Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 15 Jan 2007 10:32:53 -0800
From: "Alain Espinosa" <alainesp@...il.com>
To: john-users@...ts.openwall.com
Subject: Re: NTLM patch performance

On 1/13/07, Solar Designer <solar@...nwall.com> wrote:
> You've made sure that not a single hash got cracked during those 10
> minutes, correct?

Correct

> > Program                          1 hash                 10 hash
> > ----------------------------------------------------------------------------------
> > john_v1.7 nt_patch            1244K                  1181K
>
> The number for 10 hashes is JtR's reported effective c/s divided by 10,
> correct?

Correct

> > ppa_v1.70                        4370K                  4600K
> > saminside_v2.5.7.1         11360K                  5255K
>
> ...and what are the numbers for Simon's MMX/SSE2-enhanced NTLM support
> patch?  I'd expect something like 5000K.

benchmark: 4800K

In fact i make some improve in the NT patch (witch C code) that benchmark 5100K.
I send this code to Simon. This means that sse2 code its 4 times
slower that C code and dont help much. Searching in "Intel(r)
Pentium(r) 4 Processor Optimization" find that mmx and sse2 code
execute in parallel. We can make this:

...........................
one sse2 intruction
one mmx instruction
one sse2 intruction
one mmx instruction
...............................
and we can compute 6 words in parallel.

alain

-- 
To unsubscribe, e-mail john-users-unsubscribe@...ts.openwall.com and reply
to the automated confirmation request that will be sent to you.

Powered by blists - more mailing lists

Your e-mail address:

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