Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 29 Apr 2013 16:12:29 -0700
From: Myrice Li <>
Subject: Re: Re:password generation on GPU

Hi Sayantan,

On Sun, Apr 28, 2013 at 9:49 AM, Sayantan Datta <> wrote:
> Hi Myrice,
> On Wed, Apr 24, 2013 at 2:50 AM, Myrice Li <> wrote:
>> For more feasible password generation, I think we should start with
>> mask-mode first. In mask-mode branch, you can find mask.c, this is the
>> very initial implementation. You could improve it also. (Compare to
>> other modes, I think I miss save/restore ability and make the mode
>> more powerful and the code nicer :) )
> I went through the code in mask.c in mask-mode branch of your repo. It is
> the cpu version which we need to implement on GPU, am I  right ? I also saw
> the md5 kernel in PG-test branch. I didn't understand much of  the code in
> mask.c , probably because I don't know exactly what the mask mode does. Do
> you know anything(maybe a tutorial) which could help me getting started ?

You may refer to to
know how mask mode works. I only implement subset of it. Yes, it is
the CPU version. We may or not implement it on GPU. However, it helps
the password generation on GPU. For example, if we have pass?d?d, we
may generate pass[0-9] on CPU and let GPU generate pass[0-9][0-9]. In
my last year experiment, we should not let a GPU kernel run for a long
time to avoid freezing screen. It depends on implementation. You may
search for last year's email with subject: Mask Mode.

The mask.c should be in Mask-mode not PG-test. md5_kernel generate
password by appending [a-z0-9][a-z0-9] for a given password. It also
mixes with bitmap and the code need to be cleaned also...


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.