Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sun, 12 Aug 2012 20:31:57 +0400
From: Solar Designer <>
Subject: Re: Salted MD5 cracking problems

On Sun, Aug 12, 2012 at 03:56:53PM +0400, Vladimir Vorontsov wrote:
> Currently we do not have any solution to brute MD5(salt.pass) hashes.

How many of these do you need to try cracking?  Is the salt length fixed
(at 8?) or variable?  In fact, is the salt value fixed or do you have
multiple per-hash salts?  What cracking mode(s) would you prefer to use?

> Look forward to an internal (dev) version of the john.

The closest match to what you need is currently myrice's work-in-progress
on "fast" hashes on GPU, where he has code for raw-MD5 with hard-coded
mask for two characters on GPU (with the rest of password provided by
CPU) and with hash comparisons on GPU.  This is not optimized yet, but
it does achieve a little over 2 billion passwords/sec on 7970 when run
against up to a few thousand hashes at once.  (myrice is working on
improving the scalability to avoid the slowdown with higher hash counts,
also testing on 1 million loaded hashes.  In fact, this might be already
done - I haven't tested the latest code yet.)  There's no support for
salts in that code yet, but if you only have one salt value it can be
added easily, e.g. via external filter() in john.conf (this won't affect
the speed much since it'd be out of the loop for last two chars).

So please answer my questions above and we'll see what we can do.  Thanks.


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.