computing the run time of random algorithm based on metropolis hestings rule

How do you compute the run time of your algorithm, if you know you want n samples; your sampling method is based on metropolis heastings, so you have long does it take for the algorithm to give you the desired n samples?