Re: Random numbers generation
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Thomas Walter wrote: On Thu, 2005-12-08 at 00:22, Stuart Prescott wrote: The GSL documentation talks a bit about the different generators available and I highly recommend some of the references therein (Park Miller and L'Ecuyer in particular). http://www.gnu.org/software/gsl/manual/gsl-ref_17.html Just for completeness to the random generators mentioned above. If your system has '/dev/random' and/or '/dev/urandom' then you have also a good random to read from. This is feed by several non deterministic data available on your system. The more usage and traffic you have, the better. Those random generators have to be avoided, except in very specific cases as they are not reproducable; there is no way to replay them. Yannick -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.1 (GNU/Linux) Comment: Using GnuPG with Red Hat - http://enigmail.mozdev.org iD8DBQFDmJtoZY3s/IVpRFIRAhWVAKCZBuHVHPpzr5lHumXfBJcpDSidIQCfTlpn 72ZoHJeAXhK+ea58+TGQf4Y= =b3WV -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Random numbers generation
Hi Yannick, On Thu, 2005-12-08 at 21:45, Yannick Patois wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Thomas Walter wrote: On Thu, 2005-12-08 at 00:22, Stuart Prescott wrote: The GSL documentation talks a bit about the different generators available and I highly recommend some of the references therein (Park Miller and L'Ecuyer in particular). http://www.gnu.org/software/gsl/manual/gsl-ref_17.html Just for completeness to the random generators mentioned above. If your system has '/dev/random' and/or '/dev/urandom' then you have also a good random to read from. This is feed by several non deterministic data available on your system. The more usage and traffic you have, the better. Those random generators have to be avoided, except in very specific cases as they are not reproducable; there is no way to replay them. That's the definition of random -- you do not know what's comming next and no chance to repeat. Some random-generator algorithms are very good to show no pattern. Even in a higher dimension space. But true random has no pattern, no matter how long and good you look at. Another try may be a chaotic system, but I do not know. If you want a replay then simply remember the sequence. Or prepare the sequence by anything you want and then feed it into your application. That seems to be the best approach as the using application is independent from source of random varying from poor til true. Kind Regards, Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Random numbers generation
¿Is there any way of making differents results without having to change the seed from the outside? Hi Lisandro. setenv(3) ? Regards. -- Homepage : http://geocities.com/arhuaco The first principle is that you must not fool yourself and you are the easiest person to fool. -- Richard Feynman.
Re: Random numbers generation
Lisandro Damian Nicanor Perez Meyer [EMAIL PROTECTED] writes: ¿Is there any way of making differents results without having to change the seed from the outside? It looks like you can also call gsl_rng_set(r, some_seed); or assign directly to gsl_rng_default_seed, at least with the version of gsl available in sid (unstable). -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) Finger [EMAIL PROTECTED] (NOT a valid e-mail address) for more info. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Random numbers generation
El Mié 07 Dic 2005 15:06, David Dooling escribió: On 12/7/05, Aaron M. Ucko [EMAIL PROTECTED] wrote: Lisandro Damian Nicanor Perez Meyer [EMAIL PROTECTED] writes: ¿Is there any way of making differents results without having to change the seed from the outside? It looks like you can also call gsl_rng_set(r, some_seed); or assign directly to gsl_rng_default_seed, at least with the version of gsl available in sid (unstable). I have found the best approach is to have a default seed (probably the default that GSL uses) and allow the setting of an alternative seed via a command-line option or input file (using one of the above methods). You do not want the seed to be ``random'', as that makes bugs difficult to find. You want your output to be deterministic, even when using random numbers. dd -- David Dooling Well, thank you everyone! You're incredible! Cheers, Damian.- -- I was married by a judge. I should have asked for a jury. -- Groucho Marx
