Add timer

Author: AsTonyshment

source/source_lcao/module_operator_lcao/td_nonlocal_lcao.cpp

@@ -173,6 +173,7 @@ void hamilt::TDNonlocal<hamilt::OperatorLCAO<TK, TR>>::calculate_HR()
 
         if (use_gpu)
         {
+            ModuleBase::timer::tick("TD_Efficiency", "snap_psibeta_gpu_total");
 #ifdef __CUDA
             // GPU path: Atom-level GPU batch processing
             module_rt::gpu::snap_psibeta_atom_batch_gpu(orb_,
@@ -187,9 +188,11 @@ void hamilt::TDNonlocal<hamilt::OperatorLCAO<TK, TR>>::calculate_HR()
                                                         nlm_dim,
                                                         nlm_tot);
 #endif
+            ModuleBase::timer::tick("TD_Efficiency", "snap_psibeta_gpu_total");
         }
         else
         {
+            ModuleBase::timer::tick("TD_Efficiency", "snap_psibeta_cpu_total");
             // CPU path: OpenMP parallel over neighbors to compute nlm_tot
 #pragma omp parallel for schedule(dynamic)
             for (int ad = 0; ad < adjs.adj_num + 1; ++ad)
@@ -228,6 +231,7 @@ void hamilt::TDNonlocal<hamilt::OperatorLCAO<TK, TR>>::calculate_HR()
                     }
                 }
             }
+            ModuleBase::timer::tick("TD_Efficiency", "snap_psibeta_cpu_total");
         }
 
         // 2. calculate <psi_I|beta>D<beta|psi_{J,R}> for each pair of <IJR> atoms

source/source_lcao/module_rt/snap_psibeta_half_tddft.cpp

@@ -110,7 +110,11 @@ void snap_psibeta_half_tddft(const LCAO_Orbitals& orb,
     int required_size = calc_r ? 4 : 1;
     if (nlm.size() != required_size) nlm.resize(required_size);
 
-    if (nproj == 0) return;
+    if (nproj == 0)
+    {
+        ModuleBase::timer::tick("module_rt", "snap_psibeta_half_tddft");
+        return;
+    }
 
     // 2. Determine total number of projectors and identify active ones based on cutoff
     int natomwfc = 0;