Bugfix: Sampling efficiency in Widom insertion sample script (#4388)

As pointed out by @jonaslandsgesell, sampling only one particle insertion per MD step is terribly inefficient. Here I restore the inner loop in `samples/widom_insertion.py` to sample various particle insertion energies every time the MD integration is stopped.
espressomd · Nov 9, 2021 · b801493 · b801493
2 parents 731d76c + 181a9f5
commit b801493
Showing 1 changed file with 5 additions and 2 deletions.
diff --git a/samples/widom_insertion.py b/samples/widom_insertion.py
@@ -121,9 +121,12 @@
 particle_insertion_potential_energy_samples = []
 
 n_iterations = 500
+n_samples_per_iteration = 100
+
 for i in range(n_iterations):
-    particle_insertion_potential_energy_samples.append(
-        widom.calculate_particle_insertion_potential_energy(0))
+    for _ in range(n_samples_per_iteration):
+        particle_insertion_potential_energy_samples.append(
+            widom.calculate_particle_insertion_potential_energy(0))
     system.integrator.run(steps=500)
 
     if i % 20 == 0: