Peter Johansson <[email protected]> writes: > I'm curious what your use case is when this is an actual limitation. > FWIW, I reckon it's against the idea with gsl_rng as it was design to be > polymorphic so one can easily replace one generator with another.
...so the answer is no. Thanks, anyway. I can just continue to use MT19937/dSFMT when this feature is more important to me than being able to swap out the RNG. Anye
