[RC5] Key Rate Question
Peter A. DeNitto
denitto at alde.com
Thu May 6 16:17:18 EDT 1999
What's failed to be mentioned here is that the K6's implementation of ROTL
is not as effecient as a PII.
The problem is that the K6 does not have this function hard coded into the
CPU, in comparison to its little brother, the K5, or the Pentium or
PowerPC cpus. It is rather "wired" to run the ROTate Left function thru 4
functions that do the same together as one ROTL.
It takes more cycles to do this on a K6 than a ROTL on a Pentium.
ROTL is done 4 times per encryption round, and is about 1/3 of all
Therefore, the K6 tends to be slower because of this lack of
This is also true for True RISC CPUs (Alpha/Sparc/Mips), and all their
ROTL's are done thru software computations.
On Wed, 5 May 1999, Sean Porterfield wrote:
> Date: Wed, 05 May 1999 21:34:54 -0400
> From: Sean Porterfield <sporte at esn.net>
> Reply-To: rc5 at lists.distributed.net
> To: rc5 at lists.distributed.net
> Subject: Re: [RC5] Key Rate Question
> > Why are the Cyrix and AMD chips faster than my Intel Pentium?
> > [... snip ...]
> > But my experience has been the opposite. Anyone see similar results?
> > I have a AMD K62-450 running Solaris 7 that keys 750,000 keys/sec. I
> > expected something over a million.
> I have an AMD K6-2/350 running Linux (Red Hat 5.1) that gets around 590Kk/s
> which is comparable to my Intel PII-266 running Win95.That indicates to me
> that the Intel chip is faster.(Although when I compile, it seems faster on
> Get your website today from http://www.porterfield.net/
> Please visit a worthwhile site:
> To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
> rc5-digest subscribers replace rc5 with rc5-digest
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest
More information about the rc5