[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CA+TsHUC60W8GStRKULMV6v48nex+nZxKqwXgRGyqbDzcoRZ_kQ@mail.gmail.com>
Date: Thu, 14 Jun 2012 09:50:34 +0530
From: SAYANTAN DATTA <std2048@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: bf-opencl
Hi all,
On Thu, Jun 14, 2012 at 1:24 AM, magnum <john.magnum@...hmail.com> wrote:
> On 2012-06-13 20:17, SAYANTAN DATTA wrote:
>
>> BTW is it possible that a format passes self test but fails in actual
>> cracking?
>>
>
> This is what the Test Suite is for. BF is present in it, but you'll have
> to crate a line for bf-opencl in jtrts.dat (using the CPU line as a
> template). In case the GPU format has shorter max length than the CPU
> format, the numbers in the end (1500) should be lowered accordingly.
>
> Jim, is there any good documentation on the jtrts.dat format? I use to
> guess from existing lines and use trial'n'error in order to create new
> entries.
>
> magnum
>
>
I think we should integrate the TS with magnum-jumbo or improve the
self-test so that no error is passed undetected. Once you have modified the
code , it becomes really difficult to identify the true source of error. If
the error is detected in early phase then it is much easier to debug and
identify the source of error. Now I find that one of my very early version
of code is cracking correctly. So I would have restart from there.
Content of type "text/html" skipped
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.
