[CentOS] OT: AMD CPU actual speed....

Sat Apr 14 02:03:45 UTC 2007
Peter Arremann <loony at loonybin.org>

On Friday 13 April 2007, John Summerfield wrote:
> Peter Arremann wrote:
> > 4200 is a performance rating... Originally they modeled after a Pentium
> > III, later switched to comparing to Duron performance. With dual core
> > everything got even harder to determine - but important thing is that
> > 4200 is just a relative rating and doesn't mean 4.2Ghz.
>
> Pentium IV, not III I thought. No point comparing with Duron, it's one
> of AMD's own processors. When Intel redid its implementation, AMD kept
> on extrapolating from the older Pentium IVs so as to continue using the
> same measuring stick (and getting bigger numbers might be nice too).

Glad that thread is already marked as off topic :-) AMD first used PR numbers 
for the K5 comparing it with the P54C Pentiums... They then re-based for the 
K6. And the K7. And then for the Athlon64s... The Athlon XP was unofficially 
compared to the P4 numbers... And somewhere in the middle they compared XP 
performance against their earlier Thunderbird Athlon models as well. And then 
AMD wondered why their quantispeed nonsense never really became popular with 
consumers :) All that is up on wikipedia and many other sites... 

Anyway, this is what I remember cause a friend of mine was working at intel in 
europe at that time. AMD ran into issues in some European countries where 
competitive commercials are not allowed. That means you can't advertise your 
product as being 20% faster as something from another company. Although I 
can't remember AMD ever saying they compared to the Pentium 4 - they for a 
while had the statement saying that their performance index was a relative 
speed measure comparing to a Duron... I found a few emails in german pointing 
to that but nothing official...

Peter.