Date: Sat, 10 Apr 2004 23:44:36 +0400 From: Roman Kurakin <rik@cronyx.ru> To: Luigi Rizzo <rizzo@icir.org> Cc: Nate Lawson <nate@root.org> Subject: Re: cvs commit: src/sys/modules/random Makefile src/sys/dev/random randomdev.h randomdev_soft.c randomdev_soft.h yar Message-ID: <40784EA4.2020604@cronyx.ru> References: <4077FE63.4010602@cronyx.ru> <9455.1081606743@critter.freebsd.dk> <20040410083856.A30378@xorpc.icir.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Luigi Rizzo wrote:
>On Sat, Apr 10, 2004 at 04:19:03PM +0200, Poul-Henning Kamp wrote:
>...
>
>
>>>How about computations?
>>>
>>>
>>Well, depends what kind of computations you are talking about obviously.
>>
>>I specifically didn't list that because I have seen so many examples
>>of monte carlo analysis done with rand() that I could cry. People
>>
>>
>
>maybe just as many as those that [unvoluntarily] rely
>on uninitialized variables :)
>
My knowledge in that area may be near those one. (I am about who may use
insystem
random, not about variables :-) But as far as I know there is some
methods of estimation
that this generator is random enougth for that kind of computation. So
we may have
a set of such tests which would show how random we are. And I am sure,
most of task
didn't require supper-random-generator, but having some
fast-good-enougth without
special hardware would be very nice. Probably if we would have some
gradation here we
might have a few computation levels.
Again, this was just some thoughts. This is not my area.
Best regards,
rik
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?40784EA4.2020604>