[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <BLU0-SMTP165556D897496A3AB815207FD380@phx.gbl>
Date: Sat, 14 Apr 2012 07:57:39 +0200
From: Frank Dittrich <frank_dittrich@...mail.com>
To: john-dev@...ts.openwall.com
Subject: Re: JtR:Multi-GPU Setups
On 04/14/2012 05:48 AM, SAYANTAN DATTA wrote:
> On Sat, Apr 14, 2012 at 7:07 AM, Solar Designer <solar@...nwall.com> wrote:
>
>> For slow hashes, multiple instances of JtR are perfectly capable of
>> fully loading multiple GPUs in the same machine at once. I experimented
>> with that the other day, got over 1.5M c/s combined speed at phpass with
>> incremental mode on my AMD + Nvidia cards.
>>
>
> I have a question here. In this mode do we use two gpus to crack a single
> hash or two instances john is run on both the GPUs.
Since Alexander mentioned "multiple instances of JtR", I am pretty sure
he did start JtR twice. If you do that, you manually make sure that
there is no (or very little) overlapping in what these two instances do.
Either you use two password files (the original passwords split by
salts, so that no salt appears in both input files).
For slow hashes any many password hashes with different salts, the
overhead of generating the same password candidate twice can be neglected.
Alternatively, you try different cracking modes (e.g., the first
instance runs incremental mode for length 1-7, the second one
incremental mode for length 8).
Frank
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.
