SPIN's True Randomness can't be BUSTed

kuroneko

FWIW, an evil test sequence (10 runs, failure count listed out of 100):

23/20/28/31/17/24/24/30/30/[b]99[/b]

Yes, the last count is correct, 99. It's a rare occurrence when the cog running the test has a specific cnt alignment wrt its hub window and the driver is initialised at a specific point resulting in zero seed runs. Just thought I'd share this.

cessnapilot

@kuroneko

Thank you. 2-3 days before, after running BUST ~200 times, I noticed sometime similar behaviour of the SPIN_TriPack TRNG in long test runs using only REBOOT without hardware reset or power off/on restart. Before that usually there was some 'degradation of performance' , as in the failer sequence you presented. A power off/on cured the symptoms, but of course, did not cure the disease. I am working on a new version of the generator to solve that.

On the other hand it is good to see, that the simple tests in BUST are very sensitive to quickly pinpoint a problem in seemingly random series.

Cheers,

Istvan

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Intentionally Left Blank

heater

Clock Loop: Schrodinger's cat has been in that box since 1935. My bet is that it is dead now [noparse]:)[/noparse]

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
For me, the past is not over yet.

Leon

Something analogous is used in quantum cryptography, to detect if a message has been intercepted when passed down a fibre-optic cable.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Leon Heller
Amateur radio callsign: G1HSM

HollyMinkowski

heater said...
Clock Loop: Schrodinger's cat has been in that box since 1935. My bet is that it is dead now [noparse]:)[/noparse]

I'm going to report Schrodinger to PETA.

LoL