i_m1_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_1, y1i); - psi_i_m1_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_1, y1i); - psi_i_m1_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_1, y1i); - psi_i_m1_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_1, y1i); - psi_i_m1_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_3, y1i); - psi_i_m3_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_3, y1i); - psi_i_m3_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_3, y1i); - psi_i_m3_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_3, y1i); - psi_i_m3_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_3, y1i); - psi_i_m3_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_3, y1i); - psi_i_m3_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_3, y1i); - psi_i_m3_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_3, y1i); - psi_i_m3_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_5, y1i); - psi_i_m5_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_5, y1i); - psi_i_m5_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_5, y1i); - psi_i_m5_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_5, y1i); - psi_i_m5_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_5, y1i); - psi_i_m5_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_5, y1i); - psi_i_m5_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_5, y1i); - psi_i_m5_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_5, y1i); - psi_i_m5_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_7, y1i); - psi_i_m7_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_7, y1i); - psi_i_m7_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_7, y1i); - psi_i_m7_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_7, y1i); - psi_i_m7_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_7, y1i); - psi_i_m7_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_7, y1i); - psi_i_m7_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_7, y1i); - psi_i_m7_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_7, y1i); - psi_i_m7_m7 = _mm_abs_epi16(xmm2); - - - // Rearrange desired MF output - xmm0 = stream0_128i_in[i]; - xmm1 = stream0_128i_in[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] - y0i = _mm_unpackhi_epi64(xmm0,xmm1); - - // Rearrange desired channel magnitudes - xmm2 = ch_mag_128i[i]; // = [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2)]*(2/sqrt(10)) - xmm3 = ch_mag_128i[i+1]; // = [|h|^2(3),|h|^2(3),|h|^2(4),|h|^2(4)]*(2/sqrt(10)) - xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - ch_mag_des = _mm_unpacklo_epi64(xmm2,xmm3); - - // Rearrange interfering channel magnitudes - xmm2 = ch_mag_128i_i[i]; - xmm3 = ch_mag_128i_i[i+1]; - xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - ch_mag_int = _mm_unpacklo_epi64(xmm2,xmm3); - - y0r_one_over_sqrt_21 = _mm_mulhi_epi16(y0r, ONE_OVER_SQRT_42); - y0r_three_over_sqrt_21 = _mm_mulhi_epi16(y0r, THREE_OVER_SQRT_42); - y0r_five_over_sqrt_21 = _mm_mulhi_epi16(y0r, FIVE_OVER_SQRT_42); - y0r_five_over_sqrt_21 = _mm_slli_epi16(y0r_five_over_sqrt_21, 1); - y0r_seven_over_sqrt_21 = _mm_mulhi_epi16(y0r, SEVEN_OVER_SQRT_42); - y0r_seven_over_sqrt_21 = _mm_slli_epi16(y0r_seven_over_sqrt_21, 2); // Q2.14 - - y0i_one_over_sqrt_21 = _mm_mulhi_epi16(y0i, ONE_OVER_SQRT_42); - y0i_three_over_sqrt_21 = _mm_mulhi_epi16(y0i, THREE_OVER_SQRT_42); - y0i_five_over_sqrt_21 = _mm_mulhi_epi16(y0i, FIVE_OVER_SQRT_42); - y0i_five_over_sqrt_21 = _mm_slli_epi16(y0i_five_over_sqrt_21, 1); - y0i_seven_over_sqrt_21 = _mm_mulhi_epi16(y0i, SEVEN_OVER_SQRT_42); - y0i_seven_over_sqrt_21 = _mm_slli_epi16(y0i_seven_over_sqrt_21, 2); // Q2.14 - - y0_p_7_1 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21); - y0_p_7_3 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21); - y0_p_7_5 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21); - y0_p_7_7 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_p_5_1 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21); - y0_p_5_3 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21); - y0_p_5_5 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21); - y0_p_5_7 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_p_3_1 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21); - y0_p_3_3 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21); - y0_p_3_5 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21); - y0_p_3_7 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_p_1_1 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21); - y0_p_1_3 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21); - y0_p_1_5 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21); - y0_p_1_7 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21); - - y0_m_1_1 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21); - y0_m_1_3 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21); - y0_m_1_5 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21); - y0_m_1_7 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_m_3_1 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21); - y0_m_3_3 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21); - y0_m_3_5 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21); - y0_m_3_7 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_m_5_1 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21); - y0_m_5_3 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21); - y0_m_5_5 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21); - y0_m_5_7 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_m_7_1 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21); - y0_m_7_3 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21); - y0_m_7_5 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21); - y0_m_7_7 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21); - - interference_abs_epi16(psi_r_p7_p7, ch_mag_int, a_r_p7_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p7_p5, ch_mag_int, a_r_p7_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p7_p3, ch_mag_int, a_r_p7_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p7_p1, ch_mag_int, a_r_p7_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p7_m1, ch_mag_int, a_r_p7_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p7_m3, ch_mag_int, a_r_p7_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p7_m5, ch_mag_int, a_r_p7_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p7_m7, ch_mag_int, a_r_p7_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p5_p7, ch_mag_int, a_r_p5_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p5_p5, ch_mag_int, a_r_p5_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p5_p3, ch_mag_int, a_r_p5_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p5_p1, ch_mag_int, a_r_p5_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p5_m1, ch_mag_int, a_r_p5_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p5_m3, ch_mag_int, a_r_p5_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p5_m5, ch_mag_int, a_r_p5_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p5_m7, ch_mag_int, a_r_p5_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p3_p7, ch_mag_int, a_r_p3_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p3_p5, ch_mag_int, a_r_p3_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p3_p3, ch_mag_int, a_r_p3_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p3_p1, ch_mag_int, a_r_p3_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p3_m1, ch_mag_int, a_r_p3_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p3_m3, ch_mag_int, a_r_p3_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p3_m5, ch_mag_int, a_r_p3_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p3_m7, ch_mag_int, a_r_p3_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p1_p7, ch_mag_int, a_r_p1_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p1_p5, ch_mag_int, a_r_p1_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p1_p3, ch_mag_int, a_r_p1_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p1_p1, ch_mag_int, a_r_p1_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p1_m1, ch_mag_int, a_r_p1_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p1_m3, ch_mag_int, a_r_p1_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p1_m5, ch_mag_int, a_r_p1_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_p1_m7, ch_mag_int, a_r_p1_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_p7, ch_mag_int, a_r_m1_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_p5, ch_mag_int, a_r_m1_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_p3, ch_mag_int, a_r_m1_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_p1, ch_mag_int, a_r_m1_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_m1, ch_mag_int, a_r_m1_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_m3, ch_mag_int, a_r_m1_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_m5, ch_mag_int, a_r_m1_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m1_m7, ch_mag_int, a_r_m1_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m3_p7, ch_mag_int, a_r_m3_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m3_p5, ch_mag_int, a_r_m3_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m3_p3, ch_mag_int, a_r_m3_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m3_p1, ch_mag_int, a_r_m3_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m3_m1, ch_mag_int, a_r_m3_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m3_m3, ch_mag_int, a_r_m3_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m3_m5, ch_mag_int, a_r_m3_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m3_m7, ch_mag_int, a_r_m3_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m5_p7, ch_mag_int, a_r_m5_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m5_p5, ch_mag_int, a_r_m5_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m5_p3, ch_mag_int, a_r_m5_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m5_p1, ch_mag_int, a_r_m5_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m5_m1, ch_mag_int, a_r_m5_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m5_m3, ch_mag_int, a_r_m5_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m5_m5, ch_mag_int, a_r_m5_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m5_m7, ch_mag_int, a_r_m5_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m7_p7, ch_mag_int, a_r_m7_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m7_p5, ch_mag_int, a_r_m7_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m7_p3, ch_mag_int, a_r_m7_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m7_p1, ch_mag_int, a_r_m7_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m7_m1, ch_mag_int, a_r_m7_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m7_m3, ch_mag_int, a_r_m7_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m7_m5, ch_mag_int, a_r_m7_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_r_m7_m7, ch_mag_int, a_r_m7_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - - interference_abs_epi16(psi_i_p7_p7, ch_mag_int, a_i_p7_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p7_p5, ch_mag_int, a_i_p7_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p7_p3, ch_mag_int, a_i_p7_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p7_p1, ch_mag_int, a_i_p7_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p7_m1, ch_mag_int, a_i_p7_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p7_m3, ch_mag_int, a_i_p7_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p7_m5, ch_mag_int, a_i_p7_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p7_m7, ch_mag_int, a_i_p7_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p5_p7, ch_mag_int, a_i_p5_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p5_p5, ch_mag_int, a_i_p5_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p5_p3, ch_mag_int, a_i_p5_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p5_p1, ch_mag_int, a_i_p5_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p5_m1, ch_mag_int, a_i_p5_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p5_m3, ch_mag_int, a_i_p5_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p5_m5, ch_mag_int, a_i_p5_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p5_m7, ch_mag_int, a_i_p5_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p3_p7, ch_mag_int, a_i_p3_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p3_p5, ch_mag_int, a_i_p3_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p3_p3, ch_mag_int, a_i_p3_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p3_p1, ch_mag_int, a_i_p3_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p3_m1, ch_mag_int, a_i_p3_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p3_m3, ch_mag_int, a_i_p3_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p3_m5, ch_mag_int, a_i_p3_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p3_m7, ch_mag_int, a_i_p3_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_p7, ch_mag_int, a_i_p1_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_p5, ch_mag_int, a_i_p1_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_p3, ch_mag_int, a_i_p1_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_p1, ch_mag_int, a_i_p1_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_m1, ch_mag_int, a_i_p1_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_m3, ch_mag_int, a_i_p1_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_m5, ch_mag_int, a_i_p1_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_p1_m7, ch_mag_int, a_i_p1_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_p7, ch_mag_int, a_i_m1_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_p5, ch_mag_int, a_i_m1_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_p3, ch_mag_int, a_i_m1_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_p1, ch_mag_int, a_i_m1_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_m1, ch_mag_int, a_i_m1_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_m3, ch_mag_int, a_i_m1_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_m5, ch_mag_int, a_i_m1_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m1_m7, ch_mag_int, a_i_m1_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m3_p7, ch_mag_int, a_i_m3_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m3_p5, ch_mag_int, a_i_m3_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m3_p3, ch_mag_int, a_i_m3_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m3_p1, ch_mag_int, a_i_m3_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m3_m1, ch_mag_int, a_i_m3_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m3_m3, ch_mag_int, a_i_m3_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m3_m5, ch_mag_int, a_i_m3_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m3_m7, ch_mag_int, a_i_m3_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m5_p7, ch_mag_int, a_i_m5_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m5_p5, ch_mag_int, a_i_m5_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m5_p3, ch_mag_int, a_i_m5_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m5_p1, ch_mag_int, a_i_m5_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m5_m1, ch_mag_int, a_i_m5_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m5_m3, ch_mag_int, a_i_m5_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m5_m5, ch_mag_int, a_i_m5_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m5_m7, ch_mag_int, a_i_m5_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m7_p7, ch_mag_int, a_i_m7_p7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m7_p5, ch_mag_int, a_i_m7_p5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m7_p3, ch_mag_int, a_i_m7_p3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m7_p1, ch_mag_int, a_i_m7_p1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m7_m1, ch_mag_int, a_i_m7_m1, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m7_m3, ch_mag_int, a_i_m7_m3, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m7_m5, ch_mag_int, a_i_m7_m5, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - interference_abs_epi16(psi_i_m7_m7, ch_mag_int, a_i_m7_m7, ONE_OVER_SQRT_10_Q15, THREE_OVER_SQRT_10); - - // Calculation of a group of two terms in the bit metric involving product of psi and interference - prodsum_psi_a_epi16(psi_r_p7_p7, a_r_p7_p7, psi_i_p7_p7, a_i_p7_p7, psi_a_p7_p7); - prodsum_psi_a_epi16(psi_r_p7_p5, a_r_p7_p5, psi_i_p7_p5, a_i_p7_p5, psi_a_p7_p5); - prodsum_psi_a_epi16(psi_r_p7_p3, a_r_p7_p3, psi_i_p7_p3, a_i_p7_p3, psi_a_p7_p3); - prodsum_psi_a_epi16(psi_r_p7_p1, a_r_p7_p1, psi_i_p7_p1, a_i_p7_p1, psi_a_p7_p1); - prodsum_psi_a_epi16(psi_r_p7_m1, a_r_p7_m1, psi_i_p7_m1, a_i_p7_m1, psi_a_p7_m1); - prodsum_psi_a_epi16(psi_r_p7_m3, a_r_p7_m3, psi_i_p7_m3, a_i_p7_m3, psi_a_p7_m3); - prodsum_psi_a_epi16(psi_r_p7_m5, a_r_p7_m5, psi_i_p7_m5, a_i_p7_m5, psi_a_p7_m5); - prodsum_psi_a_epi16(psi_r_p7_m7, a_r_p7_m7, psi_i_p7_m7, a_i_p7_m7, psi_a_p7_m7); - prodsum_psi_a_epi16(psi_r_p5_p7, a_r_p5_p7, psi_i_p5_p7, a_i_p5_p7, psi_a_p5_p7); - prodsum_psi_a_epi16(psi_r_p5_p5, a_r_p5_p5, psi_i_p5_p5, a_i_p5_p5, psi_a_p5_p5); - prodsum_psi_a_epi16(psi_r_p5_p3, a_r_p5_p3, psi_i_p5_p3, a_i_p5_p3, psi_a_p5_p3); - prodsum_psi_a_epi16(psi_r_p5_p1, a_r_p5_p1, psi_i_p5_p1, a_i_p5_p1, psi_a_p5_p1); - prodsum_psi_a_epi16(psi_r_p5_m1, a_r_p5_m1, psi_i_p5_m1, a_i_p5_m1, psi_a_p5_m1); - prodsum_psi_a_epi16(psi_r_p5_m3, a_r_p5_m3, psi_i_p5_m3, a_i_p5_m3, psi_a_p5_m3); - prodsum_psi_a_epi16(psi_r_p5_m5, a_r_p5_m5, psi_i_p5_m5, a_i_p5_m5, psi_a_p5_m5); - prodsum_psi_a_epi16(psi_r_p5_m7, a_r_p5_m7, psi_i_p5_m7, a_i_p5_m7, psi_a_p5_m7); - prodsum_psi_a_epi16(psi_r_p3_p7, a_r_p3_p7, psi_i_p3_p7, a_i_p3_p7, psi_a_p3_p7); - prodsum_psi_a_epi16(psi_r_p3_p5, a_r_p3_p5, psi_i_p3_p5, a_i_p3_p5, psi_a_p3_p5); - prodsum_psi_a_epi16(psi_r_p3_p3, a_r_p3_p3, psi_i_p3_p3, a_i_p3_p3, psi_a_p3_p3); - prodsum_psi_a_epi16(psi_r_p3_p1, a_r_p3_p1, psi_i_p3_p1, a_i_p3_p1, psi_a_p3_p1); - prodsum_psi_a_epi16(psi_r_p3_m1, a_r_p3_m1, psi_i_p3_m1, a_i_p3_m1, psi_a_p3_m1); - prodsum_psi_a_epi16(psi_r_p3_m3, a_r_p3_m3, psi_i_p3_m3, a_i_p3_m3, psi_a_p3_m3); - prodsum_psi_a_epi16(psi_r_p3_m5, a_r_p3_m5, psi_i_p3_m5, a_i_p3_m5, psi_a_p3_m5); - prodsum_psi_a_epi16(psi_r_p3_m7, a_r_p3_m7, psi_i_p3_m7, a_i_p3_m7, psi_a_p3_m7); - prodsum_psi_a_epi16(psi_r_p1_p7, a_r_p1_p7, psi_i_p1_p7, a_i_p1_p7, psi_a_p1_p7); - prodsum_psi_a_epi16(psi_r_p1_p5, a_r_p1_p5, psi_i_p1_p5, a_i_p1_p5, psi_a_p1_p5); - prodsum_psi_a_epi16(psi_r_p1_p3, a_r_p1_p3, psi_i_p1_p3, a_i_p1_p3, psi_a_p1_p3); - prodsum_psi_a_epi16(psi_r_p1_p1, a_r_p1_p1, psi_i_p1_p1, a_i_p1_p1, psi_a_p1_p1); - prodsum_psi_a_epi16(psi_r_p1_m1, a_r_p1_m1, psi_i_p1_m1, a_i_p1_m1, psi_a_p1_m1); - prodsum_psi_a_epi16(psi_r_p1_m3, a_r_p1_m3, psi_i_p1_m3, a_i_p1_m3, psi_a_p1_m3); - prodsum_psi_a_epi16(psi_r_p1_m5, a_r_p1_m5, psi_i_p1_m5, a_i_p1_m5, psi_a_p1_m5); - prodsum_psi_a_epi16(psi_r_p1_m7, a_r_p1_m7, psi_i_p1_m7, a_i_p1_m7, psi_a_p1_m7); - prodsum_psi_a_epi16(psi_r_m1_p7, a_r_m1_p7, psi_i_m1_p7, a_i_m1_p7, psi_a_m1_p7); - prodsum_psi_a_epi16(psi_r_m1_p5, a_r_m1_p5, psi_i_m1_p5, a_i_m1_p5, psi_a_m1_p5); - prodsum_psi_a_epi16(psi_r_m1_p3, a_r_m1_p3, psi_i_m1_p3, a_i_m1_p3, psi_a_m1_p3); - prodsum_psi_a_epi16(psi_r_m1_p1, a_r_m1_p1, psi_i_m1_p1, a_i_m1_p1, psi_a_m1_p1); - prodsum_psi_a_epi16(psi_r_m1_m1, a_r_m1_m1, psi_i_m1_m1, a_i_m1_m1, psi_a_m1_m1); - prodsum_psi_a_epi16(psi_r_m1_m3, a_r_m1_m3, psi_i_m1_m3, a_i_m1_m3, psi_a_m1_m3); - prodsum_psi_a_epi16(psi_r_m1_m5, a_r_m1_m5, psi_i_m1_m5, a_i_m1_m5, psi_a_m1_m5); - prodsum_psi_a_epi16(psi_r_m1_m7, a_r_m1_m7, psi_i_m1_m7, a_i_m1_m7, psi_a_m1_m7); - prodsum_psi_a_epi16(psi_r_m3_p7, a_r_m3_p7, psi_i_m3_p7, a_i_m3_p7, psi_a_m3_p7); - prodsum_psi_a_epi16(psi_r_m3_p5, a_r_m3_p5, psi_i_m3_p5, a_i_m3_p5, psi_a_m3_p5); - prodsum_psi_a_epi16(psi_r_m3_p3, a_r_m3_p3, psi_i_m3_p3, a_i_m3_p3, psi_a_m3_p3); - prodsum_psi_a_epi16(psi_r_m3_p1, a_r_m3_p1, psi_i_m3_p1, a_i_m3_p1, psi_a_m3_p1); - prodsum_psi_a_epi16(psi_r_m3_m1, a_r_m3_m1, psi_i_m3_m1, a_i_m3_m1, psi_a_m3_m1); - prodsum_psi_a_epi16(psi_r_m3_m3, a_r_m3_m3, psi_i_m3_m3, a_i_m3_m3, psi_a_m3_m3); - prodsum_psi_a_epi16(psi_r_m3_m5, a_r_m3_m5, psi_i_m3_m5, a_i_m3_m5, psi_a_m3_m5); - prodsum_psi_a_epi16(psi_r_m3_m7, a_r_m3_m7, psi_i_m3_m7, a_i_m3_m7, psi_a_m3_m7); - prodsum_psi_a_epi16(psi_r_m5_p7, a_r_m5_p7, psi_i_m5_p7, a_i_m5_p7, psi_a_m5_p7); - prodsum_psi_a_epi16(psi_r_m5_p5, a_r_m5_p5, psi_i_m5_p5, a_i_m5_p5, psi_a_m5_p5); - prodsum_psi_a_epi16(psi_r_m5_p3, a_r_m5_p3, psi_i_m5_p3, a_i_m5_p3, psi_a_m5_p3); - prodsum_psi_a_epi16(psi_r_m5_p1, a_r_m5_p1, psi_i_m5_p1, a_i_m5_p1, psi_a_m5_p1); - prodsum_psi_a_epi16(psi_r_m5_m1, a_r_m5_m1, psi_i_m5_m1, a_i_m5_m1, psi_a_m5_m1); - prodsum_psi_a_epi16(psi_r_m5_m3, a_r_m5_m3, psi_i_m5_m3, a_i_m5_m3, psi_a_m5_m3); - prodsum_psi_a_epi16(psi_r_m5_m5, a_r_m5_m5, psi_i_m5_m5, a_i_m5_m5, psi_a_m5_m5); - prodsum_psi_a_epi16(psi_r_m5_m7, a_r_m5_m7, psi_i_m5_m7, a_i_m5_m7, psi_a_m5_m7); - prodsum_psi_a_epi16(psi_r_m7_p7, a_r_m7_p7, psi_i_m7_p7, a_i_m7_p7, psi_a_m7_p7); - prodsum_psi_a_epi16(psi_r_m7_p5, a_r_m7_p5, psi_i_m7_p5, a_i_m7_p5, psi_a_m7_p5); - prodsum_psi_a_epi16(psi_r_m7_p3, a_r_m7_p3, psi_i_m7_p3, a_i_m7_p3, psi_a_m7_p3); - prodsum_psi_a_epi16(psi_r_m7_p1, a_r_m7_p1, psi_i_m7_p1, a_i_m7_p1, psi_a_m7_p1); - prodsum_psi_a_epi16(psi_r_m7_m1, a_r_m7_m1, psi_i_m7_m1, a_i_m7_m1, psi_a_m7_m1); - prodsum_psi_a_epi16(psi_r_m7_m3, a_r_m7_m3, psi_i_m7_m3, a_i_m7_m3, psi_a_m7_m3); - prodsum_psi_a_epi16(psi_r_m7_m5, a_r_m7_m5, psi_i_m7_m5, a_i_m7_m5, psi_a_m7_m5); - prodsum_psi_a_epi16(psi_r_m7_m7, a_r_m7_m7, psi_i_m7_m7, a_i_m7_m7, psi_a_m7_m7); - - // Calculation of a group of two terms in the bit metric involving squares of interference - square_a_epi16(a_r_p7_p7, a_i_p7_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_p7); - square_a_epi16(a_r_p7_p5, a_i_p7_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_p5); - square_a_epi16(a_r_p7_p3, a_i_p7_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_p3); - square_a_epi16(a_r_p7_p1, a_i_p7_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_p1); - square_a_epi16(a_r_p7_m1, a_i_p7_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_m1); - square_a_epi16(a_r_p7_m3, a_i_p7_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_m3); - square_a_epi16(a_r_p7_m5, a_i_p7_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_m5); - square_a_epi16(a_r_p7_m7, a_i_p7_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p7_m7); - square_a_epi16(a_r_p5_p7, a_i_p5_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_p7); - square_a_epi16(a_r_p5_p5, a_i_p5_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_p5); - square_a_epi16(a_r_p5_p3, a_i_p5_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_p3); - square_a_epi16(a_r_p5_p1, a_i_p5_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_p1); - square_a_epi16(a_r_p5_m1, a_i_p5_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_m1); - square_a_epi16(a_r_p5_m3, a_i_p5_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_m3); - square_a_epi16(a_r_p5_m5, a_i_p5_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_m5); - square_a_epi16(a_r_p5_m7, a_i_p5_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p5_m7); - square_a_epi16(a_r_p3_p7, a_i_p3_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_p7); - square_a_epi16(a_r_p3_p5, a_i_p3_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_p5); - square_a_epi16(a_r_p3_p3, a_i_p3_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_p3); - square_a_epi16(a_r_p3_p1, a_i_p3_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_p1); - square_a_epi16(a_r_p3_m1, a_i_p3_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_m1); - square_a_epi16(a_r_p3_m3, a_i_p3_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_m3); - square_a_epi16(a_r_p3_m5, a_i_p3_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_m5); - square_a_epi16(a_r_p3_m7, a_i_p3_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p3_m7); - square_a_epi16(a_r_p1_p7, a_i_p1_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_p7); - square_a_epi16(a_r_p1_p5, a_i_p1_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_p5); - square_a_epi16(a_r_p1_p3, a_i_p1_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_p3); - square_a_epi16(a_r_p1_p1, a_i_p1_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_p1); - square_a_epi16(a_r_p1_m1, a_i_p1_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_m1); - square_a_epi16(a_r_p1_m3, a_i_p1_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_m3); - square_a_epi16(a_r_p1_m5, a_i_p1_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_m5); - square_a_epi16(a_r_p1_m7, a_i_p1_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_p1_m7); - square_a_epi16(a_r_m1_p7, a_i_m1_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_p7); - square_a_epi16(a_r_m1_p5, a_i_m1_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_p5); - square_a_epi16(a_r_m1_p3, a_i_m1_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_p3); - square_a_epi16(a_r_m1_p1, a_i_m1_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_p1); - square_a_epi16(a_r_m1_m1, a_i_m1_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_m1); - square_a_epi16(a_r_m1_m3, a_i_m1_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_m3); - square_a_epi16(a_r_m1_m5, a_i_m1_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_m5); - square_a_epi16(a_r_m1_m7, a_i_m1_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m1_m7); - square_a_epi16(a_r_m3_p7, a_i_m3_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_p7); - square_a_epi16(a_r_m3_p5, a_i_m3_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_p5); - square_a_epi16(a_r_m3_p3, a_i_m3_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_p3); - square_a_epi16(a_r_m3_p1, a_i_m3_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_p1); - square_a_epi16(a_r_m3_m1, a_i_m3_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_m1); - square_a_epi16(a_r_m3_m3, a_i_m3_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_m3); - square_a_epi16(a_r_m3_m5, a_i_m3_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_m5); - square_a_epi16(a_r_m3_m7, a_i_m3_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m3_m7); - square_a_epi16(a_r_m5_p7, a_i_m5_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_p7); - square_a_epi16(a_r_m5_p5, a_i_m5_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_p5); - square_a_epi16(a_r_m5_p3, a_i_m5_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_p3); - square_a_epi16(a_r_m5_p1, a_i_m5_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_p1); - square_a_epi16(a_r_m5_m1, a_i_m5_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_m1); - square_a_epi16(a_r_m5_m3, a_i_m5_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_m3); - square_a_epi16(a_r_m5_m5, a_i_m5_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_m5); - square_a_epi16(a_r_m5_m7, a_i_m5_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m5_m7); - square_a_epi16(a_r_m7_p7, a_i_m7_p7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_p7); - square_a_epi16(a_r_m7_p5, a_i_m7_p5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_p5); - square_a_epi16(a_r_m7_p3, a_i_m7_p3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_p3); - square_a_epi16(a_r_m7_p1, a_i_m7_p1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_p1); - square_a_epi16(a_r_m7_m1, a_i_m7_m1, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_m1); - square_a_epi16(a_r_m7_m3, a_i_m7_m3, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_m3); - square_a_epi16(a_r_m7_m5, a_i_m7_m5, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_m5); - square_a_epi16(a_r_m7_m7, a_i_m7_m7, ch_mag_int, SQRT_10_OVER_FOUR, a_sq_m7_m7); - - // Computing different multiples of ||h0||^2 - // x=1, y=1 - ch_mag_2_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,ONE_OVER_FOUR_SQRT_42); - ch_mag_2_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_2_over_42_with_sigma2,1); - // x=1, y=3 - ch_mag_10_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,FIVE_OVER_FOUR_SQRT_42); - ch_mag_10_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_10_over_42_with_sigma2,1); - // x=1, x=5 - ch_mag_26_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,THIRTEEN_OVER_FOUR_SQRT_42); - ch_mag_26_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_26_over_42_with_sigma2,1); - // x=1, y=7 - ch_mag_50_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42); - ch_mag_50_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_50_over_42_with_sigma2,1); - // x=3, y=3 - ch_mag_18_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,NINE_OVER_FOUR_SQRT_42); - ch_mag_18_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_18_over_42_with_sigma2,1); - // x=3, y=5 - ch_mag_34_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,SEVENTEEN_OVER_FOUR_SQRT_42); - ch_mag_34_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_34_over_42_with_sigma2,1); - // x=3, y=7 - ch_mag_58_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYNINE_OVER_FOUR_SQRT_42); - ch_mag_58_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_58_over_42_with_sigma2,2); - // x=5, y=5 - ch_mag_50_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42); - ch_mag_50_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_50_over_42_with_sigma2,1); - // x=5, y=7 - ch_mag_74_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,THIRTYSEVEN_OVER_FOUR_SQRT_42); - ch_mag_74_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_74_over_42_with_sigma2,2); - // x=7, y=7 - ch_mag_98_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,FORTYNINE_OVER_FOUR_SQRT_42); - ch_mag_98_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_98_over_42_with_sigma2,2); - - // Computing Metrics - xmm0 = _mm_subs_epi16(psi_a_p7_p7, a_sq_p7_p7); - xmm1 = _mm_adds_epi16(xmm0, y0_p_7_7); - bit_met_p7_p7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_p5, a_sq_p7_p5); - xmm1 = _mm_adds_epi16(xmm0, y0_p_7_5); - bit_met_p7_p5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_p3, a_sq_p7_p3); - xmm1 = _mm_adds_epi16(xmm0, y0_p_7_3); - bit_met_p7_p3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_p1, a_sq_p7_p1); - xmm1 = _mm_adds_epi16(xmm0, y0_p_7_1); - bit_met_p7_p1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_m1, a_sq_p7_m1); - xmm1 = _mm_adds_epi16(xmm0, y0_m_7_1); - bit_met_p7_m1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_m3, a_sq_p7_m3); - xmm1 = _mm_adds_epi16(xmm0, y0_m_7_3); - bit_met_p7_m3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_m5, a_sq_p7_m5); - xmm1 = _mm_adds_epi16(xmm0, y0_m_7_5); - bit_met_p7_m5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_m7, a_sq_p7_m7); - xmm1 = _mm_adds_epi16(xmm0, y0_m_7_7); - bit_met_p7_m7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_p7, a_sq_p5_p7); - xmm1 = _mm_adds_epi16(xmm0, y0_p_5_7); - bit_met_p5_p7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_p5, a_sq_p5_p5); - xmm1 = _mm_adds_epi16(xmm0, y0_p_5_5); - bit_met_p5_p5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_p3, a_sq_p5_p3); - xmm1 = _mm_adds_epi16(xmm0, y0_p_5_3); - bit_met_p5_p3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_p1, a_sq_p5_p1); - xmm1 = _mm_adds_epi16(xmm0, y0_p_5_1); - bit_met_p5_p1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_m1, a_sq_p5_m1); - xmm1 = _mm_adds_epi16(xmm0, y0_m_5_1); - bit_met_p5_m1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_m3, a_sq_p5_m3); - xmm1 = _mm_adds_epi16(xmm0, y0_m_5_3); - bit_met_p5_m3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_m5, a_sq_p5_m5); - xmm1 = _mm_adds_epi16(xmm0, y0_m_5_5); - bit_met_p5_m5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_m7, a_sq_p5_m7); - xmm1 = _mm_adds_epi16(xmm0, y0_m_5_7); - bit_met_p5_m7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_p7, a_sq_p3_p7); - xmm1 = _mm_adds_epi16(xmm0, y0_p_3_7); - bit_met_p3_p7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_p5, a_sq_p3_p5); - xmm1 = _mm_adds_epi16(xmm0, y0_p_3_5); - bit_met_p3_p5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_p3, a_sq_p3_p3); - xmm1 = _mm_adds_epi16(xmm0, y0_p_3_3); - bit_met_p3_p3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_p1, a_sq_p3_p1); - xmm1 = _mm_adds_epi16(xmm0, y0_p_3_1); - bit_met_p3_p1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_m1, a_sq_p3_m1); - xmm1 = _mm_adds_epi16(xmm0, y0_m_3_1); - bit_met_p3_m1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_m3, a_sq_p3_m3); - xmm1 = _mm_adds_epi16(xmm0, y0_m_3_3); - bit_met_p3_m3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_m5, a_sq_p3_m5); - xmm1 = _mm_adds_epi16(xmm0, y0_m_3_5); - bit_met_p3_m5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_m7, a_sq_p3_m7); - xmm1 = _mm_adds_epi16(xmm0, y0_m_3_7); - bit_met_p3_m7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_p7, a_sq_p1_p7); - xmm1 = _mm_adds_epi16(xmm0, y0_p_1_7); - bit_met_p1_p7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_p5, a_sq_p1_p5); - xmm1 = _mm_adds_epi16(xmm0, y0_p_1_5); - bit_met_p1_p5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_p3, a_sq_p1_p3); - xmm1 = _mm_adds_epi16(xmm0, y0_p_1_3); - bit_met_p1_p3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_p1, a_sq_p1_p1); - xmm1 = _mm_adds_epi16(xmm0, y0_p_1_1); - bit_met_p1_p1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_m1, a_sq_p1_m1); - xmm1 = _mm_adds_epi16(xmm0, y0_m_1_1); - bit_met_p1_m1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_m3, a_sq_p1_m3); - xmm1 = _mm_adds_epi16(xmm0, y0_m_1_3); - bit_met_p1_m3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_m5, a_sq_p1_m5); - xmm1 = _mm_adds_epi16(xmm0, y0_m_1_5); - bit_met_p1_m5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_m7, a_sq_p1_m7); - xmm1 = _mm_adds_epi16(xmm0, y0_m_1_7); - bit_met_p1_m7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - - xmm0 = _mm_subs_epi16(psi_a_m1_p7, a_sq_m1_p7); - xmm1 = _mm_subs_epi16(xmm0, y0_m_1_7); - bit_met_m1_p7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_p5, a_sq_m1_p5); - xmm1 = _mm_subs_epi16(xmm0, y0_m_1_5); - bit_met_m1_p5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_p3, a_sq_m1_p3); - xmm1 = _mm_subs_epi16(xmm0, y0_m_1_3); - bit_met_m1_p3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_p1, a_sq_m1_p1); - xmm1 = _mm_subs_epi16(xmm0, y0_m_1_1); - bit_met_m1_p1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_m1, a_sq_m1_m1); - xmm1 = _mm_subs_epi16(xmm0, y0_p_1_1); - bit_met_m1_m1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_m3, a_sq_m1_m3); - xmm1 = _mm_subs_epi16(xmm0, y0_p_1_3); - bit_met_m1_m3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_m5, a_sq_m1_m5); - xmm1 = _mm_subs_epi16(xmm0, y0_p_1_5); - bit_met_m1_m5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_m7, a_sq_m1_m7); - xmm1 = _mm_subs_epi16(xmm0, y0_p_1_7); - bit_met_m1_m7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_p7, a_sq_m3_p7); - xmm1 = _mm_subs_epi16(xmm0, y0_m_3_7); - bit_met_m3_p7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_p5, a_sq_m3_p5); - xmm1 = _mm_subs_epi16(xmm0, y0_m_3_5); - bit_met_m3_p5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_p3, a_sq_m3_p3); - xmm1 = _mm_subs_epi16(xmm0, y0_m_3_3); - bit_met_m3_p3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_p1, a_sq_m3_p1); - xmm1 = _mm_subs_epi16(xmm0, y0_m_3_1); - bit_met_m3_p1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_m1, a_sq_m3_m1); - xmm1 = _mm_subs_epi16(xmm0, y0_p_3_1); - bit_met_m3_m1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_m3, a_sq_m3_m3); - xmm1 = _mm_subs_epi16(xmm0, y0_p_3_3); - bit_met_m3_m3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_m5, a_sq_m3_m5); - xmm1 = _mm_subs_epi16(xmm0, y0_p_3_5); - bit_met_m3_m5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_m7, a_sq_m3_m7); - xmm1 = _mm_subs_epi16(xmm0, y0_p_3_7); - bit_met_m3_m7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_p7, a_sq_m5_p7); - xmm1 = _mm_subs_epi16(xmm0, y0_m_5_7); - bit_met_m5_p7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_p5, a_sq_m5_p5); - xmm1 = _mm_subs_epi16(xmm0, y0_m_5_5); - bit_met_m5_p5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_p3, a_sq_m5_p3); - xmm1 = _mm_subs_epi16(xmm0, y0_m_5_3); - bit_met_m5_p3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_p1, a_sq_m5_p1); - xmm1 = _mm_subs_epi16(xmm0, y0_m_5_1); - bit_met_m5_p1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_m1, a_sq_m5_m1); - xmm1 = _mm_subs_epi16(xmm0, y0_p_5_1); - bit_met_m5_m1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_m3, a_sq_m5_m3); - xmm1 = _mm_subs_epi16(xmm0, y0_p_5_3); - bit_met_m5_m3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_m5, a_sq_m5_m5); - xmm1 = _mm_subs_epi16(xmm0, y0_p_5_5); - bit_met_m5_m5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_m7, a_sq_m5_m7); - xmm1 = _mm_subs_epi16(xmm0, y0_p_5_7); - bit_met_m5_m7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_p7, a_sq_m7_p7); - xmm1 = _mm_subs_epi16(xmm0, y0_m_7_7); - bit_met_m7_p7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_p5, a_sq_m7_p5); - xmm1 = _mm_subs_epi16(xmm0, y0_m_7_5); - bit_met_m7_p5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_p3, a_sq_m7_p3); - xmm1 = _mm_subs_epi16(xmm0, y0_m_7_3); - bit_met_m7_p3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_p1, a_sq_m7_p1); - xmm1 = _mm_subs_epi16(xmm0, y0_m_7_1); - bit_met_m7_p1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_m1, a_sq_m7_m1); - xmm1 = _mm_subs_epi16(xmm0, y0_p_7_1); - bit_met_m7_m1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_m3, a_sq_m7_m3); - xmm1 = _mm_subs_epi16(xmm0, y0_p_7_3); - bit_met_m7_m3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_m5, a_sq_m7_m5); - xmm1 = _mm_subs_epi16(xmm0, y0_p_7_5); - bit_met_m7_m5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_m7, a_sq_m7_m7); - xmm1 = _mm_subs_epi16(xmm0, y0_p_7_7); - bit_met_m7_m7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); - - // Detection for 1st bit (LTE mapping) - // bit = 1 - xmm0 = _mm_max_epi16(bit_met_m7_p7, bit_met_m7_p5); - xmm1 = _mm_max_epi16(bit_met_m7_p3, bit_met_m7_p1); - xmm2 = _mm_max_epi16(bit_met_m7_m1, bit_met_m7_m3); - xmm3 = _mm_max_epi16(bit_met_m7_m5, bit_met_m7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m5_p7, bit_met_m5_p5); - xmm1 = _mm_max_epi16(bit_met_m5_p3, bit_met_m5_p1); - xmm2 = _mm_max_epi16(bit_met_m5_m1, bit_met_m5_m3); - xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m5_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_m3_p7, bit_met_m3_p5); - xmm1 = _mm_max_epi16(bit_met_m3_p3, bit_met_m3_p1); - xmm2 = _mm_max_epi16(bit_met_m3_m1, bit_met_m3_m3); - xmm3 = _mm_max_epi16(bit_met_m3_m5, bit_met_m3_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_m1_p7, bit_met_m1_p5); - xmm1 = _mm_max_epi16(bit_met_m1_p3, bit_met_m1_p1); - xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m1_m3); - xmm3 = _mm_max_epi16(bit_met_m1_m5, bit_met_m1_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - // bit = 0 - xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p7_p5); - xmm1 = _mm_max_epi16(bit_met_p7_p3, bit_met_p7_p1); - xmm2 = _mm_max_epi16(bit_met_p7_m1, bit_met_p7_m3); - xmm3 = _mm_max_epi16(bit_met_p7_m5, bit_met_p7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p5_p7, bit_met_p5_p5); - xmm1 = _mm_max_epi16(bit_met_p5_p3, bit_met_p5_p1); - xmm2 = _mm_max_epi16(bit_met_p5_m1, bit_met_p5_m3); - xmm3 = _mm_max_epi16(bit_met_p5_m5, bit_met_p5_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p3_p7, bit_met_p3_p5); - xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p3_p1); - xmm2 = _mm_max_epi16(bit_met_p3_m1, bit_met_p3_m3); - xmm3 = _mm_max_epi16(bit_met_p3_m5, bit_met_p3_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p1_p7, bit_met_p1_p5); - xmm1 = _mm_max_epi16(bit_met_p1_p3, bit_met_p1_p1); - xmm2 = _mm_max_epi16(bit_met_p1_m1, bit_met_p1_m3); - xmm3 = _mm_max_epi16(bit_met_p1_m5, bit_met_p1_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y0r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - // Detection for 2nd bit (LTE mapping) - // bit = 1 - xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); - xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); - xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); - xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); - xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); - xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); - xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); - xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); - xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); - xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - // bit = 0 - xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); - xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); - xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m7_p1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); - xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); - xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); - xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); - xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m7_p5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); - xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); - xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); - xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y1r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - // Detection for 3rd bit (LTE mapping) - xmm0 = _mm_max_epi16(bit_met_m7_m7, bit_met_m7_m5); - xmm1 = _mm_max_epi16(bit_met_m7_m3, bit_met_m7_m1); - xmm2 = _mm_max_epi16(bit_met_m7_p1, bit_met_m7_p3); - xmm3 = _mm_max_epi16(bit_met_m7_p5, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m5_m7, bit_met_m5_m5); - xmm1 = _mm_max_epi16(bit_met_m5_m3, bit_met_m5_m1); - xmm2 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p5_m7, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p5_m3, bit_met_p5_m1); - xmm2 = _mm_max_epi16(bit_met_p5_p1, bit_met_p5_p3); - xmm3 = _mm_max_epi16(bit_met_p5_p5, bit_met_p5_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p7_m5); - xmm1 = _mm_max_epi16(bit_met_p7_m3, bit_met_p7_m1); - xmm2 = _mm_max_epi16(bit_met_p7_p1, bit_met_p7_p3); - xmm3 = _mm_max_epi16(bit_met_p7_p5, bit_met_p7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - xmm0 = _mm_max_epi16(bit_met_m3_m7, bit_met_m3_m5); - xmm1 = _mm_max_epi16(bit_met_m3_m3, bit_met_m3_m1); - xmm2 = _mm_max_epi16(bit_met_m3_p1, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m3_p5, bit_met_m3_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m1_m7, bit_met_m1_m5); - xmm1 = _mm_max_epi16(bit_met_m1_m3, bit_met_m1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m1_p3); - xmm3 = _mm_max_epi16(bit_met_m1_p5, bit_met_m1_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p1_m7, bit_met_p1_m5); - xmm1 = _mm_max_epi16(bit_met_p1_m3, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_p1_p1, bit_met_p1_p3); - xmm3 = _mm_max_epi16(bit_met_p1_p5, bit_met_p1_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p3_m7, bit_met_p3_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p3_m1); - xmm2 = _mm_max_epi16(bit_met_p3_p1, bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_p3_p5, bit_met_p3_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y2r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - // Detection for 4th bit (LTE mapping) - xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); - xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); - xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); - xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); - xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); - xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); - xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); - xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); - xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); - xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); - xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); - xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); - xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); - xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); - xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); - xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); - xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m7_p1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); - xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); - xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y0i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - - // Detection for 5th bit (LTE mapping) - xmm0 = _mm_max_epi16(bit_met_m7_m7, bit_met_m7_m5); - xmm1 = _mm_max_epi16(bit_met_m7_m3, bit_met_m7_m1); - xmm2 = _mm_max_epi16(bit_met_m7_p1, bit_met_m7_p3); - xmm3 = _mm_max_epi16(bit_met_m7_p5, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m1_m7, bit_met_m1_m5); - xmm1 = _mm_max_epi16(bit_met_m1_m3, bit_met_m1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m1_p3); - xmm3 = _mm_max_epi16(bit_met_m1_p5, bit_met_m1_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p1_m7, bit_met_p1_m5); - xmm1 = _mm_max_epi16(bit_met_p1_m3, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_p1_p1, bit_met_p1_p3); - xmm3 = _mm_max_epi16(bit_met_p1_p5, bit_met_p1_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p7_m5); - xmm1 = _mm_max_epi16(bit_met_p7_m3, bit_met_p7_m1); - xmm2 = _mm_max_epi16(bit_met_p7_p1, bit_met_p7_p3); - xmm3 = _mm_max_epi16(bit_met_p7_p5, bit_met_p7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - xmm0 = _mm_max_epi16(bit_met_m5_m7, bit_met_m5_m5); - xmm1 = _mm_max_epi16(bit_met_m5_m3, bit_met_m5_m1); - xmm2 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m3_m7, bit_met_m3_m5); - xmm1 = _mm_max_epi16(bit_met_m3_m3, bit_met_m3_m1); - xmm2 = _mm_max_epi16(bit_met_m3_p1, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m3_p5, bit_met_m3_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p3_m7, bit_met_p3_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p3_m1); - xmm2 = _mm_max_epi16(bit_met_p3_p1, bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_p3_p5, bit_met_p3_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p5_m7, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p5_m3, bit_met_p5_m1); - xmm2 = _mm_max_epi16(bit_met_p5_p1, bit_met_p5_p3); - xmm3 = _mm_max_epi16(bit_met_p5_p5, bit_met_p5_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y1i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - // Detection for 6th bit (LTE mapping) - xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); - xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); - xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); - xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); - xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); - xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); - xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); - xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); - xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); - xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); - xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); - xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); - xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); - xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); - xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); - xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); - xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); - xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); - xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m7_p5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y2i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - - // map to output stream, difficult to do in SIMD since we have 6 16bit LLRs - // RE 1 - j = 24*i; - stream0_out[j + 0] = ((short *)&y0r)[0]; - stream0_out[j + 1] = ((short *)&y1r)[0]; - stream0_out[j + 2] = ((short *)&y2r)[0]; - stream0_out[j + 3] = ((short *)&y0i)[0]; - stream0_out[j + 4] = ((short *)&y1i)[0]; - stream0_out[j + 5] = ((short *)&y2i)[0]; - // RE 2 - stream0_out[j + 6] = ((short *)&y0r)[1]; - stream0_out[j + 7] = ((short *)&y1r)[1]; - stream0_out[j + 8] = ((short *)&y2r)[1]; - stream0_out[j + 9] = ((short *)&y0i)[1]; - stream0_out[j + 10] = ((short *)&y1i)[1]; - stream0_out[j + 11] = ((short *)&y2i)[1]; - // RE 3 - stream0_out[j + 12] = ((short *)&y0r)[2]; - stream0_out[j + 13] = ((short *)&y1r)[2]; - stream0_out[j + 14] = ((short *)&y2r)[2]; - stream0_out[j + 15] = ((short *)&y0i)[2]; - stream0_out[j + 16] = ((short *)&y1i)[2]; - stream0_out[j + 17] = ((short *)&y2i)[2]; - // RE 4 - stream0_out[j + 18] = ((short *)&y0r)[3]; - stream0_out[j + 19] = ((short *)&y1r)[3]; - stream0_out[j + 20] = ((short *)&y2r)[3]; - stream0_out[j + 21] = ((short *)&y0i)[3]; - stream0_out[j + 22] = ((short *)&y1i)[3]; - stream0_out[j + 23] = ((short *)&y2i)[3]; - // RE 5 - stream0_out[j + 24] = ((short *)&y0r)[4]; - stream0_out[j + 25] = ((short *)&y1r)[4]; - stream0_out[j + 26] = ((short *)&y2r)[4]; - stream0_out[j + 27] = ((short *)&y0i)[4]; - stream0_out[j + 28] = ((short *)&y1i)[4]; - stream0_out[j + 29] = ((short *)&y2i)[4]; - // RE 6 - stream0_out[j + 30] = ((short *)&y0r)[5]; - stream0_out[j + 31] = ((short *)&y1r)[5]; - stream0_out[j + 32] = ((short *)&y2r)[5]; - stream0_out[j + 33] = ((short *)&y0i)[5]; - stream0_out[j + 34] = ((short *)&y1i)[5]; - stream0_out[j + 35] = ((short *)&y2i)[5]; - // RE 7 - stream0_out[j + 36] = ((short *)&y0r)[6]; - stream0_out[j + 37] = ((short *)&y1r)[6]; - stream0_out[j + 38] = ((short *)&y2r)[6]; - stream0_out[j + 39] = ((short *)&y0i)[6]; - stream0_out[j + 40] = ((short *)&y1i)[6]; - stream0_out[j + 41] = ((short *)&y2i)[6]; - // RE 8 - stream0_out[j + 42] = ((short *)&y0r)[7]; - stream0_out[j + 43] = ((short *)&y1r)[7]; - stream0_out[j + 44] = ((short *)&y2r)[7]; - stream0_out[j + 45] = ((short *)&y0i)[7]; - stream0_out[j + 46] = ((short *)&y1i)[7]; - stream0_out[j + 47] = ((short *)&y2i)[7]; - -#elif defined(__arm__) || defined(__aarch64__) - -#endif - } - -#if defined(__x86_64__) || defined(__i386__) - _mm_empty(); - _m_empty(); -#endif - -} - - -int nr_dlsch_64qam_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, - int32_t **rxdataF_comp, - int32_t **rxdataF_comp_i, - int32_t **dl_ch_mag, - int32_t **dl_ch_mag_i, - int32_t **rho_i, - int16_t *dlsch_llr, - uint8_t symbol, - uint8_t first_symbol_flag, - uint16_t nb_rb, - uint16_t pbch_pss_sss_adjust, - int16_t **llr16p) -{ - - int16_t *rxF = (int16_t*)&rxdataF_comp[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *rxF_i = (int16_t*)&rxdataF_comp_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *ch_mag = (int16_t*)&dl_ch_mag[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *ch_mag_i = (int16_t*)&dl_ch_mag_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *rho = (int16_t*)&rho_i[0][(symbol*frame_parms->N_RB_DL*12)]; - int16_t *llr16; - int len; - uint8_t symbol_mod = (symbol >= (7-frame_parms->Ncp))? (symbol-(7-frame_parms->Ncp)) : symbol; - - //first symbol has different structure due to more pilots - if (first_symbol_flag == 1) { - llr16 = (int16_t*)dlsch_llr; - } else { - llr16 = (int16_t*)(*llr16p); - } - - AssertFatal(llr16!=NULL,"nr_dlsch_16qam_64qam_llr:llr is null, symbol %d\n",symbol); - - if ((symbol_mod==0) || (symbol_mod==(4-frame_parms->Ncp))) { - // if symbol has pilots - if (frame_parms->nb_antenna_ports_gNB!=1) - // in 2 antenna ports we have 8 REs per symbol per RB - len = (nb_rb*8) - (2*pbch_pss_sss_adjust/3); - else - // for 1 antenna port we have 10 REs per symbol per RB - len = (nb_rb*10) - (5*pbch_pss_sss_adjust/6); - } else { - // symbol has no pilots - len = (nb_rb*12) - pbch_pss_sss_adjust; - } - - nr_qam64_qam16((short *)rxF, - (short *)rxF_i, - (short *)ch_mag, - (short *)ch_mag_i, - (short *)llr16, - (short *)rho, - len); - - llr16 += (6*len); - *llr16p = (short *)llr16; - return(0); -} - -#if 0 -void qam64_qam64(short *stream0_in, - short *stream1_in, - short *ch_mag, - short *ch_mag_i, - short *stream0_out, - short *rho01, - int length - ) -{ - - /* - Author: S. Wagner - Date: 31-07-12 - - Input: - stream0_in: MF filter for 1st stream, i.e., y0=h0'*y - stream1_in: MF filter for 2nd stream, i.e., y1=h1'*y - ch_mag: 4*h0/sqrt(42), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc - ch_mag_i: 4*h1/sqrt(42), [Re0 Im0 Re1 Im1] s.t. Im0=Re0, Im1=Re1, etc - rho01: Channel cross correlation, i.e., h1'*h0 - - Output: - stream0_out: output LLRs for 1st stream - */ - -#if defined(__x86_64__) || defined(__i386__) - - __m128i *rho01_128i = (__m128i *)rho01; - __m128i *stream0_128i_in = (__m128i *)stream0_in; - __m128i *stream1_128i_in = (__m128i *)stream1_in; - __m128i *ch_mag_128i = (__m128i *)ch_mag; - __m128i *ch_mag_128i_i = (__m128i *)ch_mag_i; - - __m128i ONE_OVER_SQRT_42 = _mm_set1_epi16(10112); // round(1/sqrt(42)*2^16) - __m128i THREE_OVER_SQRT_42 = _mm_set1_epi16(30337); // round(3/sqrt(42)*2^16) - __m128i FIVE_OVER_SQRT_42 = _mm_set1_epi16(25281); // round(5/sqrt(42)*2^15) - __m128i SEVEN_OVER_SQRT_42 = _mm_set1_epi16(17697); // round(7/sqrt(42)*2^14) Q2.14 - __m128i ONE_OVER_SQRT_2 = _mm_set1_epi16(23170); // round(1/sqrt(2)*2^15) - __m128i ONE_OVER_SQRT_2_42 = _mm_set1_epi16(3575); // round(1/sqrt(2*42)*2^15) - __m128i THREE_OVER_SQRT_2_42 = _mm_set1_epi16(10726); // round(3/sqrt(2*42)*2^15) - __m128i FIVE_OVER_SQRT_2_42 = _mm_set1_epi16(17876); // round(5/sqrt(2*42)*2^15) - __m128i SEVEN_OVER_SQRT_2_42 = _mm_set1_epi16(25027); // round(7/sqrt(2*42)*2^15) - __m128i FORTYNINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(30969); // round(49/(4*sqrt(42))*2^14), Q2.14 - __m128i THIRTYSEVEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(23385); // round(37/(4*sqrt(42))*2^14), Q2.14 - __m128i TWENTYFIVE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(31601); // round(25/(4*sqrt(42))*2^15) - __m128i TWENTYNINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(18329); // round(29/(4*sqrt(42))*2^15), Q2.14 - __m128i SEVENTEEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(21489); // round(17/(4*sqrt(42))*2^15) - __m128i NINE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(11376); // round(9/(4*sqrt(42))*2^15) - __m128i THIRTEEN_OVER_FOUR_SQRT_42 = _mm_set1_epi16(16433); // round(13/(4*sqrt(42))*2^15) - __m128i FIVE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(6320); // round(5/(4*sqrt(42))*2^15) - __m128i ONE_OVER_FOUR_SQRT_42 = _mm_set1_epi16(1264); // round(1/(4*sqrt(42))*2^15) - __m128i SQRT_42_OVER_FOUR = _mm_set1_epi16(13272); // round(sqrt(42)/4*2^13), Q3.12 - - __m128i ch_mag_des; - __m128i ch_mag_int; - __m128i ch_mag_98_over_42_with_sigma2; - __m128i ch_mag_74_over_42_with_sigma2; - __m128i ch_mag_58_over_42_with_sigma2; - __m128i ch_mag_50_over_42_with_sigma2; - __m128i ch_mag_34_over_42_with_sigma2; - __m128i ch_mag_18_over_42_with_sigma2; - __m128i ch_mag_26_over_42_with_sigma2; - __m128i ch_mag_10_over_42_with_sigma2; - __m128i ch_mag_2_over_42_with_sigma2; - __m128i y0r_one_over_sqrt_21; - __m128i y0r_three_over_sqrt_21; - __m128i y0r_five_over_sqrt_21; - __m128i y0r_seven_over_sqrt_21; - __m128i y0i_one_over_sqrt_21; - __m128i y0i_three_over_sqrt_21; - __m128i y0i_five_over_sqrt_21; - __m128i y0i_seven_over_sqrt_21; - __m128i ch_mag_int_with_sigma2; - __m128i two_ch_mag_int_with_sigma2; - __m128i three_ch_mag_int_with_sigma2; -#elif defined(__arm__) || defined(__aarch64__) - -#endif - - int i,j; - - - for (i=0; i>2; i+=2) { - -#if defined(__x86_64__) || defined(__i386__) - - // Get rho - xmm0 = rho01_128i[i]; - xmm1 = rho01_128i[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - xmm2 = _mm_unpacklo_epi64(xmm0,xmm1); // Re(rho) - xmm3 = _mm_unpackhi_epi64(xmm0,xmm1); // Im(rho) - rho_rpi = _mm_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho) - rho_rmi = _mm_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho) - - // Compute the different rhos - rho_rpi_1_1 = _mm_mulhi_epi16(rho_rpi, ONE_OVER_SQRT_42); - rho_rmi_1_1 = _mm_mulhi_epi16(rho_rmi, ONE_OVER_SQRT_42); - rho_rpi_3_3 = _mm_mulhi_epi16(rho_rpi, THREE_OVER_SQRT_42); - rho_rmi_3_3 = _mm_mulhi_epi16(rho_rmi, THREE_OVER_SQRT_42); - rho_rpi_5_5 = _mm_mulhi_epi16(rho_rpi, FIVE_OVER_SQRT_42); - rho_rmi_5_5 = _mm_mulhi_epi16(rho_rmi, FIVE_OVER_SQRT_42); - rho_rpi_7_7 = _mm_mulhi_epi16(rho_rpi, SEVEN_OVER_SQRT_42); - rho_rmi_7_7 = _mm_mulhi_epi16(rho_rmi, SEVEN_OVER_SQRT_42); - - rho_rpi_5_5 = _mm_slli_epi16(rho_rpi_5_5, 1); - rho_rmi_5_5 = _mm_slli_epi16(rho_rmi_5_5, 1); - rho_rpi_7_7 = _mm_slli_epi16(rho_rpi_7_7, 2); - rho_rmi_7_7 = _mm_slli_epi16(rho_rmi_7_7, 2); - - xmm4 = _mm_mulhi_epi16(xmm2, ONE_OVER_SQRT_42); - xmm5 = _mm_mulhi_epi16(xmm3, ONE_OVER_SQRT_42); - xmm6 = _mm_mulhi_epi16(xmm3, THREE_OVER_SQRT_42); - xmm7 = _mm_mulhi_epi16(xmm3, FIVE_OVER_SQRT_42); - xmm8 = _mm_mulhi_epi16(xmm3, SEVEN_OVER_SQRT_42); - xmm7 = _mm_slli_epi16(xmm7, 1); - xmm8 = _mm_slli_epi16(xmm8, 2); - - rho_rpi_1_3 = _mm_adds_epi16(xmm4, xmm6); - rho_rmi_1_3 = _mm_subs_epi16(xmm4, xmm6); - rho_rpi_1_5 = _mm_adds_epi16(xmm4, xmm7); - rho_rmi_1_5 = _mm_subs_epi16(xmm4, xmm7); - rho_rpi_1_7 = _mm_adds_epi16(xmm4, xmm8); - rho_rmi_1_7 = _mm_subs_epi16(xmm4, xmm8); - - xmm4 = _mm_mulhi_epi16(xmm2, THREE_OVER_SQRT_42); - rho_rpi_3_1 = _mm_adds_epi16(xmm4, xmm5); - rho_rmi_3_1 = _mm_subs_epi16(xmm4, xmm5); - rho_rpi_3_5 = _mm_adds_epi16(xmm4, xmm7); - rho_rmi_3_5 = _mm_subs_epi16(xmm4, xmm7); - rho_rpi_3_7 = _mm_adds_epi16(xmm4, xmm8); - rho_rmi_3_7 = _mm_subs_epi16(xmm4, xmm8); - - xmm4 = _mm_mulhi_epi16(xmm2, FIVE_OVER_SQRT_42); - xmm4 = _mm_slli_epi16(xmm4, 1); - rho_rpi_5_1 = _mm_adds_epi16(xmm4, xmm5); - rho_rmi_5_1 = _mm_subs_epi16(xmm4, xmm5); - rho_rpi_5_3 = _mm_adds_epi16(xmm4, xmm6); - rho_rmi_5_3 = _mm_subs_epi16(xmm4, xmm6); - rho_rpi_5_7 = _mm_adds_epi16(xmm4, xmm8); - rho_rmi_5_7 = _mm_subs_epi16(xmm4, xmm8); - - xmm4 = _mm_mulhi_epi16(xmm2, SEVEN_OVER_SQRT_42); - xmm4 = _mm_slli_epi16(xmm4, 2); - rho_rpi_7_1 = _mm_adds_epi16(xmm4, xmm5); - rho_rmi_7_1 = _mm_subs_epi16(xmm4, xmm5); - rho_rpi_7_3 = _mm_adds_epi16(xmm4, xmm6); - rho_rmi_7_3 = _mm_subs_epi16(xmm4, xmm6); - rho_rpi_7_5 = _mm_adds_epi16(xmm4, xmm7); - rho_rmi_7_5 = _mm_subs_epi16(xmm4, xmm7); - - // Rearrange interfering MF output - xmm0 = stream1_128i_in[i]; - xmm1 = stream1_128i_in[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y1r = _mm_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)] - y1i = _mm_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)] - - // Psi_r calculation from rho_rpi or rho_rmi - xmm0 = _mm_setzero_si128(); // ZERO for abs_pi16 - xmm2 = _mm_subs_epi16(rho_rpi_7_7, y1r); - psi_r_p7_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_5, y1r); - psi_r_p7_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_3, y1r); - psi_r_p7_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_1, y1r); - psi_r_p7_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_1, y1r); - psi_r_p7_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_3, y1r); - psi_r_p7_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_5, y1r); - psi_r_p7_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_7, y1r); - psi_r_p7_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_7, y1r); - psi_r_p5_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_5, y1r); - psi_r_p5_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_3, y1r); - psi_r_p5_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_1, y1r); - psi_r_p5_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_1, y1r); - psi_r_p5_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_3, y1r); - psi_r_p5_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_5, y1r); - psi_r_p5_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_7, y1r); - psi_r_p5_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_7, y1r); - psi_r_p3_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_5, y1r); - psi_r_p3_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_3, y1r); - psi_r_p3_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_1, y1r); - psi_r_p3_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_1, y1r); - psi_r_p3_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_3, y1r); - psi_r_p3_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_5, y1r); - psi_r_p3_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_7, y1r); - psi_r_p3_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_7, y1r); - psi_r_p1_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_5, y1r); - psi_r_p1_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_3, y1r); - psi_r_p1_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_1, y1r); - psi_r_p1_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_1, y1r); - psi_r_p1_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_3, y1r); - psi_r_p1_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_5, y1r); - psi_r_p1_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_7, y1r); - psi_r_p1_m7 = _mm_abs_epi16(xmm2); - - xmm2 = _mm_adds_epi16(rho_rmi_1_7, y1r); - psi_r_m1_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_5, y1r); - psi_r_m1_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_3, y1r); - psi_r_m1_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_1, y1r); - psi_r_m1_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_1, y1r); - psi_r_m1_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_3, y1r); - psi_r_m1_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_5, y1r); - psi_r_m1_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_7, y1r); - psi_r_m1_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_7, y1r); - psi_r_m3_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_5, y1r); - psi_r_m3_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_3, y1r); - psi_r_m3_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_1, y1r); - psi_r_m3_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_1, y1r); - psi_r_m3_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_3, y1r); - psi_r_m3_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_5, y1r); - psi_r_m3_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_7, y1r); - psi_r_m3_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_7, y1r); - psi_r_m5_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_5, y1r); - psi_r_m5_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_3, y1r); - psi_r_m5_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_1, y1r); - psi_r_m5_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_1, y1r); - psi_r_m5_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_3, y1r); - psi_r_m5_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_5, y1r); - psi_r_m5_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_7, y1r); - psi_r_m5_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_7, y1r); - psi_r_m7_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_5, y1r); - psi_r_m7_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_3, y1r); - psi_r_m7_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_1, y1r); - psi_r_m7_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_1, y1r); - psi_r_m7_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_3, y1r); - psi_r_m7_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_5, y1r); - psi_r_m7_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_7, y1r); - psi_r_m7_m7 = _mm_abs_epi16(xmm2); - - // Psi_i calculation from rho_rpi or rho_rmi - xmm2 = _mm_subs_epi16(rho_rmi_7_7, y1i); - psi_i_p7_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_7, y1i); - psi_i_p7_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_7, y1i); - psi_i_p7_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_7, y1i); - psi_i_p7_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_7, y1i); - psi_i_p7_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_7, y1i); - psi_i_p7_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_7, y1i); - psi_i_p7_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_7, y1i); - psi_i_p7_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_5, y1i); - psi_i_p5_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_5, y1i); - psi_i_p5_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_5, y1i); - psi_i_p5_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_5, y1i); - psi_i_p5_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_5, y1i); - psi_i_p5_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_5, y1i); - psi_i_p5_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_5, y1i); - psi_i_p5_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_5, y1i); - psi_i_p5_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_3, y1i); - psi_i_p3_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_3, y1i); - psi_i_p3_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_3, y1i); - psi_i_p3_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_3, y1i); - psi_i_p3_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_3, y1i); - psi_i_p3_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_3, y1i); - psi_i_p3_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_3, y1i); - psi_i_p3_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_3, y1i); - psi_i_p3_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_7_1, y1i); - psi_i_p1_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_5_1, y1i); - psi_i_p1_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_3_1, y1i); - psi_i_p1_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rmi_1_1, y1i); - psi_i_p1_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_1_1, y1i); - psi_i_p1_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_3_1, y1i); - psi_i_p1_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_5_1, y1i); - psi_i_p1_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rpi_7_1, y1i); - psi_i_p1_m7 = _mm_abs_epi16(xmm2); - - xmm2 = _mm_subs_epi16(rho_rpi_7_1, y1i); - psi_i_m1_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_1, y1i); - psi_i_m1_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_1, y1i); - psi_i_m1_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_1, y1i); - psi_i_m1_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_1, y1i); - psi_i_m1_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_1, y1i); - psi_i_m1_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_1, y1i); - psi_i_m1_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_1, y1i); - psi_i_m1_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_3, y1i); - psi_i_m3_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_3, y1i); - psi_i_m3_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_3, y1i); - psi_i_m3_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_3, y1i); - psi_i_m3_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_3, y1i); - psi_i_m3_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_3, y1i); - psi_i_m3_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_3, y1i); - psi_i_m3_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_3, y1i); - psi_i_m3_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_5, y1i); - psi_i_m5_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_5, y1i); - psi_i_m5_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_5, y1i); - psi_i_m5_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_5, y1i); - psi_i_m5_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_5, y1i); - psi_i_m5_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_5, y1i); - psi_i_m5_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_5, y1i); - psi_i_m5_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_5, y1i); - psi_i_m5_m7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_7_7, y1i); - psi_i_m7_p7 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_5_7, y1i); - psi_i_m7_p5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_3_7, y1i); - psi_i_m7_p3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_subs_epi16(rho_rpi_1_7, y1i); - psi_i_m7_p1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_1_7, y1i); - psi_i_m7_m1 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_3_7, y1i); - psi_i_m7_m3 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_5_7, y1i); - psi_i_m7_m5 = _mm_abs_epi16(xmm2); - xmm2 = _mm_adds_epi16(rho_rmi_7_7, y1i); - psi_i_m7_m7 = _mm_abs_epi16(xmm2); - - - // Rearrange desired MF output - xmm0 = stream0_128i_in[i]; - xmm1 = stream0_128i_in[i+1]; - xmm0 = _mm_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm0 = _mm_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm1 = _mm_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3)); - //xmm0 = [Re(0,1) Re(2,3) Im(0,1) Im(2,3)] - //xmm1 = [Re(4,5) Re(6,7) Im(4,5) Im(6,7)] - y0r = _mm_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)] - y0i = _mm_unpackhi_epi64(xmm0,xmm1); - - // Rearrange desired channel magnitudes - xmm2 = ch_mag_128i[i]; // = [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2)]*(2/sqrt(10)) - xmm3 = ch_mag_128i[i+1]; // = [|h|^2(3),|h|^2(3),|h|^2(4),|h|^2(4)]*(2/sqrt(10)) - xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - ch_mag_des = _mm_unpacklo_epi64(xmm2,xmm3); - - // Rearrange interfering channel magnitudes - xmm2 = ch_mag_128i_i[i]; - xmm3 = ch_mag_128i_i[i+1]; - xmm2 = _mm_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm2 = _mm_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - xmm3 = _mm_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3)); - ch_mag_int = _mm_unpacklo_epi64(xmm2,xmm3); - - y0r_one_over_sqrt_21 = _mm_mulhi_epi16(y0r, ONE_OVER_SQRT_42); - y0r_three_over_sqrt_21 = _mm_mulhi_epi16(y0r, THREE_OVER_SQRT_42); - y0r_five_over_sqrt_21 = _mm_mulhi_epi16(y0r, FIVE_OVER_SQRT_42); - y0r_five_over_sqrt_21 = _mm_slli_epi16(y0r_five_over_sqrt_21, 1); - y0r_seven_over_sqrt_21 = _mm_mulhi_epi16(y0r, SEVEN_OVER_SQRT_42); - y0r_seven_over_sqrt_21 = _mm_slli_epi16(y0r_seven_over_sqrt_21, 2); // Q2.14 - - y0i_one_over_sqrt_21 = _mm_mulhi_epi16(y0i, ONE_OVER_SQRT_42); - y0i_three_over_sqrt_21 = _mm_mulhi_epi16(y0i, THREE_OVER_SQRT_42); - y0i_five_over_sqrt_21 = _mm_mulhi_epi16(y0i, FIVE_OVER_SQRT_42); - y0i_five_over_sqrt_21 = _mm_slli_epi16(y0i_five_over_sqrt_21, 1); - y0i_seven_over_sqrt_21 = _mm_mulhi_epi16(y0i, SEVEN_OVER_SQRT_42); - y0i_seven_over_sqrt_21 = _mm_slli_epi16(y0i_seven_over_sqrt_21, 2); // Q2.14 - - y0_p_7_1 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21); - y0_p_7_3 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21); - y0_p_7_5 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21); - y0_p_7_7 = _mm_adds_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_p_5_1 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21); - y0_p_5_3 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21); - y0_p_5_5 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21); - y0_p_5_7 = _mm_adds_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_p_3_1 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21); - y0_p_3_3 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21); - y0_p_3_5 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21); - y0_p_3_7 = _mm_adds_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_p_1_1 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21); - y0_p_1_3 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21); - y0_p_1_5 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21); - y0_p_1_7 = _mm_adds_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21); - - y0_m_1_1 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21); - y0_m_1_3 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21); - y0_m_1_5 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21); - y0_m_1_7 = _mm_subs_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_m_3_1 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21); - y0_m_3_3 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21); - y0_m_3_5 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21); - y0_m_3_7 = _mm_subs_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_m_5_1 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21); - y0_m_5_3 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21); - y0_m_5_5 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21); - y0_m_5_7 = _mm_subs_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21); - y0_m_7_1 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21); - y0_m_7_3 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21); - y0_m_7_5 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21); - y0_m_7_7 = _mm_subs_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21); - - // Detection of interference term - ch_mag_int_with_sigma2 = _mm_srai_epi16(ch_mag_int, 1); // *2 - two_ch_mag_int_with_sigma2 = ch_mag_int; // *4 - three_ch_mag_int_with_sigma2 = _mm_adds_epi16(ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2); // *6 - - interference_abs_64qam_epi16(psi_r_p7_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p7_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p7_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p7_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p7_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p7_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p7_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p7_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p7_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p5_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p5_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p5_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p5_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p5_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p5_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p5_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p5_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p5_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p3_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p3_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p3_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p3_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p3_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p3_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p3_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p3_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p3_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p1_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p1_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p1_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p1_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p1_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_p1_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_p1_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m1_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m1_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m3_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m3_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m3_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m3_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m3_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m3_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m3_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m3_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m3_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m5_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m5_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m5_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m5_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m5_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m5_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m5_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m5_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m5_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m7_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m7_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m7_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m7_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m7_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m7_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m7_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_r_m7_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_r_m7_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - - interference_abs_64qam_epi16(psi_i_p7_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p7_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p7_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p7_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p7_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p7_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p7_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p7_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p7_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p5_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p5_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p5_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p5_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p5_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p5_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p5_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p5_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p5_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p3_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p3_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p3_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p3_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p3_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p3_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p3_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p3_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p3_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_p1_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_p1_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m1_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m1_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m3_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m3_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m3_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m3_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m3_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m3_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m3_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m3_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m3_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m5_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m5_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m5_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m5_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m5_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m5_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m5_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m5_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m5_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m7_p7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_p7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m7_p5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_p5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m7_p3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_p3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m7_p1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_p1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m7_m1, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_m1, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m7_m3, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_m3, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m7_m5, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_m5, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - interference_abs_64qam_epi16(psi_i_m7_m7, ch_mag_int_with_sigma2, two_ch_mag_int_with_sigma2, three_ch_mag_int_with_sigma2, a_i_m7_m7, ONE_OVER_SQRT_2_42, THREE_OVER_SQRT_2_42, FIVE_OVER_SQRT_2_42, - SEVEN_OVER_SQRT_2_42); - - // Calculation of a group of two terms in the bit metric involving product of psi and interference - prodsum_psi_a_epi16(psi_r_p7_p7, a_r_p7_p7, psi_i_p7_p7, a_i_p7_p7, psi_a_p7_p7); - prodsum_psi_a_epi16(psi_r_p7_p5, a_r_p7_p5, psi_i_p7_p5, a_i_p7_p5, psi_a_p7_p5); - prodsum_psi_a_epi16(psi_r_p7_p3, a_r_p7_p3, psi_i_p7_p3, a_i_p7_p3, psi_a_p7_p3); - prodsum_psi_a_epi16(psi_r_p7_p1, a_r_p7_p1, psi_i_p7_p1, a_i_p7_p1, psi_a_p7_p1); - prodsum_psi_a_epi16(psi_r_p7_m1, a_r_p7_m1, psi_i_p7_m1, a_i_p7_m1, psi_a_p7_m1); - prodsum_psi_a_epi16(psi_r_p7_m3, a_r_p7_m3, psi_i_p7_m3, a_i_p7_m3, psi_a_p7_m3); - prodsum_psi_a_epi16(psi_r_p7_m5, a_r_p7_m5, psi_i_p7_m5, a_i_p7_m5, psi_a_p7_m5); - prodsum_psi_a_epi16(psi_r_p7_m7, a_r_p7_m7, psi_i_p7_m7, a_i_p7_m7, psi_a_p7_m7); - prodsum_psi_a_epi16(psi_r_p5_p7, a_r_p5_p7, psi_i_p5_p7, a_i_p5_p7, psi_a_p5_p7); - prodsum_psi_a_epi16(psi_r_p5_p5, a_r_p5_p5, psi_i_p5_p5, a_i_p5_p5, psi_a_p5_p5); - prodsum_psi_a_epi16(psi_r_p5_p3, a_r_p5_p3, psi_i_p5_p3, a_i_p5_p3, psi_a_p5_p3); - prodsum_psi_a_epi16(psi_r_p5_p1, a_r_p5_p1, psi_i_p5_p1, a_i_p5_p1, psi_a_p5_p1); - prodsum_psi_a_epi16(psi_r_p5_m1, a_r_p5_m1, psi_i_p5_m1, a_i_p5_m1, psi_a_p5_m1); - prodsum_psi_a_epi16(psi_r_p5_m3, a_r_p5_m3, psi_i_p5_m3, a_i_p5_m3, psi_a_p5_m3); - prodsum_psi_a_epi16(psi_r_p5_m5, a_r_p5_m5, psi_i_p5_m5, a_i_p5_m5, psi_a_p5_m5); - prodsum_psi_a_epi16(psi_r_p5_m7, a_r_p5_m7, psi_i_p5_m7, a_i_p5_m7, psi_a_p5_m7); - prodsum_psi_a_epi16(psi_r_p3_p7, a_r_p3_p7, psi_i_p3_p7, a_i_p3_p7, psi_a_p3_p7); - prodsum_psi_a_epi16(psi_r_p3_p5, a_r_p3_p5, psi_i_p3_p5, a_i_p3_p5, psi_a_p3_p5); - prodsum_psi_a_epi16(psi_r_p3_p3, a_r_p3_p3, psi_i_p3_p3, a_i_p3_p3, psi_a_p3_p3); - prodsum_psi_a_epi16(psi_r_p3_p1, a_r_p3_p1, psi_i_p3_p1, a_i_p3_p1, psi_a_p3_p1); - prodsum_psi_a_epi16(psi_r_p3_m1, a_r_p3_m1, psi_i_p3_m1, a_i_p3_m1, psi_a_p3_m1); - prodsum_psi_a_epi16(psi_r_p3_m3, a_r_p3_m3, psi_i_p3_m3, a_i_p3_m3, psi_a_p3_m3); - prodsum_psi_a_epi16(psi_r_p3_m5, a_r_p3_m5, psi_i_p3_m5, a_i_p3_m5, psi_a_p3_m5); - prodsum_psi_a_epi16(psi_r_p3_m7, a_r_p3_m7, psi_i_p3_m7, a_i_p3_m7, psi_a_p3_m7); - prodsum_psi_a_epi16(psi_r_p1_p7, a_r_p1_p7, psi_i_p1_p7, a_i_p1_p7, psi_a_p1_p7); - prodsum_psi_a_epi16(psi_r_p1_p5, a_r_p1_p5, psi_i_p1_p5, a_i_p1_p5, psi_a_p1_p5); - prodsum_psi_a_epi16(psi_r_p1_p3, a_r_p1_p3, psi_i_p1_p3, a_i_p1_p3, psi_a_p1_p3); - prodsum_psi_a_epi16(psi_r_p1_p1, a_r_p1_p1, psi_i_p1_p1, a_i_p1_p1, psi_a_p1_p1); - prodsum_psi_a_epi16(psi_r_p1_m1, a_r_p1_m1, psi_i_p1_m1, a_i_p1_m1, psi_a_p1_m1); - prodsum_psi_a_epi16(psi_r_p1_m3, a_r_p1_m3, psi_i_p1_m3, a_i_p1_m3, psi_a_p1_m3); - prodsum_psi_a_epi16(psi_r_p1_m5, a_r_p1_m5, psi_i_p1_m5, a_i_p1_m5, psi_a_p1_m5); - prodsum_psi_a_epi16(psi_r_p1_m7, a_r_p1_m7, psi_i_p1_m7, a_i_p1_m7, psi_a_p1_m7); - prodsum_psi_a_epi16(psi_r_m1_p7, a_r_m1_p7, psi_i_m1_p7, a_i_m1_p7, psi_a_m1_p7); - prodsum_psi_a_epi16(psi_r_m1_p5, a_r_m1_p5, psi_i_m1_p5, a_i_m1_p5, psi_a_m1_p5); - prodsum_psi_a_epi16(psi_r_m1_p3, a_r_m1_p3, psi_i_m1_p3, a_i_m1_p3, psi_a_m1_p3); - prodsum_psi_a_epi16(psi_r_m1_p1, a_r_m1_p1, psi_i_m1_p1, a_i_m1_p1, psi_a_m1_p1); - prodsum_psi_a_epi16(psi_r_m1_m1, a_r_m1_m1, psi_i_m1_m1, a_i_m1_m1, psi_a_m1_m1); - prodsum_psi_a_epi16(psi_r_m1_m3, a_r_m1_m3, psi_i_m1_m3, a_i_m1_m3, psi_a_m1_m3); - prodsum_psi_a_epi16(psi_r_m1_m5, a_r_m1_m5, psi_i_m1_m5, a_i_m1_m5, psi_a_m1_m5); - prodsum_psi_a_epi16(psi_r_m1_m7, a_r_m1_m7, psi_i_m1_m7, a_i_m1_m7, psi_a_m1_m7); - prodsum_psi_a_epi16(psi_r_m3_p7, a_r_m3_p7, psi_i_m3_p7, a_i_m3_p7, psi_a_m3_p7); - prodsum_psi_a_epi16(psi_r_m3_p5, a_r_m3_p5, psi_i_m3_p5, a_i_m3_p5, psi_a_m3_p5); - prodsum_psi_a_epi16(psi_r_m3_p3, a_r_m3_p3, psi_i_m3_p3, a_i_m3_p3, psi_a_m3_p3); - prodsum_psi_a_epi16(psi_r_m3_p1, a_r_m3_p1, psi_i_m3_p1, a_i_m3_p1, psi_a_m3_p1); - prodsum_psi_a_epi16(psi_r_m3_m1, a_r_m3_m1, psi_i_m3_m1, a_i_m3_m1, psi_a_m3_m1); - prodsum_psi_a_epi16(psi_r_m3_m3, a_r_m3_m3, psi_i_m3_m3, a_i_m3_m3, psi_a_m3_m3); - prodsum_psi_a_epi16(psi_r_m3_m5, a_r_m3_m5, psi_i_m3_m5, a_i_m3_m5, psi_a_m3_m5); - prodsum_psi_a_epi16(psi_r_m3_m7, a_r_m3_m7, psi_i_m3_m7, a_i_m3_m7, psi_a_m3_m7); - prodsum_psi_a_epi16(psi_r_m5_p7, a_r_m5_p7, psi_i_m5_p7, a_i_m5_p7, psi_a_m5_p7); - prodsum_psi_a_epi16(psi_r_m5_p5, a_r_m5_p5, psi_i_m5_p5, a_i_m5_p5, psi_a_m5_p5); - prodsum_psi_a_epi16(psi_r_m5_p3, a_r_m5_p3, psi_i_m5_p3, a_i_m5_p3, psi_a_m5_p3); - prodsum_psi_a_epi16(psi_r_m5_p1, a_r_m5_p1, psi_i_m5_p1, a_i_m5_p1, psi_a_m5_p1); - prodsum_psi_a_epi16(psi_r_m5_m1, a_r_m5_m1, psi_i_m5_m1, a_i_m5_m1, psi_a_m5_m1); - prodsum_psi_a_epi16(psi_r_m5_m3, a_r_m5_m3, psi_i_m5_m3, a_i_m5_m3, psi_a_m5_m3); - prodsum_psi_a_epi16(psi_r_m5_m5, a_r_m5_m5, psi_i_m5_m5, a_i_m5_m5, psi_a_m5_m5); - prodsum_psi_a_epi16(psi_r_m5_m7, a_r_m5_m7, psi_i_m5_m7, a_i_m5_m7, psi_a_m5_m7); - prodsum_psi_a_epi16(psi_r_m7_p7, a_r_m7_p7, psi_i_m7_p7, a_i_m7_p7, psi_a_m7_p7); - prodsum_psi_a_epi16(psi_r_m7_p5, a_r_m7_p5, psi_i_m7_p5, a_i_m7_p5, psi_a_m7_p5); - prodsum_psi_a_epi16(psi_r_m7_p3, a_r_m7_p3, psi_i_m7_p3, a_i_m7_p3, psi_a_m7_p3); - prodsum_psi_a_epi16(psi_r_m7_p1, a_r_m7_p1, psi_i_m7_p1, a_i_m7_p1, psi_a_m7_p1); - prodsum_psi_a_epi16(psi_r_m7_m1, a_r_m7_m1, psi_i_m7_m1, a_i_m7_m1, psi_a_m7_m1); - prodsum_psi_a_epi16(psi_r_m7_m3, a_r_m7_m3, psi_i_m7_m3, a_i_m7_m3, psi_a_m7_m3); - prodsum_psi_a_epi16(psi_r_m7_m5, a_r_m7_m5, psi_i_m7_m5, a_i_m7_m5, psi_a_m7_m5); - prodsum_psi_a_epi16(psi_r_m7_m7, a_r_m7_m7, psi_i_m7_m7, a_i_m7_m7, psi_a_m7_m7); - - // Multiply by sqrt(2) - psi_a_p7_p7 = _mm_mulhi_epi16(psi_a_p7_p7, ONE_OVER_SQRT_2); - psi_a_p7_p7 = _mm_slli_epi16(psi_a_p7_p7, 2); - psi_a_p7_p5 = _mm_mulhi_epi16(psi_a_p7_p5, ONE_OVER_SQRT_2); - psi_a_p7_p5 = _mm_slli_epi16(psi_a_p7_p5, 2); - psi_a_p7_p3 = _mm_mulhi_epi16(psi_a_p7_p3, ONE_OVER_SQRT_2); - psi_a_p7_p3 = _mm_slli_epi16(psi_a_p7_p3, 2); - psi_a_p7_p1 = _mm_mulhi_epi16(psi_a_p7_p1, ONE_OVER_SQRT_2); - psi_a_p7_p1 = _mm_slli_epi16(psi_a_p7_p1, 2); - psi_a_p7_m1 = _mm_mulhi_epi16(psi_a_p7_m1, ONE_OVER_SQRT_2); - psi_a_p7_m1 = _mm_slli_epi16(psi_a_p7_m1, 2); - psi_a_p7_m3 = _mm_mulhi_epi16(psi_a_p7_m3, ONE_OVER_SQRT_2); - psi_a_p7_m3 = _mm_slli_epi16(psi_a_p7_m3, 2); - psi_a_p7_m5 = _mm_mulhi_epi16(psi_a_p7_m5, ONE_OVER_SQRT_2); - psi_a_p7_m5 = _mm_slli_epi16(psi_a_p7_m5, 2); - psi_a_p7_m7 = _mm_mulhi_epi16(psi_a_p7_m7, ONE_OVER_SQRT_2); - psi_a_p7_m7 = _mm_slli_epi16(psi_a_p7_m7, 2); - psi_a_p5_p7 = _mm_mulhi_epi16(psi_a_p5_p7, ONE_OVER_SQRT_2); - psi_a_p5_p7 = _mm_slli_epi16(psi_a_p5_p7, 2); - psi_a_p5_p5 = _mm_mulhi_epi16(psi_a_p5_p5, ONE_OVER_SQRT_2); - psi_a_p5_p5 = _mm_slli_epi16(psi_a_p5_p5, 2); - psi_a_p5_p3 = _mm_mulhi_epi16(psi_a_p5_p3, ONE_OVER_SQRT_2); - psi_a_p5_p3 = _mm_slli_epi16(psi_a_p5_p3, 2); - psi_a_p5_p1 = _mm_mulhi_epi16(psi_a_p5_p1, ONE_OVER_SQRT_2); - psi_a_p5_p1 = _mm_slli_epi16(psi_a_p5_p1, 2); - psi_a_p5_m1 = _mm_mulhi_epi16(psi_a_p5_m1, ONE_OVER_SQRT_2); - psi_a_p5_m1 = _mm_slli_epi16(psi_a_p5_m1, 2); - psi_a_p5_m3 = _mm_mulhi_epi16(psi_a_p5_m3, ONE_OVER_SQRT_2); - psi_a_p5_m3 = _mm_slli_epi16(psi_a_p5_m3, 2); - psi_a_p5_m5 = _mm_mulhi_epi16(psi_a_p5_m5, ONE_OVER_SQRT_2); - psi_a_p5_m5 = _mm_slli_epi16(psi_a_p5_m5, 2); - psi_a_p5_m7 = _mm_mulhi_epi16(psi_a_p5_m7, ONE_OVER_SQRT_2); - psi_a_p5_m7 = _mm_slli_epi16(psi_a_p5_m7, 2); - psi_a_p3_p7 = _mm_mulhi_epi16(psi_a_p3_p7, ONE_OVER_SQRT_2); - psi_a_p3_p7 = _mm_slli_epi16(psi_a_p3_p7, 2); - psi_a_p3_p5 = _mm_mulhi_epi16(psi_a_p3_p5, ONE_OVER_SQRT_2); - psi_a_p3_p5 = _mm_slli_epi16(psi_a_p3_p5, 2); - psi_a_p3_p3 = _mm_mulhi_epi16(psi_a_p3_p3, ONE_OVER_SQRT_2); - psi_a_p3_p3 = _mm_slli_epi16(psi_a_p3_p3, 2); - psi_a_p3_p1 = _mm_mulhi_epi16(psi_a_p3_p1, ONE_OVER_SQRT_2); - psi_a_p3_p1 = _mm_slli_epi16(psi_a_p3_p1, 2); - psi_a_p3_m1 = _mm_mulhi_epi16(psi_a_p3_m1, ONE_OVER_SQRT_2); - psi_a_p3_m1 = _mm_slli_epi16(psi_a_p3_m1, 2); - psi_a_p3_m3 = _mm_mulhi_epi16(psi_a_p3_m3, ONE_OVER_SQRT_2); - psi_a_p3_m3 = _mm_slli_epi16(psi_a_p3_m3, 2); - psi_a_p3_m5 = _mm_mulhi_epi16(psi_a_p3_m5, ONE_OVER_SQRT_2); - psi_a_p3_m5 = _mm_slli_epi16(psi_a_p3_m5, 2); - psi_a_p3_m7 = _mm_mulhi_epi16(psi_a_p3_m7, ONE_OVER_SQRT_2); - psi_a_p3_m7 = _mm_slli_epi16(psi_a_p3_m7, 2); - psi_a_p1_p7 = _mm_mulhi_epi16(psi_a_p1_p7, ONE_OVER_SQRT_2); - psi_a_p1_p7 = _mm_slli_epi16(psi_a_p1_p7, 2); - psi_a_p1_p5 = _mm_mulhi_epi16(psi_a_p1_p5, ONE_OVER_SQRT_2); - psi_a_p1_p5 = _mm_slli_epi16(psi_a_p1_p5, 2); - psi_a_p1_p3 = _mm_mulhi_epi16(psi_a_p1_p3, ONE_OVER_SQRT_2); - psi_a_p1_p3 = _mm_slli_epi16(psi_a_p1_p3, 2); - psi_a_p1_p1 = _mm_mulhi_epi16(psi_a_p1_p1, ONE_OVER_SQRT_2); - psi_a_p1_p1 = _mm_slli_epi16(psi_a_p1_p1, 2); - psi_a_p1_m1 = _mm_mulhi_epi16(psi_a_p1_m1, ONE_OVER_SQRT_2); - psi_a_p1_m1 = _mm_slli_epi16(psi_a_p1_m1, 2); - psi_a_p1_m3 = _mm_mulhi_epi16(psi_a_p1_m3, ONE_OVER_SQRT_2); - psi_a_p1_m3 = _mm_slli_epi16(psi_a_p1_m3, 2); - psi_a_p1_m5 = _mm_mulhi_epi16(psi_a_p1_m5, ONE_OVER_SQRT_2); - psi_a_p1_m5 = _mm_slli_epi16(psi_a_p1_m5, 2); - psi_a_p1_m7 = _mm_mulhi_epi16(psi_a_p1_m7, ONE_OVER_SQRT_2); - psi_a_p1_m7 = _mm_slli_epi16(psi_a_p1_m7, 2); - psi_a_m1_p7 = _mm_mulhi_epi16(psi_a_m1_p7, ONE_OVER_SQRT_2); - psi_a_m1_p7 = _mm_slli_epi16(psi_a_m1_p7, 2); - psi_a_m1_p5 = _mm_mulhi_epi16(psi_a_m1_p5, ONE_OVER_SQRT_2); - psi_a_m1_p5 = _mm_slli_epi16(psi_a_m1_p5, 2); - psi_a_m1_p3 = _mm_mulhi_epi16(psi_a_m1_p3, ONE_OVER_SQRT_2); - psi_a_m1_p3 = _mm_slli_epi16(psi_a_m1_p3, 2); - psi_a_m1_p1 = _mm_mulhi_epi16(psi_a_m1_p1, ONE_OVER_SQRT_2); - psi_a_m1_p1 = _mm_slli_epi16(psi_a_m1_p1, 2); - psi_a_m1_m1 = _mm_mulhi_epi16(psi_a_m1_m1, ONE_OVER_SQRT_2); - psi_a_m1_m1 = _mm_slli_epi16(psi_a_m1_m1, 2); - psi_a_m1_m3 = _mm_mulhi_epi16(psi_a_m1_m3, ONE_OVER_SQRT_2); - psi_a_m1_m3 = _mm_slli_epi16(psi_a_m1_m3, 2); - psi_a_m1_m5 = _mm_mulhi_epi16(psi_a_m1_m5, ONE_OVER_SQRT_2); - psi_a_m1_m5 = _mm_slli_epi16(psi_a_m1_m5, 2); - psi_a_m1_m7 = _mm_mulhi_epi16(psi_a_m1_m7, ONE_OVER_SQRT_2); - psi_a_m1_m7 = _mm_slli_epi16(psi_a_m1_m7, 2); - psi_a_m3_p7 = _mm_mulhi_epi16(psi_a_m3_p7, ONE_OVER_SQRT_2); - psi_a_m3_p7 = _mm_slli_epi16(psi_a_m3_p7, 2); - psi_a_m3_p5 = _mm_mulhi_epi16(psi_a_m3_p5, ONE_OVER_SQRT_2); - psi_a_m3_p5 = _mm_slli_epi16(psi_a_m3_p5, 2); - psi_a_m3_p3 = _mm_mulhi_epi16(psi_a_m3_p3, ONE_OVER_SQRT_2); - psi_a_m3_p3 = _mm_slli_epi16(psi_a_m3_p3, 2); - psi_a_m3_p1 = _mm_mulhi_epi16(psi_a_m3_p1, ONE_OVER_SQRT_2); - psi_a_m3_p1 = _mm_slli_epi16(psi_a_m3_p1, 2); - psi_a_m3_m1 = _mm_mulhi_epi16(psi_a_m3_m1, ONE_OVER_SQRT_2); - psi_a_m3_m1 = _mm_slli_epi16(psi_a_m3_m1, 2); - psi_a_m3_m3 = _mm_mulhi_epi16(psi_a_m3_m3, ONE_OVER_SQRT_2); - psi_a_m3_m3 = _mm_slli_epi16(psi_a_m3_m3, 2); - psi_a_m3_m5 = _mm_mulhi_epi16(psi_a_m3_m5, ONE_OVER_SQRT_2); - psi_a_m3_m5 = _mm_slli_epi16(psi_a_m3_m5, 2); - psi_a_m3_m7 = _mm_mulhi_epi16(psi_a_m3_m7, ONE_OVER_SQRT_2); - psi_a_m3_m7 = _mm_slli_epi16(psi_a_m3_m7, 2); - psi_a_m5_p7 = _mm_mulhi_epi16(psi_a_m5_p7, ONE_OVER_SQRT_2); - psi_a_m5_p7 = _mm_slli_epi16(psi_a_m5_p7, 2); - psi_a_m5_p5 = _mm_mulhi_epi16(psi_a_m5_p5, ONE_OVER_SQRT_2); - psi_a_m5_p5 = _mm_slli_epi16(psi_a_m5_p5, 2); - psi_a_m5_p3 = _mm_mulhi_epi16(psi_a_m5_p3, ONE_OVER_SQRT_2); - psi_a_m5_p3 = _mm_slli_epi16(psi_a_m5_p3, 2); - psi_a_m5_p1 = _mm_mulhi_epi16(psi_a_m5_p1, ONE_OVER_SQRT_2); - psi_a_m5_p1 = _mm_slli_epi16(psi_a_m5_p1, 2); - psi_a_m5_m1 = _mm_mulhi_epi16(psi_a_m5_m1, ONE_OVER_SQRT_2); - psi_a_m5_m1 = _mm_slli_epi16(psi_a_m5_m1, 2); - psi_a_m5_m3 = _mm_mulhi_epi16(psi_a_m5_m3, ONE_OVER_SQRT_2); - psi_a_m5_m3 = _mm_slli_epi16(psi_a_m5_m3, 2); - psi_a_m5_m5 = _mm_mulhi_epi16(psi_a_m5_m5, ONE_OVER_SQRT_2); - psi_a_m5_m5 = _mm_slli_epi16(psi_a_m5_m5, 2); - psi_a_m5_m7 = _mm_mulhi_epi16(psi_a_m5_m7, ONE_OVER_SQRT_2); - psi_a_m5_m7 = _mm_slli_epi16(psi_a_m5_m7, 2); - psi_a_m7_p7 = _mm_mulhi_epi16(psi_a_m7_p7, ONE_OVER_SQRT_2); - psi_a_m7_p7 = _mm_slli_epi16(psi_a_m7_p7, 2); - psi_a_m7_p5 = _mm_mulhi_epi16(psi_a_m7_p5, ONE_OVER_SQRT_2); - psi_a_m7_p5 = _mm_slli_epi16(psi_a_m7_p5, 2); - psi_a_m7_p3 = _mm_mulhi_epi16(psi_a_m7_p3, ONE_OVER_SQRT_2); - psi_a_m7_p3 = _mm_slli_epi16(psi_a_m7_p3, 2); - psi_a_m7_p1 = _mm_mulhi_epi16(psi_a_m7_p1, ONE_OVER_SQRT_2); - psi_a_m7_p1 = _mm_slli_epi16(psi_a_m7_p1, 2); - psi_a_m7_m1 = _mm_mulhi_epi16(psi_a_m7_m1, ONE_OVER_SQRT_2); - psi_a_m7_m1 = _mm_slli_epi16(psi_a_m7_m1, 2); - psi_a_m7_m3 = _mm_mulhi_epi16(psi_a_m7_m3, ONE_OVER_SQRT_2); - psi_a_m7_m3 = _mm_slli_epi16(psi_a_m7_m3, 2); - psi_a_m7_m5 = _mm_mulhi_epi16(psi_a_m7_m5, ONE_OVER_SQRT_2); - psi_a_m7_m5 = _mm_slli_epi16(psi_a_m7_m5, 2); - psi_a_m7_m7 = _mm_mulhi_epi16(psi_a_m7_m7, ONE_OVER_SQRT_2); - psi_a_m7_m7 = _mm_slli_epi16(psi_a_m7_m7, 2); - - // Calculation of a group of two terms in the bit metric involving squares of interference - square_a_64qam_epi16(a_r_p7_p7, a_i_p7_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_p7); - square_a_64qam_epi16(a_r_p7_p5, a_i_p7_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_p5); - square_a_64qam_epi16(a_r_p7_p3, a_i_p7_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_p3); - square_a_64qam_epi16(a_r_p7_p1, a_i_p7_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_p1); - square_a_64qam_epi16(a_r_p7_m1, a_i_p7_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_m1); - square_a_64qam_epi16(a_r_p7_m3, a_i_p7_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_m3); - square_a_64qam_epi16(a_r_p7_m5, a_i_p7_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_m5); - square_a_64qam_epi16(a_r_p7_m7, a_i_p7_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p7_m7); - square_a_64qam_epi16(a_r_p5_p7, a_i_p5_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_p7); - square_a_64qam_epi16(a_r_p5_p5, a_i_p5_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_p5); - square_a_64qam_epi16(a_r_p5_p3, a_i_p5_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_p3); - square_a_64qam_epi16(a_r_p5_p1, a_i_p5_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_p1); - square_a_64qam_epi16(a_r_p5_m1, a_i_p5_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_m1); - square_a_64qam_epi16(a_r_p5_m3, a_i_p5_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_m3); - square_a_64qam_epi16(a_r_p5_m5, a_i_p5_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_m5); - square_a_64qam_epi16(a_r_p5_m7, a_i_p5_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p5_m7); - square_a_64qam_epi16(a_r_p3_p7, a_i_p3_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_p7); - square_a_64qam_epi16(a_r_p3_p5, a_i_p3_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_p5); - square_a_64qam_epi16(a_r_p3_p3, a_i_p3_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_p3); - square_a_64qam_epi16(a_r_p3_p1, a_i_p3_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_p1); - square_a_64qam_epi16(a_r_p3_m1, a_i_p3_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_m1); - square_a_64qam_epi16(a_r_p3_m3, a_i_p3_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_m3); - square_a_64qam_epi16(a_r_p3_m5, a_i_p3_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_m5); - square_a_64qam_epi16(a_r_p3_m7, a_i_p3_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p3_m7); - square_a_64qam_epi16(a_r_p1_p7, a_i_p1_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_p7); - square_a_64qam_epi16(a_r_p1_p5, a_i_p1_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_p5); - square_a_64qam_epi16(a_r_p1_p3, a_i_p1_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_p3); - square_a_64qam_epi16(a_r_p1_p1, a_i_p1_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_p1); - square_a_64qam_epi16(a_r_p1_m1, a_i_p1_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_m1); - square_a_64qam_epi16(a_r_p1_m3, a_i_p1_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_m3); - square_a_64qam_epi16(a_r_p1_m5, a_i_p1_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_m5); - square_a_64qam_epi16(a_r_p1_m7, a_i_p1_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_p1_m7); - square_a_64qam_epi16(a_r_m1_p7, a_i_m1_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_p7); - square_a_64qam_epi16(a_r_m1_p5, a_i_m1_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_p5); - square_a_64qam_epi16(a_r_m1_p3, a_i_m1_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_p3); - square_a_64qam_epi16(a_r_m1_p1, a_i_m1_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_p1); - square_a_64qam_epi16(a_r_m1_m1, a_i_m1_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_m1); - square_a_64qam_epi16(a_r_m1_m3, a_i_m1_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_m3); - square_a_64qam_epi16(a_r_m1_m5, a_i_m1_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_m5); - square_a_64qam_epi16(a_r_m1_m7, a_i_m1_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m1_m7); - square_a_64qam_epi16(a_r_m3_p7, a_i_m3_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_p7); - square_a_64qam_epi16(a_r_m3_p5, a_i_m3_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_p5); - square_a_64qam_epi16(a_r_m3_p3, a_i_m3_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_p3); - square_a_64qam_epi16(a_r_m3_p1, a_i_m3_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_p1); - square_a_64qam_epi16(a_r_m3_m1, a_i_m3_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_m1); - square_a_64qam_epi16(a_r_m3_m3, a_i_m3_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_m3); - square_a_64qam_epi16(a_r_m3_m5, a_i_m3_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_m5); - square_a_64qam_epi16(a_r_m3_m7, a_i_m3_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m3_m7); - square_a_64qam_epi16(a_r_m5_p7, a_i_m5_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_p7); - square_a_64qam_epi16(a_r_m5_p5, a_i_m5_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_p5); - square_a_64qam_epi16(a_r_m5_p3, a_i_m5_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_p3); - square_a_64qam_epi16(a_r_m5_p1, a_i_m5_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_p1); - square_a_64qam_epi16(a_r_m5_m1, a_i_m5_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_m1); - square_a_64qam_epi16(a_r_m5_m3, a_i_m5_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_m3); - square_a_64qam_epi16(a_r_m5_m5, a_i_m5_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_m5); - square_a_64qam_epi16(a_r_m5_m7, a_i_m5_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m5_m7); - square_a_64qam_epi16(a_r_m7_p7, a_i_m7_p7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_p7); - square_a_64qam_epi16(a_r_m7_p5, a_i_m7_p5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_p5); - square_a_64qam_epi16(a_r_m7_p3, a_i_m7_p3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_p3); - square_a_64qam_epi16(a_r_m7_p1, a_i_m7_p1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_p1); - square_a_64qam_epi16(a_r_m7_m1, a_i_m7_m1, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_m1); - square_a_64qam_epi16(a_r_m7_m3, a_i_m7_m3, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_m3); - square_a_64qam_epi16(a_r_m7_m5, a_i_m7_m5, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_m5); - square_a_64qam_epi16(a_r_m7_m7, a_i_m7_m7, ch_mag_int, SQRT_42_OVER_FOUR, a_sq_m7_m7); - - // Computing different multiples of ||h0||^2 - // x=1, y=1 - ch_mag_2_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,ONE_OVER_FOUR_SQRT_42); - ch_mag_2_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_2_over_42_with_sigma2,1); - // x=1, y=3 - ch_mag_10_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,FIVE_OVER_FOUR_SQRT_42); - ch_mag_10_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_10_over_42_with_sigma2,1); - // x=1, x=5 - ch_mag_26_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,THIRTEEN_OVER_FOUR_SQRT_42); - ch_mag_26_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_26_over_42_with_sigma2,1); - // x=1, y=7 - ch_mag_50_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42); - ch_mag_50_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_50_over_42_with_sigma2,1); - // x=3, y=3 - ch_mag_18_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,NINE_OVER_FOUR_SQRT_42); - ch_mag_18_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_18_over_42_with_sigma2,1); - // x=3, y=5 - ch_mag_34_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,SEVENTEEN_OVER_FOUR_SQRT_42); - ch_mag_34_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_34_over_42_with_sigma2,1); - // x=3, y=7 - ch_mag_58_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYNINE_OVER_FOUR_SQRT_42); - ch_mag_58_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_58_over_42_with_sigma2,2); - // x=5, y=5 - ch_mag_50_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42); - ch_mag_50_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_50_over_42_with_sigma2,1); - // x=5, y=7 - ch_mag_74_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,THIRTYSEVEN_OVER_FOUR_SQRT_42); - ch_mag_74_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_74_over_42_with_sigma2,2); - // x=7, y=7 - ch_mag_98_over_42_with_sigma2 = _mm_mulhi_epi16(ch_mag_des,FORTYNINE_OVER_FOUR_SQRT_42); - ch_mag_98_over_42_with_sigma2 = _mm_slli_epi16(ch_mag_98_over_42_with_sigma2,2); - - // Computing Metrics - xmm0 = _mm_subs_epi16(psi_a_p7_p7, a_sq_p7_p7); - xmm1 = _mm_adds_epi16(xmm0, y0_p_7_7); - bit_met_p7_p7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_p5, a_sq_p7_p5); - xmm1 = _mm_adds_epi16(xmm0, y0_p_7_5); - bit_met_p7_p5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_p3, a_sq_p7_p3); - xmm1 = _mm_adds_epi16(xmm0, y0_p_7_3); - bit_met_p7_p3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_p1, a_sq_p7_p1); - xmm1 = _mm_adds_epi16(xmm0, y0_p_7_1); - bit_met_p7_p1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_m1, a_sq_p7_m1); - xmm1 = _mm_adds_epi16(xmm0, y0_m_7_1); - bit_met_p7_m1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_m3, a_sq_p7_m3); - xmm1 = _mm_adds_epi16(xmm0, y0_m_7_3); - bit_met_p7_m3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_m5, a_sq_p7_m5); - xmm1 = _mm_adds_epi16(xmm0, y0_m_7_5); - bit_met_p7_m5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p7_m7, a_sq_p7_m7); - xmm1 = _mm_adds_epi16(xmm0, y0_m_7_7); - bit_met_p7_m7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_p7, a_sq_p5_p7); - xmm1 = _mm_adds_epi16(xmm0, y0_p_5_7); - bit_met_p5_p7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_p5, a_sq_p5_p5); - xmm1 = _mm_adds_epi16(xmm0, y0_p_5_5); - bit_met_p5_p5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_p3, a_sq_p5_p3); - xmm1 = _mm_adds_epi16(xmm0, y0_p_5_3); - bit_met_p5_p3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_p1, a_sq_p5_p1); - xmm1 = _mm_adds_epi16(xmm0, y0_p_5_1); - bit_met_p5_p1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_m1, a_sq_p5_m1); - xmm1 = _mm_adds_epi16(xmm0, y0_m_5_1); - bit_met_p5_m1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_m3, a_sq_p5_m3); - xmm1 = _mm_adds_epi16(xmm0, y0_m_5_3); - bit_met_p5_m3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_m5, a_sq_p5_m5); - xmm1 = _mm_adds_epi16(xmm0, y0_m_5_5); - bit_met_p5_m5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p5_m7, a_sq_p5_m7); - xmm1 = _mm_adds_epi16(xmm0, y0_m_5_7); - bit_met_p5_m7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_p7, a_sq_p3_p7); - xmm1 = _mm_adds_epi16(xmm0, y0_p_3_7); - bit_met_p3_p7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_p5, a_sq_p3_p5); - xmm1 = _mm_adds_epi16(xmm0, y0_p_3_5); - bit_met_p3_p5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_p3, a_sq_p3_p3); - xmm1 = _mm_adds_epi16(xmm0, y0_p_3_3); - bit_met_p3_p3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_p1, a_sq_p3_p1); - xmm1 = _mm_adds_epi16(xmm0, y0_p_3_1); - bit_met_p3_p1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_m1, a_sq_p3_m1); - xmm1 = _mm_adds_epi16(xmm0, y0_m_3_1); - bit_met_p3_m1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_m3, a_sq_p3_m3); - xmm1 = _mm_adds_epi16(xmm0, y0_m_3_3); - bit_met_p3_m3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_m5, a_sq_p3_m5); - xmm1 = _mm_adds_epi16(xmm0, y0_m_3_5); - bit_met_p3_m5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p3_m7, a_sq_p3_m7); - xmm1 = _mm_adds_epi16(xmm0, y0_m_3_7); - bit_met_p3_m7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_p7, a_sq_p1_p7); - xmm1 = _mm_adds_epi16(xmm0, y0_p_1_7); - bit_met_p1_p7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_p5, a_sq_p1_p5); - xmm1 = _mm_adds_epi16(xmm0, y0_p_1_5); - bit_met_p1_p5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_p3, a_sq_p1_p3); - xmm1 = _mm_adds_epi16(xmm0, y0_p_1_3); - bit_met_p1_p3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_p1, a_sq_p1_p1); - xmm1 = _mm_adds_epi16(xmm0, y0_p_1_1); - bit_met_p1_p1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_m1, a_sq_p1_m1); - xmm1 = _mm_adds_epi16(xmm0, y0_m_1_1); - bit_met_p1_m1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_m3, a_sq_p1_m3); - xmm1 = _mm_adds_epi16(xmm0, y0_m_1_3); - bit_met_p1_m3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_m5, a_sq_p1_m5); - xmm1 = _mm_adds_epi16(xmm0, y0_m_1_5); - bit_met_p1_m5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_p1_m7, a_sq_p1_m7); - xmm1 = _mm_adds_epi16(xmm0, y0_m_1_7); - bit_met_p1_m7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - - xmm0 = _mm_subs_epi16(psi_a_m1_p7, a_sq_m1_p7); - xmm1 = _mm_subs_epi16(xmm0, y0_m_1_7); - bit_met_m1_p7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_p5, a_sq_m1_p5); - xmm1 = _mm_subs_epi16(xmm0, y0_m_1_5); - bit_met_m1_p5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_p3, a_sq_m1_p3); - xmm1 = _mm_subs_epi16(xmm0, y0_m_1_3); - bit_met_m1_p3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_p1, a_sq_m1_p1); - xmm1 = _mm_subs_epi16(xmm0, y0_m_1_1); - bit_met_m1_p1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_m1, a_sq_m1_m1); - xmm1 = _mm_subs_epi16(xmm0, y0_p_1_1); - bit_met_m1_m1 = _mm_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_m3, a_sq_m1_m3); - xmm1 = _mm_subs_epi16(xmm0, y0_p_1_3); - bit_met_m1_m3 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_m5, a_sq_m1_m5); - xmm1 = _mm_subs_epi16(xmm0, y0_p_1_5); - bit_met_m1_m5 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m1_m7, a_sq_m1_m7); - xmm1 = _mm_subs_epi16(xmm0, y0_p_1_7); - bit_met_m1_m7 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_p7, a_sq_m3_p7); - xmm1 = _mm_subs_epi16(xmm0, y0_m_3_7); - bit_met_m3_p7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_p5, a_sq_m3_p5); - xmm1 = _mm_subs_epi16(xmm0, y0_m_3_5); - bit_met_m3_p5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_p3, a_sq_m3_p3); - xmm1 = _mm_subs_epi16(xmm0, y0_m_3_3); - bit_met_m3_p3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_p1, a_sq_m3_p1); - xmm1 = _mm_subs_epi16(xmm0, y0_m_3_1); - bit_met_m3_p1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_m1, a_sq_m3_m1); - xmm1 = _mm_subs_epi16(xmm0, y0_p_3_1); - bit_met_m3_m1 = _mm_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_m3, a_sq_m3_m3); - xmm1 = _mm_subs_epi16(xmm0, y0_p_3_3); - bit_met_m3_m3 = _mm_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_m5, a_sq_m3_m5); - xmm1 = _mm_subs_epi16(xmm0, y0_p_3_5); - bit_met_m3_m5 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m3_m7, a_sq_m3_m7); - xmm1 = _mm_subs_epi16(xmm0, y0_p_3_7); - bit_met_m3_m7 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_p7, a_sq_m5_p7); - xmm1 = _mm_subs_epi16(xmm0, y0_m_5_7); - bit_met_m5_p7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_p5, a_sq_m5_p5); - xmm1 = _mm_subs_epi16(xmm0, y0_m_5_5); - bit_met_m5_p5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_p3, a_sq_m5_p3); - xmm1 = _mm_subs_epi16(xmm0, y0_m_5_3); - bit_met_m5_p3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_p1, a_sq_m5_p1); - xmm1 = _mm_subs_epi16(xmm0, y0_m_5_1); - bit_met_m5_p1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_m1, a_sq_m5_m1); - xmm1 = _mm_subs_epi16(xmm0, y0_p_5_1); - bit_met_m5_m1 = _mm_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_m3, a_sq_m5_m3); - xmm1 = _mm_subs_epi16(xmm0, y0_p_5_3); - bit_met_m5_m3 = _mm_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_m5, a_sq_m5_m5); - xmm1 = _mm_subs_epi16(xmm0, y0_p_5_5); - bit_met_m5_m5 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m5_m7, a_sq_m5_m7); - xmm1 = _mm_subs_epi16(xmm0, y0_p_5_7); - bit_met_m5_m7 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_p7, a_sq_m7_p7); - xmm1 = _mm_subs_epi16(xmm0, y0_m_7_7); - bit_met_m7_p7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_p5, a_sq_m7_p5); - xmm1 = _mm_subs_epi16(xmm0, y0_m_7_5); - bit_met_m7_p5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_p3, a_sq_m7_p3); - xmm1 = _mm_subs_epi16(xmm0, y0_m_7_3); - bit_met_m7_p3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_p1, a_sq_m7_p1); - xmm1 = _mm_subs_epi16(xmm0, y0_m_7_1); - bit_met_m7_p1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_m1, a_sq_m7_m1); - xmm1 = _mm_subs_epi16(xmm0, y0_p_7_1); - bit_met_m7_m1 = _mm_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_m3, a_sq_m7_m3); - xmm1 = _mm_subs_epi16(xmm0, y0_p_7_3); - bit_met_m7_m3 = _mm_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_m5, a_sq_m7_m5); - xmm1 = _mm_subs_epi16(xmm0, y0_p_7_5); - bit_met_m7_m5 = _mm_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2); - xmm0 = _mm_subs_epi16(psi_a_m7_m7, a_sq_m7_m7); - xmm1 = _mm_subs_epi16(xmm0, y0_p_7_7); - bit_met_m7_m7 = _mm_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2); - - // Detection for 1st bit (LTE mapping) - // bit = 1 - xmm0 = _mm_max_epi16(bit_met_m7_p7, bit_met_m7_p5); - xmm1 = _mm_max_epi16(bit_met_m7_p3, bit_met_m7_p1); - xmm2 = _mm_max_epi16(bit_met_m7_m1, bit_met_m7_m3); - xmm3 = _mm_max_epi16(bit_met_m7_m5, bit_met_m7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m5_p7, bit_met_m5_p5); - xmm1 = _mm_max_epi16(bit_met_m5_p3, bit_met_m5_p1); - xmm2 = _mm_max_epi16(bit_met_m5_m1, bit_met_m5_m3); - xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m5_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_m3_p7, bit_met_m3_p5); - xmm1 = _mm_max_epi16(bit_met_m3_p3, bit_met_m3_p1); - xmm2 = _mm_max_epi16(bit_met_m3_m1, bit_met_m3_m3); - xmm3 = _mm_max_epi16(bit_met_m3_m5, bit_met_m3_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_m1_p7, bit_met_m1_p5); - xmm1 = _mm_max_epi16(bit_met_m1_p3, bit_met_m1_p1); - xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m1_m3); - xmm3 = _mm_max_epi16(bit_met_m1_m5, bit_met_m1_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - // bit = 0 - xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p7_p5); - xmm1 = _mm_max_epi16(bit_met_p7_p3, bit_met_p7_p1); - xmm2 = _mm_max_epi16(bit_met_p7_m1, bit_met_p7_m3); - xmm3 = _mm_max_epi16(bit_met_p7_m5, bit_met_p7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p5_p7, bit_met_p5_p5); - xmm1 = _mm_max_epi16(bit_met_p5_p3, bit_met_p5_p1); - xmm2 = _mm_max_epi16(bit_met_p5_m1, bit_met_p5_m3); - xmm3 = _mm_max_epi16(bit_met_p5_m5, bit_met_p5_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p3_p7, bit_met_p3_p5); - xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p3_p1); - xmm2 = _mm_max_epi16(bit_met_p3_m1, bit_met_p3_m3); - xmm3 = _mm_max_epi16(bit_met_p3_m5, bit_met_p3_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p1_p7, bit_met_p1_p5); - xmm1 = _mm_max_epi16(bit_met_p1_p3, bit_met_p1_p1); - xmm2 = _mm_max_epi16(bit_met_p1_m1, bit_met_p1_m3); - xmm3 = _mm_max_epi16(bit_met_p1_m5, bit_met_p1_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y0r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - // Detection for 2nd bit (LTE mapping) - // bit = 1 - xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); - xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); - xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); - xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); - xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); - xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); - xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); - xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); - xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); - xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - // bit = 0 - xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); - xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); - xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m7_p1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); - xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); - xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); - xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); - xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m7_p5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); - xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); - xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); - xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y1r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - // Detection for 3rd bit (LTE mapping) - xmm0 = _mm_max_epi16(bit_met_m7_m7, bit_met_m7_m5); - xmm1 = _mm_max_epi16(bit_met_m7_m3, bit_met_m7_m1); - xmm2 = _mm_max_epi16(bit_met_m7_p1, bit_met_m7_p3); - xmm3 = _mm_max_epi16(bit_met_m7_p5, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m5_m7, bit_met_m5_m5); - xmm1 = _mm_max_epi16(bit_met_m5_m3, bit_met_m5_m1); - xmm2 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p5_m7, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p5_m3, bit_met_p5_m1); - xmm2 = _mm_max_epi16(bit_met_p5_p1, bit_met_p5_p3); - xmm3 = _mm_max_epi16(bit_met_p5_p5, bit_met_p5_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p7_m5); - xmm1 = _mm_max_epi16(bit_met_p7_m3, bit_met_p7_m1); - xmm2 = _mm_max_epi16(bit_met_p7_p1, bit_met_p7_p3); - xmm3 = _mm_max_epi16(bit_met_p7_p5, bit_met_p7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - xmm0 = _mm_max_epi16(bit_met_m3_m7, bit_met_m3_m5); - xmm1 = _mm_max_epi16(bit_met_m3_m3, bit_met_m3_m1); - xmm2 = _mm_max_epi16(bit_met_m3_p1, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m3_p5, bit_met_m3_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m1_m7, bit_met_m1_m5); - xmm1 = _mm_max_epi16(bit_met_m1_m3, bit_met_m1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m1_p3); - xmm3 = _mm_max_epi16(bit_met_m1_p5, bit_met_m1_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p1_m7, bit_met_p1_m5); - xmm1 = _mm_max_epi16(bit_met_p1_m3, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_p1_p1, bit_met_p1_p3); - xmm3 = _mm_max_epi16(bit_met_p1_p5, bit_met_p1_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p3_m7, bit_met_p3_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p3_m1); - xmm2 = _mm_max_epi16(bit_met_p3_p1, bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_p3_p5, bit_met_p3_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y2r = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - // Detection for 4th bit (LTE mapping) - xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); - xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); - xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); - xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); - xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); - xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); - xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); - xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); - xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); - xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); - xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); - xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); - xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); - xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); - xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); - xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); - xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m7_p1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); - xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); - xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y0i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - - // Detection for 5th bit (LTE mapping) - xmm0 = _mm_max_epi16(bit_met_m7_m7, bit_met_m7_m5); - xmm1 = _mm_max_epi16(bit_met_m7_m3, bit_met_m7_m1); - xmm2 = _mm_max_epi16(bit_met_m7_p1, bit_met_m7_p3); - xmm3 = _mm_max_epi16(bit_met_m7_p5, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m1_m7, bit_met_m1_m5); - xmm1 = _mm_max_epi16(bit_met_m1_m3, bit_met_m1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m1_p3); - xmm3 = _mm_max_epi16(bit_met_m1_p5, bit_met_m1_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p1_m7, bit_met_p1_m5); - xmm1 = _mm_max_epi16(bit_met_p1_m3, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_p1_p1, bit_met_p1_p3); - xmm3 = _mm_max_epi16(bit_met_p1_p5, bit_met_p1_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p7_m5); - xmm1 = _mm_max_epi16(bit_met_p7_m3, bit_met_p7_m1); - xmm2 = _mm_max_epi16(bit_met_p7_p1, bit_met_p7_p3); - xmm3 = _mm_max_epi16(bit_met_p7_p5, bit_met_p7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - xmm0 = _mm_max_epi16(bit_met_m5_m7, bit_met_m5_m5); - xmm1 = _mm_max_epi16(bit_met_m5_m3, bit_met_m5_m1); - xmm2 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m5_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_m3_m7, bit_met_m3_m5); - xmm1 = _mm_max_epi16(bit_met_m3_m3, bit_met_m3_m1); - xmm2 = _mm_max_epi16(bit_met_m3_p1, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m3_p5, bit_met_m3_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p3_m7, bit_met_p3_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p3_m1); - xmm2 = _mm_max_epi16(bit_met_p3_p1, bit_met_p3_p3); - xmm3 = _mm_max_epi16(bit_met_p3_p5, bit_met_p3_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p5_m7, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p5_m3, bit_met_p5_m1); - xmm2 = _mm_max_epi16(bit_met_p5_p1, bit_met_p5_p3); - xmm3 = _mm_max_epi16(bit_met_p5_p5, bit_met_p5_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y1i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - // Detection for 6th bit (LTE mapping) - xmm0 = _mm_max_epi16(bit_met_p7_p7, bit_met_p5_p7); - xmm1 = _mm_max_epi16(bit_met_p3_p7, bit_met_p1_p7); - xmm2 = _mm_max_epi16(bit_met_m1_p7, bit_met_m3_p7); - xmm3 = _mm_max_epi16(bit_met_m5_p7, bit_met_m7_p7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p1, bit_met_p5_p1); - xmm1 = _mm_max_epi16(bit_met_p3_p1, bit_met_p1_p1); - xmm2 = _mm_max_epi16(bit_met_m1_p1, bit_met_m3_p1); - xmm3 = _mm_max_epi16(bit_met_m5_p1, bit_met_m5_p1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m1, bit_met_p5_m1); - xmm1 = _mm_max_epi16(bit_met_p3_m1, bit_met_p1_m1); - xmm2 = _mm_max_epi16(bit_met_m1_m1, bit_met_m3_m1); - xmm3 = _mm_max_epi16(bit_met_m5_m1, bit_met_m7_m1); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m7, bit_met_p5_m7); - xmm1 = _mm_max_epi16(bit_met_p3_m7, bit_met_p1_m7); - xmm2 = _mm_max_epi16(bit_met_m1_m7, bit_met_m3_m7); - xmm3 = _mm_max_epi16(bit_met_m5_m7, bit_met_m7_m7); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm4); - logmax_den_re0 = _mm_max_epi16(logmax_den_re0, xmm5); - - xmm0 = _mm_max_epi16(bit_met_p7_m5, bit_met_p5_m5); - xmm1 = _mm_max_epi16(bit_met_p3_m5, bit_met_p1_m5); - xmm2 = _mm_max_epi16(bit_met_m1_m5, bit_met_m3_m5); - xmm3 = _mm_max_epi16(bit_met_m5_m5, bit_met_m7_m5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(xmm4, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_m3, bit_met_p5_m3); - xmm1 = _mm_max_epi16(bit_met_p3_m3, bit_met_p1_m3); - xmm2 = _mm_max_epi16(bit_met_m1_m3, bit_met_m3_m3); - xmm3 = _mm_max_epi16(bit_met_m5_m3, bit_met_m7_m3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p3, bit_met_p5_p3); - xmm1 = _mm_max_epi16(bit_met_p3_p3, bit_met_p1_p3); - xmm2 = _mm_max_epi16(bit_met_m1_p3, bit_met_m3_p3); - xmm3 = _mm_max_epi16(bit_met_m5_p3, bit_met_m7_p3); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - xmm0 = _mm_max_epi16(bit_met_p7_p5, bit_met_p5_p5); - xmm1 = _mm_max_epi16(bit_met_p3_p5, bit_met_p1_p5); - xmm2 = _mm_max_epi16(bit_met_m1_p5, bit_met_m3_p5); - xmm3 = _mm_max_epi16(bit_met_m5_p5, bit_met_m7_p5); - xmm4 = _mm_max_epi16(xmm0, xmm1); - xmm5 = _mm_max_epi16(xmm2, xmm3); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm4); - logmax_num_re0 = _mm_max_epi16(logmax_num_re0, xmm5); - - y2i = _mm_subs_epi16(logmax_num_re0, logmax_den_re0); - - - // map to output stream, difficult to do in SIMD since we have 6 16bit LLRs - // RE 1 - j = 24*i; - stream0_out[j + 0] = ((short *)&y0r)[0]; - stream0_out[j + 1] = ((short *)&y1r)[0]; - stream0_out[j + 2] = ((short *)&y2r)[0]; - stream0_out[j + 3] = ((short *)&y0i)[0]; - stream0_out[j + 4] = ((short *)&y1i)[0]; - stream0_out[j + 5] = ((short *)&y2i)[0]; - // RE 2 - stream0_out[j + 6] = ((short *)&y0r)[1]; - stream0_out[j + 7] = ((short *)&y1r)[1]; - stream0_out[j + 8] = ((short *)&y2r)[1]; - stream0_out[j + 9] = ((short *)&y0i)[1]; - stream0_out[j + 10] = ((short *)&y1i)[1]; - stream0_out[j + 11] = ((short *)&y2i)[1]; - // RE 3 - stream0_out[j + 12] = ((short *)&y0r)[2]; - stream0_out[j + 13] = ((short *)&y1r)[2]; - stream0_out[j + 14] = ((short *)&y2r)[2]; - stream0_out[j + 15] = ((short *)&y0i)[2]; - stream0_out[j + 16] = ((short *)&y1i)[2]; - stream0_out[j + 17] = ((short *)&y2i)[2]; - // RE 4 - stream0_out[j + 18] = ((short *)&y0r)[3]; - stream0_out[j + 19] = ((short *)&y1r)[3]; - stream0_out[j + 20] = ((short *)&y2r)[3]; - stream0_out[j + 21] = ((short *)&y0i)[3]; - stream0_out[j + 22] = ((short *)&y1i)[3]; - stream0_out[j + 23] = ((short *)&y2i)[3]; - // RE 5 - stream0_out[j + 24] = ((short *)&y0r)[4]; - stream0_out[j + 25] = ((short *)&y1r)[4]; - stream0_out[j + 26] = ((short *)&y2r)[4]; - stream0_out[j + 27] = ((short *)&y0i)[4]; - stream0_out[j + 28] = ((short *)&y1i)[4]; - stream0_out[j + 29] = ((short *)&y2i)[4]; - // RE 6 - stream0_out[j + 30] = ((short *)&y0r)[5]; - stream0_out[j + 31] = ((short *)&y1r)[5]; - stream0_out[j + 32] = ((short *)&y2r)[5]; - stream0_out[j + 33] = ((short *)&y0i)[5]; - stream0_out[j + 34] = ((short *)&y1i)[5]; - stream0_out[j + 35] = ((short *)&y2i)[5]; - // RE 7 - stream0_out[j + 36] = ((short *)&y0r)[6]; - stream0_out[j + 37] = ((short *)&y1r)[6]; - stream0_out[j + 38] = ((short *)&y2r)[6]; - stream0_out[j + 39] = ((short *)&y0i)[6]; - stream0_out[j + 40] = ((short *)&y1i)[6]; - stream0_out[j + 41] = ((short *)&y2i)[6]; - // RE 8 - stream0_out[j + 42] = ((short *)&y0r)[7]; - stream0_out[j + 43] = ((short *)&y1r)[7]; - stream0_out[j + 44] = ((short *)&y2r)[7]; - stream0_out[j + 45] = ((short *)&y0i)[7]; - stream0_out[j + 46] = ((short *)&y1i)[7]; - stream0_out[j + 47] = ((short *)&y2i)[7]; - -#elif defined(__arm__) || defined(__aarch64__) - -#endif - - } - -#if defined(__x86_64__) || defined(__i386__) - _mm_empty(); - _m_empty(); -#endif -} -#endif - - -int nr_dlsch_64qam_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, - int32_t **rxdataF_comp, - int32_t **rxdataF_comp_i, - int32_t **dl_ch_mag, - int32_t **dl_ch_mag_i, - int32_t **rho_i, - int16_t *dlsch_llr, - uint8_t symbol, - uint32_t len, - uint8_t first_symbol_flag, - uint16_t nb_rb, - uint16_t pbch_pss_sss_adjust, - //int16_t **llr16p, - uint32_t llr_offset) -{ - - int16_t *rxF = (int16_t*)&rxdataF_comp[0][(symbol*nb_rb*12)]; - int16_t *rxF_i = (int16_t*)&rxdataF_comp_i[0][(symbol*nb_rb*12)]; - int16_t *ch_mag = (int16_t*)&dl_ch_mag[0][(symbol*nb_rb*12)]; - int16_t *ch_mag_i = (int16_t*)&dl_ch_mag_i[0][(symbol*nb_rb*12)]; - int16_t *rho = (int16_t*)&rho_i[0][(symbol*nb_rb*12)]; - int16_t *llr16; - int8_t *pllr_symbol; // pointer where llrs should filled for this ofdm symbol - - //first symbol has different structure due to more pilots - /*if (first_symbol_flag == 1) { - llr16 = (int16_t*)dlsch_llr; - } else { - llr16 = (int16_t*)(*llr16p); - }*/ - - llr16 = (int16_t*)dlsch_llr; - - AssertFatal(llr16!=NULL,"nr_dlsch_16qam_64qam_llr:llr is null, symbol %d\n",symbol); - - - pllr_symbol = (int8_t*)dlsch_llr; - pllr_symbol += llr_offset; - //printf("nr_dlsch_64qam_64qam_llr: symbol %d,nb_rb %d, len %d,pbch_pss_sss_adjust %d\n",symbol,nb_rb,len,pbch_pss_sss_adjust); - /*LOG_I(PHY,"nr_dlsch_64qam_64qam_llr [symb %d / FirstSym %d / Length %d / LLR Offset %d]: @LLR Buff %x, @LLR Buff(symb) %x, , @Compute LLR Buff(symb) %x \n", - symbol, - first_symbol_flag, - len, - llr_offset, - (int16_t*)dlsch_llr, - llr16, - pllr_symbol);*/ - - // Round length up to multiple of 16 words - uint32_t len256i = ((len+16)>>4)*16; - int32_t *rxF_256i = (int32_t*) malloc16_clear(len256i*4); - int32_t *rxF_i_256i = (int32_t*) malloc16_clear(len256i*4); - int32_t *ch_mag_256i = (int32_t*) malloc16_clear(len256i*4); - int32_t *ch_mag_i_256i = (int32_t*) malloc16_clear(len256i*4); - int32_t *rho_256i = (int32_t*) malloc16_clear(len256i*4); - - memcpy(rxF_256i, rxF, len*4); - memcpy(rxF_i_256i, rxF_i, len*4); - memcpy(ch_mag_256i, ch_mag, len*4); - memcpy(ch_mag_i_256i, ch_mag_i, len*4); - memcpy(rho_256i, rho, len*4); - -#if 0 - qam64_qam16_avx2((short *)rxF_256i, - (short *)rxF_i_256i, - (short *)ch_mag_256i, - (short *)ch_mag_i_256i, - (short *)llr16, - (short *) rho_256i, - len); -#else - qam64_qam64_avx2((int32_t *)rxF_256i, - (int32_t *)rxF_i_256i, - (int32_t *)ch_mag_256i, - (int32_t *)ch_mag_i_256i, - (int16_t *)llr16, - (int32_t *) rho_256i, - len); -#endif - - free16(rxF_256i, sizeof(rxF_256i)); - free16(rxF_i_256i, sizeof(rxF_i_256i)); - free16(ch_mag_256i, sizeof(ch_mag_256i)); - free16(ch_mag_i_256i, sizeof(ch_mag_i_256i)); - free16(rho_256i, sizeof(rho_256i)); - - llr16 += (6*len); - //*llr16p = (short *)llr16; - - return(0); -} diff --git a/openair1/PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h b/openair1/PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h index 746c4ff5ca..7432c82aac 100644 --- a/openair1/PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h +++ b/openair1/PHY/NR_UE_TRANSPORT/nr_transport_proto_ue.h @@ -87,361 +87,6 @@ int32_t nr_dlsch_qpsk_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, uint16_t pbch_pss_sss_adj, int16_t **llr128p); -/** \brief This function computes the LLRs for ML (max-logsum approximation) dual-stream QPSK/16QAM reception. - @param stream0_in Input from channel compensated (MR combined) stream 0 - @param stream1_in Input from channel compensated (MR combined) stream 1 - @param ch_mag_i Input from scaled channel magnitude square of h0'*g1 - @param stream0_out Output from LLR unit for stream0 - @param rho01 Cross-correlation between channels (MR combined) - @param length in complex channel outputs*/ -void nr_qpsk_qam16(int16_t *stream0_in, - int16_t *stream1_in, - short *ch_mag_i, - int16_t *stream0_out, - int16_t *rho01, - int32_t length); - -/** \brief This function perform LLR computation for dual-stream (QPSK/16QAM) transmission. - @param frame_parms Frame descriptor structure - @param rxdataF_comp Compensated channel output - @param rxdataF_comp_i Compensated channel output for interference - @param rho_i Correlation between channel of signal and inteference - @param dlsch_llr llr output - @param symbol OFDM symbol index in sub-frame - @param first_symbol_flag flag to indicate this is the first symbol of the dlsch - @param nb_rb number of RBs for this allocation - @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS - @param llr128p pointer to pointer to symbol in dlsch_llr*/ -int32_t nr_dlsch_qpsk_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, - int32_t **rxdataF_comp, - int32_t **rxdataF_comp_i, - int **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) - int32_t **rho_i, - int16_t *dlsch_llr, - uint8_t symbol, - uint8_t first_symbol_flag, - uint16_t nb_rb, - uint16_t pbch_pss_sss_adj, - int16_t **llr128p); - - -/** \brief This function computes the LLRs for ML (max-logsum approximation) dual-stream QPSK/64QAM reception. - @param stream0_in Input from channel compensated (MR combined) stream 0 - @param stream1_in Input from channel compensated (MR combined) stream 1 - @param ch_mag_i Input from scaled channel magnitude square of h0'*g1 - @param stream0_out Output from LLR unit for stream0 - @param rho01 Cross-correlation between channels (MR combined) - @param length in complex channel outputs*/ -void nr_qpsk_qam64(int16_t *stream0_in, - int16_t *stream1_in, - short *ch_mag_i, - int16_t *stream0_out, - int16_t *rho01, - int32_t length); - -/** \brief This function perform LLR computation for dual-stream (QPSK/64QAM) transmission. - @param frame_parms Frame descriptor structure - @param rxdataF_comp Compensated channel output - @param rxdataF_comp_i Compensated channel output for interference - @param rho_i Correlation between channel of signal and inteference - @param dlsch_llr llr output - @param symbol OFDM symbol index in sub-frame - @param first_symbol_flag flag to indicate this is the first symbol of the dlsch - @param nb_rb number of RBs for this allocation - @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS - @param llr128p pointer to pointer to symbol in dlsch_llr*/ -int32_t nr_dlsch_qpsk_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, - int32_t **rxdataF_comp, - int32_t **rxdataF_comp_i, - int **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) - int32_t **rho_i, - int16_t *dlsch_llr, - uint8_t symbol, - uint8_t first_symbol_flag, - uint16_t nb_rb, - uint16_t pbch_pss_sss_adj, - int16_t **llr128p); - - -/** \brief This function computes the LLRs for ML (max-logsum approximation) dual-stream 16QAM/QPSK reception. - @param stream0_in Input from channel compensated (MR combined) stream 0 - @param stream1_in Input from channel compensated (MR combined) stream 1 - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param stream0_out Output from LLR unit for stream0 - @param rho01 Cross-correlation between channels (MR combined) - @param length in complex channel outputs*/ -void nr_qam16_qpsk(short *stream0_in, - short *stream1_in, - short *ch_mag, - short *stream0_out, - short *rho01, - int length); -/** \brief This function perform LLR computation for dual-stream (16QAM/QPSK) transmission. - @param frame_parms Frame descriptor structure - @param rxdataF_comp Compensated channel output - @param rxdataF_comp_i Compensated channel output for interference - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param rho_i Correlation between channel of signal and inteference - @param dlsch_llr llr output - @param symbol OFDM symbol index in sub-frame - @param first_symbol_flag flag to indicate this is the first symbol of the dlsch - @param nb_rb number of RBs for this allocation - @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS - @param llr16p pointer to pointer to symbol in dlsch_llr*/ -int nr_dlsch_16qam_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, - int **rxdataF_comp, - int **rxdataF_comp_i, - int **dl_ch_mag, //|h_0|^2*(2/sqrt{10}) - int **rho_i, - short *dlsch_llr, - unsigned char symbol, - unsigned char first_symbol_flag, - unsigned short nb_rb, - uint16_t pbch_pss_sss_adjust, - short **llr16p); - -/** \brief This function computes the LLRs for ML (max-logsum approximation) dual-stream 16QAM/16QAM reception. - @param stream0_in Input from channel compensated (MR combined) stream 0 - @param stream1_in Input from channel compensated (MR combined) stream 1 - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param ch_mag_i Input from scaled channel magnitude square of h0'*g1 - @param stream0_out Output from LLR unit for stream0 - @param rho01 Cross-correlation between channels (MR combined) - @param length in complex channel outputs*/ -void nr_qam16_qam16(short *stream0_in, - short *stream1_in, - short *ch_mag, - short *ch_mag_i, - short *stream0_out, - short *rho01, - int length); - -/** \brief This function perform LLR computation for dual-stream (16QAM/16QAM) transmission. - @param frame_parms Frame descriptor structure - @param rxdataF_comp Compensated channel output - @param rxdataF_comp_i Compensated channel output for interference - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param ch_mag_i Input from scaled channel magnitude square of h0'*g1 - @param rho_i Correlation between channel of signal and inteference - @param dlsch_llr llr output - @param symbol OFDM symbol index in sub-frame - @param first_symbol_flag flag to indicate this is the first symbol of the dlsch - @param nb_rb number of RBs for this allocation - @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS - @param llr16p pointer to pointer to symbol in dlsch_llr*/ -int nr_dlsch_16qam_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, - int **rxdataF_comp, - int **rxdataF_comp_i, - int **dl_ch_mag, //|h_0|^2*(2/sqrt{10}) - int **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) - int **rho_i, - short *dlsch_llr, - unsigned char symbol, - uint32_t len, - unsigned char first_symbol_flag, - unsigned short nb_rb, - uint16_t pbch_pss_sss_adjust, - short **llr16p); - -/** \brief This function computes the LLRs for ML (max-logsum approximation) dual-stream 16QAM/64QAM reception. - @param stream0_in Input from channel compensated (MR combined) stream 0 - @param stream1_in Input from channel compensated (MR combined) stream 1 - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param ch_mag_i Input from scaled channel magnitude square of h0'*g1 - @param stream0_out Output from LLR unit for stream0 - @param rho01 Cross-correlation between channels (MR combined) - @param length in complex channel outputs*/ -void nr_qam16_qam64(short *stream0_in, - short *stream1_in, - short *ch_mag, - short *ch_mag_i, - short *stream0_out, - short *rho01, - int length); - -/** \brief This function perform LLR computation for dual-stream (16QAM/64QAM) transmission. - @param frame_parms Frame descriptor structure - @param rxdataF_comp Compensated channel output - @param rxdataF_comp_i Compensated channel output for interference - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param ch_mag_i Input from scaled channel magnitude square of h0'*g1 - @param rho_i Correlation between channel of signal and inteference - @param dlsch_llr llr output - @param symbol OFDM symbol index in sub-frame - @param first_symbol_flag flag to indicate this is the first symbol of the dlsch - @param nb_rb number of RBs for this allocation - @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS - @param llr16p pointer to pointer to symbol in dlsch_llr*/ -int nr_dlsch_16qam_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, - int **rxdataF_comp, - int **rxdataF_comp_i, - int **dl_ch_mag, //|h_0|^2*(2/sqrt{10}) - int **dl_ch_mag_i, //|h_1|^2*(2/sqrt{10}) - int **rho_i, - short *dlsch_llr, - unsigned char symbol, - unsigned char first_symbol_flag, - unsigned short nb_rb, - uint16_t pbch_pss_sss_adjust, - short **llr16p); - -/** \brief This function computes the LLRs for ML (max-logsum approximation) dual-stream 64QAM/64QAM reception. - @param stream0_in Input from channel compensated (MR combined) stream 0 - @param stream1_in Input from channel compensated (MR combined) stream 1 - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param stream0_out Output from LLR unit for stream0 - @param rho01 Cross-correlation between channels (MR combined) - @param length in complex channel outputs*/ -void nr_qam64_qpsk(short *stream0_in, - short *stream1_in, - short *ch_mag, - short *stream0_out, - short *rho01, - int length); - -/** \brief This function perform LLR computation for dual-stream (64QAM/64QAM) transmission. - @param frame_parms Frame descriptor structure - @param rxdataF_comp Compensated channel output - @param rxdataF_comp_i Compensated channel output for interference - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param rho_i Correlation between channel of signal and inteference - @param dlsch_llr llr output - @param symbol OFDM symbol index in sub-frame - @param first_symbol_flag flag to indicate this is the first symbol of the dlsch - @param nb_rb number of RBs for this allocation - @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS - @param llr16p pointer to pointer to symbol in dlsch_llr*/ -int nr_dlsch_64qam_qpsk_llr(NR_DL_FRAME_PARMS *frame_parms, - int **rxdataF_comp, - int **rxdataF_comp_i, - int **dl_ch_mag, - int **rho_i, - short *dlsch_llr, - unsigned char symbol, - unsigned char first_symbol_flag, - unsigned short nb_rb, - uint16_t pbch_pss_sss_adjust, - short **llr16p); - -/** \brief This function computes the LLRs for ML (max-logsum approximation) dual-stream 64QAM/16QAM reception. - @param stream0_in Input from channel compensated (MR combined) stream 0 - @param stream1_in Input from channel compensated (MR combined) stream 1 - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param ch_mag_i Input from scaled channel magnitude square of h0'*g1 - @param stream0_out Output from LLR unit for stream0 - @param rho01 Cross-correlation between channels (MR combined) - @param length in complex channel outputs*/ -void nr_qam64_qam16(short *stream0_in, - short *stream1_in, - short *ch_mag, - short *ch_mag_i, - short *stream0_out, - short *rho01, - int length); - -/** \brief This function computes the LLRs for ML (max-logsum approximation) dual-stream 64QAM/16QAM reception. - @param stream0_in Input from channel compensated (MR combined) stream 0 - @param stream1_in Input from channel compensated (MR combined) stream 1 - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param ch_mag_i Input from scaled channel magnitude square of h0'*g1 - @param stream0_out Output from LLR unit for stream0 - @param rho01 Cross-correlation between channels (MR combined) - @param length in complex channel outputs*/ -void qam64_qam16_avx2(short *stream0_in, - short *stream1_in, - short *ch_mag, - short *ch_mag_i, - short *stream0_out, - short *rho01, - int length); - -/** \brief This function perform LLR computation for dual-stream (64QAM/16QAM) transmission. - @param frame_parms Frame descriptor structure - @param rxdataF_comp Compensated channel output - @param rxdataF_comp_i Compensated channel output for interference - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param ch_mag_i Input from scaled channel magnitude square of h0'*g1 - @param rho_i Correlation between channel of signal and inteference - @param dlsch_llr llr output - @param symbol OFDM symbol index in sub-frame - @param first_symbol_flag flag to indicate this is the first symbol of the dlsch - @param nb_rb number of RBs for this allocation - @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS - @param llr16p pointer to pointer to symbol in dlsch_llr*/ -int nr_dlsch_64qam_16qam_llr(NR_DL_FRAME_PARMS *frame_parms, - int **rxdataF_comp, - int **rxdataF_comp_i, - int **dl_ch_mag, - int **dl_ch_mag_i, - int **rho_i, - short *dlsch_llr, - unsigned char symbol, - unsigned char first_symbol_flag, - unsigned short nb_rb, - uint16_t pbch_pss_sss_adjust, - short **llr16p); - -/** \brief This function computes the LLRs for ML (max-logsum approximation) dual-stream 64QAM/64QAM reception. - @param stream0_in Input from channel compensated (MR combined) stream 0 - @param stream1_in Input from channel compensated (MR combined) stream 1 - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param ch_mag_i Input from scaled channel magnitude square of h0'*g1 - @param stream0_out Output from LLR unit for stream0 - @param rho01 Cross-correlation between channels (MR combined) - @param length in complex channel outputs*/ -void qam64_qam64(short *stream0_in, - short *stream1_in, - short *ch_mag, - short *ch_mag_i, - short *stream0_out, - short *rho01, - int length); - -/** \brief This function computes the LLRs for ML (max-logsum approximation) dual-stream 64QAM/64QAM reception. - @param stream0_in Input from channel compensated (MR combined) stream 0 - @param stream1_in Input from channel compensated (MR combined) stream 1 - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param ch_mag_i Input from scaled channel magnitude square of h0'*g1 - @param stream0_out Output from LLR unit for stream0 - @param rho01 Cross-correlation between channels (MR combined) - @param length in complex channel outputs*/ -void qam64_qam64_avx2(int32_t *stream0_in, - int32_t *stream1_in, - int32_t *ch_mag, - int32_t *ch_mag_i, - int16_t *stream0_out, - int32_t *rho01, - int length); - -/** \brief This function perform LLR computation for dual-stream (64QAM/64QAM) transmission. - @param frame_parms Frame descriptor structure - @param rxdataF_comp Compensated channel output - @param rxdataF_comp_i Compensated channel output for interference - @param ch_mag Input from scaled channel magnitude square of h0'*g0 - @param ch_mag_i Input from scaled channel magnitude square of h0'*g1 - @param rho_i Correlation between channel of signal and inteference - @param dlsch_llr llr output - @param symbol OFDM symbol index in sub-frame - @param first_symbol_flag flag to indicate this is the first symbol of the dlsch - @param nb_rb number of RBs for this allocation - @param pbch_pss_sss_adj Number of channel bits taken by PBCH/PSS/SSS - @param llr16p pointer to pointer to symbol in dlsch_llr*/ -int nr_dlsch_64qam_64qam_llr(NR_DL_FRAME_PARMS *frame_parms, - int **rxdataF_comp, - int **rxdataF_comp_i, - int **dl_ch_mag, - int **dl_ch_mag_i, - int **rho_i, - short *dlsch_llr, - unsigned char symbol, - uint32_t len, - unsigned char first_symbol_flag, - unsigned short nb_rb, - uint16_t pbch_pss_sss_adjust, - //short **llr16p, - uint32_t llr_offset); - - /** \brief This function generates log-likelihood ratios (decoder input) for single-stream QPSK received waveforms. @param frame_parms Frame descriptor structure @param rxdataF_comp Compensated channel output @@ -518,95 +163,6 @@ void nr_dlsch_256qam_llr(NR_DL_FRAME_PARMS *frame_parms, uint8_t first_symbol_flag, uint16_t nb_rb); - -/** \fn dlsch_extract_rbs(int32_t **rxdataF, - int32_t **dl_ch_estimates, - int32_t **rxdataF_ext, - int32_t **dl_ch_estimates_ext, - unsigned char symbol - uint8_t pilots, - uint8_t config_type, - unsigned short start_rb, - unsigned short nb_rb_pdsch, - uint8_t n_dmrs_cdm_groups, - uint8_t Nl, - NR_DL_FRAME_PARMS *frame_parms, - uint16_t dlDmrsSymbPos) - \brief This function extracts the received resource blocks, both channel estimates and data symbols, - for the current allocation and for multiple layer antenna gNB transmission. - @param rxdataF Raw FFT output of received signal - @param dl_ch_estimates Channel estimates of current slot - @param rxdataF_ext FFT output for RBs in this allocation - @param dl_ch_estimates_ext Channel estimates for RBs in this allocation - @param Nl nb of antenna layers - @param symbol Symbol to extract - @param n_dmrs_cdm_groups - @param frame_parms Pointer to frame descriptor -*/ -void nr_dlsch_extract_rbs(uint32_t rxdataF_sz, - c16_t rxdataF[][rxdataF_sz], - const uint32_t rx_size, - const uint32_t pdsch_est_size, - int32_t dl_ch_estimates[][pdsch_est_size], - int32_t rxdataF_ext[][rx_size], - int32_t dl_ch_estimates_ext[][rx_size], - unsigned char symbol, - uint8_t pilots, - uint8_t config_type, - unsigned short start_rb, - unsigned short nb_rb_pdsch, - uint8_t n_dmrs_cdm_groups, - uint8_t Nl, - NR_DL_FRAME_PARMS *frame_parms, - uint16_t dlDmrsSymbPos, - int chest_time_type); - -/** \brief This function performs channel compensation (matched filtering) on the received RBs for this allocation. In addition, it computes the squared-magnitude of the channel with weightings for 16QAM/64QAM detection as well as dual-stream detection (cross-correlation) - @param rxdataF_ext Frequency-domain received signal in RBs to be demodulated - @param dl_ch_estimates_ext Frequency-domain channel estimates in RBs to be demodulated - @param dl_ch_mag First Channel magnitudes (16QAM/64QAM) - @param dl_ch_magb Second weighted Channel magnitudes (64QAM) - @param rxdataF_comp Compensated received waveform - @param rho Cross-correlation between two spatial channels on each RX antenna - @param frame_parms Pointer to frame descriptor - @param symbol Symbol on which to operate - @param first_symbol_flag set to 1 on first DLSCH symbol - @param mod_order Modulation order of allocation - @param nb_rb Number of RBs in allocation - @param output_shift Rescaling for compensated output (should be energy-normalizing) - @param phy_measurements Pointer to UE PHY measurements -*/ -void nr_dlsch_channel_compensation(uint32_t rx_size, - int32_t rxdataF_ext[][rx_size], - int32_t dl_ch_estimates_ext[][rx_size], - int32_t dl_ch_mag[][rx_size], - int32_t dl_ch_magb[][rx_size], - int32_t dl_ch_magr[][rx_size], - int32_t rxdataF_comp[][rx_size], - int ***rho, - NR_DL_FRAME_PARMS *frame_parms, - uint8_t nb_aatx, - unsigned char symbol, - int length, - uint8_t first_symbol_flag, - unsigned char mod_order, - unsigned short nb_rb, - unsigned char output_shift, - PHY_NR_MEASUREMENTS *measurements); - -void nr_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 nr_dlsch_deinterleaving(uint8_t symbol, uint8_t start_symbol, uint16_t L, @@ -614,63 +170,12 @@ void nr_dlsch_deinterleaving(uint8_t symbol, uint16_t *llr_deint, uint16_t nb_rb_pdsch); -void nr_dlsch_channel_level_median(uint32_t rx_size, - int32_t dl_ch_estimates[][rx_size], - int32_t *median, - int n_tx, - int n_rx, - int length, - int start_point); - -void nr_dlsch_detection_mrc(uint32_t rx_size, - int32_t rxdataF_comp[][rx_size], - int ***rho, - int32_t dl_ch_mag[][rx_size], - int32_t dl_ch_magb[][rx_size], - int32_t dl_ch_magr[][rx_size], - short n_tx, - short n_rx, - unsigned char symbol, - unsigned short nb_rb, - int length); - void nr_conjch0_mult_ch1(int *ch0, int *ch1, int32_t *ch0conj_ch1, unsigned short nb_rb, unsigned char output_shift0); -void nr_a_sum_b(__m128i *input_x, - __m128i *input_y, - unsigned short nb_rb); - - -/** \brief This function computes the average channel level over all allocated RBs and antennas (TX/RX) in order to compute output shift for compensated signal - @param dl_ch_estimates_ext Channel estimates in allocated RBs - @param frame_parms Pointer to frame descriptor - @param avg Pointer to average signal strength - @param pilots_flag Flag to indicate pilots in symbol - @param nb_rb Number of allocated RBs -*/ -void nr_dlsch_channel_level(uint32_t rx_size, - int32_t dl_ch_estimates[][rx_size], - NR_DL_FRAME_PARMS *frame_parms, - uint8_t n_tx, - int32_t *avg, - uint8_t symbol, - uint32_t len, - unsigned short nb_rb); - -void nr_dlsch_scale_channel(uint32_t rx_size, - int32_t dl_ch_estimates[][rx_size], - NR_DL_FRAME_PARMS *frame_parms, - uint8_t n_tx, - uint8_t n_rx, - uint8_t symbol, - uint8_t pilots, - uint32_t len, - unsigned short nb_rb); - /** \brief This is the top-level entry point for DLSCH decoding in UE. It should be replicated on several threads (on multi-core machines) corresponding to different HARQ processes. The routine first computes the segmentation information, followed by rate dematching and sub-block deinterleaving the of the @@ -877,23 +382,20 @@ int nr_rx_pdsch(PHY_VARS_NR_UE *ue, uint32_t pdsch_est_size, int32_t dl_ch_estimates[][pdsch_est_size], int16_t *llr[2], - c16_t ptrs_phase_per_slot[][NR_SYMBOLS_PER_SLOT], - int32_t ptrs_re_per_slot[][NR_SYMBOLS_PER_SLOT], uint32_t dl_valid_re[NR_SYMBOLS_PER_SLOT], - uint32_t rx_size, - int32_t dl_ch_estimates_ext[][rx_size], - int32_t rxdataF_ext[][rx_size], - int32_t rxdataF_comp[][rx_size], - int32_t dl_ch_mag[][rx_size], - int32_t dl_ch_magb[][rx_size], - int32_t dl_ch_magr[][rx_size], c16_t rxdataF[][ue->frame_parms.samples_per_slot_wCP], uint32_t llr_offset[NR_SYMBOLS_PER_SLOT], - int32_t *log2_maxh); + int32_t *log2_maxhrx_size_symbol, + int rx_size_symbol, + int nbRx, + int32_t rxdataF_comp[][nbRx][rx_size_symbol], + c16_t ptrs_phase_per_slot[][NR_SYMBOLS_PER_SLOT], + int32_t ptrs_re_per_slot[][NR_SYMBOLS_PER_SLOT]); int32_t generate_nr_prach(PHY_VARS_NR_UE *ue, uint8_t gNB_id, int frame, uint8_t slot); void dump_nrdlsch(PHY_VARS_NR_UE *ue,uint8_t gNB_id,uint8_t nr_slot_rx,unsigned int *coded_bits_per_codeword,int round, unsigned char harq_pid); +void nr_a_sum_b(c16_t *input_x, c16_t *input_y, unsigned short nb_rb); /**@}*/ #endif diff --git a/openair1/PHY/TOOLS/tools_defs.h b/openair1/PHY/TOOLS/tools_defs.h index 62945456c3..aa3ea6fc41 100644 --- a/openair1/PHY/TOOLS/tools_defs.h +++ b/openair1/PHY/TOOLS/tools_defs.h @@ -97,6 +97,20 @@ extern "C" { #define squaredMod(a) ((a).r*(a).r + (a).i*(a).i) #define csum(res, i1, i2) (res).r = (i1).r + (i2).r ; (res).i = (i1).i + (i2).i + __attribute__((always_inline)) inline c16_t c16Shift(const c16_t a, const int Shift) { + return (c16_t) { + .r = (int16_t)(a.r >> Shift), + .i = (int16_t)(a.i >> Shift) + }; + } + + __attribute__((always_inline)) inline c16_t c16addShift(const c16_t a, const c16_t b, const int Shift) { + return (c16_t) { + .r = (int16_t)((a.r + b.r) >> Shift), + .i = (int16_t)((a.i + b.i) >> Shift) + }; + } + __attribute__((always_inline)) inline c16_t c16mulShift(const c16_t a, const c16_t b, const int Shift) { return (c16_t) { .r = (int16_t)((a.r * b.r - a.i * b.i) >> Shift), diff --git a/openair1/SCHED_NR_UE/phy_procedures_nr_ue.c b/openair1/SCHED_NR_UE/phy_procedures_nr_ue.c index a68411c4cf..57b3dff776 100644 --- a/openair1/SCHED_NR_UE/phy_procedures_nr_ue.c +++ b/openair1/SCHED_NR_UE/phy_procedures_nr_ue.c @@ -577,9 +577,19 @@ int nr_ue_pdsch_procedures(PHY_VARS_NR_UE *ue, LOG_D(PHY,"[UE %d] nr_slot_rx %d, harq_pid %d (%d), rb_start %d, nb_rb %d, symbol_start %d, nb_symbols %d, DMRS mask %x, Nl %d\n", ue->Mod_id,nr_slot_rx,harq_pid,dlsch0_harq->status,pdsch_start_rb,pdsch_nb_rb,s0,s1,dlsch0->dlsch_config.dlDmrsSymbPos, dlsch0->Nl); - const uint32_t pdsch_est_size = ((ue->frame_parms.symbols_per_slot*ue->frame_parms.ofdm_symbol_size+15)/16)*16; - __attribute__ ((aligned(32))) int32_t pdsch_dl_ch_estimates[ue->frame_parms.nb_antennas_rx*dlsch0->Nl][pdsch_est_size]; - memset(pdsch_dl_ch_estimates, 0, sizeof(int32_t)*ue->frame_parms.nb_antennas_rx*dlsch0->Nl*pdsch_est_size); + const uint32_t pdsch_est_size = ((ue->frame_parms.symbols_per_slot * ue->frame_parms.ofdm_symbol_size + 15) / 16) * 16; + __attribute__((aligned(32))) int32_t pdsch_dl_ch_estimates[ue->frame_parms.nb_antennas_rx * dlsch0->Nl][pdsch_est_size]; + memset(pdsch_dl_ch_estimates, 0, sizeof(int32_t) * ue->frame_parms.nb_antennas_rx * dlsch0->Nl * pdsch_est_size); + + c16_t ptrs_phase_per_slot[ue->frame_parms.nb_antennas_rx][NR_SYMBOLS_PER_SLOT]; + memset(ptrs_phase_per_slot, 0, sizeof(ptrs_phase_per_slot)); + + int32_t ptrs_re_per_slot[ue->frame_parms.nb_antennas_rx][NR_SYMBOLS_PER_SLOT]; + memset(ptrs_re_per_slot, 0, sizeof(ptrs_re_per_slot)); + + const uint32_t rx_size_symbol = dlsch[0].dlsch_config.number_rbs * NR_NB_SC_PER_RB; + __attribute__((aligned(32))) int32_t rxdataF_comp[ue->frame_parms.nb_antennas_tx][ue->frame_parms.nb_antennas_rx][rx_size_symbol * NR_SYMBOLS_PER_SLOT]; + memset(rxdataF_comp, 0, sizeof(rxdataF_comp)); for (m = s0; m < (s0 +s1); m++) { if (dlsch0->dlsch_config.dlDmrsSymbPos & (1 << m)) { @@ -598,7 +608,7 @@ int nr_ue_pdsch_procedures(PHY_VARS_NR_UE *ue, pdsch_nb_rb, pdsch_est_size, pdsch_dl_ch_estimates, - rxdataF); + ue->frame_parms.samples_per_slot_wCP, rxdataF); #if 0 ///LOG_M: the channel estimation int nr_frame_rx = proc->frame_rx; @@ -635,35 +645,9 @@ int nr_ue_pdsch_procedures(PHY_VARS_NR_UE *ue, first_symbol_with_data++; } - c16_t ptrs_phase_per_slot[ue->frame_parms.nb_antennas_rx][NR_SYMBOLS_PER_SLOT]; - memset(ptrs_phase_per_slot, 0, ue->frame_parms.nb_antennas_rx*NR_SYMBOLS_PER_SLOT*sizeof(c16_t)); - - int32_t ptrs_re_per_slot[ue->frame_parms.nb_antennas_rx][NR_SYMBOLS_PER_SLOT]; - memset(ptrs_re_per_slot, 0, ue->frame_parms.nb_antennas_rx*NR_SYMBOLS_PER_SLOT*sizeof(c16_t)); - uint32_t dl_valid_re[NR_SYMBOLS_PER_SLOT] = {0}; uint32_t llr_offset[NR_SYMBOLS_PER_SLOT] = {0}; - const uint32_t rx_size = ((NR_SYMBOLS_PER_SLOT * dlsch[0].dlsch_config.number_rbs * NR_NB_SC_PER_RB + 15) >> 4) << 4; - - __attribute__ ((aligned(32))) int32_t dl_ch_estimates_ext[ue->frame_parms.nb_antennas_rx*dlsch0->Nl][rx_size]; - memset(dl_ch_estimates_ext, 0, ue->frame_parms.nb_antennas_rx*dlsch0->Nl*rx_size*sizeof(int32_t)); - - __attribute__ ((aligned(32))) int32_t rxdataF_ext[ue->frame_parms.nb_antennas_rx*dlsch0->Nl][rx_size]; - memset(rxdataF_ext, 0, ue->frame_parms.nb_antennas_rx*dlsch0->Nl*rx_size*sizeof(int32_t)); - - __attribute__ ((aligned(32))) int32_t rxdataF_comp[ue->frame_parms.nb_antennas_rx*dlsch0->Nl][rx_size]; - memset(rxdataF_comp, 0, ue->frame_parms.nb_antennas_rx*dlsch0->Nl*rx_size*sizeof(int32_t)); - - __attribute__ ((aligned(32))) int32_t dl_ch_mag[ue->frame_parms.nb_antennas_rx*dlsch0->Nl][rx_size]; - memset(dl_ch_mag, 0, ue->frame_parms.nb_antennas_rx*dlsch0->Nl*rx_size*sizeof(int32_t)); - - __attribute__ ((aligned(32))) int32_t dl_ch_magb[ue->frame_parms.nb_antennas_rx*dlsch0->Nl][rx_size]; - memset(dl_ch_magb, 0, ue->frame_parms.nb_antennas_rx*dlsch0->Nl*rx_size*sizeof(int32_t)); - - __attribute__ ((aligned(32))) int32_t dl_ch_magr[ue->frame_parms.nb_antennas_rx*dlsch0->Nl][rx_size]; - memset(dl_ch_magr, 0, ue->frame_parms.nb_antennas_rx*dlsch0->Nl*rx_size*sizeof(int32_t)); - int32_t log2_maxh = 0; start_meas(&ue->rx_pdsch_stats); for (m = s0; m < (s1 + s0); m++) { @@ -688,19 +672,16 @@ int nr_ue_pdsch_procedures(PHY_VARS_NR_UE *ue, pdsch_est_size, pdsch_dl_ch_estimates, llr, - ptrs_phase_per_slot, - ptrs_re_per_slot, dl_valid_re, - rx_size, - dl_ch_estimates_ext, - rxdataF_ext, - rxdataF_comp, - dl_ch_mag, - dl_ch_magb, - dl_ch_magr, rxdataF, llr_offset, - &log2_maxh) < 0) + &log2_maxh, + rx_size_symbol, + ue->frame_parms.nb_antennas_rx, + rxdataF_comp, + ptrs_phase_per_slot, + ptrs_re_per_slot) + < 0) return -1; stop_meas(&ue->dlsch_llr_stats_parallelization[slot]); @@ -711,17 +692,6 @@ int nr_ue_pdsch_procedures(PHY_VARS_NR_UE *ue, } } // CRNTI active stop_meas(&ue->rx_pdsch_stats); - - UEscopeCopy(ue, pdschRxdataF_comp, rxdataF_comp, sizeof(struct complex16), ue->frame_parms.nb_antennas_rx*dlsch0->Nl, rx_size); - - if (ue->phy_sim_pdsch_rxdataF_comp) - memcpy(ue->phy_sim_pdsch_rxdataF_comp, rxdataF_comp, sizeof(int32_t)*rx_size*ue->frame_parms.nb_antennas_rx*dlsch0->Nl); - if (ue->phy_sim_pdsch_rxdataF_ext) - memcpy(ue->phy_sim_pdsch_rxdataF_ext, rxdataF_ext, sizeof(int32_t)*rx_size*ue->frame_parms.nb_antennas_rx*dlsch0->Nl); - if (ue->phy_sim_pdsch_dl_ch_estimates_ext) - memcpy(ue->phy_sim_pdsch_dl_ch_estimates_ext, dl_ch_estimates_ext, sizeof(int32_t)*rx_size*ue->frame_parms.nb_antennas_rx*dlsch0->Nl); - if (ue->phy_sim_pdsch_dl_ch_estimates) - memcpy(ue->phy_sim_pdsch_dl_ch_estimates, dl_ch_estimates_ext, sizeof(int32_t)*rx_size*ue->frame_parms.nb_antennas_rx*dlsch0->Nl); } return 0; } @@ -909,7 +879,6 @@ bool nr_ue_dlsch_procedures(PHY_VARS_NR_UE *ue, const double N_TA_max = Ta_max * bw_scaling * tc_factor; NR_UE_MAC_INST_t *mac = get_mac_inst(0); - NR_BWP_Id_t dl_bwp = mac->current_DL_BWP.bwp_id; NR_BWP_Id_t ul_bwp = mac->current_UL_BWP.bwp_id; NR_PUSCH_TimeDomainResourceAllocationList_t *pusch_TimeDomainAllocationList = NULL; @@ -938,17 +907,8 @@ bool nr_ue_dlsch_procedures(PHY_VARS_NR_UE *ue, } long mapping_type_ul = pusch_TimeDomainAllocationList ? pusch_TimeDomainAllocationList->list.array[0]->mappingType : NR_PUSCH_TimeDomainResourceAllocation__mappingType_typeA; - NR_PDSCH_Config_t *pdsch_Config = NULL; - NR_PDSCH_TimeDomainResourceAllocationList_t *pdsch_TimeDomainAllocationList = NULL; - if(dl_bwp){ - pdsch_Config = (mac->DLbwp[dl_bwp-1] && mac->DLbwp[dl_bwp-1]->bwp_Dedicated->pdsch_Config->choice.setup) ? mac->DLbwp[dl_bwp-1]->bwp_Dedicated->pdsch_Config->choice.setup : NULL; - if (mac->DLbwp[dl_bwp-1] && mac->DLbwp[dl_bwp-1]->bwp_Dedicated->pdsch_Config->choice.setup->pdsch_TimeDomainAllocationList) - pdsch_TimeDomainAllocationList = pdsch_Config->pdsch_TimeDomainAllocationList->choice.setup; - else if (mac->DLbwp[dl_bwp-1] && mac->DLbwp[dl_bwp-1]->bwp_Common->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList) - pdsch_TimeDomainAllocationList = mac->DLbwp[dl_bwp-1]->bwp_Common->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList; - } - else if (mac->scc_SIB && mac->scc_SIB->downlinkConfigCommon.initialDownlinkBWP.pdsch_ConfigCommon->choice.setup) - pdsch_TimeDomainAllocationList = mac->scc_SIB->downlinkConfigCommon.initialDownlinkBWP.pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList; + NR_PDSCH_Config_t *pdsch_Config = mac->current_DL_BWP.pdsch_Config; + NR_PDSCH_TimeDomainResourceAllocationList_t *pdsch_TimeDomainAllocationList = mac->current_DL_BWP.tdaList_Common; long mapping_type_dl = pdsch_TimeDomainAllocationList ? pdsch_TimeDomainAllocationList->list.array[0]->mappingType : NR_PDSCH_TimeDomainResourceAllocation__mappingType_typeA; NR_DMRS_DownlinkConfig_t *NR_DMRS_dlconfig = NULL; diff --git a/openair1/SIMULATION/NR_PHY/dlsim.c b/openair1/SIMULATION/NR_PHY/dlsim.c index 1fe4229c04..ed171523de 100644 --- a/openair1/SIMULATION/NR_PHY/dlsim.c +++ b/openair1/SIMULATION/NR_PHY/dlsim.c @@ -256,9 +256,10 @@ void nr_dlsim_preprocessor(module_id_t module_id, * configuration */ current_BWP->mcsTableIdx = g_mcsTableIdx; sched_pdsch->time_domain_allocation = get_dl_tda(RC.nrmac[module_id], scc, slot); - AssertFatal(sched_pdsch->time_domain_allocation>=0,"Unable to find PDSCH time domain allocation in list\n"); + AssertFatal(sched_pdsch->time_domain_allocation >= 0,"Unable to find PDSCH time domain allocation in list\n"); - sched_pdsch->tda_info = nr_get_pdsch_tda_info(current_BWP, sched_pdsch->time_domain_allocation); + sched_pdsch->tda_info = get_dl_tda_info(current_BWP, sched_ctrl->search_space->searchSpaceType->present, sched_pdsch->time_domain_allocation, + NR_MIB__dmrs_TypeA_Position_pos2, 1, NR_RNTI_C, sched_ctrl->coreset->controlResourceSetId, false); sched_pdsch->dmrs_parms = get_dl_dmrs_params(scc, current_BWP, @@ -309,6 +310,7 @@ int NB_UE_INST = 1; int main(int argc, char **argv) { + setbuf(stdout, NULL); char c; int i,aa;//,l; double sigma2, sigma2_dB=10, SNR, snr0=-2.0, snr1=2.0; @@ -1219,9 +1221,11 @@ int main(int argc, char **argv) uint8_t nb_symb_sch = rel15->NrOfSymbols; available_bits = nr_get_G(nb_rb, nb_symb_sch, nb_re_dmrs, length_dmrs, mod_order, rel15->nrOfLayers); - if(pdu_bit_map & 0x1) { - available_bits-= (ptrsSymbPerSlot * ptrsRePerSymb *rel15->nrOfLayers* 2); - printf("[DLSIM][PTRS] Available bits are: %5u, removed PTRS bits are: %5u \n",available_bits, (ptrsSymbPerSlot * ptrsRePerSymb *rel15->nrOfLayers* 2) ); + if (pdu_bit_map & 0x1) { + available_bits -= (ptrsSymbPerSlot * ptrsRePerSymb * rel15->nrOfLayers * 2); + if (trial == 0 && round == 0) { + printf("[DLSIM][PTRS] Available bits are: %5u, removed PTRS bits are: %5u \n", available_bits, (ptrsSymbPerSlot * ptrsRePerSymb * rel15->nrOfLayers * 2)); + } } for (i = 0; i < available_bits; i++) { diff --git a/openair2/COMMON/f1ap_messages_types.h b/openair2/COMMON/f1ap_messages_types.h index 159af160ea..d8aac0fff1 100644 --- a/openair2/COMMON/f1ap_messages_types.h +++ b/openair2/COMMON/f1ap_messages_types.h @@ -356,7 +356,7 @@ typedef struct cu_to_du_rrc_information_s { uint32_t measConfig_length; }cu_to_du_rrc_information_t; -typedef struct du_to_du_rrc_information_s { +typedef struct du_to_cu_rrc_information_s { uint8_t * cellGroupConfig; uint32_t cellGroupConfig_length; uint8_t * measGapConfig; diff --git a/openair2/LAYER2/NR_MAC_COMMON/nr_mac.h b/openair2/LAYER2/NR_MAC_COMMON/nr_mac.h index fec7e90178..4af6eb02e6 100644 --- a/openair2/LAYER2/NR_MAC_COMMON/nr_mac.h +++ b/openair2/LAYER2/NR_MAC_COMMON/nr_mac.h @@ -574,9 +574,10 @@ typedef struct NR_UE_DL_BWP { uint16_t BWPStart; uint16_t initial_BWPSize; uint16_t initial_BWPStart; - NR_PDSCH_TimeDomainResourceAllocationList_t *tdaList; + NR_PDSCH_TimeDomainResourceAllocationList_t *tdaList_Common; NR_PDSCH_Config_t *pdsch_Config; NR_PDSCH_ServingCellConfig_t *pdsch_servingcellconfig; + long *pdsch_HARQ_ACK_Codebook; uint8_t mcsTableIdx; nr_dci_format_t dci_format; } NR_UE_DL_BWP_t; @@ -590,18 +591,40 @@ typedef struct NR_UE_UL_BWP { uint16_t BWPStart; uint16_t initial_BWPSize; uint16_t initial_BWPStart; + NR_RACH_ConfigCommon_t *rach_ConfigCommon; NR_PUSCH_ServingCellConfig_t *pusch_servingcellconfig; - NR_PUSCH_TimeDomainResourceAllocationList_t *tdaList; + NR_PUSCH_TimeDomainResourceAllocationList_t *tdaList_Common; + NR_ConfiguredGrantConfig_t *configuredGrantConfig; NR_PUSCH_Config_t *pusch_Config; NR_PUCCH_Config_t *pucch_Config; NR_PUCCH_ConfigCommon_t *pucch_ConfigCommon; + long *harq_ACK_SpatialBundlingPUCCH; NR_CSI_MeasConfig_t *csi_MeasConfig; NR_SRS_Config_t *srs_Config; + long *msg3_DeltaPreamble; uint8_t transform_precoding; uint8_t mcs_table; nr_dci_format_t dci_format; int max_fb_time; } NR_UE_UL_BWP_t; +typedef enum { + defaultA = 0, + defaultB = 1, + defaultC = 2 +} default_table_type_t; + +typedef enum { + typeA = 0, + typeB = 1 +} mappingType_t; + +typedef struct NR_tda_info { + mappingType_t mapping_type; + int startSymbolIndex; + int nrOfSymbols; + long k2; +} NR_tda_info_t; + #endif /*__LAYER2_MAC_H__ */ diff --git a/openair2/LAYER2/NR_MAC_COMMON/nr_mac_common.c b/openair2/LAYER2/NR_MAC_COMMON/nr_mac_common.c index 619f64dee9..7ee06c4406 100644 --- a/openair2/LAYER2/NR_MAC_COMMON/nr_mac_common.c +++ b/openair2/LAYER2/NR_MAC_COMMON/nr_mac_common.c @@ -419,79 +419,230 @@ const char table_38211_6_3_1_5_5[22][4][2] = { {{'1', '1'}, {'o', 'o'}, {'j', 'o'}, {'1', 'n'}} // tpmi 21 }; -void get_info_from_tda_tables(int default_abc, - int tda, - int dmrs_TypeA_Position, - int normal_CP, - bool *is_mapping_typeA, - int *startSymbolIndex, - int *nrOfSymbols) { +// Default PUSCH time domain resource allocation tables from 38.214 +const uint8_t table_6_1_2_1_1_2[16][4] = { + {0, 0, 0, 14}, // row index 1 + {0, 0, 0, 12}, // row index 2 + {0, 0, 0, 10}, // row index 3 + {1, 0, 2, 10}, // row index 4 + {1, 0, 4, 10}, // row index 5 + {1, 0, 4, 8}, // row index 6 + {1, 0, 4, 6}, // row index 7 + {0, 1, 0, 14}, // row index 8 + {0, 1, 0, 12}, // row index 9 + {0, 1, 0, 10}, // row index 10 + {0, 2, 0, 14}, // row index 11 + {0, 2, 0, 12}, // row index 12 + {0, 2, 0, 10}, // row index 13 + {1, 0, 8, 6}, // row index 14 + {0, 3, 0, 14}, // row index 15 + {0, 3, 0, 10} // row index 16 +}; + +const uint8_t table_6_1_2_1_1_3[16][4] = { + {0, 0, 0, 8}, // row index 1 + {0, 0, 0, 12}, // row index 2 + {0, 0, 0, 10}, // row index 3 + {1, 0, 2, 10}, // row index 4 + {1, 0, 4, 4}, // row index 5 + {1, 0, 4, 8}, // row index 6 + {1, 0, 4, 6}, // row index 7 + {0, 1, 0, 8}, // row index 8 + {0, 1, 0, 12}, // row index 9 + {0, 1, 0, 10}, // row index 10 + {0, 2, 0, 6}, // row index 11 + {0, 2, 0, 12}, // row index 12 + {0, 2, 0, 10}, // row index 13 + {1, 0, 8, 4}, // row index 14 + {0, 3, 0, 8}, // row index 15 + {0, 3, 0, 10} // row index 16 +}; + +NR_tda_info_t get_ul_tda_info(const NR_UE_UL_BWP_t *ul_bwp, int controlResourceSetId, int ss_type, nr_rnti_type_t rnti_type, int tda_index) +{ + NR_tda_info_t tda_info = {0}; + NR_PUSCH_TimeDomainResourceAllocationList_t *tdalist = get_ul_tdalist(ul_bwp, controlResourceSetId, ss_type, rnti_type); + // Definition of value j in Table 6.1.2.1.1-4 of 38.214 + int scs = ul_bwp->scs; + AssertFatal(scs >= 0 && scs < 5, "Subcarrier spacing indicatior %d invalid value\n", scs); + int j = scs == 0 ? 1 : scs; + if (tdalist) { + AssertFatal(tda_index < tdalist->list.count, "TDA index from DCI %d exceeds TDA list array size %d\n", tda_index, tdalist->list.count); + NR_PUSCH_TimeDomainResourceAllocation_t *tda = tdalist->list.array[tda_index]; + tda_info.mapping_type = tda->mappingType; + // As described in 38.331, when the field is absent the UE applies the value 1 when PUSCH SCS is 15/30KHz + // 2 when PUSCH SCS is 60KHz and 3 when PUSCH SCS is 120KHz. This equates to the parameter j. + tda_info.k2 = tda->k2 ? *tda->k2 : j; + int S, L; + SLIV2SL(tda->startSymbolAndLength, &S, &L); + tda_info.startSymbolIndex = S; + tda_info.nrOfSymbols = L; + } else { + bool normal_CP = ul_bwp->cyclicprefix ? false : true; + if (normal_CP) { + tda_info.mapping_type = table_6_1_2_1_1_2[tda_index][0]; + tda_info.k2 = table_6_1_2_1_1_2[tda_index][1] + j; + tda_info.startSymbolIndex = table_6_1_2_1_1_2[tda_index][2]; + tda_info.nrOfSymbols = table_6_1_2_1_1_2[tda_index][3]; + } else { + tda_info.mapping_type = table_6_1_2_1_1_3[tda_index][0]; + tda_info.k2 = table_6_1_2_1_1_3[tda_index][1] + j; + tda_info.startSymbolIndex = table_6_1_2_1_1_3[tda_index][2]; + tda_info.nrOfSymbols = table_6_1_2_1_1_3[tda_index][3]; + } + } + return tda_info; +} + +NR_tda_info_t get_info_from_tda_tables(default_table_type_t table_type, + int tda, + int dmrs_TypeA_Position, + int normal_CP) +{ + NR_tda_info_t tda_info = {0}; + bool is_mapping_typeA; int k0 = 0; - switch(default_abc){ - case 1: + switch(table_type){ + case defaultA: if (normal_CP){ if (dmrs_TypeA_Position){ - *is_mapping_typeA = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos3[tda][0]; + is_mapping_typeA = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos3[tda][0]; k0 = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos3[tda][1]; - *startSymbolIndex = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos3[tda][2]; - *nrOfSymbols = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos3[tda][3]; + tda_info.startSymbolIndex = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos3[tda][2]; + tda_info.nrOfSymbols = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos3[tda][3]; } else{ - *is_mapping_typeA = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos2[tda][0]; + is_mapping_typeA = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos2[tda][0]; k0 = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos2[tda][1]; - *startSymbolIndex = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos2[tda][2]; - *nrOfSymbols = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos2[tda][3]; + tda_info.startSymbolIndex = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos2[tda][2]; + tda_info.nrOfSymbols = table_5_1_2_1_1_2_time_dom_res_alloc_A_dmrs_typeA_pos2[tda][3]; } } else{ if (dmrs_TypeA_Position){ - *is_mapping_typeA = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos3[tda][0]; + is_mapping_typeA = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos3[tda][0]; k0 = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos3[tda][1]; - *startSymbolIndex = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos3[tda][2]; - *nrOfSymbols = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos3[tda][3]; + tda_info.startSymbolIndex = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos3[tda][2]; + tda_info.nrOfSymbols = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos3[tda][3]; } else{ - *is_mapping_typeA = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos2[tda][0]; + is_mapping_typeA = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos2[tda][0]; k0 = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos2[tda][1]; - *startSymbolIndex = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos2[tda][2]; - *nrOfSymbols = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos2[tda][3]; + tda_info.startSymbolIndex = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos2[tda][2]; + tda_info.nrOfSymbols = table_5_1_2_1_1_3_time_dom_res_alloc_A_extCP_dmrs_typeA_pos2[tda][3]; } } break; - case 2: + case defaultB: if (dmrs_TypeA_Position){ - *is_mapping_typeA = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos3[tda][0]; + is_mapping_typeA = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos3[tda][0]; k0 = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos3[tda][1]; - *startSymbolIndex = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos3[tda][2]; - *nrOfSymbols = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos3[tda][3]; + tda_info.startSymbolIndex = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos3[tda][2]; + tda_info.nrOfSymbols = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos3[tda][3]; } else{ - *is_mapping_typeA = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos2[tda][0]; + is_mapping_typeA = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos2[tda][0]; k0 = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos2[tda][1]; - *startSymbolIndex = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos2[tda][2]; - *nrOfSymbols = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos2[tda][3]; + tda_info.startSymbolIndex = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos2[tda][2]; + tda_info.nrOfSymbols = table_5_1_2_1_1_4_time_dom_res_alloc_B_dmrs_typeA_pos2[tda][3]; } break; - case 3: + case defaultC: if (dmrs_TypeA_Position){ - *is_mapping_typeA = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos3[tda][0]; + is_mapping_typeA = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos3[tda][0]; k0 = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos3[tda][1]; - *startSymbolIndex = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos3[tda][2]; - *nrOfSymbols = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos3[tda][3]; + tda_info.startSymbolIndex = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos3[tda][2]; + tda_info.nrOfSymbols = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos3[tda][3]; } else{ - *is_mapping_typeA = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos2[tda][0]; + is_mapping_typeA = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos2[tda][0]; k0 = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos2[tda][1]; - *startSymbolIndex = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos2[tda][2]; - *nrOfSymbols = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos2[tda][3]; + tda_info.startSymbolIndex = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos2[tda][2]; + tda_info.nrOfSymbols = table_5_1_2_1_1_5_time_dom_res_alloc_C_dmrs_typeA_pos2[tda][3]; } break; default: - AssertFatal(1==0,"Invalid default time domaing allocation type\n"); + AssertFatal(1 == 0, "Invalid default time domaing allocation type\n"); } - AssertFatal(k0==0,"Only k0 = 0 is supported\n"); + AssertFatal(k0 == 0, "Only k0 = 0 is supported\n"); + tda_info.mapping_type = is_mapping_typeA ? typeA : typeB; + return tda_info; } +default_table_type_t get_default_table_type(int mux_pattern) +{ + switch (mux_pattern) { + case 1: + return defaultA; + case 2: + return defaultB; + case 3: + return defaultC; + default : + AssertFatal(1 == 0, "Invalid multiplexing type %d\n", mux_pattern); + } +} + +NR_tda_info_t set_tda_info_from_list(NR_PDSCH_TimeDomainResourceAllocationList_t *tdalist, int tda_index) +{ + NR_tda_info_t tda_info = {0}; + AssertFatal(tda_index < tdalist->list.count, "TDA index from DCI %d exceeds TDA list array size %d\n", tda_index, tdalist->list.count); + NR_PDSCH_TimeDomainResourceAllocation_t *tda = tdalist->list.array[tda_index]; + tda_info.mapping_type = tda->mappingType; + int S, L; + SLIV2SL(tda->startSymbolAndLength, &S, &L); + tda_info.startSymbolIndex = S; + tda_info.nrOfSymbols = L; + return tda_info; +} + +NR_tda_info_t get_dl_tda_info(const NR_UE_DL_BWP_t *dl_BWP, int ss_type, int tda_index, int dmrs_typeA_pos, + int mux_pattern, nr_rnti_type_t rnti_type, int coresetid, bool sib1) +{ + NR_tda_info_t tda_info; + bool normal_CP = dl_BWP->cyclicprefix ? false : true; + // implements Table 5.1.2.1.1-1 of 38.214 + NR_PDSCH_TimeDomainResourceAllocationList_t *tdalist = get_dl_tdalist(dl_BWP, coresetid, ss_type, rnti_type); + switch (rnti_type) { + case NR_RNTI_SI: + if(sib1) { + default_table_type_t table_type = get_default_table_type(mux_pattern); + tda_info = get_info_from_tda_tables(table_type, tda_index, dmrs_typeA_pos, normal_CP); + } + else { + if(tdalist) + tda_info = set_tda_info_from_list(tdalist, tda_index); + else { + default_table_type_t table_type = get_default_table_type(mux_pattern); + tda_info = get_info_from_tda_tables(table_type, tda_index, dmrs_typeA_pos, normal_CP); + } + } + break; + case NR_RNTI_P: + if(tdalist) + tda_info = set_tda_info_from_list(tdalist, tda_index); + else { + default_table_type_t table_type = get_default_table_type(mux_pattern); + tda_info = get_info_from_tda_tables(table_type, tda_index, dmrs_typeA_pos, normal_CP); + } + break; + case NR_RNTI_C: + case NR_RNTI_CS: + case NR_RNTI_MCS_C: + case NR_RNTI_RA: + case NR_RNTI_TC: + if(tdalist) + tda_info = set_tda_info_from_list(tdalist, tda_index); + else + tda_info = get_info_from_tda_tables(defaultA, tda_index, dmrs_typeA_pos, normal_CP); + break; + default : + AssertFatal(1 == 0, "Invalid RNTI type\n"); + } + return tda_info; +} + + const char *prachfmt[]={"0","1","2","3", "A1","A2","A3","B1","B4","C0","C2","A1/B1","A2/B2","A3/B3"}; uint16_t get_NCS(uint8_t index, uint16_t format0, uint8_t restricted_set_config) { @@ -2701,38 +2852,15 @@ uint16_t get_nr_srs_offset(NR_SRS_PeriodicityAndOffset_t periodicityAndOffset) { // Set the transform precoding status according to 6.1.3 of 3GPP TS 38.214 version 16.3.0 Release 16: // - "UE procedure for applying transform precoding on PUSCH" -uint8_t get_transformPrecoding(const NR_BWP_UplinkCommon_t *initialUplinkBWP, - const NR_PUSCH_Config_t *pusch_config, - const NR_BWP_UplinkDedicated_t *ubwp, - uint8_t *dci_format, - int rnti_type, - uint8_t configuredGrant){ +uint8_t get_transformPrecoding(const NR_UE_UL_BWP_t *current_UL_BWP, nr_dci_format_t dci_format, uint8_t configuredGrant) +{ + if (configuredGrant && current_UL_BWP->configuredGrantConfig && current_UL_BWP->configuredGrantConfig->transformPrecoder) + return *current_UL_BWP->configuredGrantConfig->transformPrecoder; - if (configuredGrant) { - if (ubwp->configuredGrantConfig) { - if (ubwp->configuredGrantConfig->choice.setup->transformPrecoder) { - return *ubwp->configuredGrantConfig->choice.setup->transformPrecoder; - } - } - } - - if (rnti_type != NR_RNTI_RA && rnti_type != NR_RNTI_TC) { - if (*dci_format != NR_UL_DCI_FORMAT_0_0) { - if (pusch_config && pusch_config->transformPrecoder != NULL) { - return *pusch_config->transformPrecoder; - } - } - } - - if (initialUplinkBWP->rach_ConfigCommon->choice.setup->msg3_transformPrecoder == NULL) { - return 1; // Transformprecoding disabled - } else { - LOG_D(PHY, "MAC_COMMON: Transform Precodig enabled through msg3_transformPrecoder\n"); - return 0; // Enabled - } - - LOG_E(MAC, "In %s: could not fetch transform precoder status...\n", __FUNCTION__); - return -1; + if (dci_format == NR_UL_DCI_FORMAT_0_1 && current_UL_BWP && current_UL_BWP->pusch_Config && current_UL_BWP->pusch_Config->transformPrecoder) + return *current_UL_BWP->pusch_Config->transformPrecoder; + else + return current_UL_BWP->rach_ConfigCommon->msg3_transformPrecoder ? 0 : 1; } uint8_t get_pusch_nb_antenna_ports(NR_PUSCH_Config_t *pusch_Config, @@ -3130,6 +3258,27 @@ uint8_t compute_precoding_information(NR_PUSCH_Config_t *pusch_Config, return nbits; } +NR_PDSCH_TimeDomainResourceAllocationList_t *get_dl_tdalist(const NR_UE_DL_BWP_t *DL_BWP, int controlResourceSetId, int ss_type, nr_rnti_type_t rnti_type) +{ + if (!DL_BWP) + return NULL; + // see table 5.1.2.1.1-1 in 38.214 + if ((rnti_type == NR_RNTI_CS || rnti_type == NR_RNTI_C || rnti_type == NR_RNTI_MCS_C) && !(ss_type == NR_SearchSpace__searchSpaceType_PR_common && controlResourceSetId == 0) + && (DL_BWP->pdsch_Config && DL_BWP->pdsch_Config->pdsch_TimeDomainAllocationList)) + return DL_BWP->pdsch_Config->pdsch_TimeDomainAllocationList->choice.setup; + else + return DL_BWP->tdaList_Common; +} + +NR_PUSCH_TimeDomainResourceAllocationList_t *get_ul_tdalist(const NR_UE_UL_BWP_t *UL_BWP, int controlResourceSetId, int ss_type, nr_rnti_type_t rnti_type) +{ + if ((rnti_type == NR_RNTI_CS || rnti_type == NR_RNTI_C || rnti_type == NR_RNTI_MCS_C) && !(ss_type == NR_SearchSpace__searchSpaceType_PR_common && controlResourceSetId == 0) + && (UL_BWP->pusch_Config && UL_BWP->pusch_Config->pusch_TimeDomainAllocationList)) + return UL_BWP->pusch_Config->pusch_TimeDomainAllocationList->choice.setup; + else + return UL_BWP->tdaList_Common; +} + uint16_t get_rb_bwp_dci(nr_dci_format_t format, int ss_type, uint16_t cset0_bwp_size, @@ -3155,15 +3304,13 @@ uint16_t get_rb_bwp_dci(nr_dci_format_t format, return N_RB; } -uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, - const NR_BWP_UplinkCommon_t *initialUplinkBWP, - const NR_UE_DL_BWP_t *DL_BWP, +uint16_t nr_dci_size(const NR_UE_DL_BWP_t *DL_BWP, const NR_UE_UL_BWP_t *UL_BWP, const NR_CellGroupConfig_t *cg, dci_pdu_rel15_t *dci_pdu, nr_dci_format_t format, nr_rnti_type_t rnti_type, - int controlResourceSetId, + NR_ControlResourceSet_t *coreset, int bwp_id, int ss_type, uint16_t cset0_bwp_size, @@ -3175,43 +3322,10 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, long rbg_size_config; int num_entries = 0; - NR_UplinkConfig_t *uplinkConfig = NULL; - if (cg && cg->spCellConfig && cg->spCellConfig->spCellConfigDedicated) { - uplinkConfig = cg->spCellConfig->spCellConfigDedicated->uplinkConfig; - } - - const NR_BWP_DownlinkDedicated_t *bwpd = NULL; - const NR_BWP_UplinkDedicated_t *ubwpd = NULL; - const NR_BWP_DownlinkCommon_t *bwpc = NULL; - const NR_BWP_UplinkCommon_t *ubwpc = NULL; - NR_PDSCH_Config_t *pdsch_Config = NULL; - NR_PUSCH_Config_t *pusch_Config = NULL; - NR_PUCCH_Config_t *pucch_Config = NULL; - NR_PDCCH_Config_t *pdcch_Config = NULL; - NR_SRS_Config_t *srs_config = NULL; - if(bwp_id > 0) { - AssertFatal(cg!=NULL,"Cellgroup is null and bwp_id!=0"); - bwpd = cg->spCellConfig->spCellConfigDedicated->downlinkBWP_ToAddModList->list.array[bwp_id-1]->bwp_Dedicated; - bwpc = cg->spCellConfig->spCellConfigDedicated->downlinkBWP_ToAddModList->list.array[bwp_id-1]->bwp_Common; - ubwpd = uplinkConfig ? uplinkConfig->uplinkBWP_ToAddModList->list.array[bwp_id-1]->bwp_Dedicated : NULL; - ubwpc = uplinkConfig ? uplinkConfig->uplinkBWP_ToAddModList->list.array[bwp_id-1]->bwp_Common : NULL; - pdsch_Config = (bwpd->pdsch_Config) ? bwpd->pdsch_Config->choice.setup : NULL; - pdcch_Config = (bwpd->pdcch_Config) ? bwpd->pdcch_Config->choice.setup : NULL; - pucch_Config = (ubwpd->pucch_Config) ? ubwpd->pucch_Config->choice.setup : NULL; - pusch_Config = (ubwpd->pusch_Config) ? ubwpd->pusch_Config->choice.setup : NULL; - srs_config = (ubwpd->srs_Config) ? ubwpd->srs_Config->choice.setup : NULL; - } else if (cg) { - bwpc = initialDownlinkBWP; - ubwpc = initialUplinkBWP; - bwpd = cg->spCellConfig && cg->spCellConfig->spCellConfigDedicated ? - cg->spCellConfig->spCellConfigDedicated->initialDownlinkBWP : NULL; - ubwpd = uplinkConfig ? uplinkConfig->initialUplinkBWP : NULL; - pdsch_Config = (bwpd && bwpd->pdsch_Config) ? bwpd->pdsch_Config->choice.setup : NULL; - pdcch_Config = (bwpd && bwpd->pdcch_Config) ? bwpd->pdcch_Config->choice.setup : NULL; - pucch_Config = (ubwpd && ubwpd->pucch_Config) ? ubwpd->pucch_Config->choice.setup : NULL; - pusch_Config = (ubwpd && ubwpd->pusch_Config) ? ubwpd->pusch_Config->choice.setup : NULL; - srs_config = (ubwpd && ubwpd->srs_Config) ? ubwpd->srs_Config->choice.setup: NULL; - } + NR_PDSCH_Config_t *pdsch_Config = DL_BWP ? DL_BWP->pdsch_Config : NULL; + NR_PUSCH_Config_t *pusch_Config = UL_BWP ? UL_BWP->pusch_Config : NULL; + NR_PUCCH_Config_t *pucch_Config = UL_BWP ? UL_BWP->pucch_Config : NULL; + NR_SRS_Config_t *srs_config = UL_BWP ? UL_BWP->srs_Config : NULL; uint16_t N_RB = cset0_bwp_size; if (DL_BWP) @@ -3223,7 +3337,6 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, UL_BWP->initial_BWPSize, DL_BWP->initial_BWPSize); - int n_ul_bwp = 1,n_dl_bwp = 1; switch(format) { case NR_UL_DCI_FORMAT_0_0: /// fixed: Format identifier 1, Hop flag 1, MCS 5, NDI 1, RV 2, HARQ PID 4, PUSCH TPC 2 Time Domain assgnmt 4 --20 @@ -3253,14 +3366,11 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, size += dci_pdu->ul_sul_indicator.nbits; } // BWP Indicator - n_ul_bwp = 0; - if (cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - cg->spCellConfig->spCellConfigDedicated->uplinkConfig->uplinkBWP_ToAddModList) - n_ul_bwp = cg->spCellConfig->spCellConfigDedicated->uplinkConfig->uplinkBWP_ToAddModList->list.count; - if (n_ul_bwp < 2) - dci_pdu->bwp_indicator.nbits = n_ul_bwp; + if (UL_BWP->n_ul_bwp < 2) + dci_pdu->bwp_indicator.nbits = UL_BWP->n_ul_bwp; else dci_pdu->bwp_indicator.nbits = 2; + LOG_D(NR_MAC, "BWP indicator nbits %d, num UL BWPs %d\n", dci_pdu->bwp_indicator.nbits, UL_BWP->n_ul_bwp); size += dci_pdu->bwp_indicator.nbits; // Freq domain assignment if (pusch_Config) { @@ -3278,19 +3388,13 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, } else dci_pdu->frequency_domain_assignment.nbits = (int)ceil(log2((N_RB * (N_RB + 1)) >> 1)); - LOG_D(NR_MAC,"PUSCH Frequency Domain Assignment nbits %d, N_RB %d\n",dci_pdu->frequency_domain_assignment.nbits,N_RB); + LOG_D(NR_MAC, "PUSCH Frequency Domain Assignment nbits %d, N_RB %d\n", dci_pdu->frequency_domain_assignment.nbits, N_RB); size += dci_pdu->frequency_domain_assignment.nbits; // Time domain assignment - if (pusch_Config==NULL || pusch_Config->pusch_TimeDomainAllocationList==NULL) { - if (ubwpc->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList==NULL) - num_entries = 16; // num of entries in default table - else - num_entries = ubwpc->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList->list.count; - } - else - num_entries = pusch_Config->pusch_TimeDomainAllocationList->choice.setup->list.count; + NR_PUSCH_TimeDomainResourceAllocationList_t *tdalistul = get_ul_tdalist(UL_BWP, coreset->controlResourceSetId, ss_type, rnti_type); + num_entries = tdalistul ? tdalistul->list.count : 16; // 16 in default table dci_pdu->time_domain_assignment.nbits = (int)ceil(log2(num_entries)); - LOG_D(NR_MAC,"PUSCH Time Domain Allocation nbits %d, pusch_Config %p\n",dci_pdu->time_domain_assignment.nbits,pusch_Config); + LOG_D(NR_MAC, "PUSCH Time Domain Allocation nbits %d, pusch_Config %p\n", dci_pdu->time_domain_assignment.nbits, pusch_Config); size += dci_pdu->time_domain_assignment.nbits; // Frequency Hopping flag if (pusch_Config && @@ -3300,21 +3404,19 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, size += 1; } // 1st DAI - if (cg->physicalCellGroupConfig && cg->physicalCellGroupConfig->pdsch_HARQ_ACK_Codebook == NR_PhysicalCellGroupConfig__pdsch_HARQ_ACK_Codebook_dynamic) + if (DL_BWP->pdsch_HARQ_ACK_Codebook && *DL_BWP->pdsch_HARQ_ACK_Codebook == NR_PhysicalCellGroupConfig__pdsch_HARQ_ACK_Codebook_dynamic) dci_pdu->dai[0].nbits = 2; else dci_pdu->dai[0].nbits = 1; size += dci_pdu->dai[0].nbits; LOG_D(NR_MAC, "DAI1 nbits %d\n", dci_pdu->dai[0].nbits); // 2nd DAI - if (cg->spCellConfig->spCellConfigDedicated->pdsch_ServingCellConfig && cg->spCellConfig->spCellConfigDedicated->pdsch_ServingCellConfig->choice.setup - && cg->spCellConfig->spCellConfigDedicated->pdsch_ServingCellConfig->choice.setup->codeBlockGroupTransmission != NULL) { // TODO not sure about that + if (DL_BWP->pdsch_servingcellconfig && DL_BWP->pdsch_servingcellconfig->codeBlockGroupTransmission != NULL) { dci_pdu->dai[1].nbits = 2; size += dci_pdu->dai[1].nbits; } // SRS resource indicator - NR_PUSCH_ServingCellConfig_t *pusch_servingcellconfig = uplinkConfig && uplinkConfig->pusch_ServingCellConfig ? uplinkConfig->pusch_ServingCellConfig->choice.setup : NULL; - dci_pdu->srs_resource_indicator.nbits = compute_srs_resource_indicator(pusch_servingcellconfig, pusch_Config, srs_config, NULL, NULL); + dci_pdu->srs_resource_indicator.nbits = compute_srs_resource_indicator(UL_BWP->pusch_servingcellconfig, pusch_Config, srs_config, NULL, NULL); size += dci_pdu->srs_resource_indicator.nbits; LOG_D(NR_MAC, "dci_pdu->srs_resource_indicator.nbits %d\n", dci_pdu->srs_resource_indicator.nbits); // Precoding info and number of layers @@ -3322,7 +3424,7 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, size += dci_pdu->precoding_information.nbits; LOG_D(NR_MAC, "dci_pdu->precoding_informaiton.nbits=%d\n", dci_pdu->precoding_information.nbits); // Antenna ports - long transformPrecoder = get_transformPrecoding(initialUplinkBWP, pusch_Config, ubwpd, (uint8_t *)&format, rnti_type, 0); + long transformPrecoder = get_transformPrecoding(UL_BWP, format, 0); NR_DMRS_UplinkConfig_t *NR_DMRS_UplinkConfig = NULL; int xa = 0; int xb = 0; @@ -3348,16 +3450,15 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, dci_pdu->srs_request.nbits = 3; size += dci_pdu->srs_request.nbits; // CSI request - if (cg->spCellConfig->spCellConfigDedicated->csi_MeasConfig != NULL) { - if (cg->spCellConfig->spCellConfigDedicated->csi_MeasConfig->choice.setup->reportTriggerSize != NULL) { - dci_pdu->csi_request.nbits = *cg->spCellConfig->spCellConfigDedicated->csi_MeasConfig->choice.setup->reportTriggerSize; + if (UL_BWP->csi_MeasConfig != NULL) { + if (UL_BWP->csi_MeasConfig->reportTriggerSize != NULL) { + dci_pdu->csi_request.nbits = *UL_BWP->csi_MeasConfig->reportTriggerSize; size += dci_pdu->csi_request.nbits; } } // CBGTI - if (cg->spCellConfig->spCellConfigDedicated->uplinkConfig->pusch_ServingCellConfig && - cg->spCellConfig->spCellConfigDedicated->uplinkConfig->pusch_ServingCellConfig->choice.setup->codeBlockGroupTransmission != NULL) { - int num = cg->spCellConfig->spCellConfigDedicated->uplinkConfig->pusch_ServingCellConfig->choice.setup->codeBlockGroupTransmission->choice.setup->maxCodeBlockGroupsPerTransportBlock; + if (UL_BWP->pusch_servingcellconfig && UL_BWP->pusch_servingcellconfig->codeBlockGroupTransmission != NULL) { + int num = UL_BWP->pusch_servingcellconfig->codeBlockGroupTransmission->choice.setup->maxCodeBlockGroupsPerTransportBlock; dci_pdu->cbgti.nbits = 2 + (num<<1); size += dci_pdu->cbgti.nbits; } @@ -3403,7 +3504,7 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, break; case NR_DL_DCI_FORMAT_1_1: - LOG_D(NR_MAC,"DCI_FORMAT 1_1 : pdsch_Config %p, pucch_Config %p\n",pdsch_Config,pucch_Config); + LOG_D(NR_MAC, "DCI_FORMAT 1_1 : pdsch_Config %p, pucch_Config %p\n", pdsch_Config, pucch_Config); // General note: 0 bits condition is ignored as default nbits is 0. // Format identifier size = 1; @@ -3414,11 +3515,8 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, } // BWP Indicator - n_dl_bwp = 0; - if (cg->spCellConfig->spCellConfigDedicated->downlinkBWP_ToAddModList) - n_dl_bwp = cg->spCellConfig->spCellConfigDedicated->downlinkBWP_ToAddModList->list.count; - if (n_dl_bwp < 2) - dci_pdu->bwp_indicator.nbits = n_dl_bwp; + if (DL_BWP->n_dl_bwp < 2) + dci_pdu->bwp_indicator.nbits = DL_BWP->n_dl_bwp; else dci_pdu->bwp_indicator.nbits = 2; size += dci_pdu->bwp_indicator.nbits; @@ -3435,15 +3533,8 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, dci_pdu->frequency_domain_assignment.nbits = ((int)ceil(log2((N_RB * (N_RB + 1)) >> 1)) > numRBG) ? (int)ceil(log2((N_RB * (N_RB + 1)) >> 1)) + 1 : numRBG + 1; size += dci_pdu->frequency_domain_assignment.nbits; LOG_D(NR_MAC,"dci_pdu->frequency_domain_assignment.nbits %d (N_RB %d)\n",dci_pdu->frequency_domain_assignment.nbits,N_RB); - // Time domain assignment (see table 5.1.2.1.1-1 in 38.214 - if (pdsch_Config == NULL || pdsch_Config->pdsch_TimeDomainAllocationList==NULL) { - if (bwpc->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList==NULL) - num_entries = 16; // num of entries in default table - else - num_entries = bwpc->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList->list.count; - } - else - num_entries = pdsch_Config->pdsch_TimeDomainAllocationList->choice.setup->list.count; + NR_PDSCH_TimeDomainResourceAllocationList_t *tdalist = get_dl_tdalist(DL_BWP, coreset->controlResourceSetId, ss_type, rnti_type); + num_entries = tdalist ? tdalist->list.count : 16; // 16 in default table dci_pdu->time_domain_assignment.nbits = (int)ceil(log2(num_entries)); LOG_D(NR_MAC,"pdsch tda.nbits= %d\n",dci_pdu->time_domain_assignment.nbits); size += dci_pdu->time_domain_assignment.nbits; @@ -3485,8 +3576,7 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, // HARQ PID size += 4; // DAI - if (cg->physicalCellGroupConfig && - cg->physicalCellGroupConfig->pdsch_HARQ_ACK_Codebook == NR_PhysicalCellGroupConfig__pdsch_HARQ_ACK_Codebook_dynamic) { // FIXME in case of more than one serving cell + if (DL_BWP->pdsch_HARQ_ACK_Codebook && *DL_BWP->pdsch_HARQ_ACK_Codebook == NR_PhysicalCellGroupConfig__pdsch_HARQ_ACK_Codebook_dynamic) { // FIXME in case of more than one serving cell dci_pdu->dai[0].nbits = 2; size += dci_pdu->dai[0].nbits; } @@ -3508,15 +3598,9 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, size += dci_pdu->antenna_ports.nbits; LOG_D(NR_MAC,"dci_pdu->antenna_ports.nbits %d\n",dci_pdu->antenna_ports.nbits); // Tx Config Indication - for (int i = 0; i < pdcch_Config->controlResourceSetToAddModList->list.count; i++) { - if (pdcch_Config->controlResourceSetToAddModList->list.array[i]->controlResourceSetId == controlResourceSetId) { - long *isTciEnable = pdcch_Config->controlResourceSetToAddModList->list.array[i]->tci_PresentInDCI; - if (isTciEnable != NULL) { - dci_pdu->transmission_configuration_indication.nbits = 3; - size += dci_pdu->transmission_configuration_indication.nbits; - } - break; - } + if (coreset->tci_PresentInDCI != NULL) { + dci_pdu->transmission_configuration_indication.nbits = 3; + size += dci_pdu->transmission_configuration_indication.nbits; } // SRS request if (cg->spCellConfig->spCellConfigDedicated->supplementaryUplink==NULL) @@ -3525,16 +3609,14 @@ uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, dci_pdu->srs_request.nbits = 3; size += dci_pdu->srs_request.nbits; // CBGTI - if (cg->spCellConfig->spCellConfigDedicated->pdsch_ServingCellConfig && - cg->spCellConfig->spCellConfigDedicated->pdsch_ServingCellConfig->choice.setup && - cg->spCellConfig->spCellConfigDedicated->pdsch_ServingCellConfig->choice.setup->codeBlockGroupTransmission != NULL) { - uint8_t maxCBGperTB = (cg->spCellConfig->spCellConfigDedicated->pdsch_ServingCellConfig->choice.setup->codeBlockGroupTransmission->choice.setup->maxCodeBlockGroupsPerTransportBlock + 1) * 2; + if (DL_BWP->pdsch_servingcellconfig && DL_BWP->pdsch_servingcellconfig->codeBlockGroupTransmission != NULL) { + uint8_t maxCBGperTB = (DL_BWP->pdsch_servingcellconfig->codeBlockGroupTransmission->choice.setup->maxCodeBlockGroupsPerTransportBlock + 1) * 2; long *maxCWperDCI_rrc = pdsch_Config->maxNrofCodeWordsScheduledByDCI; uint8_t maxCW = (maxCWperDCI_rrc == NULL) ? 1 : *maxCWperDCI_rrc; dci_pdu->cbgti.nbits = maxCBGperTB * maxCW; size += dci_pdu->cbgti.nbits; // CBGFI - if (cg->spCellConfig->spCellConfigDedicated->pdsch_ServingCellConfig->choice.setup->codeBlockGroupTransmission->choice.setup->codeBlockGroupFlushIndicator) { + if (DL_BWP->pdsch_servingcellconfig->codeBlockGroupTransmission->choice.setup->codeBlockGroupFlushIndicator) { dci_pdu->cbgfi.nbits = 1; size += dci_pdu->cbgfi.nbits; } diff --git a/openair2/LAYER2/NR_MAC_COMMON/nr_mac_common.h b/openair2/LAYER2/NR_MAC_COMMON/nr_mac_common.h index 8aee2cc5a3..a6cc264c02 100644 --- a/openair2/LAYER2/NR_MAC_COMMON/nr_mac_common.h +++ b/openair2/LAYER2/NR_MAC_COMMON/nr_mac_common.h @@ -51,11 +51,6 @@ typedef enum { pusch_len2 = 2 } pusch_maxLength_t; -typedef enum { - typeA = 0, - typeB = 1 -} mappingType_t; - uint32_t get_Y(const NR_SearchSpace_t *ss, int slot, rnti_t rnti); uint8_t get_BG(uint32_t A, uint16_t R); @@ -89,15 +84,22 @@ uint8_t compute_precoding_information(NR_PUSCH_Config_t *pusch_Config, const uint8_t *nrOfLayers, uint32_t *val); -uint16_t nr_dci_size(const NR_BWP_DownlinkCommon_t *initialDownlinkBWP, - const NR_BWP_UplinkCommon_t *initialUplinkBWP, - const NR_UE_DL_BWP_t *DL_BWP, +NR_PDSCH_TimeDomainResourceAllocationList_t *get_dl_tdalist(const NR_UE_DL_BWP_t *DL_BWP, int controlResourceSetId, int ss_type, nr_rnti_type_t rnti_type); + +NR_PUSCH_TimeDomainResourceAllocationList_t *get_ul_tdalist(const NR_UE_UL_BWP_t *UL_BWP, int controlResourceSetId, int ss_type, nr_rnti_type_t rnti_type); + +NR_tda_info_t get_ul_tda_info(const NR_UE_UL_BWP_t *ul_bwp, int controlResourceSetId, int ss_type, nr_rnti_type_t rnti_type, int tda_index); + +NR_tda_info_t get_dl_tda_info(const NR_UE_DL_BWP_t *dl_BWP, int ss_type, int tda_index, int dmrs_typeA_pos, + int mux_pattern, nr_rnti_type_t rnti_type, int coresetid, bool sib1); + +uint16_t nr_dci_size(const NR_UE_DL_BWP_t *DL_BWP, const NR_UE_UL_BWP_t *UL_BWP, const NR_CellGroupConfig_t *cg, dci_pdu_rel15_t *dci_pdu, nr_dci_format_t format, nr_rnti_type_t rnti_type, - int controlResourceSetId, + NR_ControlResourceSet_t *coreset, int bwp_id, int ss_type, uint16_t cset0_bwp_size, @@ -220,13 +222,12 @@ void get_type0_PDCCH_CSS_config_parameters(NR_Type0_PDCCH_CSS_config_t *type0_PD uint16_t get_ssb_start_symbol(const long band, NR_SubcarrierSpacing_t scs, int i_ssb); -void get_info_from_tda_tables(int default_abc, - int tda, - int dmrs_TypeA_Position, - int normal_CP, - bool *is_mapping_typeA, - int *startSymbolIndex, - int *nrOfSymbols); +NR_tda_info_t get_info_from_tda_tables(default_table_type_t table_type, + int tda, + int dmrs_TypeA_Position, + int normal_CP); + +default_table_type_t get_default_table_type(int mux_pattern); void fill_coresetZero(NR_ControlResourceSet_t *coreset0, NR_Type0_PDCCH_CSS_config_t *type0_PDCCH_CSS_config); void fill_searchSpaceZero(NR_SearchSpace_t *ss0, NR_Type0_PDCCH_CSS_config_t *type0_PDCCH_CSS_config); @@ -273,12 +274,7 @@ bool set_ul_ptrs_values(NR_PTRS_UplinkConfig_t *ul_ptrs_config, @param rnti_type rnti type @param configuredGrant indicates whether a configured grant was received or not @returns transformPrecoding value */ -uint8_t get_transformPrecoding(const NR_BWP_UplinkCommon_t *initialUplinkBWP, - const NR_PUSCH_Config_t *pusch_config, - const NR_BWP_UplinkDedicated_t *ubwp, - uint8_t *dci_format, - int rnti_type, - uint8_t configuredGrant); +uint8_t get_transformPrecoding(const NR_UE_UL_BWP_t *current_UL_BWP, nr_dci_format_t dci_format, uint8_t configuredGrant); void nr_mac_gNB_rrc_ul_failure(const module_id_t Mod_instP, const int CC_idP, diff --git a/openair2/LAYER2/NR_MAC_UE/config_ue.c b/openair2/LAYER2/NR_MAC_UE/config_ue.c index ac48e17b62..5b5587a7e7 100644 --- a/openair2/LAYER2/NR_MAC_UE/config_ue.c +++ b/openair2/LAYER2/NR_MAC_UE/config_ue.c @@ -547,27 +547,68 @@ void configure_current_BWP(NR_UE_MAC_INST_t *mac, NR_UE_UL_BWP_t *UL_BWP = &mac->current_UL_BWP; NR_BWP_t dl_genericParameters = {0}; NR_BWP_t ul_genericParameters = {0}; + NR_BWP_DownlinkCommon_t *bwp_dlcommon = NULL; + NR_BWP_UplinkCommon_t *bwp_ulcommon = NULL; + DL_BWP->n_dl_bwp = 0; + UL_BWP->n_ul_bwp = 0; if(scc) { DL_BWP->bwp_id = 0; UL_BWP->bwp_id = 0; - dl_genericParameters = scc->downlinkConfigCommon.initialDownlinkBWP.genericParameters; + bwp_dlcommon = &scc->downlinkConfigCommon.initialDownlinkBWP; + bwp_ulcommon = &scc->uplinkConfigCommon->initialUplinkBWP; + dl_genericParameters = bwp_dlcommon->genericParameters; if(scc->uplinkConfigCommon) ul_genericParameters = scc->uplinkConfigCommon->initialUplinkBWP.genericParameters; else - ul_genericParameters = scc->downlinkConfigCommon.initialDownlinkBWP.genericParameters; + ul_genericParameters = bwp_dlcommon->genericParameters; + + DL_BWP->pdsch_Config = NULL; + if (bwp_dlcommon->pdsch_ConfigCommon) + DL_BWP->tdaList_Common = bwp_dlcommon->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList; + if (bwp_ulcommon->pusch_ConfigCommon) { + UL_BWP->tdaList_Common = bwp_ulcommon->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList; + UL_BWP->msg3_DeltaPreamble = bwp_ulcommon->pusch_ConfigCommon->choice.setup->msg3_DeltaPreamble; + } + if (bwp_ulcommon->pucch_ConfigCommon) + UL_BWP->pucch_ConfigCommon = bwp_ulcommon->pucch_ConfigCommon->choice.setup; + if (bwp_ulcommon->rach_ConfigCommon) + UL_BWP->rach_ConfigCommon = bwp_ulcommon->rach_ConfigCommon->choice.setup; } + if(cell_group_config) { + if (cell_group_config->physicalCellGroupConfig) { + DL_BWP->pdsch_HARQ_ACK_Codebook = &cell_group_config->physicalCellGroupConfig->pdsch_HARQ_ACK_Codebook; + UL_BWP->harq_ACK_SpatialBundlingPUCCH = cell_group_config->physicalCellGroupConfig->harq_ACK_SpatialBundlingPUCCH; + } if (cell_group_config->spCellConfig && cell_group_config->spCellConfig->spCellConfigDedicated) { struct NR_ServingCellConfig *spCellConfigDedicated = cell_group_config->spCellConfig->spCellConfigDedicated; + UL_BWP->csi_MeasConfig = spCellConfigDedicated->csi_MeasConfig ? spCellConfigDedicated->csi_MeasConfig->choice.setup : NULL; + UL_BWP->pusch_servingcellconfig = + spCellConfigDedicated->uplinkConfig && spCellConfigDedicated->uplinkConfig->pusch_ServingCellConfig ? spCellConfigDedicated->uplinkConfig->pusch_ServingCellConfig->choice.setup : NULL; + DL_BWP->pdsch_servingcellconfig = spCellConfigDedicated->pdsch_ServingCellConfig ? spCellConfigDedicated->pdsch_ServingCellConfig->choice.setup : NULL; + if (spCellConfigDedicated->firstActiveDownlinkBWP_Id) DL_BWP->bwp_id = *spCellConfigDedicated->firstActiveDownlinkBWP_Id; if (spCellConfigDedicated->uplinkConfig->firstActiveUplinkBWP_Id) UL_BWP->bwp_id = *spCellConfigDedicated->uplinkConfig->firstActiveUplinkBWP_Id; + if (mac->scc) { + bwp_dlcommon = mac->scc->downlinkConfigCommon->initialDownlinkBWP; + bwp_ulcommon = mac->scc->uplinkConfigCommon->initialUplinkBWP; + } + else if (mac->scc_SIB) { + bwp_dlcommon = &mac->scc_SIB->downlinkConfigCommon.initialDownlinkBWP; + bwp_ulcommon = &mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP; + } + else + AssertFatal(false, "Either SCC or SCC SIB should be non-NULL\n"); + NR_BWP_Downlink_t *bwp_downlink = NULL; const struct NR_ServingCellConfig__downlinkBWP_ToAddModList *bwpList = spCellConfigDedicated->downlinkBWP_ToAddModList; + if (bwpList) + DL_BWP->n_dl_bwp = bwpList->list.count; if (bwpList && DL_BWP->bwp_id > 0) { for (int i = 0; i < bwpList->list.count; i++) { bwp_downlink = bwpList->list.array[i]; @@ -576,17 +617,21 @@ void configure_current_BWP(NR_UE_MAC_INST_t *mac, } AssertFatal(bwp_downlink != NULL,"Couldn't find DLBWP corresponding to BWP ID %ld\n", DL_BWP->bwp_id); dl_genericParameters = bwp_downlink->bwp_Common->genericParameters; + DL_BWP->pdsch_Config = bwp_downlink->bwp_Dedicated->pdsch_Config->choice.setup; + DL_BWP->tdaList_Common = bwp_downlink->bwp_Common->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList; } else { - if(mac->scc) { - dl_genericParameters = mac->scc->downlinkConfigCommon->initialDownlinkBWP->genericParameters; - } - if(mac->scc_SIB) { - dl_genericParameters = mac->scc_SIB->downlinkConfigCommon.initialDownlinkBWP.genericParameters; - } + dl_genericParameters = bwp_dlcommon->genericParameters; + DL_BWP->pdsch_Config = spCellConfigDedicated->initialDownlinkBWP->pdsch_Config->choice.setup; + DL_BWP->tdaList_Common = bwp_dlcommon->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList; } + + UL_BWP->msg3_DeltaPreamble = bwp_ulcommon->pusch_ConfigCommon->choice.setup->msg3_DeltaPreamble; + NR_BWP_Uplink_t *bwp_uplink = NULL; const struct NR_UplinkConfig__uplinkBWP_ToAddModList *ubwpList = spCellConfigDedicated->uplinkConfig->uplinkBWP_ToAddModList; + if (ubwpList) + UL_BWP->n_ul_bwp = ubwpList->list.count; if (ubwpList && UL_BWP->bwp_id > 0) { for (int i = 0; i < ubwpList->list.count; i++) { bwp_uplink = ubwpList->list.array[i]; @@ -595,14 +640,26 @@ void configure_current_BWP(NR_UE_MAC_INST_t *mac, } AssertFatal(bwp_uplink != NULL,"Couldn't find ULBWP corresponding to BWP ID %ld\n",UL_BWP->bwp_id); ul_genericParameters = bwp_uplink->bwp_Common->genericParameters; + UL_BWP->tdaList_Common = bwp_uplink->bwp_Common->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList; + UL_BWP->pusch_Config = bwp_uplink->bwp_Dedicated->pusch_Config->choice.setup; + UL_BWP->pucch_Config = bwp_uplink->bwp_Dedicated->pucch_Config->choice.setup; + UL_BWP->srs_Config = bwp_uplink->bwp_Dedicated->srs_Config->choice.setup; + UL_BWP->configuredGrantConfig = bwp_uplink->bwp_Dedicated->configuredGrantConfig ? bwp_uplink->bwp_Dedicated->configuredGrantConfig->choice.setup : NULL; + if (bwp_uplink->bwp_Common->pucch_ConfigCommon) + UL_BWP->pucch_ConfigCommon = bwp_uplink->bwp_Common->pucch_ConfigCommon->choice.setup; + if (bwp_uplink->bwp_Common->rach_ConfigCommon) + UL_BWP->rach_ConfigCommon = bwp_uplink->bwp_Common->rach_ConfigCommon->choice.setup; } else { - if(mac->scc) { - ul_genericParameters = mac->scc->uplinkConfigCommon->initialUplinkBWP->genericParameters; - } - if(mac->scc_SIB) { - ul_genericParameters = mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.genericParameters; - } + UL_BWP->tdaList_Common = bwp_ulcommon->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList; + UL_BWP->pusch_Config = spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pusch_Config->choice.setup; + UL_BWP->pucch_Config = spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup; + UL_BWP->srs_Config = spCellConfigDedicated->uplinkConfig->initialUplinkBWP->srs_Config->choice.setup; + UL_BWP->configuredGrantConfig = + spCellConfigDedicated->uplinkConfig->initialUplinkBWP->configuredGrantConfig ? spCellConfigDedicated->uplinkConfig->initialUplinkBWP->configuredGrantConfig->choice.setup : NULL; + ul_genericParameters = bwp_ulcommon->genericParameters; + if (bwp_ulcommon->rach_ConfigCommon) + UL_BWP->rach_ConfigCommon = bwp_ulcommon->rach_ConfigCommon->choice.setup; } } else diff --git a/openair2/LAYER2/NR_MAC_UE/mac_proto.h b/openair2/LAYER2/NR_MAC_UE/mac_proto.h index bf6b89f71c..eb773672d4 100644 --- a/openair2/LAYER2/NR_MAC_UE/mac_proto.h +++ b/openair2/LAYER2/NR_MAC_UE/mac_proto.h @@ -251,29 +251,8 @@ void config_dci_pdu(NR_UE_MAC_INST_t *mac, fapi_nr_dl_config_dci_dl_pdu_rel15_t void ue_dci_configuration(NR_UE_MAC_INST_t *mac, fapi_nr_dl_config_request_t *dl_config, frame_t frame, int slot); -void get_bwp_info(NR_UE_MAC_INST_t *mac, - int dl_bwp_id, - int ul_bwp_id, - NR_BWP_DownlinkDedicated_t **bwpd, - NR_BWP_DownlinkCommon_t **bwpc, - NR_BWP_UplinkDedicated_t **ubwpd, - NR_BWP_UplinkCommon_t **ubwpc); - NR_BWP_DownlinkCommon_t *get_bwp_downlink_common(NR_UE_MAC_INST_t *mac, NR_BWP_Id_t dl_bwp_id); -NR_PUSCH_TimeDomainResourceAllocationList_t *choose_ul_tda_list(const NR_PUSCH_Config_t *pusch_Config,NR_PUSCH_ConfigCommon_t *pusch_ConfigCommon); -NR_PDSCH_TimeDomainResourceAllocationList_t *choose_dl_tda_list(NR_PDSCH_Config_t *pdsch_Config,NR_PDSCH_ConfigCommon_t *pdsch_ConfigCommon); - -int8_t nr_ue_process_dci_time_dom_resource_assignment(NR_UE_MAC_INST_t *mac, - NR_PUSCH_TimeDomainResourceAllocationList_t *pusch_TimeDomainAllocationList, - NR_PDSCH_TimeDomainResourceAllocationList_t *pdsch_TimeDomainAllocationList, - nfapi_nr_ue_pusch_pdu_t *pusch_config_pdu, - fapi_nr_dl_config_dlsch_pdu_rel15_t *dlsch_config_pdu, - int *mappingtype, - uint8_t time_domain_ind, - int default_abc, - bool use_default); - uint8_t nr_ue_get_sdu(module_id_t module_idP, int cc_id, frame_t frameP, @@ -475,7 +454,7 @@ int16_t compute_nr_SSB_PL(NR_UE_MAC_INST_t *mac, short ssb_rsrp_dBm); // - in which ULSCH should be scheduled. K2 is configured in RRC configuration. // PUSCH Msg3 scheduler: // - scheduled by RAR UL grant according to 8.3 of TS 38.213 -int nr_ue_pusch_scheduler(NR_UE_MAC_INST_t *mac, uint8_t is_Msg3, frame_t current_frame, int current_slot, frame_t *frame_tx, int *slot_tx, uint8_t tda_id); +int nr_ue_pusch_scheduler(NR_UE_MAC_INST_t *mac, uint8_t is_Msg3, frame_t current_frame, int current_slot, frame_t *frame_tx, int *slot_tx, long k2); int get_rnti_type(NR_UE_MAC_INST_t *mac, uint16_t rnti); @@ -487,11 +466,6 @@ int get_rnti_type(NR_UE_MAC_INST_t *mac, uint16_t rnti); // - 6.1.4.2 of TS 38.214 // - 6.4.1.1.1 of TS 38.211 // - 6.3.1.7 of 38.211 -int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, - nfapi_nr_ue_pusch_pdu_t *pusch_config_pdu, - dci_pdu_rel15_t *dci, - RAR_grant_t *rar_grant, - uint16_t rnti, - uint8_t *dci_format); +int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, NR_tda_info_t *tda_info, nfapi_nr_ue_pusch_pdu_t *pusch_config_pdu, dci_pdu_rel15_t *dci, RAR_grant_t *rar_grant, uint16_t rnti, uint8_t *dci_format); #endif /** @}*/ diff --git a/openair2/LAYER2/NR_MAC_UE/nr_ra_procedures.c b/openair2/LAYER2/NR_MAC_UE/nr_ra_procedures.c index e4bbb546d8..f871796230 100644 --- a/openair2/LAYER2/NR_MAC_UE/nr_ra_procedures.c +++ b/openair2/LAYER2/NR_MAC_UE/nr_ra_procedures.c @@ -138,8 +138,8 @@ void init_RA(module_id_t mod_id, NR_PRACH_RESOURCES_t *prach_resources, NR_RACH_ConfigCommon_t *nr_rach_ConfigCommon, NR_RACH_ConfigGeneric_t *rach_ConfigGeneric, - NR_RACH_ConfigDedicated_t *rach_ConfigDedicated) { - + NR_RACH_ConfigDedicated_t *rach_ConfigDedicated) +{ NR_UE_MAC_INST_t *mac = get_mac_inst(mod_id); RA_config_t *ra = &mac->ra; @@ -278,16 +278,13 @@ void init_RA(module_id_t mod_id, } } - /* TS 38.321 subclause 7.3 - return DELTA_PREAMBLE values in dB */ -int8_t nr_get_DELTA_PREAMBLE(module_id_t mod_id, int CC_id, uint16_t prach_format){ - +int8_t nr_get_DELTA_PREAMBLE(module_id_t mod_id, int CC_id, uint16_t prach_format) +{ NR_UE_MAC_INST_t *mac = get_mac_inst(mod_id); - NR_RACH_ConfigCommon_t *nr_rach_ConfigCommon = (mac->scc!=NULL) ? - mac->scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup: - mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.rach_ConfigCommon->choice.setup; + NR_RACH_ConfigCommon_t *nr_rach_ConfigCommon = mac->current_UL_BWP.rach_ConfigCommon; NR_SubcarrierSpacing_t scs = *nr_rach_ConfigCommon->msg1_SubcarrierSpacing; - int prach_sequence_length = (mac->scc!=NULL)?(mac->scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup->prach_RootSequenceIndex.present - 1) : (mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.rach_ConfigCommon->choice.setup->prach_RootSequenceIndex.present-1); + int prach_sequence_length = nr_rach_ConfigCommon->prach_RootSequenceIndex.present - 1; uint8_t prachConfigIndex, mu; AssertFatal(CC_id == 0, "Transmission on secondary CCs is not supported yet\n"); @@ -385,13 +382,13 @@ int8_t nr_get_DELTA_PREAMBLE(module_id_t mod_id, int CC_id, uint16_t prach_forma // - RA_PREAMBLE_POWER_RAMPING_STEP dB // - POWER_OFFSET_2STEP_RA dB // returns receivedTargerPower in dBm -int nr_get_Po_NOMINAL_PUSCH(NR_PRACH_RESOURCES_t *prach_resources, module_id_t mod_id, uint8_t CC_id){ - +int nr_get_Po_NOMINAL_PUSCH(NR_PRACH_RESOURCES_t *prach_resources, module_id_t mod_id, uint8_t CC_id) +{ NR_UE_MAC_INST_t *mac = get_mac_inst(mod_id); int8_t receivedTargerPower; int8_t delta_preamble; - NR_RACH_ConfigCommon_t *nr_rach_ConfigCommon = (mac->scc != NULL) ? mac->scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup: mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.rach_ConfigCommon->choice.setup; + NR_RACH_ConfigCommon_t *nr_rach_ConfigCommon = mac->current_UL_BWP.rach_ConfigCommon; long preambleReceivedTargetPower = nr_rach_ConfigCommon->rach_ConfigGeneric.preambleReceivedTargetPower; delta_preamble = nr_get_DELTA_PREAMBLE(mod_id, CC_id, prach_resources->prach_format); @@ -402,8 +399,8 @@ int nr_get_Po_NOMINAL_PUSCH(NR_PRACH_RESOURCES_t *prach_resources, module_id_t m return receivedTargerPower; } -void ssb_rach_config(RA_config_t *ra, NR_PRACH_RESOURCES_t *prach_resources, NR_RACH_ConfigCommon_t *nr_rach_ConfigCommon){ - +void ssb_rach_config(RA_config_t *ra, NR_PRACH_RESOURCES_t *prach_resources, NR_RACH_ConfigCommon_t *nr_rach_ConfigCommon) +{ // Determine the SSB to RACH mapping ratio // ======================================= @@ -489,14 +486,11 @@ void ra_preambles_config(NR_PRACH_RESOURCES_t *prach_resources, NR_UE_MAC_INST_t unsigned short int *seed = (unsigned short int*)&tmp; RA_config_t *ra = &mac->ra; - NR_RACH_ConfigCommon_t *setup; - if (mac->scc) setup = mac->scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup; - else setup = mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.rach_ConfigCommon->choice.setup; + NR_RACH_ConfigCommon_t *setup = mac->current_UL_BWP.rach_ConfigCommon; NR_RACH_ConfigGeneric_t *rach_ConfigGeneric = &setup->rach_ConfigGeneric; - NR_BWP_UplinkCommon_t *initialUplinkBWP = (mac->scc) ? mac->scc->uplinkConfigCommon->initialUplinkBWP : &mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP; - if (initialUplinkBWP->pusch_ConfigCommon->choice.setup->msg3_DeltaPreamble){ - deltaPreamble_Msg3 = (*initialUplinkBWP->pusch_ConfigCommon->choice.setup->msg3_DeltaPreamble) * 2; // dB + if (mac->current_UL_BWP.msg3_DeltaPreamble) { + deltaPreamble_Msg3 = (*mac->current_UL_BWP.msg3_DeltaPreamble) * 2; // dB LOG_D(MAC, "In %s: deltaPreamble_Msg3 set to %ld\n", __FUNCTION__, deltaPreamble_Msg3); } @@ -661,9 +655,7 @@ void nr_get_prach_resources(module_id_t mod_id, NR_UE_MAC_INST_t *mac = get_mac_inst(mod_id); RA_config_t *ra = &mac->ra; - NR_RACH_ConfigCommon_t *nr_rach_ConfigCommon = (mac->scc)? - mac->scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup : - mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.rach_ConfigCommon->choice.setup; + NR_RACH_ConfigCommon_t *nr_rach_ConfigCommon = mac->current_UL_BWP.rach_ConfigCommon; LOG_D(MAC, "In %s: getting PRACH resources frame (first_Msg3 %d)\n", __FUNCTION__, ra->first_Msg3); @@ -701,12 +693,8 @@ void nr_Msg3_transmitted(module_id_t mod_id, uint8_t CC_id, frame_t frameP, slot NR_UE_MAC_INST_t *mac = get_mac_inst(mod_id); RA_config_t *ra = &mac->ra; - NR_RACH_ConfigCommon_t *nr_rach_ConfigCommon = (mac->scc) ? - mac->scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup: - mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.rach_ConfigCommon->choice.setup; - long mu = (mac->scc) ? - mac->scc->downlinkConfigCommon->frequencyInfoDL->scs_SpecificCarrierList.list.array[0]->subcarrierSpacing : - mac->scc_SIB->downlinkConfigCommon.frequencyInfoDL.scs_SpecificCarrierList.list.array[0]->subcarrierSpacing; + NR_RACH_ConfigCommon_t *nr_rach_ConfigCommon = mac->current_UL_BWP.rach_ConfigCommon; + long mu = mac->current_UL_BWP.scs; int subframes_per_slot = nr_slots_per_frame[mu]/10; // start contention resolution timer (cnt in slots) @@ -934,9 +922,7 @@ void nr_get_RA_window(NR_UE_MAC_INST_t *mac){ uint8_t mu, ra_ResponseWindow; RA_config_t *ra = &mac->ra; - NR_RACH_ConfigCommon_t *setup; - if (mac->scc) setup = mac->scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup; - else setup = mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.rach_ConfigCommon->choice.setup; + NR_RACH_ConfigCommon_t *setup = mac->current_UL_BWP.rach_ConfigCommon; AssertFatal(&setup->rach_ConfigGeneric != NULL, "In %s: FATAL! rach_ConfigGeneric is NULL...\n", __FUNCTION__); NR_RACH_ConfigGeneric_t *rach_ConfigGeneric = &setup->rach_ConfigGeneric; long scs = (mac->scc) ? diff --git a/openair2/LAYER2/NR_MAC_UE/nr_ue_dci_configuration.c b/openair2/LAYER2/NR_MAC_UE/nr_ue_dci_configuration.c index 2815a42c16..9d691514ed 100644 --- a/openair2/LAYER2/NR_MAC_UE/nr_ue_dci_configuration.c +++ b/openair2/LAYER2/NR_MAC_UE/nr_ue_dci_configuration.c @@ -93,20 +93,9 @@ void config_dci_pdu(NR_UE_MAC_INST_t *mac, fapi_nr_dl_config_dci_dl_pdu_rel15_t uint16_t monitoringSymbolsWithinSlot = 0; int sps = 0; - AssertFatal(mac->scc == NULL || mac->scc_SIB == NULL, "both scc and scc_SIB cannot be non-null\n"); - NR_UE_DL_BWP_t *current_DL_BWP = &mac->current_DL_BWP; NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; NR_BWP_Id_t dl_bwp_id = current_DL_BWP ? current_DL_BWP->bwp_id : 0; - NR_ServingCellConfigCommon_t *scc = mac->scc; - NR_ServingCellConfigCommonSIB_t *scc_SIB = mac->scc_SIB; - NR_BWP_DownlinkCommon_t *initialDownlinkBWP=NULL; - NR_BWP_UplinkCommon_t *initialUplinkBWP=NULL; - - if (scc!=NULL || scc_SIB != NULL) { - initialDownlinkBWP = scc!=NULL ? scc->downlinkConfigCommon->initialDownlinkBWP : &scc_SIB->downlinkConfigCommon.initialDownlinkBWP; - initialUplinkBWP = scc!=NULL ? scc->uplinkConfigCommon->initialUplinkBWP : &scc_SIB->uplinkConfigCommon->initialUplinkBWP; - } NR_SearchSpace_t *ss; NR_ControlResourceSet_t *coreset; @@ -173,24 +162,24 @@ void config_dci_pdu(NR_UE_MAC_INST_t *mac, fapi_nr_dl_config_dci_dl_pdu_rel15_t // computing alternative size for padding dci_pdu_rel15_t temp_pdu; if(dci_format == NR_DL_DCI_FORMAT_1_0) - alt_size = nr_dci_size(initialDownlinkBWP,initialUplinkBWP, - current_DL_BWP, current_UL_BWP, - mac->cg, &temp_pdu, - NR_UL_DCI_FORMAT_0_0, rnti_type, coreset_id, dl_bwp_id, - ss->searchSpaceType->present, mac->type0_PDCCH_CSS_config.num_rbs, 0); + alt_size = + nr_dci_size(current_DL_BWP, current_UL_BWP, mac->cg, &temp_pdu, NR_UL_DCI_FORMAT_0_0, rnti_type, coreset, dl_bwp_id, ss->searchSpaceType->present, mac->type0_PDCCH_CSS_config.num_rbs, 0); if(dci_format == NR_UL_DCI_FORMAT_0_0) - alt_size = nr_dci_size(initialDownlinkBWP,initialUplinkBWP, - current_DL_BWP, current_UL_BWP, - mac->cg, &temp_pdu, - NR_DL_DCI_FORMAT_1_0, rnti_type, coreset_id, dl_bwp_id, - ss->searchSpaceType->present, mac->type0_PDCCH_CSS_config.num_rbs, 0); + alt_size = + nr_dci_size(current_DL_BWP, current_UL_BWP, mac->cg, &temp_pdu, NR_DL_DCI_FORMAT_1_0, rnti_type, coreset, dl_bwp_id, ss->searchSpaceType->present, mac->type0_PDCCH_CSS_config.num_rbs, 0); } - rel15->dci_length_options[i] = nr_dci_size(initialDownlinkBWP,initialUplinkBWP, - current_DL_BWP, current_UL_BWP, - mac->cg, &mac->def_dci_pdu_rel15[dci_format], - dci_format, NR_RNTI_TC, coreset_id, dl_bwp_id, - ss->searchSpaceType->present, mac->type0_PDCCH_CSS_config.num_rbs, alt_size); + rel15->dci_length_options[i] = nr_dci_size(current_DL_BWP, + current_UL_BWP, + mac->cg, + &mac->def_dci_pdu_rel15[dci_format], + dci_format, + NR_RNTI_TC, + coreset, + dl_bwp_id, + ss->searchSpaceType->present, + mac->type0_PDCCH_CSS_config.num_rbs, + alt_size); rel15->BWPStart = coreset_id == 0 ? mac->type0_PDCCH_CSS_config.cset_start_rb : current_DL_BWP->BWPStart; rel15->BWPSize = coreset_id == 0 ? mac->type0_PDCCH_CSS_config.num_rbs : current_DL_BWP->BWPSize; diff --git a/openair2/LAYER2/NR_MAC_UE/nr_ue_procedures.c b/openair2/LAYER2/NR_MAC_UE/nr_ue_procedures.c index 17dde4c9c9..02ad48a655 100644 --- a/openair2/LAYER2/NR_MAC_UE/nr_ue_procedures.c +++ b/openair2/LAYER2/NR_MAC_UE/nr_ue_procedures.c @@ -180,51 +180,6 @@ void nr_ue_init_mac(module_id_t module_idP) { } } -void get_bwp_info(NR_UE_MAC_INST_t *mac, - int dl_bwp_id, - int ul_bwp_id, - NR_BWP_DownlinkDedicated_t **bwpd, - NR_BWP_DownlinkCommon_t **bwpc, - NR_BWP_UplinkDedicated_t **ubwpd, - NR_BWP_UplinkCommon_t **ubwpc) { - - if (dl_bwp_id > 0) { - AssertFatal(mac->DLbwp[dl_bwp_id-1]!=NULL,"mac->DLbwp[%d] is null, shouldn't be\n", (int)dl_bwp_id-1); - *bwpd = mac->DLbwp[dl_bwp_id-1]->bwp_Dedicated; - } else { - if (mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->initialDownlinkBWP) - *bwpd = mac->cg->spCellConfig->spCellConfigDedicated->initialDownlinkBWP; - } - - *bwpc = get_bwp_downlink_common(mac, dl_bwp_id); - AssertFatal(*bwpc!=NULL,"bwpc shouldn't be null\n"); - - if (ul_bwp_id > 0) { - AssertFatal(mac->ULbwp[ul_bwp_id-1]!=NULL,"mac->ULbwp[%d] is null, shouldn't be\n", - ul_bwp_id-1); - *ubwpd = mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated; - if (mac->ULbwp[ul_bwp_id-1]->bwp_Common) *ubwpc = mac->ULbwp[ul_bwp_id-1]->bwp_Common; - else if (mac->scc) *ubwpc = mac->scc->uplinkConfigCommon->initialUplinkBWP; - else if (mac->scc_SIB) *ubwpc = &mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP; - AssertFatal(*bwpc!=NULL,"bwpc shouldn't be null\n"); - - } - else { - if (mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP) - *ubwpd = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP; - if (mac->scc) *ubwpc = mac->scc->uplinkConfigCommon->initialUplinkBWP; - else if (mac->scc_SIB) *ubwpc = &mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP; - AssertFatal(*ubwpc!=NULL,"ubwpc shouldn't be null\n"); - } -} - NR_BWP_DownlinkCommon_t *get_bwp_downlink_common(NR_UE_MAC_INST_t *mac, NR_BWP_Id_t dl_bwp_id) { NR_BWP_DownlinkCommon_t *bwp_Common = NULL; if (dl_bwp_id > 0 && mac->cg->spCellConfig->spCellConfigDedicated->downlinkBWP_ToAddModList) { @@ -237,33 +192,8 @@ NR_BWP_DownlinkCommon_t *get_bwp_downlink_common(NR_UE_MAC_INST_t *mac, NR_BWP_I return bwp_Common; } -NR_PDSCH_TimeDomainResourceAllocationList_t *choose_dl_tda_list(NR_PDSCH_Config_t *pdsch_Config,NR_PDSCH_ConfigCommon_t *pdsch_ConfigCommon) { - - NR_PDSCH_TimeDomainResourceAllocationList_t *pdsch_TimeDomainAllocationList=NULL; - - if (pdsch_Config && - pdsch_Config->pdsch_TimeDomainAllocationList) - pdsch_TimeDomainAllocationList = pdsch_Config->pdsch_TimeDomainAllocationList->choice.setup; - else if (pdsch_ConfigCommon->pdsch_TimeDomainAllocationList) - pdsch_TimeDomainAllocationList = pdsch_ConfigCommon->pdsch_TimeDomainAllocationList; - - return(pdsch_TimeDomainAllocationList); -} - -NR_PUSCH_TimeDomainResourceAllocationList_t *choose_ul_tda_list(const NR_PUSCH_Config_t *pusch_Config,NR_PUSCH_ConfigCommon_t *pusch_ConfigCommon) { - - NR_PUSCH_TimeDomainResourceAllocationList_t *pusch_TimeDomainAllocationList=NULL; - - if (pusch_Config && - pusch_Config->pusch_TimeDomainAllocationList) - pusch_TimeDomainAllocationList = pusch_Config->pusch_TimeDomainAllocationList->choice.setup; - else if (pusch_ConfigCommon->pusch_TimeDomainAllocationList) - pusch_TimeDomainAllocationList = pusch_ConfigCommon->pusch_TimeDomainAllocationList; - - return(pusch_TimeDomainAllocationList); -} - -int get_rnti_type(NR_UE_MAC_INST_t *mac, uint16_t rnti){ +int get_rnti_type(NR_UE_MAC_INST_t *mac, uint16_t rnti) +{ RA_config_t *ra = &mac->ra; int rnti_type; @@ -481,138 +411,6 @@ int8_t nr_ue_process_dci_freq_dom_resource_assignment(nfapi_nr_ue_pusch_pdu_t *p return 0; } -int8_t nr_ue_process_dci_time_dom_resource_assignment(NR_UE_MAC_INST_t *mac, - NR_PUSCH_TimeDomainResourceAllocationList_t *pusch_TimeDomainAllocationList, - NR_PDSCH_TimeDomainResourceAllocationList_t *pdsch_TimeDomainAllocationList, - nfapi_nr_ue_pusch_pdu_t *pusch_config_pdu, - fapi_nr_dl_config_dlsch_pdu_rel15_t *dlsch_config_pdu, - int *mapping_type, - uint8_t time_domain_ind, - int default_abc, - bool use_default){ - - int dmrs_typeA_pos = (mac->scc != NULL) ? mac->scc->dmrs_TypeA_Position : mac->mib->dmrs_TypeA_Position; - -// uint8_t k_offset=0; - int sliv_S=0; - int sliv_L=0; - uint8_t mu_pusch = 1; - - // definition table j Table 6.1.2.1.1-4 - uint8_t j = (mu_pusch==3)?3:(mu_pusch==2)?2:1; - uint8_t table_6_1_2_1_1_2_time_dom_res_alloc_A[16][3]={ // for PUSCH from TS 38.214 subclause 6.1.2.1.1 - {j, 0,14}, // row index 1 - {j, 0,12}, // row index 2 - {j, 0,10}, // row index 3 - {j, 2,10}, // row index 4 - {j, 4,10}, // row index 5 - {j, 4,8}, // row index 6 - {j, 4,6}, // row index 7 - {j+1,0,14}, // row index 8 - {j+1,0,12}, // row index 9 - {j+1,0,10}, // row index 10 - {j+2,0,14}, // row index 11 - {j+2,0,12}, // row index 12 - {j+2,0,10}, // row index 13 - {j, 8,6}, // row index 14 - {j+3,0,14}, // row index 15 - {j+3,0,10} // row index 16 - }; - /*uint8_t table_6_1_2_1_1_3_time_dom_res_alloc_A_extCP[16][3]={ // for PUSCH from TS 38.214 subclause 6.1.2.1.1 - {j, 0,8}, // row index 1 - {j, 0,12}, // row index 2 - {j, 0,10}, // row index 3 - {j, 2,10}, // row index 4 - {j, 4,4}, // row index 5 - {j, 4,8}, // row index 6 - {j, 4,6}, // row index 7 - {j+1,0,8}, // row index 8 - {j+1,0,12}, // row index 9 - {j+1,0,10}, // row index 10 - {j+2,0,6}, // row index 11 - {j+2,0,12}, // row index 12 - {j+2,0,10}, // row index 13 - {j, 8,4}, // row index 14 - {j+3,0,8}, // row index 15 - {j+3,0,10} // row index 16 - };*/ - - /* - * TS 38.214 subclause 5.1.2.1 Resource allocation in time domain (downlink) - */ - if(dlsch_config_pdu != NULL){ - if (pdsch_TimeDomainAllocationList && use_default==false) { - - if (time_domain_ind >= pdsch_TimeDomainAllocationList->list.count) { - LOG_E(MAC, "time_domain_ind %d >= pdsch->TimeDomainAllocationList->list.count %d\n", - time_domain_ind, pdsch_TimeDomainAllocationList->list.count); - dlsch_config_pdu->start_symbol = 0; - dlsch_config_pdu->number_symbols = 0; - return -1; - } - - int startSymbolAndLength = pdsch_TimeDomainAllocationList->list.array[time_domain_ind]->startSymbolAndLength; - int S,L; - SLIV2SL(startSymbolAndLength,&S,&L); - dlsch_config_pdu->start_symbol=S; - dlsch_config_pdu->number_symbols=L; - - LOG_D(MAC,"SLIV = %i\n", startSymbolAndLength); - LOG_D(MAC,"start_symbol = %i\n", dlsch_config_pdu->start_symbol); - LOG_D(MAC,"number_symbols = %i\n", dlsch_config_pdu->number_symbols); - - } - else {// Default configuration from tables - - bool is_typeA; - get_info_from_tda_tables(default_abc, - time_domain_ind, - dmrs_typeA_pos, - 1, // normal CP - &is_typeA, - &sliv_S, - &sliv_L); - *mapping_type = is_typeA? typeA : typeB; - dlsch_config_pdu->number_symbols = sliv_L; - dlsch_config_pdu->start_symbol = sliv_S; - } - } /* - * TS 38.214 subclause 6.1.2.1 Resource allocation in time domain (uplink) - */ - if(pusch_config_pdu != NULL){ - if (pusch_TimeDomainAllocationList && use_default==false) { - if (time_domain_ind >= pusch_TimeDomainAllocationList->list.count) { - LOG_E(NR_MAC, "time_domain_ind %d >= pusch->TimeDomainAllocationList->list.count %d\n", - time_domain_ind, pusch_TimeDomainAllocationList->list.count); - pusch_config_pdu->start_symbol_index=0; - pusch_config_pdu->nr_of_symbols=0; - return -1; - } - - LOG_D(NR_MAC,"Filling Time-Domain Allocation from pusch_TimeDomainAllocationList\n"); - int startSymbolAndLength = pusch_TimeDomainAllocationList->list.array[time_domain_ind]->startSymbolAndLength; - int S,L; - SLIV2SL(startSymbolAndLength,&S,&L); - pusch_config_pdu->start_symbol_index=S; - pusch_config_pdu->nr_of_symbols=L; - } - else { - LOG_D(NR_MAC,"Filling Time-Domain Allocation from tables\n"); -// k_offset = table_6_1_2_1_1_2_time_dom_res_alloc_A[time_domain_ind-1][0]; - sliv_S = table_6_1_2_1_1_2_time_dom_res_alloc_A[time_domain_ind][1]; - sliv_L = table_6_1_2_1_1_2_time_dom_res_alloc_A[time_domain_ind][2]; - // k_offset = table_6_1_2_1_1_3_time_dom_res_alloc_A_extCP[nr_pdci_info_extracted->time_dom_resource_assignment][0]; - // sliv_S = table_6_1_2_1_1_3_time_dom_res_alloc_A_extCP[nr_pdci_info_extracted->time_dom_resource_assignment][1]; - // sliv_L = table_6_1_2_1_1_3_time_dom_res_alloc_A_extCP[nr_pdci_info_extracted->time_dom_resource_assignment][2]; - pusch_config_pdu->nr_of_symbols = sliv_L; - pusch_config_pdu->start_symbol_index = sliv_S; - } - LOG_D(NR_MAC,"start_symbol = %i\n", pusch_config_pdu->start_symbol_index); - LOG_D(NR_MAC,"number_symbols = %i\n", pusch_config_pdu->nr_of_symbols); - } - return 0; -} - int nr_ue_process_dci_indication_pdu(module_id_t module_id,int cc_id, int gNB_index, frame_t frame, int slot, fapi_nr_dci_indication_pdu_t *dci) { NR_UE_MAC_INST_t *mac = get_mac_inst(module_id); @@ -634,6 +432,7 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr uint16_t rnti = dci_ind->rnti; uint8_t dci_format = dci_ind->dci_format; + int coreset_type = dci_ind->coreset_type == NFAPI_NR_CSET_CONFIG_PDCCH_CONFIG; // 0 for coreset0, 1 otherwise int ret = 0; int pucch_res_set_cnt = 0, valid = 0; frame_t frame_tx = 0; @@ -645,11 +444,11 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr uint8_t is_Msg3 = 0; NR_UE_DL_BWP_t *current_DL_BWP = &mac->current_DL_BWP; NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; - NR_BWP_Id_t dl_bwp_id = current_DL_BWP->bwp_id; - NR_BWP_Id_t ul_bwp_id = current_UL_BWP->bwp_id; - int default_abc = 1; + int mux_pattern = 1; LOG_D(MAC, "In %s: Processing received DCI format %s\n", __FUNCTION__, dci_formats[dci_format]); + NR_tda_info_t tda_info = {0}; + NR_PUCCH_Config_t *pucch_Config = current_UL_BWP->pucch_Config; switch(dci_format){ case NR_UL_DCI_FORMAT_0_0: { @@ -674,7 +473,11 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr // - SUL_IND_0_0 // Schedule PUSCH - ret = nr_ue_pusch_scheduler(mac, is_Msg3, frame, slot, &frame_tx, &slot_tx, dci->time_domain_assignment.val); + tda_info = get_ul_tda_info(current_UL_BWP, coreset_type, dci_ind->ss_type, get_rnti_type(mac, rnti), dci->time_domain_assignment.val); + if (tda_info.nrOfSymbols == 0) + ret = -1; + else + ret = nr_ue_pusch_scheduler(mac, is_Msg3, frame, slot, &frame_tx, &slot_tx, tda_info.k2); if (ret != -1){ @@ -693,9 +496,9 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr pthread_mutex_unlock(&ul_config->mutex_ul_config); // Config PUSCH PDU - ret = nr_config_pusch_pdu(mac, pusch_config_pdu, dci, NULL, rnti, &dci_format); - } - + ret = nr_config_pusch_pdu(mac, &tda_info, pusch_config_pdu, dci, NULL, rnti, &dci_format); + } else + LOG_E(MAC, "Cannot schedule PUSCH\n"); break; } @@ -734,7 +537,11 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr // - SRS_RESOURCE_IND // Schedule PUSCH - ret = nr_ue_pusch_scheduler(mac, is_Msg3, frame, slot, &frame_tx, &slot_tx, dci->time_domain_assignment.val); + tda_info = get_ul_tda_info(current_UL_BWP, coreset_type, dci_ind->ss_type, get_rnti_type(mac, rnti), dci->time_domain_assignment.val); + if (tda_info.nrOfSymbols == 0) + ret = -1; + else + ret = nr_ue_pusch_scheduler(mac, is_Msg3, frame, slot, &frame_tx, &slot_tx, tda_info.k2); if (ret != -1){ @@ -755,8 +562,9 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr pthread_mutex_unlock(&ul_config->mutex_ul_config); // Config PUSCH PDU - ret = nr_config_pusch_pdu(mac, pusch_config_pdu, dci, NULL, rnti, &dci_format); - } else AssertFatal(1==0,"Cannot schedule PUSCH\n"); + ret = nr_config_pusch_pdu(mac, &tda_info, pusch_config_pdu, dci, NULL, rnti, &dci_format); + } else + LOG_E(MAC, "Cannot schedule PUSCH\n"); break; } @@ -815,8 +623,8 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr dl_config->dl_config_list[dl_config->number_pdus].dlsch_config_pdu.rnti = rnti; fapi_nr_dl_config_dlsch_pdu_rel15_t *dlsch_config_pdu_1_0 = &dl_config->dl_config_list[dl_config->number_pdus].dlsch_config_pdu.dlsch_config_rel15; - NR_PDSCH_Config_t *pdsch_config= (dl_bwp_id>0 && mac->DLbwp[dl_bwp_id-1]) ? mac->DLbwp[dl_bwp_id-1]->bwp_Dedicated->pdsch_Config->choice.setup : NULL; - int is_common=0; + NR_PDSCH_Config_t *pdsch_config = current_DL_BWP ? current_DL_BWP->pdsch_Config : NULL; + int is_common = 0; if (dci_ind->ss_type == NR_SearchSpace__searchSpaceType_PR_common) { dlsch_config_pdu_1_0->BWPSize = mac->type0_PDCCH_CSS_config.num_rbs ? mac->type0_PDCCH_CSS_config.num_rbs : current_DL_BWP->initial_BWPSize; dlsch_config_pdu_1_0->BWPStart = dci_ind->cset_start; @@ -827,7 +635,7 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr } if(rnti == SI_RNTI) { NR_Type0_PDCCH_CSS_config_t type0_PDCCH_CSS_config = mac->type0_PDCCH_CSS_config; - default_abc = type0_PDCCH_CSS_config.type0_pdcch_ss_mux_pattern; + mux_pattern = type0_PDCCH_CSS_config.type0_pdcch_ss_mux_pattern; dl_config->dl_config_list[dl_config->number_pdus].pdu_type = FAPI_NR_DL_CONFIG_TYPE_SI_DLSCH; dlsch_config_pdu_1_0->SubcarrierSpacing = mac->mib->subCarrierSpacingCommon; if (pdsch_config) pdsch_config->dmrs_DownlinkForPDSCH_MappingTypeA->choice.setup->dmrs_AdditionalPosition = NULL; // For PDSCH with mapping type A, the UE shall assume dmrs-AdditionalPosition='pos2' @@ -838,17 +646,10 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr } else { dl_config->dl_config_list[dl_config->number_pdus].pdu_type = FAPI_NR_DL_CONFIG_TYPE_DLSCH; } - if( (ra->RA_window_cnt >= 0 && rnti == ra->ra_rnti) || (rnti == ra->t_crnti) ) { - if (mac->scc == NULL) { // use coreset0 - is_common=1; - } - if (!get_softmodem_params()->sa) { // NSA mode is not using the Initial BWP - pdsch_config = mac->DLbwp[dl_bwp_id-1]->bwp_Dedicated->pdsch_Config->choice.setup; - } - } else if (dl_bwp_id>0 && mac->DLbwp[dl_bwp_id-1]) { - pdsch_config = mac->DLbwp[dl_bwp_id-1]->bwp_Dedicated->pdsch_Config->choice.setup; - } else if (mac->scc) - pdsch_config = NULL; + if ((ra->RA_window_cnt >= 0 && rnti == ra->ra_rnti) || (rnti == ra->t_crnti)) { + if (mac->scc == NULL) // use coreset0 + is_common = 1; + } } /* IDENTIFIER_DCI_FORMATS */ @@ -858,36 +659,18 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr return -1; } - - NR_PDSCH_TimeDomainResourceAllocationList_t *pdsch_TimeDomainAllocationList = NULL; - if (dl_bwp_id>0 && - mac->DLbwp[dl_bwp_id-1] && - mac->DLbwp[dl_bwp_id-1]->bwp_Dedicated && - mac->DLbwp[dl_bwp_id-1]->bwp_Dedicated->pdsch_Config && - mac->DLbwp[dl_bwp_id-1]->bwp_Dedicated->pdsch_Config->choice.setup->pdsch_TimeDomainAllocationList) - pdsch_TimeDomainAllocationList = mac->DLbwp[dl_bwp_id-1]->bwp_Dedicated->pdsch_Config->choice.setup->pdsch_TimeDomainAllocationList->choice.setup; - else if (dl_bwp_id>0 && mac->DLbwp[dl_bwp_id-1] && mac->DLbwp[dl_bwp_id-1]->bwp_Common->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList) - pdsch_TimeDomainAllocationList = mac->DLbwp[dl_bwp_id-1]->bwp_Common->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList; - else if (mac->scc_SIB && mac->scc_SIB->downlinkConfigCommon.initialDownlinkBWP.pdsch_ConfigCommon->choice.setup) - pdsch_TimeDomainAllocationList = mac->scc_SIB->downlinkConfigCommon.initialDownlinkBWP.pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList; - - int mappingtype; /* TIME_DOM_RESOURCE_ASSIGNMENT */ - if (nr_ue_process_dci_time_dom_resource_assignment(mac,NULL,pdsch_TimeDomainAllocationList, - NULL,dlsch_config_pdu_1_0,&mappingtype, - dci->time_domain_assignment.val, - default_abc,rnti==SI_RNTI) < 0) { - LOG_W(MAC, "[%d.%d] Invalid time_domain_assignment. Possibly due to false DCI. Ignoring DCI!\n", frame, slot); - return -1; - } - if(pdsch_TimeDomainAllocationList && rnti!=SI_RNTI) - mappingtype = pdsch_TimeDomainAllocationList->list.array[dci->time_domain_assignment.val]->mappingType; + int dmrs_typeA_pos = (mac->scc != NULL) ? mac->scc->dmrs_TypeA_Position : mac->mib->dmrs_TypeA_Position; + // TODO need to differentiate SI_RNTI between SIB1 and other SIB + NR_tda_info_t tda_info = get_dl_tda_info(current_DL_BWP, dci_ind->ss_type, dci->time_domain_assignment.val, + dmrs_typeA_pos, mux_pattern, get_rnti_type(mac, rnti), coreset_type, rnti == SI_RNTI); + + dlsch_config_pdu_1_0->number_symbols = tda_info.nrOfSymbols; + dlsch_config_pdu_1_0->start_symbol = tda_info.startSymbolIndex; struct NR_DMRS_DownlinkConfig *dl_dmrs_config = NULL; - if(dl_bwp_id>0 && mac->DLbwp[dl_bwp_id-1] != NULL) - dl_dmrs_config = (mappingtype == typeA) ? - mac->DLbwp[dl_bwp_id-1]->bwp_Dedicated->pdsch_Config->choice.setup->dmrs_DownlinkForPDSCH_MappingTypeA->choice.setup : - mac->DLbwp[dl_bwp_id-1]->bwp_Dedicated->pdsch_Config->choice.setup->dmrs_DownlinkForPDSCH_MappingTypeB->choice.setup; + if (pdsch_config) + dl_dmrs_config = (tda_info.mapping_type == typeA) ? pdsch_config->dmrs_DownlinkForPDSCH_MappingTypeA->choice.setup : pdsch_config->dmrs_DownlinkForPDSCH_MappingTypeB->choice.setup; dlsch_config_pdu_1_0->nscid = 0; if(dl_dmrs_config && dl_dmrs_config->scramblingID0) @@ -901,7 +684,7 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr (get_softmodem_params()->nsa) ? mac->scc->dmrs_TypeA_Position : mac->mib->dmrs_TypeA_Position, dlsch_config_pdu_1_0->number_symbols, dlsch_config_pdu_1_0->start_symbol, - mappingtype, + tda_info.mapping_type, 1); dlsch_config_pdu_1_0->dmrsConfigType = (dl_dmrs_config != NULL) ? @@ -971,36 +754,16 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr if (dci->tpc == 3) dlsch_config_pdu_1_0->accumulated_delta_PUCCH = 3; // Sanity check for pucch_resource_indicator value received to check for false DCI. valid = 0; - if (ul_bwp_id > 0 && - mac->ULbwp[ul_bwp_id-1] && - mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated && - mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pucch_Config && - mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup&& - mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup->resourceSetToAddModList) { - pucch_res_set_cnt = mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup->resourceSetToAddModList->list.count; + if (pucch_Config && pucch_Config->resourceSetToAddModList) { + pucch_res_set_cnt = pucch_Config->resourceSetToAddModList->list.count; for (int id = 0; id < pucch_res_set_cnt; id++) { - if (dci->pucch_resource_indicator < mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup->resourceSetToAddModList->list.array[id]->resourceList.list.count) { - valid = 1; - break; - } - } - } - else if (mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup->resourceSetToAddModList){ - pucch_res_set_cnt = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup->resourceSetToAddModList->list.count; - for (int id = 0; id < pucch_res_set_cnt; id++) { - if (dci->pucch_resource_indicator < mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup->resourceSetToAddModList->list.array[id]->resourceList.list.count) { + if (dci->pucch_resource_indicator < pucch_Config->resourceSetToAddModList->list.array[id]->resourceList.list.count) { valid = 1; break; } } - } else valid=1; + } else + valid = 1; if (!valid) { LOG_W(MAC, "[%d.%d] pucch_resource_indicator value %d is out of bounds. Possibly due to false DCI. Ignoring DCI!\n", frame, slot, dci->pucch_resource_indicator); return -1; @@ -1084,16 +847,8 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr LOG_W(NR_MAC,"[%d.%d] bwp_indicator %d > NR_MAX_NUM_BWP Possibly due to false DCI. Ignoring DCI!\n", frame, slot,dci->bwp_indicator.val); return -1; } - NR_BWP_Id_t dl_bwp_id = mac->current_DL_BWP.bwp_id; - NR_BWP_Id_t ul_bwp_id = mac->current_UL_BWP.bwp_id; - NR_PDSCH_Config_t *pdsch_Config=NULL; - NR_BWP_DownlinkDedicated_t *bwpd=NULL; - NR_BWP_DownlinkCommon_t *bwpc=NULL; - NR_BWP_UplinkDedicated_t *ubwpd=NULL; - NR_BWP_UplinkCommon_t *ubwpc=NULL; - get_bwp_info(mac,dl_bwp_id,ul_bwp_id,&bwpd,&bwpc,&ubwpd,&ubwpc); + NR_PDSCH_Config_t *pdsch_Config = current_DL_BWP->pdsch_Config; - pdsch_Config = bwpd->pdsch_Config->choice.setup; dl_config->dl_config_list[dl_config->number_pdus].pdu_type = FAPI_NR_DL_CONFIG_TYPE_DLSCH; dl_config->dl_config_list[dl_config->number_pdus].dlsch_config_pdu.rnti = rnti; @@ -1113,18 +868,14 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr return -1; } /* TIME_DOM_RESOURCE_ASSIGNMENT */ - int mappingtype; - NR_PDSCH_TimeDomainResourceAllocationList_t *pdsch_TimeDomainAllocationList = choose_dl_tda_list(pdsch_Config,bwpc->pdsch_ConfigCommon->choice.setup); - if (nr_ue_process_dci_time_dom_resource_assignment(mac,NULL,pdsch_TimeDomainAllocationList, - NULL,dlsch_config_pdu_1_1,&mappingtype, - dci->time_domain_assignment.val,0,false) < 0) { - LOG_W(MAC, "[%d.%d] Invalid time_domain_assignment. Possibly due to false DCI. Ignoring DCI!\n", frame, slot); - return -1; - } - if(pdsch_TimeDomainAllocationList) - mappingtype = pdsch_TimeDomainAllocationList->list.array[dci->time_domain_assignment.val]->mappingType; + int dmrs_typeA_pos = (mac->scc != NULL) ? mac->scc->dmrs_TypeA_Position : mac->mib->dmrs_TypeA_Position; + NR_tda_info_t tda_info = get_dl_tda_info(current_DL_BWP, dci_ind->ss_type, dci->time_domain_assignment.val, + dmrs_typeA_pos, mux_pattern, get_rnti_type(mac, rnti), coreset_type, false); - struct NR_DMRS_DownlinkConfig *dl_dmrs_config = (mappingtype == typeA) ? + dlsch_config_pdu_1_1->number_symbols = tda_info.nrOfSymbols; + dlsch_config_pdu_1_1->start_symbol = tda_info.startSymbolIndex; + + struct NR_DMRS_DownlinkConfig *dl_dmrs_config = (tda_info.mapping_type == typeA) ? pdsch_Config->dmrs_DownlinkForPDSCH_MappingTypeA->choice.setup : pdsch_Config->dmrs_DownlinkForPDSCH_MappingTypeB->choice.setup; @@ -1194,9 +945,9 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr // Sanity check for pucch_resource_indicator value received to check for false DCI. valid = 0; - pucch_res_set_cnt = ubwpd->pucch_Config->choice.setup->resourceSetToAddModList->list.count; + pucch_res_set_cnt = pucch_Config->resourceSetToAddModList->list.count; for (int id = 0; id < pucch_res_set_cnt; id++) { - if (dci->pucch_resource_indicator < ubwpd->pucch_Config->choice.setup->resourceSetToAddModList->list.array[id]->resourceList.list.count) { + if (dci->pucch_resource_indicator < pucch_Config->resourceSetToAddModList->list.array[id]->resourceList.list.count) { valid = 1; break; } @@ -1311,7 +1062,7 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr mac->scc? mac->scc->dmrs_TypeA_Position:mac->mib->dmrs_TypeA_Position, dlsch_config_pdu_1_1->number_symbols, dlsch_config_pdu_1_1->start_symbol, - mappingtype, + tda_info.mapping_type, dlsch_config_pdu_1_1->n_front_load_symb); /* TCI */ @@ -1332,8 +1083,7 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr /* PDSCH_TO_HARQ_FEEDBACK_TIME_IND */ // according to TS 38.213 Table 9.2.3-1 - uint8_t feedback_ti = - ubwpd->pucch_Config->choice.setup->dl_DataToUL_ACK->list.array[dci->pdsch_to_harq_feedback_timing_indicator.val][0]; + uint8_t feedback_ti = pucch_Config->dl_DataToUL_ACK->list.array[dci->pdsch_to_harq_feedback_timing_indicator.val][0]; AssertFatal(feedback_ti>=DURATION_RX_TO_TX,"PDSCH to HARQ feedback time (%d) cannot be less than DURATION_RX_TO_TX (%d). Min feedback time set in config file (min_rxtxtime).\n", feedback_ti,DURATION_RX_TO_TX); @@ -1374,7 +1124,7 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr nb_rb_oh, 0, Nl); // TBS_LBRM according to section 5.4.2.1 of 38.212 - long *maxMIMO_Layers = mac->cg->spCellConfig->spCellConfigDedicated->pdsch_ServingCellConfig->choice.setup->ext1->maxMIMO_Layers; + long *maxMIMO_Layers = current_DL_BWP->pdsch_servingcellconfig->ext1->maxMIMO_Layers; AssertFatal (maxMIMO_Layers != NULL,"Option with max MIMO layers not configured is not supported\n"); int nl_tbslbrm = *maxMIMO_Layers < 4 ? *maxMIMO_Layers : 4; int bw_tbslbrm = get_dlbw_tbslbrm(current_DL_BWP->initial_BWPSize, mac->cg); @@ -1414,25 +1164,7 @@ int8_t nr_ue_process_dci(module_id_t module_id, int cc_id, uint8_t gNB_index, fr break; } - - if(rnti == SI_RNTI){ - - // }else if(rnti == mac->ra_rnti){ - - }else if(rnti == P_RNTI){ - - }else{ // c-rnti - - /// check if this is pdcch order - //dci->random_access_preamble_index; - //dci->ss_pbch_index; - //dci->prach_mask_index; - - /// else normal DL-SCH grant - } - return ret; - } int8_t nr_ue_process_csirs_measurements(module_id_t module_id, @@ -1496,7 +1228,7 @@ void nr_ue_configure_pucch(NR_UE_MAC_INST_t *mac, int O_CRC = 0; //FIXME uint16_t O_uci = O_CSI + O_ACK; NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; - NR_BWP_Id_t bwp_id = current_UL_BWP->bwp_id; + NR_UE_DL_BWP_t *current_DL_BWP = &mac->current_DL_BWP; NR_PUCCH_FormatConfig_t *pucchfmt; long *pusch_id = NULL; long *id0 = NULL; @@ -1512,10 +1244,10 @@ void nr_ue_configure_pucch(NR_UE_MAC_INST_t *mac, // only for ack/nack if (pucch->initial_pucch_id > -1 && pucch->pucch_resource == NULL) { - - pucch_pdu->format_type = initial_pucch_resource[pucch->initial_pucch_id].format; - pucch_pdu->start_symbol_index = initial_pucch_resource[pucch->initial_pucch_id].startingSymbolIndex; - pucch_pdu->nr_of_symbols = initial_pucch_resource[pucch->initial_pucch_id].nrofSymbols; + int pucch_resourcecommon = *current_UL_BWP->pucch_ConfigCommon->pucch_ResourceCommon; + pucch_pdu->format_type = initial_pucch_resource[pucch_resourcecommon].format; + pucch_pdu->start_symbol_index = initial_pucch_resource[pucch_resourcecommon].startingSymbolIndex; + pucch_pdu->nr_of_symbols = initial_pucch_resource[pucch_resourcecommon].nrofSymbols; pucch_pdu->bwp_size = current_UL_BWP->BWPSize; pucch_pdu->bwp_start = current_UL_BWP->BWPStart; @@ -1523,19 +1255,20 @@ void nr_ue_configure_pucch(NR_UE_MAC_INST_t *mac, pucch_pdu->prb_size = 1; // format 0 or 1 int RB_BWP_offset; if (pucch->initial_pucch_id == 15) - RB_BWP_offset = pucch_pdu->bwp_size>>2; + RB_BWP_offset = pucch_pdu->bwp_size >> 2; else - RB_BWP_offset = initial_pucch_resource[pucch->initial_pucch_id].PRB_offset; + RB_BWP_offset = initial_pucch_resource[pucch_resourcecommon].PRB_offset; - int N_CS = initial_pucch_resource[pucch->initial_pucch_id].nb_CS_indexes; - pucch_pdu->prb_start = RB_BWP_offset + (pucch->initial_pucch_id/N_CS); - if (pucch->initial_pucch_id>>3 == 0) { - pucch_pdu->second_hop_prb = pucch_pdu->bwp_size - 1 - RB_BWP_offset - (pucch->initial_pucch_id/N_CS); - pucch_pdu->initial_cyclic_shift = initial_pucch_resource[pucch->initial_pucch_id].initial_CS_indexes[pucch->initial_pucch_id%N_CS]; - } - else { - pucch_pdu->second_hop_prb = pucch_pdu->bwp_size - 1 - RB_BWP_offset - ((pucch->initial_pucch_id - 8)/N_CS); - pucch_pdu->initial_cyclic_shift = initial_pucch_resource[pucch->initial_pucch_id].initial_CS_indexes[(pucch->initial_pucch_id - 8)%N_CS]; + int N_CS = initial_pucch_resource[pucch_resourcecommon].nb_CS_indexes; + + if (pucch->initial_pucch_id >> 3 == 0) { + pucch_pdu->prb_start = RB_BWP_offset + (pucch->initial_pucch_id / N_CS); + pucch_pdu->second_hop_prb = pucch_pdu->bwp_size - 1 - RB_BWP_offset - (pucch->initial_pucch_id / N_CS); + pucch_pdu->initial_cyclic_shift = initial_pucch_resource[pucch_resourcecommon].initial_CS_indexes[pucch->initial_pucch_id % N_CS]; + } else { + pucch_pdu->prb_start = pucch_pdu->bwp_size - 1 - RB_BWP_offset - ((pucch->initial_pucch_id - 8) / N_CS); + pucch_pdu->second_hop_prb = RB_BWP_offset + ((pucch->initial_pucch_id - 8) / N_CS); + pucch_pdu->initial_cyclic_shift = initial_pucch_resource[pucch_resourcecommon].initial_CS_indexes[(pucch->initial_pucch_id - 8) % N_CS]; } pucch_pdu->freq_hop_flag = 1; pucch_pdu->time_domain_occ_idx = 0; @@ -1551,48 +1284,25 @@ void nr_ue_configure_pucch(NR_UE_MAC_INST_t *mac, NR_PUCCH_Resource_t *pucchres = pucch->pucch_resource; - if (mac->cg && - mac->cg->physicalCellGroupConfig && - (mac->cg->physicalCellGroupConfig->harq_ACK_SpatialBundlingPUCCH != NULL || - mac->cg->physicalCellGroupConfig->pdsch_HARQ_ACK_Codebook != 1)) { + if (current_UL_BWP->harq_ACK_SpatialBundlingPUCCH != NULL || *current_DL_BWP->pdsch_HARQ_ACK_Codebook != 1) { LOG_E(MAC,"PUCCH Unsupported cell group configuration\n"); return; - } - else if (mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->pdsch_ServingCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated->pdsch_ServingCellConfig->choice.setup && - mac->cg->spCellConfig->spCellConfigDedicated->pdsch_ServingCellConfig->choice.setup->codeBlockGroupTransmission != NULL) { + } else if (current_DL_BWP && current_DL_BWP->pdsch_servingcellconfig && current_DL_BWP->pdsch_servingcellconfig->codeBlockGroupTransmission != NULL) { LOG_E(MAC,"PUCCH Unsupported code block group for serving cell config\n"); return; } - NR_PUCCH_Config_t *pucch_Config; - if (bwp_id>0 && - mac->ULbwp[bwp_id-1] && - mac->ULbwp[bwp_id-1]->bwp_Dedicated && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup) { - NR_PUSCH_Config_t *pusch_Config = mac->ULbwp[bwp_id-1]->bwp_Dedicated->pusch_Config->choice.setup; + + NR_PUSCH_Config_t *pusch_Config = current_UL_BWP ? current_UL_BWP->pusch_Config : NULL; + if (pusch_Config) { pusch_id = pusch_Config->dataScramblingIdentityPUSCH; if (pusch_Config->dmrs_UplinkForPUSCH_MappingTypeA != NULL) id0 = pusch_Config->dmrs_UplinkForPUSCH_MappingTypeA->choice.setup->transformPrecodingDisabled->scramblingID0; else if (pusch_Config->dmrs_UplinkForPUSCH_MappingTypeB != NULL) id0 = pusch_Config->dmrs_UplinkForPUSCH_MappingTypeB->choice.setup->transformPrecodingDisabled->scramblingID0; - else *id0 = mac->physCellId; - pucch_Config = mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup; } - else if (bwp_id==0 && - mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup) { - pucch_Config = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup; - } - else AssertFatal(1==0,"no pucch_Config\n"); + + NR_PUCCH_Config_t *pucch_Config = current_UL_BWP->pucch_Config; + AssertFatal(pucch_Config, "no pucch_Config\n"); pucch_pdu->bwp_size = current_UL_BWP->BWPSize; pucch_pdu->bwp_start = current_UL_BWP->BWPStart; @@ -1642,8 +1352,8 @@ void nr_ue_configure_pucch(NR_UE_MAC_INST_t *mac, pucch_pdu->n_bit = O_uci+O_SR; pucch_pdu->nr_of_symbols = pucchres->format.choice.format2->nrofSymbols; pucch_pdu->start_symbol_index = pucchres->format.choice.format2->startingSymbolIndex; - pucch_pdu->data_scrambling_id = pusch_id!= NULL ? *pusch_id : mac->physCellId; - pucch_pdu->dmrs_scrambling_id = id0!= NULL ? *id0 : mac->physCellId; + pucch_pdu->data_scrambling_id = pusch_id != NULL ? *pusch_id : mac->physCellId; + pucch_pdu->dmrs_scrambling_id = id0 != NULL ? *id0 : mac->physCellId; pucch_pdu->prb_size = compute_pucch_prb_size(2,pucchres->format.choice.format2->nrofPRBs, O_uci+O_SR,O_CSI,pucch_Config->format2->choice.setup->maxCodeRate, 2,pucchres->format.choice.format2->nrofSymbols,8); @@ -1653,7 +1363,7 @@ void nr_ue_configure_pucch(NR_UE_MAC_INST_t *mac, pucch_pdu->n_bit = O_uci+O_SR; pucch_pdu->nr_of_symbols = pucchres->format.choice.format3->nrofSymbols; pucch_pdu->start_symbol_index = pucchres->format.choice.format3->startingSymbolIndex; - pucch_pdu->data_scrambling_id = pusch_id!= NULL ? *pusch_id : mac->physCellId; + pucch_pdu->data_scrambling_id = pusch_id != NULL ? *pusch_id : mac->physCellId; if (pucch_Config->format3 == NULL) { pucch_pdu->pi_2bpsk = 0; pucch_pdu->add_dmrs_flag = 0; @@ -1712,14 +1422,8 @@ void nr_ue_configure_pucch(NR_UE_MAC_INST_t *mac, } else AssertFatal(1==0,"problem with pucch configuration\n"); - NR_PUCCH_ConfigCommon_t *pucch_ConfigCommon; - if (bwp_id>0 && - mac->ULbwp[bwp_id-1] && - mac->ULbwp[bwp_id-1]->bwp_Common && - mac->ULbwp[bwp_id-1]->bwp_Common->pucch_ConfigCommon) - pucch_ConfigCommon = mac->ULbwp[bwp_id-1]->bwp_Common->pucch_ConfigCommon->choice.setup; - else if (mac->scc) pucch_ConfigCommon = mac->scc->uplinkConfigCommon->initialUplinkBWP->pucch_ConfigCommon->choice.setup; - else pucch_ConfigCommon = mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.pucch_ConfigCommon->choice.setup; + NR_PUCCH_ConfigCommon_t *pucch_ConfigCommon = current_UL_BWP->pucch_ConfigCommon; + if (pucch_ConfigCommon->hoppingId != NULL) pucch_pdu->hopping_id = *pucch_ConfigCommon->hoppingId; else @@ -1759,11 +1463,9 @@ int16_t get_pucch_tx_power_ue(NR_UE_MAC_INST_t *mac, int subframe_number, int O_ACK, int O_SR, int O_CSI, int O_CRC) { - + NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; int PUCCH_POWER_DEFAULT = 0; - int16_t P_O_NOMINAL_PUCCH; - if (mac->scc) P_O_NOMINAL_PUCCH = *mac->scc->uplinkConfigCommon->initialUplinkBWP->pucch_ConfigCommon->choice.setup->p0_nominal; - else P_O_NOMINAL_PUCCH = *mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.pucch_ConfigCommon->choice.setup->p0_nominal; + int16_t P_O_NOMINAL_PUCCH = *current_UL_BWP->pucch_ConfigCommon->p0_nominal; struct NR_PUCCH_PowerControl *power_config = pucch_Config->pucch_PowerControl; @@ -1907,7 +1609,7 @@ int get_deltatf(uint16_t nb_of_prbs, int find_pucch_resource_set(NR_UE_MAC_INST_t *mac, int uci_size) { int pucch_resource_set_id = 0; - NR_BWP_Id_t bwp_id = mac->current_DL_BWP.bwp_id; + NR_PUCCH_Config_t *pucch_Config = mac->current_UL_BWP.pucch_Config; //long *pucch_max_pl_bits = NULL; @@ -1920,23 +1622,7 @@ int find_pucch_resource_set(NR_UE_MAC_INST_t *mac, int uci_size) { */ /* look for the first resource set which supports uci_size number of bits for payload */ while (pucch_resource_set_id < MAX_NB_OF_PUCCH_RESOURCE_SETS) { - if ((bwp_id>0 && - mac->ULbwp[bwp_id-1] && - mac->ULbwp[bwp_id-1]->bwp_Dedicated && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup->resourceSetToAddModList && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup->resourceSetToAddModList->list.array[pucch_resource_set_id] != NULL) || - (bwp_id==0 && - mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup->resourceSetToAddModList && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup->resourceSetToAddModList->list.array[pucch_resource_set_id] != NULL)) { + if (pucch_Config && pucch_Config->resourceSetToAddModList && pucch_Config->resourceSetToAddModList->list.array[pucch_resource_set_id] != NULL) { // PUCCH with format0 can be up to 3 bits (2 ack/nacks + 1 sr is 3 max bits) if (uci_size <= 3) { pucch_resource_set_id = 0; @@ -1975,34 +1661,12 @@ int find_pucch_resource_set(NR_UE_MAC_INST_t *mac, int uci_size) { * *********************************************************************/ -void select_pucch_resource(NR_UE_MAC_INST_t *mac, - PUCCH_sched_t *pucch) { - +void select_pucch_resource(NR_UE_MAC_INST_t *mac, PUCCH_sched_t *pucch) +{ NR_PUCCH_ResourceId_t *current_resource_id = NULL; - NR_BWP_Id_t bwp_id = mac->current_UL_BWP.bwp_id; + NR_PUCCH_Config_t *pucch_Config = mac->current_UL_BWP.pucch_Config; int n_list; - - if (pucch->is_common == 1 || - (bwp_id == 0 && - mac->cg == NULL) || - (bwp_id == 0 && - mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup->resourceSetToAddModList && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup->resourceSetToAddModList->list.array[0] == NULL) || - (mac->ULbwp[bwp_id-1] && - mac->ULbwp[bwp_id-1]->bwp_Dedicated && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup->resourceSetToAddModList && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup->resourceSetToAddModList->list.array[0] == NULL) - ){ - + if (pucch->is_common == 1 || !pucch_Config || !pucch_Config->resourceSetToAddModList || pucch_Config->resourceSetToAddModList->list.array[0] == NULL) { /* see TS 38.213 9.2.1 PUCCH Resource Sets */ int delta_PRI = pucch->resource_indicator; int n_CCE_0 = pucch->n_CCE; @@ -2013,21 +1677,9 @@ void select_pucch_resource(NR_UE_MAC_INST_t *mac, int r_PUCCH = ((2 * n_CCE_0)/N_CCE_0) + (2 * delta_PRI); pucch->initial_pucch_id = r_PUCCH; pucch->pucch_resource = NULL; - } - else { - struct NR_PUCCH_Config__resourceSetToAddModList *resourceSetToAddModList = NULL; - struct NR_PUCCH_Config__resourceToAddModList *resourceToAddModList = NULL; - if (bwp_id > 0 && mac->ULbwp[bwp_id-1]) { - AssertFatal(mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup->resourceSetToAddModList!=NULL, - "mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup->resourceSetToAddModList is null\n"); - resourceSetToAddModList = mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup->resourceSetToAddModList; - resourceToAddModList = mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup->resourceToAddModList; - } - else if (bwp_id == 0 && mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup->resourceSetToAddModList!=NULL) { - resourceSetToAddModList = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup->resourceSetToAddModList; - resourceToAddModList = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup->resourceToAddModList; - } - + } else { + struct NR_PUCCH_Config__resourceSetToAddModList *resourceSetToAddModList = pucch_Config->resourceSetToAddModList; + struct NR_PUCCH_Config__resourceToAddModList *resourceToAddModList = pucch_Config->resourceToAddModList; n_list = resourceSetToAddModList->list.count; if (pucch->resource_set_id > n_list) { LOG_E(MAC,"Invalid PUCCH resource set id %d\n",pucch->resource_set_id); @@ -2107,19 +1759,8 @@ uint8_t get_downlink_ack(NR_UE_MAC_INST_t *mac, NR_UE_DL_BWP_t *current_DL_BWP = &mac->current_DL_BWP; NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; - NR_BWP_Id_t dl_bwp_id = current_DL_BWP->bwp_id; - NR_BWP_Id_t ul_bwp_id = current_UL_BWP->bwp_id; - NR_BWP_DownlinkDedicated_t *bwpd=NULL; - NR_BWP_DownlinkCommon_t *bwpc=NULL; - NR_BWP_UplinkDedicated_t *ubwpd=NULL; - NR_BWP_UplinkCommon_t *ubwpc=NULL; - get_bwp_info(mac,dl_bwp_id,ul_bwp_id,&bwpd,&bwpc,&ubwpd,&ubwpc); - if (bwpd && - bwpd->pdsch_Config && - bwpd->pdsch_Config->choice.setup && - bwpd->pdsch_Config->choice.setup->maxNrofCodeWordsScheduledByDCI && - bwpd->pdsch_Config->choice.setup->maxNrofCodeWordsScheduledByDCI[0] == 2) { + if (current_DL_BWP && current_DL_BWP->pdsch_Config && current_DL_BWP->pdsch_Config->maxNrofCodeWordsScheduledByDCI && current_DL_BWP->pdsch_Config->maxNrofCodeWordsScheduledByDCI[0] == 2) { two_transport_blocks = true; number_of_code_word = 2; } @@ -2144,8 +1785,7 @@ uint8_t get_downlink_ack(NR_UE_MAC_INST_t *mac, sched_frame = (sched_frame + 1) % 1024; } AssertFatal(sched_slot < slots_per_frame, "sched_slot was calculated incorrect %d\n", sched_slot); - LOG_D(PHY,"HARQ pid %d is active for %d.%d (dl_slot %d, feedback_to_ul %d, is_common %d\n", - dl_harq_pid, sched_frame,sched_slot,current_harq->dl_slot,current_harq->feedback_to_ul,current_harq->is_common); + LOG_D(PHY, "HARQ pid %d is active for %d.%d (dl_slot %d, feedback_to_ul %d\n", dl_harq_pid, sched_frame, sched_slot, current_harq->dl_slot, current_harq->feedback_to_ul); /* check if current tx slot should transmit downlink acknowlegment */ if (sched_frame == frame && sched_slot == slot) { if (get_softmodem_params()->emulate_l1) { @@ -2209,10 +1849,8 @@ uint8_t get_downlink_ack(NR_UE_MAC_INST_t *mac, U_DAI_c = number_harq_feedback/number_of_code_word; N_m_c_rx = number_harq_feedback; int N_SPS_c = 0; /* FFS TODO_NR multicells and SPS are not supported at the moment */ - if (mac->cg != NULL && - mac->cg->physicalCellGroupConfig != NULL && - mac->cg->physicalCellGroupConfig->harq_ACK_SpatialBundlingPUCCH != NULL) { - int N_TB_max_DL = bwpd->pdsch_Config->choice.setup->maxNrofCodeWordsScheduledByDCI[0]; + if (current_UL_BWP->harq_ACK_SpatialBundlingPUCCH != NULL) { + int N_TB_max_DL = current_DL_BWP->pdsch_Config->maxNrofCodeWordsScheduledByDCI[0]; pucch->n_HARQ_ACK = (((V_DAI_m_DL - U_DAI_c)%4) * N_TB_max_DL) + N_m_c_rx + N_SPS_c; LOG_D(MAC, "PUCCH power n(%d) = ( V(%d) - U(%d) )mod4 * N_TB(%d) + N(%d) \n", pucch->n_HARQ_ACK, V_DAI_m_DL, U_DAI_c, N_TB_max_DL, N_m_c_rx); } @@ -2305,28 +1943,10 @@ bool trigger_periodic_scheduling_request(NR_UE_MAC_INST_t *mac, int slot) { NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; - NR_BWP_Id_t bwp_id = current_UL_BWP->bwp_id; - NR_PUCCH_Config_t *pucch_Config = NULL; + NR_PUCCH_Config_t *pucch_Config = current_UL_BWP->pucch_Config; const int n_slots_frame = nr_slots_per_frame[current_UL_BWP->scs]; - if (bwp_id>0 && - mac->ULbwp[bwp_id-1] && - mac->ULbwp[bwp_id-1]->bwp_Dedicated && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup) { - pucch_Config = mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup; - } - else if (bwp_id==0 && - mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup) { - pucch_Config = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup; - } if(!pucch_Config || !pucch_Config->schedulingRequestResourceToAddModList || pucch_Config->schedulingRequestResourceToAddModList->list.count==0) @@ -2419,15 +2039,11 @@ uint8_t nr_get_csi_measurements(NR_UE_MAC_INST_t *mac, NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; NR_BWP_Id_t bwp_id = current_UL_BWP->bwp_id; - NR_PUCCH_Config_t *pucch_Config = NULL; + NR_PUCCH_Config_t *pucch_Config = current_UL_BWP->pucch_Config; int csi_bits = 0; - if(mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->csi_MeasConfig) { - - NR_CSI_MeasConfig_t *csi_measconfig = mac->cg->spCellConfig->spCellConfigDedicated->csi_MeasConfig->choice.setup; + if (current_UL_BWP->csi_MeasConfig) { + NR_CSI_MeasConfig_t *csi_measconfig = current_UL_BWP->csi_MeasConfig; for (int csi_report_id = 0; csi_report_id < csi_measconfig->csi_ReportConfigToAddModList->list.count; csi_report_id++){ NR_CSI_ReportConfig_t *csirep = csi_measconfig->csi_ReportConfigToAddModList->list.array[csi_report_id]; @@ -2441,21 +2057,6 @@ uint8_t nr_get_csi_measurements(NR_UE_MAC_INST_t *mac, int period, offset; csi_period_offset(csirep, NULL, &period, &offset); - if (bwp_id>0 && - mac->ULbwp[bwp_id-1] && - mac->ULbwp[bwp_id-1]->bwp_Dedicated && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup) { - pucch_Config = mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup; - } - else if (bwp_id==0 && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup) { - pucch_Config = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup; - } - const int n_slots_frame = nr_slots_per_frame[current_UL_BWP->scs]; if (((n_slots_frame*frame + slot - offset)%period) == 0 && pucch_Config) { LOG_D(NR_MAC, "Preparing CSI report in frame %d slot %d CSI report ID %d\n", frame, slot, csi_report_id); @@ -4074,7 +3675,13 @@ int nr_ue_process_rar(nr_downlink_indication_t *dl_info, NR_UL_TIME_ALIGNMENT_t #endif // Schedule Msg3 - ret = nr_ue_pusch_scheduler(mac, is_Msg3, frame, slot, &frame_tx, &slot_tx, rar_grant.Msg3_t_alloc); + NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; + NR_tda_info_t tda_info = get_ul_tda_info(current_UL_BWP, *ra->ss->controlResourceSetId, ra->ss->searchSpaceType->present, NR_RNTI_RA, rar_grant.Msg3_t_alloc); + if (tda_info.nrOfSymbols == 0) { + LOG_E(MAC, "Cannot schedule Msg3. Something wrong in TDA information\n"); + return -1; + } + ret = nr_ue_pusch_scheduler(mac, is_Msg3, frame, slot, &frame_tx, &slot_tx, tda_info.k2); if (ret != -1){ @@ -4102,7 +3709,7 @@ int nr_ue_process_rar(nr_downlink_indication_t *dl_info, NR_UL_TIME_ALIGNMENT_t pthread_mutex_unlock(&ul_config->mutex_ul_config); // Config Msg3 PDU - nr_config_pusch_pdu(mac, pusch_config_pdu, NULL, &rar_grant, rnti, NULL); + nr_config_pusch_pdu(mac, &tda_info, pusch_config_pdu, NULL, &rar_grant, rnti, NULL); } } else { diff --git a/openair2/LAYER2/NR_MAC_UE/nr_ue_scheduler.c b/openair2/LAYER2/NR_MAC_UE/nr_ue_scheduler.c index 813292f462..a6acf1fa77 100644 --- a/openair2/LAYER2/NR_MAC_UE/nr_ue_scheduler.c +++ b/openair2/LAYER2/NR_MAC_UE/nr_ue_scheduler.c @@ -100,49 +100,6 @@ void fill_scheduled_response(nr_scheduled_response_t *scheduled_response, } -/* - * This function returns the slot offset K2 corresponding to a given time domain - * indication value from RRC configuration. - */ -long get_k2(NR_UE_MAC_INST_t *mac, uint8_t time_domain_ind) { - long k2 = -1; - - NR_BWP_Id_t ul_bwp_id = mac->current_UL_BWP.bwp_id; - // Get K2 from RRC configuration - NR_PUSCH_Config_t *pusch_config= ul_bwp_id > 0 && mac->ULbwp[ul_bwp_id-1] ? mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pusch_Config->choice.setup : NULL; - NR_PUSCH_TimeDomainResourceAllocationList_t *pusch_TimeDomainAllocationList = NULL; - if (pusch_config && pusch_config->pusch_TimeDomainAllocationList) { - pusch_TimeDomainAllocationList = pusch_config->pusch_TimeDomainAllocationList->choice.setup; - } - else if (ul_bwp_id > 0 && - mac->ULbwp[ul_bwp_id-1] && - mac->ULbwp[ul_bwp_id-1]->bwp_Common&& - mac->ULbwp[ul_bwp_id-1]->bwp_Common->pusch_ConfigCommon&& - mac->ULbwp[ul_bwp_id-1]->bwp_Common->pusch_ConfigCommon->choice.setup && - mac->ULbwp[ul_bwp_id-1]->bwp_Common->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList) { - pusch_TimeDomainAllocationList = mac->ULbwp[ul_bwp_id-1]->bwp_Common->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList; - } - else if (mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList) - pusch_TimeDomainAllocationList=mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList; - else AssertFatal(1==0,"need to fall back to default PUSCH time-domain allocations\n"); - - if (pusch_TimeDomainAllocationList) { - if (time_domain_ind >= pusch_TimeDomainAllocationList->list.count) { - LOG_E(NR_MAC, "time_domain_ind %d >= pusch->TimeDomainAllocationList->list.count %d\n", - time_domain_ind, pusch_TimeDomainAllocationList->list.count); - return -1; - } - k2 = *pusch_TimeDomainAllocationList->list.array[time_domain_ind]->k2; - } - - AssertFatal(k2 >= DURATION_RX_TO_TX, - "Slot offset K2 (%ld) cannot be less than DURATION_RX_TO_TX (%d). K2 set according to min_rxtxtime in config file.\n", - k2,DURATION_RX_TO_TX); - - LOG_D(NR_MAC, "get_k2(): k2 is %ld\n", k2); - return k2; -} - /* * This function returns the UL config corresponding to a given UL slot * from MAC instance . @@ -179,37 +136,13 @@ fapi_nr_ul_config_request_t *get_ul_config_request(NR_UE_MAC_INST_t *mac, int sl return &mac->ul_config_request[index]; } -void ul_layers_config(NR_UE_MAC_INST_t * mac, nfapi_nr_ue_pusch_pdu_t *pusch_config_pdu, dci_pdu_rel15_t *dci) { +void ul_layers_config(NR_UE_MAC_INST_t *mac, nfapi_nr_ue_pusch_pdu_t *pusch_config_pdu, dci_pdu_rel15_t *dci, nr_dci_format_t dci_format) +{ + NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; + NR_SRS_Config_t *srs_config = current_UL_BWP->srs_Config; + NR_PUSCH_Config_t *pusch_Config = current_UL_BWP->pusch_Config; - NR_ServingCellConfigCommon_t *scc = mac->scc; - NR_BWP_Id_t ul_bwp_id = mac->current_UL_BWP.bwp_id; - NR_BWP_UplinkDedicated_t *ubwpd=NULL; - NR_SRS_Config_t *srs_config = NULL; - - if (mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP) { - ubwpd = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP; - srs_config = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->srs_Config->choice.setup; - } - - NR_PUSCH_Config_t *pusch_Config = ul_bwp_id > 0 && mac->ULbwp[ul_bwp_id-1] ? - mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pusch_Config->choice.setup : - (ubwpd? - ubwpd->pusch_Config->choice.setup: - NULL); - - long transformPrecoder; - if (pusch_Config && pusch_Config->transformPrecoder) - transformPrecoder = *pusch_Config->transformPrecoder; - else { - if(scc && scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup->msg3_transformPrecoder) - transformPrecoder = NR_PUSCH_Config__transformPrecoder_enabled; - else - transformPrecoder = NR_PUSCH_Config__transformPrecoder_disabled; - } + long transformPrecoder = get_transformPrecoding(current_UL_BWP, dci_format, 0); pusch_config_pdu->transform_precoding = transformPrecoder; /* PRECOD_NBR_LAYERS */ @@ -336,33 +269,14 @@ void ul_layers_config(NR_UE_MAC_INST_t * mac, nfapi_nr_ue_pusch_pdu_t *pusch_con } // todo: this function shall be reviewed completely because of the many comments left by the author -void ul_ports_config(NR_UE_MAC_INST_t *mac, int *n_front_load_symb, nfapi_nr_ue_pusch_pdu_t *pusch_config_pdu, dci_pdu_rel15_t *dci) { - +void ul_ports_config(NR_UE_MAC_INST_t *mac, int *n_front_load_symb, nfapi_nr_ue_pusch_pdu_t *pusch_config_pdu, dci_pdu_rel15_t *dci, nr_dci_format_t dci_format) +{ uint8_t rank = pusch_config_pdu->nrOfLayers; - NR_BWP_Id_t ul_bwp_id = mac->current_UL_BWP.bwp_id; - NR_ServingCellConfigCommon_t *scc = mac->scc; - NR_BWP_UplinkDedicated_t *ubwpd=NULL; - - if (mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP) - ubwpd = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP; - - NR_PUSCH_Config_t *pusch_Config = ul_bwp_id > 0 && mac->ULbwp[ul_bwp_id-1] ? mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pusch_Config->choice.setup : (ubwpd?ubwpd->pusch_Config->choice.setup:NULL); + NR_PUSCH_Config_t *pusch_Config = mac->current_UL_BWP.pusch_Config; AssertFatal(pusch_Config!=NULL,"pusch_Config shouldn't be null\n"); - long transformPrecoder; - if (pusch_Config->transformPrecoder) - transformPrecoder = *pusch_Config->transformPrecoder; - else { - if(scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup->msg3_transformPrecoder) - transformPrecoder = NR_PUSCH_Config__transformPrecoder_enabled; - else - transformPrecoder = NR_PUSCH_Config__transformPrecoder_disabled; - } + long transformPrecoder = get_transformPrecoding(&mac->current_UL_BWP, dci_format, 0); long *max_length = NULL; long *dmrs_type = NULL; LOG_D(NR_MAC,"transformPrecoder %s\n",transformPrecoder==NR_PUSCH_Config__transformPrecoder_disabled?"disabled":"enabled"); @@ -535,18 +449,11 @@ void ul_ports_config(NR_UE_MAC_INST_t *mac, int *n_front_load_symb, nfapi_nr_ue_ // - 6.1.4.2 of TS 38.214 // - 6.4.1.1.1 of TS 38.211 // - 6.3.1.7 of 38.211 -int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, - nfapi_nr_ue_pusch_pdu_t *pusch_config_pdu, - dci_pdu_rel15_t *dci, - RAR_grant_t *rar_grant, - uint16_t rnti, - uint8_t *dci_format) +int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, NR_tda_info_t *tda_info, nfapi_nr_ue_pusch_pdu_t *pusch_config_pdu, dci_pdu_rel15_t *dci, RAR_grant_t *rar_grant, uint16_t rnti, uint8_t *dci_format) { int f_alloc; int mask; - int StartSymbolIndex; - int NrOfSymbols; uint8_t nb_dmrs_re_per_rb; uint16_t l_prime_mask = 0; @@ -555,7 +462,6 @@ int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, int rnti_type = get_rnti_type(mac, rnti); NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; - NR_BWP_Id_t ul_bwp_id = current_UL_BWP->bwp_id; // Common configuration pusch_config_pdu->dmrs_config_type = pusch_dmrs_type1; @@ -563,36 +469,14 @@ int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, pusch_config_pdu->nrOfLayers = 1; pusch_config_pdu->Tpmi = 0; pusch_config_pdu->rnti = rnti; - NR_BWP_UplinkCommon_t *initialUplinkBWP; - if (mac->scc) initialUplinkBWP = mac->scc->uplinkConfigCommon->initialUplinkBWP; - else initialUplinkBWP = &mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP; pusch_dmrs_AdditionalPosition_t add_pos = pusch_dmrs_pos2; int dmrslength = 1; + NR_PUSCH_Config_t *pusch_Config = current_UL_BWP->pusch_Config; if (rar_grant) { // Note: for Msg3 or MsgA PUSCH transmission the N_PRB_oh is always set to 0 - NR_BWP_Uplink_t *ubwp = ul_bwp_id > 0 ? mac->ULbwp[ul_bwp_id - 1] : NULL; - NR_BWP_UplinkDedicated_t *ibwp; - int startSymbolAndLength,mappingtype; - NR_PUSCH_Config_t *pusch_Config=NULL; - if (mac->cg && ubwp && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP) { - - ibwp = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP; - pusch_Config = ibwp->pusch_Config->choice.setup; - startSymbolAndLength = ubwp->bwp_Common->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList->list.array[rar_grant->Msg3_t_alloc]->startSymbolAndLength; - mappingtype = ubwp->bwp_Common->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList->list.array[rar_grant->Msg3_t_alloc]->mappingType; - - } - else { - startSymbolAndLength = initialUplinkBWP->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList->list.array[rar_grant->Msg3_t_alloc]->startSymbolAndLength; - mappingtype = initialUplinkBWP->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList->list.array[rar_grant->Msg3_t_alloc]->mappingType; - } int ibwp_start = current_UL_BWP->initial_BWPStart; int ibwp_size = current_UL_BWP->initial_BWPSize; int abwp_start = current_UL_BWP->BWPStart; @@ -623,14 +507,14 @@ int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, //pusch_config_pdu->rb_start += ibwp_start - abwp_start; // Time domain allocation - SLIV2SL(startSymbolAndLength, &StartSymbolIndex, &NrOfSymbols); - pusch_config_pdu->start_symbol_index = StartSymbolIndex; - pusch_config_pdu->nr_of_symbols = NrOfSymbols; + pusch_config_pdu->start_symbol_index = tda_info->startSymbolIndex; + pusch_config_pdu->nr_of_symbols = tda_info->nrOfSymbols; - l_prime_mask = get_l_prime(NrOfSymbols, mappingtype, add_pos, dmrslength, StartSymbolIndex, mac->scc ? mac->scc->dmrs_TypeA_Position : mac->mib->dmrs_TypeA_Position); - LOG_D(NR_MAC, "MSG3 start_sym:%d NR Symb:%d mappingtype:%d, DMRS_MASK:%x\n", pusch_config_pdu->start_symbol_index, pusch_config_pdu->nr_of_symbols, mappingtype, l_prime_mask); + l_prime_mask = + get_l_prime(tda_info->nrOfSymbols, tda_info->mapping_type, add_pos, dmrslength, tda_info->startSymbolIndex, mac->scc ? mac->scc->dmrs_TypeA_Position : mac->mib->dmrs_TypeA_Position); + LOG_D(NR_MAC, "MSG3 start_sym:%d NR Symb:%d mappingtype:%d, DMRS_MASK:%x\n", pusch_config_pdu->start_symbol_index, pusch_config_pdu->nr_of_symbols, tda_info->mapping_type, l_prime_mask); - #ifdef DEBUG_MSG3 +#ifdef DEBUG_MSG3 LOG_D(NR_MAC, "In %s BWP assignment (BWP (start %d, size %d) \n", __FUNCTION__, pusch_config_pdu->bwp_start, pusch_config_pdu->bwp_size); #endif @@ -648,10 +532,10 @@ int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, pusch_config_pdu->scid = 0; // Transform precoding according to 6.1.3 UE procedure for applying transform precoding on PUSCH in 38.214 - pusch_config_pdu->transform_precoding = get_transformPrecoding(initialUplinkBWP, pusch_Config, NULL, NULL, NR_RNTI_TC, 0); // TBR fix rnti and take out + pusch_config_pdu->transform_precoding = get_transformPrecoding(current_UL_BWP, NR_UL_DCI_FORMAT_0_0, 0); // as if it was DCI 0_0 // Resource allocation in frequency domain according to 6.1.2.2 in TS 38.214 - pusch_config_pdu->resource_alloc = (mac->cg) ? pusch_Config->resourceAllocation : 1; + pusch_config_pdu->resource_alloc = 1; //// Completing PUSCH PDU pusch_config_pdu->mcs_table = 0; @@ -669,13 +553,7 @@ int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, pusch_config_pdu->tbslbrm = 0; } else if (dci) { - NR_BWP_Id_t dl_bwp_id = mac->current_DL_BWP.bwp_id; NR_BWP_Id_t ul_bwp_id = mac->current_UL_BWP.bwp_id; - NR_BWP_DownlinkDedicated_t *bwpd=NULL; - NR_BWP_DownlinkCommon_t *bwpc=NULL; - NR_BWP_UplinkDedicated_t *ubwpd=NULL; - NR_BWP_UplinkCommon_t *ubwpc=NULL; - get_bwp_info(mac,dl_bwp_id,ul_bwp_id,&bwpd,&bwpc,&ubwpd,&ubwpc); int target_ss; bool valid_ptrs_setup = 0; @@ -683,8 +561,6 @@ int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, pusch_config_pdu->bwp_start = current_UL_BWP->BWPStart; pusch_config_pdu->bwp_size = current_UL_BWP->BWPSize; - const NR_PUSCH_Config_t *pusch_Config = ubwpd? ubwpd->pusch_Config->choice.setup : NULL; - // Basic sanity check for MCS value to check for a false or erroneous DCI if (dci->mcs > 28) { LOG_W(NR_MAC, "MCS value %d out of bounds! Possibly due to false DCI. Ignoring DCI!\n", dci->mcs); @@ -693,7 +569,7 @@ int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, /* Transform precoding */ if (rnti_type != NR_RNTI_CS || (rnti_type == NR_RNTI_CS && dci->ndi == 1)) { - pusch_config_pdu->transform_precoding = get_transformPrecoding(initialUplinkBWP, pusch_Config, NULL, dci_format, rnti_type, 0); + pusch_config_pdu->transform_precoding = get_transformPrecoding(current_UL_BWP, *dci_format, 0); } /*DCI format-related configuration*/ @@ -703,11 +579,9 @@ int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, } else if (*dci_format == NR_UL_DCI_FORMAT_0_1) { - get_bwp_info(mac,dl_bwp_id,ul_bwp_id,&bwpd,&bwpc,&ubwpd,&ubwpc); - target_ss = NR_SearchSpace__searchSpaceType_PR_ue_Specific; - ul_layers_config(mac, pusch_config_pdu, dci); - ul_ports_config(mac, &dmrslength, pusch_config_pdu, dci); + ul_layers_config(mac, pusch_config_pdu, dci, *dci_format); + ul_ports_config(mac, &dmrslength, pusch_config_pdu, dci, *dci_format); } else { @@ -716,10 +590,7 @@ int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, } - NR_PUSCH_ConfigCommon_t *pusch_ConfigCommon = ubwpc->pusch_ConfigCommon->choice.setup; - NR_PUSCH_TimeDomainResourceAllocationList_t *pusch_TimeDomainAllocationList = choose_ul_tda_list(pusch_Config,pusch_ConfigCommon); - - int mappingtype = pusch_TimeDomainAllocationList->list.array[dci->time_domain_assignment.val]->mappingType; + int mappingtype = tda_info->mapping_type; NR_DMRS_UplinkConfig_t *NR_DMRS_ulconfig = NULL; if(pusch_Config) { @@ -774,10 +645,9 @@ int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, if (nr_ue_process_dci_freq_dom_resource_assignment(pusch_config_pdu, NULL, current_UL_BWP->BWPSize, 0, dci->frequency_domain_assignment.val) < 0){ return -1; } - /* TIME_DOM_RESOURCE_ASSIGNMENT */ - if (nr_ue_process_dci_time_dom_resource_assignment(mac, pusch_TimeDomainAllocationList, NULL, pusch_config_pdu, NULL, NULL, dci->time_domain_assignment.val,0,false) < 0) { - return -1; - } + + pusch_config_pdu->start_symbol_index = tda_info->startSymbolIndex; + pusch_config_pdu->nr_of_symbols = tda_info->nrOfSymbols; /* FREQ_HOPPING_FLAG */ if ((pusch_Config!=NULL) && (pusch_Config->frequencyHopping!=NULL) && (pusch_Config->resourceAllocation != NR_PUSCH_Config__resourceAllocation_resourceAllocationType0)){ @@ -837,23 +707,13 @@ int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, } // Num PRB Overhead from PUSCH-ServingCellConfig - if (mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->pusch_ServingCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->pusch_ServingCellConfig->choice.setup->xOverhead) - N_PRB_oh = *mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->pusch_ServingCellConfig->choice.setup->xOverhead; + if (current_UL_BWP->pusch_servingcellconfig && current_UL_BWP->pusch_servingcellconfig->xOverhead) + N_PRB_oh = *current_UL_BWP->pusch_servingcellconfig->xOverhead; + else + N_PRB_oh = 0; - else N_PRB_oh = 0; - - if (mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->pusch_ServingCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->pusch_ServingCellConfig->choice.setup->rateMatching) { - long *maxMIMO_Layers = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->pusch_ServingCellConfig->choice.setup->ext1->maxMIMO_Layers; + if (current_UL_BWP->pusch_servingcellconfig && current_UL_BWP->pusch_servingcellconfig->rateMatching) { + long *maxMIMO_Layers = current_UL_BWP->pusch_servingcellconfig->ext1->maxMIMO_Layers; if (!maxMIMO_Layers) maxMIMO_Layers = pusch_Config ? pusch_Config->maxRank : NULL; AssertFatal (maxMIMO_Layers != NULL,"Option with max MIMO layers not configured is not supported\n"); @@ -861,33 +721,30 @@ int nr_config_pusch_pdu(NR_UE_MAC_INST_t *mac, pusch_config_pdu->tbslbrm = nr_compute_tbslbrm(pusch_config_pdu->mcs_table, bw_tbslbrm, *maxMIMO_Layers); - } - else + } else pusch_config_pdu->tbslbrm = 0; /* PTRS */ - if (ul_bwp_id > 0 && - mac->ULbwp[ul_bwp_id-1] && - mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated && - mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pusch_Config && - mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pusch_Config->choice.setup && - mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pusch_Config->choice.setup->dmrs_UplinkForPUSCH_MappingTypeB && - mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pusch_Config->choice.setup->dmrs_UplinkForPUSCH_MappingTypeB->choice.setup->phaseTrackingRS) { + if (pusch_Config && pusch_Config->dmrs_UplinkForPUSCH_MappingTypeB && pusch_Config->dmrs_UplinkForPUSCH_MappingTypeB->choice.setup->phaseTrackingRS) { if (pusch_config_pdu->transform_precoding == NR_PUSCH_Config__transformPrecoder_disabled) { nfapi_nr_ue_ptrs_ports_t ptrs_ports_list; pusch_config_pdu->pusch_ptrs.ptrs_ports_list = &ptrs_ports_list; - valid_ptrs_setup = set_ul_ptrs_values(mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->pusch_Config->choice.setup->dmrs_UplinkForPUSCH_MappingTypeB->choice.setup->phaseTrackingRS->choice.setup, - pusch_config_pdu->rb_size, pusch_config_pdu->mcs_index, pusch_config_pdu->mcs_table, - &pusch_config_pdu->pusch_ptrs.ptrs_freq_density,&pusch_config_pdu->pusch_ptrs.ptrs_time_density, - &pusch_config_pdu->pusch_ptrs.ptrs_ports_list->ptrs_re_offset,&pusch_config_pdu->pusch_ptrs.num_ptrs_ports, - &pusch_config_pdu->pusch_ptrs.ul_ptrs_power, pusch_config_pdu->nr_of_symbols); + valid_ptrs_setup = set_ul_ptrs_values(pusch_Config->dmrs_UplinkForPUSCH_MappingTypeB->choice.setup->phaseTrackingRS->choice.setup, + pusch_config_pdu->rb_size, + pusch_config_pdu->mcs_index, + pusch_config_pdu->mcs_table, + &pusch_config_pdu->pusch_ptrs.ptrs_freq_density, + &pusch_config_pdu->pusch_ptrs.ptrs_time_density, + &pusch_config_pdu->pusch_ptrs.ptrs_ports_list->ptrs_re_offset, + &pusch_config_pdu->pusch_ptrs.num_ptrs_ports, + &pusch_config_pdu->pusch_ptrs.ul_ptrs_power, + pusch_config_pdu->nr_of_symbols); if(valid_ptrs_setup==true) { pusch_config_pdu->pdu_bit_map |= PUSCH_PDU_BITMAP_PUSCH_PTRS; } LOG_D(NR_MAC, "UL PTRS values: PTRS time den: %d, PTRS freq den: %d\n", pusch_config_pdu->pusch_ptrs.ptrs_time_density, pusch_config_pdu->pusch_ptrs.ptrs_freq_density); } } - } LOG_D(NR_MAC, "In %s: received UL grant (rb_start %d, rb_size %d, start_symbol_index %d, nr_of_symbols %d) for RNTI type %s \n", @@ -937,21 +794,8 @@ bool nr_ue_periodic_srs_scheduling(module_id_t mod_id, frame_t frame, slot_t slo NR_UE_MAC_INST_t *mac = get_mac_inst(mod_id); NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; - NR_BWP_Id_t ul_bwp_id = current_UL_BWP->bwp_id; - NR_SRS_Config_t *srs_config = NULL; - if (ul_bwp_id > 0 && mac->ULbwp[ul_bwp_id-1]) { - if (mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated && - mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->srs_Config) { - srs_config = mac->ULbwp[ul_bwp_id-1]->bwp_Dedicated->srs_Config->choice.setup; - } - } else if (mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP) { - srs_config = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->srs_Config->choice.setup; - } + NR_SRS_Config_t *srs_config = current_UL_BWP->srs_Config; if (!srs_config) { return false; @@ -1559,30 +1403,25 @@ int nr_get_sf_retxBSRTimer(uint8_t sf_offset) { // PUSCH Msg3 scheduler: // - scheduled by RAR UL grant according to 8.3 of TS 38.213 // Note: Msg3 tx in the uplink symbols of mixed slot -int nr_ue_pusch_scheduler(NR_UE_MAC_INST_t *mac, - uint8_t is_Msg3, - frame_t current_frame, - int current_slot, - frame_t *frame_tx, - int *slot_tx, - uint8_t tda_id){ +int nr_ue_pusch_scheduler(NR_UE_MAC_INST_t *mac, uint8_t is_Msg3, frame_t current_frame, int current_slot, frame_t *frame_tx, int *slot_tx, long k2) +{ + AssertFatal(k2 >= DURATION_RX_TO_TX, + "Slot offset K2 (%ld) cannot be less than DURATION_RX_TO_TX (%d). Please set min_rxtxtime at least to %d in gNB config file or gNBs.[0].min_rxtxtime=%d via command line.\n", + k2, + DURATION_RX_TO_TX, + DURATION_RX_TO_TX, + DURATION_RX_TO_TX); int delta = 0; NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; - NR_BWP_Id_t ul_bwp_id = current_UL_BWP->bwp_id; // Get the numerology to calculate the Tx frame and slot int mu = current_UL_BWP->scs; - NR_PUSCH_TimeDomainResourceAllocationList_t *pusch_TimeDomainAllocationList = ul_bwp_id > 0 && mac->ULbwp[ul_bwp_id-1] ? - mac->ULbwp[ul_bwp_id-1]->bwp_Common->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList: - mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList; // k2 as per 3GPP TS 38.214 version 15.9.0 Release 15 ch 6.1.2.1.1 // PUSCH time domain resource allocation is higher layer configured from uschTimeDomainAllocationList in either pusch-ConfigCommon - int k2; if (is_Msg3) { - k2 = *pusch_TimeDomainAllocationList->list.array[tda_id]->k2; switch (mu) { case 0: @@ -1599,9 +1438,12 @@ int nr_ue_pusch_scheduler(NR_UE_MAC_INST_t *mac, break; } - AssertFatal((k2+delta) >= DURATION_RX_TO_TX, - "Slot offset (%d) for Msg3 cannot be less than DURATION_RX_TO_TX (%d)\n", - k2+delta,DURATION_RX_TO_TX); + AssertFatal((k2 + delta) >= DURATION_RX_TO_TX, + "Slot offset (%ld) for Msg3 cannot be less than DURATION_RX_TO_TX (%d). Please set min_rxtxtime at least to %d in gNB config file or gNBs.[0].min_rxtxtime=%d via command line.\n", + k2, + DURATION_RX_TO_TX, + DURATION_RX_TO_TX, + DURATION_RX_TO_TX); *slot_tx = (current_slot + k2 + delta) % nr_slots_per_frame[mu]; if (current_slot + k2 + delta > nr_slots_per_frame[mu]){ @@ -1612,10 +1454,8 @@ int nr_ue_pusch_scheduler(NR_UE_MAC_INST_t *mac, } else { - // Get slot offset K2 which will be used to calculate TX slot - k2 = get_k2(mac, tda_id); if (k2 < 0) { // This can happen when a false DCI is received - LOG_W(PHY,"%d.%d. Received k2 %d, tda_id %d\n",current_frame,current_slot,k2,tda_id); + LOG_W(PHY, "%d.%d. Received k2 %ld\n", current_frame, current_slot, k2); return -1; } @@ -1625,10 +1465,9 @@ int nr_ue_pusch_scheduler(NR_UE_MAC_INST_t *mac, } - LOG_D(NR_MAC, "In %s: currently at [%d.%d] UL transmission in [%d.%d] (k2 %d delta %d)\n", __FUNCTION__, current_frame, current_slot, *frame_tx, *slot_tx, k2, delta); + LOG_D(NR_MAC, "[%04d.%02d] UL transmission in [%04d.%02d] (k2 %ld delta %d)\n", current_frame, current_slot, *frame_tx, *slot_tx, k2, delta); return 0; - } // Build the list of all the valid RACH occasions in the maximum association pattern period according to the PRACH config @@ -1649,7 +1488,7 @@ static void build_ro_list(NR_UE_MAC_INST_t *mac) { uint8_t format2 = 0xff; int nb_fdm; - uint8_t config_index, mu; + uint8_t config_index; int msg1_FDM; uint8_t prach_conf_period_idx; @@ -1657,20 +1496,16 @@ static void build_ro_list(NR_UE_MAC_INST_t *mac) { uint8_t prach_conf_period_frame_idx; int64_t *prach_config_info_p; - NR_RACH_ConfigCommon_t *setup = (mac->scc) ? - mac->scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup: - mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.rach_ConfigCommon->choice.setup; + NR_RACH_ConfigCommon_t *setup = mac->current_UL_BWP.rach_ConfigCommon; NR_RACH_ConfigGeneric_t *rach_ConfigGeneric = &setup->rach_ConfigGeneric; config_index = rach_ConfigGeneric->prach_ConfigurationIndex; - if (setup->msg1_SubcarrierSpacing) { + int mu; + if (setup->msg1_SubcarrierSpacing) mu = *setup->msg1_SubcarrierSpacing; - } else if(mac->scc) { - mu = mac->scc->downlinkConfigCommon->frequencyInfoDL->scs_SpecificCarrierList.list.array[0]->subcarrierSpacing; - } else { - mu = get_softmodem_params()->numerology; - } + else + mu = mac->current_UL_BWP.scs; msg1_FDM = rach_ConfigGeneric->msg1_FDM; @@ -1905,9 +1740,7 @@ static void map_ssb_to_ro(NR_UE_MAC_INST_t *mac) { // Map SSBs to PRACH occasions // =========================== // WIP: Assumption: No PRACH occasion is rejected because of a conflict with SSBs or TDD_UL_DL_ConfigurationCommon schedule - NR_RACH_ConfigCommon_t *setup = (mac->scc) ? - mac->scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup: - mac->scc_SIB->uplinkConfigCommon->initialUplinkBWP.rach_ConfigCommon->choice.setup; + NR_RACH_ConfigCommon_t *setup = mac->current_UL_BWP.rach_ConfigCommon; NR_RACH_ConfigCommon__ssb_perRACH_OccasionAndCB_PreamblesPerSSB_PR ssb_perRACH_config = setup->ssb_perRACH_OccasionAndCB_PreamblesPerSSB->present; bool multiple_ssb_per_ro; // true if more than one or exactly one SSB per RACH occasion, false if more than one RO per SSB @@ -2317,27 +2150,7 @@ void nr_ue_pucch_scheduler(module_id_t module_idP, frame_t frameP, int slotP, vo // ACKNACK O_ACK = get_downlink_ack(mac, frameP, slotP, &pucch); - NR_BWP_Id_t bwp_id = mac->current_UL_BWP.bwp_id; - NR_PUCCH_Config_t *pucch_Config = NULL; - - if (bwp_id>0 && - mac->ULbwp[bwp_id-1] && - mac->ULbwp[bwp_id-1]->bwp_Dedicated && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config && - mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup) { - pucch_Config = mac->ULbwp[bwp_id-1]->bwp_Dedicated->pucch_Config->choice.setup; - } - else if (bwp_id==0 && - mac->cg && - mac->cg->spCellConfig && - mac->cg->spCellConfig->spCellConfigDedicated && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config && - mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup) { - pucch_Config = mac->cg->spCellConfig->spCellConfigDedicated->uplinkConfig->initialUplinkBWP->pucch_Config->choice.setup; - } - + NR_PUCCH_Config_t *pucch_Config = mac->current_UL_BWP.pucch_Config; // if multiplexing of HARQ and CSI is not possible, transmit only HARQ bits if ((O_ACK != 0) && (O_CSI != 0) && @@ -2377,7 +2190,6 @@ void nr_ue_pucch_scheduler(module_id_t module_idP, frame_t frameP, int slotP, vo &pucch, pucch_pdu, O_SR, O_ACK, O_CSI); - LOG_D(NR_MAC, "Configuring pucch, is_common = %d\n", pucch.is_common); nr_scheduled_response_t scheduled_response; fill_scheduled_response(&scheduled_response, NULL, ul_config, NULL, module_idP, 0 /*TBR fix*/, frameP, slotP, phy_data); if(mac->if_module != NULL && mac->if_module->scheduled_response != NULL) @@ -2391,13 +2203,12 @@ void nr_schedule_csi_for_im(NR_UE_MAC_INST_t *mac, int frame, int slot) { if (mac->ra.ra_state != RA_SUCCEEDED && get_softmodem_params()->phy_test == 0) return; - NR_CellGroupConfig_t *CellGroup = mac->cg; + NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; - if (!CellGroup || !CellGroup->spCellConfig || !CellGroup->spCellConfig->spCellConfigDedicated || - !CellGroup->spCellConfig->spCellConfigDedicated->csi_MeasConfig) + if (!current_UL_BWP->csi_MeasConfig) return; - NR_CSI_MeasConfig_t *csi_measconfig = CellGroup->spCellConfig->spCellConfigDedicated->csi_MeasConfig->choice.setup; + NR_CSI_MeasConfig_t *csi_measconfig = current_UL_BWP->csi_MeasConfig; if (csi_measconfig->csi_IM_ResourceToAddModList == NULL) return; @@ -2449,13 +2260,12 @@ void nr_schedule_csirs_reception(NR_UE_MAC_INST_t *mac, int frame, int slot) { if (mac->ra.ra_state != RA_SUCCEEDED && get_softmodem_params()->phy_test == 0) return; - NR_CellGroupConfig_t *CellGroup = mac->cg; + NR_UE_UL_BWP_t *current_UL_BWP = &mac->current_UL_BWP; - if (!CellGroup || !CellGroup->spCellConfig || !CellGroup->spCellConfig->spCellConfigDedicated || - !CellGroup->spCellConfig->spCellConfigDedicated->csi_MeasConfig) + if (!current_UL_BWP->csi_MeasConfig) return; - NR_CSI_MeasConfig_t *csi_measconfig = CellGroup->spCellConfig->spCellConfigDedicated->csi_MeasConfig->choice.setup; + NR_CSI_MeasConfig_t *csi_measconfig = current_UL_BWP->csi_MeasConfig; if (csi_measconfig->nzp_CSI_RS_ResourceToAddModList == NULL) return; @@ -2636,9 +2446,7 @@ void nr_ue_prach_scheduler(module_id_t module_idP, frame_t frameP, sub_frame_t s NR_ServingCellConfigCommon_t *scc = mac->scc; NR_ServingCellConfigCommonSIB_t *scc_SIB = mac->scc_SIB; - NR_RACH_ConfigCommon_t *setup; - if (scc!=NULL) setup = scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup; - else setup = scc_SIB->uplinkConfigCommon->initialUplinkBWP.rach_ConfigCommon->choice.setup; + NR_RACH_ConfigCommon_t *setup = mac->current_UL_BWP.rach_ConfigCommon; NR_RACH_ConfigGeneric_t *rach_ConfigGeneric = &setup->rach_ConfigGeneric; NR_TDD_UL_DL_ConfigCommon_t *tdd_config = scc==NULL ? scc_SIB->tdd_UL_DL_ConfigurationCommon : scc->tdd_UL_DL_ConfigurationCommon; diff --git a/openair2/LAYER2/NR_MAC_gNB/config.c b/openair2/LAYER2/NR_MAC_gNB/config.c index f55a3f1a27..08a3f291fb 100644 --- a/openair2/LAYER2/NR_MAC_gNB/config.c +++ b/openair2/LAYER2/NR_MAC_gNB/config.c @@ -474,9 +474,11 @@ int rrc_mac_config_req_gNB(module_id_t Mod_idP, if (scc != NULL ) { AssertFatal((scc->ssb_PositionsInBurst->present > 0) && (scc->ssb_PositionsInBurst->present < 4), "SSB Bitmap type %d is not valid\n",scc->ssb_PositionsInBurst->present); - const int n = nr_slots_per_frame[*scc->ssbSubcarrierSpacing]; - RC.nrmac[Mod_idP]->common_channels[0].vrb_map_UL = - calloc(n * MAX_BWP_SIZE, sizeof(uint16_t)); + int n = nr_slots_per_frame[*scc->ssbSubcarrierSpacing]; + if (*scc->ssbSubcarrierSpacing == 0) + n <<= 1; // to have enough room for feedback possibly beyond the frame we need a larger array at 15kHz SCS + RC.nrmac[Mod_idP]->common_channels[0].vrb_map_UL = calloc(n * MAX_BWP_SIZE, sizeof(uint16_t)); + RC.nrmac[Mod_idP]->vrb_map_UL_size = n; AssertFatal(RC.nrmac[Mod_idP]->common_channels[0].vrb_map_UL, "could not allocate memory for RC.nrmac[]->common_channels[0].vrb_map_UL\n"); diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler.c index 3009dc6b6c..895f0bd0dc 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler.c @@ -59,40 +59,40 @@ const uint8_t nr_rv_round_map[4] = { 0, 2, 3, 1 }; uint16_t nr_pdcch_order_table[6] = { 31, 31, 511, 2047, 2047, 8191 }; -uint8_t vnf_first_sched_entry = 1; - void clear_nr_nfapi_information(gNB_MAC_INST * gNB, int CC_idP, frame_t frameP, sub_frame_t slotP) { + NR_ServingCellConfigCommon_t *scc = gNB->common_channels->ServingCellConfigCommon; const int num_slots = nr_slots_per_frame[*scc->ssbSubcarrierSpacing]; - UL_tti_req_ahead_initialization(gNB, scc, num_slots, CC_idP, frameP); + UL_tti_req_ahead_initialization(gNB, scc, num_slots, CC_idP, frameP, slotP, *scc->ssbSubcarrierSpacing); nfapi_nr_dl_tti_request_t *DL_req = &gNB->DL_req[0]; nfapi_nr_dl_tti_pdcch_pdu_rel15_t **pdcch = (nfapi_nr_dl_tti_pdcch_pdu_rel15_t **)gNB->pdcch_pdu_idx[CC_idP]; - nfapi_nr_ul_tti_request_t *future_ul_tti_req = &gNB->UL_tti_req_ahead[CC_idP][(slotP + num_slots - 1) % num_slots]; nfapi_nr_ul_dci_request_t *UL_dci_req = &gNB->UL_dci_req[0]; nfapi_nr_tx_data_request_t *TX_req = &gNB->TX_req[0]; gNB->pdu_index[CC_idP] = 0; - DL_req[CC_idP].SFN = frameP; - DL_req[CC_idP].Slot = slotP; + DL_req[CC_idP].SFN = frameP; + DL_req[CC_idP].Slot = slotP; DL_req[CC_idP].dl_tti_request_body.nPDUs = 0; - DL_req[CC_idP].dl_tti_request_body.nGroup = 0; - //DL_req[CC_idP].dl_tti_request_body.transmission_power_pcfich = 6000; + DL_req[CC_idP].dl_tti_request_body.nGroup = 0; memset(pdcch, 0, sizeof(*pdcch) * MAX_NUM_CORESET); - UL_dci_req[CC_idP].SFN = frameP; - UL_dci_req[CC_idP].Slot = slotP; - UL_dci_req[CC_idP].numPdus = 0; + UL_dci_req[CC_idP].SFN = frameP; + UL_dci_req[CC_idP].Slot = slotP; + UL_dci_req[CC_idP].numPdus = 0; /* advance last round's future UL_tti_req to be ahead of current frame/slot */ - future_ul_tti_req->SFN = (slotP == 0 ? frameP : frameP + 1) % 1024; - LOG_D(NR_MAC, "In %s: UL_tti_req_ahead SFN.slot = %d.%d for slot %d \n", __FUNCTION__, future_ul_tti_req->SFN, future_ul_tti_req->Slot, (slotP + num_slots - 1) % num_slots); + const int size = gNB->UL_tti_req_ahead_size; + const int prev_slot = frameP * num_slots + slotP + size - 1; + nfapi_nr_ul_tti_request_t *future_ul_tti_req = &gNB->UL_tti_req_ahead[CC_idP][prev_slot % size]; + future_ul_tti_req->SFN = (prev_slot / num_slots) % 1024; + LOG_D(NR_MAC, "%d.%d UL_tti_req_ahead SFN.slot = %d.%d for index %d \n", frameP, slotP, future_ul_tti_req->SFN, future_ul_tti_req->Slot, prev_slot % size); /* future_ul_tti_req->Slot is fixed! */ future_ul_tti_req->n_pdus = 0; future_ul_tti_req->n_ulsch = 0; @@ -101,9 +101,9 @@ void clear_nr_nfapi_information(gNB_MAC_INST * gNB, /* UL_tti_req is a simple pointer into the current UL_tti_req_ahead, i.e., * it walks over UL_tti_req_ahead in a circular fashion */ - gNB->UL_tti_req[CC_idP] = &gNB->UL_tti_req_ahead[CC_idP][slotP]; - - TX_req[CC_idP].Number_of_PDUs = 0; + const int current_index = ul_buffer_index(frameP, slotP, *scc->ssbSubcarrierSpacing, gNB->UL_tti_req_ahead_size); + gNB->UL_tti_req[CC_idP] = &gNB->UL_tti_req_ahead[CC_idP][current_index]; + TX_req[CC_idP].Number_of_PDUs = 0; } @@ -113,9 +113,10 @@ bool is_xlsch_in_slot(uint64_t bitmap, sub_frame_t slot) { void gNB_dlsch_ulsch_scheduler(module_id_t module_idP, frame_t frame, - sub_frame_t slot){ + sub_frame_t slot) +{ - protocol_ctxt_t ctxt={0}; + protocol_ctxt_t ctxt = {0}; PROTOCOL_CTXT_SET_BY_MODULE_ID(&ctxt, module_idP, ENB_FLAG_YES, NOT_A_RNTI, frame, slot,module_idP); gNB_MAC_INST *gNB = RC.nrmac[module_idP]; @@ -135,7 +136,7 @@ void gNB_dlsch_ulsch_scheduler(module_id_t module_idP, gNB->frame = frame; gNB->slot = slot; - start_meas(&RC.nrmac[module_idP]->eNB_scheduler); + start_meas(&gNB->eNB_scheduler); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_gNB_DLSCH_ULSCH_SCHEDULER,VCD_FUNCTION_IN); pdcp_run(&ctxt); @@ -155,32 +156,18 @@ void gNB_dlsch_ulsch_scheduler(module_id_t module_idP, memset(cc[CC_id].vrb_map, 0, sizeof(uint16_t) * MAX_BWP_SIZE); // clear last scheduled slot's content (only)! const int num_slots = nr_slots_per_frame[*scc->ssbSubcarrierSpacing]; - const int last_slot = (slot + num_slots - 1) % num_slots; + const int size = gNB->vrb_map_UL_size; + const int prev_slot = frame * num_slots + slot + size - 1; uint16_t *vrb_map_UL = cc[CC_id].vrb_map_UL; - memcpy(&vrb_map_UL[last_slot * MAX_BWP_SIZE], &RC.nrmac[module_idP]->ulprbbl, sizeof(uint16_t) * MAX_BWP_SIZE); + memcpy(&vrb_map_UL[prev_slot % size * MAX_BWP_SIZE], &gNB->ulprbbl, sizeof(uint16_t) * MAX_BWP_SIZE); - clear_nr_nfapi_information(RC.nrmac[module_idP], CC_id, frame, slot); + clear_nr_nfapi_information(gNB, CC_id, frame, slot); - /*VNF first entry into scheduler. Since frame numbers for future_ul_tti_req of some future slots - will not be set before we encounter them, set them here */ - - if (NFAPI_MODE == NFAPI_MODE_VNF){ - if(vnf_first_sched_entry == 1) - { - for (int i = 0; iUL_tti_req_ahead[CC_id][i].SFN = (frame + 1) % 1024; - else - gNB->UL_tti_req_ahead[CC_id][i].SFN = frame; - } - vnf_first_sched_entry = 0; - } - } } if ((slot == 0) && (frame & 127) == 0) { char stats_output[16000] = {0}; - dump_mac_stats(RC.nrmac[module_idP], stats_output, sizeof(stats_output), true); + dump_mac_stats(gNB, stats_output, sizeof(stats_output), true); LOG_I(NR_MAC, "Frame.Slot %d.%d\n%s\n", frame, slot, stats_output); } @@ -232,11 +219,11 @@ void gNB_dlsch_ulsch_scheduler(module_id_t module_idP, nr_schedule_ue_spec(module_idP, frame, slot); stop_meas(&gNB->schedule_dlsch); - nr_sr_reporting(RC.nrmac[module_idP], frame, slot); + nr_sr_reporting(gNB, frame, slot); - nr_schedule_pucch(RC.nrmac[module_idP], frame, slot); + nr_schedule_pucch(gNB, frame, slot); - stop_meas(&RC.nrmac[module_idP]->eNB_scheduler); + stop_meas(&gNB->eNB_scheduler); VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_gNB_DLSCH_ULSCH_SCHEDULER,VCD_FUNCTION_OUT); } diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_RA.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_RA.c index 6d968a513e..06d58bae57 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_RA.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_RA.c @@ -245,24 +245,25 @@ void schedule_nr_prach(module_id_t module_idP, frame_t frameP, sub_frame_t slotP gNB_MAC_INST *gNB = RC.nrmac[module_idP]; NR_COMMON_channels_t *cc = gNB->common_channels; NR_ServingCellConfigCommon_t *scc = cc->ServingCellConfigCommon; - nfapi_nr_ul_tti_request_t *UL_tti_req = &RC.nrmac[module_idP]->UL_tti_req_ahead[0][slotP]; + int mu; + if (scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup->msg1_SubcarrierSpacing) + mu = *scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup->msg1_SubcarrierSpacing; + else + mu = scc->downlinkConfigCommon->frequencyInfoDL->scs_SpecificCarrierList.list.array[0]->subcarrierSpacing; + int index = ul_buffer_index(frameP, slotP, mu, gNB->UL_tti_req_ahead_size); + nfapi_nr_ul_tti_request_t *UL_tti_req = &RC.nrmac[module_idP]->UL_tti_req_ahead[0][index]; nfapi_nr_config_request_scf_t *cfg = &RC.nrmac[module_idP]->config[0]; if (is_nr_UL_slot(scc->tdd_UL_DL_ConfigurationCommon, slotP, cc->frame_type)) { uint8_t config_index = scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup->rach_ConfigGeneric.prach_ConfigurationIndex; - uint8_t mu,N_dur,N_t_slot,start_symbol = 0,N_RA_slot; + uint8_t N_dur, N_t_slot, start_symbol = 0, N_RA_slot; uint16_t RA_sfn_index = -1; uint8_t config_period = 1; uint16_t format; int slot_index = 0; uint16_t prach_occasion_id = -1; - if (scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup->msg1_SubcarrierSpacing) - mu = *scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup->msg1_SubcarrierSpacing; - else - mu = scc->downlinkConfigCommon->frequencyInfoDL->scs_SpecificCarrierList.list.array[0]->subcarrierSpacing; - int bwp_start = NRRIV2PRBOFFSET(scc->uplinkConfigCommon->initialUplinkBWP->genericParameters.locationAndBandwidth, MAX_BWP_SIZE); uint8_t fdm = cfg->prach_config.num_prach_fd_occasions.value; @@ -391,7 +392,8 @@ void schedule_nr_prach(module_id_t module_idP, frame_t frameP, sub_frame_t slotP const uint8_t mu_pusch = scc->uplinkConfigCommon->frequencyInfoUL->scs_SpecificCarrierList.list.array[0]->subcarrierSpacing; const int16_t N_RA_RB = get_N_RA_RB(cfg->prach_config.prach_sub_c_spacing.value, mu_pusch); - uint16_t *vrb_map_UL = &cc->vrb_map_UL[slotP * MAX_BWP_SIZE]; + index = ul_buffer_index(frameP, slotP, mu, gNB->vrb_map_UL_size); + uint16_t *vrb_map_UL = &cc->vrb_map_UL[index * MAX_BWP_SIZE]; for (int i = 0; i < N_RA_RB * fdm; ++i) vrb_map_UL[bwp_start + rach_ConfigGeneric->msg1_FrequencyStart + i] |= SL_to_bitmap(start_symbol, N_t_slot*N_dur); } @@ -700,13 +702,13 @@ void nr_generate_Msg3_retransmission(module_id_t module_idP, int CC_id, frame_t NR_ServingCellConfigCommon_t *scc = cc->ServingCellConfigCommon; NR_UE_UL_BWP_t *ul_bwp = &ra->UL_BWP; - NR_PUSCH_TimeDomainResourceAllocationList_t *pusch_TimeDomainAllocationList = ul_bwp->tdaList; + NR_PUSCH_TimeDomainResourceAllocationList_t *pusch_TimeDomainAllocationList = ul_bwp->tdaList_Common; int mu = ul_bwp->scs; uint8_t K2 = *pusch_TimeDomainAllocationList->list.array[ra->Msg3_tda_id]->k2; const int sched_frame = frame + (slot + K2 >= nr_slots_per_frame[mu]); const int sched_slot = (slot + K2) % nr_slots_per_frame[mu]; - if (is_xlsch_in_slot(RC.nrmac[module_idP]->ulsch_slot_bitmap[sched_slot / 64], sched_slot)) { + if (is_xlsch_in_slot(nr_mac->ulsch_slot_bitmap[sched_slot / 64], sched_slot)) { // beam association for FR2 int16_t *tdd_beam_association = nr_mac->tdd_beam_association; if (*scc->downlinkConfigCommon->frequencyInfoDL->frequencyBandList.list.array[0] >= 257) { @@ -729,7 +731,8 @@ void nr_generate_Msg3_retransmission(module_id_t module_idP, int CC_id, frame_t SLIV2SL(startSymbolAndLength, &StartSymbolIndex, &NrOfSymbols); int mappingtype = pusch_TimeDomainAllocationList->list.array[ra->Msg3_tda_id]->mappingType; - uint16_t *vrb_map_UL = &RC.nrmac[module_idP]->common_channels[CC_id].vrb_map_UL[sched_slot * MAX_BWP_SIZE]; + int buffer_index = ul_buffer_index(sched_frame, sched_slot, mu, nr_mac->vrb_map_UL_size); + uint16_t *vrb_map_UL = &nr_mac->common_channels[CC_id].vrb_map_UL[buffer_index * MAX_BWP_SIZE]; const int BWPSize = ul_bwp->initial_BWPSize; const int BWPStart = ul_bwp->initial_BWPStart; @@ -749,7 +752,8 @@ void nr_generate_Msg3_retransmission(module_id_t module_idP, int CC_id, frame_t LOG_I(NR_MAC, "[gNB %d][RAPROC] Frame %d, Slot %d : CC_id %d Scheduling retransmission of Msg3 in (%d,%d)\n", module_idP, frame, slot, CC_id, sched_frame, sched_slot); - nfapi_nr_ul_tti_request_t *future_ul_tti_req = &RC.nrmac[module_idP]->UL_tti_req_ahead[CC_id][sched_slot]; + buffer_index = ul_buffer_index(sched_frame, sched_slot, mu, nr_mac->UL_tti_req_ahead_size); + nfapi_nr_ul_tti_request_t *future_ul_tti_req = &nr_mac->UL_tti_req_ahead[CC_id][buffer_index]; AssertFatal(future_ul_tti_req->SFN == sched_frame && future_ul_tti_req->Slot == sched_slot, "future UL_tti_req's frame.slot %d.%d does not match PUSCH %d.%d\n", @@ -871,12 +875,14 @@ void nr_get_Msg3alloc(module_id_t module_id, sub_frame_t current_slot, frame_t current_frame, NR_RA_t *ra, - int16_t *tdd_beam_association) { + int16_t *tdd_beam_association) +{ // msg3 is scheduled in mixed slot in the following TDD period uint16_t msg3_nb_rb = 8; // sdu has 6 or 8 bytes - frame_type_t frame_type = RC.nrmac[module_id]->common_channels->frame_type; + gNB_MAC_INST *mac = RC.nrmac[module_id]; + frame_type_t frame_type = mac->common_channels->frame_type; NR_UE_UL_BWP_t *ul_bwp = &ra->UL_BWP; @@ -888,7 +894,7 @@ void nr_get_Msg3alloc(module_id_t module_id, int Msg3maxsymb = 14, Msg3start = 0; ra->Msg3_tda_id = 16; // initialization to a value above limit - NR_PUSCH_TimeDomainResourceAllocationList_t *pusch_TimeDomainAllocationList = ul_bwp->tdaList; + NR_PUSCH_TimeDomainResourceAllocationList_t *pusch_TimeDomainAllocationList = ul_bwp->tdaList_Common; const NR_TDD_UL_DL_Pattern_t *tdd = scc->tdd_UL_DL_ConfigurationCommon ? &scc->tdd_UL_DL_ConfigurationCommon->pattern1 : NULL; const int n_slots_frame = nr_slots_per_frame[mu]; @@ -915,7 +921,7 @@ void nr_get_Msg3alloc(module_id_t module_id, abs_slot = (current_slot + k2 + DELTA[mu]); int temp_slot = abs_slot % nr_slots_per_frame[mu]; // msg3 slot according to 8.3 in 38.213 if ((temp_slot % nb_slots_per_period) == msg3_slot && - is_xlsch_in_slot(RC.nrmac[module_id]->ulsch_slot_bitmap[temp_slot / 64], temp_slot) && + is_xlsch_in_slot(mac->ulsch_slot_bitmap[temp_slot / 64], temp_slot) && StartSymbolIndex == Msg3start && NrOfSymbols <= Msg3maxsymb) { ra->Msg3_tda_id = i; @@ -953,8 +959,8 @@ void nr_get_Msg3alloc(module_id_t module_id, } LOG_I(NR_MAC, "[RAPROC] Msg3 slot %d: current slot %u Msg3 frame %u k2 %u Msg3_tda_id %u\n", ra->Msg3_slot, current_slot, ra->Msg3_frame, k2,ra->Msg3_tda_id); - uint16_t *vrb_map_UL = - &RC.nrmac[module_id]->common_channels[CC_id].vrb_map_UL[ra->Msg3_slot * MAX_BWP_SIZE]; + const int buffer_index = ul_buffer_index(ra->Msg3_frame, ra->Msg3_slot, mu, mac->vrb_map_UL_size); + uint16_t *vrb_map_UL = &mac->common_channels[CC_id].vrb_map_UL[buffer_index * MAX_BWP_SIZE]; int bwpSize = ul_bwp->initial_BWPSize; int bwpStart = ul_bwp->initial_BWPStart; @@ -1080,8 +1086,10 @@ void nr_add_msg3(module_id_t module_idP, int CC_id, frame_t frameP, sub_frame_t return; } + const int scs = ul_bwp->scs; const uint16_t mask = SL_to_bitmap(ra->msg3_startsymb, ra->msg3_nrsymb); - uint16_t *vrb_map_UL = &RC.nrmac[module_idP]->common_channels[CC_id].vrb_map_UL[ra->Msg3_slot * MAX_BWP_SIZE]; + int buffer_index = ul_buffer_index(ra->Msg3_frame, ra->Msg3_slot, scs, mac->vrb_map_UL_size); + uint16_t *vrb_map_UL = &RC.nrmac[module_idP]->common_channels[CC_id].vrb_map_UL[buffer_index * MAX_BWP_SIZE]; for (int i = 0; i < ra->msg3_nb_rb; ++i) { AssertFatal(!(vrb_map_UL[i + ra->msg3_first_rb + ra->msg3_bwp_start] & mask), "RB %d in %4d.%2d is already taken, cannot allocate Msg3!\n", @@ -1092,8 +1100,8 @@ void nr_add_msg3(module_id_t module_idP, int CC_id, frame_t frameP, sub_frame_t } LOG_D(NR_MAC, "[gNB %d][RAPROC] Frame %d, Slot %d : CC_id %d RA is active, Msg3 in (%d,%d)\n", module_idP, frameP, slotP, CC_id, ra->Msg3_frame, ra->Msg3_slot); - - nfapi_nr_ul_tti_request_t *future_ul_tti_req = &RC.nrmac[module_idP]->UL_tti_req_ahead[CC_id][ra->Msg3_slot]; + buffer_index = ul_buffer_index(ra->Msg3_frame, ra->Msg3_slot, scs, mac->UL_tti_req_ahead_size); + nfapi_nr_ul_tti_request_t *future_ul_tti_req = &RC.nrmac[module_idP]->UL_tti_req_ahead[CC_id][buffer_index]; AssertFatal(future_ul_tti_req->SFN == ra->Msg3_frame && future_ul_tti_req->Slot == ra->Msg3_slot, "future UL_tti_req's frame.slot %d.%d does not match PUSCH %d.%d\n", @@ -1107,10 +1115,9 @@ void nr_add_msg3(module_id_t module_idP, int CC_id, frame_t frameP, sub_frame_t memset(pusch_pdu, 0, sizeof(nfapi_nr_pusch_pdu_t)); const int ibwp_size = ul_bwp->initial_BWPSize; - const int scs = ul_bwp->scs; const int fh = (ul_bwp->pusch_Config && ul_bwp->pusch_Config->frequencyHopping) ? 1 : 0; - const int startSymbolAndLength = ul_bwp->tdaList->list.array[ra->Msg3_tda_id]->startSymbolAndLength; - const int mappingtype = ul_bwp->tdaList->list.array[ra->Msg3_tda_id]->mappingType; + const int startSymbolAndLength = ul_bwp->tdaList_Common->list.array[ra->Msg3_tda_id]->startSymbolAndLength; + const int mappingtype = ul_bwp->tdaList_Common->list.array[ra->Msg3_tda_id]->mappingType; LOG_D(NR_MAC, "Frame %d, Slot %d Adding Msg3 UL Config Request for (%d,%d) : (%d,%d,%d) for rnti: %d\n", frameP, @@ -1162,14 +1169,14 @@ void nr_generate_Msg2(module_id_t module_idP, int CC_id, frame_t frameP, sub_fra BWPSize = type0_PDCCH_CSS_config->num_rbs; } + NR_ControlResourceSet_t *coreset = ra->coreset; + AssertFatal(coreset != NULL,"Coreset cannot be null for RA-Msg2\n"); + const int coresetid = coreset->controlResourceSetId; // Calculate number of symbols int time_domain_assignment = get_dl_tda(nr_mac, scc, slotP); - NR_tda_info_t tda_info = nr_get_pdsch_tda_info(dl_bwp, - time_domain_assignment); - - NR_ControlResourceSet_t *coreset = ra->coreset; - - AssertFatal(coreset!=NULL,"Coreset cannot be null for RA-Msg2\n"); + int mux_pattern = type0_PDCCH_CSS_config ? type0_PDCCH_CSS_config->type0_pdcch_ss_mux_pattern : 1; + NR_tda_info_t tda_info = get_dl_tda_info(dl_bwp, ss->searchSpaceType->present, time_domain_assignment, + scc->dmrs_TypeA_Position, mux_pattern, NR_RNTI_RA, coresetid, false); uint16_t *vrb_map = cc[CC_id].vrb_map; for (int i = 0; (i < rbSize) && (rbStart <= (BWPSize - rbSize)); i++) { @@ -1209,7 +1216,6 @@ void nr_generate_Msg2(module_id_t module_idP, int CC_id, frame_t frameP, sub_fra // look up the PDCCH PDU for this CC, BWP, and CORESET. If it does not exist, create it. This is especially // important if we have multiple RAs, and the DLSCH has to reuse them, so we need to mark them - const int coresetid = coreset->controlResourceSetId; nfapi_nr_dl_tti_pdcch_pdu_rel15_t *pdcch_pdu_rel15 = nr_mac->pdcch_pdu_idx[CC_id][coresetid]; if (!pdcch_pdu_rel15) { nfapi_nr_dl_tti_request_pdu_t *dl_tti_pdcch_pdu = &dl_req->dl_tti_pdu_list[dl_req->nPDUs]; @@ -1481,7 +1487,10 @@ void nr_generate_Msg4(module_id_t module_idP, int CC_id, frame_t frameP, sub_fra } uint8_t time_domain_assignment = get_dl_tda(nr_mac, scc, slotP); - NR_tda_info_t msg4_tda = nr_get_pdsch_tda_info(dl_bwp, time_domain_assignment); + int mux_pattern = type0_PDCCH_CSS_config ? type0_PDCCH_CSS_config->type0_pdcch_ss_mux_pattern : 1; + NR_tda_info_t msg4_tda = get_dl_tda_info(dl_bwp, ss->searchSpaceType->present, time_domain_assignment, + scc->dmrs_TypeA_Position, mux_pattern, NR_RNTI_TC, coreset->controlResourceSetId, false); + NR_pdsch_dmrs_t dmrs_info = get_dl_dmrs_params(scc, dl_bwp, &msg4_tda, @@ -1540,7 +1549,7 @@ void nr_generate_Msg4(module_id_t module_idP, int CC_id, frame_t frameP, sub_fra LOG_D(NR_MAC,"[RAPROC] Msg4 r_pucch %d (CCEIndex %d, delta_PRI %d)\n", r_pucch, CCEIndex, delta_PRI); int alloc = nr_acknack_scheduling(nr_mac, UE, frameP, slotP, r_pucch, 1); - if (alloc<0) { + if (alloc < 0) { LOG_D(NR_MAC,"Couldn't find a pucch allocation for ack nack (msg4) in frame %d slot %d\n",frameP,slotP); return; } diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_bch.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_bch.c index dc2467e904..8db40a26c7 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_bch.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_bch.c @@ -569,23 +569,11 @@ void schedule_nr_sib1(module_id_t module_idP, frame_t frameP, sub_frame_t slotP) for (int k=0;ktype0_pdcch_ss_mux_pattern); + // assuming normal CP + NR_tda_info_t tda_info = get_info_from_tda_tables(table_type, time_domain_allocation, gNB_mac->common_channels->ServingCellConfigCommon->dmrs_TypeA_Position, true); - get_info_from_tda_tables(type0_PDCCH_CSS_config->type0_pdcch_ss_mux_pattern, - time_domain_allocation, - gNB_mac->common_channels->ServingCellConfigCommon->dmrs_TypeA_Position, - 1, &is_typeA, - &startSymbolIndex, &nrOfSymbols); - - AssertFatal((startSymbolIndex+nrOfSymbols)<14,"SIB1 TDA %d would cause overlap with CSI-RS. Please select a different SIB1 TDA.\n",time_domain_allocation); - - NR_tda_info_t tda_info = { - .mapping_type = is_typeA ? typeA : typeB, - .startSymbolIndex = startSymbolIndex, - .nrOfSymbols = nrOfSymbols - }; + AssertFatal((tda_info.startSymbolIndex + tda_info.nrOfSymbols) < 14, "SIB1 TDA %d would cause overlap with CSI-RS. Please select a different SIB1 TDA.\n", time_domain_allocation); NR_pdsch_dmrs_t dmrs_parms = get_dl_dmrs_params(scc, NULL, @@ -603,7 +591,7 @@ void schedule_nr_sib1(module_id_t module_idP, frame_t frameP, sub_frame_t slotP) nfapi_nr_dl_tti_request_body_t *dl_req = &gNB_mac->DL_req[CC_id].dl_tti_request_body; int pdu_index = gNB_mac->pdu_index[0]++; - nr_fill_nfapi_dl_sib1_pdu(module_idP, dl_req, pdu_index, type0_PDCCH_CSS_config, TBS, startSymbolIndex, nrOfSymbols); + nr_fill_nfapi_dl_sib1_pdu(module_idP, dl_req, pdu_index, type0_PDCCH_CSS_config, TBS, tda_info.startSymbolIndex, tda_info.nrOfSymbols); const int ntx_req = gNB_mac->TX_req[CC_id].Number_of_PDUs; nfapi_nr_pdu_t *tx_req = &gNB_mac->TX_req[CC_id].pdu_list[ntx_req]; diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_dlsch.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_dlsch.c index 7cd98a5e28..08611fccd5 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_dlsch.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_dlsch.c @@ -411,7 +411,9 @@ bool allocate_dl_retransmission(module_id_t module_id, /* Check first whether the old TDA can be reused * this helps allocate retransmission when TDA changes (e.g. new nrOfSymbols > old nrOfSymbols) */ - NR_tda_info_t temp_tda = nr_get_pdsch_tda_info(dl_bwp, tda); + NR_tda_info_t temp_tda = get_dl_tda_info(dl_bwp, sched_ctrl->search_space->searchSpaceType->present, tda, + scc->dmrs_TypeA_Position, 1, NR_RNTI_C, coresetid, false); + bool reuse_old_tda = (retInfo->tda_info.startSymbolIndex == temp_tda.startSymbolIndex) && (retInfo->tda_info.nrOfSymbols <= temp_tda.nrOfSymbols); LOG_D(NR_MAC, "[UE %x] %s old TDA, %s number of layers\n", UE->rnti, @@ -729,7 +731,9 @@ void pf_dl(module_id_t module_id, sched_pdsch->time_domain_allocation = get_dl_tda(mac, scc, slot); AssertFatal(sched_pdsch->time_domain_allocation>=0,"Unable to find PDSCH time domain allocation in list\n"); - sched_pdsch->tda_info = nr_get_pdsch_tda_info(dl_bwp, sched_pdsch->time_domain_allocation); + sched_pdsch->tda_info = get_dl_tda_info(dl_bwp, sched_ctrl->search_space->searchSpaceType->present, sched_pdsch->time_domain_allocation, + scc->dmrs_TypeA_Position, 1, NR_RNTI_C, coresetid, false); + NR_tda_info_t *tda_info = &sched_pdsch->tda_info; const uint16_t slbitmap = SL_to_bitmap(tda_info->startSymbolIndex, tda_info->nrOfSymbols); @@ -798,13 +802,12 @@ void nr_fr1_dlsch_preprocessor(module_id_t module_id, frame_t frame, sub_frame_t NR_UE_DL_BWP_t *current_BWP = &UE->current_DL_BWP; const int tda = get_dl_tda(RC.nrmac[module_id], scc, slot); int startSymbolIndex, nrOfSymbols; - const struct NR_PDSCH_TimeDomainResourceAllocationList *tdaList = current_BWP->tdaList; + const int coresetid = sched_ctrl->coreset->controlResourceSetId; + const struct NR_PDSCH_TimeDomainResourceAllocationList *tdaList = get_dl_tdalist(current_BWP, coresetid, sched_ctrl->search_space->searchSpaceType->present, NR_RNTI_C); AssertFatal(tda < tdaList->list.count, "time_domain_allocation %d>=%d\n", tda, tdaList->list.count); const int startSymbolAndLength = tdaList->list.array[tda]->startSymbolAndLength; SLIV2SL(startSymbolAndLength, &startSymbolIndex, &nrOfSymbols); - const int coresetid = sched_ctrl->coreset->controlResourceSetId; - const uint16_t bwpSize = coresetid == 0 ? RC.nrmac[module_id]->cset0_bwp_size : current_BWP->BWPSize; const uint16_t BWPStart = coresetid == 0 ? RC.nrmac[module_id]->cset0_bwp_start : current_BWP->BWPStart; diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_phytest.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_phytest.c index e2b4542a80..96756d41cd 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_phytest.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_phytest.c @@ -59,7 +59,9 @@ void nr_preprocessor_phytest(module_id_t module_id, const int CC_id = 0; const int tda = get_dl_tda(RC.nrmac[module_id], scc, slot); - NR_tda_info_t tda_info = nr_get_pdsch_tda_info(dl_bwp, tda); + NR_tda_info_t tda_info = get_dl_tda_info(dl_bwp, sched_ctrl->search_space->searchSpaceType->present, tda, + scc->dmrs_TypeA_Position, 1, NR_RNTI_C, sched_ctrl->coreset->controlResourceSetId, false); + sched_ctrl->sched_pdsch.tda_info = tda_info; sched_ctrl->sched_pdsch.time_domain_allocation = tda; @@ -205,7 +207,7 @@ bool nr_ul_preprocessor_phytest(module_id_t module_id, frame_t frame, sub_frame_ NR_UE_UL_BWP_t *ul_bwp = &UE->current_UL_BWP; const int mu = ul_bwp->scs; - const struct NR_PUSCH_TimeDomainResourceAllocationList *tdaList = ul_bwp->tdaList; + NR_PUSCH_TimeDomainResourceAllocationList_t *tdaList = get_ul_tdalist(ul_bwp, sched_ctrl->coreset->controlResourceSetId, sched_ctrl->search_space->searchSpaceType->present, NR_RNTI_C); const int temp_tda = get_ul_tda(nr_mac, scc, frame, slot); if (temp_tda < 0) return false; @@ -213,7 +215,7 @@ bool nr_ul_preprocessor_phytest(module_id_t module_id, frame_t frame, sub_frame_ "time domain assignment %d >= %d\n", temp_tda, tdaList->list.count); - int K2 = get_K2(ul_bwp->tdaList, temp_tda, mu); + int K2 = get_K2(tdaList, temp_tda, mu); const int sched_frame = frame + (slot + K2 >= nr_slots_per_frame[mu]); const int sched_slot = (slot + K2) % nr_slots_per_frame[mu]; const int tda = get_ul_tda(nr_mac, scc, sched_frame, sched_slot); @@ -240,11 +242,11 @@ bool nr_ul_preprocessor_phytest(module_id_t module_id, frame_t frame, sub_frame_ else rbSize = target_ul_bw; - NR_tda_info_t tda_info = nr_get_pusch_tda_info(ul_bwp, tda); + NR_tda_info_t tda_info = get_ul_tda_info(ul_bwp, sched_ctrl->coreset->controlResourceSetId, sched_ctrl->search_space->searchSpaceType->present, NR_RNTI_C, tda); sched_ctrl->sched_pusch.tda_info = tda_info; - uint16_t *vrb_map_UL = - &RC.nrmac[module_id]->common_channels[CC_id].vrb_map_UL[sched_slot * MAX_BWP_SIZE]; + const int buffer_index = ul_buffer_index(sched_frame, sched_slot, mu, nr_mac->vrb_map_UL_size); + uint16_t *vrb_map_UL = &nr_mac->common_channels[CC_id].vrb_map_UL[buffer_index * MAX_BWP_SIZE]; for (int i = rbStart; i < rbStart + rbSize; ++i) { if ((vrb_map_UL[i+BWPStart] & SL_to_bitmap(tda_info.startSymbolIndex, tda_info.nrOfSymbols)) != 0) { LOG_E(MAC, diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_primitives.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_primitives.c index 4814250fd9..fea0531bee 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_primitives.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_primitives.c @@ -563,30 +563,6 @@ bool nr_find_nb_rb(uint16_t Qm, return *tbs >= bytes && *nb_rb <= nb_rb_max; } -NR_tda_info_t nr_get_pdsch_tda_info(const NR_UE_DL_BWP_t *dl_bwp, - const int tda) { - - NR_tda_info_t tda_info = {0}; - NR_PDSCH_TimeDomainResourceAllocationList_t *tdaList = dl_bwp->tdaList; - AssertFatal(tda < tdaList->list.count, "time_domain_allocation %d>=%d\n", tda, tdaList->list.count); - tda_info.mapping_type = tdaList->list.array[tda]->mappingType; - const int startSymbolAndLength = tdaList->list.array[tda]->startSymbolAndLength; - SLIV2SL(startSymbolAndLength, &tda_info.startSymbolIndex, &tda_info.nrOfSymbols); - return tda_info; -} - -NR_tda_info_t nr_get_pusch_tda_info(const NR_UE_UL_BWP_t *ul_bwp, - const int tda) { - - NR_tda_info_t tda_info = {0}; - NR_PUSCH_TimeDomainResourceAllocationList_t *tdaList = ul_bwp->tdaList; - AssertFatal(tda < tdaList->list.count, "time_domain_allocation %d>=%d\n", tda, tdaList->list.count); - tda_info.mapping_type = tdaList->list.array[tda]->mappingType; - const int startSymbolAndLength = tdaList->list.array[tda]->startSymbolAndLength; - SLIV2SL(startSymbolAndLength, &tda_info.startSymbolIndex, &tda_info.nrOfSymbols); - return tda_info; -} - NR_pusch_dmrs_t get_ul_dmrs_params(const NR_ServingCellConfigCommon_t *scc, const NR_UE_UL_BWP_t *ul_bwp, const NR_tda_info_t *tda_info, @@ -1078,12 +1054,8 @@ void set_r_pucch_parms(int rsetindex, *start_symbol_index = default_pucch_firstsymb[rsetindex]; } -void prepare_dci(const NR_CellGroupConfig_t *CellGroup, - const NR_UE_DL_BWP_t *current_BWP, - const NR_ControlResourceSet_t *coreset, - dci_pdu_rel15_t *dci_pdu_rel15, - nr_dci_format_t format) { - +void prepare_dci(const NR_CellGroupConfig_t *CellGroup, const NR_UE_DL_BWP_t *current_BWP, const NR_ControlResourceSet_t *coreset, dci_pdu_rel15_t *dci_pdu_rel15, nr_dci_format_t format) +{ AssertFatal(CellGroup!=NULL,"CellGroup shouldn't be null here\n"); const NR_PDSCH_Config_t *pdsch_Config = current_BWP ? current_BWP->pdsch_Config : NULL; @@ -1157,7 +1129,6 @@ void prepare_dci(const NR_CellGroupConfig_t *CellGroup, } } - void fill_dci_pdu_rel15(const NR_ServingCellConfigCommon_t *scc, const NR_CellGroupConfig_t *CellGroup, const NR_UE_DL_BWP_t *current_DL_BWP, @@ -1177,7 +1148,6 @@ void fill_dci_pdu_rel15(const NR_ServingCellConfigCommon_t *scc, *dci_pdu = 0; uint16_t alt_size = 0; uint16_t N_RB; - const int controlResourceSetId = *ss->controlResourceSetId; if(current_DL_BWP) { N_RB = get_rb_bwp_dci(dci_format, ss->searchSpaceType->present, @@ -1190,28 +1160,16 @@ void fill_dci_pdu_rel15(const NR_ServingCellConfigCommon_t *scc, // computing alternative size for padding dci_pdu_rel15_t temp_pdu; if(dci_format == NR_DL_DCI_FORMAT_1_0) - alt_size = nr_dci_size(scc->downlinkConfigCommon->initialDownlinkBWP, - scc->uplinkConfigCommon->initialUplinkBWP, - current_DL_BWP, current_UL_BWP, - CellGroup, &temp_pdu, NR_UL_DCI_FORMAT_0_0, rnti_type, - controlResourceSetId, bwp_id, ss->searchSpaceType->present, cset0_bwp_size, 0); + alt_size = nr_dci_size(current_DL_BWP, current_UL_BWP, CellGroup, &temp_pdu, NR_UL_DCI_FORMAT_0_0, rnti_type, coreset, bwp_id, ss->searchSpaceType->present, cset0_bwp_size, 0); if(dci_format == NR_UL_DCI_FORMAT_0_0) - alt_size = nr_dci_size(scc->downlinkConfigCommon->initialDownlinkBWP, - scc->uplinkConfigCommon->initialUplinkBWP, - current_DL_BWP, current_UL_BWP, - CellGroup, &temp_pdu, NR_DL_DCI_FORMAT_1_0, rnti_type, - controlResourceSetId, bwp_id, ss->searchSpaceType->present, cset0_bwp_size, 0); + alt_size = nr_dci_size(current_DL_BWP, current_UL_BWP, CellGroup, &temp_pdu, NR_DL_DCI_FORMAT_1_0, rnti_type, coreset, bwp_id, ss->searchSpaceType->present, cset0_bwp_size, 0); } else N_RB = cset0_bwp_size; - int dci_size = nr_dci_size(scc->downlinkConfigCommon->initialDownlinkBWP, - scc->uplinkConfigCommon->initialUplinkBWP, - current_DL_BWP, current_UL_BWP, - CellGroup, dci_pdu_rel15, dci_format, rnti_type, controlResourceSetId, - bwp_id, ss->searchSpaceType->present, cset0_bwp_size, alt_size); + int dci_size = nr_dci_size(current_DL_BWP, current_UL_BWP, CellGroup, dci_pdu_rel15, dci_format, rnti_type, coreset, bwp_id, ss->searchSpaceType->present, cset0_bwp_size, alt_size); pdcch_dci_pdu->PayloadSizeBits = dci_size; AssertFatal(dci_size <= 64, "DCI sizes above 64 bits not yet supported"); if (dci_format == NR_DL_DCI_FORMAT_1_1 || dci_format == NR_UL_DCI_FORMAT_0_1) @@ -2087,13 +2045,17 @@ void configure_UE_BWP(gNB_MAC_INST *nr_mac, NR_BWP_Uplink_t *ul_bwp = NULL; NR_BWP_DownlinkDedicated_t *bwpd = NULL; NR_BWP_UplinkDedicated_t *ubwpd = NULL; - DL_BWP->n_dl_bwp = 1; - UL_BWP->n_ul_bwp = 1; + // number of additional BWPs (excluding initial BWP) + DL_BWP->n_dl_bwp = 0; + UL_BWP->n_ul_bwp = 0; int old_dl_bwp_id = DL_BWP->bwp_id; int old_ul_bwp_id = UL_BWP->bwp_id; int target_ss; + if (CellGroup && CellGroup->physicalCellGroupConfig) + DL_BWP->pdsch_HARQ_ACK_Codebook = &CellGroup->physicalCellGroupConfig->pdsch_HARQ_ACK_Codebook; + if (CellGroup && CellGroup->spCellConfig && CellGroup->spCellConfig->spCellConfigDedicated) { @@ -2102,6 +2064,7 @@ void configure_UE_BWP(gNB_MAC_INST *nr_mac, DL_BWP->pdsch_servingcellconfig = servingCellConfig->pdsch_ServingCellConfig ? servingCellConfig->pdsch_ServingCellConfig->choice.setup : NULL; UL_BWP->pusch_servingcellconfig = servingCellConfig->uplinkConfig && servingCellConfig->uplinkConfig->pusch_ServingCellConfig ? servingCellConfig->uplinkConfig->pusch_ServingCellConfig->choice.setup : NULL; + target_ss = NR_SearchSpace__searchSpaceType_PR_ue_Specific; if(UE && UE->Msg3_dcch_dtch) { @@ -2167,6 +2130,7 @@ void configure_UE_BWP(gNB_MAC_INST *nr_mac, ubwpd = servingCellConfig->uplinkConfig->initialUplinkBWP; DL_BWP->pdsch_Config = bwpd->pdsch_Config->choice.setup; + UL_BWP->configuredGrantConfig = ubwpd->configuredGrantConfig ? ubwpd->configuredGrantConfig->choice.setup : NULL; UL_BWP->pusch_Config = ubwpd->pusch_Config->choice.setup; UL_BWP->pucch_Config = ubwpd->pucch_Config->choice.setup; UL_BWP->srs_Config = ubwpd->srs_Config->choice.setup; @@ -2180,6 +2144,7 @@ void configure_UE_BWP(gNB_MAC_INST *nr_mac, UL_BWP->pusch_Config = NULL; UL_BWP->pucch_Config = NULL; UL_BWP->csi_MeasConfig = NULL; + UL_BWP->configuredGrantConfig = NULL; } if (old_dl_bwp_id != DL_BWP->bwp_id) @@ -2189,14 +2154,14 @@ void configure_UE_BWP(gNB_MAC_INST *nr_mac, // TDA lists if (DL_BWP->bwp_id>0) - DL_BWP->tdaList = dl_bwp->bwp_Common->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList; + DL_BWP->tdaList_Common = dl_bwp->bwp_Common->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList; else - DL_BWP->tdaList = scc->downlinkConfigCommon->initialDownlinkBWP->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList; + DL_BWP->tdaList_Common = scc->downlinkConfigCommon->initialDownlinkBWP->pdsch_ConfigCommon->choice.setup->pdsch_TimeDomainAllocationList; if(UL_BWP->bwp_id>0) - UL_BWP->tdaList = ul_bwp->bwp_Common->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList; + UL_BWP->tdaList_Common = ul_bwp->bwp_Common->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList; else - UL_BWP->tdaList = scc->uplinkConfigCommon->initialUplinkBWP->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList; + UL_BWP->tdaList_Common = scc->uplinkConfigCommon->initialUplinkBWP->pusch_ConfigCommon->choice.setup->pusch_TimeDomainAllocationList; // setting generic parameters NR_BWP_t dl_genericParameters = (DL_BWP->bwp_id > 0 && dl_bwp) ? @@ -2226,10 +2191,13 @@ void configure_UE_BWP(gNB_MAC_INST *nr_mac, else UL_BWP->transform_precoding = *UL_BWP->pusch_Config->transformPrecoder; - if(UL_BWP->bwp_id>0) + if (UL_BWP->bwp_id > 0) { UL_BWP->pucch_ConfigCommon = ul_bwp->bwp_Common->pucch_ConfigCommon->choice.setup; - else + UL_BWP->rach_ConfigCommon = ul_bwp->bwp_Common->rach_ConfigCommon->choice.setup; + } else { UL_BWP->pucch_ConfigCommon = scc->uplinkConfigCommon->initialUplinkBWP->pucch_ConfigCommon->choice.setup; + UL_BWP->rach_ConfigCommon = scc->uplinkConfigCommon->initialUplinkBWP->rach_ConfigCommon->choice.setup; + } if(UE) { @@ -2893,20 +2861,33 @@ void schedule_nr_bwp_switch(module_id_t module_id, } } -void UL_tti_req_ahead_initialization(gNB_MAC_INST * gNB, NR_ServingCellConfigCommon_t *scc, int n, int CCid, frame_t frameP) { +int ul_buffer_index(int frame, int slot, int scs, int size) +{ + const int abs_slot = frame * nr_slots_per_frame[scs] + slot; + return abs_slot % size; +} - if(gNB->UL_tti_req_ahead[CCid]) return; +void UL_tti_req_ahead_initialization(gNB_MAC_INST * gNB, NR_ServingCellConfigCommon_t *scc, int n, int CCid, frame_t frameP, int slotP, int scs) +{ - gNB->UL_tti_req_ahead[CCid] = calloc(n, sizeof(nfapi_nr_ul_tti_request_t)); - AssertFatal(gNB->UL_tti_req_ahead[CCid], - "could not allocate memory for RC.nrmac[]->UL_tti_req_ahead[]\n"); + if(gNB->UL_tti_req_ahead[CCid]) + return; + + int size = n; + if (scs == 0) + size <<= 1; // to have enough room for feedback possibly beyond the frame we need a larger array at 15kHz SCS + + gNB->UL_tti_req_ahead_size = size; + gNB->UL_tti_req_ahead[CCid] = calloc(size, sizeof(nfapi_nr_ul_tti_request_t)); + AssertFatal(gNB->UL_tti_req_ahead[CCid], "could not allocate memory for RC.nrmac[]->UL_tti_req_ahead[]\n"); /* fill in slot/frame numbers: slot is fixed, frame will be updated by scheduler * consider that scheduler runs sl_ahead: the first sl_ahead slots are * already "in the past" and thus we put frame 1 instead of 0! */ - for (int i = 0; i < n; ++i) { - nfapi_nr_ul_tti_request_t *req = &gNB->UL_tti_req_ahead[CCid][i]; - req->SFN = frameP + (i < (gNB->if_inst->sl_ahead-1)); - req->Slot = i; + for (int i = 0; i < size; ++i) { + int abs_slot = frameP * n + slotP + i; + nfapi_nr_ul_tti_request_t *req = &gNB->UL_tti_req_ahead[CCid][abs_slot % size]; + req->SFN = abs_slot / n; + req->Slot = abs_slot % n; } } diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_srs.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_srs.c index ef98c82eab..8a85448c7f 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_srs.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_srs.c @@ -137,7 +137,7 @@ void nr_srs_ri_computation(const nfapi_nr_srs_normalized_channel_iq_matrix_t *nr } -void nr_configure_srs(nfapi_nr_srs_pdu_t *srs_pdu, int slot, int module_id, int CC_id, NR_UE_info_t *UE, NR_SRS_ResourceSet_t *srs_resource_set, NR_SRS_Resource_t *srs_resource) +void nr_configure_srs(nfapi_nr_srs_pdu_t *srs_pdu, int slot, int module_id, int CC_id, NR_UE_info_t *UE, NR_SRS_ResourceSet_t *srs_resource_set, NR_SRS_Resource_t *srs_resource, int buffer_index) { NR_UE_UL_BWP_t *current_BWP = &UE->current_UL_BWP; @@ -190,15 +190,17 @@ void nr_configure_srs(nfapi_nr_srs_pdu_t *srs_pdu, int slot, int module_id, int srs_pdu->beamforming.prg_size = 1; } - uint16_t *vrb_map_UL = &RC.nrmac[module_id]->common_channels[CC_id].vrb_map_UL[slot * MAX_BWP_SIZE]; + uint16_t *vrb_map_UL = &RC.nrmac[module_id]->common_channels[CC_id].vrb_map_UL[buffer_index * MAX_BWP_SIZE]; uint64_t mask = SL_to_bitmap(13 - srs_pdu->time_start_position, srs_pdu->num_symbols); for (int i = 0; i < srs_pdu->bwp_size; ++i) vrb_map_UL[i + srs_pdu->bwp_start] |= mask; } -void nr_fill_nfapi_srs(int module_id, int CC_id, NR_UE_info_t* UE, sub_frame_t slot, NR_SRS_ResourceSet_t *srs_resource_set, NR_SRS_Resource_t *srs_resource) { +void nr_fill_nfapi_srs(int module_id, int CC_id, NR_UE_info_t* UE, int frame, int slot, NR_SRS_ResourceSet_t *srs_resource_set, NR_SRS_Resource_t *srs_resource) +{ - nfapi_nr_ul_tti_request_t *future_ul_tti_req = &RC.nrmac[module_id]->UL_tti_req_ahead[0][slot]; + int index = ul_buffer_index(frame, slot, UE->current_UL_BWP.scs, RC.nrmac[module_id]->UL_tti_req_ahead_size); + nfapi_nr_ul_tti_request_t *future_ul_tti_req = &RC.nrmac[module_id]->UL_tti_req_ahead[0][index]; AssertFatal(future_ul_tti_req->n_pdus < sizeof(future_ul_tti_req->pdus_list) / sizeof(future_ul_tti_req->pdus_list[0]), "Invalid future_ul_tti_req->n_pdus %d\n", future_ul_tti_req->n_pdus); @@ -207,8 +209,8 @@ void nr_fill_nfapi_srs(int module_id, int CC_id, NR_UE_info_t* UE, sub_frame_t s nfapi_nr_srs_pdu_t *srs_pdu = &future_ul_tti_req->pdus_list[future_ul_tti_req->n_pdus].srs_pdu; memset(srs_pdu, 0, sizeof(nfapi_nr_srs_pdu_t)); future_ul_tti_req->n_pdus += 1; - - nr_configure_srs(srs_pdu, slot, module_id, CC_id, UE, srs_resource_set, srs_resource); + index = ul_buffer_index(frame, slot, UE->current_UL_BWP.scs, RC.nrmac[module_id]->vrb_map_UL_size); + nr_configure_srs(srs_pdu, slot, module_id, CC_id, UE, srs_resource_set, srs_resource, index); } /******************************************************************* @@ -274,11 +276,12 @@ void nr_schedule_srs(int module_id, frame_t frame, int slot) continue; } - const int num_tda = current_BWP->tdaList->list.count; + NR_PUSCH_TimeDomainResourceAllocationList_t *tdaList = get_ul_tdalist(current_BWP, sched_ctrl->coreset->controlResourceSetId, sched_ctrl->search_space->searchSpaceType->present, NR_RNTI_C); + const int num_tda = tdaList->list.count; int max_k2 = 0; // avoid last one in the list (for msg3) for (int i = 0; i < num_tda - 1; i++) { - int k2 = get_K2(current_BWP->tdaList, i, current_BWP->scs); + int k2 = get_K2(tdaList, i, current_BWP->scs); max_k2 = k2 > max_k2 ? k2 : max_k2; } @@ -293,7 +296,7 @@ void nr_schedule_srs(int module_id, frame_t frame, int slot) // Check if UE will transmit the SRS in this frame if ((sched_frame * n_slots_frame + sched_slot - offset) % period == 0) { LOG_D(NR_MAC," %d.%d Scheduling SRS reception for %d.%d\n", frame, slot, sched_frame, sched_slot); - nr_fill_nfapi_srs(module_id, CC_id, UE, sched_slot, srs_resource_set, srs_resource); + nr_fill_nfapi_srs(module_id, CC_id, UE, sched_frame, sched_slot, srs_resource_set, srs_resource); sched_ctrl->sched_srs.frame = sched_frame; sched_ctrl->sched_srs.slot = sched_slot; sched_ctrl->sched_srs.srs_scheduled = true; diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_uci.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_uci.c index 5caea0c2c7..1a5561e05b 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_uci.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_uci.c @@ -45,7 +45,8 @@ static void nr_fill_nfapi_pucch(gNB_MAC_INST *nrmac, NR_UE_info_t* UE) { - nfapi_nr_ul_tti_request_t *future_ul_tti_req = &nrmac->UL_tti_req_ahead[0][pucch->ul_slot]; + const int index = ul_buffer_index(pucch->frame, pucch->ul_slot, UE->current_UL_BWP.scs, nrmac->UL_tti_req_ahead_size); + nfapi_nr_ul_tti_request_t *future_ul_tti_req = &nrmac->UL_tti_req_ahead[0][index]; if (future_ul_tti_req->SFN != pucch->frame || future_ul_tti_req->Slot != pucch->ul_slot) LOG_W(MAC, "Current %d.%d : future UL_tti_req's frame.slot %4d.%2d does not match PUCCH %4d.%2d\n", @@ -182,9 +183,10 @@ void nr_schedule_pucch(gNB_MAC_INST *nrmac, void nr_csi_meas_reporting(int Mod_idP, frame_t frame, - sub_frame_t slot) { - - UE_iterator(RC.nrmac[Mod_idP]->UE_info.list, UE ) { + sub_frame_t slot) +{ + gNB_MAC_INST *nrmac = RC.nrmac[Mod_idP]; + UE_iterator(nrmac->UE_info.list, UE ) { NR_UE_sched_ctrl_t *sched_ctrl = &UE->UE_sched_ctrl; NR_UE_UL_BWP_t *ul_bwp = &UE->current_UL_BWP; const int n_slots_frame = nr_slots_per_frame[ul_bwp->scs]; @@ -216,7 +218,7 @@ void nr_csi_meas_reporting(int Mod_idP, if ((sched_frame*n_slots_frame + sched_slot - offset)%period != 0) continue; - AssertFatal(is_xlsch_in_slot(RC.nrmac[Mod_idP]->ulsch_slot_bitmap[sched_slot / 64], sched_slot), "CSI reporting slot %d is not set for an uplink slot\n", sched_slot); + AssertFatal(is_xlsch_in_slot(nrmac->ulsch_slot_bitmap[sched_slot / 64], sched_slot), "CSI reporting slot %d is not set for an uplink slot\n", sched_slot); LOG_D(NR_MAC, "CSI reporting in frame %d slot %d CSI report ID %ld\n", sched_frame, sched_slot, csirep->reportConfigId); const NR_PUCCH_ResourceSet_t *pucchresset = pucch_Config->resourceSetToAddModList->list.array[1]; // set with formats >1 @@ -228,9 +230,9 @@ void nr_csi_meas_reporting(int Mod_idP, AssertFatal(res_index < n, "CSI pucch resource %ld not found among PUCCH resources\n", pucchcsires->pucch_Resource); - const NR_ServingCellConfigCommon_t *scc = RC.nrmac[Mod_idP]->common_channels[0].ServingCellConfigCommon; + const NR_ServingCellConfigCommon_t *scc = nrmac->common_channels[0].ServingCellConfigCommon; const NR_TDD_UL_DL_Pattern_t *tdd = scc->tdd_UL_DL_ConfigurationCommon ? &scc->tdd_UL_DL_ConfigurationCommon->pattern1 : NULL; - AssertFatal(tdd || RC.nrmac[Mod_idP]->common_channels[0].frame_type == FDD, "Dynamic TDD not handled yet\n"); + AssertFatal(tdd || nrmac->common_channels[0].frame_type == FDD, "Dynamic TDD not handled yet\n"); const int pucch_index = get_pucch_index(sched_frame, sched_slot, n_slots_frame, tdd, sched_ctrl->sched_pucch_size); NR_sched_pucch_t *curr_pucch = &sched_ctrl->sched_pucch[pucch_index]; AssertFatal(curr_pucch->active == false, "CSI structure is scheduled in advance. It should be free!\n"); @@ -244,7 +246,8 @@ void nr_csi_meas_reporting(int Mod_idP, int bwp_start = ul_bwp->BWPStart; // going through the list of PUCCH resources to find the one indexed by resource_id - uint16_t *vrb_map_UL = &RC.nrmac[Mod_idP]->common_channels[0].vrb_map_UL[sched_slot * MAX_BWP_SIZE]; + const int index = ul_buffer_index(sched_frame, sched_slot, ul_bwp->scs, nrmac->vrb_map_UL_size); + uint16_t *vrb_map_UL = &nrmac->common_channels[0].vrb_map_UL[index * MAX_BWP_SIZE]; const int m = pucch_Config->resourceToAddModList->list.count; for (int j = 0; j < m; j++) { NR_PUCCH_Resource_t *pucchres = pucch_Config->resourceToAddModList->list.array[j]; @@ -1015,14 +1018,10 @@ void handle_nr_uci_pucch_2_3_4(module_id_t mod_id, } } -void set_pucch_allocation(const NR_PUCCH_Config_t *pucch_Config, - const NR_ServingCellConfigCommon_t *scc, - const int r_pucch, - const int bwp_size, - NR_sched_pucch_t *pucch) +void set_pucch_allocation(const NR_UE_UL_BWP_t *ul_bwp, const int r_pucch, const int bwp_size, NR_sched_pucch_t *pucch) { if(r_pucch<0){ - const NR_PUCCH_Resource_t *resource = pucch_Config->resourceToAddModList->list.array[0]; + const NR_PUCCH_Resource_t *resource = ul_bwp->pucch_Config->resourceToAddModList->list.array[0]; DevAssert(resource->format.present == NR_PUCCH_Resource__format_PR_format0); pucch->second_hop_prb = resource->secondHopPRB!= NULL ? *resource->secondHopPRB : 0; pucch->nr_of_symb = resource->format.choice.format0->nrofSymbols; @@ -1030,7 +1029,7 @@ void set_pucch_allocation(const NR_PUCCH_Config_t *pucch_Config, pucch->prb_start = resource->startingPRB; } else{ - int rsetindex = *scc->uplinkConfigCommon->initialUplinkBWP->pucch_ConfigCommon->choice.setup->pucch_ResourceCommon; + int rsetindex = *ul_bwp->pucch_ConfigCommon->pucch_ResourceCommon; set_r_pucch_parms(rsetindex, r_pucch, bwp_size, @@ -1155,15 +1154,16 @@ int nr_acknack_scheduling(gNB_MAC_INST *mac, } else { // unoccupied occasion // checking if in ul_slot the resources potentially to be assigned to this PUCCH are available - set_pucch_allocation(pucch_Config, scc, r_pucch, bwp_size, curr_pucch); - uint16_t *vrb_map_UL = &mac->common_channels[CC_id].vrb_map_UL[pucch_slot * MAX_BWP_SIZE]; + set_pucch_allocation(ul_bwp, r_pucch, bwp_size, curr_pucch); + const int index = ul_buffer_index(pucch_frame, pucch_slot, ul_bwp->scs, mac->vrb_map_UL_size); + uint16_t *vrb_map_UL = &mac->common_channels[CC_id].vrb_map_UL[index * MAX_BWP_SIZE]; bool ret = test_pucch0_vrb_occupation(curr_pucch, vrb_map_UL, bwp_start, bwp_size); if(!ret) { LOG_D(NR_MAC, "DL %4d.%2d, UL_ACK %4d.%2d PRB resources for this occasion are already occupied, move to the following occasion\n", - frame,slot,curr_pucch->frame,curr_pucch->ul_slot); + frame, slot, pucch_frame, pucch_slot); continue; } // allocating a new PUCCH structure for this occasion @@ -1176,7 +1176,7 @@ int nr_acknack_scheduling(gNB_MAC_INST *mac, curr_pucch->r_pucch=r_pucch; LOG_D(NR_MAC, "DL %4d.%2d, UL_ACK %4d.%2d Scheduling ACK/NACK in PUCCH %d with timing indicator %d DAI %d\n", - frame,slot,curr_pucch->frame,curr_pucch->ul_slot,pucch_index,f,curr_pucch->dai_c); + frame, slot, curr_pucch->frame, curr_pucch->ul_slot, pucch_index, f, curr_pucch->dai_c); // blocking resources for current PUCCH in VRB map set_pucch0_vrb_occupation(curr_pucch, vrb_map_UL, bwp_start); @@ -1246,10 +1246,11 @@ void nr_sr_reporting(gNB_MAC_INST *nrmac, frame_t SFN, sub_frame_t slot) continue; } else { - uint16_t *vrb_map_UL = &nrmac->common_channels[CC_id].vrb_map_UL[slot * MAX_BWP_SIZE]; + const int index = ul_buffer_index(SFN, slot, ul_bwp->scs, nrmac->vrb_map_UL_size); + uint16_t *vrb_map_UL = &nrmac->common_channels[CC_id].vrb_map_UL[index * MAX_BWP_SIZE]; const int bwp_start = ul_bwp->BWPStart; const int bwp_size = ul_bwp->BWPSize; - set_pucch_allocation(pucch_Config, scc, -1, bwp_size, curr_pucch); + set_pucch_allocation(ul_bwp, -1, bwp_size, curr_pucch); bool ret = test_pucch0_vrb_occupation(curr_pucch, vrb_map_UL, bwp_start, diff --git a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_ulsch.c b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_ulsch.c index a3c5a2e95c..25089b2edf 100644 --- a/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_ulsch.c +++ b/openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_ulsch.c @@ -1424,13 +1424,14 @@ static bool allocate_ul_retransmission(gNB_MAC_INST *nrmac, const int CC_id = 0; NR_UE_sched_ctrl_t *sched_ctrl = &UE->UE_sched_ctrl; NR_sched_pusch_t *retInfo = &sched_ctrl->ul_harq_processes[harq_pid].sched_pusch; + NR_UE_UL_BWP_t *ul_bwp = &UE->current_UL_BWP; int rbStart = 0; // wrt BWP start - const uint16_t bwpSize = UE->current_UL_BWP.BWPSize; + const uint16_t bwpSize = ul_bwp->BWPSize; const uint8_t nrOfLayers = retInfo->nrOfLayers; LOG_D(NR_MAC,"retInfo->time_domain_allocation = %d, tda = %d\n", retInfo->time_domain_allocation, tda); LOG_D(NR_MAC,"tbs %d\n",retInfo->tb_size); - NR_tda_info_t tda_info = nr_get_pusch_tda_info(&UE->current_UL_BWP, tda); + NR_tda_info_t tda_info = get_ul_tda_info(ul_bwp, sched_ctrl->coreset->controlResourceSetId, sched_ctrl->search_space->searchSpaceType->present, NR_RNTI_C, tda); bool reuse_old_tda = (retInfo->tda_info.startSymbolIndex == tda_info.startSymbolIndex) && (retInfo->tda_info.nrOfSymbols <= tda_info.nrOfSymbols); if (reuse_old_tda && nrOfLayers == retInfo->nrOfLayers) { /* Check the resource is enough for retransmission */ @@ -1450,7 +1451,7 @@ static bool allocate_ul_retransmission(gNB_MAC_INST *nrmac, LOG_D(NR_MAC, "%s(): retransmission keeping TDA %d and TBS %d\n", __func__, tda, retInfo->tb_size); } else { NR_pusch_dmrs_t dmrs_info = get_ul_dmrs_params(scc, - &UE->current_UL_BWP, + ul_bwp, &tda_info, nrOfLayers); /* the retransmission will use a different time domain allocation, check @@ -1666,7 +1667,7 @@ void pf_ul(module_id_t module_id, sched_pusch->nrOfLayers = sched_ctrl->srs_feedback.ul_ri + 1; sched_pusch->time_domain_allocation = get_ul_tda(nrmac, scc, sched_pusch->frame, sched_pusch->slot); - sched_pusch->tda_info = nr_get_pusch_tda_info(current_BWP, sched_pusch->time_domain_allocation); + sched_pusch->tda_info = get_ul_tda_info(current_BWP, sched_ctrl->coreset->controlResourceSetId, sched_ctrl->search_space->searchSpaceType->present, NR_RNTI_C, sched_pusch->time_domain_allocation); sched_pusch->dmrs_info = get_ul_dmrs_params(scc, current_BWP, &sched_pusch->tda_info, @@ -1766,7 +1767,7 @@ void pf_ul(module_id_t module_id, sched_pusch->nrOfLayers = sched_ctrl->srs_feedback.ul_ri + 1; sched_pusch->time_domain_allocation = get_ul_tda(nrmac, scc, sched_pusch->frame, sched_pusch->slot); - sched_pusch->tda_info = nr_get_pusch_tda_info(current_BWP, sched_pusch->time_domain_allocation); + sched_pusch->tda_info = get_ul_tda_info(current_BWP, sched_ctrl->coreset->controlResourceSetId, sched_ctrl->search_space->searchSpaceType->present, NR_RNTI_C, sched_pusch->time_domain_allocation); sched_pusch->dmrs_info = get_ul_dmrs_params(scc, current_BWP, &sched_pusch->tda_info, @@ -1868,13 +1869,14 @@ bool nr_fr1_ulsch_preprocessor(module_id_t module_id, frame_t frame, sub_frame_t NR_UE_UL_BWP_t *current_BWP = &nr_mac->UE_info.list[0]->current_UL_BWP; int mu = current_BWP->scs; const int temp_tda = get_ul_tda(nr_mac, scc, frame, slot); - int K2 = get_K2(current_BWP->tdaList, temp_tda, mu); + NR_PUSCH_TimeDomainResourceAllocationList_t *tdaList = get_ul_tdalist(current_BWP, sched_ctrl->coreset->controlResourceSetId, sched_ctrl->search_space->searchSpaceType->present, NR_RNTI_C); + int K2 = get_K2(tdaList, temp_tda, mu); const int sched_frame = (frame + (slot + K2 >= nr_slots_per_frame[mu])) & 1023; const int sched_slot = (slot + K2) % nr_slots_per_frame[mu]; const int tda = get_ul_tda(nr_mac, scc, sched_frame, sched_slot); if (tda < 0) return false; - DevAssert(K2 == get_K2(current_BWP->tdaList, tda, mu)); + DevAssert(K2 == get_K2(tdaList, tda, mu)); if (!is_xlsch_in_slot(nr_mac->ulsch_slot_bitmap[sched_slot / 64], sched_slot)) return false; @@ -1883,9 +1885,9 @@ bool nr_fr1_ulsch_preprocessor(module_id_t module_id, frame_t frame, sub_frame_t sched_ctrl->sched_pusch.frame = sched_frame; UE_iterator(nr_mac->UE_info.list, UE2) { NR_UE_sched_ctrl_t *sched_ctrl = &UE2->UE_sched_ctrl; - AssertFatal(K2 == get_K2(current_BWP->tdaList, tda, mu), + AssertFatal(K2 == get_K2(tdaList, tda, mu), "Different K2, %d(UE%d) != %ld(UE%04x)\n", - K2, 0, get_K2(current_BWP->tdaList, tda, mu), UE2->rnti); + K2, 0, get_K2(tdaList, tda, mu), UE2->rnti); sched_ctrl->sched_pusch.slot = sched_slot; sched_ctrl->sched_pusch.frame = sched_frame; } @@ -1893,13 +1895,13 @@ bool nr_fr1_ulsch_preprocessor(module_id_t module_id, frame_t frame, sub_frame_t /* Change vrb_map_UL to rballoc_mask: check which symbols per RB (in * vrb_map_UL) overlap with the "default" tda and exclude those RBs. * Calculate largest contiguous RBs */ - uint16_t *vrb_map_UL = - &RC.nrmac[module_id]->common_channels[CC_id].vrb_map_UL[sched_slot * MAX_BWP_SIZE]; + const int index = ul_buffer_index(sched_frame, sched_slot, mu, nr_mac->vrb_map_UL_size); + uint16_t *vrb_map_UL = &nr_mac->common_channels[CC_id].vrb_map_UL[index * MAX_BWP_SIZE]; const uint16_t bwpSize = current_BWP->BWPSize; const uint16_t bwpStart = current_BWP->BWPStart; - const int startSymbolAndLength = current_BWP->tdaList->list.array[tda]->startSymbolAndLength; + const int startSymbolAndLength = tdaList->list.array[tda]->startSymbolAndLength; int startSymbolIndex, nrOfSymbols; SLIV2SL(startSymbolAndLength, &startSymbolIndex, &nrOfSymbols); const uint16_t symb = SL_to_bitmap(startSymbolIndex, nrOfSymbols); @@ -1981,12 +1983,12 @@ void nr_schedule_ulsch(module_id_t module_id, frame_t frame, sub_frame_t slot) LOG_D(NR_MAC, "Current slot %d is NOT DL slot, cannot schedule DCI0 for UL data\n", slot); return; } - bool do_sched = RC.nrmac[module_id]->pre_processor_ul(module_id, frame, slot); + bool do_sched = nr_mac->pre_processor_ul(module_id, frame, slot); if (!do_sched) return; const int CC_id = 0; - nfapi_nr_ul_dci_request_t *ul_dci_req = &RC.nrmac[module_id]->UL_dci_req[CC_id]; + nfapi_nr_ul_dci_request_t *ul_dci_req = &nr_mac->UL_dci_req[CC_id]; ul_dci_req->SFN = frame; ul_dci_req->Slot = slot; /* a PDCCH PDU groups DCIs per BWP and CORESET. Save a pointer to each @@ -1995,9 +1997,9 @@ void nr_schedule_ulsch(module_id_t module_id, frame_t frame, sub_frame_t slot) nfapi_nr_dl_tti_pdcch_pdu_rel15_t *pdcch_pdu_coreset[MAX_NUM_CORESET] = {0}; - NR_ServingCellConfigCommon_t *scc = RC.nrmac[module_id]->common_channels[0].ServingCellConfigCommon; - NR_UEs_t *UE_info = &RC.nrmac[module_id]->UE_info; - const NR_SIB1_t *sib1 = RC.nrmac[module_id]->common_channels[0].sib1 ? RC.nrmac[module_id]->common_channels[0].sib1->message.choice.c1->choice.systemInformationBlockType1 : NULL; + NR_ServingCellConfigCommon_t *scc = nr_mac->common_channels[0].ServingCellConfigCommon; + NR_UEs_t *UE_info = &nr_mac->UE_info; + const NR_SIB1_t *sib1 = nr_mac->common_channels[0].sib1 ? nr_mac->common_channels[0].sib1->message.choice.c1->choice.systemInformationBlockType1 : NULL; UE_iterator( UE_info->list, UE) { NR_UE_sched_ctrl_t *sched_ctrl = &UE->UE_sched_ctrl; if (sched_ctrl->ul_failure == 1 && get_softmodem_params()->phy_test==0) continue; @@ -2104,7 +2106,8 @@ void nr_schedule_ulsch(module_id_t module_id, frame_t frame, sub_frame_t slot) sched_ctrl->tpc0); /* PUSCH in a later slot, but corresponding DCI now! */ - nfapi_nr_ul_tti_request_t *future_ul_tti_req = &RC.nrmac[module_id]->UL_tti_req_ahead[0][sched_pusch->slot]; + const int index = ul_buffer_index(sched_pusch->frame, sched_pusch->slot, current_BWP->scs, nr_mac->UL_tti_req_ahead_size); + nfapi_nr_ul_tti_request_t *future_ul_tti_req = &nr_mac->UL_tti_req_ahead[0][index]; if (future_ul_tti_req->SFN != sched_pusch->frame || future_ul_tti_req->Slot != sched_pusch->slot) LOG_W(MAC, "%d.%d future UL_tti_req's frame.slot %d.%d does not match PUSCH %d.%d\n", diff --git a/openair2/LAYER2/NR_MAC_gNB/mac_proto.h b/openair2/LAYER2/NR_MAC_gNB/mac_proto.h index 7e9a66ca7f..ff74e34c4a 100644 --- a/openair2/LAYER2/NR_MAC_gNB/mac_proto.h +++ b/openair2/LAYER2/NR_MAC_gNB/mac_proto.h @@ -277,11 +277,7 @@ void fill_dci_pdu_rel15(const NR_ServingCellConfigCommon_t *scc, NR_ControlResourceSet_t *coreset, uint16_t cset0_bwp_size); -void prepare_dci(const NR_CellGroupConfig_t *CellGroup, - const NR_UE_DL_BWP_t *dl_bwp, - const NR_ControlResourceSet_t *coreset, - dci_pdu_rel15_t *dci_pdu_rel15, - nr_dci_format_t format); +void prepare_dci(const NR_CellGroupConfig_t *CellGroup, const NR_UE_DL_BWP_t *current_BWP, const NR_ControlResourceSet_t *coreset, dci_pdu_rel15_t *dci_pdu_rel15, nr_dci_format_t format); void set_r_pucch_parms(int rsetindex, int r_pucch, @@ -307,12 +303,6 @@ long get_K2(NR_PUSCH_TimeDomainResourceAllocationList_t *tdaList, int time_domain_assignment, int mu); -NR_tda_info_t nr_get_pdsch_tda_info(const NR_UE_DL_BWP_t *dl_bwp, - const int tda); - -NR_tda_info_t nr_get_pusch_tda_info(const NR_UE_UL_BWP_t *ul_bwp, - const int tda); - NR_pusch_dmrs_t get_ul_dmrs_params(const NR_ServingCellConfigCommon_t *scc, const NR_UE_UL_BWP_t *ul_bwp, const NR_tda_info_t *tda_info, @@ -482,7 +472,9 @@ int get_mcs_from_bler(const NR_bler_options_t *bler_options, int max_mcs, frame_t frame); -void UL_tti_req_ahead_initialization(gNB_MAC_INST * gNB, NR_ServingCellConfigCommon_t *scc, int n, int CCid, frame_t frameP); +int ul_buffer_index(int frame, int slot, int scs, int size); + +void UL_tti_req_ahead_initialization(gNB_MAC_INST * gNB, NR_ServingCellConfigCommon_t *scc, int n, int CCid, frame_t frameP, int slotP, int scs); void nr_sr_reporting(gNB_MAC_INST *nrmac, frame_t frameP, sub_frame_t slotP); diff --git a/openair2/LAYER2/NR_MAC_gNB/nr_mac_gNB.h b/openair2/LAYER2/NR_MAC_gNB/nr_mac_gNB.h index 22c1619ea6..b4fa43c40f 100644 --- a/openair2/LAYER2/NR_MAC_gNB/nr_mac_gNB.h +++ b/openair2/LAYER2/NR_MAC_gNB/nr_mac_gNB.h @@ -334,12 +334,6 @@ typedef struct NR_sched_pucch { int start_symb; } NR_sched_pucch_t; -typedef struct NR_tda_info { - mappingType_t mapping_type; - int startSymbolIndex; - int nrOfSymbols; -} NR_tda_info_t; - typedef struct NR_pusch_dmrs { uint8_t N_PRB_DMRS; uint8_t num_dmrs_symb; @@ -739,6 +733,8 @@ typedef struct gNB_MAC_INST_s { /// NFAPI UL TTI Request Structure for future TTIs, dynamically allocated /// because length depends on number of slots nfapi_nr_ul_tti_request_t *UL_tti_req_ahead[NFAPI_CC_MAX]; + int UL_tti_req_ahead_size; + int vrb_map_UL_size; /// NFAPI HI/DCI0 Config Request Structure nfapi_nr_ul_dci_request_t UL_dci_req[NFAPI_CC_MAX]; /// NFAPI DL PDU structure diff --git a/openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.c b/openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.c index eb1de74d20..27adaa39bc 100644 --- a/openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.c +++ b/openair2/NR_UE_PHY_INTERFACE/NR_IF_Module.c @@ -667,6 +667,7 @@ static void fill_dci_from_dl_config(nr_downlink_indication_t*dl_ind, fapi_nr_dl_ if (rel15_dci->dci_length_options[j] == dl_ind->dci_ind->dci_list[k].payloadSize) { dl_ind->dci_ind->dci_list[k].dci_format = rel15_dci->dci_format_options[j]; dl_ind->dci_ind->dci_list[k].ss_type = rel15_dci->dci_type_options[j]; + dl_ind->dci_ind->dci_list[k].coreset_type = rel15_dci->coreset.CoreSetType; LOG_D(NR_PHY, "format assigned dl_ind->dci_ind->dci_list[k].dci_format %d\n", dl_ind->dci_ind->dci_list[k].dci_format); } diff --git a/openair2/RRC/NR/nr_rrc_config.c b/openair2/RRC/NR/nr_rrc_config.c index c0d3b730be..a38f9c3972 100644 --- a/openair2/RRC/NR/nr_rrc_config.c +++ b/openair2/RRC/NR/nr_rrc_config.c @@ -760,15 +760,12 @@ void nr_rrc_config_ul_tda(NR_ServingCellConfigCommon_t *scc, int min_fb_delay){ void set_dl_DataToUL_ACK(NR_PUCCH_Config_t *pucch_Config, int min_feedback_time, NR_SubcarrierSpacing_t subcarrierSpacing) { pucch_Config->dl_DataToUL_ACK = calloc(1,sizeof(*pucch_Config->dl_DataToUL_ACK)); - const int n_slots_frame = slotsperframe[subcarrierSpacing]; long *delay[8]; for (int i = 0; i < 8; i++) { int curr_delay = i + min_feedback_time; - if (curr_delay <= n_slots_frame) { - delay[i] = calloc(1,sizeof(*delay[i])); - *delay[i] = curr_delay; - asn1cSeqAdd(&pucch_Config->dl_DataToUL_ACK->list,delay[i]); - } + delay[i] = calloc(1,sizeof(*delay[i])); + *delay[i] = curr_delay; + asn1cSeqAdd(&pucch_Config->dl_DataToUL_ACK->list,delay[i]); } } diff --git a/openair2/RRC/NR/rrc_gNB.c b/openair2/RRC/NR/rrc_gNB.c index 2deeea6e1f..c686bdb81b 100644 --- a/openair2/RRC/NR/rrc_gNB.c +++ b/openair2/RRC/NR/rrc_gNB.c @@ -1673,6 +1673,7 @@ void rrc_gNB_process_RRCReestablishmentComplete(const protocol_ctxt_t *const ctx ue_context_pP->ue_context.masterCellGroup->spCellConfig = ue_context_pP->ue_context.spCellConfigReestablishment; ue_context_pP->ue_context.spCellConfigReestablishment = NULL; cellGroupConfig->spCellConfig = ue_context_pP->ue_context.masterCellGroup->spCellConfig; + cellGroupConfig->physicalCellGroupConfig = ue_context_pP->ue_context.masterCellGroup->physicalCellGroupConfig; fill_mastercellGroupConfig(cellGroupConfig, ue_context_pP->ue_context.masterCellGroup, rrc->um_on_default_drb, (drb_id_to_setup_start < 2) ? 1 : 0, drb_id_to_setup_start, nb_drb_to_setup, drb_priority); diff --git a/targets/PROJECTS/GENERIC-NR-5GC/CONF/gnb.sa.band41.fr1.52PRB.usrpb210.conf b/targets/PROJECTS/GENERIC-NR-5GC/CONF/gnb.sa.band41.fr1.52PRB.usrpb210.conf new file mode 100644 index 0000000000..fdf159d5d9 --- /dev/null +++ b/targets/PROJECTS/GENERIC-NR-5GC/CONF/gnb.sa.band41.fr1.52PRB.usrpb210.conf @@ -0,0 +1,288 @@ +# This file is derived from ./phy-demo/oai/targets/PROJECTS/GENERIC-NR-5GC/CONF/gnb.sa.band78.fr1.24PRB.usrpb210.conf +# ./nr-uesoftmodem -r 24 --numerology 1 --band 41 -C 2593320000 --ssb 24 --rfsim --sa + +Active_gNBs = ( "gNB-OAI"); +# Asn1_verbosity, choice in: none, info, annoying +Asn1_verbosity = "none"; + +gNBs = +( + { + ////////// Identification parameters: + gNB_ID = 0xe00; + gNB_name = "gNB-OAI"; + + // Tracking area code, 0x0000 and 0xfffe are reserved values + tracking_area_code = 1; + plmn_list = ({ mcc = 208; mnc = 99; mnc_length = 2; snssaiList = ({ sst = 1 }) }); + + nr_cellid = 12345678L; + + ////////// Physical parameters: + + do_CSIRS = 1; + do_SRS = 1; + sib1_tda = 15; + min_rxtxtime = 6; + + pdcch_ConfigSIB1 = ( + { + controlResourceSetZero = 6; + searchSpaceZero = 0; + } + ); + + servingCellConfigCommon = ( + { + #spCellConfigCommon + + physCellId = 0; + +# downlinkConfigCommon + #frequencyInfoDL + # this is 2574.270 MHz + 53 PRBs@30kHz SCS (same as initial BWP) + absoluteFrequencySSB = 518670; + dl_frequencyBand = 41; + # this is 3600 MHz + dl_absoluteFrequencyPointA = 517734; + #scs-SpecificCarrierList + dl_offstToCarrier = 0; +# subcarrierSpacing +# 0=kHz15, 1=kHz30, 2=kHz60, 3=kHz120 + dl_subcarrierSpacing = 0; + dl_carrierBandwidth = 52; + #initialDownlinkBWP + #genericParameters + # this is RBstart=0,L=106 (275*(L-1))+RBstart + initialDLBWPlocationAndBandwidth = 14025; # 6366 12925 12956 28875 12952 +# subcarrierSpacing +# 0=kHz15, 1=kHz30, 2=kHz60, 3=kHz120 + initialDLBWPsubcarrierSpacing = 0; + #pdcch-ConfigCommon + initialDLBWPcontrolResourceSetZero = 6; + initialDLBWPsearchSpaceZero = 0; + + #uplinkConfigCommon + #frequencyInfoUL + ul_frequencyBand = 41; + #scs-SpecificCarrierList + ul_offstToCarrier = 0; +# subcarrierSpacing +# 0=kHz15, 1=kHz30, 2=kHz60, 3=kHz120 + ul_subcarrierSpacing = 0; + ul_carrierBandwidth = 52; + pMax = 20; + #initialUplinkBWP + #genericParameters + initialULBWPlocationAndBandwidth = 14025; +# subcarrierSpacing +# 0=kHz15, 1=kHz30, 2=kHz60, 3=kHz120 + initialULBWPsubcarrierSpacing = 0; + #rach-ConfigCommon + #rach-ConfigGeneric + prach_ConfigurationIndex = 98; +#prach_msg1_FDM +#0 = one, 1=two, 2=four, 3=eight + prach_msg1_FDM = 0; + prach_msg1_FrequencyStart = 0; + zeroCorrelationZoneConfig = 13; + preambleReceivedTargetPower = -96; +#preamblTransMax (0...10) = (3,4,5,6,7,8,10,20,50,100,200) + preambleTransMax = 6; +#powerRampingStep +# 0=dB0,1=dB2,2=dB4,3=dB6 + powerRampingStep = 1; +#ra_ReponseWindow +#1,2,4,8,10,20,40,80 + ra_ResponseWindow = 4; +#ssb_perRACH_OccasionAndCB_PreamblesPerSSB_PR +#1=oneeighth,2=onefourth,3=half,4=one,5=two,6=four,7=eight,8=sixteen + ssb_perRACH_OccasionAndCB_PreamblesPerSSB_PR = 4; +#oneHalf (0..15) 4,8,12,16,...60,64 + ssb_perRACH_OccasionAndCB_PreamblesPerSSB = 15; +#ra_ContentionResolutionTimer +#(0..7) 8,16,24,32,40,48,56,64 + ra_ContentionResolutionTimer = 7; + rsrp_ThresholdSSB = 19; +#prach-RootSequenceIndex_PR +#1 = 839, 2 = 139 + prach_RootSequenceIndex_PR = 2; + prach_RootSequenceIndex = 1; + # SCS for msg1, can only be 15 for 30 kHz < 6 GHz, takes precendence over the one derived from prach-ConfigIndex + # + msg1_SubcarrierSpacing = 0, +# restrictedSetConfig +# 0=unrestricted, 1=restricted type A, 2=restricted type B + restrictedSetConfig = 0, + + msg3_DeltaPreamble = 1; + p0_NominalWithGrant =-90; + +# pucch-ConfigCommon setup : +# pucchGroupHopping +# 0 = neither, 1= group hopping, 2=sequence hopping + pucchGroupHopping = 0; + hoppingId = 40; + p0_nominal = -90; +# ssb_PositionsInBurs_BitmapPR +# 1=short, 2=medium, 3=long + ssb_PositionsInBurst_PR = 2; + ssb_PositionsInBurst_Bitmap = 1; + +# ssb_periodicityServingCell +# 0 = ms5, 1=ms10, 2=ms20, 3=ms40, 4=ms80, 5=ms160, 6=spare2, 7=spare1 + ssb_periodicityServingCell = 2; + +# dmrs_TypeA_position +# 0 = pos2, 1 = pos3 + dmrs_TypeA_Position = 0; + +# subcarrierSpacing +# 0=kHz15, 1=kHz30, 2=kHz60, 3=kHz120 + subcarrierSpacing = 0; + + + #tdd-UL-DL-ConfigurationCommon +# subcarrierSpacing +# 0=kHz15, 1=kHz30, 2=kHz60, 3=kHz120 + referenceSubcarrierSpacing = 0; + # pattern1 + # dl_UL_TransmissionPeriodicity + # 0=ms0p5, 1=ms0p625, 2=ms1, 3=ms1p25, 4=ms2, 5=ms2p5, 6=ms5, 7=ms10 + dl_UL_TransmissionPeriodicity = 7; + nrofDownlinkSlots = 7; + nrofDownlinkSymbols = 6; + nrofUplinkSlots = 2; + nrofUplinkSymbols = 4; + + ssPBCH_BlockPower = -25; + } + + ); + + + # ------- SCTP definitions + SCTP : + { + # Number of streams to use in input/output + SCTP_INSTREAMS = 2; + SCTP_OUTSTREAMS = 2; + }; + + + ////////// AMF parameters: + amf_ip_address = ( { ipv4 = "192.168.70.132"; + ipv6 = "192:168:30::17"; + active = "yes"; + preference = "ipv4"; + } + ); + + + NETWORK_INTERFACES : + { + GNB_INTERFACE_NAME_FOR_NG_AMF = "demo-oai"; + GNB_IPV4_ADDRESS_FOR_NG_AMF = "192.168.70.129/24"; + GNB_INTERFACE_NAME_FOR_NGU = "demo-oai"; + GNB_IPV4_ADDRESS_FOR_NGU = "192.168.70.129/24"; + GNB_PORT_FOR_S1U = 2152; # Spec 2152 + }; + + } +); + +MACRLCs = ( +{ + num_cc = 1; + tr_s_preference = "local_L1"; + tr_n_preference = "local_RRC"; + pusch_TargetSNRx10 = 150; + pucch_TargetSNRx10 = 200; + ulsch_max_frame_inactivity = 1; +} +); + +L1s = ( +{ + num_cc = 1; + tr_n_preference = "local_mac"; + pusch_proc_threads = 8; + prach_dtx_threshold = 120; + pucch0_dtx_threshold = 150; + ofdm_offset_divisor = 8; #set this to UINT_MAX for offset 0 +} +); + +RUs = ( +{ + local_rf = "yes" + nb_tx = 1 + nb_rx = 1 + att_tx = 0 + att_rx = 0; + bands = [41]; + max_pdschReferenceSignalPower = -27; + max_rxgain = 114; + sf_extension = 0; + eNB_instances = [0]; + #beamforming 1x4 matrix: + bf_weights = [0x00007fff, 0x0000, 0x0000, 0x0000]; + clock_src = "internal"; +} +); + +THREAD_STRUCT = ( +{ + #three config for level of parallelism "PARALLEL_SINGLE_THREAD", "PARALLEL_RU_L1_SPLIT", or "PARALLEL_RU_L1_TRX_SPLIT" + parallel_config = "PARALLEL_SINGLE_THREAD"; + #two option for worker "WORKER_DISABLE" or "WORKER_ENABLE" + worker_config = "WORKER_ENABLE"; +} +); + +rfsimulator : +{ + serveraddr = "server"; + serverport = "4043"; + options = (); #("saviq"); or/and "chanmod" + modelname = "AWGN"; + IQfile = "/tmp/rfsimulator.iqs"; +}; + +security = { + # preferred ciphering algorithms + # the first one of the list that an UE supports in chosen + # valid values: nea0, nea1, nea2, nea3 + ciphering_algorithms = ( "nea0" ); + + # preferred integrity algorithms + # the first one of the list that an UE supports in chosen + # valid values: nia0, nia1, nia2, nia3 + integrity_algorithms = ( "nia2", "nia0" ); + + # setting 'drb_ciphering' to "no" disables ciphering for DRBs, no matter + # what 'ciphering_algorithms' configures; same thing for 'drb_integrity' + drb_ciphering = "yes"; + drb_integrity = "no"; +}; + +log_config : +{ + global_log_level ="info"; + global_log_verbosity ="medium"; + hw_log_level ="info"; + hw_log_verbosity ="medium"; + phy_log_level ="info"; + phy_log_verbosity ="medium"; + mac_log_level ="info"; + mac_log_verbosity ="high"; + rlc_log_level ="info"; + rlc_log_verbosity ="medium"; + pdcp_log_level ="info"; + pdcp_log_verbosity ="medium"; + rrc_log_level ="info"; + rrc_log_verbosity ="medium"; + ngap_log_level ="debug"; + ngap_log_verbosity ="medium"; +}; +