On two conflicting estimates of the Peierls stress