OpenEuphoria

Re: 32-bit random numbers

• Posted by "Igor Kachan" <kinz at peterlink.ru> Jul 10, 2004
• 589 views

Hi Al,

----------
> From: Al Getz <guest at RapidEuphoria.com>
> To: EUforum at topica.com
> Subject: Re: 32-bit random numbers
> Sent: 10 jul 2004 y. 21:10
> 
> posted by: Al Getz <Xaxo at aol.com>
> 
> 
> Hi igor,
> 
> "K * (rand(N) - 1) + rand(K) - 1"
> 
> It looks flat, but what is your intended advantage over
> N*(rand(N)-1)+rand(N)-1, {N=#1000}
> ??

Al, the proposed function is:

constant K = 1073741823 -- max EU integer
function rand_atom(integer N)
    return K * (rand(N) - 1) + rand(K) - 1
end function

Did you see in explanation that N << K ? 

This parameter (N) is the key
to understanding of flat character
of this distribution. We need the
maximum range of integer random
atoms with step = 1 in EU.
This is our task now.

And I am just trying to solve this
concrete task as far as I understand it.

Yes, formulas are similar, but they
are *different*, Al.

I can say nothing about the distribution
on your formula. I just see it is not flat.

But you do see just now that proposed new
function seems to give the flat distribution.

> Notice you can also extend that one:
> 
> N*(rand(N)-1)+rand(N)-1
> 
> goes to:
> 
> N*N*(rand(N)-1)+N*(rand(N)-1)+rand(N)-1
> 
> if N is an integer multiple of the number
> of terms minus 1,
> or:
> 
> N*N*(rand(N2)-1)+N*(rand(N2)-1)+rand(N2)-1
> 
> if N2=number of terms minus 1 .
> 
> This gives the ability to get higher 
> and higher random numbers 

OK OK OK, Al, what a problem, use please your
functions as you want for the suited tasks,
and tell us what were the distributions.

> 
> Take care,
> Al

Regards,
Igor Kachan
kinz at peterlink.ru