Here We Go Again

AMD is beginning to hype its 45nm chips.

You’d think they would have learned their lesson last year, and indeed they did learn something from last year, though the wrong lesson.

Last year, they spit out what at least looked to be specifics, and took a heavy hit when they failed to meet them.

This time around, they’re trying to hype the product without saying anything important they could be held accountable for.

They say things like the 45nm will consume less power and have better performance. Wow, that tells us a lot! Less power and better performance than what? How much less power, how much more performance?

Oh, they do say that the core frequencies of these chips are “expected” to “greatly increase.” Well, “expect” can cover a lot of ground. You can expect to get a paycheck after work; you can expect the Tooth Fairy to show up. The difference between the two is that you have a lot more reason to expect the first than the second, and your expectations are a lot more specific.

Let’s not even dwell on what “greatly increase” means. Given that the K10 bar is currently 2.3GHz, that’s a pretty low bar to jump over.

You might say, “It’s too early for AMD to get specific about what its 45nm chips can do.” There’s probably something to be said for that, but if that’s the case, why are they hyping something they don’t have? Have they forgotten last year already?

Nor is it all that early for AMD to have an idea as to what these chips can do. They’re supposed to start making these things by the end of June or so; they’re not leaving themselves much time between first silicon and production silicon. Maybe early February is too early, but by late April, they ought to have a pretty good idea what they can do, and if they don’t tell us what that is by then, that ought to set off warning bells in your head.

You might also say, “We don’t have specifics on Nehalem, either.” This is true, but does anyone doubt we’ll see specifics by around Computex in June, just as we did for C2D and Penryn, and if we don’t, there’s likely to be something amiss?

The point is not to berate AMD for not having complete benchmarks out now, but to alert you that these and all other AMD comments on their 45nm chips have been a cluster of meaningless fluff. Not too big a deal yet, but if we keep getting more of the same in April and May and June, that’s a very strong indication that it’s going to be 2007 all over again.

Don’t get fooled again.

Ed


Be the first to comment

Leave a Reply