[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20100119155106.GA7564@openwall.com>
Date: Tue, 19 Jan 2010 18:51:06 +0300
From: Solar Designer <solar@...nwall.com>
To: john-users@...ts.openwall.com
Subject: Re: JtR 1.7.4 and jumbo patch update
Matt,
On Fri, Jan 15, 2010 at 12:17:35AM -0500, Charles Weir wrote:
> I applied the john-1.7.4-jumbo-5.diff patch to the base 1.7.4 install,
[...]
> words: 10495949352 time: 0:00:39:31 100.00% (ETA: Thu Jan 14 23:02:37
> 2010) w/s: 4426K current: 9zzzzzzzthi$
>
> Analysis: It appears that the bug has been fixed. In addition, it ran
> significantly faster this time, (shaving a full 10 minutes off the
> running time). While I wasn't running any additional programs during
> any of the tests, I don't know if this was due to my computer, the
> changes made in the update, or applying the jumbo patch vs. the base
> install. Regardless, it's now almost twice as fast as version 1.7.3.4
> when comparing guess generation speed.
This extra speedup (vs. clean 1.7.4) should be due to JimF's loading of
the entire wordlist into the process address space (as opposed to
reading it from the operating system for each rule separately, which is
what clean JtR does). This code is currently in the jumbo patch.
I intend to reimplement it in an official manner in a new development
version. My changes between 1.7.3.4 and 1.7.4 were orthogonal to those
made in the jumbo and JimF patches, such that even greater speedup is
achieved when both kinds of changes are implemented at once - which is
what you're seeing.
After 1.7.4, I have started to make changes that are similar in spirit
(and sometimes in code) to those previously implemented in the patches.
I have included some of those into 1.7.4.2. So there's little speedup
to be expected with 1.7.4.2 (and likely with some further 1.7.4.x
development versions) when you're applying the jumbo patch anyway, but
the code becomes "official", which means many things: availability (to
those who are not using the patch), maintenance (now as part of JtR),
quality, completeness, portability (some of these were relevant to the
changes made so far, some I expect to be relevant to future changes).
> http://reusablesec.blogspot.com
You have some curious and relevant stuff in your blog. I think you
should be posting about some of your analyses in here.
Thanks,
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.
