A Matter of Faith

Even those who should know better are still thinking “MHz = MHz.” –Ed

Intel is going to wipe the floor with AMD.

I’m looking at my emails, and I’m in shock.

A majority of you (IF the emails are representative) can’t let go of MHz. You just can’t let go. You can recite
all the relevant facts, but they don’t matter, deep down inside. This is your scoreboard, and no one is going to take it away
from you. Period. End of discussion.

When I say “scoreboard,” I literally mean “scoreboard,” as this person put it:

“Reverting to the PR rating is openly acknowledging that Intel “beat them”,
at least it is in my mind.”

This is a pretty common sentiment.

There seems to be a common feeling that AMD is “changing the rules of the game.”

They promised that they would be the first to break the 1 GHz barrier. And they did. And they
rejoiced. Now that Intel has surpassed the 2Ghz barrier, they are crying
FOUL. . . .

I . . . agree that they have raped the Mhz rating. But I feel that AMD is crying FOUL because they are being beaten in the
Mhz war that they created. That is not right.

This particular person happens to be an Intel fan, but the same sentiment is also being expressed by those not so inclined.
It doesn’t seem to register deep down inside that this is a response to Intel “changing the rules of the game.”

Or you have people just calling this a “sales gimmick.” It doesn’t register that Intel’s doing the same thing (especially matching a high performing PIV with an SDRAM motherboard).

Nor does it seem to matter at all what AMD would use as a basis for its PR equivalent. ANYTHING besides MHz is a lie
and a fraud. The concept that you can devalue Mhz (which Intel has effectively done) just does not sink in.

I’m using terms like “deep down inside” and “sink in” for a reason. Even those people who recite all the right facts and say they understand still think
there’s something wrong somewhere in this equation. It’s like they don’t really believe what they’re saying.

This is not an intellectual reaction. It’s visceral; it’s the gut feeling that “this can’t be right.”

That Rock You’re Clinging To Has Been Hollowed Out

Since the beginning of the PC era, people have regarded MHz as a rock to cling to.

Most of the time, it has been, but not all of the time.

The reality is MHz is only one half of the formula. All MHz tells you is how often a processor does something. It does not tell you how much gets done each time.

During long stretches of time, x86 processors did do roughly the same amount of work per cycle. When that’s the case, MHz works well enough.

But that is not always the case. On occasion, Intel has “devalued” the Mhz by decreasing the work done per cycle. They did so with the Pentium, they’ve done it more with the current PIV.

This is and of itself is not evil. It’s really a design decision. For instance, the old Alphas basically did the same thing, less work more often. That’s the major reason why they could get 600Mhz out of a .35 micron Alpha while Intel could only get 300Mhz from the Klameth.

On the opposite extreme, PowerPC processors do more work per cycle than “normal” x86s, roughly a third more. The Maclunatics aren’t lunatics for saying “their” MHz are worth more than Intel’s, they’re lunatics for
grossly exaggerating how much more.

So long as you realize that CPU performance means Speed X Work and that the value of “Work” can be changed, this is no big deal.

It’s when you can’t or won’t accept that formula that there’s a problem and you leave yourself open to be played for saps by those who do know the formula and can change the second half of it.

And if Intel can pull the wool at least halfway down most of your more technically sophisticated than the average person’s eyes (or at least make you believe AMD is somehow at fault), what chance does the average person have?

We’ll continue to talk about this over time, because this IS the major computer story over the next few months, one that will be played out millions and millions of times, one person at a time.

But if what I’m seeing is fairly accurate and continues to hold true, AMD is in a heap of trouble.

Email Ed

Be the first to comment

Leave a Reply