[RC5] oddball DES vs. CPU speeds

linas at linas.org linas at linas.org
Wed Jan 14 12:55:26 EST 1998


Hi,

I am seeing some bizarre cracking speeds with the new DES-II cores;
makes me wonder if there are bugs?  My Cyrix-133 which was much
faster than my pentium-133 on RC5-64 is much slower on DES.  And
my PPro-200, which was faster than both pentiums combined, is only
a little bit faster now.

Specifically,  after doing 1 block (of 2^30 keys) of DES, I get:

Pentium-133 -- 611kkeys/sec
Cyrix-133    -- 478 kkeys/sec
PPro-200    -- 851 kkeys/sec

This doesn't make sense -- the PPro isn't even running 200/133 times
faster than the pentium!!!

Compare this to the rc564 rates I used to observe:
Pentium-133 -- 168 kkeys/sec
Cyrix-133    -- 247 kkeys/sec
PPro-200   -- 517 kkeys/sec

The OS is Linux,  the processor types were manually selected, the client
was rc5des381-linux-x86-mt.tar.gz

-----------------
BTW, 
No, its not "the DES algorithm".  The Pentium Pro has a dramatically 
different CPU architecture than the Penitum.  It should be 3-4 times
faster for almost everything.  It was 3-4 times faster when runnning
the DESCHAL code, so the Bovine results don't make technical sense.

--
Linas Vepstas   -- linas at linas.org.spam.stopper -- http://linas.org/

--
To unsubcribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list