Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20120629010233.GA26645@openwall.com>
Date: Fri, 29 Jun 2012 05:02:33 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: sha512crypt-opencl and sha512crypt-cuda test

On Thu, Jun 28, 2012 at 09:57:57PM -0300, Claudio Andr? wrote:
> Em 28-06-2012 21:47, Solar Designer escreveu:
> >I've just tested sha512crypt-opencl and sha512crypt-cuda on:
> >
> >$6$saltstring$fgNTR89zXnDUV97U5dkWayBBRaB0WIBnu6s4T7T8Tz1SbUyewwiHjho25yWVkph2p18CmUkqXh4aIyjPnxdgl0
> >
> >They crack the password "john" fine, albeit a lot slower than CPU.
> >Hopefully, their speed would improve with a longer run.

> Have you removed the self-test (time)? As magnum sad recently, the 
> "timer" should start after self-tests (it impacts GPU formats).

You're right, we probably have this issue now.  The timer starts after
loading of the hashes, but before the self-test.  Additionally, in
"batch mode" the self-test is re-run when proceeding to next pass
(single crack -> wordlist -> incremental).

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.