[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20210302193416.GA31634@openwall.com>
Date: Tue, 2 Mar 2021 20:34:16 +0100
From: Solar Designer <solar@...nwall.com>
To: john-users@...ts.openwall.com
Subject: Re: Implementing mixed mask attack
On Tue, Mar 02, 2021 at 08:12:02PM +0100, Micha?? Majchrowicz wrote:
> OK. No problem I have written simple filter that counts how many
> uppercase characters are in candidate and reject those above two. I
> run with command like this:
> john --format=descrypt -1=?l?d?u --min-len=1 --mask=?1?1?1?1?1?1?1?1
> test.pwd --external=myfilter
>
> However instead of gaining anything I got slower speed :( Looks like
FWIW, the reported speed would become lower with any kind of filtering
even if the filtering helps reduce the total running time.
> generation of those candidates (which in such case it's pointless)
> becomes significant operation so still generating all those mask (with
> two uppercase letters for example) and running dozes and dozens of
> sessions is the only way :(
The filtering is reasonably usable for much slower hashes and/or with a
large number of different salts. For descrypt with few different salts,
it's just too much overhead.
It looks like you're sort of trying to do what incremental mode does on
its own - test more likely patterns first. Have you tried simply
letting the default incremental mode run for a long while, and see how
it fares against your masks? For a fair comparison, give both runs the
same set of hashes without anything in the pot file(s) yet. Try this
without splitting it across your many nodes yet, to remove the effects
of that from comparison of the modes' basic functionality.
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.
