I used both in the Monte Carlo calculations for my PhD thesis back around 1990. Much better convergence than generic quasi-random numbers.