fixed correlation computation in pucch_rx.c

openair1/PHY/NR_TRANSPORT/pucch_rx.c

@@ -53,7 +53,7 @@
 
 #include "T.h"
 
-#define DEBUG_NR_PUCCH_RX 1
+//#define DEBUG_NR_PUCCH_RX 1
 
 NR_gNB_PUCCH_t *new_gNB_pucch(void){
     NR_gNB_PUCCH_t *pucch;
@@ -1515,31 +1515,50 @@ void nr_decode_pucch2(PHY_VARS_gNB *gNB,
 
 #endif
             prod_re[aa] = simde_mm256_hadds_epi16(prod_re[aa],prod_re[aa]);// 0+1
+#ifdef DEBUG_NR_PUCCH_RX
+            printf("0.prod_re[%d] => (%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d)\n",aa,
+                   ((int16_t*)&prod_re[aa])[0],((int16_t*)&prod_re[aa])[1],((int16_t*)&prod_re[aa])[2],((int16_t*)&prod_re[aa])[3],
+                   ((int16_t*)&prod_re[aa])[4],((int16_t*)&prod_re[aa])[5],((int16_t*)&prod_re[aa])[6],((int16_t*)&prod_re[aa])[7],
+                   ((int16_t*)&prod_re[aa])[8],((int16_t*)&prod_re[aa])[9],((int16_t*)&prod_re[aa])[10],((int16_t*)&prod_re[aa])[11],
+                   ((int16_t*)&prod_re[aa])[12],((int16_t*)&prod_re[aa])[13],((int16_t*)&prod_re[aa])[14],((int16_t*)&prod_re[aa])[15]);
+#endif	    
             prod_im[aa] = simde_mm256_hadds_epi16(prod_im[aa],prod_im[aa]);
             prod_re[aa] = simde_mm256_hadds_epi16(prod_re[aa],prod_re[aa]);// 0+1+2+3
+#ifdef DEBUG_NR_PUCCH_RX
+            printf("1.prod_re[%d] => (%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d)\n",aa,
+                   ((int16_t*)&prod_re[aa])[0],((int16_t*)&prod_re[aa])[1],((int16_t*)&prod_re[aa])[2],((int16_t*)&prod_re[aa])[3],
+                   ((int16_t*)&prod_re[aa])[4],((int16_t*)&prod_re[aa])[5],((int16_t*)&prod_re[aa])[6],((int16_t*)&prod_re[aa])[7],
+                   ((int16_t*)&prod_re[aa])[8],((int16_t*)&prod_re[aa])[9],((int16_t*)&prod_re[aa])[10],((int16_t*)&prod_re[aa])[11],
+                   ((int16_t*)&prod_re[aa])[12],((int16_t*)&prod_re[aa])[13],((int16_t*)&prod_re[aa])[14],((int16_t*)&prod_re[aa])[15]);
+#endif	    
             prod_im[aa] = simde_mm256_hadds_epi16(prod_im[aa],prod_im[aa]);
             prod_re[aa] = simde_mm256_hadds_epi16(prod_re[aa],prod_re[aa]);// 0+1+2+3+4+5+6+7
-            prod_im[aa] = simde_mm256_hadds_epi16(prod_im[aa],prod_im[aa]);
-            prod_re[aa] = simde_mm256_hadds_epi16(prod_re[aa],prod_re[aa]);// 0+1+2+3+4+5+6+7+8+9+10+11+12+13+14+15
+#ifdef DEBUG_NR_PUCCH_RX
+            printf("2.prod_re[%d] => (%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d)\n",aa,
+                   ((int16_t*)&prod_re[aa])[0],((int16_t*)&prod_re[aa])[1],((int16_t*)&prod_re[aa])[2],((int16_t*)&prod_re[aa])[3],
+                   ((int16_t*)&prod_re[aa])[4],((int16_t*)&prod_re[aa])[5],((int16_t*)&prod_re[aa])[6],((int16_t*)&prod_re[aa])[7],
+                   ((int16_t*)&prod_re[aa])[8],((int16_t*)&prod_re[aa])[9],((int16_t*)&prod_re[aa])[10],((int16_t*)&prod_re[aa])[11],
+                   ((int16_t*)&prod_re[aa])[12],((int16_t*)&prod_re[aa])[13],((int16_t*)&prod_re[aa])[14],((int16_t*)&prod_re[aa])[15]);
+#endif	    
             prod_im[aa] = simde_mm256_hadds_epi16(prod_im[aa],prod_im[aa]);
           }
           int64_t corr_re=0,corr_im=0;
 
 
           for (int aa=0;aa<Prx;aa++) {
+
+            corr_re = ( corr32_re[symb][group][aa]+((int16_t*)(&prod_re[aa]))[0]+((int16_t*)(&prod_re[aa]))[8]);
+            corr_im = ( corr32_im[symb][group][aa]+((int16_t*)(&prod_im[aa]))[0]+((int16_t*)(&prod_im[aa]))[8]);
 #ifdef DEBUG_NR_PUCCH_RX
             printf("pucch2 cw %d group %d aa %d: (%d,%d)+(%d,%d) = (%d,%d)\n",cw,group,aa,
               corr32_re[symb][group][aa],corr32_im[symb][group][aa],
-              ((int16_t*)(&prod_re[aa]))[0],
-              ((int16_t*)(&prod_im[aa]))[0],
-              corr32_re[symb][group][aa]+((int16_t*)(&prod_re[aa]))[0],
-              corr32_im[symb][group][aa]+((int16_t*)(&prod_im[aa]))[0]);
+              ((int16_t*)(&prod_re[aa]))[0]+((int16_t*)(&prod_re[aa]))[8],
+              ((int16_t*)(&prod_im[aa]))[0]+((int16_t*)(&prod_im[aa]))[8],
+	      corr_re,corr_im
+              );
 
 #endif
 
-            corr_re = ( corr32_re[symb][group][aa]+((int16_t*)(&prod_re[aa]))[0]);
-            corr_im = ( corr32_im[symb][group][aa]+((int16_t*)(&prod_im[aa]))[0]);
-
             corr_tmp += corr_re*corr_re + corr_im*corr_im;
           } // aa loop
         }// group loop
@@ -1547,11 +1566,14 @@ void nr_decode_pucch2(PHY_VARS_gNB *gNB,
       if (corr_tmp > corr) {
          corr = corr_tmp;
          cw_ML=cw;
+#ifdef DEBUG_NR_PUCCH_RX
+         printf("slot %d PUCCH2 cw_ML %d, corr %llu\n",slot,cw_ML,corr);
+#endif
       }
     } // cw loop
     corr_dB = dB_fixed64((uint64_t)corr);
 #ifdef DEBUG_NR_PUCCH_RX
-    printf("slot %d PUCCH2 cw_ML %d, metric %d dB\n",slot,cw_ML,corr_dB);
+    printf("slot %d PUCCH2 cw_ML %d, metric %d \n",slot,cw_ML,corr_dB);
 #endif
     decodedPayload[0]=(uint64_t)cw_ML;
   }