diff --git a/cmake_targets/CMakeLists.txt b/cmake_targets/CMakeLists.txt
index e4d3acf4b1c976540a95e6cbd033bfa26b9d9ecd..0bf63cfb1ac626240f429314bf0ca312c72e56b2 100644
--- a/cmake_targets/CMakeLists.txt
+++ b/cmake_targets/CMakeLists.txt
@@ -1034,6 +1034,7 @@ set(PHY_SRC
${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/dlsch_modulation.c
${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/dlsch_demodulation.c
${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/dlsch_llr_computation.c
+ ${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/dlsch_llr_computation_avx2.c
${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/power_control.c
${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/dlsch_decoding.c
${OPENAIR1_DIR}/PHY/LTE_TRANSPORT/dlsch_scrambling.c
diff --git a/openair1/PHY/INIT/lte_init.c b/openair1/PHY/INIT/lte_init.c
index d0556a0235c042a9805cd595af26a51a2ae1cbc2..de83506bb31aed698f0c41fa95ff49468c204955 100755
--- a/openair1/PHY/INIT/lte_init.c
+++ b/openair1/PHY/INIT/lte_init.c
@@ -948,6 +948,11 @@ void phy_config_dedicated_ue(uint8_t Mod_id,int CC_id,uint8_t eNB_id,
// fill cqi parameters for periodic CQI reporting
get_cqipmiri_params(phy_vars_ue,eNB_id);
+ // disable MIB SIB decoding once we are on connected mode
+ LOG_I(PHY,"Disabling SIB MIB decoding \n");
+ phy_vars_ue->decode_SIB = 0;
+ phy_vars_ue->decode_MIB = 0;
+
}
void phy_config_cba_rnti (module_id_t Mod_id,int CC_id,eNB_flag_t eNB_flag, uint8_t index, rnti_t cba_rnti, uint8_t cba_group_id, uint8_t num_active_cba_groups)
@@ -1306,6 +1311,10 @@ int phy_init_lte_ue(PHY_VARS_UE *ue,
ue->high_speed_flag = 1;
ue->ch_est_alpha = 24576;
+ // enable MIB/SIB decoding by default
+ ue->decode_MIB = 1;
+ ue->decode_SIB = 1;
+
init_prach_tables(839);
diff --git a/openair1/PHY/LTE_ESTIMATION/lte_dl_bf_channel_estimation.c b/openair1/PHY/LTE_ESTIMATION/lte_dl_bf_channel_estimation.c
index 02b5102ae544a16b04e527ac9aed30fc1fc84009..7e112342ce15980c0e199cc930c6428ea8347db7 100644
--- a/openair1/PHY/LTE_ESTIMATION/lte_dl_bf_channel_estimation.c
+++ b/openair1/PHY/LTE_ESTIMATION/lte_dl_bf_channel_estimation.c
@@ -51,10 +51,10 @@ int lte_dl_bf_channel_estimation(PHY_VARS_UE *phy_vars_ue,
int uespec_pilot[300];
LTE_DL_FRAME_PARMS *frame_parms = &phy_vars_ue->frame_parms;
- LTE_UE_DLSCH_t **dlsch_ue = phy_vars_ue->dlsch[eNB_id];
+ LTE_UE_DLSCH_t **dlsch_ue = phy_vars_ue->dlsch[(Ns>>1)&0x1][eNB_id];
LTE_DL_UE_HARQ_t *dlsch0_harq;
- harq_pid = dlsch_ue[0]->current_harq_pid;
+ harq_pid = dlsch_ue[0]->current_harq_pid;
dlsch0_harq = dlsch_ue[0]->harq_processes[harq_pid];
if (((frame_parms->Ncp == NORMAL) && (symbol>=7)) ||
diff --git a/openair1/PHY/LTE_TRANSPORT/dci.c b/openair1/PHY/LTE_TRANSPORT/dci.c
index 49fe90bf15c0f58668d8e12f010aa4b46a2fe842..af25e33ff74dbc70ea7d4d46a6cdfcbf41fad1a1 100644
--- a/openair1/PHY/LTE_TRANSPORT/dci.c
+++ b/openair1/PHY/LTE_TRANSPORT/dci.c
@@ -2893,15 +2893,15 @@ void dci_decoding_procedure0(LTE_UE_PDCCH **pdcch_vars,
break;
case 2:
- *CCEmap|=(0x03<<(CCEind&0x1f));
+ *CCEmap|=(1<<(CCEind&0x1f));
break;
case 4:
- *CCEmap|=(0x0f<<(CCEind&0x1f));
+ *CCEmap|=(1<<(CCEind&0x1f));
break;
case 8:
- *CCEmap|=(0xff<<(CCEind&0x1f));
+ *CCEmap|=(1<<(CCEind&0x1f));
break;
}
@@ -3149,7 +3149,7 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0) ,
ra_rnti,
P_RNTI,
2,
@@ -3177,7 +3177,7 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
2,
@@ -3209,7 +3209,7 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
P_RNTI,
ra_rnti,
3,
@@ -3237,7 +3237,7 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
3,
@@ -3271,7 +3271,7 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
0,
@@ -3300,7 +3300,7 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
1,
@@ -3333,7 +3333,7 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
2,
@@ -3362,7 +3362,7 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
3,
@@ -3395,7 +3395,7 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
0,
@@ -3426,7 +3426,7 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
1,
@@ -3458,7 +3458,7 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
2,
@@ -3490,7 +3490,7 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
3,
@@ -3518,16 +3518,18 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
} else if (tmode == 3) {
+ LOG_D(PHY," Now check UE_SPEC format 2A_2A search aggregation 1\n");
// Now check UE_SPEC format 2A_2A search spaces at aggregation 1
+ old_dci_cnt=dci_cnt;
dci_decoding_procedure0(pdcch_vars,0,mode,
subframe,
dci_alloc,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
- P_RNTI,
+ P_RNTI,
0,
format1A,
format1A,
@@ -3542,21 +3544,25 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
&CCEmap1,
&CCEmap2);
+ LOG_D(PHY," format 2A_2A search CCEmap0 %x, format0_found %d, format_c_found %d \n", CCEmap0, format0_found, format_c_found);
if ((CCEmap0==0xffff)||
((format0_found==1)&&(format_c_found==1)))
return(dci_cnt);
+ LOG_D(PHY," format 2A_2A search dci_cnt %d, old_dci_cn t%d \n", dci_cnt, old_dci_cnt);
if (dci_cnt>old_dci_cnt)
return(dci_cnt);
// Now check UE_SPEC format 2 search spaces at aggregation 2
+ LOG_D(PHY," Now check UE_SPEC format 2A_2A search aggregation 2\n");
+ old_dci_cnt=dci_cnt;
dci_decoding_procedure0(pdcch_vars,0,mode,
subframe,
dci_alloc,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
1,
@@ -3577,19 +3583,22 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
((format0_found==1)&&(format_c_found==1)))
return(dci_cnt);
+ LOG_D(PHY," format 2A_2A search dci_cnt %d, old_dci_cn t%d \n", dci_cnt, old_dci_cnt);
if (dci_cnt>old_dci_cnt)
return(dci_cnt);
// Now check UE_SPEC format 2_2A search spaces at aggregation 4
+ LOG_D(PHY," Now check UE_SPEC format 2_2A search spaces at aggregation 4 \n");
+ old_dci_cnt=dci_cnt;
dci_decoding_procedure0(pdcch_vars,0,mode,
subframe,
dci_alloc,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
- P_RNTI,
+ P_RNTI,
2,
format1A,
format1A,
@@ -3608,18 +3617,21 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
((format0_found==1)&&(format_c_found==1)))
return(dci_cnt);
+ LOG_D(PHY," format 2A_2A search dci_cnt %d, old_dci_cn t%d \n", dci_cnt, old_dci_cnt);
if (dci_cnt>old_dci_cnt)
return(dci_cnt);
//#ifdef ALL_AGGREGATION
// Now check UE_SPEC format 2_2A search spaces at aggregation 8
+ LOG_D(PHY," Now check UE_SPEC format 2_2A search spaces at aggregation 8 \n");
+ old_dci_cnt=dci_cnt;
dci_decoding_procedure0(pdcch_vars,0,mode,
subframe,
dci_alloc,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
3,
@@ -3636,16 +3648,24 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
&CCEmap1,
&CCEmap2);
//#endif
+ if ((CCEmap0==0xffff)||
+ ((format0_found==1)&&(format_c_found==1)))
+ return(dci_cnt);
+
+ LOG_D(PHY," format 2A_2A search dci_cnt %d, old_dci_cn t%d \n", dci_cnt, old_dci_cnt);
+ if (dci_cnt>old_dci_cnt)
+ return(dci_cnt);
} else if (tmode == 4) {
// Now check UE_SPEC format 2_2A search spaces at aggregation 1
+ old_dci_cnt=dci_cnt;
dci_decoding_procedure0(pdcch_vars,0,mode,
subframe,
dci_alloc,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
0,
@@ -3670,13 +3690,14 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
return(dci_cnt);
// Now check UE_SPEC format 2 search spaces at aggregation 2
+ old_dci_cnt=dci_cnt;
dci_decoding_procedure0(pdcch_vars,0,mode,
subframe,
dci_alloc,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
1,
@@ -3701,13 +3722,14 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
return(dci_cnt);
// Now check UE_SPEC format 2_2A search spaces at aggregation 4
+ old_dci_cnt=dci_cnt;
dci_decoding_procedure0(pdcch_vars,0,mode,
subframe,
dci_alloc,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
2,
@@ -3733,13 +3755,14 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
//#ifdef ALL_AGGREGATION
// Now check UE_SPEC format 2_2A search spaces at aggregation 8
+ old_dci_cnt=dci_cnt;
dci_decoding_procedure0(pdcch_vars,0,mode,
subframe,
dci_alloc,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
3,
@@ -3762,13 +3785,14 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
#ifdef DEBUG_DCI_DECODING
LOG_I(PHY," MU-MIMO check UE_SPEC format 1E_2A_M10PRB\n");
#endif
+ old_dci_cnt=dci_cnt;
dci_decoding_procedure0(pdcch_vars,0,mode,
subframe,
dci_alloc,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
0,
@@ -3794,13 +3818,14 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
return(dci_cnt);
// Now check UE_SPEC format 1E_2A_M10PRB search spaces aggregation 2
+ old_dci_cnt=dci_cnt;
dci_decoding_procedure0(pdcch_vars,0,mode,
subframe,
dci_alloc,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
1,
@@ -3825,13 +3850,14 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
return(dci_cnt);
// Now check UE_SPEC format 1E_2A_M10PRB search spaces aggregation 4
+ old_dci_cnt=dci_cnt;
dci_decoding_procedure0(pdcch_vars,0,mode,
subframe,
dci_alloc,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
2,
@@ -3858,13 +3884,14 @@ uint16_t dci_decoding_procedure(PHY_VARS_UE *ue,
//#ifdef ALL_AGGREGATION
// Now check UE_SPEC format 1E_2A_M10PRB search spaces at aggregation 8
+ old_dci_cnt=dci_cnt;
dci_decoding_procedure0(pdcch_vars,0,mode,
subframe,
dci_alloc,
eNB_id,
frame_parms,
mi,
- SI_RNTI,
+ ((ue->decode_SIB == 1) ? SI_RNTI : 0),
ra_rnti,
P_RNTI,
3,
diff --git a/openair1/PHY/LTE_TRANSPORT/dci_tools.c b/openair1/PHY/LTE_TRANSPORT/dci_tools.c
index 9f4749cdc60a69f8bf3d7281a4a533891d49dcae..c4e379b45ab1b42c1db7b9622f67402be0f0b737 100644
--- a/openair1/PHY/LTE_TRANSPORT/dci_tools.c
+++ b/openair1/PHY/LTE_TRANSPORT/dci_tools.c
@@ -4763,22 +4763,26 @@ int check_dci_format1_1a_coherency(DCI_format_t dci_format,
uint16_t si_rnti,
uint16_t ra_rnti,
uint16_t p_rnti,
+ uint32_t frame,
+ uint8_t subframe,
DCI_INFO_EXTRACTED_t *pdci_info_extarcted,
LTE_DL_UE_HARQ_t *pdlsch0_harq)
{
uint8_t harq_pid = pdci_info_extarcted->harq_pid;
uint32_t rballoc = pdci_info_extarcted->rballoc;
uint8_t mcs1 = pdci_info_extarcted->mcs1;
- uint8_t rv1 = pdci_info_extarcted->rv1;
- uint8_t ndi1 = pdci_info_extarcted->ndi1;
uint8_t TPC = pdci_info_extarcted->TPC;
uint8_t rah = pdci_info_extarcted->rah;
+#ifdef DEBUG_DCI
+ uint8_t rv1 = pdci_info_extarcted->rv1;
+ uint8_t ndi1 = pdci_info_extarcted->ndi1;
+#endif
uint8_t NPRB = 0;
long long int RIV_max = 0;
#ifdef DEBUG_DCI
- LOG_I(PHY,"[DCI-FORMAT-1-1A] dci_format %d\n", dci_format);
+ LOG_I(PHY,"[DCI-FORMAT-1-1A] AbsSubframe %d.%d dci_format %d\n", frame, subframe, dci_format);
LOG_I(PHY,"[DCI-FORMAT-1-1A] rnti %x\n", rnti);
LOG_I(PHY,"[DCI-FORMAT-1-1A] harq_pid %d\n", harq_pid);
LOG_I(PHY,"[DCI-FORMAT-1-1A] rah %d\n", rah);
@@ -4983,24 +4987,30 @@ int check_dci_format2_2a_coherency(DCI_format_t dci_format,
uint8_t rv2 = pdci_info_extarcted->rv2;
uint8_t harq_pid = pdci_info_extarcted->harq_pid;
uint32_t rballoc = pdci_info_extarcted->rballoc;
+
+#ifdef DEBUG_DCI
uint8_t ndi1 = pdci_info_extarcted->ndi1;
uint8_t ndi2 = pdci_info_extarcted->ndi2;
+#endif
uint8_t NPRB = 0;
long long RIV_max = 0;
+#ifdef DEBUG_DCI
LOG_I(PHY, "extarcted dci - dci_format %d \n", dci_format);
+ LOG_I(PHY, "extarcted dci - rnti %d \n", rnti);
LOG_I(PHY, "extarcted dci - rah %d \n", rah);
LOG_I(PHY, "extarcted dci - mcs1 %d \n", mcs1);
LOG_I(PHY, "extarcted dci - mcs2 %d \n", mcs2);
LOG_I(PHY, "extarcted dci - rv1 %d \n", rv1);
LOG_I(PHY, "extarcted dci - rv2 %d \n", rv2);
- LOG_I(PHY, "extarcted dci - ndi1 %d \n", ndi1);
- LOG_I(PHY, "extarcted dci - ndi2 %d \n", ndi2);
+ //LOG_I(PHY, "extarcted dci - ndi1 %d \n", ndi1);
+ //LOG_I(PHY, "extarcted dci - ndi2 %d \n", ndi2);
LOG_I(PHY, "extarcted dci - rballoc %x \n", rballoc);
- LOG_I(PHY, "extarcted dci - harq pif %d \n", harq_pid);
+ LOG_I(PHY, "extarcted dci - harq pid %d \n", harq_pid);
LOG_I(PHY, "extarcted dci - round0 %d \n", pdlsch0_harq->round);
LOG_I(PHY, "extarcted dci - round1 %d \n", pdlsch1_harq->round);
+#endif
// I- check dci content minimum coherency
if(harq_pid >8)
@@ -5043,6 +5053,21 @@ int check_dci_format2_2a_coherency(DCI_format_t dci_format,
}*/
+ if((pdlsch0_harq->round == 0) && (rv1 > 0))
+ {
+ // DCI false detection
+ LOG_I(PHY,"bad rv1\n");
+ return(0);
+ }
+
+ if((pdlsch1_harq->round == 0) && (rv2 > 0))
+ {
+ // DCI false detection
+ LOG_I(PHY,"bad rv2\n");
+ return(0);
+ }
+
+
switch (N_RB_DL) {
case 6:
if (rah == 0)
@@ -5606,6 +5631,7 @@ void compute_precoding_info_format2A(uint8_t tpmi,
void prepare_dl_decoding_format2_2A(DCI_format_t dci_format,
DCI_INFO_EXTRACTED_t *pdci_info_extarcted,
LTE_DL_FRAME_PARMS *frame_parms,
+ uint16_t rnti,
uint8_t subframe,
LTE_DL_UE_HARQ_t *dlsch0_harq,
LTE_DL_UE_HARQ_t *dlsch1_harq,
@@ -5663,15 +5689,17 @@ void prepare_dl_decoding_format2_2A(DCI_format_t dci_format,
dlsch1_harq->dl_power_off = 1;
pdlsch0->current_harq_pid = harq_pid;
- pdlsch0->harq_ack[subframe].harq_id = harq_pid;
+ pdlsch0->harq_ack[subframe].harq_id = harq_pid;
pdlsch1->current_harq_pid = harq_pid;
- pdlsch1->harq_ack[subframe].harq_id = harq_pid;
+ pdlsch1->harq_ack[subframe].harq_id = harq_pid;
// assume two CW are active
dlsch0_harq->status = ACTIVE;
dlsch1_harq->status = ACTIVE;
pdlsch0->active = 1;
pdlsch1->active = 1;
+ pdlsch0->rnti = rnti;
+ pdlsch1->rnti = rnti;
if (TB0_active && TB1_active && tbswap==1) {
@@ -5682,20 +5710,16 @@ void prepare_dl_decoding_format2_2A(DCI_format_t dci_format,
if (TB0_active==0) {
dlsch0_harq->status = SCH_IDLE;
pdlsch0->active = 0;
-#ifdef DEBUG_HARQ
+ #ifdef DEBUG_HARQ
printf("[DCI UE]: TB0 is deactivated, retransmit TB1 transmit in TM6\n");
-#endif
+ #endif
}
if (TB1_active==0) {
dlsch1_harq->status = SCH_IDLE;
pdlsch1->active = 0;
-#ifdef DEBUG_HARQ
- printf("[DCI UE]: TB1 is deactivated, retransmit TB0 transmit in TM6\n");
-#endif
}
-
#ifdef DEBUG_HARQ
printf("[DCI UE]: dlsch0_harq status %d , dlsch1_harq status %d\n", dlsch0_harq->status, dlsch1_harq->status);
#endif
@@ -5727,6 +5751,9 @@ void prepare_dl_decoding_format2_2A(DCI_format_t dci_format,
dlsch1_harq->rb_alloc_odd[3] = dlsch0_harq->rb_alloc_odd[3];
dlsch1_harq->nb_rb = dlsch0_harq->nb_rb;
+
+ //dlsch0_harq->Nl = 1;
+ //dlsch1_harq->Nl = 1;
}
} else if ((TB0_active == 0) && (TB1_active == 1)){
@@ -5774,38 +5801,57 @@ void prepare_dl_decoding_format2_2A(DCI_format_t dci_format,
if ((ndi1!=dlsch0_harq->DCINdi) || (dlsch0_harq->first_tx==1)) {
dlsch0_harq->round = 0;
+ //LOG_I(PHY,"[UE] DLSCH: New Data Indicator CW0 subframe %d (pid %d, round %d)\n",
+ // subframe,harq_pid,dlsch0_harq->round);
if ( dlsch0_harq->first_tx==1) {
LOG_D(PHY,"Format 2 DCI First TX0: Clearing flag\n");
dlsch0_harq->first_tx = 0;
}
}else{
if(dlsch0_harq->round == 0) {
+#if 0
// skip pdsch decoding and report ack
dlsch0_harq->status = SCH_IDLE;
pdlsch0->active = 0;
pdlsch0->harq_ack[subframe].ack = 1;
pdlsch0->harq_ack[subframe].harq_id = harq_pid;
pdlsch0->harq_ack[subframe].send_harq_status = 1;
+#endif
}
}
- dlsch0_harq->TBS = TBStable[get_I_TBS(dlsch0_harq->mcs)][dlsch0_harq->nb_rb-1];
- if(dlsch0_harq->Nl == 2)
- dlsch0_harq->TBS = TBStable[get_I_TBS(dlsch0_harq->mcs)][(dlsch0_harq->nb_rb<<1)-1];
- if (mcs1 <= 28)
+ // if Imcs in [29..31] TBS is assumed to be as determined from DCI transported in the latest
+ // PDCCH for the same trasport block using Imcs in [0 .. 28]
+ if(dlsch0_harq->mcs <= 28)
+ {
+ dlsch0_harq->TBS = TBStable[get_I_TBS(dlsch0_harq->mcs)][dlsch0_harq->nb_rb-1];
+ LOG_D(PHY,"[UE] DLSCH: New TBS CW0 subframe %d (pid %d, round %d) TBS %d \n",
+ subframe,harq_pid,dlsch0_harq->round, dlsch0_harq->TBS);
+ }
+ else
+ {
+ LOG_D(PHY,"[UE] DLSCH: Keep the same TBS CW0 subframe %d (pid %d, round %d) TBS %d \n",
+ subframe,harq_pid,dlsch0_harq->round, dlsch0_harq->TBS);
+ }
+ //if(dlsch0_harq->Nl == 2)
+ //dlsch0_harq->TBS = TBStable[get_I_TBS(dlsch0_harq->mcs)][(dlsch0_harq->nb_rb<<1)-1];
+ if (mcs1 <= 28)
dlsch0_harq->Qm = get_Qm(mcs1);
- else if (mcs1<=31)
+ else if (mcs1<=31)
dlsch0_harq->Qm = (mcs1-28)<<1;
}
if (TB1_active) {
if ((ndi2!=dlsch1_harq->DCINdi) || (dlsch1_harq->first_tx==1)) {
dlsch1_harq->round = 0;
+ //LOG_I(PHY,"[UE] DLSCH: New Data Indicator CW1 subframe %d (pid %d, round %d)\n",
+ // subframe,harq_pid,dlsch0_harq->round);
if (dlsch1_harq->first_tx==1) {
LOG_D(PHY,"Format 2 DCI First TX1: Clearing flag\n");
dlsch1_harq->first_tx = 0;
}
}else{
+#if 0
if(dlsch1_harq->round == 0) {
// skip pdsch decoding and report ack
dlsch1_harq->status = SCH_IDLE;
@@ -5814,15 +5860,25 @@ void prepare_dl_decoding_format2_2A(DCI_format_t dci_format,
pdlsch1->harq_ack[subframe].harq_id = harq_pid;
pdlsch1->harq_ack[subframe].send_harq_status = 1;
}
+#endif
}
- dlsch1_harq->TBS = TBStable[get_I_TBS(dlsch1_harq->mcs)][dlsch1_harq->nb_rb-1];
- if(dlsch0_harq->Nl == 2)
- dlsch0_harq->TBS = TBStable[get_I_TBS(dlsch0_harq->mcs)][(dlsch0_harq->nb_rb<<1)-1];
-
- if (mcs2 <= 28)
+ // if Imcs in [29..31] TBS is assumed to be as determined from DCI transported in the latest
+ // PDCCH for the same trasport block using Imcs in [0 .. 28]
+ if(dlsch1_harq->mcs <= 28)
+ {
+ dlsch1_harq->TBS = TBStable[get_I_TBS(dlsch1_harq->mcs)][dlsch1_harq->nb_rb-1];
+ LOG_D(PHY,"[UE] DLSCH: New TBS CW1 subframe %d (pid %d, round %d) TBS %d \n",
+ subframe,harq_pid,dlsch1_harq->round, dlsch1_harq->TBS);
+ }
+ else
+ {
+ LOG_D(PHY,"[UE] DLSCH: Keep the same TBS CW1 subframe %d (pid %d, round %d) TBS %d \n",
+ subframe,harq_pid,dlsch1_harq->round, dlsch1_harq->TBS);
+ }
+ if (mcs2 <= 28)
dlsch1_harq->Qm = get_Qm(mcs2);
- else if (mcs1<=31)
+ else if (mcs1<=31)
dlsch1_harq->Qm = (mcs2-28)<<1;
}
@@ -5920,7 +5976,7 @@ int generate_ue_dlsch_params_from_dci(int frame,
tc_rnti,
si_rnti,
ra_rnti,
- p_rnti,
+ p_rnti,frame,subframe,
&dci_info_extarcted,
dlsch0_harq);
if(status == 0)
@@ -6020,7 +6076,7 @@ int generate_ue_dlsch_params_from_dci(int frame,
tc_rnti,
si_rnti,
ra_rnti,
- p_rnti,
+ p_rnti,frame,subframe,
&dci_info_extarcted,
dlsch0_harq);
if(status == 0)
@@ -6047,7 +6103,7 @@ int generate_ue_dlsch_params_from_dci(int frame,
case format2:
{
// extract dci infomation
- LOG_I(PHY,"[DCI-format2] extract dci infomation \n");
+ //LOG_I(PHY,"[DCI-format2] AbsSubframe %d.%d extract dci infomation \n", frame, subframe);
extract_dci2_info(frame_parms->N_RB_DL,
frame_type,
frame_parms->nb_antenna_ports_eNB,
@@ -6069,7 +6125,7 @@ int generate_ue_dlsch_params_from_dci(int frame,
dlsch0_harq = dlsch0->harq_processes[harq_pid];
dlsch1_harq = dlsch1->harq_processes[harq_pid];
- LOG_I(PHY,"[DCI-format2] check dci content \n");
+ //LOG_I(PHY,"[DCI-format2] check dci content \n");
status = check_dci_format2_2a_coherency(format2,
frame_parms->N_RB_DL,
&dci_info_extarcted,
@@ -6083,10 +6139,11 @@ int generate_ue_dlsch_params_from_dci(int frame,
return(-1);
// dci is correct ==> update internal structure and prepare dl decoding
- LOG_I(PHY,"[DCI-format2] update internal structure and prepare dl decoding \n");
+ //LOG_I(PHY,"[DCI-format2] update internal structure and prepare dl decoding \n");
prepare_dl_decoding_format2_2A(format2,
&dci_info_extarcted,
frame_parms,
+ rnti,
subframe,
dlsch0_harq,
dlsch1_harq,
@@ -6099,7 +6156,7 @@ int generate_ue_dlsch_params_from_dci(int frame,
case format2A:
{
// extract dci infomation
- LOG_I(PHY,"[DCI-format2A] extract dci infomation \n");
+ //LOG_I(PHY,"[DCI-format2] AbsSubframe %d.%d extract dci infomation \n", frame%1024, subframe);
extract_dci2A_info(frame_parms->N_RB_DL,
frame_type,
frame_parms->nb_antenna_ports_eNB,
@@ -6107,10 +6164,10 @@ int generate_ue_dlsch_params_from_dci(int frame,
&dci_info_extarcted);
// check dci content
- LOG_I(PHY,"[DCI-format2A] check dci content \n");
- LOG_I(PHY,"[DCI-format2A] tb_swap %d harq_pid %d\n", dci_info_extarcted.tb_swap, dci_info_extarcted.harq_pid);
- dlsch[0]->active = 0;
- dlsch[1]->active = 0;
+ //LOG_I(PHY,"[DCI-format2A] check dci content \n");
+ //LOG_I(PHY,"[DCI-format2A] tb_swap %d harq_pid %d\n", dci_info_extarcted.tb_swap, dci_info_extarcted.harq_pid);
+ //dlsch[0]->active = 0;
+ //dlsch[1]->active = 0;
if (dci_info_extarcted.tb_swap == 0) {
dlsch0 = dlsch[0];
@@ -6122,7 +6179,7 @@ int generate_ue_dlsch_params_from_dci(int frame,
dlsch0_harq = dlsch0->harq_processes[dci_info_extarcted.harq_pid];
dlsch1_harq = dlsch1->harq_processes[dci_info_extarcted.harq_pid];
- LOG_I(PHY,"[DCI-format2A] check dci content \n");
+ //LOG_I(PHY,"[DCI-format2A] check dci content \n");
status = check_dci_format2_2a_coherency(format2A,
frame_parms->N_RB_DL,
&dci_info_extarcted,
@@ -6136,10 +6193,11 @@ int generate_ue_dlsch_params_from_dci(int frame,
return(-1);
// dci is correct ==> update internal structure and prepare dl decoding
- LOG_I(PHY,"[DCI-format2A] update internal structure and prepare dl decoding \n");
+ //LOG_I(PHY,"[DCI-format2A] update internal structure and prepare dl decoding \n");
prepare_dl_decoding_format2_2A(format2A,
&dci_info_extarcted,
frame_parms,
+ rnti,
subframe,
dlsch0_harq,
dlsch1_harq,
@@ -7142,7 +7200,7 @@ int generate_ue_ulsch_params_from_dci(void *dci_pdu,
uint8_t transmission_mode = ue->transmission_mode[eNB_id];
ANFBmode_t AckNackFBMode;
LTE_UE_ULSCH_t *ulsch = ue->ulsch[eNB_id];
- LTE_UE_DLSCH_t **dlsch = ue->dlsch[0];
+ LTE_UE_DLSCH_t **dlsch = ue->dlsch[subframe&0x1][0];
PHY_MEASUREMENTS *meas = &ue->measurements;
LTE_DL_FRAME_PARMS *frame_parms = &ue->frame_parms;
// uint32_t current_dlsch_cqi = ue->current_dlsch_cqi[eNB_id];
@@ -7947,7 +8005,7 @@ int generate_ue_ulsch_params_from_dci(void *dci_pdu,
if (frame_parms->frame_type == FDD) {
int dl_subframe = (subframe<4) ? (subframe+6) : (subframe-4);
- if (ue->dlsch[eNB_id][0]->harq_ack[dl_subframe].send_harq_status>0) { // we have downlink transmission
+ if (ue->dlsch[dl_subframe&0x1][eNB_id][0]->harq_ack[dl_subframe].send_harq_status>0) { // we have downlink transmission
ulsch->harq_processes[harq_pid]->O_ACK = 1;
} else {
ulsch->harq_processes[harq_pid]->O_ACK = 0;
diff --git a/openair1/PHY/LTE_TRANSPORT/defs.h b/openair1/PHY/LTE_TRANSPORT/defs.h
index 3b13c01c52ccf0b2b02e69772aa38ec61487d3e5..f6293bf1b2d73689d6769a1270bfe0ad53fc418b 100755
--- a/openair1/PHY/LTE_TRANSPORT/defs.h
+++ b/openair1/PHY/LTE_TRANSPORT/defs.h
@@ -708,7 +708,7 @@ typedef struct {
int16_t sqrt_rho_a;
/// amplitude of PDSCH (compared to RS) in symbols containing pilots
int16_t sqrt_rho_b;
- /// Current HARQ process id
+ /// Current HARQ process id threadRx Odd and threadRx Even
uint8_t current_harq_pid;
/// Current subband antenna selection
uint32_t antenna_alloc;
diff --git a/openair1/PHY/LTE_TRANSPORT/dlsch_decoding.c b/openair1/PHY/LTE_TRANSPORT/dlsch_decoding.c
index 04794e3ffcba4fcf6540f31cffe594416774b23f..d683decc9f57230760206b91795621eaa585de8e 100644
--- a/openair1/PHY/LTE_TRANSPORT/dlsch_decoding.c
+++ b/openair1/PHY/LTE_TRANSPORT/dlsch_decoding.c
@@ -38,6 +38,7 @@
#include "SIMULATION/TOOLS/defs.h"
//#define DEBUG_DLSCH_DECODING
+extern double cpuf;
void free_ue_dlsch(LTE_UE_DLSCH_t *dlsch)
{
@@ -270,7 +271,7 @@ uint32_t dlsch_decoding(PHY_VARS_UE *phy_vars_ue,
return(max_turbo_iterations);
}*/
- /*harq_pid = dlsch->current_harq_pid;
+ /*harq_pid = dlsch->current_harq_pid[subframe&0x1];
if (harq_pid >= 8) {
printf("dlsch_decoding.c: Illegal harq_pid %d\n",harq_pid);
return(max_turbo_iterations);
@@ -341,6 +342,8 @@ uint32_t dlsch_decoding(PHY_VARS_UE *phy_vars_ue,
printf("Segmentation: C %d, Cminus %d, Kminus %d, Kplus %d\n",harq_process->C,harq_process->Cminus,harq_process->Kminus,harq_process->Kplus);
#endif
+ opp_enabled=1;
+
for (r=0; r<harq_process->C; r++) {
@@ -376,7 +379,7 @@ uint32_t dlsch_decoding(PHY_VARS_UE *phy_vars_ue,
(r==0) ? harq_process->F : 0);
#ifdef DEBUG_DLSCH_DECODING
- LOG_I(PHY,"HARQ_PID %d Rate Matching Segment %d (coded bits %d,unpunctured/repeated bits %d, TBS %d, mod_order %d, nb_rb %d, Nl %d, rv %d, round %d)...\n",
+ LOG_D(PHY,"HARQ_PID %d Rate Matching Segment %d (coded bits %d,unpunctured/repeated bits %d, TBS %d, mod_order %d, nb_rb %d, Nl %d, rv %d, round %d)...\n",
harq_pid,r, G,
Kr*3,
harq_process->TBS,
@@ -459,6 +462,10 @@ uint32_t dlsch_decoding(PHY_VARS_UE *phy_vars_ue,
#if 1
if (err_flag == 0) {
+ LOG_D(PHY, "turbo algo Kr=%d cb_cnt=%d C=%d nbRB=%d TBSInput=%d TBSHarq=%d TBSplus24=%d mcs=%d Qm=%d RIV=%d round=%d maxIter %d\n",
+ Kr,r,harq_process->C,harq_process->nb_rb,A,harq_process->TBS,
+ harq_process->B,harq_process->mcs,harq_process->Qm,harq_process->rvidx,harq_process->round,dlsch->max_turbo_iterations);
+
if (llr8_flag) {
AssertFatal (Kr >= 256, "turbo algo issue Kr=%d cb_cnt=%d C=%d nbRB=%d TBSInput=%d TBSHarq=%d TBSplus24=%d mcs=%d Qm=%d RIV=%d round=%d\n",
Kr,r,harq_process->C,harq_process->nb_rb,A,harq_process->TBS,harq_process->B,harq_process->mcs,harq_process->Qm,harq_process->rvidx,harq_process->round);
@@ -605,6 +612,13 @@ uint32_t dlsch_decoding(PHY_VARS_UE *phy_vars_ue,
&phy_vars_ue->dlsch_tc_intl2_stats); //(is_crnti==0)?harq_pid:harq_pid+1);
stop_meas(dlsch_turbo_decoding_stats);
+ /*printf("Segmentation: C %d r %d, dlsch_rate_unmatching_stats %5.3f dlsch_deinterleaving_stats %5.3f dlsch_turbo_decoding_stats %5.3f \n",
+ harq_process->C,
+ r,
+ dlsch_rate_unmatching_stats->p_time/(cpuf*1000.0),
+ dlsch_deinterleaving_stats->p_time/(cpuf*1000.0),
+ dlsch_turbo_decoding_stats->p_time/(cpuf*1000.0));*/
+
}
}
}
@@ -636,25 +650,28 @@ uint32_t dlsch_decoding(PHY_VARS_UE *phy_vars_ue,
harq_process->round++;
- if(is_crnti)
- {
- LOG_D(PHY,"[UE %d] DLSCH: Setting NACK for subframe %d (pid %d, round %d, TBS %d)\n",phy_vars_ue->Mod_id,subframe,harq_pid,harq_process->round,harq_process->TBS);
- }
// printf("Rate: [UE %d] DLSCH: Setting NACK for subframe %d (pid %d, round %d)\n",phy_vars_ue->Mod_id,subframe,harq_pid,harq_process->round);
if (harq_process->round >= dlsch->Mdlharq) {
harq_process->status = SCH_IDLE;
+ harq_process->round = 0;
+ }
+ if(is_crnti)
+ {
+ LOG_D(PHY,"[UE %d] DLSCH: Setting NACK for subframe %d (pid %d, pid status %d, round %d/Max %d, TBS %d)\n",
+ phy_vars_ue->Mod_id,subframe,harq_pid,harq_process->status,harq_process->round,dlsch->Mdlharq,harq_process->TBS);
}
return((1+dlsch->max_turbo_iterations));
} else {
- LOG_D(PHY,"[UE %d] DLSCH: Setting ACK for SFN/SF %d/%d (pid %d, round %d, subframe %d)\n",
- phy_vars_ue->Mod_id, frame_rx_prev, subframe_rx_prev, harq_pid, harq_process->round, subframe);
harq_process->status = SCH_IDLE;
harq_process->round = 0;
dlsch->harq_ack[subframe].ack = 1;
dlsch->harq_ack[subframe].harq_id = harq_pid;
dlsch->harq_ack[subframe].send_harq_status = 1;
+ LOG_D(PHY,"[UE %d] DLSCH: Setting ACK for SFN/SF %d/%d (pid %d, pid status %d, round %d, subframe %d)\n",
+ phy_vars_ue->Mod_id, frame_rx_prev, subframe_rx_prev, harq_pid, harq_process->status, harq_process->round, subframe);
+
if(is_crnti)
{
LOG_D(PHY,"[UE %d] DLSCH: Setting ACK for subframe %d (pid %d, round %d, TBS %d)\n",phy_vars_ue->Mod_id,subframe,harq_pid,harq_process->round,harq_process->TBS);
@@ -908,7 +925,7 @@ uint32_t dlsch_decoding_emul(PHY_VARS_UE *phy_vars_ue,
break;
case PDSCH: // TB0
- dlsch_ue = phy_vars_ue->dlsch[eNB_id][0];
+ dlsch_ue = phy_vars_ue->dlsch[subframe&0x1][eNB_id][0];
harq_pid = dlsch_ue->current_harq_pid;
ue_id= (uint32_t)find_ue((int16_t)phy_vars_ue->pdcch_vars[(uint32_t)eNB_id]->crnti,PHY_vars_eNB_g[eNB_id2][CC_id]);
DevAssert( ue_id != (uint32_t)-1 );
@@ -954,7 +971,7 @@ uint32_t dlsch_decoding_emul(PHY_VARS_UE *phy_vars_ue,
break;
case PDSCH1: { // TB1
- dlsch_ue = phy_vars_ue->dlsch[eNB_id][1];
+ dlsch_ue = phy_vars_ue->dlsch[subframe&0x1][eNB_id][1];
harq_pid = dlsch_ue->current_harq_pid;
int8_t UE_id = find_ue( phy_vars_ue->pdcch_vars[eNB_id]->crnti, PHY_vars_eNB_g[eNB_id2][CC_id] );
DevAssert( UE_id != -1 );
@@ -1008,7 +1025,7 @@ uint32_t dlsch_decoding_emul(PHY_VARS_UE *phy_vars_ue,
break;
default:
- dlsch_ue = phy_vars_ue->dlsch[eNB_id][0];
+ dlsch_ue = phy_vars_ue->dlsch[subframe&0x1][eNB_id][0];
LOG_E(PHY,"dlsch_decoding_emul: FATAL, unknown DLSCH_id %d\n",dlsch_id);
dlsch_ue->last_iteration_cnt = 1+dlsch_ue->max_turbo_iterations;
return(1+dlsch_ue->max_turbo_iterations);
diff --git a/openair1/PHY/LTE_TRANSPORT/dlsch_demodulation.c b/openair1/PHY/LTE_TRANSPORT/dlsch_demodulation.c
index 601eacc6205143468931e31b6526a58dfba962c9..72b486c04202891f9d600fad11a185d918dfcb43 100644
--- a/openair1/PHY/LTE_TRANSPORT/dlsch_demodulation.c
+++ b/openair1/PHY/LTE_TRANSPORT/dlsch_demodulation.c
@@ -42,6 +42,7 @@
#define NOCYGWIN_STATIC
#endif
+extern int16_t dlsch_demod_shift;
//#define DEBUG_HARQ
//#undef LOG_D
@@ -138,16 +139,18 @@ int rx_pdsch(PHY_VARS_UE *ue,
case PDSCH:
pdsch_vars = &ue->pdsch_vars[subframe&0x1][eNB_id];
- dlsch = ue->dlsch[eNB_id];
+ dlsch = ue->dlsch[subframe&0x1][eNB_id];
+ LOG_D(PHY,"AbsSubframe %d.%d / Sym %d harq_pid %d, harq status %d.%d \n",
+ frame,subframe,symbol,harq_pid,
+ dlsch[0]->harq_processes[harq_pid]->status,
+ dlsch[1]->harq_processes[harq_pid]->status);
+
if ((dlsch[0]->harq_processes[harq_pid]->status == ACTIVE) &&
(dlsch[1]->harq_processes[harq_pid]->status == ACTIVE)){
codeword_TB0 = dlsch[0]->harq_processes[harq_pid]->codeword;
codeword_TB1 = dlsch[1]->harq_processes[harq_pid]->codeword;
dlsch0_harq = dlsch[codeword_TB0]->harq_processes[harq_pid];
dlsch1_harq = dlsch[codeword_TB1]->harq_processes[harq_pid];
-#ifdef DEBUG_HARQ
- printf("I am assuming both CW active\n");
-#endif
}
else if ((dlsch[0]->harq_processes[harq_pid]->status == ACTIVE) &&
(dlsch[1]->harq_processes[harq_pid]->status != ACTIVE) ) {
@@ -159,8 +162,8 @@ int rx_pdsch(PHY_VARS_UE *ue,
else if ((dlsch[0]->harq_processes[harq_pid]->status != ACTIVE) &&
(dlsch[1]->harq_processes[harq_pid]->status == ACTIVE) ){
codeword_TB1 = dlsch[1]->harq_processes[harq_pid]->codeword;
- dlsch0_harq = dlsch[1]->harq_processes[harq_pid];
- dlsch1_harq = NULL;
+ dlsch0_harq = dlsch[1]->harq_processes[harq_pid];
+ dlsch1_harq = NULL;
codeword_TB0 = -1;
}
else {
@@ -336,7 +339,7 @@ int rx_pdsch(PHY_VARS_UE *ue,
LOG_W(PHY,"dlsch_demodulation: beamforming mode not supported yet.\n");
}
- // printf("nb_rb = %d, eNB_id %d\n",nb_rb,eNB_id);
+ //printf("nb_rb = %d, eNB_id %d\n",nb_rb,eNB_id);
if (nb_rb==0) {
LOG_D(PHY,"dlsch_demodulation.c: nb_rb=0\n");
return(-1);
@@ -357,12 +360,18 @@ int rx_pdsch(PHY_VARS_UE *ue,
symbol,
nb_rb);
- if ((dlsch0_harq->mimo_mode<DUALSTREAM_UNIFORM_PRECODING1) && (rx_type==rx_IC_single_stream) && (eNB_id_i==ue->n_connected_eNB) && (dlsch0_harq->dl_power_off==0)) // TM5 two-user
+ if ((dlsch0_harq->mimo_mode<DUALSTREAM_UNIFORM_PRECODING1) &&
+ (rx_type==rx_IC_single_stream) &&
+ (eNB_id_i==ue->n_connected_eNB) &&
+ (dlsch0_harq->dl_power_off==0)
+ ) // TM5 two-user
+ {
dlsch_scale_channel(pdsch_vars[eNB_id_i]->dl_ch_estimates_ext,
frame_parms,
dlsch,
symbol,
nb_rb);
+ }
if (first_symbol_flag==1) {
if (beamforming_mode==0){
@@ -392,21 +401,23 @@ int rx_pdsch(PHY_VARS_UE *ue,
nb_rb,
dlsch0_harq->mimo_mode);
+ LOG_D(PHY,"Channel Level TM34 avg_0 %d, avg_1 %d, rx_type %d, rx_standard %d, interf_unaw_shift %d \n", avg_0[0],
+ avg_1[0], rx_type, rx_standard, interf_unaw_shift);
if (rx_type>rx_standard) {
- avg_0[0] = (log2_approx(avg_0[0])/2) -13 + interf_unaw_shift;
- avg_1[0] = (log2_approx(avg_1[0])/2) -13 + interf_unaw_shift;
+ avg_0[0] = (log2_approx(avg_0[0])/2) + dlsch_demod_shift;// + 2 ;//+ 4;
+ avg_1[0] = (log2_approx(avg_1[0])/2) + dlsch_demod_shift;// + 2 ;//+ 4;
pdsch_vars[eNB_id]->log2_maxh0 = cmax(avg_0[0],0);
pdsch_vars[eNB_id]->log2_maxh1 = cmax(avg_1[0],0);
- //printf("TM4 I-A log2_maxh0 = %d\n", lte_ue_pdsch_vars[eNB_id]->log2_maxh0);
- //printf("TM4 I-A log2_maxh1 = %d\n", lte_ue_pdsch_vars[eNB_id]->log2_maxh1);
+ //printf("TM4 I-A log2_maxh0 = %d\n", pdsch_vars[eNB_id]->log2_maxh0);
+ //printf("TM4 I-A log2_maxh1 = %d\n", pdsch_vars[eNB_id]->log2_maxh1);
}
else {
avg_0[0] = (log2_approx(avg_0[0])/2) - 13 + interf_unaw_shift;
avg_1[0] = (log2_approx(avg_1[0])/2) - 13 + interf_unaw_shift;
pdsch_vars[eNB_id]->log2_maxh0 = cmax(avg_0[0],0);
pdsch_vars[eNB_id]->log2_maxh1 = cmax(avg_1[0],0);
- //printf("TM4 I-UA log2_maxh0 = %d\n", lte_ue_pdsch_vars[eNB_id]->log2_maxh0);
- //printf("TM4 I-UA log2_maxh1 = %d\n", lte_ue_pdsch_vars[eNB_id]->log2_maxh1);
+ //printf("TM4 I-UA log2_maxh0 = %d\n", pdsch_vars[eNB_id]->log2_maxh0);
+ //printf("TM4 I-UA log2_maxh1 = %d\n", pdsch_vars[eNB_id]->log2_maxh1);
}
}
else if (dlsch0_harq->mimo_mode<DUALSTREAM_UNIFORM_PRECODING1) {// single-layer precoding (TM5, TM6)
@@ -448,8 +459,11 @@ int rx_pdsch(PHY_VARS_UE *ue,
symbol,
nb_rb);
#ifdef DEBUG_PHY
- LOG_D(PHY,"[DLSCH] log2_maxh = %d (%d,%d)\n",pdsch_vars[eNB_id]->log2_maxh,avg[0],avgs);
- LOG_D(PHY,"[DLSCH] mimo_mode = %d\n", dlsch0_harq->mimo_mode);
+ LOG_I(PHY,"[DLSCH] log2_maxh = %d [log2_maxh0 %d log2_maxh1 %d] (%d,%d)\n",pdsch_vars[eNB_id]->log2_maxh,
+ pdsch_vars[eNB_id]->log2_maxh0,
+ pdsch_vars[eNB_id]->log2_maxh1,
+ avg[0],avgs);
+ LOG_I(PHY,"[DLSCH] mimo_mode = %d\n", dlsch0_harq->mimo_mode);
#endif
}
@@ -1023,6 +1037,44 @@ int rx_pdsch(PHY_VARS_UE *ue,
}
}
+// Please keep it: useful for debugging
+#if 0
+ if( (symbol == 13) && (dlsch0_harq->mimo_mode == 2) )
+ {
+ LOG_E(PHY,"Dump Phy Chan Est \n");
+ if(subframe&0x1)
+ {
+#if 1
+ //write_output("rxdataF0.m" , "rxdataF0", &common_vars->common_vars_rx_data_per_thread[subframe&0x1].rxdataF[0][0],14*frame_parms->ofdm_symbol_size,1,1);
+ //write_output("rxdataF1.m" , "rxdataF1", &common_vars->common_vars_rx_data_per_thread[subframe&0x1].rxdataF[0][0],14*frame_parms->ofdm_symbol_size,1,1);
+ //write_output("dl_ch_estimates00.m", "dl_ch_estimates00", &common_vars->common_vars_rx_data_per_thread[subframe&0x1].dl_ch_estimates[eNB_id][0][0],14*frame_parms->ofdm_symbol_size,1,1);
+ //write_output("dl_ch_estimates01.m", "dl_ch_estimates01", &common_vars->common_vars_rx_data_per_thread[subframe&0x1].dl_ch_estimates[eNB_id][1][0],14*frame_parms->ofdm_symbol_size,1,1);
+ //write_output("dl_ch_estimates10.m", "dl_ch_estimates10", &common_vars->common_vars_rx_data_per_thread[subframe&0x1].dl_ch_estimates[eNB_id][2][0],14*frame_parms->ofdm_symbol_size,1,1);
+ //write_output("dl_ch_estimates11.m", "dl_ch_estimates11", &common_vars->common_vars_rx_data_per_thread[subframe&0x1].dl_ch_estimates[eNB_id][3][0],14*frame_parms->ofdm_symbol_size,1,1);
+
+
+ //write_output("rxdataF_ext00.m" , "rxdataF_ext00", &pdsch_vars[eNB_id]->rxdataF_ext[0][0],14*frame_parms->N_RB_DL*12,1,1);
+ //write_output("rxdataF_ext01.m" , "rxdataF_ext01", &pdsch_vars[eNB_id]->rxdataF_ext[1][0],14*frame_parms->N_RB_DL*12,1,1);
+ //write_output("rxdataF_ext10.m" , "rxdataF_ext10", &pdsch_vars[eNB_id]->rxdataF_ext[2][0],14*frame_parms->N_RB_DL*12,1,1);
+ //write_output("rxdataF_ext11.m" , "rxdataF_ext11", &pdsch_vars[eNB_id]->rxdataF_ext[3][0],14*frame_parms->N_RB_DL*12,1,1);
+ write_output("dl_ch_estimates_ext00.m", "dl_ch_estimates_ext00", &pdsch_vars[eNB_id]->dl_ch_estimates_ext[0][0],14*frame_parms->N_RB_DL*12,1,1);
+ write_output("dl_ch_estimates_ext01.m", "dl_ch_estimates_ext01", &pdsch_vars[eNB_id]->dl_ch_estimates_ext[1][0],14*frame_parms->N_RB_DL*12,1,1);
+ write_output("dl_ch_estimates_ext10.m", "dl_ch_estimates_ext10", &pdsch_vars[eNB_id]->dl_ch_estimates_ext[2][0],14*frame_parms->N_RB_DL*12,1,1);
+ write_output("dl_ch_estimates_ext11.m", "dl_ch_estimates_ext11", &pdsch_vars[eNB_id]->dl_ch_estimates_ext[3][0],14*frame_parms->N_RB_DL*12,1,1);
+ write_output("rxdataF_comp00.m","rxdataF_comp00", &pdsch_vars[eNB_id]->rxdataF_comp0[0][0],14*frame_parms->N_RB_DL*12,1,1);
+ write_output("rxdataF_comp01.m","rxdataF_comp01", &pdsch_vars[eNB_id]->rxdataF_comp0[1][0],14*frame_parms->N_RB_DL*12,1,1);
+ write_output("rxdataF_comp10.m","rxdataF_comp10", &pdsch_vars[eNB_id]->rxdataF_comp1[harq_pid][round][0][0],14*frame_parms->N_RB_DL*12,1,1);
+ write_output("rxdataF_comp11.m","rxdataF_comp11", &pdsch_vars[eNB_id]->rxdataF_comp1[harq_pid][round][1][0],14*frame_parms->N_RB_DL*12,1,1);
+#endif
+ write_output("llr0.m","llr0", &pdsch_vars[eNB_id]->llr[0][0],(14*nb_rb*12*dlsch1_harq->Qm) - 4*(nb_rb*4*dlsch1_harq->Qm),1,0);
+ write_output("llr1.m","llr1", &pdsch_vars[eNB_id]->llr[1][0],(14*nb_rb*12*dlsch1_harq->Qm) - 4*(nb_rb*4*dlsch1_harq->Qm),1,0);
+
+
+ AssertFatal(0," ");
+ }
+
+ }
+#endif
#if T_TRACER
T(T_UE_PHY_PDSCH_IQ, T_INT(eNB_id), T_INT(ue->Mod_id), T_INT(frame%1024),
@@ -1609,9 +1661,7 @@ void prec2A_TM3_128(__m128i *ch0,__m128i *ch1) {
__m128i tmp0,tmp1;
- // sqrt(2) is already taken into account in computation sqrt_rho_a, sqrt_rho_b,
- //so divide by 2 is replaced by divide by sqrt(2).
-
+ //_mm_mulhi_epi16
// print_shorts("prec2A_TM3 ch0 (before):",ch0);
// print_shorts("prec2A_TM3 ch1 (before):",ch1);
@@ -1622,17 +1672,22 @@ void prec2A_TM3_128(__m128i *ch0,__m128i *ch1) {
ch0[0] = _mm_adds_epi16(ch0[0],tmp1);
ch1[0] = _mm_subs_epi16(tmp0,tmp1);
-
- // print_shorts("prec2A_TM3 ch0 (mid):",&tmp0);
- // print_shorts("prec2A_TM3 ch1 (mid):",ch1);
-
ch0[0] = _mm_mulhi_epi16(ch0[0],amp);
ch0[0] = _mm_slli_epi16(ch0[0],1);
+
ch1[0] = _mm_mulhi_epi16(ch1[0],amp);
ch1[0] = _mm_slli_epi16(ch1[0],1);
- // ch0[0] = _mm_srai_epi16(ch0[0],1);
- // ch1[0] = _mm_srai_epi16(ch1[0],1);
+ // print_shorts("prec2A_TM3 ch0 (mid):",&tmp0);
+ // print_shorts("prec2A_TM3 ch1 (mid):",ch1);
+
+ //ch0[0] = _mm_mulhi_epi16(ch0[0],amp);
+ //ch0[0] = _mm_slli_epi16(ch0[0],1);
+ //ch1[0] = _mm_mulhi_epi16(ch1[0],amp);
+ //ch1[0] = _mm_slli_epi16(ch1[0],1);
+
+ //ch0[0] = _mm_srai_epi16(ch0[0],1);
+ //ch1[0] = _mm_srai_epi16(ch1[0],1);
// print_shorts("prec2A_TM3 ch0 (after):",ch0);
// print_shorts("prec2A_TM3 ch1 (after):",ch1);
diff --git a/openair1/PHY/LTE_TRANSPORT/dlsch_llr_computation.c b/openair1/PHY/LTE_TRANSPORT/dlsch_llr_computation.c
index 0b73e2b02390a060e3774838e42fa4ab7cfa1533..b08836762cdbbf4111cc9f32b1460c891718d752 100644
--- a/openair1/PHY/LTE_TRANSPORT/dlsch_llr_computation.c
+++ b/openair1/PHY/LTE_TRANSPORT/dlsch_llr_computation.c
@@ -8831,6 +8831,37 @@ int dlsch_64qam_64qam_llr(LTE_DL_FRAME_PARMS *frame_parms,
len = (nb_rb*12) - pbch_pss_sss_adjust;
}
+#ifdef __AVX2__
+
+ // 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);
+
+ 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);
+
+ 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));
+
+#else
qam64_qam64((short *)rxF,
(short *)rxF_i,
(short *)ch_mag,
@@ -8838,6 +8869,7 @@ int dlsch_64qam_64qam_llr(LTE_DL_FRAME_PARMS *frame_parms,
(short *)llr16,
(short *)rho,
len);
+#endif
llr16 += (6*len);
*llr16p = (short *)llr16;
diff --git a/openair1/PHY/LTE_TRANSPORT/dlsch_llr_computation_avx2.c b/openair1/PHY/LTE_TRANSPORT/dlsch_llr_computation_avx2.c
new file mode 100644
index 0000000000000000000000000000000000000000..cda5ad0f55ee8afbe37bf4d3536cea17fb3ed575
--- /dev/null
+++ b/openair1/PHY/LTE_TRANSPORT/dlsch_llr_computation_avx2.c
@@ -0,0 +1,4034 @@
+ /*
+ * Licensed to the OpenAirInterface (OAI) Software Alliance under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The OpenAirInterface Software Alliance licenses this file to You under
+ * the OAI Public License, Version 1.0 (the "License"); you may not use this file
+ * except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * http://www.openairinterface.org/?page_id=698
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ *-------------------------------------------------------------------------------
+ * For more information about the OpenAirInterface (OAI) Software Alliance:
+ * contact@openairinterface.org
+ */
+
+/*! \file PHY/LTE_TRANSPORT/dlsch_llr_computation.c
+ * \brief Top-level routines for LLR computation of the PDSCH physical channel from 36-211, V8.6 2009-03
+ * \author R. Knopp, F. Kaltenberger,A. Bhamri, S. Aubert, S. Wagner, X Jiang
+ * \date 2011
+ * \version 0.1
+ * \company Eurecom
+ * \email: knopp@eurecom.fr,florian.kaltenberger@eurecom.fr,ankit.bhamri@eurecom.fr,sebastien.aubert@eurecom.fr, sebastian.wagner@eurecom.fr
+ * \note
+ * \warning
+ */
+
+#include "PHY/defs.h"
+#include "PHY/TOOLS/defs.h"
+#include "PHY/extern.h"
+#include "defs.h"
+#include "extern.h"
+#include "PHY/sse_intrin.h"
+
+int16_t ones256[16] __attribute__ ((aligned(32))) = {0xffff,0xffff,0xffff,0xffff,0xffff,0xffff,0xffff,0xffff,0xffff,0xffff,0xffff,0xffff,0xffff,0xffff,0xffff,0xffff};
+
+static __m256i rho_rpi __attribute__ ((aligned(32)));
+static __m256i rho_rmi __attribute__ ((aligned(32)));
+static __m256i rho_rpi_1_1 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_1_3 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_1_5 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_1_7 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_3_1 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_3_3 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_3_5 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_3_7 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_5_1 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_5_3 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_5_5 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_5_7 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_7_1 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_7_3 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_7_5 __attribute__ ((aligned(32)));
+static __m256i rho_rpi_7_7 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_1_1 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_1_3 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_1_5 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_1_7 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_3_1 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_3_3 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_3_5 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_3_7 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_5_1 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_5_3 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_5_5 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_5_7 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_7_1 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_7_3 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_7_5 __attribute__ ((aligned(32)));
+static __m256i rho_rmi_7_7 __attribute__ ((aligned(32)));
+
+static __m256i psi_r_m7_m7 __attribute__ ((aligned(32)));
+static __m256i psi_r_m7_m5 __attribute__ ((aligned(32)));
+static __m256i psi_r_m7_m3 __attribute__ ((aligned(32)));
+static __m256i psi_r_m7_m1 __attribute__ ((aligned(32)));
+static __m256i psi_r_m7_p1 __attribute__ ((aligned(32)));
+static __m256i psi_r_m7_p3 __attribute__ ((aligned(32)));
+static __m256i psi_r_m7_p5 __attribute__ ((aligned(32)));
+static __m256i psi_r_m7_p7 __attribute__ ((aligned(32)));
+static __m256i psi_r_m5_m7 __attribute__ ((aligned(32)));
+static __m256i psi_r_m5_m5 __attribute__ ((aligned(32)));
+static __m256i psi_r_m5_m3 __attribute__ ((aligned(32)));
+static __m256i psi_r_m5_m1 __attribute__ ((aligned(32)));
+static __m256i psi_r_m5_p1 __attribute__ ((aligned(32)));
+static __m256i psi_r_m5_p3 __attribute__ ((aligned(32)));
+static __m256i psi_r_m5_p5 __attribute__ ((aligned(32)));
+static __m256i psi_r_m5_p7 __attribute__ ((aligned(32)));
+static __m256i psi_r_m3_m7 __attribute__ ((aligned(32)));
+static __m256i psi_r_m3_m5 __attribute__ ((aligned(32)));
+static __m256i psi_r_m3_m3 __attribute__ ((aligned(32)));
+static __m256i psi_r_m3_m1 __attribute__ ((aligned(32)));
+static __m256i psi_r_m3_p1 __attribute__ ((aligned(32)));
+static __m256i psi_r_m3_p3 __attribute__ ((aligned(32)));
+static __m256i psi_r_m3_p5 __attribute__ ((aligned(32)));
+static __m256i psi_r_m3_p7 __attribute__ ((aligned(32)));
+static __m256i psi_r_m1_m7 __attribute__ ((aligned(32)));
+static __m256i psi_r_m1_m5 __attribute__ ((aligned(32)));
+static __m256i psi_r_m1_m3 __attribute__ ((aligned(32)));
+static __m256i psi_r_m1_m1 __attribute__ ((aligned(32)));
+static __m256i psi_r_m1_p1 __attribute__ ((aligned(32)));
+static __m256i psi_r_m1_p3 __attribute__ ((aligned(32)));
+static __m256i psi_r_m1_p5 __attribute__ ((aligned(32)));
+static __m256i psi_r_m1_p7 __attribute__ ((aligned(32)));
+static __m256i psi_r_p1_m7 __attribute__ ((aligned(32)));
+static __m256i psi_r_p1_m5 __attribute__ ((aligned(32)));
+static __m256i psi_r_p1_m3 __attribute__ ((aligned(32)));
+static __m256i psi_r_p1_m1 __attribute__ ((aligned(32)));
+static __m256i psi_r_p1_p1 __attribute__ ((aligned(32)));
+static __m256i psi_r_p1_p3 __attribute__ ((aligned(32)));
+static __m256i psi_r_p1_p5 __attribute__ ((aligned(32)));
+static __m256i psi_r_p1_p7 __attribute__ ((aligned(32)));
+static __m256i psi_r_p3_m7 __attribute__ ((aligned(32)));
+static __m256i psi_r_p3_m5 __attribute__ ((aligned(32)));
+static __m256i psi_r_p3_m3 __attribute__ ((aligned(32)));
+static __m256i psi_r_p3_m1 __attribute__ ((aligned(32)));
+static __m256i psi_r_p3_p1 __attribute__ ((aligned(32)));
+static __m256i psi_r_p3_p3 __attribute__ ((aligned(32)));
+static __m256i psi_r_p3_p5 __attribute__ ((aligned(32)));
+static __m256i psi_r_p3_p7 __attribute__ ((aligned(32)));
+static __m256i psi_r_p5_m7 __attribute__ ((aligned(32)));
+static __m256i psi_r_p5_m5 __attribute__ ((aligned(32)));
+static __m256i psi_r_p5_m3 __attribute__ ((aligned(32)));
+static __m256i psi_r_p5_m1 __attribute__ ((aligned(32)));
+static __m256i psi_r_p5_p1 __attribute__ ((aligned(32)));
+static __m256i psi_r_p5_p3 __attribute__ ((aligned(32)));
+static __m256i psi_r_p5_p5 __attribute__ ((aligned(32)));
+static __m256i psi_r_p5_p7 __attribute__ ((aligned(32)));
+static __m256i psi_r_p7_m7 __attribute__ ((aligned(32)));
+static __m256i psi_r_p7_m5 __attribute__ ((aligned(32)));
+static __m256i psi_r_p7_m3 __attribute__ ((aligned(32)));
+static __m256i psi_r_p7_m1 __attribute__ ((aligned(32)));
+static __m256i psi_r_p7_p1 __attribute__ ((aligned(32)));
+static __m256i psi_r_p7_p3 __attribute__ ((aligned(32)));
+static __m256i psi_r_p7_p5 __attribute__ ((aligned(32)));
+static __m256i psi_r_p7_p7 __attribute__ ((aligned(32)));
+
+static __m256i psi_i_m7_m7 __attribute__ ((aligned(32)));
+static __m256i psi_i_m7_m5 __attribute__ ((aligned(32)));
+static __m256i psi_i_m7_m3 __attribute__ ((aligned(32)));
+static __m256i psi_i_m7_m1 __attribute__ ((aligned(32)));
+static __m256i psi_i_m7_p1 __attribute__ ((aligned(32)));
+static __m256i psi_i_m7_p3 __attribute__ ((aligned(32)));
+static __m256i psi_i_m7_p5 __attribute__ ((aligned(32)));
+static __m256i psi_i_m7_p7 __attribute__ ((aligned(32)));
+static __m256i psi_i_m5_m7 __attribute__ ((aligned(32)));
+static __m256i psi_i_m5_m5 __attribute__ ((aligned(32)));
+static __m256i psi_i_m5_m3 __attribute__ ((aligned(32)));
+static __m256i psi_i_m5_m1 __attribute__ ((aligned(32)));
+static __m256i psi_i_m5_p1 __attribute__ ((aligned(32)));
+static __m256i psi_i_m5_p3 __attribute__ ((aligned(32)));
+static __m256i psi_i_m5_p5 __attribute__ ((aligned(32)));
+static __m256i psi_i_m5_p7 __attribute__ ((aligned(32)));
+static __m256i psi_i_m3_m7 __attribute__ ((aligned(32)));
+static __m256i psi_i_m3_m5 __attribute__ ((aligned(32)));
+static __m256i psi_i_m3_m3 __attribute__ ((aligned(32)));
+static __m256i psi_i_m3_m1 __attribute__ ((aligned(32)));
+static __m256i psi_i_m3_p1 __attribute__ ((aligned(32)));
+static __m256i psi_i_m3_p3 __attribute__ ((aligned(32)));
+static __m256i psi_i_m3_p5 __attribute__ ((aligned(32)));
+static __m256i psi_i_m3_p7 __attribute__ ((aligned(32)));
+static __m256i psi_i_m1_m7 __attribute__ ((aligned(32)));
+static __m256i psi_i_m1_m5 __attribute__ ((aligned(32)));
+static __m256i psi_i_m1_m3 __attribute__ ((aligned(32)));
+static __m256i psi_i_m1_m1 __attribute__ ((aligned(32)));
+static __m256i psi_i_m1_p1 __attribute__ ((aligned(32)));
+static __m256i psi_i_m1_p3 __attribute__ ((aligned(32)));
+static __m256i psi_i_m1_p5 __attribute__ ((aligned(32)));
+static __m256i psi_i_m1_p7 __attribute__ ((aligned(32)));
+static __m256i psi_i_p1_m7 __attribute__ ((aligned(32)));
+static __m256i psi_i_p1_m5 __attribute__ ((aligned(32)));
+static __m256i psi_i_p1_m3 __attribute__ ((aligned(32)));
+static __m256i psi_i_p1_m1 __attribute__ ((aligned(32)));
+static __m256i psi_i_p1_p1 __attribute__ ((aligned(32)));
+static __m256i psi_i_p1_p3 __attribute__ ((aligned(32)));
+static __m256i psi_i_p1_p5 __attribute__ ((aligned(32)));
+static __m256i psi_i_p1_p7 __attribute__ ((aligned(32)));
+static __m256i psi_i_p3_m7 __attribute__ ((aligned(32)));
+static __m256i psi_i_p3_m5 __attribute__ ((aligned(32)));
+static __m256i psi_i_p3_m3 __attribute__ ((aligned(32)));
+static __m256i psi_i_p3_m1 __attribute__ ((aligned(32)));
+static __m256i psi_i_p3_p1 __attribute__ ((aligned(32)));
+static __m256i psi_i_p3_p3 __attribute__ ((aligned(32)));
+static __m256i psi_i_p3_p5 __attribute__ ((aligned(32)));
+static __m256i psi_i_p3_p7 __attribute__ ((aligned(32)));
+static __m256i psi_i_p5_m7 __attribute__ ((aligned(32)));
+static __m256i psi_i_p5_m5 __attribute__ ((aligned(32)));
+static __m256i psi_i_p5_m3 __attribute__ ((aligned(32)));
+static __m256i psi_i_p5_m1 __attribute__ ((aligned(32)));
+static __m256i psi_i_p5_p1 __attribute__ ((aligned(32)));
+static __m256i psi_i_p5_p3 __attribute__ ((aligned(32)));
+static __m256i psi_i_p5_p5 __attribute__ ((aligned(32)));
+static __m256i psi_i_p5_p7 __attribute__ ((aligned(32)));
+static __m256i psi_i_p7_m7 __attribute__ ((aligned(32)));
+static __m256i psi_i_p7_m5 __attribute__ ((aligned(32)));
+static __m256i psi_i_p7_m3 __attribute__ ((aligned(32)));
+static __m256i psi_i_p7_m1 __attribute__ ((aligned(32)));
+static __m256i psi_i_p7_p1 __attribute__ ((aligned(32)));
+static __m256i psi_i_p7_p3 __attribute__ ((aligned(32)));
+static __m256i psi_i_p7_p5 __attribute__ ((aligned(32)));
+static __m256i psi_i_p7_p7 __attribute__ ((aligned(32)));
+
+static __m256i a_r_m7_m7 __attribute__ ((aligned(32)));
+static __m256i a_r_m7_m5 __attribute__ ((aligned(32)));
+static __m256i a_r_m7_m3 __attribute__ ((aligned(32)));
+static __m256i a_r_m7_m1 __attribute__ ((aligned(32)));
+static __m256i a_r_m7_p1 __attribute__ ((aligned(32)));
+static __m256i a_r_m7_p3 __attribute__ ((aligned(32)));
+static __m256i a_r_m7_p5 __attribute__ ((aligned(32)));
+static __m256i a_r_m7_p7 __attribute__ ((aligned(32)));
+static __m256i a_r_m5_m7 __attribute__ ((aligned(32)));
+static __m256i a_r_m5_m5 __attribute__ ((aligned(32)));
+static __m256i a_r_m5_m3 __attribute__ ((aligned(32)));
+static __m256i a_r_m5_m1 __attribute__ ((aligned(32)));
+static __m256i a_r_m5_p1 __attribute__ ((aligned(32)));
+static __m256i a_r_m5_p3 __attribute__ ((aligned(32)));
+static __m256i a_r_m5_p5 __attribute__ ((aligned(32)));
+static __m256i a_r_m5_p7 __attribute__ ((aligned(32)));
+static __m256i a_r_m3_m7 __attribute__ ((aligned(32)));
+static __m256i a_r_m3_m5 __attribute__ ((aligned(32)));
+static __m256i a_r_m3_m3 __attribute__ ((aligned(32)));
+static __m256i a_r_m3_m1 __attribute__ ((aligned(32)));
+static __m256i a_r_m3_p1 __attribute__ ((aligned(32)));
+static __m256i a_r_m3_p3 __attribute__ ((aligned(32)));
+static __m256i a_r_m3_p5 __attribute__ ((aligned(32)));
+static __m256i a_r_m3_p7 __attribute__ ((aligned(32)));
+static __m256i a_r_m1_m7 __attribute__ ((aligned(32)));
+static __m256i a_r_m1_m5 __attribute__ ((aligned(32)));
+static __m256i a_r_m1_m3 __attribute__ ((aligned(32)));
+static __m256i a_r_m1_m1 __attribute__ ((aligned(32)));
+static __m256i a_r_m1_p1 __attribute__ ((aligned(32)));
+static __m256i a_r_m1_p3 __attribute__ ((aligned(32)));
+static __m256i a_r_m1_p5 __attribute__ ((aligned(32)));
+static __m256i a_r_m1_p7 __attribute__ ((aligned(32)));
+static __m256i a_r_p1_m7 __attribute__ ((aligned(32)));
+static __m256i a_r_p1_m5 __attribute__ ((aligned(32)));
+static __m256i a_r_p1_m3 __attribute__ ((aligned(32)));
+static __m256i a_r_p1_m1 __attribute__ ((aligned(32)));
+static __m256i a_r_p1_p1 __attribute__ ((aligned(32)));
+static __m256i a_r_p1_p3 __attribute__ ((aligned(32)));
+static __m256i a_r_p1_p5 __attribute__ ((aligned(32)));
+static __m256i a_r_p1_p7 __attribute__ ((aligned(32)));
+static __m256i a_r_p3_m7 __attribute__ ((aligned(32)));
+static __m256i a_r_p3_m5 __attribute__ ((aligned(32)));
+static __m256i a_r_p3_m3 __attribute__ ((aligned(32)));
+static __m256i a_r_p3_m1 __attribute__ ((aligned(32)));
+static __m256i a_r_p3_p1 __attribute__ ((aligned(32)));
+static __m256i a_r_p3_p3 __attribute__ ((aligned(32)));
+static __m256i a_r_p3_p5 __attribute__ ((aligned(32)));
+static __m256i a_r_p3_p7 __attribute__ ((aligned(32)));
+static __m256i a_r_p5_m7 __attribute__ ((aligned(32)));
+static __m256i a_r_p5_m5 __attribute__ ((aligned(32)));
+static __m256i a_r_p5_m3 __attribute__ ((aligned(32)));
+static __m256i a_r_p5_m1 __attribute__ ((aligned(32)));
+static __m256i a_r_p5_p1 __attribute__ ((aligned(32)));
+static __m256i a_r_p5_p3 __attribute__ ((aligned(32)));
+static __m256i a_r_p5_p5 __attribute__ ((aligned(32)));
+static __m256i a_r_p5_p7 __attribute__ ((aligned(32)));
+static __m256i a_r_p7_m7 __attribute__ ((aligned(32)));
+static __m256i a_r_p7_m5 __attribute__ ((aligned(32)));
+static __m256i a_r_p7_m3 __attribute__ ((aligned(32)));
+static __m256i a_r_p7_m1 __attribute__ ((aligned(32)));
+static __m256i a_r_p7_p1 __attribute__ ((aligned(32)));
+static __m256i a_r_p7_p3 __attribute__ ((aligned(32)));
+static __m256i a_r_p7_p5 __attribute__ ((aligned(32)));
+static __m256i a_r_p7_p7 __attribute__ ((aligned(32)));
+
+static __m256i a_i_m7_m7 __attribute__ ((aligned(32)));
+static __m256i a_i_m7_m5 __attribute__ ((aligned(32)));
+static __m256i a_i_m7_m3 __attribute__ ((aligned(32)));
+static __m256i a_i_m7_m1 __attribute__ ((aligned(32)));
+static __m256i a_i_m7_p1 __attribute__ ((aligned(32)));
+static __m256i a_i_m7_p3 __attribute__ ((aligned(32)));
+static __m256i a_i_m7_p5 __attribute__ ((aligned(32)));
+static __m256i a_i_m7_p7 __attribute__ ((aligned(32)));
+static __m256i a_i_m5_m7 __attribute__ ((aligned(32)));
+static __m256i a_i_m5_m5 __attribute__ ((aligned(32)));
+static __m256i a_i_m5_m3 __attribute__ ((aligned(32)));
+static __m256i a_i_m5_m1 __attribute__ ((aligned(32)));
+static __m256i a_i_m5_p1 __attribute__ ((aligned(32)));
+static __m256i a_i_m5_p3 __attribute__ ((aligned(32)));
+static __m256i a_i_m5_p5 __attribute__ ((aligned(32)));
+static __m256i a_i_m5_p7 __attribute__ ((aligned(32)));
+static __m256i a_i_m3_m7 __attribute__ ((aligned(32)));
+static __m256i a_i_m3_m5 __attribute__ ((aligned(32)));
+static __m256i a_i_m3_m3 __attribute__ ((aligned(32)));
+static __m256i a_i_m3_m1 __attribute__ ((aligned(32)));
+static __m256i a_i_m3_p1 __attribute__ ((aligned(32)));
+static __m256i a_i_m3_p3 __attribute__ ((aligned(32)));
+static __m256i a_i_m3_p5 __attribute__ ((aligned(32)));
+static __m256i a_i_m3_p7 __attribute__ ((aligned(32)));
+static __m256i a_i_m1_m7 __attribute__ ((aligned(32)));
+static __m256i a_i_m1_m5 __attribute__ ((aligned(32)));
+static __m256i a_i_m1_m3 __attribute__ ((aligned(32)));
+static __m256i a_i_m1_m1 __attribute__ ((aligned(32)));
+static __m256i a_i_m1_p1 __attribute__ ((aligned(32)));
+static __m256i a_i_m1_p3 __attribute__ ((aligned(32)));
+static __m256i a_i_m1_p5 __attribute__ ((aligned(32)));
+static __m256i a_i_m1_p7 __attribute__ ((aligned(32)));
+static __m256i a_i_p1_m7 __attribute__ ((aligned(32)));
+static __m256i a_i_p1_m5 __attribute__ ((aligned(32)));
+static __m256i a_i_p1_m3 __attribute__ ((aligned(32)));
+static __m256i a_i_p1_m1 __attribute__ ((aligned(32)));
+static __m256i a_i_p1_p1 __attribute__ ((aligned(32)));
+static __m256i a_i_p1_p3 __attribute__ ((aligned(32)));
+static __m256i a_i_p1_p5 __attribute__ ((aligned(32)));
+static __m256i a_i_p1_p7 __attribute__ ((aligned(32)));
+static __m256i a_i_p3_m7 __attribute__ ((aligned(32)));
+static __m256i a_i_p3_m5 __attribute__ ((aligned(32)));
+static __m256i a_i_p3_m3 __attribute__ ((aligned(32)));
+static __m256i a_i_p3_m1 __attribute__ ((aligned(32)));
+static __m256i a_i_p3_p1 __attribute__ ((aligned(32)));
+static __m256i a_i_p3_p3 __attribute__ ((aligned(32)));
+static __m256i a_i_p3_p5 __attribute__ ((aligned(32)));
+static __m256i a_i_p3_p7 __attribute__ ((aligned(32)));
+static __m256i a_i_p5_m7 __attribute__ ((aligned(32)));
+static __m256i a_i_p5_m5 __attribute__ ((aligned(32)));
+static __m256i a_i_p5_m3 __attribute__ ((aligned(32)));
+static __m256i a_i_p5_m1 __attribute__ ((aligned(32)));
+static __m256i a_i_p5_p1 __attribute__ ((aligned(32)));
+static __m256i a_i_p5_p3 __attribute__ ((aligned(32)));
+static __m256i a_i_p5_p5 __attribute__ ((aligned(32)));
+static __m256i a_i_p5_p7 __attribute__ ((aligned(32)));
+static __m256i a_i_p7_m7 __attribute__ ((aligned(32)));
+static __m256i a_i_p7_m5 __attribute__ ((aligned(32)));
+static __m256i a_i_p7_m3 __attribute__ ((aligned(32)));
+static __m256i a_i_p7_m1 __attribute__ ((aligned(32)));
+static __m256i a_i_p7_p1 __attribute__ ((aligned(32)));
+static __m256i a_i_p7_p3 __attribute__ ((aligned(32)));
+static __m256i a_i_p7_p5 __attribute__ ((aligned(32)));
+static __m256i a_i_p7_p7 __attribute__ ((aligned(32)));
+
+static __m256i psi_a_m7_m7 __attribute__ ((aligned(32)));
+static __m256i psi_a_m7_m5 __attribute__ ((aligned(32)));
+static __m256i psi_a_m7_m3 __attribute__ ((aligned(32)));
+static __m256i psi_a_m7_m1 __attribute__ ((aligned(32)));
+static __m256i psi_a_m7_p1 __attribute__ ((aligned(32)));
+static __m256i psi_a_m7_p3 __attribute__ ((aligned(32)));
+static __m256i psi_a_m7_p5 __attribute__ ((aligned(32)));
+static __m256i psi_a_m7_p7 __attribute__ ((aligned(32)));
+static __m256i psi_a_m5_m7 __attribute__ ((aligned(32)));
+static __m256i psi_a_m5_m5 __attribute__ ((aligned(32)));
+static __m256i psi_a_m5_m3 __attribute__ ((aligned(32)));
+static __m256i psi_a_m5_m1 __attribute__ ((aligned(32)));
+static __m256i psi_a_m5_p1 __attribute__ ((aligned(32)));
+static __m256i psi_a_m5_p3 __attribute__ ((aligned(32)));
+static __m256i psi_a_m5_p5 __attribute__ ((aligned(32)));
+static __m256i psi_a_m5_p7 __attribute__ ((aligned(32)));
+static __m256i psi_a_m3_m7 __attribute__ ((aligned(32)));
+static __m256i psi_a_m3_m5 __attribute__ ((aligned(32)));
+static __m256i psi_a_m3_m3 __attribute__ ((aligned(32)));
+static __m256i psi_a_m3_m1 __attribute__ ((aligned(32)));
+static __m256i psi_a_m3_p1 __attribute__ ((aligned(32)));
+static __m256i psi_a_m3_p3 __attribute__ ((aligned(32)));
+static __m256i psi_a_m3_p5 __attribute__ ((aligned(32)));
+static __m256i psi_a_m3_p7 __attribute__ ((aligned(32)));
+static __m256i psi_a_m1_m7 __attribute__ ((aligned(32)));
+static __m256i psi_a_m1_m5 __attribute__ ((aligned(32)));
+static __m256i psi_a_m1_m3 __attribute__ ((aligned(32)));
+static __m256i psi_a_m1_m1 __attribute__ ((aligned(32)));
+static __m256i psi_a_m1_p1 __attribute__ ((aligned(32)));
+static __m256i psi_a_m1_p3 __attribute__ ((aligned(32)));
+static __m256i psi_a_m1_p5 __attribute__ ((aligned(32)));
+static __m256i psi_a_m1_p7 __attribute__ ((aligned(32)));
+static __m256i psi_a_p1_m7 __attribute__ ((aligned(32)));
+static __m256i psi_a_p1_m5 __attribute__ ((aligned(32)));
+static __m256i psi_a_p1_m3 __attribute__ ((aligned(32)));
+static __m256i psi_a_p1_m1 __attribute__ ((aligned(32)));
+static __m256i psi_a_p1_p1 __attribute__ ((aligned(32)));
+static __m256i psi_a_p1_p3 __attribute__ ((aligned(32)));
+static __m256i psi_a_p1_p5 __attribute__ ((aligned(32)));
+static __m256i psi_a_p1_p7 __attribute__ ((aligned(32)));
+static __m256i psi_a_p3_m7 __attribute__ ((aligned(32)));
+static __m256i psi_a_p3_m5 __attribute__ ((aligned(32)));
+static __m256i psi_a_p3_m3 __attribute__ ((aligned(32)));
+static __m256i psi_a_p3_m1 __attribute__ ((aligned(32)));
+static __m256i psi_a_p3_p1 __attribute__ ((aligned(32)));
+static __m256i psi_a_p3_p3 __attribute__ ((aligned(32)));
+static __m256i psi_a_p3_p5 __attribute__ ((aligned(32)));
+static __m256i psi_a_p3_p7 __attribute__ ((aligned(32)));
+static __m256i psi_a_p5_m7 __attribute__ ((aligned(32)));
+static __m256i psi_a_p5_m5 __attribute__ ((aligned(32)));
+static __m256i psi_a_p5_m3 __attribute__ ((aligned(32)));
+static __m256i psi_a_p5_m1 __attribute__ ((aligned(32)));
+static __m256i psi_a_p5_p1 __attribute__ ((aligned(32)));
+static __m256i psi_a_p5_p3 __attribute__ ((aligned(32)));
+static __m256i psi_a_p5_p5 __attribute__ ((aligned(32)));
+static __m256i psi_a_p5_p7 __attribute__ ((aligned(32)));
+static __m256i psi_a_p7_m7 __attribute__ ((aligned(32)));
+static __m256i psi_a_p7_m5 __attribute__ ((aligned(32)));
+static __m256i psi_a_p7_m3 __attribute__ ((aligned(32)));
+static __m256i psi_a_p7_m1 __attribute__ ((aligned(32)));
+static __m256i psi_a_p7_p1 __attribute__ ((aligned(32)));
+static __m256i psi_a_p7_p3 __attribute__ ((aligned(32)));
+static __m256i psi_a_p7_p5 __attribute__ ((aligned(32)));
+static __m256i psi_a_p7_p7 __attribute__ ((aligned(32)));
+
+static __m256i a_sq_m7_m7 __attribute__ ((aligned(32)));
+static __m256i a_sq_m7_m5 __attribute__ ((aligned(32)));
+static __m256i a_sq_m7_m3 __attribute__ ((aligned(32)));
+static __m256i a_sq_m7_m1 __attribute__ ((aligned(32)));
+static __m256i a_sq_m7_p1 __attribute__ ((aligned(32)));
+static __m256i a_sq_m7_p3 __attribute__ ((aligned(32)));
+static __m256i a_sq_m7_p5 __attribute__ ((aligned(32)));
+static __m256i a_sq_m7_p7 __attribute__ ((aligned(32)));
+static __m256i a_sq_m5_m7 __attribute__ ((aligned(32)));
+static __m256i a_sq_m5_m5 __attribute__ ((aligned(32)));
+static __m256i a_sq_m5_m3 __attribute__ ((aligned(32)));
+static __m256i a_sq_m5_m1 __attribute__ ((aligned(32)));
+static __m256i a_sq_m5_p1 __attribute__ ((aligned(32)));
+static __m256i a_sq_m5_p3 __attribute__ ((aligned(32)));
+static __m256i a_sq_m5_p5 __attribute__ ((aligned(32)));
+static __m256i a_sq_m5_p7 __attribute__ ((aligned(32)));
+static __m256i a_sq_m3_m7 __attribute__ ((aligned(32)));
+static __m256i a_sq_m3_m5 __attribute__ ((aligned(32)));
+static __m256i a_sq_m3_m3 __attribute__ ((aligned(32)));
+static __m256i a_sq_m3_m1 __attribute__ ((aligned(32)));
+static __m256i a_sq_m3_p1 __attribute__ ((aligned(32)));
+static __m256i a_sq_m3_p3 __attribute__ ((aligned(32)));
+static __m256i a_sq_m3_p5 __attribute__ ((aligned(32)));
+static __m256i a_sq_m3_p7 __attribute__ ((aligned(32)));
+static __m256i a_sq_m1_m7 __attribute__ ((aligned(32)));
+static __m256i a_sq_m1_m5 __attribute__ ((aligned(32)));
+static __m256i a_sq_m1_m3 __attribute__ ((aligned(32)));
+static __m256i a_sq_m1_m1 __attribute__ ((aligned(32)));
+static __m256i a_sq_m1_p1 __attribute__ ((aligned(32)));
+static __m256i a_sq_m1_p3 __attribute__ ((aligned(32)));
+static __m256i a_sq_m1_p5 __attribute__ ((aligned(32)));
+static __m256i a_sq_m1_p7 __attribute__ ((aligned(32)));
+static __m256i a_sq_p1_m7 __attribute__ ((aligned(32)));
+static __m256i a_sq_p1_m5 __attribute__ ((aligned(32)));
+static __m256i a_sq_p1_m3 __attribute__ ((aligned(32)));
+static __m256i a_sq_p1_m1 __attribute__ ((aligned(32)));
+static __m256i a_sq_p1_p1 __attribute__ ((aligned(32)));
+static __m256i a_sq_p1_p3 __attribute__ ((aligned(32)));
+static __m256i a_sq_p1_p5 __attribute__ ((aligned(32)));
+static __m256i a_sq_p1_p7 __attribute__ ((aligned(32)));
+static __m256i a_sq_p3_m7 __attribute__ ((aligned(32)));
+static __m256i a_sq_p3_m5 __attribute__ ((aligned(32)));
+static __m256i a_sq_p3_m3 __attribute__ ((aligned(32)));
+static __m256i a_sq_p3_m1 __attribute__ ((aligned(32)));
+static __m256i a_sq_p3_p1 __attribute__ ((aligned(32)));
+static __m256i a_sq_p3_p3 __attribute__ ((aligned(32)));
+static __m256i a_sq_p3_p5 __attribute__ ((aligned(32)));
+static __m256i a_sq_p3_p7 __attribute__ ((aligned(32)));
+static __m256i a_sq_p5_m7 __attribute__ ((aligned(32)));
+static __m256i a_sq_p5_m5 __attribute__ ((aligned(32)));
+static __m256i a_sq_p5_m3 __attribute__ ((aligned(32)));
+static __m256i a_sq_p5_m1 __attribute__ ((aligned(32)));
+static __m256i a_sq_p5_p1 __attribute__ ((aligned(32)));
+static __m256i a_sq_p5_p3 __attribute__ ((aligned(32)));
+static __m256i a_sq_p5_p5 __attribute__ ((aligned(32)));
+static __m256i a_sq_p5_p7 __attribute__ ((aligned(32)));
+static __m256i a_sq_p7_m7 __attribute__ ((aligned(32)));
+static __m256i a_sq_p7_m5 __attribute__ ((aligned(32)));
+static __m256i a_sq_p7_m3 __attribute__ ((aligned(32)));
+static __m256i a_sq_p7_m1 __attribute__ ((aligned(32)));
+static __m256i a_sq_p7_p1 __attribute__ ((aligned(32)));
+static __m256i a_sq_p7_p3 __attribute__ ((aligned(32)));
+static __m256i a_sq_p7_p5 __attribute__ ((aligned(32)));
+static __m256i a_sq_p7_p7 __attribute__ ((aligned(32)));
+
+static __m256i bit_met_m7_m7 __attribute__ ((aligned(32)));
+static __m256i bit_met_m7_m5 __attribute__ ((aligned(32)));
+static __m256i bit_met_m7_m3 __attribute__ ((aligned(32)));
+static __m256i bit_met_m7_m1 __attribute__ ((aligned(32)));
+static __m256i bit_met_m7_p1 __attribute__ ((aligned(32)));
+static __m256i bit_met_m7_p3 __attribute__ ((aligned(32)));
+static __m256i bit_met_m7_p5 __attribute__ ((aligned(32)));
+static __m256i bit_met_m7_p7 __attribute__ ((aligned(32)));
+static __m256i bit_met_m5_m7 __attribute__ ((aligned(32)));
+static __m256i bit_met_m5_m5 __attribute__ ((aligned(32)));
+static __m256i bit_met_m5_m3 __attribute__ ((aligned(32)));
+static __m256i bit_met_m5_m1 __attribute__ ((aligned(32)));
+static __m256i bit_met_m5_p1 __attribute__ ((aligned(32)));
+static __m256i bit_met_m5_p3 __attribute__ ((aligned(32)));
+static __m256i bit_met_m5_p5 __attribute__ ((aligned(32)));
+static __m256i bit_met_m5_p7 __attribute__ ((aligned(32)));
+static __m256i bit_met_m3_m7 __attribute__ ((aligned(32)));
+static __m256i bit_met_m3_m5 __attribute__ ((aligned(32)));
+static __m256i bit_met_m3_m3 __attribute__ ((aligned(32)));
+static __m256i bit_met_m3_m1 __attribute__ ((aligned(32)));
+static __m256i bit_met_m3_p1 __attribute__ ((aligned(32)));
+static __m256i bit_met_m3_p3 __attribute__ ((aligned(32)));
+static __m256i bit_met_m3_p5 __attribute__ ((aligned(32)));
+static __m256i bit_met_m3_p7 __attribute__ ((aligned(32)));
+static __m256i bit_met_m1_m7 __attribute__ ((aligned(32)));
+static __m256i bit_met_m1_m5 __attribute__ ((aligned(32)));
+static __m256i bit_met_m1_m3 __attribute__ ((aligned(32)));
+static __m256i bit_met_m1_m1 __attribute__ ((aligned(32)));
+static __m256i bit_met_m1_p1 __attribute__ ((aligned(32)));
+static __m256i bit_met_m1_p3 __attribute__ ((aligned(32)));
+static __m256i bit_met_m1_p5 __attribute__ ((aligned(32)));
+static __m256i bit_met_m1_p7 __attribute__ ((aligned(32)));
+static __m256i bit_met_p1_m7 __attribute__ ((aligned(32)));
+static __m256i bit_met_p1_m5 __attribute__ ((aligned(32)));
+static __m256i bit_met_p1_m3 __attribute__ ((aligned(32)));
+static __m256i bit_met_p1_m1 __attribute__ ((aligned(32)));
+static __m256i bit_met_p1_p1 __attribute__ ((aligned(32)));
+static __m256i bit_met_p1_p3 __attribute__ ((aligned(32)));
+static __m256i bit_met_p1_p5 __attribute__ ((aligned(32)));
+static __m256i bit_met_p1_p7 __attribute__ ((aligned(32)));
+static __m256i bit_met_p3_m7 __attribute__ ((aligned(32)));
+static __m256i bit_met_p3_m5 __attribute__ ((aligned(32)));
+static __m256i bit_met_p3_m3 __attribute__ ((aligned(32)));
+static __m256i bit_met_p3_m1 __attribute__ ((aligned(32)));
+static __m256i bit_met_p3_p1 __attribute__ ((aligned(32)));
+static __m256i bit_met_p3_p3 __attribute__ ((aligned(32)));
+static __m256i bit_met_p3_p5 __attribute__ ((aligned(32)));
+static __m256i bit_met_p3_p7 __attribute__ ((aligned(32)));
+static __m256i bit_met_p5_m7 __attribute__ ((aligned(32)));
+static __m256i bit_met_p5_m5 __attribute__ ((aligned(32)));
+static __m256i bit_met_p5_m3 __attribute__ ((aligned(32)));
+static __m256i bit_met_p5_m1 __attribute__ ((aligned(32)));
+static __m256i bit_met_p5_p1 __attribute__ ((aligned(32)));
+static __m256i bit_met_p5_p3 __attribute__ ((aligned(32)));
+static __m256i bit_met_p5_p5 __attribute__ ((aligned(32)));
+static __m256i bit_met_p5_p7 __attribute__ ((aligned(32)));
+static __m256i bit_met_p7_m7 __attribute__ ((aligned(32)));
+static __m256i bit_met_p7_m5 __attribute__ ((aligned(32)));
+static __m256i bit_met_p7_m3 __attribute__ ((aligned(32)));
+static __m256i bit_met_p7_m1 __attribute__ ((aligned(32)));
+static __m256i bit_met_p7_p1 __attribute__ ((aligned(32)));
+static __m256i bit_met_p7_p3 __attribute__ ((aligned(32)));
+static __m256i bit_met_p7_p5 __attribute__ ((aligned(32)));
+static __m256i bit_met_p7_p7 __attribute__ ((aligned(32)));
+
+static __m256i y0_p_1_1 __attribute__ ((aligned(32)));
+static __m256i y0_p_1_3 __attribute__ ((aligned(32)));
+static __m256i y0_p_1_5 __attribute__ ((aligned(32)));
+static __m256i y0_p_1_7 __attribute__ ((aligned(32)));
+static __m256i y0_p_3_1 __attribute__ ((aligned(32)));
+static __m256i y0_p_3_3 __attribute__ ((aligned(32)));
+static __m256i y0_p_3_5 __attribute__ ((aligned(32)));
+static __m256i y0_p_3_7 __attribute__ ((aligned(32)));
+static __m256i y0_p_5_1 __attribute__ ((aligned(32)));
+static __m256i y0_p_5_3 __attribute__ ((aligned(32)));
+static __m256i y0_p_5_5 __attribute__ ((aligned(32)));
+static __m256i y0_p_5_7 __attribute__ ((aligned(32)));
+static __m256i y0_p_7_1 __attribute__ ((aligned(32)));
+static __m256i y0_p_7_3 __attribute__ ((aligned(32)));
+static __m256i y0_p_7_5 __attribute__ ((aligned(32)));
+static __m256i y0_p_7_7 __attribute__ ((aligned(32)));
+static __m256i y0_m_1_1 __attribute__ ((aligned(32)));
+static __m256i y0_m_1_3 __attribute__ ((aligned(32)));
+static __m256i y0_m_1_5 __attribute__ ((aligned(32)));
+static __m256i y0_m_1_7 __attribute__ ((aligned(32)));
+static __m256i y0_m_3_1 __attribute__ ((aligned(32)));
+static __m256i y0_m_3_3 __attribute__ ((aligned(32)));
+static __m256i y0_m_3_5 __attribute__ ((aligned(32)));
+static __m256i y0_m_3_7 __attribute__ ((aligned(32)));
+static __m256i y0_m_5_1 __attribute__ ((aligned(32)));
+static __m256i y0_m_5_3 __attribute__ ((aligned(32)));
+static __m256i y0_m_5_5 __attribute__ ((aligned(32)));
+static __m256i y0_m_5_7 __attribute__ ((aligned(32)));
+static __m256i y0_m_7_1 __attribute__ ((aligned(32)));
+static __m256i y0_m_7_3 __attribute__ ((aligned(32)));
+static __m256i y0_m_7_5 __attribute__ ((aligned(32)));
+static __m256i y0_m_7_7 __attribute__ ((aligned(32)));
+
+static __m256i xmm0 __attribute__ ((aligned(32)));
+static __m256i xmm1 __attribute__ ((aligned(32)));
+static __m256i xmm2 __attribute__ ((aligned(32)));
+static __m256i xmm3 __attribute__ ((aligned(32)));
+static __m256i xmm4 __attribute__ ((aligned(32)));
+static __m256i xmm5 __attribute__ ((aligned(32)));
+static __m256i xmm6 __attribute__ ((aligned(32)));
+static __m256i xmm7 __attribute__ ((aligned(32)));
+static __m256i xmm8 __attribute__ ((aligned(32)));
+
+static __m256i y0r __attribute__ ((aligned(32)));
+static __m256i y0i __attribute__ ((aligned(32)));
+static __m256i y1r __attribute__ ((aligned(32)));
+static __m256i y1i __attribute__ ((aligned(32)));
+static __m256i y2r __attribute__ ((aligned(32)));
+static __m256i y2i __attribute__ ((aligned(32)));
+
+static __m256i logmax_num_re0 __attribute__ ((aligned(32)));
+static __m256i logmax_den_re0 __attribute__ ((aligned(32)));
+
+static __m256i tmp_result __attribute__ ((aligned(32)));
+static __m256i tmp_result2 __attribute__ ((aligned(32)));
+static __m256i tmp_result3 __attribute__ ((aligned(32)));
+static __m256i tmp_result4 __attribute__ ((aligned(32)));
+
+//==============================================================================================
+// Auxiliary Makros
+
+// calculate interference magnitude
+#define interference_abs_epi16(psi,int_ch_mag,int_mag,c1,c2) tmp_result = _mm256_cmpgt_epi16(int_ch_mag,psi); tmp_result2 = _mm256_xor_si256(tmp_result,(*(__m256i*)&ones256[0])); tmp_result = _mm256_and_si256(tmp_result,c1); tmp_result2 = _mm256_and_si256(tmp_result2,c2); int_mag = _mm256_or_si256(tmp_result,tmp_result2);
+
+// calculate interference magnitude
+// tmp_result = ones in shorts corr. to interval 2<=x<=4, tmp_result2 interval < 2, tmp_result3 interval 4<x<6 and tmp_result4 interval x>6
+#define interference_abs_64qam_epi16(psi,int_ch_mag,int_two_ch_mag,int_three_ch_mag,a,c1,c3,c5,c7) tmp_result = _mm256_cmpgt_epi16(int_two_ch_mag,psi); tmp_result3 = _mm256_xor_si256(tmp_result,(*(__m256i*)&ones256[0])); tmp_result2 = _mm256_cmpgt_epi16(int_ch_mag,psi); tmp_result = _mm256_xor_si256(tmp_result,tmp_result2); tmp_result4 = _mm256_cmpgt_epi16(psi,int_three_ch_mag); tmp_result3 = _mm256_xor_si256(tmp_result3,tmp_result4); tmp_result = _mm256_and_si256(tmp_result,c3); tmp_result2 = _mm256_and_si256(tmp_result2,c1); tmp_result3 = _mm256_and_si256(tmp_result3,c5); tmp_result4 = _mm256_and_si256(tmp_result4,c7); tmp_result = _mm256_or_si256(tmp_result,tmp_result2); tmp_result3 = _mm256_or_si256(tmp_result3,tmp_result4); a = _mm256_or_si256(tmp_result,tmp_result3);
+
+// calculates psi_a = psi_r*a_r + psi_i*a_i
+#define prodsum_psi_a_epi16(psi_r,a_r,psi_i,a_i,psi_a) tmp_result = _mm256_mulhi_epi16(psi_r,a_r); tmp_result = _mm256_slli_epi16(tmp_result,1); tmp_result2 = _mm256_mulhi_epi16(psi_i,a_i); tmp_result2 = _mm256_slli_epi16(tmp_result2,1); psi_a = _mm256_adds_epi16(tmp_result,tmp_result2);
+
+// 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 = _mm256_mulhi_epi16(a_r,a_r); tmp_result = _mm256_slli_epi16(tmp_result,1); tmp_result = _mm256_mulhi_epi16(tmp_result,scale_factor); tmp_result = _mm256_slli_epi16(tmp_result,1); tmp_result = _mm256_mulhi_epi16(tmp_result,int_ch_mag); tmp_result = _mm256_slli_epi16(tmp_result,1); tmp_result2 = _mm256_mulhi_epi16(a_i,a_i); tmp_result2 = _mm256_slli_epi16(tmp_result2,1); tmp_result2 = _mm256_mulhi_epi16(tmp_result2,scale_factor); tmp_result2 = _mm256_slli_epi16(tmp_result2,1); tmp_result2 = _mm256_mulhi_epi16(tmp_result2,int_ch_mag); tmp_result2 = _mm256_slli_epi16(tmp_result2,1); a_sq = _mm256_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 = _mm256_mulhi_epi16(a_r,a_r); tmp_result = _mm256_slli_epi16(tmp_result,1); tmp_result = _mm256_mulhi_epi16(tmp_result,scale_factor); tmp_result = _mm256_slli_epi16(tmp_result,3); tmp_result = _mm256_mulhi_epi16(tmp_result,int_ch_mag); tmp_result = _mm256_slli_epi16(tmp_result,1); tmp_result2 = _mm256_mulhi_epi16(a_i,a_i); tmp_result2 = _mm256_slli_epi16(tmp_result2,1); tmp_result2 = _mm256_mulhi_epi16(tmp_result2,scale_factor); tmp_result2 = _mm256_slli_epi16(tmp_result2,3); tmp_result2 = _mm256_mulhi_epi16(tmp_result2,int_ch_mag); tmp_result2 = _mm256_slli_epi16(tmp_result2,1); a_sq = _mm256_adds_epi16(tmp_result,tmp_result2);
+
+void seperate_real_imag_parts(__m256i *out_re,
+ __m256i *out_im,
+ __m256i in0,
+ __m256i in1)
+{
+ __m256i tmp0;
+ __m256i tmp1;
+
+ in0 = _mm256_shufflelo_epi16(in0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ in0 = _mm256_shufflehi_epi16(in0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ in0 = _mm256_shuffle_epi32(in0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+
+ in1 = _mm256_shufflelo_epi16(in1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ in1 = _mm256_shufflehi_epi16(in1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ in1 = _mm256_shuffle_epi32(in1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+
+ //in0 = [Re(0,1,2,3) Im(0,1,2,3) Re(4,5,6,7) Im(4,5,6,7)]
+ //in0 = [Re(8,9,10,11) Im(8,9,10,11) Re(12,13,14,15) Im(12,13,14,15)]
+
+ tmp0 = _mm256_unpacklo_epi64(in0, in1);
+ //axmm2 = [Re(0,1,2,3) Re(8,9,10,11) Re(4,5,6,7) Re(12,13,14,15)]
+ tmp0 = _mm256_permute4x64_epi64(tmp0,0xd8); // Re(rho)
+
+ tmp1 = _mm256_unpackhi_epi64(in0, in1);
+ //axmm3 = [Im(0,1,2,3) Im(8,9,10,11) Im(4,5,6,7) Im(12,13,14,15)]
+ tmp1 = _mm256_permute4x64_epi64(tmp1,0xd8); // Im(rho)
+
+ *out_re = tmp0;
+ *out_im = tmp1;
+}
+
+void qam64_qam16_avx2(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__)
+
+ __m256i *rho01_256i = (__m256i *)rho01;
+ __m256i *stream0_256i_in = (__m256i *)stream0_in;
+ __m256i *stream1_256i_in = (__m256i *)stream1_in;
+ __m256i *ch_mag_256i = (__m256i *)ch_mag;
+ __m256i *ch_mag_256i_i = (__m256i *)ch_mag_i;
+
+ __m256i ONE_OVER_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(10112)); // round(1/sqrt(42)*2^16)
+ __m256i THREE_OVER_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(30337)); // round(3/sqrt(42)*2^16)
+ __m256i FIVE_OVER_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(25281)); // round(5/sqrt(42)*2^15)
+ __m256i SEVEN_OVER_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(17697)); // round(5/sqrt(42)*2^15)
+ __m256i FORTYNINE_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(30969)); // round(49/(4*sqrt(42))*2^14), Q2.14
+ __m256i THIRTYSEVEN_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(23385)); // round(37/(4*sqrt(42))*2^14), Q2.14
+ __m256i TWENTYFIVE_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(31601)); // round(25/(4*sqrt(42))*2^15)
+ __m256i TWENTYNINE_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(18329)); // round(29/(4*sqrt(42))*2^15), Q2.14
+ __m256i SEVENTEEN_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(21489)); // round(17/(4*sqrt(42))*2^15)
+ __m256i NINE_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(11376)); // round(9/(4*sqrt(42))*2^15)
+ __m256i THIRTEEN_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(16433)); // round(13/(4*sqrt(42))*2^15)
+ __m256i FIVE_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(6320)); // round(5/(4*sqrt(42))*2^15)
+ __m256i ONE_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(1264)); // round(1/(4*sqrt(42))*2^15)
+ __m256i ONE_OVER_SQRT_10_Q15 = _mm256_broadcastw_epi16(_mm_set1_epi16(10362)); // round(1/sqrt(10)*2^15)
+ __m256i THREE_OVER_SQRT_10 = _mm256_broadcastw_epi16(_mm_set1_epi16(31086)); // round(3/sqrt(10)*2^15)
+ __m256i SQRT_10_OVER_FOUR = _mm256_broadcastw_epi16(_mm_set1_epi16(25905)); // round(sqrt(10)/4*2^15)
+
+
+ __m256i ch_mag_int;
+ __m256i ch_mag_des;
+ __m256i ch_mag_98_over_42_with_sigma2;
+ __m256i ch_mag_74_over_42_with_sigma2;
+ __m256i ch_mag_58_over_42_with_sigma2;
+ __m256i ch_mag_50_over_42_with_sigma2;
+ __m256i ch_mag_34_over_42_with_sigma2;
+ __m256i ch_mag_18_over_42_with_sigma2;
+ __m256i ch_mag_26_over_42_with_sigma2;
+ __m256i ch_mag_10_over_42_with_sigma2;
+ __m256i ch_mag_2_over_42_with_sigma2;
+ __m256i y0r_one_over_sqrt_21;
+ __m256i y0r_three_over_sqrt_21;
+ __m256i y0r_five_over_sqrt_21;
+ __m256i y0r_seven_over_sqrt_21;
+ __m256i y0i_one_over_sqrt_21;
+ __m256i y0i_three_over_sqrt_21;
+ __m256i y0i_five_over_sqrt_21;
+ __m256i y0i_seven_over_sqrt_21;
+
+#elif defined(__arm__)
+
+#endif
+ int i,j;
+ uint32_t len256 = (length)>>3;
+
+ for (i=0; i<len256; 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)
+ */
+ seperate_real_imag_parts(&xmm2, &xmm3, rho01_256i[i], rho01_256i[i+1]);
+
+ rho_rpi = _mm256_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho)
+ rho_rmi = _mm256_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho)
+
+ // Compute the different rhos
+ rho_rpi_1_1 = _mm256_mulhi_epi16(rho_rpi, ONE_OVER_SQRT_42);
+ rho_rmi_1_1 = _mm256_mulhi_epi16(rho_rmi, ONE_OVER_SQRT_42);
+ rho_rpi_3_3 = _mm256_mulhi_epi16(rho_rpi, THREE_OVER_SQRT_42);
+ rho_rmi_3_3 = _mm256_mulhi_epi16(rho_rmi, THREE_OVER_SQRT_42);
+ rho_rpi_5_5 = _mm256_mulhi_epi16(rho_rpi, FIVE_OVER_SQRT_42);
+ rho_rmi_5_5 = _mm256_mulhi_epi16(rho_rmi, FIVE_OVER_SQRT_42);
+ rho_rpi_7_7 = _mm256_mulhi_epi16(rho_rpi, SEVEN_OVER_SQRT_42);
+ rho_rmi_7_7 = _mm256_mulhi_epi16(rho_rmi, SEVEN_OVER_SQRT_42);
+
+ rho_rpi_5_5 = _mm256_slli_epi16(rho_rpi_5_5, 1);
+ rho_rmi_5_5 = _mm256_slli_epi16(rho_rmi_5_5, 1);
+ rho_rpi_7_7 = _mm256_slli_epi16(rho_rpi_7_7, 2);
+ rho_rmi_7_7 = _mm256_slli_epi16(rho_rmi_7_7, 2);
+
+ xmm4 = _mm256_mulhi_epi16(xmm2, ONE_OVER_SQRT_42);
+ xmm5 = _mm256_mulhi_epi16(xmm3, ONE_OVER_SQRT_42);
+ xmm6 = _mm256_mulhi_epi16(xmm3, THREE_OVER_SQRT_42);
+ xmm7 = _mm256_mulhi_epi16(xmm3, FIVE_OVER_SQRT_42);
+ xmm8 = _mm256_mulhi_epi16(xmm3, SEVEN_OVER_SQRT_42);
+ xmm7 = _mm256_slli_epi16(xmm7, 1);
+ xmm8 = _mm256_slli_epi16(xmm8, 2);
+
+ rho_rpi_1_3 = _mm256_adds_epi16(xmm4, xmm6);
+ rho_rmi_1_3 = _mm256_subs_epi16(xmm4, xmm6);
+ rho_rpi_1_5 = _mm256_adds_epi16(xmm4, xmm7);
+ rho_rmi_1_5 = _mm256_subs_epi16(xmm4, xmm7);
+ rho_rpi_1_7 = _mm256_adds_epi16(xmm4, xmm8);
+ rho_rmi_1_7 = _mm256_subs_epi16(xmm4, xmm8);
+
+ xmm4 = _mm256_mulhi_epi16(xmm2, THREE_OVER_SQRT_42);
+ rho_rpi_3_1 = _mm256_adds_epi16(xmm4, xmm5);
+ rho_rmi_3_1 = _mm256_subs_epi16(xmm4, xmm5);
+ rho_rpi_3_5 = _mm256_adds_epi16(xmm4, xmm7);
+ rho_rmi_3_5 = _mm256_subs_epi16(xmm4, xmm7);
+ rho_rpi_3_7 = _mm256_adds_epi16(xmm4, xmm8);
+ rho_rmi_3_7 = _mm256_subs_epi16(xmm4, xmm8);
+
+ xmm4 = _mm256_mulhi_epi16(xmm2, FIVE_OVER_SQRT_42);
+ xmm4 = _mm256_slli_epi16(xmm4, 1);
+ rho_rpi_5_1 = _mm256_adds_epi16(xmm4, xmm5);
+ rho_rmi_5_1 = _mm256_subs_epi16(xmm4, xmm5);
+ rho_rpi_5_3 = _mm256_adds_epi16(xmm4, xmm6);
+ rho_rmi_5_3 = _mm256_subs_epi16(xmm4, xmm6);
+ rho_rpi_5_7 = _mm256_adds_epi16(xmm4, xmm8);
+ rho_rmi_5_7 = _mm256_subs_epi16(xmm4, xmm8);
+
+ xmm4 = _mm256_mulhi_epi16(xmm2, SEVEN_OVER_SQRT_42);
+ xmm4 = _mm256_slli_epi16(xmm4, 2);
+ rho_rpi_7_1 = _mm256_adds_epi16(xmm4, xmm5);
+ rho_rmi_7_1 = _mm256_subs_epi16(xmm4, xmm5);
+ rho_rpi_7_3 = _mm256_adds_epi16(xmm4, xmm6);
+ rho_rmi_7_3 = _mm256_subs_epi16(xmm4, xmm6);
+ rho_rpi_7_5 = _mm256_adds_epi16(xmm4, xmm7);
+ rho_rmi_7_5 = _mm256_subs_epi16(xmm4, xmm7);
+
+ // Rearrange interfering MF output
+ /*
+ xmm0 = stream1_128i_in[i];
+ xmm1 = stream1_128i_in[i+1];
+ xmm0 = _mm256_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm0 = _mm256_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm0 = _mm256_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm1 = _mm256_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm1 = _mm256_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm1 = _mm256_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 = _mm256_unpacklo_epi64(xmm0,xmm1); //[y1r(1),y1r(2),y1r(3),y1r(4)]
+ y1i = _mm256_unpackhi_epi64(xmm0,xmm1); //[y1i(1),y1i(2),y1i(3),y1i(4)]
+ */
+
+ seperate_real_imag_parts(&y1r, &y1i, stream1_256i_in[i], stream1_256i_in[i+1]);
+
+ // Psi_r calculation from rho_rpi or rho_rmi
+ xmm0 = _mm256_broadcastw_epi16(_mm_set1_epi16(0));// ZERO for abs_pi16
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_7, y1r);
+ psi_r_p7_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_5, y1r);
+ psi_r_p7_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_3, y1r);
+ psi_r_p7_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_1, y1r);
+ psi_r_p7_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_1, y1r);
+ psi_r_p7_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_3, y1r);
+ psi_r_p7_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_5, y1r);
+ psi_r_p7_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_7, y1r);
+ psi_r_p7_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_7, y1r);
+ psi_r_p5_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_5, y1r);
+ psi_r_p5_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_3, y1r);
+ psi_r_p5_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_1, y1r);
+ psi_r_p5_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_1, y1r);
+ psi_r_p5_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_3, y1r);
+ psi_r_p5_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_5, y1r);
+ psi_r_p5_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_7, y1r);
+ psi_r_p5_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_7, y1r);
+ psi_r_p3_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_5, y1r);
+ psi_r_p3_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_3, y1r);
+ psi_r_p3_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_1, y1r);
+ psi_r_p3_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_1, y1r);
+ psi_r_p3_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_3, y1r);
+ psi_r_p3_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_5, y1r);
+ psi_r_p3_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_7, y1r);
+ psi_r_p3_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_7, y1r);
+ psi_r_p1_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_5, y1r);
+ psi_r_p1_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_3, y1r);
+ psi_r_p1_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_1, y1r);
+ psi_r_p1_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_1, y1r);
+ psi_r_p1_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_3, y1r);
+ psi_r_p1_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_5, y1r);
+ psi_r_p1_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_7, y1r);
+ psi_r_p1_m7 = _mm256_abs_epi16(xmm2);
+
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_7, y1r);
+ psi_r_m1_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_5, y1r);
+ psi_r_m1_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_3, y1r);
+ psi_r_m1_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_1, y1r);
+ psi_r_m1_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_1, y1r);
+ psi_r_m1_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_3, y1r);
+ psi_r_m1_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_5, y1r);
+ psi_r_m1_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_7, y1r);
+ psi_r_m1_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_7, y1r);
+ psi_r_m3_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_5, y1r);
+ psi_r_m3_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_3, y1r);
+ psi_r_m3_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_1, y1r);
+ psi_r_m3_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_1, y1r);
+ psi_r_m3_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_3, y1r);
+ psi_r_m3_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_5, y1r);
+ psi_r_m3_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_7, y1r);
+ psi_r_m3_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_7, y1r);
+ psi_r_m5_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_5, y1r);
+ psi_r_m5_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_3, y1r);
+ psi_r_m5_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_1, y1r);
+ psi_r_m5_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_1, y1r);
+ psi_r_m5_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_3, y1r);
+ psi_r_m5_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_5, y1r);
+ psi_r_m5_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_7, y1r);
+ psi_r_m5_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_7, y1r);
+ psi_r_m7_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_5, y1r);
+ psi_r_m7_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_3, y1r);
+ psi_r_m7_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_1, y1r);
+ psi_r_m7_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_1, y1r);
+ psi_r_m7_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_3, y1r);
+ psi_r_m7_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_5, y1r);
+ psi_r_m7_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_7, y1r);
+ psi_r_m7_m7 = _mm256_abs_epi16(xmm2);
+
+ // Psi_i calculation from rho_rpi or rho_rmi
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_7, y1i);
+ psi_i_p7_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_7, y1i);
+ psi_i_p7_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_7, y1i);
+ psi_i_p7_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_7, y1i);
+ psi_i_p7_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_7, y1i);
+ psi_i_p7_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_7, y1i);
+ psi_i_p7_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_7, y1i);
+ psi_i_p7_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_7, y1i);
+ psi_i_p7_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_5, y1i);
+ psi_i_p5_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_5, y1i);
+ psi_i_p5_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_5, y1i);
+ psi_i_p5_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_5, y1i);
+ psi_i_p5_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_5, y1i);
+ psi_i_p5_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_5, y1i);
+ psi_i_p5_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_5, y1i);
+ psi_i_p5_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_5, y1i);
+ psi_i_p5_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_3, y1i);
+ psi_i_p3_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_3, y1i);
+ psi_i_p3_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_3, y1i);
+ psi_i_p3_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_3, y1i);
+ psi_i_p3_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_3, y1i);
+ psi_i_p3_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_3, y1i);
+ psi_i_p3_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_3, y1i);
+ psi_i_p3_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_3, y1i);
+ psi_i_p3_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_1, y1i);
+ psi_i_p1_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_1, y1i);
+ psi_i_p1_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_1, y1i);
+ psi_i_p1_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_1, y1i);
+ psi_i_p1_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_1, y1i);
+ psi_i_p1_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_1, y1i);
+ psi_i_p1_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_1, y1i);
+ psi_i_p1_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_1, y1i);
+ psi_i_p1_m7 = _mm256_abs_epi16(xmm2);
+
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_1, y1i);
+ psi_i_m1_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_1, y1i);
+ psi_i_m1_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_1, y1i);
+ psi_i_m1_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_1, y1i);
+ psi_i_m1_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_1, y1i);
+ psi_i_m1_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_1, y1i);
+ psi_i_m1_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_1, y1i);
+ psi_i_m1_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_1, y1i);
+ psi_i_m1_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_3, y1i);
+ psi_i_m3_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_3, y1i);
+ psi_i_m3_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_3, y1i);
+ psi_i_m3_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_3, y1i);
+ psi_i_m3_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_3, y1i);
+ psi_i_m3_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_3, y1i);
+ psi_i_m3_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_3, y1i);
+ psi_i_m3_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_3, y1i);
+ psi_i_m3_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_5, y1i);
+ psi_i_m5_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_5, y1i);
+ psi_i_m5_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_5, y1i);
+ psi_i_m5_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_5, y1i);
+ psi_i_m5_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_5, y1i);
+ psi_i_m5_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_5, y1i);
+ psi_i_m5_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_5, y1i);
+ psi_i_m5_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_5, y1i);
+ psi_i_m5_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_7, y1i);
+ psi_i_m7_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_7, y1i);
+ psi_i_m7_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_7, y1i);
+ psi_i_m7_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_7, y1i);
+ psi_i_m7_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_7, y1i);
+ psi_i_m7_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_7, y1i);
+ psi_i_m7_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_7, y1i);
+ psi_i_m7_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_7, y1i);
+ psi_i_m7_m7 = _mm256_abs_epi16(xmm2);
+
+/*
+ // Rearrange desired MF output
+ xmm0 = stream0_128i_in[i];
+ xmm1 = stream0_128i_in[i+1];
+ xmm0 = _mm256_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm0 = _mm256_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm0 = _mm256_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm1 = _mm256_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm1 = _mm256_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm1 = _mm256_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 = _mm256_unpacklo_epi64(xmm0,xmm1); // = [y0r(1),y0r(2),y0r(3),y0r(4)]
+ y0i = _mm256_unpackhi_epi64(xmm0,xmm1);
+*/
+ seperate_real_imag_parts(&y0r, &y0i, stream0_256i_in[i], stream0_256i_in[i+1]);
+
+ /*
+ // 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 = _mm256_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm2 = _mm256_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm2 = _mm256_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm3 = _mm256_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm3 = _mm256_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm3 = _mm256_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ ch_mag_des = _mm256_unpacklo_epi64(xmm2,xmm3);
+ */
+
+ seperate_real_imag_parts(&ch_mag_des, &xmm2, ch_mag_256i[i], ch_mag_256i[i+1]);
+
+ // Rearrange interfering channel magnitudes
+ /*
+ xmm2 = ch_mag_128i_i[i];
+ xmm3 = ch_mag_128i_i[i+1];
+ xmm2 = _mm256_shufflelo_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm2 = _mm256_shufflehi_epi16(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm2 = _mm256_shuffle_epi32(xmm2,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm3 = _mm256_shufflelo_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm3 = _mm256_shufflehi_epi16(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm3 = _mm256_shuffle_epi32(xmm3,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ ch_mag_int = _mm256_unpacklo_epi64(xmm2,xmm3);
+ */
+
+ seperate_real_imag_parts(&ch_mag_int, &xmm2, ch_mag_256i_i[i], ch_mag_256i_i[i+1]);
+
+ y0r_one_over_sqrt_21 = _mm256_mulhi_epi16(y0r, ONE_OVER_SQRT_42);
+ y0r_three_over_sqrt_21 = _mm256_mulhi_epi16(y0r, THREE_OVER_SQRT_42);
+ y0r_five_over_sqrt_21 = _mm256_mulhi_epi16(y0r, FIVE_OVER_SQRT_42);
+ y0r_five_over_sqrt_21 = _mm256_slli_epi16(y0r_five_over_sqrt_21, 1);
+ y0r_seven_over_sqrt_21 = _mm256_mulhi_epi16(y0r, SEVEN_OVER_SQRT_42);
+ y0r_seven_over_sqrt_21 = _mm256_slli_epi16(y0r_seven_over_sqrt_21, 2); // Q2.14
+
+ y0i_one_over_sqrt_21 = _mm256_mulhi_epi16(y0i, ONE_OVER_SQRT_42);
+ y0i_three_over_sqrt_21 = _mm256_mulhi_epi16(y0i, THREE_OVER_SQRT_42);
+ y0i_five_over_sqrt_21 = _mm256_mulhi_epi16(y0i, FIVE_OVER_SQRT_42);
+ y0i_five_over_sqrt_21 = _mm256_slli_epi16(y0i_five_over_sqrt_21, 1);
+ y0i_seven_over_sqrt_21 = _mm256_mulhi_epi16(y0i, SEVEN_OVER_SQRT_42);
+ y0i_seven_over_sqrt_21 = _mm256_slli_epi16(y0i_seven_over_sqrt_21, 2); // Q2.14
+
+ y0_p_7_1 = _mm256_adds_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_p_7_3 = _mm256_adds_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_p_7_5 = _mm256_adds_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_p_7_7 = _mm256_adds_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21);
+ y0_p_5_1 = _mm256_adds_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_p_5_3 = _mm256_adds_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_p_5_5 = _mm256_adds_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_p_5_7 = _mm256_adds_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21);
+ y0_p_3_1 = _mm256_adds_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_p_3_3 = _mm256_adds_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_p_3_5 = _mm256_adds_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_p_3_7 = _mm256_adds_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21);
+ y0_p_1_1 = _mm256_adds_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_p_1_3 = _mm256_adds_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_p_1_5 = _mm256_adds_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_p_1_7 = _mm256_adds_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21);
+
+ y0_m_1_1 = _mm256_subs_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_m_1_3 = _mm256_subs_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_m_1_5 = _mm256_subs_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_m_1_7 = _mm256_subs_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21);
+ y0_m_3_1 = _mm256_subs_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_m_3_3 = _mm256_subs_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_m_3_5 = _mm256_subs_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_m_3_7 = _mm256_subs_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21);
+ y0_m_5_1 = _mm256_subs_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_m_5_3 = _mm256_subs_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_m_5_5 = _mm256_subs_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_m_5_7 = _mm256_subs_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21);
+ y0_m_7_1 = _mm256_subs_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_m_7_3 = _mm256_subs_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_m_7_5 = _mm256_subs_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_m_7_7 = _mm256_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 = _mm256_mulhi_epi16(ch_mag_des,ONE_OVER_FOUR_SQRT_42);
+ ch_mag_2_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_2_over_42_with_sigma2,1);
+ // x=1, y=3
+ ch_mag_10_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,FIVE_OVER_FOUR_SQRT_42);
+ ch_mag_10_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_10_over_42_with_sigma2,1);
+ // x=1, x=5
+ ch_mag_26_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,THIRTEEN_OVER_FOUR_SQRT_42);
+ ch_mag_26_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_26_over_42_with_sigma2,1);
+ // x=1, y=7
+ ch_mag_50_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42);
+ ch_mag_50_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_50_over_42_with_sigma2,1);
+ // x=3, y=3
+ ch_mag_18_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,NINE_OVER_FOUR_SQRT_42);
+ ch_mag_18_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_18_over_42_with_sigma2,1);
+ // x=3, y=5
+ ch_mag_34_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,SEVENTEEN_OVER_FOUR_SQRT_42);
+ ch_mag_34_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_34_over_42_with_sigma2,1);
+ // x=3, y=7
+ ch_mag_58_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,TWENTYNINE_OVER_FOUR_SQRT_42);
+ ch_mag_58_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_58_over_42_with_sigma2,2);
+ // x=5, y=5
+ ch_mag_50_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42);
+ ch_mag_50_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_50_over_42_with_sigma2,1);
+ // x=5, y=7
+ ch_mag_74_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,THIRTYSEVEN_OVER_FOUR_SQRT_42);
+ ch_mag_74_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_74_over_42_with_sigma2,2);
+ // x=7, y=7
+ ch_mag_98_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,FORTYNINE_OVER_FOUR_SQRT_42);
+ ch_mag_98_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_98_over_42_with_sigma2,2);
+
+ // Computing Metrics
+ xmm0 = _mm256_subs_epi16(psi_a_p7_p7, a_sq_p7_p7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_7_7);
+ bit_met_p7_p7 = _mm256_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_p5, a_sq_p7_p5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_7_5);
+ bit_met_p7_p5 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_p3, a_sq_p7_p3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_7_3);
+ bit_met_p7_p3 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_p1, a_sq_p7_p1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_7_1);
+ bit_met_p7_p1 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_m1, a_sq_p7_m1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_7_1);
+ bit_met_p7_m1 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_m3, a_sq_p7_m3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_7_3);
+ bit_met_p7_m3 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_m5, a_sq_p7_m5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_7_5);
+ bit_met_p7_m5 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_m7, a_sq_p7_m7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_7_7);
+ bit_met_p7_m7 = _mm256_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_p7, a_sq_p5_p7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_5_7);
+ bit_met_p5_p7 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_p5, a_sq_p5_p5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_5_5);
+ bit_met_p5_p5 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_p3, a_sq_p5_p3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_5_3);
+ bit_met_p5_p3 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_p1, a_sq_p5_p1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_5_1);
+ bit_met_p5_p1 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_m1, a_sq_p5_m1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_5_1);
+ bit_met_p5_m1 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_m3, a_sq_p5_m3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_5_3);
+ bit_met_p5_m3 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_m5, a_sq_p5_m5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_5_5);
+ bit_met_p5_m5 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_m7, a_sq_p5_m7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_5_7);
+ bit_met_p5_m7 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_p7, a_sq_p3_p7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_3_7);
+ bit_met_p3_p7 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_p5, a_sq_p3_p5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_3_5);
+ bit_met_p3_p5 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_p3, a_sq_p3_p3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_3_3);
+ bit_met_p3_p3 = _mm256_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_p1, a_sq_p3_p1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_3_1);
+ bit_met_p3_p1 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_m1, a_sq_p3_m1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_3_1);
+ bit_met_p3_m1 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_m3, a_sq_p3_m3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_3_3);
+ bit_met_p3_m3 = _mm256_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_m5, a_sq_p3_m5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_3_5);
+ bit_met_p3_m5 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_m7, a_sq_p3_m7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_3_7);
+ bit_met_p3_m7 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_p7, a_sq_p1_p7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_1_7);
+ bit_met_p1_p7 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_p5, a_sq_p1_p5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_1_5);
+ bit_met_p1_p5 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_p3, a_sq_p1_p3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_1_3);
+ bit_met_p1_p3 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_p1, a_sq_p1_p1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_1_1);
+ bit_met_p1_p1 = _mm256_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_m1, a_sq_p1_m1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_1_1);
+ bit_met_p1_m1 = _mm256_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_m3, a_sq_p1_m3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_1_3);
+ bit_met_p1_m3 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_m5, a_sq_p1_m5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_1_5);
+ bit_met_p1_m5 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_m7, a_sq_p1_m7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_1_7);
+ bit_met_p1_m7 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+
+ xmm0 = _mm256_subs_epi16(psi_a_m1_p7, a_sq_m1_p7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_1_7);
+ bit_met_m1_p7 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_p5, a_sq_m1_p5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_1_5);
+ bit_met_m1_p5 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_p3, a_sq_m1_p3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_1_3);
+ bit_met_m1_p3 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_p1, a_sq_m1_p1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_1_1);
+ bit_met_m1_p1 = _mm256_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_m1, a_sq_m1_m1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_1_1);
+ bit_met_m1_m1 = _mm256_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_m3, a_sq_m1_m3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_1_3);
+ bit_met_m1_m3 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_m5, a_sq_m1_m5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_1_5);
+ bit_met_m1_m5 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_m7, a_sq_m1_m7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_1_7);
+ bit_met_m1_m7 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_p7, a_sq_m3_p7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_3_7);
+ bit_met_m3_p7 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_p5, a_sq_m3_p5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_3_5);
+ bit_met_m3_p5 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_p3, a_sq_m3_p3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_3_3);
+ bit_met_m3_p3 = _mm256_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_p1, a_sq_m3_p1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_3_1);
+ bit_met_m3_p1 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_m1, a_sq_m3_m1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_3_1);
+ bit_met_m3_m1 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_m3, a_sq_m3_m3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_3_3);
+ bit_met_m3_m3 = _mm256_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_m5, a_sq_m3_m5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_3_5);
+ bit_met_m3_m5 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_m7, a_sq_m3_m7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_3_7);
+ bit_met_m3_m7 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_p7, a_sq_m5_p7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_5_7);
+ bit_met_m5_p7 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_p5, a_sq_m5_p5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_5_5);
+ bit_met_m5_p5 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_p3, a_sq_m5_p3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_5_3);
+ bit_met_m5_p3 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_p1, a_sq_m5_p1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_5_1);
+ bit_met_m5_p1 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_m1, a_sq_m5_m1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_5_1);
+ bit_met_m5_m1 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_m3, a_sq_m5_m3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_5_3);
+ bit_met_m5_m3 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_m5, a_sq_m5_m5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_5_5);
+ bit_met_m5_m5 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_m7, a_sq_m5_m7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_5_7);
+ bit_met_m5_m7 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_p7, a_sq_m7_p7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_7_7);
+ bit_met_m7_p7 = _mm256_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_p5, a_sq_m7_p5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_7_5);
+ bit_met_m7_p5 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_p3, a_sq_m7_p3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_7_3);
+ bit_met_m7_p3 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_p1, a_sq_m7_p1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_7_1);
+ bit_met_m7_p1 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_m1, a_sq_m7_m1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_7_1);
+ bit_met_m7_m1 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_m3, a_sq_m7_m3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_7_3);
+ bit_met_m7_m3 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_m5, a_sq_m7_m5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_7_5);
+ bit_met_m7_m5 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_m7, a_sq_m7_m7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_7_7);
+ bit_met_m7_m7 = _mm256_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2);
+
+ // Detection for 1st bit (LTE mapping)
+ // bit = 1
+ xmm0 = _mm256_max_epi16(bit_met_m7_p7, bit_met_m7_p5);
+ xmm1 = _mm256_max_epi16(bit_met_m7_p3, bit_met_m7_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m7_m1, bit_met_m7_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m7_m5, bit_met_m7_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m5_p7, bit_met_m5_p5);
+ xmm1 = _mm256_max_epi16(bit_met_m5_p3, bit_met_m5_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m5_m1, bit_met_m5_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m5, bit_met_m5_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m3_p7, bit_met_m3_p5);
+ xmm1 = _mm256_max_epi16(bit_met_m3_p3, bit_met_m3_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m3_m1, bit_met_m3_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m3_m5, bit_met_m3_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m1_p7, bit_met_m1_p5);
+ xmm1 = _mm256_max_epi16(bit_met_m1_p3, bit_met_m1_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m1, bit_met_m1_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m1_m5, bit_met_m1_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+
+ // bit = 0
+ xmm0 = _mm256_max_epi16(bit_met_p7_p7, bit_met_p7_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p7_p3, bit_met_p7_p1);
+ xmm2 = _mm256_max_epi16(bit_met_p7_m1, bit_met_p7_m3);
+ xmm3 = _mm256_max_epi16(bit_met_p7_m5, bit_met_p7_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p5_p7, bit_met_p5_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p5_p3, bit_met_p5_p1);
+ xmm2 = _mm256_max_epi16(bit_met_p5_m1, bit_met_p5_m3);
+ xmm3 = _mm256_max_epi16(bit_met_p5_m5, bit_met_p5_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p3_p7, bit_met_p3_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p3, bit_met_p3_p1);
+ xmm2 = _mm256_max_epi16(bit_met_p3_m1, bit_met_p3_m3);
+ xmm3 = _mm256_max_epi16(bit_met_p3_m5, bit_met_p3_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p1_p7, bit_met_p1_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p1_p3, bit_met_p1_p1);
+ xmm2 = _mm256_max_epi16(bit_met_p1_m1, bit_met_p1_m3);
+ xmm3 = _mm256_max_epi16(bit_met_p1_m5, bit_met_p1_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+
+ y0r = _mm256_subs_epi16(logmax_num_re0, logmax_den_re0);
+
+ // Detection for 2nd bit (LTE mapping)
+ // bit = 1
+ xmm0 = _mm256_max_epi16(bit_met_p7_m1, bit_met_p5_m1);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m1, bit_met_p1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m1, bit_met_m3_m1);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m1, bit_met_m7_m1);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m3, bit_met_p5_m3);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m3, bit_met_p1_m3);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m3, bit_met_m3_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m3, bit_met_m7_m3);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m5, bit_met_p5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m5, bit_met_p1_m5);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m5, bit_met_m3_m5);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m5, bit_met_m7_m5);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m7, bit_met_p5_m7);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m7, bit_met_p1_m7);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m7, bit_met_m3_m7);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m7, bit_met_m7_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+
+ // bit = 0
+ xmm0 = _mm256_max_epi16(bit_met_p7_p1, bit_met_p5_p1);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p1, bit_met_p1_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p1, bit_met_m3_p1);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p1, bit_met_m7_p1);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p3, bit_met_p5_p3);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p3, bit_met_p1_p3);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p3, bit_met_m3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p3, bit_met_m7_p3);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p5, bit_met_p5_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p5, bit_met_p1_p5);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p5, bit_met_m3_p5);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p5, bit_met_m7_p5);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p7, bit_met_p5_p7);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p7, bit_met_p1_p7);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p7, bit_met_m3_p7);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p7, bit_met_m7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+
+ y1r = _mm256_subs_epi16(logmax_num_re0, logmax_den_re0);
+
+ // Detection for 3rd bit (LTE mapping)
+ xmm0 = _mm256_max_epi16(bit_met_m7_m7, bit_met_m7_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m7_m3, bit_met_m7_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m7_p1, bit_met_m7_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m7_p5, bit_met_m7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m5_m7, bit_met_m5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m5_m3, bit_met_m5_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m5_p1, bit_met_m5_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p5, bit_met_m5_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p5_m7, bit_met_p5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p5_m3, bit_met_p5_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p5_p1, bit_met_p5_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p5_p5, bit_met_p5_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m7, bit_met_p7_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p7_m3, bit_met_p7_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p7_p1, bit_met_p7_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p7_p5, bit_met_p7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+
+ xmm0 = _mm256_max_epi16(bit_met_m3_m7, bit_met_m3_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m3_m3, bit_met_m3_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m3_p1, bit_met_m3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m3_p5, bit_met_m3_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m1_m7, bit_met_m1_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m1_m3, bit_met_m1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p1, bit_met_m1_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m1_p5, bit_met_m1_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p1_m7, bit_met_p1_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p1_m3, bit_met_p1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p1_p1, bit_met_p1_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p1_p5, bit_met_p1_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p3_m7, bit_met_p3_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m3, bit_met_p3_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p3_p1, bit_met_p3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p3_p5, bit_met_p3_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+
+ y2r = _mm256_subs_epi16(logmax_num_re0, logmax_den_re0);
+
+ // Detection for 4th bit (LTE mapping)
+ xmm0 = _mm256_max_epi16(bit_met_p7_p7, bit_met_p5_p7);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p7, bit_met_p1_p7);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p7, bit_met_m3_p7);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p7, bit_met_m7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p5, bit_met_p5_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p5, bit_met_p1_p5);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p5, bit_met_m3_p5);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p5, bit_met_m5_p5);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m5, bit_met_p5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m5, bit_met_p1_m5);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m5, bit_met_m3_m5);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m5, bit_met_m7_m5);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m7, bit_met_p5_m7);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m7, bit_met_p1_m7);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m7, bit_met_m3_m7);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m7, bit_met_m7_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+
+ xmm0 = _mm256_max_epi16(bit_met_p7_m1, bit_met_p5_m1);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m1, bit_met_p1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m1, bit_met_m3_m1);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m1, bit_met_m7_m1);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m3, bit_met_p5_m3);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m3, bit_met_p1_m3);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m3, bit_met_m3_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m3, bit_met_m7_m3);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p1, bit_met_p5_p1);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p1, bit_met_p1_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p1, bit_met_m3_p1);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p1, bit_met_m7_p1);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p3, bit_met_p5_p3);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p3, bit_met_p1_p3);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p3, bit_met_m3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p3, bit_met_m7_p3);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+
+ y0i = _mm256_subs_epi16(logmax_num_re0, logmax_den_re0);
+
+
+ // Detection for 5th bit (LTE mapping)
+ xmm0 = _mm256_max_epi16(bit_met_m7_m7, bit_met_m7_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m7_m3, bit_met_m7_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m7_p1, bit_met_m7_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m7_p5, bit_met_m7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m1_m7, bit_met_m1_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m1_m3, bit_met_m1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p1, bit_met_m1_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m1_p5, bit_met_m1_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p1_m7, bit_met_p1_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p1_m3, bit_met_p1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p1_p1, bit_met_p1_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p1_p5, bit_met_p1_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m7, bit_met_p7_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p7_m3, bit_met_p7_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p7_p1, bit_met_p7_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p7_p5, bit_met_p7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+
+ xmm0 = _mm256_max_epi16(bit_met_m5_m7, bit_met_m5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m5_m3, bit_met_m5_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m5_p1, bit_met_m5_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p5, bit_met_m5_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m3_m7, bit_met_m3_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m3_m3, bit_met_m3_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m3_p1, bit_met_m3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m3_p5, bit_met_m3_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p3_m7, bit_met_p3_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m3, bit_met_p3_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p3_p1, bit_met_p3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p3_p5, bit_met_p3_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p5_m7, bit_met_p5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p5_m3, bit_met_p5_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p5_p1, bit_met_p5_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p5_p5, bit_met_p5_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+
+ y1i = _mm256_subs_epi16(logmax_num_re0, logmax_den_re0);
+
+ // Detection for 6th bit (LTE mapping)
+ xmm0 = _mm256_max_epi16(bit_met_p7_p7, bit_met_p5_p7);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p7, bit_met_p1_p7);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p7, bit_met_m3_p7);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p7, bit_met_m7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p1, bit_met_p5_p1);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p1, bit_met_p1_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p1, bit_met_m3_p1);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p1, bit_met_m5_p1);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m1, bit_met_p5_m1);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m1, bit_met_p1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m1, bit_met_m3_m1);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m1, bit_met_m7_m1);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m7, bit_met_p5_m7);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m7, bit_met_p1_m7);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m7, bit_met_m3_m7);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m7, bit_met_m7_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+
+ xmm0 = _mm256_max_epi16(bit_met_p7_m5, bit_met_p5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m5, bit_met_p1_m5);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m5, bit_met_m3_m5);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m5, bit_met_m7_m5);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m3, bit_met_p5_m3);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m3, bit_met_p1_m3);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m3, bit_met_m3_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m3, bit_met_m7_m3);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p3, bit_met_p5_p3);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p3, bit_met_p1_p3);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p3, bit_met_m3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p3, bit_met_m7_p3);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p5, bit_met_p5_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p5, bit_met_p1_p5);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p5, bit_met_m3_p5);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p5, bit_met_m7_p5);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+
+ y2i = _mm256_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 = 48*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];
+
+ // RE 9
+ stream0_out[j + 48] = ((short *)&y0r)[8];
+ stream0_out[j + 49] = ((short *)&y1r)[8];
+ stream0_out[j + 50] = ((short *)&y2r)[8];
+ stream0_out[j + 51] = ((short *)&y0i)[8];
+ stream0_out[j + 52] = ((short *)&y1i)[8];
+ stream0_out[j + 53] = ((short *)&y2i)[8];
+ // RE 10
+ stream0_out[j + 54] = ((short *)&y0r)[9];
+ stream0_out[j + 55] = ((short *)&y1r)[9];
+ stream0_out[j + 56] = ((short *)&y2r)[9];
+ stream0_out[j + 57] = ((short *)&y0i)[9];
+ stream0_out[j + 58] = ((short *)&y1i)[9];
+ stream0_out[j + 59] = ((short *)&y2i)[9];
+ // RE 11
+ stream0_out[j + 60] = ((short *)&y0r)[10];
+ stream0_out[j + 61] = ((short *)&y1r)[10];
+ stream0_out[j + 62] = ((short *)&y2r)[10];
+ stream0_out[j + 63] = ((short *)&y0i)[10];
+ stream0_out[j + 64] = ((short *)&y1i)[10];
+ stream0_out[j + 65] = ((short *)&y2i)[10];
+ // RE 12
+ stream0_out[j + 66] = ((short *)&y0r)[11];
+ stream0_out[j + 67] = ((short *)&y1r)[11];
+ stream0_out[j + 68] = ((short *)&y2r)[11];
+ stream0_out[j + 69] = ((short *)&y0i)[11];
+ stream0_out[j + 70] = ((short *)&y1i)[11];
+ stream0_out[j + 71] = ((short *)&y2i)[11];
+ // RE 13
+ stream0_out[j + 72] = ((short *)&y0r)[12];
+ stream0_out[j + 73] = ((short *)&y1r)[12];
+ stream0_out[j + 74] = ((short *)&y2r)[12];
+ stream0_out[j + 75] = ((short *)&y0i)[12];
+ stream0_out[j + 76] = ((short *)&y1i)[12];
+ stream0_out[j + 77] = ((short *)&y2i)[12];
+ // RE 14
+ stream0_out[j + 78] = ((short *)&y0r)[13];
+ stream0_out[j + 79] = ((short *)&y1r)[13];
+ stream0_out[j + 80] = ((short *)&y2r)[13];
+ stream0_out[j + 81] = ((short *)&y0i)[13];
+ stream0_out[j + 82] = ((short *)&y1i)[13];
+ stream0_out[j + 83] = ((short *)&y2i)[13];
+ // RE 15
+ stream0_out[j + 84] = ((short *)&y0r)[14];
+ stream0_out[j + 85] = ((short *)&y1r)[14];
+ stream0_out[j + 86] = ((short *)&y2r)[14];
+ stream0_out[j + 87] = ((short *)&y0i)[14];
+ stream0_out[j + 88] = ((short *)&y1i)[14];
+ stream0_out[j + 89] = ((short *)&y2i)[14];
+ // RE 16
+ stream0_out[j + 90] = ((short *)&y0r)[15];
+ stream0_out[j + 91] = ((short *)&y1r)[15];
+ stream0_out[j + 92] = ((short *)&y2r)[15];
+ stream0_out[j + 93] = ((short *)&y0i)[15];
+ stream0_out[j + 94] = ((short *)&y1i)[15];
+ stream0_out[j + 95] = ((short *)&y2i)[15];
+
+#elif defined(__arm__)
+
+#endif
+ }
+
+#if defined(__x86_64__) || defined(__i386__)
+ _mm_empty();
+ _m_empty();
+#endif
+
+}
+
+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
+ )
+{
+
+ /*
+ Author: S. Wagner
+ Date: 28-02-17
+
+ 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__)
+
+ __m256i *rho01_256i = (__m256i *)rho01;
+ __m256i *stream0_256i_in = (__m256i *)stream0_in;
+ __m256i *stream1_256i_in = (__m256i *)stream1_in;
+ __m256i *ch_mag_256i = (__m256i *)ch_mag;
+ __m256i *ch_mag_256i_i = (__m256i *)ch_mag_i;
+
+ __m256i ONE_OVER_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(10112)); // round(1/sqrt(42)*2^16)
+ __m256i THREE_OVER_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(30337)); // round(3/sqrt(42)*2^16)
+ __m256i FIVE_OVER_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(25281)); // round(5/sqrt(42)*2^15)
+ __m256i SEVEN_OVER_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(17697)); // round(7/sqrt(42)*2^14) Q2.14
+ __m256i ONE_OVER_SQRT_2 = _mm256_broadcastw_epi16(_mm_set1_epi16(23170)); // round(1/sqrt(2)*2^15)
+ __m256i ONE_OVER_SQRT_2_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(3575)); // round(1/sqrt(2*42)*2^15)
+ __m256i THREE_OVER_SQRT_2_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(10726)); // round(3/sqrt(2*42)*2^15)
+ __m256i FIVE_OVER_SQRT_2_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(17876)); // round(5/sqrt(2*42)*2^15)
+ __m256i SEVEN_OVER_SQRT_2_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(25027)); // round(7/sqrt(2*42)*2^15)
+ __m256i FORTYNINE_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(30969)); // round(49/(4*sqrt(42))*2^14), Q2.14
+ __m256i THIRTYSEVEN_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(23385)); // round(37/(4*sqrt(42))*2^14), Q2.14
+ __m256i TWENTYFIVE_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(31601)); // round(25/(4*sqrt(42))*2^15)
+ __m256i TWENTYNINE_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(18329)); // round(29/(4*sqrt(42))*2^15), Q2.14
+ __m256i SEVENTEEN_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(21489)); // round(17/(4*sqrt(42))*2^15)
+ __m256i NINE_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(11376)); // round(9/(4*sqrt(42))*2^15)
+ __m256i THIRTEEN_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(16433)); // round(13/(4*sqrt(42))*2^15)
+ __m256i FIVE_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(6320)); // round(5/(4*sqrt(42))*2^15)
+ __m256i ONE_OVER_FOUR_SQRT_42 = _mm256_broadcastw_epi16(_mm_set1_epi16(1264)); // round(1/(4*sqrt(42))*2^15)
+ __m256i SQRT_42_OVER_FOUR = _mm256_broadcastw_epi16(_mm_set1_epi16(13272)); // round(sqrt(42)/4*2^13), Q3.12
+
+ __m256i ch_mag_des;
+ __m256i ch_mag_int;
+ __m256i ch_mag_98_over_42_with_sigma2;
+ __m256i ch_mag_74_over_42_with_sigma2;
+ __m256i ch_mag_58_over_42_with_sigma2;
+ __m256i ch_mag_50_over_42_with_sigma2;
+ __m256i ch_mag_34_over_42_with_sigma2;
+ __m256i ch_mag_18_over_42_with_sigma2;
+ __m256i ch_mag_26_over_42_with_sigma2;
+ __m256i ch_mag_10_over_42_with_sigma2;
+ __m256i ch_mag_2_over_42_with_sigma2;
+ __m256i y0r_one_over_sqrt_21;
+ __m256i y0r_three_over_sqrt_21;
+ __m256i y0r_five_over_sqrt_21;
+ __m256i y0r_seven_over_sqrt_21;
+ __m256i y0i_one_over_sqrt_21;
+ __m256i y0i_three_over_sqrt_21;
+ __m256i y0i_five_over_sqrt_21;
+ __m256i y0i_seven_over_sqrt_21;
+ __m256i ch_mag_int_with_sigma2;
+ __m256i two_ch_mag_int_with_sigma2;
+ __m256i three_ch_mag_int_with_sigma2;
+#elif defined(__arm__)
+
+#endif
+
+ int i,j;
+ uint32_t len256 = (length)>>3;
+
+ for (i=0; i<len256; i+=2) {
+
+#if defined(__x86_64__) || defined(__i386__)
+
+ // Get rho
+ /*
+ xmm0 = rho01_256i[i];
+ xmm1 = rho01_256i[i+1];
+ xmm0 = _mm256_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm0 = _mm256_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm0 = _mm256_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+
+ xmm1 = _mm256_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm1 = _mm256_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm1 = _mm256_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+
+ //xmm0 = [Re(0,1,2,3) Im(0,1,2,3) Re(4,5,6,7) Im(4,5,6,7)]
+ //xmm0 = [Re(8,9,10,11) Im(8,9,10,11) Re(12,13,14,15) Im(12,13,14,15)]
+
+ xmm2 = _mm256_unpacklo_epi64(xmm0, xmm1);
+ //xmm2 = [Re(0,1,2,3) Re(8,9,10,11) Re(4,5,6,7) Re(12,13,14,15)]
+ xmm2 = _mm256_permute4x64_epi64(xmm2,0xd8); // Re(rho)
+
+ xmm3 = _mm256_unpackhi_epi64(xmm0, xmm1);
+ //xmm3 = [Im(0,1,2,3) Im(8,9,10,11) Im(4,5,6,7) Im(12,13,14,15)]
+ xmm3 = _mm256_permute4x64_epi64(xmm3,0xd8); // Im(rho)
+ */
+
+ seperate_real_imag_parts(&xmm2, &xmm3, rho01_256i[i], rho01_256i[i+1]);
+
+ rho_rpi = _mm256_adds_epi16(xmm2,xmm3); // rho = Re(rho) + Im(rho)
+ rho_rmi = _mm256_subs_epi16(xmm2,xmm3); // rho* = Re(rho) - Im(rho)
+
+ // Compute the different rhos
+ rho_rpi_1_1 = _mm256_mulhi_epi16(rho_rpi, ONE_OVER_SQRT_42);
+ rho_rmi_1_1 = _mm256_mulhi_epi16(rho_rmi, ONE_OVER_SQRT_42);
+ rho_rpi_3_3 = _mm256_mulhi_epi16(rho_rpi, THREE_OVER_SQRT_42);
+ rho_rmi_3_3 = _mm256_mulhi_epi16(rho_rmi, THREE_OVER_SQRT_42);
+ rho_rpi_5_5 = _mm256_mulhi_epi16(rho_rpi, FIVE_OVER_SQRT_42);
+ rho_rmi_5_5 = _mm256_mulhi_epi16(rho_rmi, FIVE_OVER_SQRT_42);
+ rho_rpi_7_7 = _mm256_mulhi_epi16(rho_rpi, SEVEN_OVER_SQRT_42);
+ rho_rmi_7_7 = _mm256_mulhi_epi16(rho_rmi, SEVEN_OVER_SQRT_42);
+
+ rho_rpi_5_5 = _mm256_slli_epi16(rho_rpi_5_5, 1);
+ rho_rmi_5_5 = _mm256_slli_epi16(rho_rmi_5_5, 1);
+ rho_rpi_7_7 = _mm256_slli_epi16(rho_rpi_7_7, 2);
+ rho_rmi_7_7 = _mm256_slli_epi16(rho_rmi_7_7, 2);
+
+ xmm4 = _mm256_mulhi_epi16(xmm2, ONE_OVER_SQRT_42);
+ xmm5 = _mm256_mulhi_epi16(xmm3, ONE_OVER_SQRT_42);
+ xmm6 = _mm256_mulhi_epi16(xmm3, THREE_OVER_SQRT_42);
+ xmm7 = _mm256_mulhi_epi16(xmm3, FIVE_OVER_SQRT_42);
+ xmm8 = _mm256_mulhi_epi16(xmm3, SEVEN_OVER_SQRT_42);
+ xmm7 = _mm256_slli_epi16(xmm7, 1);
+ xmm8 = _mm256_slli_epi16(xmm8, 2);
+
+ rho_rpi_1_3 = _mm256_adds_epi16(xmm4, xmm6);
+ rho_rmi_1_3 = _mm256_subs_epi16(xmm4, xmm6);
+ rho_rpi_1_5 = _mm256_adds_epi16(xmm4, xmm7);
+ rho_rmi_1_5 = _mm256_subs_epi16(xmm4, xmm7);
+ rho_rpi_1_7 = _mm256_adds_epi16(xmm4, xmm8);
+ rho_rmi_1_7 = _mm256_subs_epi16(xmm4, xmm8);
+
+ xmm4 = _mm256_mulhi_epi16(xmm2, THREE_OVER_SQRT_42);
+ rho_rpi_3_1 = _mm256_adds_epi16(xmm4, xmm5);
+ rho_rmi_3_1 = _mm256_subs_epi16(xmm4, xmm5);
+ rho_rpi_3_5 = _mm256_adds_epi16(xmm4, xmm7);
+ rho_rmi_3_5 = _mm256_subs_epi16(xmm4, xmm7);
+ rho_rpi_3_7 = _mm256_adds_epi16(xmm4, xmm8);
+ rho_rmi_3_7 = _mm256_subs_epi16(xmm4, xmm8);
+
+ xmm4 = _mm256_mulhi_epi16(xmm2, FIVE_OVER_SQRT_42);
+ xmm4 = _mm256_slli_epi16(xmm4, 1);
+ rho_rpi_5_1 = _mm256_adds_epi16(xmm4, xmm5);
+ rho_rmi_5_1 = _mm256_subs_epi16(xmm4, xmm5);
+ rho_rpi_5_3 = _mm256_adds_epi16(xmm4, xmm6);
+ rho_rmi_5_3 = _mm256_subs_epi16(xmm4, xmm6);
+ rho_rpi_5_7 = _mm256_adds_epi16(xmm4, xmm8);
+ rho_rmi_5_7 = _mm256_subs_epi16(xmm4, xmm8);
+
+ xmm4 = _mm256_mulhi_epi16(xmm2, SEVEN_OVER_SQRT_42);
+ xmm4 = _mm256_slli_epi16(xmm4, 2);
+ rho_rpi_7_1 = _mm256_adds_epi16(xmm4, xmm5);
+ rho_rmi_7_1 = _mm256_subs_epi16(xmm4, xmm5);
+ rho_rpi_7_3 = _mm256_adds_epi16(xmm4, xmm6);
+ rho_rmi_7_3 = _mm256_subs_epi16(xmm4, xmm6);
+ rho_rpi_7_5 = _mm256_adds_epi16(xmm4, xmm7);
+ rho_rmi_7_5 = _mm256_subs_epi16(xmm4, xmm7);
+
+ // Rearrange interfering MF output
+ /*
+ xmm0 = stream1_256i_in[i];
+ xmm1 = stream1_256i_in[i+1];
+ xmm0 = _mm256_shufflelo_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm0 = _mm256_shufflehi_epi16(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm0 = _mm256_shuffle_epi32(xmm0,0xd8); //_MM_SHUFFLE(0,2,1,3));
+
+ xmm1 = _mm256_shufflelo_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm1 = _mm256_shufflehi_epi16(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+ xmm1 = _mm256_shuffle_epi32(xmm1,0xd8); //_MM_SHUFFLE(0,2,1,3));
+
+ y1r = _mm256_unpacklo_epi64(xmm0, xmm1);
+ y1r = _mm256_permute4x64_epi64(y1r,0xd8); // Re(y1)
+
+ y1i = _mm256_unpackhi_epi64(xmm0, xmm1);
+ y1i = _mm256_permute4x64_epi64(y1i,0xd8); // Im(y1)
+ */
+
+ seperate_real_imag_parts(&y1r, &y1i, stream1_256i_in[i], stream1_256i_in[i+1]);
+
+ // Psi_r calculation from rho_rpi or rho_rmi
+ xmm0 = _mm256_broadcastw_epi16(_mm_set1_epi16(0));// ZERO for abs_pi16
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_7, y1r);
+
+ psi_r_p7_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_5, y1r);
+ psi_r_p7_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_3, y1r);
+ psi_r_p7_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_1, y1r);
+ psi_r_p7_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_1, y1r);
+ psi_r_p7_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_3, y1r);
+ psi_r_p7_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_5, y1r);
+ psi_r_p7_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_7, y1r);
+ psi_r_p7_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_7, y1r);
+ psi_r_p5_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_5, y1r);
+ psi_r_p5_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_3, y1r);
+ psi_r_p5_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_1, y1r);
+ psi_r_p5_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_1, y1r);
+ psi_r_p5_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_3, y1r);
+ psi_r_p5_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_5, y1r);
+ psi_r_p5_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_7, y1r);
+ psi_r_p5_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_7, y1r);
+ psi_r_p3_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_5, y1r);
+ psi_r_p3_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_3, y1r);
+ psi_r_p3_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_1, y1r);
+ psi_r_p3_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_1, y1r);
+ psi_r_p3_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_3, y1r);
+ psi_r_p3_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_5, y1r);
+ psi_r_p3_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_7, y1r);
+ psi_r_p3_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_7, y1r);
+ psi_r_p1_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_5, y1r);
+ psi_r_p1_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_3, y1r);
+ psi_r_p1_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_1, y1r);
+ psi_r_p1_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_1, y1r);
+ psi_r_p1_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_3, y1r);
+ psi_r_p1_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_5, y1r);
+ psi_r_p1_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_7, y1r);
+ psi_r_p1_m7 = _mm256_abs_epi16(xmm2);
+
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_7, y1r);
+ psi_r_m1_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_5, y1r);
+ psi_r_m1_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_3, y1r);
+ psi_r_m1_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_1, y1r);
+ psi_r_m1_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_1, y1r);
+ psi_r_m1_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_3, y1r);
+ psi_r_m1_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_5, y1r);
+ psi_r_m1_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_7, y1r);
+ psi_r_m1_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_7, y1r);
+ psi_r_m3_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_5, y1r);
+ psi_r_m3_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_3, y1r);
+ psi_r_m3_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_1, y1r);
+ psi_r_m3_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_1, y1r);
+ psi_r_m3_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_3, y1r);
+ psi_r_m3_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_5, y1r);
+ psi_r_m3_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_7, y1r);
+ psi_r_m3_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_7, y1r);
+ psi_r_m5_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_5, y1r);
+ psi_r_m5_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_3, y1r);
+ psi_r_m5_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_1, y1r);
+ psi_r_m5_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_1, y1r);
+ psi_r_m5_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_3, y1r);
+ psi_r_m5_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_5, y1r);
+ psi_r_m5_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_7, y1r);
+ psi_r_m5_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_7, y1r);
+ psi_r_m7_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_5, y1r);
+ psi_r_m7_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_3, y1r);
+ psi_r_m7_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_1, y1r);
+ psi_r_m7_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_1, y1r);
+ psi_r_m7_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_3, y1r);
+ psi_r_m7_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_5, y1r);
+ psi_r_m7_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_7, y1r);
+ psi_r_m7_m7 = _mm256_abs_epi16(xmm2);
+
+ // Psi_i calculation from rho_rpi or rho_rmi
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_7, y1i);
+ psi_i_p7_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_7, y1i);
+ psi_i_p7_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_7, y1i);
+ psi_i_p7_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_7, y1i);
+ psi_i_p7_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_7, y1i);
+ psi_i_p7_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_7, y1i);
+ psi_i_p7_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_7, y1i);
+ psi_i_p7_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_7, y1i);
+ psi_i_p7_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_5, y1i);
+ psi_i_p5_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_5, y1i);
+ psi_i_p5_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_5, y1i);
+ psi_i_p5_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_5, y1i);
+ psi_i_p5_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_5, y1i);
+ psi_i_p5_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_5, y1i);
+ psi_i_p5_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_5, y1i);
+ psi_i_p5_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_5, y1i);
+ psi_i_p5_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_3, y1i);
+ psi_i_p3_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_3, y1i);
+ psi_i_p3_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_3, y1i);
+ psi_i_p3_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_3, y1i);
+ psi_i_p3_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_3, y1i);
+ psi_i_p3_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_3, y1i);
+ psi_i_p3_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_3, y1i);
+ psi_i_p3_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_3, y1i);
+ psi_i_p3_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_7_1, y1i);
+ psi_i_p1_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_5_1, y1i);
+ psi_i_p1_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_3_1, y1i);
+ psi_i_p1_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rmi_1_1, y1i);
+ psi_i_p1_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_1_1, y1i);
+ psi_i_p1_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_3_1, y1i);
+ psi_i_p1_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_5_1, y1i);
+ psi_i_p1_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rpi_7_1, y1i);
+ psi_i_p1_m7 = _mm256_abs_epi16(xmm2);
+
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_1, y1i);
+ psi_i_m1_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_1, y1i);
+ psi_i_m1_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_1, y1i);
+ psi_i_m1_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_1, y1i);
+ psi_i_m1_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_1, y1i);
+ psi_i_m1_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_1, y1i);
+ psi_i_m1_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_1, y1i);
+ psi_i_m1_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_1, y1i);
+ psi_i_m1_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_3, y1i);
+ psi_i_m3_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_3, y1i);
+ psi_i_m3_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_3, y1i);
+ psi_i_m3_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_3, y1i);
+ psi_i_m3_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_3, y1i);
+ psi_i_m3_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_3, y1i);
+ psi_i_m3_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_3, y1i);
+ psi_i_m3_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_3, y1i);
+ psi_i_m3_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_5, y1i);
+ psi_i_m5_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_5, y1i);
+ psi_i_m5_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_5, y1i);
+ psi_i_m5_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_5, y1i);
+ psi_i_m5_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_5, y1i);
+ psi_i_m5_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_5, y1i);
+ psi_i_m5_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_5, y1i);
+ psi_i_m5_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_5, y1i);
+ psi_i_m5_m7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_7_7, y1i);
+ psi_i_m7_p7 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_5_7, y1i);
+ psi_i_m7_p5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_3_7, y1i);
+ psi_i_m7_p3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_subs_epi16(rho_rpi_1_7, y1i);
+ psi_i_m7_p1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_1_7, y1i);
+ psi_i_m7_m1 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_3_7, y1i);
+ psi_i_m7_m3 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_5_7, y1i);
+ psi_i_m7_m5 = _mm256_abs_epi16(xmm2);
+ xmm2 = _mm256_adds_epi16(rho_rmi_7_7, y1i);
+ psi_i_m7_m7 = _mm256_abs_epi16(xmm2);
+
+ /*
+ // Rearrange desired MF output
+ xmm0 = stream0_256i_in[i];
+ xmm1 = stream0_256i_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);
+ */
+ seperate_real_imag_parts(&y0r, &y0i, stream0_256i_in[i], stream0_256i_in[i+1]);
+
+ // Rearrange desired channel magnitudes
+ // [|h|^2(1),|h|^2(1),|h|^2(2),|h|^2(2),...,,|h|^2(7),|h|^2(7)]*(2/sqrt(10))
+ /*
+ xmm2 = ch_mag_256i[i];
+ xmm3 = ch_mag_256i[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_des = _mm_unpacklo_epi64(xmm2,xmm3);
+ */
+ // xmm2 is dummy variable that contains the same values as ch_mag_des
+ seperate_real_imag_parts(&ch_mag_des, &xmm2, ch_mag_256i[i], ch_mag_256i[i+1]);
+
+
+ // Rearrange interfering channel magnitudes
+ /*
+ xmm2 = ch_mag_256i_i[i];
+ xmm3 = ch_mag_256i_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);
+ */
+ seperate_real_imag_parts(&ch_mag_int, &xmm2, ch_mag_256i_i[i], ch_mag_256i_i[i+1]);
+
+ y0r_one_over_sqrt_21 = _mm256_mulhi_epi16(y0r, ONE_OVER_SQRT_42);
+ y0r_three_over_sqrt_21 = _mm256_mulhi_epi16(y0r, THREE_OVER_SQRT_42);
+ y0r_five_over_sqrt_21 = _mm256_mulhi_epi16(y0r, FIVE_OVER_SQRT_42);
+ y0r_five_over_sqrt_21 = _mm256_slli_epi16(y0r_five_over_sqrt_21, 1);
+ y0r_seven_over_sqrt_21 = _mm256_mulhi_epi16(y0r, SEVEN_OVER_SQRT_42);
+ y0r_seven_over_sqrt_21 = _mm256_slli_epi16(y0r_seven_over_sqrt_21, 2); // Q2.14
+
+ y0i_one_over_sqrt_21 = _mm256_mulhi_epi16(y0i, ONE_OVER_SQRT_42);
+ y0i_three_over_sqrt_21 = _mm256_mulhi_epi16(y0i, THREE_OVER_SQRT_42);
+ y0i_five_over_sqrt_21 = _mm256_mulhi_epi16(y0i, FIVE_OVER_SQRT_42);
+ y0i_five_over_sqrt_21 = _mm256_slli_epi16(y0i_five_over_sqrt_21, 1);
+ y0i_seven_over_sqrt_21 = _mm256_mulhi_epi16(y0i, SEVEN_OVER_SQRT_42);
+ y0i_seven_over_sqrt_21 = _mm256_slli_epi16(y0i_seven_over_sqrt_21, 2); // Q2.14
+
+
+ y0_p_7_1 = _mm256_adds_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_p_7_3 = _mm256_adds_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_p_7_5 = _mm256_adds_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_p_7_7 = _mm256_adds_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21);
+ y0_p_5_1 = _mm256_adds_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_p_5_3 = _mm256_adds_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_p_5_5 = _mm256_adds_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_p_5_7 = _mm256_adds_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21);
+ y0_p_3_1 = _mm256_adds_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_p_3_3 = _mm256_adds_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_p_3_5 = _mm256_adds_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_p_3_7 = _mm256_adds_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21);
+ y0_p_1_1 = _mm256_adds_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_p_1_3 = _mm256_adds_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_p_1_5 = _mm256_adds_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_p_1_7 = _mm256_adds_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21);
+
+ y0_m_1_1 = _mm256_subs_epi16(y0r_one_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_m_1_3 = _mm256_subs_epi16(y0r_one_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_m_1_5 = _mm256_subs_epi16(y0r_one_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_m_1_7 = _mm256_subs_epi16(y0r_one_over_sqrt_21, y0i_seven_over_sqrt_21);
+ y0_m_3_1 = _mm256_subs_epi16(y0r_three_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_m_3_3 = _mm256_subs_epi16(y0r_three_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_m_3_5 = _mm256_subs_epi16(y0r_three_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_m_3_7 = _mm256_subs_epi16(y0r_three_over_sqrt_21, y0i_seven_over_sqrt_21);
+ y0_m_5_1 = _mm256_subs_epi16(y0r_five_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_m_5_3 = _mm256_subs_epi16(y0r_five_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_m_5_5 = _mm256_subs_epi16(y0r_five_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_m_5_7 = _mm256_subs_epi16(y0r_five_over_sqrt_21, y0i_seven_over_sqrt_21);
+ y0_m_7_1 = _mm256_subs_epi16(y0r_seven_over_sqrt_21, y0i_one_over_sqrt_21);
+ y0_m_7_3 = _mm256_subs_epi16(y0r_seven_over_sqrt_21, y0i_three_over_sqrt_21);
+ y0_m_7_5 = _mm256_subs_epi16(y0r_seven_over_sqrt_21, y0i_five_over_sqrt_21);
+ y0_m_7_7 = _mm256_subs_epi16(y0r_seven_over_sqrt_21, y0i_seven_over_sqrt_21);
+
+ // Detection of interference term
+ ch_mag_int_with_sigma2 = _mm256_srai_epi16(ch_mag_int, 1); // *2
+ two_ch_mag_int_with_sigma2 = ch_mag_int; // *4
+ three_ch_mag_int_with_sigma2 = _mm256_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 = _mm256_mulhi_epi16(psi_a_p7_p7, ONE_OVER_SQRT_2);
+ psi_a_p7_p7 = _mm256_slli_epi16(psi_a_p7_p7, 2);
+ psi_a_p7_p5 = _mm256_mulhi_epi16(psi_a_p7_p5, ONE_OVER_SQRT_2);
+ psi_a_p7_p5 = _mm256_slli_epi16(psi_a_p7_p5, 2);
+ psi_a_p7_p3 = _mm256_mulhi_epi16(psi_a_p7_p3, ONE_OVER_SQRT_2);
+ psi_a_p7_p3 = _mm256_slli_epi16(psi_a_p7_p3, 2);
+ psi_a_p7_p1 = _mm256_mulhi_epi16(psi_a_p7_p1, ONE_OVER_SQRT_2);
+ psi_a_p7_p1 = _mm256_slli_epi16(psi_a_p7_p1, 2);
+ psi_a_p7_m1 = _mm256_mulhi_epi16(psi_a_p7_m1, ONE_OVER_SQRT_2);
+ psi_a_p7_m1 = _mm256_slli_epi16(psi_a_p7_m1, 2);
+ psi_a_p7_m3 = _mm256_mulhi_epi16(psi_a_p7_m3, ONE_OVER_SQRT_2);
+ psi_a_p7_m3 = _mm256_slli_epi16(psi_a_p7_m3, 2);
+ psi_a_p7_m5 = _mm256_mulhi_epi16(psi_a_p7_m5, ONE_OVER_SQRT_2);
+ psi_a_p7_m5 = _mm256_slli_epi16(psi_a_p7_m5, 2);
+ psi_a_p7_m7 = _mm256_mulhi_epi16(psi_a_p7_m7, ONE_OVER_SQRT_2);
+ psi_a_p7_m7 = _mm256_slli_epi16(psi_a_p7_m7, 2);
+ psi_a_p5_p7 = _mm256_mulhi_epi16(psi_a_p5_p7, ONE_OVER_SQRT_2);
+ psi_a_p5_p7 = _mm256_slli_epi16(psi_a_p5_p7, 2);
+ psi_a_p5_p5 = _mm256_mulhi_epi16(psi_a_p5_p5, ONE_OVER_SQRT_2);
+ psi_a_p5_p5 = _mm256_slli_epi16(psi_a_p5_p5, 2);
+ psi_a_p5_p3 = _mm256_mulhi_epi16(psi_a_p5_p3, ONE_OVER_SQRT_2);
+ psi_a_p5_p3 = _mm256_slli_epi16(psi_a_p5_p3, 2);
+ psi_a_p5_p1 = _mm256_mulhi_epi16(psi_a_p5_p1, ONE_OVER_SQRT_2);
+ psi_a_p5_p1 = _mm256_slli_epi16(psi_a_p5_p1, 2);
+ psi_a_p5_m1 = _mm256_mulhi_epi16(psi_a_p5_m1, ONE_OVER_SQRT_2);
+ psi_a_p5_m1 = _mm256_slli_epi16(psi_a_p5_m1, 2);
+ psi_a_p5_m3 = _mm256_mulhi_epi16(psi_a_p5_m3, ONE_OVER_SQRT_2);
+ psi_a_p5_m3 = _mm256_slli_epi16(psi_a_p5_m3, 2);
+ psi_a_p5_m5 = _mm256_mulhi_epi16(psi_a_p5_m5, ONE_OVER_SQRT_2);
+ psi_a_p5_m5 = _mm256_slli_epi16(psi_a_p5_m5, 2);
+ psi_a_p5_m7 = _mm256_mulhi_epi16(psi_a_p5_m7, ONE_OVER_SQRT_2);
+ psi_a_p5_m7 = _mm256_slli_epi16(psi_a_p5_m7, 2);
+ psi_a_p3_p7 = _mm256_mulhi_epi16(psi_a_p3_p7, ONE_OVER_SQRT_2);
+ psi_a_p3_p7 = _mm256_slli_epi16(psi_a_p3_p7, 2);
+ psi_a_p3_p5 = _mm256_mulhi_epi16(psi_a_p3_p5, ONE_OVER_SQRT_2);
+ psi_a_p3_p5 = _mm256_slli_epi16(psi_a_p3_p5, 2);
+ psi_a_p3_p3 = _mm256_mulhi_epi16(psi_a_p3_p3, ONE_OVER_SQRT_2);
+ psi_a_p3_p3 = _mm256_slli_epi16(psi_a_p3_p3, 2);
+ psi_a_p3_p1 = _mm256_mulhi_epi16(psi_a_p3_p1, ONE_OVER_SQRT_2);
+ psi_a_p3_p1 = _mm256_slli_epi16(psi_a_p3_p1, 2);
+ psi_a_p3_m1 = _mm256_mulhi_epi16(psi_a_p3_m1, ONE_OVER_SQRT_2);
+ psi_a_p3_m1 = _mm256_slli_epi16(psi_a_p3_m1, 2);
+ psi_a_p3_m3 = _mm256_mulhi_epi16(psi_a_p3_m3, ONE_OVER_SQRT_2);
+ psi_a_p3_m3 = _mm256_slli_epi16(psi_a_p3_m3, 2);
+ psi_a_p3_m5 = _mm256_mulhi_epi16(psi_a_p3_m5, ONE_OVER_SQRT_2);
+ psi_a_p3_m5 = _mm256_slli_epi16(psi_a_p3_m5, 2);
+ psi_a_p3_m7 = _mm256_mulhi_epi16(psi_a_p3_m7, ONE_OVER_SQRT_2);
+ psi_a_p3_m7 = _mm256_slli_epi16(psi_a_p3_m7, 2);
+ psi_a_p1_p7 = _mm256_mulhi_epi16(psi_a_p1_p7, ONE_OVER_SQRT_2);
+ psi_a_p1_p7 = _mm256_slli_epi16(psi_a_p1_p7, 2);
+ psi_a_p1_p5 = _mm256_mulhi_epi16(psi_a_p1_p5, ONE_OVER_SQRT_2);
+ psi_a_p1_p5 = _mm256_slli_epi16(psi_a_p1_p5, 2);
+ psi_a_p1_p3 = _mm256_mulhi_epi16(psi_a_p1_p3, ONE_OVER_SQRT_2);
+ psi_a_p1_p3 = _mm256_slli_epi16(psi_a_p1_p3, 2);
+ psi_a_p1_p1 = _mm256_mulhi_epi16(psi_a_p1_p1, ONE_OVER_SQRT_2);
+ psi_a_p1_p1 = _mm256_slli_epi16(psi_a_p1_p1, 2);
+ psi_a_p1_m1 = _mm256_mulhi_epi16(psi_a_p1_m1, ONE_OVER_SQRT_2);
+ psi_a_p1_m1 = _mm256_slli_epi16(psi_a_p1_m1, 2);
+ psi_a_p1_m3 = _mm256_mulhi_epi16(psi_a_p1_m3, ONE_OVER_SQRT_2);
+ psi_a_p1_m3 = _mm256_slli_epi16(psi_a_p1_m3, 2);
+ psi_a_p1_m5 = _mm256_mulhi_epi16(psi_a_p1_m5, ONE_OVER_SQRT_2);
+ psi_a_p1_m5 = _mm256_slli_epi16(psi_a_p1_m5, 2);
+ psi_a_p1_m7 = _mm256_mulhi_epi16(psi_a_p1_m7, ONE_OVER_SQRT_2);
+ psi_a_p1_m7 = _mm256_slli_epi16(psi_a_p1_m7, 2);
+ psi_a_m1_p7 = _mm256_mulhi_epi16(psi_a_m1_p7, ONE_OVER_SQRT_2);
+ psi_a_m1_p7 = _mm256_slli_epi16(psi_a_m1_p7, 2);
+ psi_a_m1_p5 = _mm256_mulhi_epi16(psi_a_m1_p5, ONE_OVER_SQRT_2);
+ psi_a_m1_p5 = _mm256_slli_epi16(psi_a_m1_p5, 2);
+ psi_a_m1_p3 = _mm256_mulhi_epi16(psi_a_m1_p3, ONE_OVER_SQRT_2);
+ psi_a_m1_p3 = _mm256_slli_epi16(psi_a_m1_p3, 2);
+ psi_a_m1_p1 = _mm256_mulhi_epi16(psi_a_m1_p1, ONE_OVER_SQRT_2);
+ psi_a_m1_p1 = _mm256_slli_epi16(psi_a_m1_p1, 2);
+ psi_a_m1_m1 = _mm256_mulhi_epi16(psi_a_m1_m1, ONE_OVER_SQRT_2);
+ psi_a_m1_m1 = _mm256_slli_epi16(psi_a_m1_m1, 2);
+ psi_a_m1_m3 = _mm256_mulhi_epi16(psi_a_m1_m3, ONE_OVER_SQRT_2);
+ psi_a_m1_m3 = _mm256_slli_epi16(psi_a_m1_m3, 2);
+ psi_a_m1_m5 = _mm256_mulhi_epi16(psi_a_m1_m5, ONE_OVER_SQRT_2);
+ psi_a_m1_m5 = _mm256_slli_epi16(psi_a_m1_m5, 2);
+ psi_a_m1_m7 = _mm256_mulhi_epi16(psi_a_m1_m7, ONE_OVER_SQRT_2);
+ psi_a_m1_m7 = _mm256_slli_epi16(psi_a_m1_m7, 2);
+ psi_a_m3_p7 = _mm256_mulhi_epi16(psi_a_m3_p7, ONE_OVER_SQRT_2);
+ psi_a_m3_p7 = _mm256_slli_epi16(psi_a_m3_p7, 2);
+ psi_a_m3_p5 = _mm256_mulhi_epi16(psi_a_m3_p5, ONE_OVER_SQRT_2);
+ psi_a_m3_p5 = _mm256_slli_epi16(psi_a_m3_p5, 2);
+ psi_a_m3_p3 = _mm256_mulhi_epi16(psi_a_m3_p3, ONE_OVER_SQRT_2);
+ psi_a_m3_p3 = _mm256_slli_epi16(psi_a_m3_p3, 2);
+ psi_a_m3_p1 = _mm256_mulhi_epi16(psi_a_m3_p1, ONE_OVER_SQRT_2);
+ psi_a_m3_p1 = _mm256_slli_epi16(psi_a_m3_p1, 2);
+ psi_a_m3_m1 = _mm256_mulhi_epi16(psi_a_m3_m1, ONE_OVER_SQRT_2);
+ psi_a_m3_m1 = _mm256_slli_epi16(psi_a_m3_m1, 2);
+ psi_a_m3_m3 = _mm256_mulhi_epi16(psi_a_m3_m3, ONE_OVER_SQRT_2);
+ psi_a_m3_m3 = _mm256_slli_epi16(psi_a_m3_m3, 2);
+ psi_a_m3_m5 = _mm256_mulhi_epi16(psi_a_m3_m5, ONE_OVER_SQRT_2);
+ psi_a_m3_m5 = _mm256_slli_epi16(psi_a_m3_m5, 2);
+ psi_a_m3_m7 = _mm256_mulhi_epi16(psi_a_m3_m7, ONE_OVER_SQRT_2);
+ psi_a_m3_m7 = _mm256_slli_epi16(psi_a_m3_m7, 2);
+ psi_a_m5_p7 = _mm256_mulhi_epi16(psi_a_m5_p7, ONE_OVER_SQRT_2);
+ psi_a_m5_p7 = _mm256_slli_epi16(psi_a_m5_p7, 2);
+ psi_a_m5_p5 = _mm256_mulhi_epi16(psi_a_m5_p5, ONE_OVER_SQRT_2);
+ psi_a_m5_p5 = _mm256_slli_epi16(psi_a_m5_p5, 2);
+ psi_a_m5_p3 = _mm256_mulhi_epi16(psi_a_m5_p3, ONE_OVER_SQRT_2);
+ psi_a_m5_p3 = _mm256_slli_epi16(psi_a_m5_p3, 2);
+ psi_a_m5_p1 = _mm256_mulhi_epi16(psi_a_m5_p1, ONE_OVER_SQRT_2);
+ psi_a_m5_p1 = _mm256_slli_epi16(psi_a_m5_p1, 2);
+ psi_a_m5_m1 = _mm256_mulhi_epi16(psi_a_m5_m1, ONE_OVER_SQRT_2);
+ psi_a_m5_m1 = _mm256_slli_epi16(psi_a_m5_m1, 2);
+ psi_a_m5_m3 = _mm256_mulhi_epi16(psi_a_m5_m3, ONE_OVER_SQRT_2);
+ psi_a_m5_m3 = _mm256_slli_epi16(psi_a_m5_m3, 2);
+ psi_a_m5_m5 = _mm256_mulhi_epi16(psi_a_m5_m5, ONE_OVER_SQRT_2);
+ psi_a_m5_m5 = _mm256_slli_epi16(psi_a_m5_m5, 2);
+ psi_a_m5_m7 = _mm256_mulhi_epi16(psi_a_m5_m7, ONE_OVER_SQRT_2);
+ psi_a_m5_m7 = _mm256_slli_epi16(psi_a_m5_m7, 2);
+ psi_a_m7_p7 = _mm256_mulhi_epi16(psi_a_m7_p7, ONE_OVER_SQRT_2);
+ psi_a_m7_p7 = _mm256_slli_epi16(psi_a_m7_p7, 2);
+ psi_a_m7_p5 = _mm256_mulhi_epi16(psi_a_m7_p5, ONE_OVER_SQRT_2);
+ psi_a_m7_p5 = _mm256_slli_epi16(psi_a_m7_p5, 2);
+ psi_a_m7_p3 = _mm256_mulhi_epi16(psi_a_m7_p3, ONE_OVER_SQRT_2);
+ psi_a_m7_p3 = _mm256_slli_epi16(psi_a_m7_p3, 2);
+ psi_a_m7_p1 = _mm256_mulhi_epi16(psi_a_m7_p1, ONE_OVER_SQRT_2);
+ psi_a_m7_p1 = _mm256_slli_epi16(psi_a_m7_p1, 2);
+ psi_a_m7_m1 = _mm256_mulhi_epi16(psi_a_m7_m1, ONE_OVER_SQRT_2);
+ psi_a_m7_m1 = _mm256_slli_epi16(psi_a_m7_m1, 2);
+ psi_a_m7_m3 = _mm256_mulhi_epi16(psi_a_m7_m3, ONE_OVER_SQRT_2);
+ psi_a_m7_m3 = _mm256_slli_epi16(psi_a_m7_m3, 2);
+ psi_a_m7_m5 = _mm256_mulhi_epi16(psi_a_m7_m5, ONE_OVER_SQRT_2);
+ psi_a_m7_m5 = _mm256_slli_epi16(psi_a_m7_m5, 2);
+ psi_a_m7_m7 = _mm256_mulhi_epi16(psi_a_m7_m7, ONE_OVER_SQRT_2);
+ psi_a_m7_m7 = _mm256_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 = _mm256_mulhi_epi16(ch_mag_des,ONE_OVER_FOUR_SQRT_42);
+ ch_mag_2_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_2_over_42_with_sigma2,1);
+ // x=1, y=3
+ ch_mag_10_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,FIVE_OVER_FOUR_SQRT_42);
+ ch_mag_10_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_10_over_42_with_sigma2,1);
+ // x=1, x=5
+ ch_mag_26_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,THIRTEEN_OVER_FOUR_SQRT_42);
+ ch_mag_26_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_26_over_42_with_sigma2,1);
+ // x=1, y=7
+ ch_mag_50_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42);
+ ch_mag_50_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_50_over_42_with_sigma2,1);
+ // x=3, y=3
+ ch_mag_18_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,NINE_OVER_FOUR_SQRT_42);
+ ch_mag_18_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_18_over_42_with_sigma2,1);
+ // x=3, y=5
+ ch_mag_34_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,SEVENTEEN_OVER_FOUR_SQRT_42);
+ ch_mag_34_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_34_over_42_with_sigma2,1);
+ // x=3, y=7
+ ch_mag_58_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,TWENTYNINE_OVER_FOUR_SQRT_42);
+ ch_mag_58_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_58_over_42_with_sigma2,2);
+ // x=5, y=5
+ ch_mag_50_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,TWENTYFIVE_OVER_FOUR_SQRT_42);
+ ch_mag_50_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_50_over_42_with_sigma2,1);
+ // x=5, y=7
+ ch_mag_74_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,THIRTYSEVEN_OVER_FOUR_SQRT_42);
+ ch_mag_74_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_74_over_42_with_sigma2,2);
+ // x=7, y=7
+ ch_mag_98_over_42_with_sigma2 = _mm256_mulhi_epi16(ch_mag_des,FORTYNINE_OVER_FOUR_SQRT_42);
+ ch_mag_98_over_42_with_sigma2 = _mm256_slli_epi16(ch_mag_98_over_42_with_sigma2,2);
+
+ // Computing Metrics
+ xmm0 = _mm256_subs_epi16(psi_a_p7_p7, a_sq_p7_p7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_7_7);
+ bit_met_p7_p7 = _mm256_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_p5, a_sq_p7_p5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_7_5);
+ bit_met_p7_p5 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_p3, a_sq_p7_p3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_7_3);
+ bit_met_p7_p3 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_p1, a_sq_p7_p1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_7_1);
+ bit_met_p7_p1 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_m1, a_sq_p7_m1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_7_1);
+ bit_met_p7_m1 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_m3, a_sq_p7_m3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_7_3);
+ bit_met_p7_m3 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_m5, a_sq_p7_m5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_7_5);
+ bit_met_p7_m5 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p7_m7, a_sq_p7_m7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_7_7);
+ bit_met_p7_m7 = _mm256_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_p7, a_sq_p5_p7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_5_7);
+ bit_met_p5_p7 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_p5, a_sq_p5_p5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_5_5);
+ bit_met_p5_p5 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_p3, a_sq_p5_p3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_5_3);
+ bit_met_p5_p3 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_p1, a_sq_p5_p1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_5_1);
+ bit_met_p5_p1 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_m1, a_sq_p5_m1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_5_1);
+ bit_met_p5_m1 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_m3, a_sq_p5_m3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_5_3);
+ bit_met_p5_m3 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_m5, a_sq_p5_m5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_5_5);
+ bit_met_p5_m5 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p5_m7, a_sq_p5_m7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_5_7);
+ bit_met_p5_m7 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_p7, a_sq_p3_p7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_3_7);
+ bit_met_p3_p7 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_p5, a_sq_p3_p5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_3_5);
+ bit_met_p3_p5 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_p3, a_sq_p3_p3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_3_3);
+ bit_met_p3_p3 = _mm256_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_p1, a_sq_p3_p1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_3_1);
+ bit_met_p3_p1 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_m1, a_sq_p3_m1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_3_1);
+ bit_met_p3_m1 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_m3, a_sq_p3_m3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_3_3);
+ bit_met_p3_m3 = _mm256_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_m5, a_sq_p3_m5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_3_5);
+ bit_met_p3_m5 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p3_m7, a_sq_p3_m7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_3_7);
+ bit_met_p3_m7 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_p7, a_sq_p1_p7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_1_7);
+ bit_met_p1_p7 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_p5, a_sq_p1_p5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_1_5);
+ bit_met_p1_p5 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_p3, a_sq_p1_p3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_1_3);
+ bit_met_p1_p3 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_p1, a_sq_p1_p1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_p_1_1);
+ bit_met_p1_p1 = _mm256_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_m1, a_sq_p1_m1);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_1_1);
+ bit_met_p1_m1 = _mm256_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_m3, a_sq_p1_m3);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_1_3);
+ bit_met_p1_m3 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_m5, a_sq_p1_m5);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_1_5);
+ bit_met_p1_m5 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_p1_m7, a_sq_p1_m7);
+ xmm1 = _mm256_adds_epi16(xmm0, y0_m_1_7);
+ bit_met_p1_m7 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+
+ xmm0 = _mm256_subs_epi16(psi_a_m1_p7, a_sq_m1_p7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_1_7);
+ bit_met_m1_p7 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_p5, a_sq_m1_p5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_1_5);
+ bit_met_m1_p5 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_p3, a_sq_m1_p3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_1_3);
+ bit_met_m1_p3 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_p1, a_sq_m1_p1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_1_1);
+ bit_met_m1_p1 = _mm256_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_m1, a_sq_m1_m1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_1_1);
+ bit_met_m1_m1 = _mm256_subs_epi16(xmm1, ch_mag_2_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_m3, a_sq_m1_m3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_1_3);
+ bit_met_m1_m3 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_m5, a_sq_m1_m5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_1_5);
+ bit_met_m1_m5 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m1_m7, a_sq_m1_m7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_1_7);
+ bit_met_m1_m7 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_p7, a_sq_m3_p7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_3_7);
+ bit_met_m3_p7 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_p5, a_sq_m3_p5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_3_5);
+ bit_met_m3_p5 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_p3, a_sq_m3_p3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_3_3);
+ bit_met_m3_p3 = _mm256_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_p1, a_sq_m3_p1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_3_1);
+ bit_met_m3_p1 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_m1, a_sq_m3_m1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_3_1);
+ bit_met_m3_m1 = _mm256_subs_epi16(xmm1, ch_mag_10_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_m3, a_sq_m3_m3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_3_3);
+ bit_met_m3_m3 = _mm256_subs_epi16(xmm1, ch_mag_18_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_m5, a_sq_m3_m5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_3_5);
+ bit_met_m3_m5 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m3_m7, a_sq_m3_m7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_3_7);
+ bit_met_m3_m7 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_p7, a_sq_m5_p7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_5_7);
+ bit_met_m5_p7 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_p5, a_sq_m5_p5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_5_5);
+ bit_met_m5_p5 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_p3, a_sq_m5_p3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_5_3);
+ bit_met_m5_p3 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_p1, a_sq_m5_p1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_5_1);
+ bit_met_m5_p1 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_m1, a_sq_m5_m1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_5_1);
+ bit_met_m5_m1 = _mm256_subs_epi16(xmm1, ch_mag_26_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_m3, a_sq_m5_m3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_5_3);
+ bit_met_m5_m3 = _mm256_subs_epi16(xmm1, ch_mag_34_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_m5, a_sq_m5_m5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_5_5);
+ bit_met_m5_m5 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m5_m7, a_sq_m5_m7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_5_7);
+ bit_met_m5_m7 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_p7, a_sq_m7_p7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_7_7);
+ bit_met_m7_p7 = _mm256_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_p5, a_sq_m7_p5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_7_5);
+ bit_met_m7_p5 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_p3, a_sq_m7_p3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_7_3);
+ bit_met_m7_p3 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_p1, a_sq_m7_p1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_m_7_1);
+ bit_met_m7_p1 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_m1, a_sq_m7_m1);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_7_1);
+ bit_met_m7_m1 = _mm256_subs_epi16(xmm1, ch_mag_50_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_m3, a_sq_m7_m3);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_7_3);
+ bit_met_m7_m3 = _mm256_subs_epi16(xmm1, ch_mag_58_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_m5, a_sq_m7_m5);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_7_5);
+ bit_met_m7_m5 = _mm256_subs_epi16(xmm1, ch_mag_74_over_42_with_sigma2);
+ xmm0 = _mm256_subs_epi16(psi_a_m7_m7, a_sq_m7_m7);
+ xmm1 = _mm256_subs_epi16(xmm0, y0_p_7_7);
+ bit_met_m7_m7 = _mm256_subs_epi16(xmm1, ch_mag_98_over_42_with_sigma2);
+
+ // Detection for 1st bit (LTE mapping)
+ // bit = 1
+ xmm0 = _mm256_max_epi16(bit_met_m7_p7, bit_met_m7_p5);
+ xmm1 = _mm256_max_epi16(bit_met_m7_p3, bit_met_m7_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m7_m1, bit_met_m7_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m7_m5, bit_met_m7_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m5_p7, bit_met_m5_p5);
+ xmm1 = _mm256_max_epi16(bit_met_m5_p3, bit_met_m5_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m5_m1, bit_met_m5_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m5, bit_met_m5_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m3_p7, bit_met_m3_p5);
+ xmm1 = _mm256_max_epi16(bit_met_m3_p3, bit_met_m3_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m3_m1, bit_met_m3_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m3_m5, bit_met_m3_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m1_p7, bit_met_m1_p5);
+ xmm1 = _mm256_max_epi16(bit_met_m1_p3, bit_met_m1_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m1, bit_met_m1_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m1_m5, bit_met_m1_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+
+ // bit = 0
+ xmm0 = _mm256_max_epi16(bit_met_p7_p7, bit_met_p7_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p7_p3, bit_met_p7_p1);
+ xmm2 = _mm256_max_epi16(bit_met_p7_m1, bit_met_p7_m3);
+ xmm3 = _mm256_max_epi16(bit_met_p7_m5, bit_met_p7_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p5_p7, bit_met_p5_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p5_p3, bit_met_p5_p1);
+ xmm2 = _mm256_max_epi16(bit_met_p5_m1, bit_met_p5_m3);
+ xmm3 = _mm256_max_epi16(bit_met_p5_m5, bit_met_p5_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p3_p7, bit_met_p3_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p3, bit_met_p3_p1);
+ xmm2 = _mm256_max_epi16(bit_met_p3_m1, bit_met_p3_m3);
+ xmm3 = _mm256_max_epi16(bit_met_p3_m5, bit_met_p3_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p1_p7, bit_met_p1_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p1_p3, bit_met_p1_p1);
+ xmm2 = _mm256_max_epi16(bit_met_p1_m1, bit_met_p1_m3);
+ xmm3 = _mm256_max_epi16(bit_met_p1_m5, bit_met_p1_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+
+ y0r = _mm256_subs_epi16(logmax_num_re0, logmax_den_re0);
+
+ // Detection for 2nd bit (LTE mapping)
+ // bit = 1
+ xmm0 = _mm256_max_epi16(bit_met_p7_m1, bit_met_p5_m1);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m1, bit_met_p1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m1, bit_met_m3_m1);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m1, bit_met_m7_m1);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m3, bit_met_p5_m3);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m3, bit_met_p1_m3);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m3, bit_met_m3_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m3, bit_met_m7_m3);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m5, bit_met_p5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m5, bit_met_p1_m5);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m5, bit_met_m3_m5);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m5, bit_met_m7_m5);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m7, bit_met_p5_m7);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m7, bit_met_p1_m7);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m7, bit_met_m3_m7);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m7, bit_met_m7_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+
+ // bit = 0
+ xmm0 = _mm256_max_epi16(bit_met_p7_p1, bit_met_p5_p1);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p1, bit_met_p1_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p1, bit_met_m3_p1);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p1, bit_met_m7_p1);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p3, bit_met_p5_p3);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p3, bit_met_p1_p3);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p3, bit_met_m3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p3, bit_met_m7_p3);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p5, bit_met_p5_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p5, bit_met_p1_p5);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p5, bit_met_m3_p5);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p5, bit_met_m7_p5);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p7, bit_met_p5_p7);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p7, bit_met_p1_p7);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p7, bit_met_m3_p7);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p7, bit_met_m7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+
+ y1r = _mm256_subs_epi16(logmax_num_re0, logmax_den_re0);
+
+ // Detection for 3rd bit (LTE mapping)
+ xmm0 = _mm256_max_epi16(bit_met_m7_m7, bit_met_m7_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m7_m3, bit_met_m7_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m7_p1, bit_met_m7_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m7_p5, bit_met_m7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m5_m7, bit_met_m5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m5_m3, bit_met_m5_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m5_p1, bit_met_m5_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p5, bit_met_m5_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p5_m7, bit_met_p5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p5_m3, bit_met_p5_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p5_p1, bit_met_p5_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p5_p5, bit_met_p5_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m7, bit_met_p7_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p7_m3, bit_met_p7_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p7_p1, bit_met_p7_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p7_p5, bit_met_p7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+
+ xmm0 = _mm256_max_epi16(bit_met_m3_m7, bit_met_m3_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m3_m3, bit_met_m3_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m3_p1, bit_met_m3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m3_p5, bit_met_m3_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m1_m7, bit_met_m1_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m1_m3, bit_met_m1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p1, bit_met_m1_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m1_p5, bit_met_m1_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p1_m7, bit_met_p1_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p1_m3, bit_met_p1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p1_p1, bit_met_p1_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p1_p5, bit_met_p1_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p3_m7, bit_met_p3_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m3, bit_met_p3_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p3_p1, bit_met_p3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p3_p5, bit_met_p3_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+
+ y2r = _mm256_subs_epi16(logmax_num_re0, logmax_den_re0);
+
+ // Detection for 4th bit (LTE mapping)
+ xmm0 = _mm256_max_epi16(bit_met_p7_p7, bit_met_p5_p7);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p7, bit_met_p1_p7);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p7, bit_met_m3_p7);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p7, bit_met_m7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p5, bit_met_p5_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p5, bit_met_p1_p5);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p5, bit_met_m3_p5);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p5, bit_met_m5_p5);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m5, bit_met_p5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m5, bit_met_p1_m5);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m5, bit_met_m3_m5);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m5, bit_met_m7_m5);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m7, bit_met_p5_m7);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m7, bit_met_p1_m7);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m7, bit_met_m3_m7);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m7, bit_met_m7_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+
+ xmm0 = _mm256_max_epi16(bit_met_p7_m1, bit_met_p5_m1);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m1, bit_met_p1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m1, bit_met_m3_m1);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m1, bit_met_m7_m1);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m3, bit_met_p5_m3);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m3, bit_met_p1_m3);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m3, bit_met_m3_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m3, bit_met_m7_m3);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p1, bit_met_p5_p1);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p1, bit_met_p1_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p1, bit_met_m3_p1);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p1, bit_met_m7_p1);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p3, bit_met_p5_p3);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p3, bit_met_p1_p3);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p3, bit_met_m3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p3, bit_met_m7_p3);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+
+ y0i = _mm256_subs_epi16(logmax_num_re0, logmax_den_re0);
+
+
+ // Detection for 5th bit (LTE mapping)
+ xmm0 = _mm256_max_epi16(bit_met_m7_m7, bit_met_m7_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m7_m3, bit_met_m7_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m7_p1, bit_met_m7_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m7_p5, bit_met_m7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m1_m7, bit_met_m1_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m1_m3, bit_met_m1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p1, bit_met_m1_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m1_p5, bit_met_m1_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p1_m7, bit_met_p1_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p1_m3, bit_met_p1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p1_p1, bit_met_p1_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p1_p5, bit_met_p1_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m7, bit_met_p7_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p7_m3, bit_met_p7_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p7_p1, bit_met_p7_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p7_p5, bit_met_p7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+
+ xmm0 = _mm256_max_epi16(bit_met_m5_m7, bit_met_m5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m5_m3, bit_met_m5_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m5_p1, bit_met_m5_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p5, bit_met_m5_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_m3_m7, bit_met_m3_m5);
+ xmm1 = _mm256_max_epi16(bit_met_m3_m3, bit_met_m3_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m3_p1, bit_met_m3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m3_p5, bit_met_m3_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p3_m7, bit_met_p3_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m3, bit_met_p3_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p3_p1, bit_met_p3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p3_p5, bit_met_p3_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p5_m7, bit_met_p5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p5_m3, bit_met_p5_m1);
+ xmm2 = _mm256_max_epi16(bit_met_p5_p1, bit_met_p5_p3);
+ xmm3 = _mm256_max_epi16(bit_met_p5_p5, bit_met_p5_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+
+ y1i = _mm256_subs_epi16(logmax_num_re0, logmax_den_re0);
+
+ // Detection for 6th bit (LTE mapping)
+ xmm0 = _mm256_max_epi16(bit_met_p7_p7, bit_met_p5_p7);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p7, bit_met_p1_p7);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p7, bit_met_m3_p7);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p7, bit_met_m7_p7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p1, bit_met_p5_p1);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p1, bit_met_p1_p1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p1, bit_met_m3_p1);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p1, bit_met_m5_p1);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m1, bit_met_p5_m1);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m1, bit_met_p1_m1);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m1, bit_met_m3_m1);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m1, bit_met_m7_m1);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m7, bit_met_p5_m7);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m7, bit_met_p1_m7);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m7, bit_met_m3_m7);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m7, bit_met_m7_m7);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm4);
+ logmax_den_re0 = _mm256_max_epi16(logmax_den_re0, xmm5);
+
+ xmm0 = _mm256_max_epi16(bit_met_p7_m5, bit_met_p5_m5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m5, bit_met_p1_m5);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m5, bit_met_m3_m5);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m5, bit_met_m7_m5);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(xmm4, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_m3, bit_met_p5_m3);
+ xmm1 = _mm256_max_epi16(bit_met_p3_m3, bit_met_p1_m3);
+ xmm2 = _mm256_max_epi16(bit_met_m1_m3, bit_met_m3_m3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_m3, bit_met_m7_m3);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p3, bit_met_p5_p3);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p3, bit_met_p1_p3);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p3, bit_met_m3_p3);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p3, bit_met_m7_p3);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+ xmm0 = _mm256_max_epi16(bit_met_p7_p5, bit_met_p5_p5);
+ xmm1 = _mm256_max_epi16(bit_met_p3_p5, bit_met_p1_p5);
+ xmm2 = _mm256_max_epi16(bit_met_m1_p5, bit_met_m3_p5);
+ xmm3 = _mm256_max_epi16(bit_met_m5_p5, bit_met_m7_p5);
+ xmm4 = _mm256_max_epi16(xmm0, xmm1);
+ xmm5 = _mm256_max_epi16(xmm2, xmm3);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm4);
+ logmax_num_re0 = _mm256_max_epi16(logmax_num_re0, xmm5);
+
+ y2i = _mm256_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 = 48*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];
+
+ // RE 9
+ stream0_out[j + 48] = ((short *)&y0r)[8];
+ stream0_out[j + 49] = ((short *)&y1r)[8];
+ stream0_out[j + 50] = ((short *)&y2r)[8];
+ stream0_out[j + 51] = ((short *)&y0i)[8];
+ stream0_out[j + 52] = ((short *)&y1i)[8];
+ stream0_out[j + 53] = ((short *)&y2i)[8];
+ // RE 10
+ stream0_out[j + 54] = ((short *)&y0r)[9];
+ stream0_out[j + 55] = ((short *)&y1r)[9];
+ stream0_out[j + 56] = ((short *)&y2r)[9];
+ stream0_out[j + 57] = ((short *)&y0i)[9];
+ stream0_out[j + 58] = ((short *)&y1i)[9];
+ stream0_out[j + 59] = ((short *)&y2i)[9];
+ // RE 11
+ stream0_out[j + 60] = ((short *)&y0r)[10];
+ stream0_out[j + 61] = ((short *)&y1r)[10];
+ stream0_out[j + 62] = ((short *)&y2r)[10];
+ stream0_out[j + 63] = ((short *)&y0i)[10];
+ stream0_out[j + 64] = ((short *)&y1i)[10];
+ stream0_out[j + 65] = ((short *)&y2i)[10];
+ // RE 12
+ stream0_out[j + 66] = ((short *)&y0r)[11];
+ stream0_out[j + 67] = ((short *)&y1r)[11];
+ stream0_out[j + 68] = ((short *)&y2r)[11];
+ stream0_out[j + 69] = ((short *)&y0i)[11];
+ stream0_out[j + 70] = ((short *)&y1i)[11];
+ stream0_out[j + 71] = ((short *)&y2i)[11];
+ // RE 13
+ stream0_out[j + 72] = ((short *)&y0r)[12];
+ stream0_out[j + 73] = ((short *)&y1r)[12];
+ stream0_out[j + 74] = ((short *)&y2r)[12];
+ stream0_out[j + 75] = ((short *)&y0i)[12];
+ stream0_out[j + 76] = ((short *)&y1i)[12];
+ stream0_out[j + 77] = ((short *)&y2i)[12];
+ // RE 14
+ stream0_out[j + 78] = ((short *)&y0r)[13];
+ stream0_out[j + 79] = ((short *)&y1r)[13];
+ stream0_out[j + 80] = ((short *)&y2r)[13];
+ stream0_out[j + 81] = ((short *)&y0i)[13];
+ stream0_out[j + 82] = ((short *)&y1i)[13];
+ stream0_out[j + 83] = ((short *)&y2i)[13];
+ // RE 15
+ stream0_out[j + 84] = ((short *)&y0r)[14];
+ stream0_out[j + 85] = ((short *)&y1r)[14];
+ stream0_out[j + 86] = ((short *)&y2r)[14];
+ stream0_out[j + 87] = ((short *)&y0i)[14];
+ stream0_out[j + 88] = ((short *)&y1i)[14];
+ stream0_out[j + 89] = ((short *)&y2i)[14];
+ // RE 16
+ stream0_out[j + 90] = ((short *)&y0r)[15];
+ stream0_out[j + 91] = ((short *)&y1r)[15];
+ stream0_out[j + 92] = ((short *)&y2r)[15];
+ stream0_out[j + 93] = ((short *)&y0i)[15];
+ stream0_out[j + 94] = ((short *)&y1i)[15];
+ stream0_out[j + 95] = ((short *)&y2i)[15];
+
+#elif defined(__arm__)
+
+#endif
+
+ }
+
+#if defined(__x86_64__) || defined(__i386__)
+ _mm_empty();
+ _m_empty();
+#endif
+}
diff --git a/openair1/PHY/LTE_TRANSPORT/print_stats.c b/openair1/PHY/LTE_TRANSPORT/print_stats.c
index c8d4292e8d03c8866f6a053ce1c7a45d37b591ad..c3e40e1e72f323b108d740b7d5db815f76cf8cd1 100644
--- a/openair1/PHY/LTE_TRANSPORT/print_stats.c
+++ b/openair1/PHY/LTE_TRANSPORT/print_stats.c
@@ -107,9 +107,9 @@ int dump_ue_stats(PHY_VARS_UE *ue, UE_rxtx_proc_t *proc,char* buffer, int length
len += sprintf(&buffer[len], "[UE PROC] Po_PUCCH = %d dBm (Po_NOMINAL_PUCCH %d dBm, g_pucch %d dB)\n",
get_PL(ue->Mod_id,ue->CC_id,0)+
ue->frame_parms.ul_power_control_config_common.p0_NominalPUCCH+
- ue->dlsch[0][0]->g_pucch,
+ ue->dlsch[0][0][0]->g_pucch,
ue->frame_parms.ul_power_control_config_common.p0_NominalPUCCH,
- ue->dlsch[0][0]->g_pucch);
+ ue->dlsch[0][0][0]->g_pucch);
}
//for (eNB=0;eNB<NUMBER_OF_eNB_MAX;eNB++) {
for (eNB=0; eNB<1; eNB++) {
@@ -482,24 +482,24 @@ int dump_ue_stats(PHY_VARS_UE *ue, UE_rxtx_proc_t *proc,char* buffer, int length
len += sprintf(&buffer[len], "[UE PROC] Mode 6 Wideband CQI eNB %d : %d dB\n",eNB,ue->measurements.precoded_cqi_dB[eNB][0]);
for (harq_pid=0;harq_pid<8;harq_pid++) {
- len+=sprintf(&buffer[len],"[UE PROC] eNB %d: CW 0 harq_pid %d, mcs %d:",eNB,harq_pid,ue->dlsch[0][0]->harq_processes[harq_pid]->mcs);
+ len+=sprintf(&buffer[len],"[UE PROC] eNB %d: CW 0 harq_pid %d, mcs %d:",eNB,harq_pid,ue->dlsch[0][0][0]->harq_processes[harq_pid]->mcs);
for (round=0;round<8;round++)
len+=sprintf(&buffer[len],"%d/%d ",
- ue->dlsch[0][0]->harq_processes[harq_pid]->errors[round],
- ue->dlsch[0][0]->harq_processes[harq_pid]->trials[round]);
+ ue->dlsch[0][0][0]->harq_processes[harq_pid]->errors[round],
+ ue->dlsch[0][0][0]->harq_processes[harq_pid]->trials[round]);
len+=sprintf(&buffer[len],"\n");
}
- if (ue->dlsch[0] && ue->dlsch[0][0] && ue->dlsch[0][1]) {
- len += sprintf(&buffer[len], "[UE PROC] Saved PMI for DLSCH eNB %d : %jx (%p)\n",eNB,pmi2hex_2Ar1(ue->dlsch[0][0]->pmi_alloc),ue->dlsch[0][0]);
+ if (ue->dlsch[0][0] && ue->dlsch[0][0][0] && ue->dlsch[0][0][1]) {
+ len += sprintf(&buffer[len], "[UE PROC] Saved PMI for DLSCH eNB %d : %jx (%p)\n",eNB,pmi2hex_2Ar1(ue->dlsch[0][0][0]->pmi_alloc),ue->dlsch[0][0][0]);
- len += sprintf(&buffer[len], "[UE PROC] eNB %d: dl_power_off = %d\n",eNB,ue->dlsch[0][0]->harq_processes[0]->dl_power_off);
+ len += sprintf(&buffer[len], "[UE PROC] eNB %d: dl_power_off = %d\n",eNB,ue->dlsch[0][0][0]->harq_processes[0]->dl_power_off);
for (harq_pid=0;harq_pid<8;harq_pid++) {
- len+=sprintf(&buffer[len],"[UE PROC] eNB %d: CW 1 harq_pid %d, mcs %d:",eNB,harq_pid,ue->dlsch[0][1]->harq_processes[0]->mcs);
+ len+=sprintf(&buffer[len],"[UE PROC] eNB %d: CW 1 harq_pid %d, mcs %d:",eNB,harq_pid,ue->dlsch[0][0][1]->harq_processes[0]->mcs);
for (round=0;round<8;round++)
len+=sprintf(&buffer[len],"%d/%d ",
- ue->dlsch[0][1]->harq_processes[harq_pid]->errors[round],
- ue->dlsch[0][1]->harq_processes[harq_pid]->trials[round]);
+ ue->dlsch[0][0][1]->harq_processes[harq_pid]->errors[round],
+ ue->dlsch[0][0][1]->harq_processes[harq_pid]->trials[round]);
len+=sprintf(&buffer[len],"\n");
}
}
diff --git a/openair1/PHY/LTE_TRANSPORT/proto.h b/openair1/PHY/LTE_TRANSPORT/proto.h
index 5fc8dea9867cee8eac0abf65256d5d0179628afc..fe7991e343ab200cf96d8d5f7a34f66a29a8fba0 100644
--- a/openair1/PHY/LTE_TRANSPORT/proto.h
+++ b/openair1/PHY/LTE_TRANSPORT/proto.h
@@ -704,6 +704,22 @@ void qam64_qam16(short *stream0_in,
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
@@ -746,6 +762,22 @@ void qam64_qam64(short *stream0_in,
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
diff --git a/openair1/PHY/LTE_TRANSPORT/ulsch_coding.c b/openair1/PHY/LTE_TRANSPORT/ulsch_coding.c
index 9571ea6cd52823d8409a2781bdd1c2ab5fc7c6e4..dbb7a16ac9cbb983a650b5811e9cfc594e2245e3 100644
--- a/openair1/PHY/LTE_TRANSPORT/ulsch_coding.c
+++ b/openair1/PHY/LTE_TRANSPORT/ulsch_coding.c
@@ -233,7 +233,7 @@ uint32_t ulsch_encoding(uint8_t *a,
LTE_DL_FRAME_PARMS *frame_parms=&ue->frame_parms;
PHY_MEASUREMENTS *meas = &ue->measurements;
LTE_UE_ULSCH_t *ulsch=ue->ulsch[eNB_id];
- LTE_UE_DLSCH_t **dlsch = ue->dlsch[eNB_id];
+ LTE_UE_DLSCH_t **dlsch = ue->dlsch[0][eNB_id];
uint16_t rnti = 0xffff;
if (!ulsch) {
@@ -966,7 +966,7 @@ int ulsch_encoding_emul(uint8_t *ulsch_buffer,
{
LTE_UE_ULSCH_t *ulsch = ue->ulsch[eNB_id];
- LTE_UE_DLSCH_t **dlsch = ue->dlsch[eNB_id];
+ LTE_UE_DLSCH_t **dlsch = ue->dlsch[0][eNB_id];
PHY_MEASUREMENTS *meas = &ue->measurements;
uint8_t tmode = ue->transmission_mode[eNB_id];
uint16_t rnti=ue->pdcch_vars[eNB_id]->crnti;
diff --git a/openair1/PHY/LTE_TRANSPORT/ulsch_decoding.c b/openair1/PHY/LTE_TRANSPORT/ulsch_decoding.c
index ca67ccdb3c9c34dbf01b7752201ebce456b34017..bb2a5b3e224652d3fc62c6f2594743d88dc30cc1 100644
--- a/openair1/PHY/LTE_TRANSPORT/ulsch_decoding.c
+++ b/openair1/PHY/LTE_TRANSPORT/ulsch_decoding.c
@@ -2066,7 +2066,7 @@ uint32_t ulsch_decoding_emul(PHY_VARS_eNB *eNB, eNB_rxtx_proc_t *proc,
// get local ue's ack
if ((UE_index >= oai_emulation.info.first_ue_local) ||(UE_index <(oai_emulation.info.first_ue_local+oai_emulation.info.nb_ue_local))) {
get_ack(&eNB->frame_parms,
- PHY_vars_UE_g[UE_id][CC_id]->dlsch[0][0]->harq_ack,
+ PHY_vars_UE_g[UE_id][CC_id]->dlsch[0][0][0]->harq_ack,
subframe,
eNB->ulsch[UE_index]->harq_processes[harq_pid]->o_ACK,0);
} else { // get remote UEs' ack
diff --git a/openair1/PHY/Makefile.inc b/openair1/PHY/Makefile.inc
index 1586f353c0f623853f6149ca91f313630847b92a..90094b31f8605946ba6c18f31eb070b4ff99f484 100644
--- a/openair1/PHY/Makefile.inc
+++ b/openair1/PHY/Makefile.inc
@@ -7,6 +7,7 @@ PHY_OBJS += $(TOP_DIR)/PHY/LTE_TRANSPORT/dlsch_coding.o
PHY_OBJS += $(TOP_DIR)/PHY/LTE_TRANSPORT/dlsch_modulation.o
PHY_OBJS += $(TOP_DIR)/PHY/LTE_TRANSPORT/dlsch_demodulation.o
PHY_OBJS += $(TOP_DIR)/PHY/LTE_TRANSPORT/dlsch_llr_computation.o
+PHY_OBJS += $(TOP_DIR)/PHY/LTE_TRANSPORT/dlsch_llr_computation_avx2.o
PHY_OBJS += $(TOP_DIR)/PHY/LTE_TRANSPORT/power_control.o
PHY_OBJS += $(TOP_DIR)/PHY/LTE_TRANSPORT/dlsch_decoding.o
PHY_OBJS += $(TOP_DIR)/PHY/LTE_TRANSPORT/dlsch_scrambling.o
diff --git a/openair1/PHY/TOOLS/lte_phy_scope.c b/openair1/PHY/TOOLS/lte_phy_scope.c
index ca08f875190d18a8def154faadd11d18418299c2..7a533464530f337d4aeba89722164670a830be63 100644
--- a/openair1/PHY/TOOLS/lte_phy_scope.c
+++ b/openair1/PHY/TOOLS/lte_phy_scope.c
@@ -510,16 +510,16 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form,
int beamforming_mode = phy_vars_ue->transmission_mode[eNB_id]>6 ? phy_vars_ue->transmission_mode[eNB_id] : 0;
- if (phy_vars_ue->dlsch[eNB_id][0]!=NULL) {
- harq_pid = phy_vars_ue->dlsch[eNB_id][0]->current_harq_pid;
+ if (phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]!=NULL) {
+ harq_pid = phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]->current_harq_pid;
if (harq_pid>=8)
return;
- mcs = phy_vars_ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->mcs;
+ mcs = phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->mcs;
// Button 0
- if(!phy_vars_ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->dl_power_off) {
+ if(!phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->dl_power_off) {
// we are in TM5
fl_show_object(form->button_0);
}
@@ -530,12 +530,12 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form,
}
// coded_bits_per_codeword = frame_parms->N_RB_DL*12*get_Qm(mcs)*(frame_parms->symbols_per_tti);
- if (phy_vars_ue->dlsch[eNB_id][0]!=NULL) {
+ if (phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]!=NULL) {
coded_bits_per_codeword = get_G(frame_parms,
- phy_vars_ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->nb_rb,
- phy_vars_ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->rb_alloc_even,
+ phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->nb_rb,
+ phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->rb_alloc_even,
get_Qm(mcs),
- phy_vars_ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->Nl,
+ phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->Nl,
num_pdcch_symbols,
frame,
subframe,
diff --git a/openair1/PHY/TOOLS/lte_phy_scope_tm4.c b/openair1/PHY/TOOLS/lte_phy_scope_tm4.c
index 60c5ace2c5379471d7746421daad79f9a1844cfc..6807f1f6f48a6de4ee057c9c2341b80084526448 100755
--- a/openair1/PHY/TOOLS/lte_phy_scope_tm4.c
+++ b/openair1/PHY/TOOLS/lte_phy_scope_tm4.c
@@ -448,11 +448,11 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form,
int mcs1=0;
unsigned char harq_pid = 0;
int beamforming_mode = phy_vars_ue->transmission_mode[eNB_id]>6 ? phy_vars_ue->transmission_mode[eNB_id] : 0;
- if (phy_vars_ue->dlsch[eNB_id][0]!=NULL) {
- harq_pid = phy_vars_ue->dlsch[eNB_id][0]->current_harq_pid;
+ if (phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]!=NULL) {
+ harq_pid = phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]->current_harq_pid;
if (harq_pid>=8)
return;
- mcs0 = phy_vars_ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->mcs;
+ mcs0 = phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->mcs;
// Button 0
/*
if(!phy_vars_ue->dlsch_ue[eNB_id][0]->harq_processes[harq_pid]->dl_power_off) {
@@ -461,23 +461,23 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form,
}
*/
}
- if (phy_vars_ue->dlsch[eNB_id][1]!=NULL) {
- harq_pid = phy_vars_ue->dlsch[eNB_id][1]->current_harq_pid;
+ if (phy_vars_ue->dlsch[subframe&0x1][eNB_id][1]!=NULL) {
+ harq_pid = phy_vars_ue->dlsch[subframe&0x1][eNB_id][1]->current_harq_pid;
if (harq_pid>=8)
return;
- mcs1 = phy_vars_ue->dlsch[eNB_id][1]->harq_processes[harq_pid]->mcs;
+ mcs1 = phy_vars_ue->dlsch[subframe&0x1][eNB_id][1]->harq_processes[harq_pid]->mcs;
}
if (phy_vars_ue->pdcch_vars[eNB_id]!=NULL) {
num_pdcch_symbols = phy_vars_ue->pdcch_vars[eNB_id]->num_pdcch_symbols;
}
// coded_bits_per_codeword = frame_parms->N_RB_DL*12*get_Qm(mcs)*(frame_parms->symbols_per_tti);
- if (phy_vars_ue->dlsch[eNB_id][0]!=NULL) {
+ if (phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]!=NULL) {
mod0 = get_Qm(mcs0);
coded_bits_per_codeword0 = get_G(frame_parms,
- phy_vars_ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->nb_rb,
- phy_vars_ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->rb_alloc_even,
+ phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->nb_rb,
+ phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->rb_alloc_even,
get_Qm(mcs0),
- phy_vars_ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->Nl,
+ phy_vars_ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->Nl,
num_pdcch_symbols,
frame,
subframe,
@@ -486,13 +486,13 @@ void phy_scope_UE(FD_lte_phy_scope_ue *form,
coded_bits_per_codeword0 = 0; //frame_parms->N_RB_DL*12*get_Qm(mcs)*(frame_parms->symbols_per_tti);
mod0=0;
}
- if (phy_vars_ue->dlsch[eNB_id][1]!=NULL) {
+ if (phy_vars_ue->dlsch[subframe&0x1][eNB_id][1]!=NULL) {
mod1 = get_Qm(mcs1);
coded_bits_per_codeword1 = get_G(frame_parms,
- phy_vars_ue->dlsch[eNB_id][1]->harq_processes[harq_pid]->nb_rb,
- phy_vars_ue->dlsch[eNB_id][1]->harq_processes[harq_pid]->rb_alloc_even,
+ phy_vars_ue->dlsch[subframe&0x1][eNB_id][1]->harq_processes[harq_pid]->nb_rb,
+ phy_vars_ue->dlsch[subframe&0x1][eNB_id][1]->harq_processes[harq_pid]->rb_alloc_even,
get_Qm(mcs1),
- phy_vars_ue->dlsch[eNB_id][1]->harq_processes[harq_pid]->Nl,
+ phy_vars_ue->dlsch[subframe&0x1][eNB_id][1]->harq_processes[harq_pid]->Nl,
num_pdcch_symbols,
frame,
subframe,
diff --git a/openair1/PHY/defs.h b/openair1/PHY/defs.h
index d70bfb3f722df0bcf526266d633b932bb534057b..5c1b1158ba45a62f9a206e45022a746a5bd23db7 100644
--- a/openair1/PHY/defs.h
+++ b/openair1/PHY/defs.h
@@ -706,7 +706,7 @@ typedef struct {
LTE_DL_FRAME_PARMS frame_parms_before_ho;
LTE_UE_COMMON common_vars;
- LTE_UE_PDSCH *pdsch_vars[2][NUMBER_OF_CONNECTED_eNB_MAX+1];
+ LTE_UE_PDSCH *pdsch_vars[2][NUMBER_OF_CONNECTED_eNB_MAX+1]; // two RxTx Threads
LTE_UE_PDSCH_FLP *pdsch_vars_flp[NUMBER_OF_CONNECTED_eNB_MAX+1];
LTE_UE_PDSCH *pdsch_vars_SI[NUMBER_OF_CONNECTED_eNB_MAX+1];
LTE_UE_PDSCH *pdsch_vars_ra[NUMBER_OF_CONNECTED_eNB_MAX+1];
@@ -715,7 +715,7 @@ typedef struct {
LTE_UE_PBCH *pbch_vars[NUMBER_OF_CONNECTED_eNB_MAX];
LTE_UE_PDCCH *pdcch_vars[NUMBER_OF_CONNECTED_eNB_MAX];
LTE_UE_PRACH *prach_vars[NUMBER_OF_CONNECTED_eNB_MAX];
- LTE_UE_DLSCH_t *dlsch[NUMBER_OF_CONNECTED_eNB_MAX][2];
+ LTE_UE_DLSCH_t *dlsch[2][NUMBER_OF_CONNECTED_eNB_MAX][2]; // two RxTx Threads
LTE_UE_ULSCH_t *ulsch[NUMBER_OF_CONNECTED_eNB_MAX];
LTE_UE_DLSCH_t *dlsch_SI[NUMBER_OF_CONNECTED_eNB_MAX];
LTE_UE_DLSCH_t *dlsch_ra[NUMBER_OF_CONNECTED_eNB_MAX];
@@ -802,6 +802,8 @@ typedef struct {
uint8_t prach_cnt;
uint8_t prach_PreambleIndex;
// uint8_t prach_timer;
+ uint8_t decode_SIB;
+ uint8_t decode_MIB;
int rx_offset; /// Timing offset
int rx_offset_diff; /// Timing adjustment for ofdm symbol0 on HW USRP
int timing_advance; ///timing advance signalled from eNB
@@ -872,7 +874,7 @@ typedef struct {
time_stats_t phy_proc;
time_stats_t phy_proc_tx;
- time_stats_t phy_proc_rx;
+ time_stats_t phy_proc_rx[2];
uint32_t use_ia_receiver;
@@ -885,6 +887,10 @@ typedef struct {
time_stats_t ulsch_interleaving_stats;
time_stats_t ulsch_multiplexing_stats;
+ time_stats_t generic_stat;
+ time_stats_t pdsch_procedures_stat;
+ time_stats_t dlsch_procedures_stat;
+
time_stats_t ofdm_demod_stats;
time_stats_t dlsch_rx_pdcch_stats;
time_stats_t rx_dft_stats;
diff --git a/openair1/SCHED/phy_procedures_lte_ue.c b/openair1/SCHED/phy_procedures_lte_ue.c
index cd8fbdbc8eec0b3130e91eb6645f9ca69f6f8278..4a75d5d9e7aedd9fdf70e7a889747be0df93ce8d 100644
--- a/openair1/SCHED/phy_procedures_lte_ue.c
+++ b/openair1/SCHED/phy_procedures_lte_ue.c
@@ -75,6 +75,7 @@ fifo_dump_emos_UE emos_dump_UE;
extern int oai_exit;
+extern double cpuf;
@@ -90,10 +91,10 @@ void dump_dlsch(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uint8_t subf
uint8_t nsymb = (ue->frame_parms.Ncp == 0) ? 14 : 12;
coded_bits_per_codeword = get_G(&ue->frame_parms,
- ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->nb_rb,
- ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->rb_alloc_even,
- ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->Qm,
- ue->dlsch[eNB_id][0]->harq_processes[harq_pid]->Nl,
+ ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->nb_rb,
+ ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->rb_alloc_even,
+ ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->Qm,
+ ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[harq_pid]->Nl,
ue->pdcch_vars[eNB_id]->num_pdcch_symbols,
proc->frame_rx,
subframe,
@@ -268,37 +269,39 @@ void phy_reset_ue(uint8_t Mod_id,uint8_t CC_id,uint8_t eNB_index)
uint8_t i,j,k,s;
PHY_VARS_UE *ue = PHY_vars_UE_g[Mod_id][CC_id];
- //[NUMBER_OF_CONNECTED_eNB_MAX][2];
- for(i=0; i<NUMBER_OF_CONNECTED_eNB_MAX; i++) {
- for(j=0; j<2; j++) {
- //DL HARQ
- if(ue->dlsch[i][j]) {
- for(k=0; k<NUMBER_OF_HARQ_PID_MAX && ue->dlsch[i][j]->harq_processes[k]; k++) {
- ue->dlsch[i][j]->harq_processes[k]->status = SCH_IDLE;
- for (s=0; s<10; s++) {
- // reset ACK/NACK bit to DTX for all subframes s = 0..9
- ue->dlsch[i][j]->harq_ack[s].ack = 2;
- ue->dlsch[i][j]->harq_ack[s].send_harq_status = 0;
- ue->dlsch[i][j]->harq_ack[s].vDAI_UL = 0xff;
- ue->dlsch[i][j]->harq_ack[s].vDAI_DL = 0xff;
+ //[NUMBER_OF_RX_THREAD=2][NUMBER_OF_CONNECTED_eNB_MAX][2];
+ for(int l=0; l<2; l++) {
+ for(i=0; i<NUMBER_OF_CONNECTED_eNB_MAX; i++) {
+ for(j=0; j<2; j++) {
+ //DL HARQ
+ if(ue->dlsch[l][i][j]) {
+ for(k=0; k<NUMBER_OF_HARQ_PID_MAX && ue->dlsch[l][i][j]->harq_processes[k]; k++) {
+ ue->dlsch[l][i][j]->harq_processes[k]->status = SCH_IDLE;
+ for (s=0; s<10; s++) {
+ // reset ACK/NACK bit to DTX for all subframes s = 0..9
+ ue->dlsch[l][i][j]->harq_ack[s].ack = 2;
+ ue->dlsch[l][i][j]->harq_ack[s].send_harq_status = 0;
+ ue->dlsch[l][i][j]->harq_ack[s].vDAI_UL = 0xff;
+ ue->dlsch[l][i][j]->harq_ack[s].vDAI_DL = 0xff;
+ }
+ }
+ }
}
- }
- }
- }
- //UL HARQ
- if(ue->ulsch[i]) {
- for(k=0; k<NUMBER_OF_HARQ_PID_MAX && ue->ulsch[i]->harq_processes[k]; k++) {
- ue->ulsch[i]->harq_processes[k]->status = SCH_IDLE;
- //Set NDIs for all UL HARQs to 0
- // ue->ulsch[i]->harq_processes[k]->Ndi = 0;
+ //UL HARQ
+ if(ue->ulsch[i]) {
+ for(k=0; k<NUMBER_OF_HARQ_PID_MAX && ue->ulsch[i]->harq_processes[k]; k++) {
+ ue->ulsch[i]->harq_processes[k]->status = SCH_IDLE;
+ //Set NDIs for all UL HARQs to 0
+ // ue->ulsch[i]->harq_processes[k]->Ndi = 0;
- }
- }
+ }
+ }
- // flush Msg3 buffer
- ue->ulsch_Msg3_active[i] = 0;
+ // flush Msg3 buffer
+ ue->ulsch_Msg3_active[i] = 0;
+ }
}
}
@@ -327,7 +330,8 @@ void ra_succeeded(uint8_t Mod_id,uint8_t CC_id,uint8_t eNB_index)
for (i=0; i<8; i++) {
if (PHY_vars_UE_g[Mod_id][CC_id]->ulsch[eNB_index]->harq_processes[i]) {
PHY_vars_UE_g[Mod_id][CC_id]->ulsch[eNB_index]->harq_processes[i]->status=IDLE;
- PHY_vars_UE_g[Mod_id][CC_id]->dlsch[eNB_index][0]->harq_processes[i]->round=0;
+ PHY_vars_UE_g[Mod_id][CC_id]->dlsch[0][eNB_index][0]->harq_processes[i]->round=0;
+ PHY_vars_UE_g[Mod_id][CC_id]->dlsch[1][eNB_index][0]->harq_processes[i]->round=0;
}
}
@@ -638,7 +642,7 @@ void ue_compute_srs_occasion(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id
uint8_t pucch_ack_payload[2];
if (get_ack(&ue->frame_parms,
- ue->dlsch[eNB_id][0]->harq_ack,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->harq_ack,
subframe_tx,pucch_ack_payload,0) > 0)
{
is_sr_an_subframe = 1;
@@ -790,6 +794,7 @@ PUCCH_FMT_t get_pucch_format(lte_frame_type_t frame_type,
return pucch_format2;
}
}
+ return pucch_format1a;
}
uint16_t get_n1_pucch(PHY_VARS_UE *ue,
UE_rxtx_proc_t *proc,
@@ -929,11 +934,11 @@ uint16_t get_n1_pucch(PHY_VARS_UE *ue,
n1_pucch1 = get_Np(frame_parms->N_RB_DL,nCCE1,1) + nCCE1 + frame_parms->pucch_config_common.n1PUCCH_AN;
// set ACK/NAK to values if not DTX
- if (ue->dlsch[eNB_id][0]->harq_ack[(6+last_dl)%10].send_harq_status>0) // n-6 // subframe 6 is to be ACK/NAKed
- harq_ack1 = ue->dlsch[eNB_id][0]->harq_ack[(6+last_dl)%10].ack;
+ if (ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->harq_ack[(6+last_dl)%10].send_harq_status>0) // n-6 // subframe 6 is to be ACK/NAKed
+ harq_ack1 = ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->harq_ack[(6+last_dl)%10].ack;
- if (ue->dlsch[eNB_id][0]->harq_ack[5+last_dl].send_harq_status>0) // n-6 // subframe 5 is to be ACK/NAKed
- harq_ack0 = ue->dlsch[eNB_id][0]->harq_ack[5+last_dl].ack;
+ if (ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->harq_ack[5+last_dl].send_harq_status>0) // n-6 // subframe 5 is to be ACK/NAKed
+ harq_ack0 = ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->harq_ack[5+last_dl].ack;
if (harq_ack1!=2) { // n-6 // subframe 6,8,0 and maybe 5,7,9 is to be ACK/NAKed
@@ -1040,10 +1045,10 @@ void ulsch_common_procedures(PHY_VARS_UE *ue, UE_rxtx_proc_t *proc, uint8_t empt
int subframe_tx = proc->subframe_tx;
int frame_tx = proc->frame_tx;
int ulsch_start;
- int overflow=0;
#if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR)
+ int overflow=0;
int k,l;
- int dummy_tx_buffer[frame_parms->samples_per_tti] __attribute__((aligned(16)));
+ int dummy_tx_buffer[3840*4] __attribute__((aligned(16)));
#endif
VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX_ULSCH_COMMON,VCD_FUNCTION_IN);
@@ -1069,22 +1074,18 @@ void ulsch_common_procedures(PHY_VARS_UE *ue, UE_rxtx_proc_t *proc, uint8_t empt
ulsch_start = (frame_parms->samples_per_tti*subframe_tx)-ue->N_TA_offset; //-ue->timing_advance;
#endif //else EXMIMO
-//#if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR)
+#if defined(EXMIMO) || defined(OAI_USRP) || defined(OAI_BLADERF) || defined(OAI_LMSSDR)
if (empty_subframe)
{
//#if 1
overflow = ulsch_start - 9*frame_parms->samples_per_tti;
for (aa=0; aa<frame_parms->nb_antennas_tx; aa++) {
- if (overflow > 0)
- {
- memset(&ue->common_vars.txdata[aa][ulsch_start],0,4*(frame_parms->samples_per_tti-overflow));
- memset(&ue->common_vars.txdata[aa][0],0,4*overflow);
- }
- else
- {
- memset(&ue->common_vars.txdata[aa][ulsch_start],0,4*frame_parms->samples_per_tti);
- }
+ memset(&ue->common_vars.txdata[aa][ulsch_start],0,
+ 4*cmin(frame_parms->samples_per_tti-overflow,frame_parms->samples_per_tti));
+
+ if (overflow> 0)
+ memset(&ue->common_vars.txdata[aa][0],0,4*overflow);
}
/*#else
overflow = ulsch_start - 9*frame_parms->samples_per_tti;
@@ -1102,7 +1103,7 @@ void ulsch_common_procedures(PHY_VARS_UE *ue, UE_rxtx_proc_t *proc, uint8_t empt
#endif*/
return;
}
-//#endif
+#endif
if ((frame_tx%100) == 0)
LOG_D(PHY,"[UE %d] Frame %d, subframe %d: ulsch_start = %d (rxoff %d, HW TA %d, timing advance %d, TA_offset %d\n",
@@ -1336,7 +1337,8 @@ void ue_ulsch_uespec_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB
uint8_t ulsch_input_buffer[5477] __attribute__ ((aligned(32)));
uint8_t access_mode;
uint8_t Nbundled=0;
- uint8_t ack_status=0;
+ uint8_t ack_status_cw0=0;
+ uint8_t ack_status_cw1=0;
VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX_ULSCH_UESPEC,VCD_FUNCTION_IN);
@@ -1437,11 +1439,16 @@ void ue_ulsch_uespec_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB
ue->ulsch[eNB_id]->harq_processes[harq_pid]->round = 0;
}
- ack_status = reset_ack(&ue->frame_parms,
- ue->dlsch[eNB_id][0]->harq_ack,
- subframe_tx,
- ue->ulsch[eNB_id]->o_ACK,0);
- Nbundled = ack_status;
+ ack_status_cw0 = reset_ack(&ue->frame_parms,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->harq_ack,
+ subframe_tx,
+ ue->ulsch[eNB_id]->o_ACK,0);
+ ack_status_cw1 = reset_ack(&ue->frame_parms,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][1]->harq_ack,
+ subframe_tx,
+ ue->ulsch[eNB_id]->o_ACK,1);
+
+ Nbundled = ack_status_cw0;
first_rb = ue->ulsch[eNB_id]->harq_processes[harq_pid]->first_rb;
nb_rb = ue->ulsch[eNB_id]->harq_processes[harq_pid]->nb_rb;
@@ -1449,31 +1456,31 @@ void ue_ulsch_uespec_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB
- if (ack_status > 0) {
+ if (ack_status_cw0 > 0) {
// check if we received a PDSCH at subframe_tx - 4
// ==> send ACK/NACK on PUSCH
- ue->ulsch[eNB_id]->harq_processes[harq_pid]->O_ACK = 1;
+ ue->ulsch[eNB_id]->harq_processes[harq_pid]->O_ACK = ack_status_cw0 + ack_status_cw1;
#if T_TRACER
if(ue->ulsch[eNB_id]->o_ACK[0])
{
LOG_I(PHY,"PUSCH ACK\n");
- T(T_UE_PHY_DLSCH_UE_ACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(subframe_tx), T_INT(Mod_id), T_INT(ue->dlsch[eNB_id][0]->rnti),
- T_INT(ue->dlsch[eNB_id][0]->current_harq_pid));
+ T(T_UE_PHY_DLSCH_UE_ACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(subframe_tx), T_INT(Mod_id), T_INT(ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->rnti),
+ T_INT(ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->current_harq_pid));
}
else
{
LOG_I(PHY,"PUSCH NACK\n");
- T(T_UE_PHY_DLSCH_UE_NACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(subframe_tx), T_INT(Mod_id), T_INT(ue->dlsch[eNB_id][0]->rnti),
- T_INT(ue->dlsch[eNB_id][0]->current_harq_pid));
+ T(T_UE_PHY_DLSCH_UE_NACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(subframe_tx), T_INT(Mod_id), T_INT(ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->rnti),
+ T_INT(ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->current_harq_pid));
}
#endif
- LOG_D(PHY,"[UE %d][PDSCH %x] AbsSubFrame %d.%d Generating ACK (%d,%d) for %d bits on PUSCH\n",
+ LOG_I(PHY,"[UE %d][PDSCH %x] AbsSubFrame %d.%d Generating ACK (%d,%d) for %d bits on PUSCH\n",
Mod_id,
ue->ulsch[eNB_id]->rnti,
- frame_tx%1024,subframe_tx,
+ frame_tx,subframe_tx,
ue->ulsch[eNB_id]->o_ACK[0],ue->ulsch[eNB_id]->o_ACK[1],
ue->ulsch[eNB_id]->harq_processes[harq_pid]->O_ACK);
}
@@ -1530,6 +1537,7 @@ void ue_ulsch_uespec_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB
mac_xface->macphy_exit("Error in ulsch_coding");
VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX, VCD_FUNCTION_OUT);
stop_meas(&ue->phy_proc_tx);
+ //printf("------FULL TX PROC : %5.2f ------\n",ue->phy_proc_tx.p_time/(cpuf*1000.0));
return;
}
}
@@ -1636,8 +1644,8 @@ void ue_ulsch_uespec_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB
T(T_UE_PHY_PUSCH_TX_POWER, T_INT(eNB_id),T_INT(Mod_id), T_INT(frame_tx%1024), T_INT(subframe_tx),T_INT(ue->tx_power_dBm[subframe_tx]),
T_INT(tx_amp),T_INT(ue->ulsch[eNB_id]->f_pusch),T_INT(get_PL(Mod_id,0,eNB_id)),T_INT(nb_rb));
#endif
- LOG_D(PHY,"[UE %d][PUSCH %d] AbsSubFrame %d.%d, generating PUSCH, Po_PUSCH: %d dBm (max %d dBm), amp %d\n",
- Mod_id,harq_pid,frame_tx%1024,subframe_tx,ue->tx_power_dBm[subframe_tx],ue->tx_power_max_dBm, tx_amp);
+ LOG_D(PHY,"[UE %d][PUSCH %d] Frame %d subframe %d, generating PUSCH, Po_PUSCH: %d dBm (max %d dBm), amp %d\n",
+ Mod_id,harq_pid,frame_tx,subframe_tx,ue->tx_power_dBm[subframe_tx],ue->tx_power_max_dBm, tx_amp);
start_meas(&ue->ulsch_modulation_stats);
ulsch_modulation(ue->common_vars.txdataF,
tx_amp,
@@ -1770,13 +1778,13 @@ void get_pucch_param(PHY_VARS_UE *ue,
{
pucch_resource[0] = get_n1_pucch(ue,
proc,
- ue->dlsch[eNB_id][0]->harq_ack,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->harq_ack,
eNB_id,
ack_payload,
SR);
pucch_payload[0] = ack_payload[0];
- //pucch_payload[1] = ack_payload[1];
- pucch_payload[1] = 1;
+ pucch_payload[1] = ack_payload[1];
+ //pucch_payload[1] = 1;
}
break;
@@ -1785,7 +1793,7 @@ void get_pucch_param(PHY_VARS_UE *ue,
pucch_resource[0] = ue->cqi_report_config[eNB_id].CQI_ReportPeriodic.cqi_PUCCH_ResourceIndex;
if(cqi_report)
{
- pucch_payload[0] = get_pucch2_cqi(ue,eNB_id,plength);
+ pucch_payload[0] = get_pucch2_cqi(ue,eNB_id,(int*)plength);
}
else
{
@@ -1806,14 +1814,12 @@ void ue_pucch_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
uint8_t pucch_ack_payload[2];
- uint8_t n1_pucch,n2_pucch;
+ uint8_t n2_pucch;
uint16_t pucch_resource;
ANFBmode_t bundling_flag;
PUCCH_FMT_t format;
uint8_t SR_payload;
- uint16_t CQI_payload;
- uint16_t RI_payload;
uint8_t pucch_payload[2];
uint16_t len;
@@ -1824,13 +1830,11 @@ void ue_pucch_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
int CC_id = ue->CC_id;
int tx_amp;
int16_t Po_PUCCH;
- uint8_t ack_status=0;
uint8_t ack_status_cw0=0;
uint8_t ack_status_cw1=0;
uint8_t nb_cw=0;
uint8_t cqi_status=0;
uint8_t ri_status=0;
- uint8_t ack_sr_generated = 0;
VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_TX_PUCCH,VCD_FUNCTION_IN);
@@ -1891,13 +1895,13 @@ void ue_pucch_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
}
ack_status_cw0 = reset_ack(&ue->frame_parms,
- ue->dlsch[eNB_id][0]->harq_ack,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->harq_ack,
subframe_tx,
pucch_ack_payload,
0);
ack_status_cw1 = reset_ack(&ue->frame_parms,
- ue->dlsch[eNB_id][1]->harq_ack,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][1]->harq_ack,
subframe_tx,
pucch_ack_payload,
1);
@@ -1936,11 +1940,10 @@ void ue_pucch_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
SR_payload,
cqi_status,
&pucch_resource,
- &pucch_payload,
+ (uint8_t *)&pucch_payload,
&len);
- LOG_D(PHY,"PUCCH feedback AbsSubframe %d.%d SR %d NbCW %d AckNack %d.%d CQI %d RI %d format %d pucch_resource %d pucch_payload %d %d \n",
- frame_tx%1024, subframe_tx, SR_payload, nb_cw, pucch_ack_payload[0], pucch_ack_payload[1], cqi_status, ri_status, format, pucch_resource,pucch_payload[0],pucch_payload[1]);
+ LOG_D(PHY,"PUCCH feedback AbsSubframe %d.%d SR %d NbCW %d AckNack %d.%d CQI %d RI %d format %d pucch_resource %d pucch_payload %d %d \n", frame_tx, subframe_tx, SR_payload, nb_cw, pucch_ack_payload[0], pucch_ack_payload[1], cqi_status, ri_status, format, pucch_resource,pucch_payload[0],pucch_payload[1]);
// Part - IV
@@ -1971,13 +1974,13 @@ void ue_pucch_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
#endif
#if T_TRACER
T(T_UE_PHY_PUCCH_TX_POWER, T_INT(eNB_id),T_INT(Mod_id), T_INT(frame_tx%1024), T_INT(subframe_tx),T_INT(ue->tx_power_dBm[subframe_tx]),
- T_INT(tx_amp),T_INT(ue->dlsch[eNB_id][0]->g_pucch),T_INT(get_PL(ue->Mod_id,ue->CC_id,eNB_id)));
+ T_INT(tx_amp),T_INT(ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->g_pucch),T_INT(get_PL(ue->Mod_id,ue->CC_id,eNB_id)));
#endif
if(format == pucch_format1)
{
LOG_D(PHY,"[UE %d][SR %x] AbsSubframe %d.%d Generating PUCCH 1 (SR for PUSCH), an_srs_simultanous %d, shorten_pucch %d, n1_pucch %d, Po_PUCCH %d\n",
Mod_id,
- ue->dlsch[eNB_id][0]->rnti,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->rnti,
frame_tx, subframe_tx,
frame_parms->soundingrs_ul_config_common.ackNackSRS_SimultaneousTransmission,
isShortenPucch,
@@ -1989,7 +1992,7 @@ void ue_pucch_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
if (SR_payload>0) {
LOG_D(PHY,"[UE %d][SR %x] AbsSubFrame %d.%d Generating PUCCH %s payload %d,%d (with SR for PUSCH), an_srs_simultanous %d, shorten_pucch %d, n1_pucch %d, Po_PUCCH %d, amp %d\n",
Mod_id,
- ue->dlsch[eNB_id][0]->rnti,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->rnti,
frame_tx % 1024, subframe_tx,
(format == pucch_format1a? "1a": (
format == pucch_format1b? "1b" : "??")),
@@ -2002,7 +2005,7 @@ void ue_pucch_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
} else {
LOG_D(PHY,"[UE %d][PDSCH %x] AbsSubFrame %d.%d rx_offset_diff: %d, Generating PUCCH %s, an_srs_simultanous %d, shorten_pucch %d, n1_pucch %d, b[0]=%d,b[1]=%d (SR_Payload %d), Po_PUCCH %d, amp %d\n",
Mod_id,
- ue->dlsch[eNB_id][0]->rnti,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->rnti,
frame_tx%1024, subframe_tx,ue->rx_offset_diff,
(format == pucch_format1a? "1a": (
format == pucch_format1b? "1b" : "??")),
@@ -2017,13 +2020,13 @@ void ue_pucch_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
#if T_TRACER
if(pucch_payload[0])
{
- T(T_UE_PHY_DLSCH_UE_ACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(subframe_tx), T_INT(Mod_id), T_INT(ue->dlsch[eNB_id][0]->rnti),
- T_INT(ue->dlsch[eNB_id][0]->current_harq_pid));
+ T(T_UE_PHY_DLSCH_UE_ACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(subframe_tx), T_INT(Mod_id), T_INT(ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->rnti),
+ T_INT(ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->current_harq_pid));
}
else
{
- T(T_UE_PHY_DLSCH_UE_NACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(subframe_tx), T_INT(Mod_id), T_INT(ue->dlsch[eNB_id][0]->rnti),
- T_INT(ue->dlsch[eNB_id][0]->current_harq_pid));
+ T(T_UE_PHY_DLSCH_UE_NACK, T_INT(eNB_id), T_INT(frame_tx%1024), T_INT(subframe_tx), T_INT(Mod_id), T_INT(ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->rnti),
+ T_INT(ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->current_harq_pid));
}
#endif
@@ -2076,12 +2079,12 @@ void ue_pucch_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
#endif
#if T_TRACER
T(T_UE_PHY_PUCCH_TX_POWER, T_INT(eNB_id),T_INT(Mod_id), T_INT(frame_tx%1024), T_INT(subframe_tx),T_INT(ue->tx_power_dBm[subframe_tx]),
- T_INT(tx_amp),T_INT(ue->dlsch[eNB_id][0]->g_pucch),T_INT(get_PL(ue->Mod_id,ue->CC_id,eNB_id)));
+ T_INT(tx_amp),T_INT(ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->g_pucch),T_INT(get_PL(ue->Mod_id,ue->CC_id,eNB_id)));
#endif
LOG_D(PHY,"[UE %d][RNTI %x] AbsSubFrame %d.%d Generating PUCCH 2 (RI or CQI), n2_pucch %d, Po_PUCCH %d, isShortenPucch %d, amp %d\n",
Mod_id,
- ue->dlsch[eNB_id][0]->rnti,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->rnti,
frame_tx%1024, subframe_tx,
n2_pucch,
Po_PUCCH,
@@ -2106,13 +2109,13 @@ void ue_pucch_procedures(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
case pucch_format2a:
LOG_I(PHY,"[UE %d][RNTI %x] AbsSubFrame %d.%d Generating PUCCH 2a (RI or CQI) Ack/Nack 1bit \n",
Mod_id,
- ue->dlsch[eNB_id][0]->rnti,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->rnti,
frame_tx%1024, subframe_tx);
break;
case pucch_format2b:
LOG_I(PHY,"[UE %d][RNTI %x] AbsSubFrame %d.%d Generating PUCCH 2b (RI or CQI) Ack/Nack 2bits\n",
Mod_id,
- ue->dlsch[eNB_id][0]->rnti,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->rnti,
frame_tx%1024, subframe_tx);
break;
default:
@@ -2254,7 +2257,7 @@ void phy_procedures_UE_TX(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,ui
// reset DL ACK/NACK status
reset_ack(&ue->frame_parms,
- ue->dlsch[eNB_id][0]->harq_ack,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->harq_ack,
subframe_tx,
ue->ulsch[eNB_id]->o_ACK,0);
@@ -2451,8 +2454,8 @@ void phy_procedures_emos_UE_RX(PHY_VARS_UE *ue,uint8_t last_slot,uint8_t eNB_id)
emos_dump_UE.total_TBS_last = ue->total_TBS_last[eNB_id];
emos_dump_UE.bitrate = ue->bitrate[eNB_id];
emos_dump_UE.total_received_bits = ue->total_received_bits[eNB_id];
- emos_dump_UE.pmi_saved = ue->dlsch[eNB_id][0]->pmi_alloc;
- emos_dump_UE.mcs = ue->dlsch[eNB_id][0]->harq_processes[ue->dlsch[eNB_id][0]->current_harq_pid]->mcs;
+ emos_dump_UE.pmi_saved = ue->dlsch[subframe&0x1][eNB_id][0]->pmi_alloc;
+ emos_dump_UE.mcs = ue->dlsch[subframe&0x1][eNB_id][0]->harq_processes[ue->dlsch[subframe&0x1][eNB_id][0]->current_harq_pid]->mcs;
emos_dump_UE.use_ia_receiver = openair_daq_vars.use_ia_receiver;
bytes = rtf_put(CHANSOUNDER_FIFO_MINOR, &emos_dump_UE, sizeof(fifo_dump_emos_UE));
@@ -2859,7 +2862,7 @@ int ue_pdcch_procedures(uint8_t eNB_id,PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint
(void *)&dci_alloc_rx[i].dci_pdu,
ue->pdcch_vars[eNB_id]->crnti,
dci_alloc_rx[i].format,
- ue->dlsch[eNB_id],
+ ue->dlsch[subframe_rx&0x1][eNB_id],
&ue->frame_parms,
ue->pdsch_config_dedicated,
SI_RNTI,
@@ -2876,7 +2879,7 @@ int ue_pdcch_procedures(uint8_t eNB_id,PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint
(dci_alloc_rx[i].format == format2A) ||
(dci_alloc_rx[i].format == format2B))
{
- ue->dlsch[eNB_id][0]->g_pucch += ue->dlsch[eNB_id][0]->harq_processes[ue->dlsch[eNB_id][0]->current_harq_pid]->delta_PUCCH;
+ ue->dlsch[subframe_rx&0x1][eNB_id][0]->g_pucch += ue->dlsch[subframe_rx&0x1][eNB_id][0]->harq_processes[ue->dlsch[subframe_rx&0x1][eNB_id][0]->current_harq_pid]->delta_PUCCH;
}
ue->dlsch_received[eNB_id]++;
@@ -2884,7 +2887,7 @@ int ue_pdcch_procedures(uint8_t eNB_id,PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint
#ifdef DEBUG_PHY_PROC
LOG_D(PHY,"[UE %d] Generated UE DLSCH C_RNTI format %d\n",ue->Mod_id,dci_alloc_rx[i].format);
dump_dci(&ue->frame_parms, &dci_alloc_rx[i]);
- LOG_D(PHY,"[UE %d] *********** dlsch->active in subframe %d=> %d\n",ue->Mod_id,subframe_rx,ue->dlsch[eNB_id][0]->active);
+ LOG_D(PHY,"[UE %d] *********** dlsch->active in subframe %d=> %d\n",ue->Mod_id,subframe_rx,ue->dlsch[subframe_rx&0x1][eNB_id][0]->active);
#endif
// we received a CRNTI, so we're in PUSCH
@@ -3262,7 +3265,14 @@ void ue_pdsch_procedures(PHY_VARS_UE *ue, UE_rxtx_proc_t *proc, int eNB_id, PDSC
dual_stream_UE = 1;
eNB_id_i = ue->n_connected_eNB;
i_mod = dlsch0->harq_processes[harq_pid]->Qm;
- } else {
+ }
+ else if((pdsch==PDSCH) && (ue->transmission_mode[eNB_id]==3))
+ {
+ dual_stream_UE = rx_IC_dual_stream;
+ eNB_id_i = eNB_id;
+ i_mod = 0;
+ }
+ else {
dual_stream_UE = 0;
eNB_id_i = eNB_id+1;
i_mod = 0;
@@ -3401,11 +3411,34 @@ void ue_dlsch_procedures(PHY_VARS_UE *ue,
int harq_pid;
int frame_rx = proc->frame_rx;
int subframe_rx = proc->subframe_rx;
- int ret=0;
+ int ret=0, ret1=0;
int CC_id = ue->CC_id;
LTE_UE_PDSCH *pdsch_vars;
+ uint8_t is_cw0_active = 0;
+ uint8_t is_cw1_active = 0;
+
+ if (dlsch0==NULL)
+ AssertFatal(0,"dlsch0 should be defined at this level \n");
+
+ harq_pid = dlsch0->current_harq_pid;
+ is_cw0_active = dlsch0->harq_processes[harq_pid]->status;
+
+ if(dlsch1)
+ is_cw1_active = dlsch1->harq_processes[harq_pid]->status;
+
+ LOG_D(PHY,"AbsSubframe %d.%d Start Turbo Decoder for CW0 [harq_pid %d] ? %d \n", frame_rx%1024, subframe_rx, harq_pid, is_cw0_active);
+ LOG_D(PHY,"AbsSubframe %d.%d Start Turbo Decoder for CW1 [harq_pid %d] ? %d \n", frame_rx%1024, subframe_rx, harq_pid, is_cw1_active);
- if (dlsch0 && (!dlsch1)) {
+ if(is_cw0_active && is_cw1_active)
+ {
+ dlsch0->Kmimo = 2;
+ dlsch1->Kmimo = 2;
+ }
+ else
+ {
+ dlsch0->Kmimo = 1;
+ }
+ if (1) {
switch (pdsch) {
case SI_PDSCH:
pdsch_vars = ue->pdsch_vars_SI[eNB_id];
@@ -3431,8 +3464,6 @@ void ue_dlsch_procedures(PHY_VARS_UE *ue,
break;
}
-
- harq_pid = dlsch0->current_harq_pid;
if (frame_rx < *dlsch_errors)
*dlsch_errors=0;
@@ -3449,6 +3480,7 @@ void ue_dlsch_procedures(PHY_VARS_UE *ue,
if (abstraction_flag == 0) {
+ // start turbo decode for CW 0
dlsch0->harq_processes[harq_pid]->G = get_G(&ue->frame_parms,
dlsch0->harq_processes[harq_pid]->nb_rb,
dlsch0->harq_processes[harq_pid]->rb_alloc_even,
@@ -3468,6 +3500,13 @@ void ue_dlsch_procedures(PHY_VARS_UE *ue,
subframe_rx<<1);
stop_meas(&ue->dlsch_unscrambling_stats);
+ //LOG_I(PHY,"start turbo decode for CW 0 --> nb_rb %d \n", dlsch0->harq_processes[harq_pid]->nb_rb);
+ //LOG_I(PHY,"start turbo decode for CW 0 --> rb_alloc_even %x \n", dlsch0->harq_processes[harq_pid]->rb_alloc_even);
+ //LOG_I(PHY,"start turbo decode for CW 0 --> Qm %d \n", dlsch0->harq_processes[harq_pid]->Qm);
+ //LOG_I(PHY,"start turbo decode for CW 0 --> Nl %d \n", dlsch0->harq_processes[harq_pid]->Nl);
+ //LOG_I(PHY,"start turbo decode for CW 0 --> G %d \n", dlsch0->harq_processes[harq_pid]->G);
+ //LOG_I(PHY,"start turbo decode for CW 0 --> Kmimo %d \n", dlsch0->Kmimo);
+
start_meas(&ue->dlsch_decoding_stats);
ret = dlsch_decoding(ue,
pdsch_vars->llr[0],
@@ -3480,6 +3519,60 @@ void ue_dlsch_procedures(PHY_VARS_UE *ue,
pdsch==PDSCH?1:0,
dlsch0->harq_processes[harq_pid]->TBS>256?1:0);
stop_meas(&ue->dlsch_decoding_stats);
+
+ //printf(" --> Unscrambling for CW0 %5.3f\n",
+ // (ue->dlsch_unscrambling_stats.p_time)/(cpuf*1000.0));
+ //printf(" --> Turbo Decoding for CW0 %5.3f\n",
+ // (ue->dlsch_decoding_stats.p_time)/(cpuf*1000.0));
+
+ if(is_cw1_active)
+ {
+ // start turbo decode for CW 1
+ dlsch1->harq_processes[harq_pid]->G = get_G(&ue->frame_parms,
+ dlsch1->harq_processes[harq_pid]->nb_rb,
+ dlsch1->harq_processes[harq_pid]->rb_alloc_even,
+ dlsch1->harq_processes[harq_pid]->Qm,
+ dlsch1->harq_processes[harq_pid]->Nl,
+ ue->pdcch_vars[eNB_id]->num_pdcch_symbols,
+ frame_rx,
+ subframe_rx,
+ ue->transmission_mode[eNB_id]<7?0:ue->transmission_mode[eNB_id]);
+
+ start_meas(&ue->dlsch_unscrambling_stats);
+ dlsch_unscrambling(&ue->frame_parms,
+ 0,
+ dlsch1,
+ dlsch1->harq_processes[harq_pid]->G,
+ pdsch_vars->llr[1],
+ 1,
+ subframe_rx<<1);
+ stop_meas(&ue->dlsch_unscrambling_stats);
+
+ //LOG_I(PHY,"start turbo decode for CW 1 --> nb_rb %d \n", dlsch1->harq_processes[harq_pid]->nb_rb);
+ //LOG_I(PHY,"start turbo decode for CW 1 --> rb_alloc_even %x \n", dlsch1->harq_processes[harq_pid]->rb_alloc_even);
+ //LOG_I(PHY,"start turbo decode for CW 1 --> Qm %d \n", dlsch1->harq_processes[harq_pid]->Qm);
+ //LOG_I(PHY,"start turbo decode for CW 1 --> Nl %d \n", dlsch1->harq_processes[harq_pid]->Nl);
+ //LOG_I(PHY,"start turbo decode for CW 1 --> G %d \n", dlsch1->harq_processes[harq_pid]->G);
+ //LOG_I(PHY,"start turbo decode for CW 1 --> Kmimo %d \n", dlsch1->Kmimo);
+
+ start_meas(&ue->dlsch_decoding_stats);
+ ret1 = dlsch_decoding(ue,
+ pdsch_vars->llr[1],
+ &ue->frame_parms,
+ dlsch1,
+ dlsch1->harq_processes[harq_pid],
+ frame_rx,
+ subframe_rx,
+ harq_pid,
+ pdsch==PDSCH?1:0,
+ dlsch1->harq_processes[harq_pid]->TBS>256?1:0);
+ stop_meas(&ue->dlsch_decoding_stats);
+
+ //printf(" --> Unscrambling for CW1 %5.3f\n",
+ // (ue->dlsch_unscrambling_stats.p_time)/(cpuf*1000.0));
+ //printf(" --> Turbo Decoding for CW1 %5.3f\n",
+ // (ue->dlsch_decoding_stats.p_time)/(cpuf*1000.0));
+ }
}
else {
@@ -3492,12 +3585,13 @@ void ue_dlsch_procedures(PHY_VARS_UE *ue,
#endif
}
+ // Check CRC for CW 0
if (ret == (1+dlsch0->max_turbo_iterations)) {
*dlsch_errors=*dlsch_errors+1;
if(dlsch0->rnti != 0xffff)
{
- LOG_D(PHY,"[UE %d][PDSCH %x/%d] Frame %d subframe %d DLSCH in error (rv %d,mcs %d,TBS %d)\n",
+ LOG_D(PHY,"[UE %d][PDSCH %x/%d] AbsSubframe %d.%d : DLSCH CW0 in error (rv %d,mcs %d,TBS %d)\n",
ue->Mod_id,dlsch0->rnti,
harq_pid,frame_rx,subframe_rx,
dlsch0->harq_processes[harq_pid]->rvidx,
@@ -3509,7 +3603,7 @@ void ue_dlsch_procedures(PHY_VARS_UE *ue,
} else {
if(dlsch0->rnti != 0xffff)
{
- LOG_D(PHY,"[UE %d][PDSCH %x/%d] Frame %d subframe %d: Received DLSCH (rv %d,mcs %d,TBS %d)\n",
+ LOG_D(PHY,"[UE %d][PDSCH %x/%d] AbsSubframe %d.%d : Received DLSCH CW0 (rv %d,mcs %d,TBS %d)\n",
ue->Mod_id,dlsch0->rnti,
harq_pid,frame_rx,subframe_rx,
dlsch0->harq_processes[harq_pid]->rvidx,
@@ -3574,6 +3668,49 @@ void ue_dlsch_procedures(PHY_VARS_UE *ue,
dlsch0->harq_processes[dlsch0->current_harq_pid]->TBS;
}
+ // Check CRC for CW 1
+ if(is_cw1_active)
+ {
+ if (ret1 == (1+dlsch0->max_turbo_iterations)) {
+ LOG_D(PHY,"[UE %d][PDSCH %x/%d] Frame %d subframe %d DLSCH CW1 in error (rv %d,mcs %d,TBS %d)\n",
+ ue->Mod_id,dlsch0->rnti,
+ harq_pid,frame_rx,subframe_rx,
+ dlsch0->harq_processes[harq_pid]->rvidx,
+ dlsch0->harq_processes[harq_pid]->mcs,
+ dlsch0->harq_processes[harq_pid]->TBS);
+
+ } else {
+ LOG_D(PHY,"[UE %d][PDSCH %x/%d] Frame %d subframe %d: Received DLSCH CW1 (rv %d,mcs %d,TBS %d)\n",
+ ue->Mod_id,dlsch0->rnti,
+ harq_pid,frame_rx,subframe_rx,
+ dlsch0->harq_processes[harq_pid]->rvidx,
+ dlsch0->harq_processes[harq_pid]->mcs,
+ dlsch0->harq_processes[harq_pid]->TBS);
+
+
+ if (ue->mac_enabled == 1) {
+ switch (pdsch) {
+ case PDSCH:
+ if(is_cw1_active)
+ mac_xface->ue_send_sdu(ue->Mod_id,
+ CC_id,
+ frame_rx,
+ subframe_rx,
+ dlsch1->harq_processes[dlsch1->current_harq_pid]->b,
+ dlsch1->harq_processes[dlsch1->current_harq_pid]->TBS>>3,
+ eNB_id);
+ break;
+ case SI_PDSCH:
+ case P_PDSCH:
+ case RA_PDSCH:
+ case PDSCH1:
+ case PMCH:
+ AssertFatal(0,"exiting");
+ break;
+ }
+ }
+ }
+ }
#ifdef DEBUG_PHY_PROC
@@ -3622,16 +3759,18 @@ int phy_procedures_UE_RX(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
T_BUFFER(&ue->common_vars.rxdata[0][subframe_rx*ue->frame_parms.samples_per_tti],
ue->frame_parms.samples_per_tti * 4));
- start_meas(&ue->phy_proc_rx);
+ // start timers
+ start_meas(&ue->phy_proc_rx[subframe_rx&0x1]);
+ start_meas(&ue->generic_stat);
pmch_flag = is_pmch_subframe(frame_rx,subframe_rx,&ue->frame_parms) ? 1 : 0;
// deactivate reception until we scan pdcch
- if (ue->dlsch[eNB_id][0])
- ue->dlsch[eNB_id][0]->active = 0;
- if (ue->dlsch[eNB_id][1])
- ue->dlsch[eNB_id][1]->active = 0;
+ if (ue->dlsch[subframe_rx&0x1][eNB_id][0])
+ ue->dlsch[subframe_rx&0x1][eNB_id][0]->active = 0;
+ if (ue->dlsch[subframe_rx&0x1][eNB_id][1])
+ ue->dlsch[subframe_rx&0x1][eNB_id][1]->active = 0;
if (ue->dlsch_SI[eNB_id])
ue->dlsch_SI[eNB_id]->active = 0;
@@ -3717,15 +3856,18 @@ int phy_procedures_UE_RX(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
0);
// first slot has been processed (FFTs + Channel Estimation, PCFICH/PHICH/PDCCH)
-
+ stop_meas(&ue->generic_stat);
+ //printf("[SFN %d] Slot0: FFT + Channel Estimate + PCFICH/PHICH/PDCCH %5.2f \n",subframe_rx,ue->generic_stat.p_time/(cpuf*1000.0));
+
+ start_meas(&ue->generic_stat);
// do procedures for C-RNTI
- if (ue->dlsch[eNB_id][0]->active == 1) {
+ if (ue->dlsch[subframe_rx&0x1][eNB_id][0]->active == 1) {
VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_IN);
ue_pdsch_procedures(ue,
proc,
eNB_id,
PDSCH,
- ue->dlsch[eNB_id][0],
+ ue->dlsch[subframe_rx&0x1][eNB_id][0],
NULL,
ue->pdcch_vars[eNB_id]->num_pdcch_symbols,
ue->frame_parms.symbols_per_tti>>1,
@@ -3809,35 +3951,50 @@ int phy_procedures_UE_RX(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
}
} // not an S-subframe
+ stop_meas(&ue->generic_stat);
+ //printf("[SFN %d] Slot1: FFT + Channel Estimate + Pdsch Proc Slot0 %5.2f \n",subframe_rx,ue->generic_stat.p_time/(cpuf*1000.0));
+
// run pbch procedures if subframe is 0
- if (subframe_rx == 0)
+ if ( (subframe_rx == 0) && (ue->decode_MIB == 1))
+ {
ue_pbch_procedures(eNB_id,ue,proc,abstraction_flag);
+ }
// do procedures for C-RNTI
- if (ue->dlsch[eNB_id][0]->active == 1) {
+ if (ue->dlsch[subframe_rx&0x1][eNB_id][0]->active == 1) {
VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_IN);
+ start_meas(&ue->pdsch_procedures_stat);
ue_pdsch_procedures(ue,
proc,
eNB_id,
PDSCH,
- ue->dlsch[eNB_id][0],
+ ue->dlsch[subframe_rx&0x1][eNB_id][0],
NULL,
1+(ue->frame_parms.symbols_per_tti>>1),
ue->frame_parms.symbols_per_tti-1,
abstraction_flag);
+ stop_meas(&ue->pdsch_procedures_stat);
+
+ start_meas(&ue->dlsch_procedures_stat);
ue_dlsch_procedures(ue,
proc,
eNB_id,
PDSCH,
- ue->dlsch[eNB_id][0],
- NULL,
+ ue->dlsch[subframe_rx&0x1][eNB_id][0],
+ ue->dlsch[subframe_rx&0x1][eNB_id][1],
&ue->dlsch_errors[eNB_id],
mode,
abstraction_flag);
+ stop_meas(&ue->dlsch_procedures_stat);
+ //printf("[SFN %d] Slot1: Pdsch Proc %5.2f\n",subframe_rx,ue->pdsch_procedures_stat.p_time/(cpuf*1000.0));
+ //printf("[SFN %d] Slot0 Slot1: Dlsch Proc %5.2f\n",subframe_rx,ue->dlsch_procedures_stat.p_time/(cpuf*1000.0));
+
VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PDSCH_PROC, VCD_FUNCTION_OUT);
}
+ start_meas(&ue->generic_stat);
+
// do procedures for SI-RNTI
if ((ue->dlsch_SI[eNB_id]) && (ue->dlsch_SI[eNB_id]->active == 1)) {
ue_pdsch_procedures(ue,
@@ -3932,7 +4089,8 @@ int phy_procedures_UE_RX(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
}
-
+ stop_meas(&ue->generic_stat);
+ //printf("after tubo until end of Rx %5.2f \n",ue->generic_stat.p_time/(cpuf*1000.0));
#ifdef EMOS
phy_procedures_emos_UE_RX(ue,slot,eNB_id);
@@ -3940,7 +4098,10 @@ int phy_procedures_UE_RX(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t eNB_id,uin
VCD_SIGNAL_DUMPER_DUMP_FUNCTION_BY_NAME(VCD_SIGNAL_DUMPER_FUNCTIONS_PHY_PROCEDURES_UE_RX, VCD_FUNCTION_OUT);
- stop_meas(&ue->phy_proc_rx);
+ stop_meas(&ue->phy_proc_rx[subframe_rx&0x1]);
+
+ //printf("------FULL RX PROC [SFN %d]: %5.2f ------\n",subframe_rx,ue->phy_proc_rx[subframe_rx&0x1].p_time/(cpuf*1000.0));
+
return (0);
}
diff --git a/openair1/SCHED/pucch_pc.c b/openair1/SCHED/pucch_pc.c
index 5bc03bb58de4f8e0653410e68e091d05cb6780dc..ae831238622a91fef15d58e9be8d7d0f65ae28a5 100644
--- a/openair1/SCHED/pucch_pc.c
+++ b/openair1/SCHED/pucch_pc.c
@@ -51,7 +51,7 @@ int16_t pucch_power_cntl(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t subframe,u
Po_PUCCH = get_PL(ue->Mod_id,ue->CC_id,eNB_id)+
ue->frame_parms.ul_power_control_config_common.p0_NominalPUCCH+
- ue->dlsch[eNB_id][0]->g_pucch;
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->g_pucch;
switch (pucch_fmt) {
case pucch_format1:
@@ -90,19 +90,19 @@ int16_t pucch_power_cntl(PHY_VARS_UE *ue,UE_rxtx_proc_t *proc,uint8_t subframe,u
if (pucch_fmt!=pucch_format1) {
LOG_D(PHY,"[UE %d][PDSCH %x] AbsSubframe %d.%d: Po_PUCCH %d dBm : Po_NOMINAL_PUCCH %d dBm, PL %d dB, g_pucch %d dB\n",
ue->Mod_id,
- ue->dlsch[eNB_id][0]->rnti,proc->frame_tx%1024,subframe,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->rnti,proc->frame_tx%1024,subframe,
Po_PUCCH,
ue->frame_parms.ul_power_control_config_common.p0_NominalPUCCH,
get_PL(ue->Mod_id,ue->CC_id,eNB_id),
- ue->dlsch[eNB_id][0]->g_pucch);
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->g_pucch);
} else {
LOG_D(PHY,"[UE %d][SR %x] AbsSubframe %d.%d: Po_PUCCH %d dBm : Po_NOMINAL_PUCCH %d dBm, PL %d dB g_pucch %d dB\n",
ue->Mod_id,
- ue->dlsch[eNB_id][0]->rnti,proc->frame_tx%1024,subframe,
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->rnti,proc->frame_tx%1024,subframe,
Po_PUCCH,
ue->frame_parms.ul_power_control_config_common.p0_NominalPUCCH,
get_PL(ue->Mod_id,ue->CC_id,eNB_id),
- ue->dlsch[eNB_id][0]->g_pucch);
+ ue->dlsch[proc->subframe_rx&0x1][eNB_id][0]->g_pucch);
}
return(Po_PUCCH);
diff --git a/openair3/NAS/TOOLS/ue_tcl_test.conf b/openair3/NAS/TOOLS/ue_tcl_test.conf
new file mode 100644
index 0000000000000000000000000000000000000000..05db6be4d12130603fec8d272bf102158d8f23a4
--- /dev/null
+++ b/openair3/NAS/TOOLS/ue_tcl_test.conf
@@ -0,0 +1,114 @@
+# List of known PLMNS
+PLMN: {
+ PLMN0: {
+ FULLNAME="Test network";
+ SHORTNAME="OAI4G";
+ MNC="01";
+ MCC="001";
+
+ };
+ PLMN1: {
+ FULLNAME="SFR France";
+ SHORTNAME="SFR";
+ MNC="10";
+ MCC="208";
+
+ };
+ PLMN2: {
+ FULLNAME="SFR France";
+ SHORTNAME="SFR";
+ MNC="11";
+ MCC="208";
+ };
+ PLMN3: {
+ FULLNAME="SFR France";
+ SHORTNAME="SFR";
+ MNC="13";
+ MCC="208";
+ };
+ PLMN4: {
+ FULLNAME="OAI LTEBOX";
+ SHORTNAME="OAIALU";
+ MNC="93";
+ MCC="208";
+ };
+ PLMN5: {
+ FULLNAME="T-Mobile USA";
+ SHORTNAME="T-Mobile";
+ MNC="280";
+ MCC="310";
+ };
+ PLMN6: {
+ FULLNAME="FICTITIOUS USA";
+ SHORTNAME="FICTITIO";
+ MNC="028";
+ MCC="310";
+ };
+ PLMN7: {
+ FULLNAME="Vodafone Italia";
+ SHORTNAME="VODAFONE";
+ MNC="10";
+ MCC="222";
+ };
+ PLMN8: {
+ FULLNAME="Vodafone Spain";
+ SHORTNAME="VODAFONE";
+ MNC="01";
+ MCC="214";
+ };
+ PLMN9: {
+ FULLNAME="Vodafone Spain";
+ SHORTNAME="VODAFONE";
+ MNC="06";
+ MCC="214";
+ };
+ PLMN10: {
+ FULLNAME="Vodafone Germ";
+ SHORTNAME="VODAFONE";
+ MNC="02";
+ MCC="262";
+ };
+ PLMN11: {
+ FULLNAME="Vodafone Germ";
+ SHORTNAME="VODAFONE";
+ MNC="04";
+ MCC="262";
+ };
+};
+
+UE0:
+{
+ USER: {
+ IMEI="356113022094149";
+ MANUFACTURER="EURECOM";
+ MODEL="LTE Android PC";
+ PIN="0000";
+ };
+
+ SIM: {
+ MSIN="000001234";
+ USIM_API_K="000102030405060708090A0B0C0D0E0F";
+ OPC="C42449363BBAD02B66D16BC975D77CC1";
+ MSISDN="000000000000";//"33611123456";
+ };
+
+ # Home PLMN Selector with Access Technology
+ HPLMN= "00101";
+
+ # User controlled PLMN Selector with Access Technology
+ UCPLMN_LIST = ();
+
+ # Operator PLMN List
+ OPLMN_LIST = ("00101", "20810", "20811", "20813", "20893", "310280", "310028");
+
+ # Operator controlled PLMN Selector with Access Technology
+ OCPLMN_LIST = ("22210", "21401", "21406", "26202", "26204");
+
+ # Forbidden plmns
+ FPLMN_LIST = ();
+
+ # List of Equivalent HPLMNs
+#TODO: UE does not connect if set, to be fixed in the UE
+# EHPLMN_LIST= ("20811", "20813");
+ EHPLMN_LIST= ();
+};
diff --git a/targets/RT/USER/lte-softmodem.c b/targets/RT/USER/lte-softmodem.c
index 8303d88218fbe0fbf83ab827133e3b29d355f5de..5761577259eb8365881dacd72c994a6707b788e4 100644
--- a/targets/RT/USER/lte-softmodem.c
+++ b/targets/RT/USER/lte-softmodem.c
@@ -151,6 +151,8 @@ uint8_t usim_test = 0;
uint8_t nb_antenna_tx = 1;
uint8_t nb_antenna_rx = 1;
+int16_t dlsch_demod_shift = 0;
+
char ref[128] = "internal";
char channels[128] = "0";
@@ -635,6 +637,7 @@ static void get_options (int argc, char **argv) {
LONG_OPTION_THREADIQ,
LONG_OPTION_THREADODDSUBFRAME,
LONG_OPTION_THREADEVENSUBFRAME,
+ LONG_OPTION_DEMOD_SHIFT,
#if T_TRACER
LONG_OPTION_T_PORT,
LONG_OPTION_T_NOWAIT,
@@ -670,6 +673,7 @@ static void get_options (int argc, char **argv) {
{"threadIQ", required_argument, NULL, LONG_OPTION_THREADIQ},
{"threadOddSubframe", required_argument, NULL, LONG_OPTION_THREADODDSUBFRAME},
{"threadEvenSubframe", required_argument, NULL, LONG_OPTION_THREADEVENSUBFRAME},
+ {"dlsch-demod-shift", required_argument, NULL, LONG_OPTION_DEMOD_SHIFT},
#if T_TRACER
{"T_port", required_argument, 0, LONG_OPTION_T_PORT},
{"T_nowait", no_argument, 0, LONG_OPTION_T_NOWAIT},
@@ -800,7 +804,9 @@ static void get_options (int argc, char **argv) {
case LONG_OPTION_THREADEVENSUBFRAME:
threads.even=atoi(optarg);
break;
-
+ case LONG_OPTION_DEMOD_SHIFT:
+ dlsch_demod_shift = atof(optarg);
+ break;
#if T_TRACER
case LONG_OPTION_T_PORT: {
extern int T_port;
diff --git a/targets/RT/USER/lte-ue.c b/targets/RT/USER/lte-ue.c
index 3b6ea9e5c9166b098675de288347bc34f6405e15..49093ed2def4f22334c7870ff358abc938ce70f3 100644
--- a/targets/RT/USER/lte-ue.c
+++ b/targets/RT/USER/lte-ue.c
@@ -56,6 +56,8 @@
#include "T.h"
+extern double cpuf;
+
#define FRAME_PERIOD 100000000ULL
#define DAQ_PERIOD 66667ULL
#define FIFO_PRIORITY 40
@@ -538,6 +540,9 @@ static void *UE_thread_rxn_txnp4(void *arg) {
}
phy_procedures_UE_RX( UE, proc, 0, 0, UE->mode, no_relay, NULL );
}
+
+ start_meas(&UE->generic_stat);
+
if (UE->mac_enabled==1) {
ret = mac_xface->ue_scheduler(UE->Mod_id,
@@ -567,6 +572,9 @@ static void *UE_thread_rxn_txnp4(void *arg) {
UE->Mod_id, proc->frame_rx, proc->subframe_tx,txt );
}
}
+
+ stop_meas(&UE->generic_stat);
+
// Prepare the future Tx data
if ((subframe_select( &UE->frame_parms, proc->subframe_tx) == SF_UL) ||
diff --git a/targets/SIMU/USER/init_lte.c b/targets/SIMU/USER/init_lte.c
index 8ff274be14d14c9e618ba6d9f8b7b22709109a3a..a99ccf2c09f40866ec31acd82fc74c65886e758e 100644
--- a/targets/SIMU/USER/init_lte.c
+++ b/targets/SIMU/USER/init_lte.c
@@ -161,30 +161,32 @@ PHY_VARS_UE* init_lte_UE(LTE_DL_FRAME_PARMS *frame_parms,
memcpy(&(PHY_vars_UE->frame_parms), frame_parms, sizeof(LTE_DL_FRAME_PARMS));
phy_init_lte_ue(PHY_vars_UE,1,abstraction_flag);
- for (i=0; i<NUMBER_OF_CONNECTED_eNB_MAX; i++) {
- for (j=0; j<2; j++) {
- PHY_vars_UE->dlsch[i][j] = new_ue_dlsch(1,NUMBER_OF_HARQ_PID_MAX,NSOFT,MAX_TURBO_ITERATIONS,frame_parms->N_RB_DL, abstraction_flag);
+ for (int l=0; l<2; l++) {
+ for (i=0; i<NUMBER_OF_CONNECTED_eNB_MAX; i++) {
+ for (j=0; j<2; j++) {
+ PHY_vars_UE->dlsch[l][i][j] = new_ue_dlsch(1,NUMBER_OF_HARQ_PID_MAX,NSOFT,MAX_TURBO_ITERATIONS,frame_parms->N_RB_DL, abstraction_flag);
- if (!PHY_vars_UE->dlsch[i][j]) {
- LOG_E(PHY,"Can't get ue dlsch structures\n");
- exit(-1);
- } else
- LOG_D(PHY,"dlsch[%d][%d] => %p\n",UE_id,i,PHY_vars_UE->dlsch[i][j]);
- }
+ if (!PHY_vars_UE->dlsch[l][i][j]) {
+ LOG_E(PHY,"Can't get ue dlsch structures\n");
+ exit(-1);
+ } else
+ LOG_D(PHY,"dlsch[%d][%d] => %p\n",UE_id,i,PHY_vars_UE->dlsch[l][i][j]);
+ }
- PHY_vars_UE->ulsch[i] = new_ue_ulsch(frame_parms->N_RB_UL, abstraction_flag);
+ PHY_vars_UE->ulsch[i] = new_ue_ulsch(frame_parms->N_RB_UL, abstraction_flag);
- if (!PHY_vars_UE->ulsch[i]) {
- LOG_E(PHY,"Can't get ue ulsch structures\n");
- exit(-1);
- }
+ if (!PHY_vars_UE->ulsch[i]) {
+ LOG_E(PHY,"Can't get ue ulsch structures\n");
+ exit(-1);
+ }
- PHY_vars_UE->dlsch_SI[i] = new_ue_dlsch(1,1,NSOFT,MAX_TURBO_ITERATIONS,frame_parms->N_RB_DL, abstraction_flag);
- PHY_vars_UE->dlsch_ra[i] = new_ue_dlsch(1,1,NSOFT,MAX_TURBO_ITERATIONS,frame_parms->N_RB_DL, abstraction_flag);
+ PHY_vars_UE->dlsch_SI[i] = new_ue_dlsch(1,1,NSOFT,MAX_TURBO_ITERATIONS,frame_parms->N_RB_DL, abstraction_flag);
+ PHY_vars_UE->dlsch_ra[i] = new_ue_dlsch(1,1,NSOFT,MAX_TURBO_ITERATIONS,frame_parms->N_RB_DL, abstraction_flag);
- PHY_vars_UE->transmission_mode[i] = frame_parms->nb_antenna_ports_eNB==1 ? 1 : 2;
+ PHY_vars_UE->transmission_mode[i] = frame_parms->nb_antenna_ports_eNB==1 ? 1 : 2;
+ }
}
PHY_vars_UE->frame_parms.pucch_config_common.deltaPUCCH_Shift = 1;