[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <401fca91564eacb9fb19cd0f932ff32f@smtp.hushmail.com>
Date: Tue, 19 Mar 2013 03:58:25 +0100
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: Idea to increase plaintext length for GPU based hashes
On 19 Mar, 2013, at 3:05 , Lukas Odzioba <lukas.odzioba@...il.com> wrote:
> 2013/3/19 magnum <john.magnum@...hmail.com>:
>> I've had similar thoughts. For some reason I rejected the idea so far because I did not think it will fly IRL. I'm not sure how I came to that conclusion, we should definitely try it.
>
> Maybe because it could break mem read alignment - at least that's the
> reason that I didn't try.
> Until we try we wouldn't be sure.
We could keep alignment by 4 though. That is probably a good idea for many reasons. With a max length of 64 or more and an actual length of around 10, the savings are still great. The average word length in the rockyou list is 9.75 including the garbage lines.
BTW we could stuff length + offset as 8 + 24 bits, in one int. This shaves 4 bytes per key and still avoids using char.
magnum
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.
