Computational performance of route planning algorithms has become increasingly important in recent real navigation applications with many simultaneous route requests. Navigation applications should recommend routes as quickly as possible and preferably with some added value. This paper presents a performance evaluation of the main part of probabilistic time-dependent route planning algorithm. The main part of the algorithm computes the full probability distribution of travel time on routes with Monte Carlo simulation. Experiments show the performance of the algorithm and suggest real possibilities of use in modern navigation applications.