Beefy Boxes and Bandwidth Generously Provided by pair Networks
There's more than one way to do things

Re: Random numbers Perl ok! (I think)

by a-mused (Initiate)
on Oct 23, 2003 at 13:31 UTC ( #301577=note: print w/replies, xml ) Need Help??

in reply to Random numbers Perl ok! (I think)

In addition to Abigail's (as usual ;) correct assertion that Perl uses the OS' underpinnings to do (pseudo)random number generation, do note that as good as the PRNG is with drand48 (on most modern linux distributions), 48 bits is not sufficent for crypographic needs. For most computing purposes, it's sufficient - just not crypto. For that, there are several packages which work out a bit better - as have been mentioned earlier.

The problem doesn't stem from needing to draw 2G random numbers in a single run but how random the sequence is for those 2G numbers over multiple runs of smaller draws over time. Besides, to a human, 2G can seem like a big number but to a computer it's *teeny*.

As an introduction into the visualization of PRNG data, I'd recommend having a look at this where they've analyzed the pseudorandomness in the TCP sequences of various stack implimentations.

Kind Regards,

Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://301577]
[choroba]: Most coffees? Depends on the country :-)
[choroba]: oh, you mean a cup with no water.
[Discipulus]: you are rigth, gory details but i remember from dunnowhere that concetrate caffein is more effective than brown water almost coffe..
[LanX]: prepared both kinds a 100 times and thinks the amount of grind code is about the same
Discipulus less effective i mean

How do I use this? | Other CB clients
Other Users?
Others drinking their drinks and smoking their pipes about the Monastery: (10)
As of 2017-01-16 10:55 GMT
Find Nodes?
    Voting Booth?
    Do you watch meteor showers?

    Results (148 votes). Check out past polls.