[RC5] Clocks per Key on Pentium vs PPC604
trif at serv.net
Tue Jan 6 15:36:55 EST 1998
I'm not a PPC expert, but I do know that the rotl instruction on a PPC
is always 1 cycle long, whereas the rotl instruction on x86's is 2
cycles when rotating by the cl register, 1 cycle otherwise. A
substantial number of the rotates are by register, so there's some
of your cycle savings right there.
I know there is an rc5-coders list for those doing overall coding
on clients, but would it be possible to set up a publically available
list for those of us hacking on the publically released cores? This
stuff is probably pretty dry and boring for the majority of the
rc5 mailing list.
On Tue, 6 Jan 1998 kmclaurin at akashic.com wrote:
> Maybe this has already been discussed but I was wondering if there is an
> simple explanation of why the PPC appears to be twice as efficient as the
> Pentium on key processing.
> For example, a Pent133 takes 1503 Secs to process 2^28 keys which results in
> 744 HzSec (?cycles?) per key while a 120MHz PPC takes 694 Secs for 2^28 keys
> resulting in 310 cycles/key. The P133 is running under W95 and the time is
> taken from Wintop. The PPC is a Mac using OS8 with the time taken with all
> other applications shut down (387kk/sec). The Pentium is using a 6401 client
> while the Mac is using a 6402 version.
> I haven't looked at the internal structure of the 604 and Pentium (54 I
> think) for several years but it looks like the 604 PPC has twice as many
> integer units or 'pipelines' as the Pentium?? Or is there something else
> going on like how many clock cycles it takes to execute an instruction as in
> the 2 Tstates for the Motorola cpu's vs 3 to 5 T states for the Intel/Zilog
> cpus (real old stuff like 6800 vs Z80).
To unsubcribe, send 'unsubscribe rc5' to majordomo at llamas.net
rc5-digest subscribers replace rc5 with rc5-digest
More information about the rc5