Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Mon, 28 May 2012 01:24:11 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: mschapv2 conversion

On Sun, May 27, 2012 at 11:08:53PM +0200, magnum wrote:
> OK, that seems to work. But as long as it breaks the normal builds, I
> can't commit it. Could you try to get it work for at least linux-x86-sse2?
> 
> This might be just an arch.h issue though. Solar?

No, it'll be a bit more work.  The extra function added to DES_bs_b.c
will also need to be added to the assembly files.  I might get to that
in June.

> Speeds before:
> Benchmarking: MSCHAPv2 C/R MD4 DES [mschapv2]... DONE
> Many salts:	2304K c/s real, 2317K c/s virtual
> Only one salt:	1638K c/s real, 1638K c/s virtual
> 
> After:
> Benchmarking: MSCHAPv2 C/R MD4 DES [mschapv2]... DONE
> Many salts:	4861K c/s real, 4861K c/s virtual
> Only one salt:	2466K c/s real, 2466K c/s virtual

Looks good to me given that MD4 has remained the same (not vectorized)
and that this is a generic build.  One of the next steps (for this
format) is to switch to using sse-intrinsics.c code for MD4.

Alexander

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.