Re: Question about RNGs used in oct files, and multiple instances of Octave

[ghaverla]

If one looks at the English Wikipedia page on comparison of hardware
random number generators, the second last entry is from Flying Stone
Technology of Japan, for a USB based hardware RNG for $35, $36 or $37
(without enclosure, with heat shrink tube for enclosure, or "fancy"
enclosure).  It apparently operates at about 588 kbit/s and it has been
tested:

http://www.cacert.at/cgi-bin/rngresults

It generates bits via ADC quantization error (NeuG-1.0).

http://www.gniibe.org/memo/development/gnuk/rng/neug.html
http://www.seeedstudio.com/wiki/FST-01

Can a person combine that with haveged, egd (this is a Perl script) or
similar to get faster production of RNs?  Among some methods mentioned,
was to use the hardware RNG to generate starting seeds for a PRNG.
Draw a seed, feed it into PRNG, generate N RNs, repeat.

For N not being too large, even if two (or more) RNGs in a cluster
became synchronized (same internal state), at the next re-seeding they
would likely not be synchronized any more.

That list of test results has prices in it (usually), so you might be
able to roll your own.

If you had multiple FST-01 units in a cluster, I would still want to
compare their outputs to see if there are correlations.  If you open a
door (to provide ventilation), do bit patterns change?  (periodic
heating would probably be easiest to use)

Gord