diff --git a/src/force.c b/src/force.c
index 0347713..6756c5d 100644
--- a/src/force.c
+++ b/src/force.c
@@ -59,6 +59,7 @@ double computeForce(Parameter *param, Atom *atom, Neighbor *neighbor, int first_
         fz[i] = 0.0;
     }
 
+    LIKWID_MARKER_START("force");
     #pragma omp parallel for
     for(int i = 0; i < Nlocal; i++) {
         neighs = &neighbor->neighbors[i * neighbor->maxneighs];
@@ -131,6 +132,7 @@ double computeForce(Parameter *param, Atom *atom, Neighbor *neighbor, int first_
         TRACER_PRINT(fz[i], 'R');
         TRACER_PRINT(fz[i], 'W');
     }
+    LIKWID_MARKER_STOP("force");
 
     double E = getTimeStamp();
     TRACER_END;
diff --git a/src/main-stub.c b/src/main-stub.c
index c577f6a..8d3a8c0 100644
--- a/src/main-stub.c
+++ b/src/main-stub.c
@@ -211,11 +211,9 @@ int main(int argc, const char *argv[]) {
 
     double S, E;
     S = getTimeStamp();
-    LIKWID_MARKER_START("force");
     for(int i = 0; i < param.ntimes; i++) {
         computeForce(&param, atom, &neighbor, 0);
     }
-    LIKWID_MARKER_STOP("force");
     E = getTimeStamp();
     double T_accum = E-S;
     const double atoms_updates_per_sec = (double)(atom->Nlocal) / T_accum * (double)(param.ntimes * NEIGHBORS_LOOP_RUNS);