Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20150917152345.GA16626@openwall.com>
Date: Thu, 17 Sep 2015 18:23:45 +0300
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: prince mode and rules

On Thu, Sep 17, 2015 at 03:31:26PM +0200, magnum wrote:
> Ok, this depends a lot of what rules are used. The 8M p/s was the 
> default wordlist rules. Best64 is a lot faster:
> 
> $ ../run/john -stdout -prince -max-run:10 -v:1 -ru:best64
> 220957380p 0:00:00:10 0.00% 21493Kp/s aquame22me123

I guess much of the speed difference comes from our rules trying to
avoid producing duplicates, whereas best64 doesn't care.  Our approach
is good for slow hashes or when there are many salts, and best64's is
good for fast hashes.

Maybe we need an option that, when set, would optimize out the various
reject commands from the rules.  And we could set it by default when
cracking a fast unsalted hash.

Unfortunately, if we reuse the exact same approach that we use for
squeezing out no-ops, we'd have performance impact for when the reject
commands are left intact.  That's because those same code paths would
now check the pass number (optimization or actual cracking).  Maybe we
need to introduce a separate loop/switch for the optimization pass.

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.