phc-discussions - Re: [PHC] Reworked KDF available on github for feedback: NOELKDF

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAOLP8p56tJW+sXXk8DsnpMTVG+cC_x4Hw8q-SCmAM9ZU6dkNOQ@mail.gmail.com>
Date: Sat, 4 Jan 2014 17:17:52 -0500
From: Bill Cox <waywardgeek@...il.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] Reworked KDF available on github for feedback: NOELKDF

On Sat, Jan 4, 2014 at 3:07 PM, Solar Designer Your system time usage is a
lot lower.
>
>
> What OS and version are you running?
>

Manjaro Linux, which is based on Arch.

No, this sounds wrong.  Per that table, it's 16 instructions peak decode
> rate per chip for either chip.  You should compare "Eight Core" FX
> (that's AMD's marketing for the 4-module chips) vs. "Quad Core" i7.
>

I found an AMD FX 8 core CPU at work and benchmarked there.  I almost
exactly reproduced your data.  The difference has to be the machines.  I'm
just not sure what the difference is.


> > [...] I think my server has mood swings, probably related
> > to temperature.  By the time it runs 2 cores that long, it's already
> > scaling back the clock.
>
> Hardly.  In my experience, recent Intel desktop CPUs almost always run
> at max turbo frequency possible for the core count in use, with no
> scaling down, and they definitely can't heat up that much in just a few
> seconds (it'd take minutes) - well, unless you have no heatsink or
> something (but then you'd have bigger problems). ;-)


It may be just the minecraft servers.  It's technically my son's machine,
so I shouldn't complain.


> Thanks for the offer.  Unfortunately, I barely have time to complete our
> own submission by the end of January, so I definitely won't work on
> another one directly at least until then.  Also, it's not as much about
> winning as it is about having better KDFs / hashing schemes among those
> selected as winners, if any.  I am happy to help you improve yours, such
> as by providing advice in this discussion, and your benchmark results,
> etc. may also indirectly help improve escrypt.
>

Thanks for all the help so far.  I can't see how escript doesn't make it to
the next stage.  It just makes too much sense to have an improved version
of what we already know well.  Like I said, escript or better, and I'm a
happy camper.  What we really need is adoption by the various tools like
OpenSSL and TrueCrypt.  In the meantime, I'm having some fun with noelkdf.

Bill

Content of type "text/html" skipped