Does the HP Prime have HWRNG?

[TheLastMillennial]

As I was comparing the HP Prime to a TI-89 ti, I discovered that the TI-89 has a Hardware Random Number Generator. Although I would assume the HP Prime would have such a feature, I'm not certain and I didn't see it documented anywhere. I was wondering if someone could clear this up?

[Peet]

(01-13-2021 05:35 AM)TheLastMillennial Wrote:  I didn't see it documented anywhere. I was wondering if someone could clear this up?

from the manual: random()

[StephenG1CMZ]

I think random() is a pseudo-random number generator, rather than a hardware random number generator.

[C.Ret]

No doubt, even the in-display [ [?] Help ] key state it.

Quote:RANDOM([a],[b][c])

Returns a pseudo-random number generated using a seed value, and updates the seed value.

[Peet]

A few questions:

What did the TO mean, a) true random or b) hardcoded funktion?
If he means "a", how does this work on the TI?
Is the sequence "randseed() random()" a true/hardware random number generator?

[KeithB]

No, a hardware random number generator (Or TRNG) does not need a seed. On the TI cc1310 it uses: "The TRNG is built on 24 ring oscillators that create unpredictable output to feed a complex nonlinear combinatorial circuit.That post-processing of the output data is required to obtain cryptographically secure random data."

[Peet]

(01-13-2021 03:57 PM)KeithB Wrote:  a hardware random number generator (Or TRNG) does not need a seed.

It was a philosophical question. Isn't a PRN with a time-based seed not a TRN?

(01-13-2021 03:57 PM)KeithB Wrote:  On the TI cc1310 ...

Is this in the way the TI-89ti does it? I can't find any info to TRNG on this calculator.

[KeithB]

"It was a philosophical question. Isn't a PRN with a time-based seed not a TRN?"

Nope. It still has a period, and is deterministic. Cryptologists are very, very picky about their RNG's.

I have no idea what the TI calculator uses. I just happen to be familiar with the CC1310.

My google-foo failed looking up the TI-89 ti. Everything I see says it is a Motorola 68000 processor, which would mean no TRNG.

[Peet]

(01-13-2021 07:33 PM)KeithB Wrote:  My google-foo failed looking up the TI-89 ti. Everything I see says it is a Motorola 68000 processor, which would mean no TRNG.

I also found nothing about TRNG on the TI-89ti. So I assumed the TO may asked not about true random but a hardcoded function and asked for more information.

(01-13-2021 07:33 PM)KeithB Wrote:  It still has a period, and is deterministic.

assumption (philosophical not science):
randseed() without parameter uses the current time as seed. The start time of a manually started program cannot be determined or reproduced. So the first random() may be considered as a true random number. No clock in the world ist 100% accurate. If you use a new randseed() without parameter after the time when the inaccuracy of one's own clock occurred it also isn't a period.

[KeithB]

I am pretty sure that the NSA has the algorithm for every calculator's random function. Even changing the seed regularly will not protect your generator.

[Peet]

(01-13-2021 10:05 PM)KeithB Wrote:  I am pretty sure that the NSA has the algorithm for every calculator's random function.

On paper, scanned with Xerox

[Wolfgang]

(01-13-2021 07:33 PM)KeithB Wrote:  "It was a philosophical question. Isn't a PRN with a time-based seed not a TRN?"

Nope. It still has a period, and is deterministic. Cryptologists are very, very picky about their RNG's.

Some stuff:
https://csrc.nist.gov/projects/random-bit-generation

https://www.picoquant.com/scientific/pro...duct-study

[ggauny@live.fr]

Hi Wolfgang my friend,

Happy to read you!Do you have tried my little routine "Advent Sunday" ?
May be you want to add to your DFX.

Take care of you.

[Wolfgang]

(01-19-2021 12:26 PM)ggauny@live.fr Wrote:  May be you want to add to your DFX.

Cher Gerard,

I take a look at first advent sunday.

My HP11C purchased in the early eighties does generate Pseudo Random Numbers passing Knuth's spectral test (manual page 54). For my purposes enough.

BTW:
PM and Mail are ending in nirvana? ;-) No answers since last summer...


Best wishes,
Wolfgang

[TheLastMillennial]

Thank you for your answers everyone! I don't know what kind of HWRNG the TI-89 has, I was just curious if the HP Prime had one at all. It's interesting it doesn't have one but I don't expect that to be a problem at all. I completely forgot how well documented the help menu is, I didn't even think to check it on the RANDOM command!

[John Keith]

Another relevant point: Home on the prime uses the same linear congruential PRNG as other 12-digit calculators have since the 71b. CAS mode uses the Mersenne Twister algorithm.

[KeithB]

Which is why I wrote a Mersenne Twister for the Prime. 8^)

[Wolfgang]

(01-27-2021 09:46 PM)KeithB Wrote:  Which is why I wrote a Mersenne Twister for the Prime. 8^)

Wow, please show us the code!
Do you use TT800 or MT19937?
I'm interested in initialization, speed, and purposes.

https://github.com/python/cpython/blob/3.../random.py
https://docs.python.org/3/library/random.html

[KeithB]

I just searched the Prime Library for "Twister"
https://www.hpmuseum.org/forum/thread-7716.html

Of course, some folks suggested improvements. 8^)

[Wolfgang]

(01-28-2021 02:53 PM)KeithB Wrote:  I just searched the Prime Library for "Twister"
https://www.hpmuseum.org/forum/thread-7716.html

Of course, some folks suggested improvements. 8^)

OMG, you are right! Sorry!
Thank you! Very nice work!

A very nice article:
http://h20331.www2.hp.com/hpsub/download...ors_V1.pdf

From Page 4: The Randomness of HP by Nelson&Shammas,
Appendix A HP's Pseudorandom Number Generator, Appendix B Selected PRNG Programs