**Subject:** Re: System clock resolution and random numbers

**To:** Perry E. Metzger *<perry@piermont.com>*

**From:** John C. Hayward *<John.C.Hayward@wheaton.edu>*

**List:** current-users

**Date:** 05/01/1997 01:24:00
On Wed, 30 Apr 1997, Perry E. Metzger wrote:
>
> "John C. Hayward" writes:
> > Use random - it has a larger facility than 2^32 (from man 3 random)
> > ===
> > The random() function uses a non-linear additive feedback random number
> > generator employing a default table of size 31 long integers to return
> > successive pseudo-random numbers in the range from 0 to (2**31)-1.
> > The period of this random number generator is very large, approximately
> > 16*((2**31)-1).
> > ===
> > This should be large enough for _any_ application.
I should have said for the 256 long integer which has a period of > 2^69.
>
> Not for cryptography. Not by a long shot.
>
> Also, that random number generator pretty much sucks for many classes
> of monte carlo simulations.
If you compute at the rate of 1 billion events per second (~2^30) for
a year (~2^25 seconds) it would take 2^14 years to cycle for the 256 long
integer state.
johnh...
>
> Perry
>