[RC5] SGI weird performance

Phil Gregory pgreg430 at neors.cat.cc.md.us
Fri Jun 5 00:17:01 EDT 1998


-----BEGIN PGP SIGNED MESSAGE-----

> A Penitum II 300 Mhz gives around 800 000 k/sec
> 
> A Silicon Graphics Origin 2000 with four MIPS processor gives around 250 000 k/sec
> 
> A Silicon Graphics O2 R5000 gives around 140 000 k/sec

I know (well, I'm pretty sure, at least) that I've seen this in the FAQ.
The difference is that the CPUs have different instruction sets.  The x86
CPUs have a specific instruction (I think it's a 'rotate left' or
something like that) that many other CPUs don't.  Thus, the reason that
x86s do much better in the RC5 contests is not that they're faster, but
that they do the decryption more efficiently.  (But when we hit DES, watch
all of the Alpha teams scream to the top of the stats!)

                                ---------------
#define BITCOUNT(x)	(((BX_(x) + (BX_(x) >> 4)) & 0x0f0f0f0f) % 255)
#define BX_(x)		((x) - (((x) >> 1) & 0x77777777)		\
			     - (((x) >> 2) & 0x33333333)		\
			     - (((x) >> 3) & 0x11111111))

/* Counts the number of bits in a word. */
                                ---------------
                                 Phil Gregory
                                     DNRC
                              phil_g at geocities.com
                        -------------------------------
                         (E-mail me for my public key)


-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQCVAwUBNXdjMalF2m7Yx1z1AQENUwP9FW1zENFnzZTUmsBEMWcSovrjrv5d5Ehb
tYKfHqsNjPiRinP2LtpUPeA0DrDAK0yqC26/Luhqfmw5DlrzPu54zHl1g5abSfXs
9LoFj0OPouMq+f813i3j29XPPyzz/3L6J+2TUF7cO3+RCrXL5cAxPBScgzsrSl+5
LxRSNDZvF6w=
=eZkN
-----END PGP SIGNATURE-----
--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list