Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Sun, 1 Apr 2012 09:43:35 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: GPU acceleration - what to implement?

Hi Michael,

On Sat, Mar 31, 2012 at 07:41:35PM +0100, Michael Pappas wrote:
> Back from easter holidays, and I was wondering on how to start up with 
> contributing to GPU accelerated hashes. Given that I am currently 
> learning OpenCL (no prior experience), I think that a possible path 
> would be the following:
> 
> 1. Keep working on OpenCL
> 2. Study the source code of one of the already implemented hashes
> 3. Implement a hash that John does not support yet
> 4. Improve performance of my implementation / other implementations as I 
> gain further experience with OpenCL / GPGPU optimisation techniques

Sounds fine.

> I am currently looking at opencl_rawsha1_fmt.c. I have found some 
> documentation on format.h but any additional information would be 
> greatly appreciated. I'll probably get back to you with more specific 
> questions though.

OK.  Please note that raw SHA-1 is a "fast" hash, so the implementation
currently in JtR is inefficient.

> Regarding on which protocol to implement, do you have any particular 
> recommendations? I've noticed that John does not support WPA/WPA2, but 
> I'm not sure whether they fit into the GPU aspect.

They do.  In fact, this has been mentioned on this list this month.
I suggest that you start by reading john-dev postings for March to see
what has already been suggested and who is already working on what.
Just because someone has started work on something does not mean you
shouldn't pick the same task, though.  It is entirely possible that the
other person will fail, or that your implementation will be better, or
that optimizations from the two implementations could be merged
together - and even if none of this applies, you'd nevertheless learn
from this experience (including from the comparison to someone else's
approach to the task).

Thanks,

Alexander

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ