[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Mon, 28 Jun 2010 13:33:21 +0400
From: Solar Designer <solar@...nwall.com>
To: john-users@...ts.openwall.com
Subject: Re: faster DES on Atom
A minor correction:
On Mon, Jun 28, 2010 at 01:15:55PM +0400, Solar Designer wrote:
> ... when working on this code in 2006 I only encountered CPUs that
> executed both SSE and SSE2 versions of the code at the same speed (AMD
> CPUs) and those that executed the SSE2 code much faster (Intel CPUs),
> which is why the decision to go with SSE2 only was made.
I think I managed to recall the specific CPU classes mentioned in braces
above incorrectly. Plain SSE only performed well on Intel Pentium 4
(and on P4'ish Xeon and Celeron). Both AMD and non-P4 Intel CPUs ran
SSE2 faster than plain SSE (speaking of bitwise ops only, indeed).
Alexander
Powered by blists - more mailing lists
Powered by Openwall GNU/*/Linux -
Powered by OpenVZ
