Megahertz don't matter!

[HolyGrenade]

<A HREF="http://www.bea.doc.gov/bea/papers/mpuvqa.pdf" target="_new">http://www.bea.doc.gov/bea/papers/mpuvqa.pdf</A>
Its a big pdf with scary equations that supposedly shows megahertz independent ways of marking CPUs.

Heres an easier read from the inq:

THE US DEPARTMENT OF LABOR is telling us what we all knew for a long time but both Intel and AMD denied for years.
Megahertz don't matter.

An article on the Producer Price Indexes pages says that the Megahertz rating is "more confusing than enlightening" and appears to support AMD by saying that one chip running at a particular MHz rating can easily outperform other chips based at the same MHz.

And it calls for objective benchmarketing to steer your way through the maze. Well, we'd all like that, wouldn't we.

It points out different architectures give different results anyway, making the process harder, and underlines its conclusions by saying that other elements of a PC have vastly incresaed in performance anyway.

The article begins to lose us a bit when it starts talking about "hedonic modelling" and introducting very scarely looking equations like the following, but hey, we're just hack journalists here.

<A HREF="http://www.theinquirer.net/14060202.htm" target="_new">http://www.theinquirer.net/14060202.htm</A>

<b><font color=blue>Hand of God is a distant memory as the foot of <font color=red>Beckham</font color=red> strikes.</font color=blue></b>

 

[cellbiogeek]

MHz don't matter...unless a processor with a lower MHz rating performs faster than a processor with a higher MHz rating. Then, I would suppose they matter.

 

[Guest]

in that case PR rating is noway. i don't think so.

 

[dttdar]

Does that mean that Intel should create some type of rating system for their P4's. For instance, their 2.0 ghz performs close the the Athlon XP 2000 which is 1.73. So should Intels 2ghz P4 be called something like a P4 1700+? If that's the case, then wouldn't the XP rating have to change?

It's all quite confusing if you ask me.

----------------
sorry, but I'm still stuck in the Pentium III world

 

[FatBurger]

Uhh...why exactly is the Department of Labor doing all this?

<i>cellbiogeek says:</i>
MHz don't matter...unless a processor with a lower MHz rating performs faster than a processor with a higher MHz rating. Then, I would suppose they matter.

Was that supposed to make sense?

<font color=blue>Hi mom!</font color=blue>

 

[HolyGrenade]

I ask the same question.


<b><font color=blue>Hand of God is a distant memory as the foot of <font color=red>Beckham</font color=red> strikes.</font color=blue></b>

 

[cellbiogeek]

Doh! This board needs a "you're bassackwards" check for dislexics like me (as well as a grammar check and spell check). What I was intending to get across is that, though clockspeed alone is not the only factor contributing to a processors performance, the fact that Intel can ramp up so high in clock speed, has allowed them to produce CPUs that perform better than any released AMD processor. So one can't make the statement, MHz don't matter. They're just not the only thing which matters (which everyone knows but the statement just doesn't sound very contaversial)

 

[siliconjon]

MHz is a very relative term. It only applies to CPU's of the same architecture when you use it for performance comparison.



I'm not really human, I just play one on TV

 

[Crashman]

Yes, but all else being equal, the higher frequency will always perform better. If you want to argue, read the first half of that sentence.

What's the frequency, Kenneth?

 

[cellbiogeek]

Oh, I agree fully. That's what I was trying to say with my first statement except I'm backwards and worded it completely in reverse.

 

[HolyGrenade]

bingo!

<b><font color=blue>Hand of God is a distant memory as the foot of <font color=red>Beckham</font color=red> strikes.</font color=blue></b>

 

[zengeos]

Ahh but that's the issue now isn't it. All else is NOT equal...

Mark-

<font color=blue>When all else fails, throw your computer out the window!!!</font color=blue>