Re: Question about RNGs used in oct files, and multiple instances of Octave

[ghaverla]

On Fri, 30 May 2014 10:08:40 -0700 (PDT)
Michael Creel <address@hidden> wrote:

> I understand that setting the seed could be useful for other reasons,
> but I'm mostly interested that the sequences generated on the
> separate instances of Octave should not be dependent. I also know
> that there are sophisticated solutions for generating random numbers
> on clusters. However, with the long period of the MT generator, I
> think that different initial starting points will be sufficient for
> my purposes. If I'm mistaken about that, though, I'd like to know.

If you are using multiple generators in parallel, if at some point the
internal states of some generators become the same, they are now
dependent.  They are perfectly correlated with each other.

It is also possible for generators to become in step with each other,
which is just as bad (one sequence becomes the other sequence after a
lag).

I would suggest that if you are running a stochastic program on a
cluster of machines, and need to be sure about randomness, you get
hardware random number generators to generate true random numbers.  If
nothing else, they are probably faster (or could be).

Gord