Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 12 Apr 2011 03:24:24 +0200
From: Ɓukasz Odzioba <lukas.odzioba@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: sha256 format patches

>Right, but your current code will likely bump into another bottleneck
>first.  I think it already does on my system.
You might be right.

>> Slow sha256 calculates 380 c/s - 35kB/s PCI-e transfer

>I think you wrongly substituted the on-CPU speed here (not GPU), but
>that didn't affect your conclusion.
That's my mistke it should be around 2500 c/s as on the bottom of my post.

I also was wrong about 64bit integers in cuda (nv is not boasting of
it). According to this document
http://developer.download.nvidia.com/compute/cuda/3_0/toolkit/docs/ptx_isa_2.0.pdf
- page 64, this should't be a problem.
Bitslice implementation looks very interesting. I'll do further research on it.

oclHashcat results are impressive.
Thanks for *1mod.diff !
Lukas

Powered by blists - more mailing lists

Your e-mail address:

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