On May 21, David Brownlee wrote:
> 	I've tried running the rc5des from www.distributed.net on a 233Mhz
> 	DNARD system, among others, and the numbers were:
> 
> 	Machine		CPU     MHz	OS	rc5 Kkeys/sec
> 	Sun IPX		sparc	 40	1.3.1	 30
> 	Digital Hinote	486/DX	 50	1.3.1	 49
> 	Dell		Pentium	133	1.3.1	189
>         Thinkpad 560	Pentium	133	1.3.1	189
> 	PC Clone	PentMMX	200	1.3.1	260
> 	DNARD(arm32)	SA-110  233	1.3	289
> 
> 	These are pure CPU numbers (for example the IPX is faster using
> 	X than the Hinote, which this does not show), and the SA-110 does
> 	not have any hardware FP, but in the CATS board I'd expect it to
> 	be none to shabby (modulus any specific NetBSD/arm32 issues).

  It's more than that...Using the rc5 stuff as any sort of benchmark
is a very bad idea.  In day-to-day use, as you've stated above, an IPX
blows the doors off a 486, yet the 486 gets better rc5 rates.  Here's
an excerpt from distributed.net's rc5 faq that explains why:

> Why are Intel and PowerPC computers so much faster than other
> platforms?
> 
>   Integral to the mathematics of the RC5 algorithm are 32-bit rotate
> operations. For whatever reason, the designers of the x86 and the
> PowerPC architectures decided to implement the rotate function as a
> hardware instruction. Many other CPUs do not have built-in hardware
> rotate instructions and must emulate the operation by (at the very
> least) two shifts and a logical OR. This handicap is why many
> non-Intel and non-PowerPC computers run RC5 slower than one might
> expect based on real-world benchmarks. It is also the main reason why
> the RC5 client is a poor benchmark to use in determining the speed or
> performance of a particular CPU.


                               Regards,
                               -Dave McGuire
                                mcguire@neurotica.com