[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CABob6ipS3U-XWR47ocv9PRiQMZUTn22JeU7DiwVY9QARZUcQvw@mail.gmail.com>
Date: Thu, 25 Aug 2011 09:05:53 +0200
From: Lukas Odzioba <lukas.odzioba@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: Lukas's Status Report - #15 of 15
2011/8/25 Solar Designer <solar@...nwall.com>:
> Lukas -
>
> On Mon, Aug 22, 2011 at 10:02:42PM +0200, Lukas Odzioba wrote:
>> I've uploaded mscash2cuda-0 to wiki.
>
> In addition to my previous message:
>
> I've just tried this on hashes-all.txt-1.mscash2-uncracked from the
> contest - and no hashes are loaded with "--format=mscash2cuda".
> Looking at the code, I see that you need to adjust valid() and add
> split() in order to load hashes formatted in that way as well.
Thanks! I'll take care of that.
>
> The iteration count is fixed at 10240. In -jumbo's mscash2_fmt_plug.c,
> this is now configurable and optionally encoded with the hashes. Maybe
> you need to support that as well, unless doing so would have performance
> impact (quite possibly it would if the loop gets unrolled). This is not
> terribly important as I think we have not actually seen MSCash2 hashes
> with non-default iteration count yet.
No it won't have any performance impact. Main loop is not unrolled,
because code would be too large. I was aware of configurable iteration
count in jumbo but personally I think it's against YAGNI principle,
but I can add it for consistency.
>
> In the Makefile, it'd be nice to have this commented-out line for Yellow
> Dog Linux (in all of your CUDA patches):
>
> #CUDA64PATH = /usr/lib64/cuda32
>
> I think the mscash2 stuff should get into the next revision of allcuda.
Good idea. I'll add path to 32 and 64bit libs in YD.
I agree, that was my intention.
Lukas
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.
