Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 1 May 2012 10:25:36 +0530
From: Dhiru Kholia <dhiru.kholia@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: JtR compilation warnings on OS X 10.7.3 with Xcode 4.3.2

On Tue, May 1, 2012 at 5:03 AM, Solar Designer <solar@...nwall.com> wrote:
> On Mon, Apr 30, 2012 at 07:41:47PM +0530, Dhiru Kholia wrote:
> Actually, my question was about strlen(message).  For a long message, a
> single call to SHA1_Update() results in multiple invocations of the
> compression function.  Well, I guess the message was short - which is
> what we usually care about in JtR.

The message was "test message" which is short.
> Wow.  This is more impressive.  What CPU are you on?  Can you try to
> make this change for XSHA512 and benchmark it before/after, including
> with an OpenMP-enabled build?  (We don't currently have OpenMP support
> for raw SHA-512.  By the way, this is fairly easy to correct.)

I am running OS X 10.7.3 in VMware Workstation using Core 2 Duo E6750
CPU. I am not sure if the benchmark numbers can be trusted. (I am
using the relative gain as a sign that CommonCrypto is faster for SHA
family of hash functions than OpenSSL). I will benchmark XSHA512 with
OpenMP today (is CommonCrypto OpenMP friendly?).

$ ../run/john -format=xsha512 -t # before
Benchmarking: Mac OS X 10.7+ salted SHA-512 [64/64]... DONE
Many salts:	1048K c/s real, 1059K c/s virtual
Only one salt:	1027K c/s real, 1047K c/s virtual

$ ../run/john -format=xsha512 -t # after
Benchmarking: Mac OS X 10.7+ salted SHA-512 [64/64]... DONE
Many salts:	1278K c/s real, 1291K c/s virtual
Only one salt:	1244K c/s real, 1244K c/s virtual

Erik,

Can you run some benchmarks on some real hardware before and after my
CommonCrypto changes. Thanks!

-- 
Cheers,
Dhiru

Powered by blists - more mailing lists

Your e-mail address:

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