Conor said:
And? It's been a common naming system used by AMD since the 90's where
it refers to the performance equivalence rather than the clock speed.
Clock speed hasn't been a reliable indicator of performance for years.
Roughly speaking:
Performance = Clock_Speed x Instructions_Per_Clock_Cycle
Since AMD had a larger Instructions_Per_Clock_Cycle, by a factor
of about 1.5x or so, they were always losing out in clock speed
comparisons. So AMD made up their "P.R. rating" system, to counteract
that trend. The P.R rating, is a way of quoting the "effective"
speed of the processor (i.e. when compared to Intel P4), and
helps take into account the superior Instructions_Per_Clock_Cycle.
Now that Core2 Duo processors exist, they have the very same problem.
The new Intel design has also managed to boost Instructions_Per_Clock_Cycle.
A new Core2 Duo at 2GHz, is roughly equivalent in integer performance,
to maybe a 3.2-3.4GHz Pentium 4 processor. So you cannot compare
the clock speeds of families of processor directly, and need some
way to measure the parallelism (Instructions_Per_Clock_Cycle)
as well.
I find the best way, is to view various benchmarks, to see the
apparent advantage.
Paul