This will undoubtedly be a very stupid question, but... here goes.

From what I've read, an AMD cpu like the Athlon XP 2400+ actually runs at a
clockspeed slower than 2.4GHz. However, I also understand that the AMD cpu
mentioned is considered the "equivalent" of an Intel P4 2.4GHz, which runs
at a clockspeed of 2.4GHz. How is this possible?

From what I've read, the reason for this is that the AMD cpu can do more
processes in one clock cycle compared to the Intel cpu. I don't know how
valid that statement is. IMO, if this is true, the AMD cpu gives you more
bang for your buck.

Anyhow, can anyone explain the concept to me?

Thanks
Pearlf

Pearlf wrote:

> This will undoubtedly be a very stupid question, but... here goes.
>
> From what I've read, an AMD cpu like the Athlon XP 2400+ actually runs at a
> clockspeed slower than 2.4GHz. However, I also understand that the AMD cpu
> mentioned is considered the "equivalent" of an Intel P4 2.4GHz, which runs
> at a clockspeed of 2.4GHz. How is this possible?

It does more work per clock tick.


> From what I've read, the reason for this is that the AMD cpu can do more
> processes in one clock cycle compared to the Intel cpu. I don't know how
> valid that statement is. IMO, if this is true, the AMD cpu gives you more
> bang for your buck.
>
> Anyhow, can anyone explain the concept to me?


Different chip architectures can do different amounts of work per clock
tick. In other words, MHz is not a good comparison of speed between
different models of CPU. An Athlon XP can do more work per clock cycle
than an older "Thunderbird" Athlon. That's where the "XP" rating comes
from. An Athlon XP 2400+ is given that rating because it has a
performance level equivalent to a Thunderbird Athlon if it were running
at 2400MHz. (It just so happens that the XP rating is a pretty
close comparison to a P4's MHz rating)

But just because an Athlon does more per clock cycle, that does not
inherently mean that the Athlon gives you more bang for the buck.
(though AMD usually is the more cost effective option). Compare prices
and compare benchmarks for the relevant applications that you'll be
using. Make your decision from there.

-WD

"Will Dormann" <(E-Mail Removed)> wrote in message
news:kMBnb.175032$(E-Mail Removed)...
> Pearlf wrote:
>
<snip>
>
> Different chip architectures can do different amounts of work per clock
> tick. In other words, MHz is not a good comparison of speed between
> different models of CPU. An Athlon XP can do more work per clock cycle
> than an older "Thunderbird" Athlon. That's where the "XP" rating comes
> from. An Athlon XP 2400+ is given that rating because it has a
> performance level equivalent to a Thunderbird Athlon if it were running
> at 2400MHz. (It just so happens that the XP rating is a pretty
> close comparison to a P4's MHz rating)

From benchmarks I've seen at Tom's Hardware, it seems the XP's rating is not
a good comparison to the P4 MHz rating when looking at 3D performance. That
contradicts the "Overall 3D gaming" benchmarks offered on AMD's website, and
the latter is audited by PriceWaterhouseCoopers. I would appreciate it if
you could recommmend some good websites where I can find impartial
benchmarks.

I will mainly use the intended pc for DTP (word processing, photo editing),
internet browsing, writing discs and gaming. IMO, any system with sufficient
RAM and good processor will be able to take care of the main things I intend
to do (DTP), so I focus more on the gaming side.

> But just because an Athlon does more per clock cycle, that does not
> inherently mean that the Athlon gives you more bang for the buck.
> (though AMD usually is the more cost effective option). Compare prices
> and compare benchmarks for the relevant applications that you'll be
> using. Make your decision from there.
>
> -WD

Thanks for the information. Found a white paper on AMD's website addressing
the matter. Would make things much easier if there was some universal index
to benchmark performance.

Pearlf

Pearlf wrote:
> Thanks for the information. Found a white paper on AMD's website addressing
> the matter. Would make things much easier if there was some universal index
> to benchmark performance.


In an ideal world, yes, that'd make it very easy. But unfortunately
some processors excel in some areas, but are not so great in others.
So to give a generic, all-encompasing number rating to a processor isn't
quite practical. For an overly simplified example, say Processor #1
has twice the floating point performance of Processor #2, but the
integer performance is half that of Processor #2. How would you give
it a rating?

At the time that the Athlon XP processors were new, the XP rating
closely matched the Pentium 4 numbers. (And thus the confusion over
what the rating actually meant). I haven't been following too closely,
but lately I think there's a bit of a difference in the numbers.
(giving the advangate to the P4)

First check out prices for the CPU and what you can afford.
SharkyExtreme.com has always had a weekly CPU price list. Then do
some google searches for benchmarks of those CPUs and the particular
applications you plan on using. Tomshardware.com is known for
comparing a broad range of hardware at one time. Their reputation
isn't always squeaky-clean, though. Anandtech.com might be one to
check out.


-WD