In recent tests, the best distribution estimator was a geometric mean because 
the elapsed-time distribution was log-normal.

If each iteration is not too short (to avoid timing inaccuracies) and not too 
long (to avoid spending a day benchmarking), you'd better get each elapsed-time 
individually and perform proper statistics on them.

What you most certainly want is the mode of the distribution of these values.

Reply via email to