maki/openairinterface5g
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

NR UE: further simplify and optimize the usage of the PDCCH LLR array

Browse Source 
The initial version of this patch was provided by "Laurent THOMAS <laurent.thomas@open-cells.com>"
This commit is contained in:
Thomas Schlichter
2025-08-08 17:41:15 +02:00
parent a6ccdd1340
commit 4de0e54a3a
1 changed files with 16 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files
openair1/PHY/NR_UE_TRANSPORT/dci_nr.c
+16 -19
View File
@@ -70,11 +70,10 @@
// after removing the 3 DMRS RE, the RB contains 9 RE with PDCCH // after removing the 3 DMRS RE, the RB contains 9 RE with PDCCH
#define RE_PER_RB_OUT_DMRS 9 #define RE_PER_RB_OUT_DMRS 9
static void nr_pdcch_demapping_deinterleaving(c16_t *llr, static void nr_pdcch_demapping_deinterleaving(uint32_t coreset_nbr_rb,
c16_t llr[][coreset_nbr_rb * RE_PER_RB_OUT_DMRS],
c16_t *e_rx, c16_t *e_rx,
uint8_t coreset_time_dur, uint8_t coreset_time_dur,
uint8_t start_symbol,
uint32_t coreset_nbr_rb,
uint8_t reg_bundle_size_L, uint8_t reg_bundle_size_L,
uint8_t coreset_interleaver_size_R, uint8_t coreset_interleaver_size_R,
uint8_t n_shift, uint8_t n_shift,
@@ -168,15 +167,14 @@ static void nr_pdcch_demapping_deinterleaving(c16_t *llr,
} }
int rb_count = 0; int rb_count = 0;
const int data_sc = RE_PER_RB_OUT_DMRS; // 9 sub-carriers with data per PRB
for (int c_id = 0; c_id < number_of_candidates; c_id++) { for (int c_id = 0; c_id < number_of_candidates; c_id++) {
for (int symbol_idx = start_symbol; symbol_idx < start_symbol + coreset_time_dur; symbol_idx++) { for (int symbol_idx = 0; symbol_idx < coreset_time_dur; symbol_idx++) {
for (int cce_count = 0; cce_count < L[c_id]; cce_count++) { for (int cce_count = 0; cce_count < L[c_id]; cce_count++) {
for (int k = 0; k < NR_NB_REG_PER_CCE / reg_bundle_size_L; k++) { // loop over REG bundles for (int k = 0; k < NR_NB_REG_PER_CCE / reg_bundle_size_L; k++) { // loop over REG bundles
int f = f_bundle_j_list_ord[c_id][k + NR_NB_REG_PER_CCE * cce_count / reg_bundle_size_L]; int f = f_bundle_j_list_ord[c_id][k + NR_NB_REG_PER_CCE * cce_count / reg_bundle_size_L];
c16_t *in = llr + (f * B_rb + symbol_idx * coreset_nbr_rb) * data_sc; c16_t *in = llr[symbol_idx] + f * B_rb * RE_PER_RB_OUT_DMRS;
// loop over the RBs of the bundle // loop over the RBs of the bundle
memcpy(e_rx + data_sc * rb_count, in, B_rb * data_sc * sizeof(*e_rx)); memcpy(e_rx + RE_PER_RB_OUT_DMRS * rb_count, in, B_rb * RE_PER_RB_OUT_DMRS * sizeof(*e_rx));
rb_count += B_rb; rb_count += B_rb;
} }
} }
@@ -347,18 +345,18 @@ void nr_rx_pdcch(PHY_VARS_NR_UE *ue,
get_coreset_rballoc(rel15->coreset.frequency_domain_resource,&n_rb,&rb_offset); get_coreset_rballoc(rel15->coreset.frequency_domain_resource,&n_rb,&rb_offset);
const int antRx = frame_parms->nb_antennas_rx; const int antRx = frame_parms->nb_antennas_rx;
// Pointer to llrs, 4-bit resolution. // Pointer to llrs
int32_t llr_size = (rel15->coreset.duration + rel15->coreset.StartSymbolIndex) * n_rb * RE_PER_RB_OUT_DMRS; const int symb_size = n_rb * RE_PER_RB_OUT_DMRS;
c16_t llr[llr_size]; c16_t llr[rel15->coreset.duration][symb_size];
LOG_D(NR_PHY_DCI, LOG_D(NR_PHY_DCI,
"pdcch coreset: freq %x, n_rb %d, rb_offset %d\n", "pdcch coreset: freq %x, n_rb %d, rb_offset %d\n",
rel15->coreset.frequency_domain_resource[0], rel15->coreset.frequency_domain_resource[0],
n_rb, n_rb,
rb_offset); rb_offset);
for (int s = rel15->coreset.StartSymbolIndex; s < rel15->coreset.StartSymbolIndex + rel15->coreset.duration; s++) { for (int s = 0; s < rel15->coreset.duration; s++) {
LOG_D(NR_PHY_DCI, "in nr_pdcch_extract_rbs_single(rxdataF -> rxdataF_ext || dl_ch_estimates -> dl_ch_estimates_ext)\n"); LOG_D(NR_PHY_DCI, "in nr_pdcch_extract_rbs_single(rxdataF -> rxdataF_ext || dl_ch_estimates -> dl_ch_estimates_ext)\n");
const int arraySz = ceil_mod(n_rb * RE_PER_RB_OUT_DMRS, 32); const int arraySz = ceil_mod(symb_size, 32);
__attribute__((aligned(32))) c16_t rxdataF_ext[antRx][arraySz]; __attribute__((aligned(32))) c16_t rxdataF_ext[antRx][arraySz];
__attribute__((aligned(32))) c16_t pdcch_dl_ch_estimates_ext[antRx][arraySz]; __attribute__((aligned(32))) c16_t pdcch_dl_ch_estimates_ext[antRx][arraySz];
nr_pdcch_extract_rbs_single(ue->frame_parms.samples_per_slot_wCP, nr_pdcch_extract_rbs_single(ue->frame_parms.samples_per_slot_wCP,
@@ -369,7 +367,7 @@ void nr_rx_pdcch(PHY_VARS_NR_UE *ue,
arraySz, arraySz,
rxdataF_ext, rxdataF_ext,
pdcch_dl_ch_estimates_ext, pdcch_dl_ch_estimates_ext,
s, rel15->coreset.StartSymbolIndex + s,
frame_parms, frame_parms,
rel15->coreset.frequency_domain_resource, rel15->coreset.frequency_domain_resource,
rel15->BWPStart); rel15->BWPStart);
@@ -396,20 +394,19 @@ void nr_rx_pdcch(PHY_VARS_NR_UE *ue,
antRx, antRx,
log2_maxh); // log2_maxh+I0_shift log2_maxh); // log2_maxh+I0_shift
UEscopeCopy(ue, pdcchRxdataF_comp, rxdataF_comp, sizeof(struct complex16), antRx, n_rb * RE_PER_RB_OUT_DMRS, 0); UEscopeCopy(ue, pdcchRxdataF_comp, rxdataF_comp, sizeof(struct complex16), antRx, symb_size, 0);
if (antRx > 1) if (antRx > 1)
nr_pdcch_detection_mrc(arraySz, rxdataF_comp); nr_pdcch_detection_mrc(arraySz, rxdataF_comp);
nr_pdcch_llr(n_rb * RE_PER_RB_OUT_DMRS, rxdataF_comp[0], llr + s * n_rb * RE_PER_RB_OUT_DMRS); nr_pdcch_llr(symb_size, rxdataF_comp[0], llr[s]);
} }
UEscopeCopy(ue, pdcchLlr, llr, sizeof(c16_t), 1, llr_size, 0); UEscopeCopy(ue, pdcchLlr, llr, sizeof(c16_t), 1, rel15->coreset.duration * symb_size, 0);
nr_pdcch_demapping_deinterleaving(llr, nr_pdcch_demapping_deinterleaving(n_rb,
llr,
pdcch_e_rx, pdcch_e_rx,
rel15->coreset.duration, rel15->coreset.duration,
rel15->coreset.StartSymbolIndex,
n_rb,
rel15->coreset.RegBundleSize, rel15->coreset.RegBundleSize,
rel15->coreset.InterleaverSize, rel15->coreset.InterleaverSize,
rel15->coreset.ShiftIndex, rel15->coreset.ShiftIndex,