From the Intel page, this is apparently already part of each Pentium. I don't know if the latest Pentia have this capability or not. http://www.intel.com/design/security/rng/rnghow.htm "The Intel Random Number Generator was designed in accordance with two critical design criteria: * The RNG must be based on a source that's truly random and based on properties native to silicon circuitry. * The design must preserve this randomness while providing the needed performance and circuit stability across temperature and other environmental effects. To meet these criteria and others, the Intel RNG uses white noise as the random source, then processes the data through a variety of circuit elements. Intel RNG is available in Intel® chipsets, beginning with the Intel® 810 Chipset for the Intel® Celeron_ processor. This capability is also planned for the next-generation Intel® 8xx chipset for the Pentium® III processor." At 10:44 AM 2/23/2004, Steve Gray wrote:
Perhaps a simpler way to get a truly random bit stream is to amplify and digitize the inherent random electrical noise produced by a resistor at a nonzero temperature. This is not a new idea, obviously, and is probably used in some practical applications.
Steve Gray
----- Original Message ----- From: "Eugene Salamin" <gene_salamin@yahoo.com> To: "math-fun" <math-fun@mailman.xmission.com> Sent: Monday, February 23, 2004 10:10 AM Subject: Re: [math-fun] preprint: uncertainty principle in AIT
I failed to find in the paper any connection between Godel's incompleteness and randomness in physics, other than a stated conjecture that such exists.
I see no application of Godel incompleteness to physics. Questions such as whether the whole of physics is finitely describable are unanswerable, since, unlike mathematics, physics is limited by what can be measured rather than by what can be postulated.
But here is something to contemplate. A polarizing beam splitter is a device that splits an input light beam into its horizontally and vertically polarized components, and both of these beams are output. Unlike Polaroid sheet, there is no absorption. If the input light is linearly polarized at 45 degrees, the output beams have equal power. Measure each output beam with a detector that can detect individual photons. For each horizontally polarized photon, output a zero; for each vertically polarized photon, output a one. Now you have a random bit stream. Because of alignment and material imperfections, the probability of 0 and 1 cannot be made precisely 1/2. Computer scientists expend a tremendous effort trying to design computational pseudorandom number generators. But through quantum mechanics, you get it for free as a gift from nature.
Gene
--- Mike Stay <staym@clear.net.nz> wrote:
Greg Chaitin defined Omega_U as the halting probability for a self-delimiting Turing machine U. (Self-delimiting means that the domain of U is a prefix-free set; that is, no program is the prefix of another.) I noticed that Chaitin's theorem about the (un)computability of bits of Omega_U can be recast as an uncertainty principle. My professor and I have written a paper; a preprint is available at
http://www.cs.auckland.ac.nz/CDMTCS/researchreports/235cris.pdf
I'd appreciate any comments.
Thanks!
-- Mike Stay staym@clear.net.nz http://www.cs.auckland.ac.nz/~msta039