[RC5] Performance on P-IV

Ryan Malayter rmalayter at bai.org
Thu May 10 14:54:01 EDT 2001


The G4 is by far the most efficient processor out there for RC5 right now in
terms of keys/clock. This is a result of the AltiVec-optimized code. I'm not
close enough to the dnet P4 project to know, but who knows what some great
coder could make of the P4's updated SIMD architecture?

-----Original Message-----
From: RogerKIWI at aol.com [mailto:RogerKIWI at aol.com]
Sent: Thursday, May 10, 2001 11:40 AM
To: rc5 at lists.distributed.net
Subject: Re: [RC5] Performance on P-IV



In a message dated 5/10/01 12:17:54 PM, rmalayter at bai.org writes:

>That said, a P4 can be very fast when code is optimized for it. And RC5
>cracking is a very specialized use of a processor - it is in no way
>indicative of the processor's general performance potential. All of the
>cores in the current "official" dnet client are specifically hand-assembled
>for the P3 or Athalon architectures. Similarly hand-optimized P4 cores
>might make it a faster dnet machine than anything else out there.

Hmmmm my 450 MHz G4 runs just a whisker over 4,000,000 keys/sec per CPU. (I 
have a dual processor, so run just under 8 MegaKeys/sec when I'm not doing 
much else.

That's 8,889 keys per MHz.

>From the original message

>My AMD Athlon 1000 (1 GHz) scores a friggin' absolutely awesome
3,538,971.00
>Hahaha a 1GHz Athlon is almost TWICE as fast as a stinking P-IV 1400 which
>has 400MHz more under ít's belt. 

This means the Athlon under Linux runs 3,539 keys/MHz.

which is what the site http://www1.distributed.net/~fordbr says, showing
that 
the Athlon is by far the most efficient of the x86 type chips, but still not

even managing half of what the G4 PowerPC chip can do with optimised code.

plink plink

Roger
--
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 mailing list