[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 22 Nov 2013 11:17:01 -0600
From: Richard Miles <richard.k.miles@...glemail.com>
To: john-users@...ts.openwall.com
Subject: Re: Questions and suggestions to build a home cracking
box. :)
Hi Alexander,
First of all, thanks for your answer, really helpful.
On Wed, Nov 20, 2013 at 9:37 PM, Solar Designer <solar@...nwall.com> wrote:
> On Wed, Nov 20, 2013 at 08:24:26AM -0600, Richard Miles wrote:
> > I tried OMP before but it was not very stable, sometimes it worked,
> > sometimes not, so I would like to avoid it, except if it's stable now. :)
>
> This is news to me, unless you were on Windows and used other than one
> of our own builds. Can you describe those stability issues, please? -
> including what version/build of JtR you were using and the OS you were on.
>
> OpenMP is usually not the most efficient way to parallelize JtR, but it
> is stable, with the only known exception being that Cygwin shortcoming
> (patched in the version of cygwin1.dll that we're distributing with JtR).
>
I tried it ~ 8 months ago with the version available at Magnum repository.
I don't have the VM anymore to check exact version, I apologize.
But the issue was weird, sometimes OMP appeared to work, sometimes not,
very strange. Sometimes I had to reboot the VM a few times and worked,
sometimes not. That's what I gave up of it and called "unstable".
But for sure I will test it again with this new box that I'm planning to
build. :)
>
> > Between OMP and Fork what is faster?
>
> In terms of c/s rate, --fork is faster. For slow hashes, --fork is a
> bit faster. For fast hashes, --fork is a lot faster.
>
> In terms of passwords cracked per second, it depends. OpenMP uses just
> one stream of candidate passwords, which JtR tries to keep in an optimal
> order. --fork uses multiple streams, so the order in which candidate
> passwords are tested may be (slightly) less optimal. In fact, in some
> cases some of the child processes may terminate significantly sooner
> than others.
>
>
Hummm... just to have a better idea. What you mean by significantly? Are we
talking about minutes or hours?
> Overall, I recommend --fork, but if its drawbacks start to annoy you and
> you're attacking a slow hash type, you may use OpenMP for a more
> pleasant experience (other than on Windows with buggy Cygwin), at the
> expense of slightly lower c/s rate (e.g., on a certain machine md5crypt
> with OpenMP is 215k c/s, with --fork 225k c/s combined for 8 processes).
>
Interesting, great information, thanks.
Does it affect both markov and incremental? What about wordlists attacks
with and without rules? The bottleneck is the HD speed?
>
> > Both support all password hashes
> > available on your github (https://github.com/magnumripper/JohnTheRipper
> )?
>
> No. OpenMP support is included for a subset of hash types only. --fork
> is supported for all (in bleeding-jumbo branch).
>
More one point for fork :)
Thanks a lot.
>
> Alexander
>
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
