[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20150916194741.GA12716@openwall.com>
Date: Wed, 16 Sep 2015 22:47:41 +0300
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: fast hash processing bottlenecks
On Wed, Sep 16, 2015 at 09:37:54PM +0200, magnum wrote:
> On 2015-09-16 21:31, Solar Designer wrote:
> >A "rules first" mode would help greatly (apply each rule to current
> >word, then advance to next word; right now, we do it the other way
> >around, which works better for probability-optimized rulesets and slow
> >hashes).
>
> Our current "prince + rules" implementation does that (it has to) and
> it's *very* slow - while prince without rules is faster than wordlist
> mode. Did I implement it in a bad way?
That's puzzling. Without looking, I'd guess that you did. There's no
reason I can think of why "rules first" would run slower than "words
first". It should improve cache hit rate, with few other differences.
There might be something specific to how the PRINCE and rules
implementations interact, though.
> First or not, I think the rules engine has some low hanging fruit for
> you now that you are up to speed ;-)
I'm afraid I already removed all very low hanging fruit in 2009-2010,
when Jim challenged me to it. Also, it's only temporarily that I am up
to speed: I got many non-JtR tasks I need to switch to about now.
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.
