Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 08 Mar 2012 14:34:52 +0100
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: OpenCL KPC and LWS

On 03/08/2012 01:07 PM, I wrote:
> From bench.c the benchmarking is made just knowing function pointers to
> init(), set_salt(), set_key() and crypt_all(). The only difference here
> is we want to modulate lws and kpc.
> 
> For the lws enumeration I think we just need a pointer to crypt_kernel.
> 
> For kpc you do a lot of specific things - but these are just mimicing
> crypt_all(). Why not actually use crypt_all(num) instead? This (and
> using set_key() and set_salt() instead of memsets) will ensure we
> actually measure the same way as the real deal.
> 
> I will experiment a little with this.

OK, the problem with this approach is we can't use queue profiling. I
have this working using other means but I'm not sure how good it is.

magnum

Powered by blists - more mailing lists

Your e-mail address:

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