Frankly I'd be surprised if any "real" hardware random was more statistically random than say the Mersenne Twister (at multiple bit-widths say from 1 to 64). One of the most sensitive monte-carlo method uses is chaos-game rendering of IFS fractals - for example even generation over time of a standard Sierpinski triangle and the Mersenne works great - with the added bonus that the determinism means exactly the same image can be reproduced from the single seed. On 1 Oct 2013, at 02:44, Eugene Salamin wrote:
Skyrim appears to be a game. So use a pseudo if you need the same sequence. For serious scientific calculations, does anyone have a idea of how many bytes would need to be recorded?
For cryptography I would not trust a software algorithm. Even for a hardware device, you would want to be sure the NSA did not influence the design. Avoid high-level integrated devices in the critical path, and interface through standard off-the-shelf components.
How random are hardware-generated sequences? With biases removed and waiting out correlation times, they are completely random. Even then, a little bit of bias or correlation won't compromise cryptography.
-- Gene
________________________________ From: David Makin <makinmagic@tiscali.co.uk> To: math-fun <math-fun@mailman.xmission.com> Sent: Monday, September 30, 2013 6:11 PM Subject: Re: [math-fun] linear psu-random generators with linear "feed-in"
You coud record it - but then you could have memory issues - e.g. recording a whole 24 hour's play's worth of randoms for something like Skyrim is not going to be trivial (I love to be able to playback).
On 1 Oct 2013, at 01:39, Dan Asimov wrote:
I had the same thought as David, but then I thought: Why not record the sequence of hardware-generated numbers for future use if needed.
Just how random are hardward-generated sequences, anyway?
--Dan
On 2013-09-30, at 3:37 PM, David Makin wrote:
Precisely because hardware generated is *too* random since it's effectively non-deterministic, one cannot reproduce precisely the same sequence from a given seed - being able to do so is extremely useful in software terms.
On 30 Sep 2013, at 21:34, Eugene Salamin wrote:
These days, why would anyone bother with software generated random numbers? Use hardware generation, amplified noise or quantum detection. Bypass the requirement of algorithm verification. Bypass the problem of a limited number of seeds.
_______________________________________________ math-fun mailing list math-fun@mailman.xmission.com http://mailman.xmission.com/cgi-bin/mailman/listinfo/math-fun
The meaning and purpose of life is to give life purpose and meaning. The instigation of violence indicates a lack of spirituality.