If the performance of Prescotts goes down, what does that do to AMD’s PR system?
It seems to me that AMD would be perfectly justified in yanking it upward, too.
If Intel devalues the MHz again, and a 4GHz Prescott only equals, say, a 3.5GHz Northwood, it would seem quite justified.
It would create a few problems, though.
At least initially, most PIVs will still be Northwoods, so for a while, you’ll have A64s beating Prescotts at a certain speed, but not Northwoods.
Then again, Intel rather has that problem too, doesn’t it?
You have a problem with identification. If today’s 3400+ becomes tomorrow’s 3700+, how do you sell the old 3400+s?
On the other hand, there are problems if AMD does nothing, problems much more likely to get AMD’s attention.
AMD has tied its desktop Hammer pricing to the price of the “equivalent” PIV. Once the Prescott PIV gets to 3.8GHz or so, Intel could turn the pricing barrels on AMD for down at the 3.2GHz level or so.
That costs AMD money, and AMD probably won’t like that.
What Will AMD Do?
AMD is supposed to introduce socket 939 and probably its last 130nm Hammers at the end of March.
It is likely that at that time, AMD will also hop up the FSB (equivalent) from 200MHz to 250MHz.
What AMD will probably do is hop up the PR at the same time, and included in that increase will be an adjustment for the Prescott devaluation. They’ve done things like that with XPs before, which is a good reason to think they’ll do it again.
They’ll then shift rather rapidly over to 250MHz CPUs.
This will avoid most of the transition problems mentioned above.
Don’t Get Fooled Again
This will be the second time Intel has effectively devalued the MHz. The first time was with the introduction of the PIV. That created the gap between an Intel “MHz” and an AMD “MHz.”
It proved to be a very slick move on Intel’s part, because when AMD reintroduced PR to adjust for the devalued Intel MHz, many blamed, not Intel for devaluing the MHz, but AMD for making the necessary adjustment.
Many thought MHz represented some absolute standard of performance, that a MHz is a MHz is a MHz and that’s simply flat-out wrong, no more so than RPM tells you how powerful an engine is.
Unfortunately, many out there, including far too many who ought to know better, believe that, which is why AMD has to do this. Not to sneak one over on people, but to prevent one from being snuck over on people who “know” better.
No, on this one, AMD will be right. They should make an adjustment, and anyone who doesn’t think so or expresses some form of disapproval simply doesn’t know what he or she is talking about.
Of course, AMD could adjust too much, but that’s a different story. Up to now, at worst they’ve bent a few numbers a bit, but on the whole, they’ve put out legit PR numbers.