[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: streams and substreams

This page is part of the web mail archives of SRFI 27 from before July 7th, 2015. The new archives for SRFI 27 contain all messages, not just those from before July 7th, 2015.

> > A question about the substream issue: Why are there several streams and 
> > within each streams several substreams? (What is the rationale of having
> > exactly two levels in the relations of the virtual streams? One could also 
> > just
> > have several streams or also provide subsubstreams etc.)
> I don't know the answer to this question.  There is an example in Section
> 4 of streams001.pdf, but I cannot construct the general principle from the
> example.

I thought a bit more about this.

It would appear to require a good bit of analysis and computation to choose
good offset parameters to obtain independent sub$^n$-streams; this is
mentioned in L'Ecuyer's paper in the section where he explains how he chose
the particular offsets to obtain good random substreams.

Thus, it would appear to be difficult to divide the stream of random numbers
into an arbitrary number of substreams "on the fly", as it were; perhaps
streams and substreams are just a compromise that fills the practical need
of a particular set of applications, which, in this case, I believe are
simulations of various kinds.