[talk] Testing Hardware RNG

[George Rosamond]

On 1/24/25 11:39, jpb wrote:
> On Fri, 24 Jan 2025 11:09:01 -0500
> George Rosamond <george at ceetonetechnology.com> wrote:
> 
>> On 1/24/25 09:47, Isaac (.ike) Levy wrote:
>>> Hi All,
>>>
>>> Does anyone have any good reccomendations for tools or process
>>> testing to find good RNG entropy?  Is Diehard still a thing,
>>> (Dieharder?)
>>>
>>> Context:
>>> I recall quite some years back, folks here testing some hardware
>>> RNG's with me.
>>>
>>> On a lark, a hardware person I know just whipped up some hardware
>>> RNG prototypes, still playing with them, (usb thingies spewing
>>> bits).
>>>
>>> Anyhow, it's been a while, and I'm interested in testing and test
>>> methodology, but can't for the life of me remember how we tested
>>> quality of entropy before. 
>>
>> Others can reply to lines above, but I do vividly remember discussing
>> RNG testing for FIPS after some meeting....
>>
>> I'll dive in a little deeper later, but I know there's others with
>> experiences on this.
>>
>> g
>>
> 
> [snip]
> 
> https://henric.org/ may be informative.

So it looks like math/dieharder is up-to-date in both FreeBSD and
OpenBSD ports at 3.31.1.

For "cheaper" jobs and looking purely at output, ent
(https://www.fourmilab.ch/random/) is in most ports trees.

There are more qualified people on this list who should reply...

g