Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 14 Oct 2006 10:03:48 -0400
From:      Lowell Gilbert <lgusenet@be-well.ilk.org>
To:        soralx@cydem.org
Cc:        freebsd-hardware@freebsd.org
Subject:   Re: Quiet computer
Message-ID:  <44wt738057.fsf@be-well.ilk.org>
In-Reply-To: <200610140308.00451.soralx@cydem.org> (soralx@cydem.org's message of "Sat, 14 Oct 2006 03:07:59 -0700")
References:  <78ED28FACE63744386D68D8A9D1CF5D4209C94@MAIL.corp.lumeta.com> <20061012075101.Y5008@ketralnis.com> <200610140308.00451.soralx@cydem.org>

next in thread | previous in thread | raw e-mail | index | archive | help
soralx@cydem.org writes:

>> > mentioned cpu, but a 100x speed difference? That doesn't seem realistic
>> > to me (although if those are valid results, I'd be pretty happy with
>> > that)...
>> 
>> Well, the Via Padlock has a hardware random-number generator, so the
>> idea was to test that. It doesn't claim to be fast, just to be truly random 
>
> Precisely. However, speed of the crypto engine should be directly
> proportional to a peak speed of the RNG,

That statement makes no sense to me.  Why would the RNG be relevant
after the session keys are established?

>                                          so I thought that it's
> slow speed might confirm that the engine is being used, and is
> simply slow.

I don't really care how fast the crypto engine is on my Via system.  I
just care that it offloads the ALU.  I haven't gotten around to
proving whether (and by how much) it does so.

> Also, we didn't really test where are those random bits coming from :P
> Is the TRNG used by default, or some tinkering's in order to make it work?
>
> BTW... `ubench`? :)

Not impressive.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?44wt738057.fsf>