We’ve talked about Intel’s Atom and nVidia’s Tegra (there’s also the Via Nano) processors and why they’re important as precursors to the tinyCPUs running the tinyPCs in the decade ahead.
We’ve also wondered where AMD was in all this.
Unfortunately, if you look at the specs, you’ll see that this isn’t a “new” CPU at all. Compare the specs shown in the link above to this AMD page and you’ll see that this “new” CPU is just a slowed-down, cut-down Sempron. They underclocked the chip, slowed down HT, then cut out a memory channel and the pins supporting it. Otherwise, it’s the same thing.
I suppose that’s better than nothing, and it will probably perform as well or better than the initial Atoms (albeit using more power to do it).
But this processor perfectly illustrates what is wrong with AMD. Yes, this cutdown Sempron is probably the best stop-gap AMD could come up with, but that’s all it is.
It won’t start showing up until well after Atom and Tegra machines make their debut, even though Intel and nVidia did far more work on those processors than AMD will do on this. It looks like AMD got “blindsided” again, even though Intel mentioned Atom last fall. Why did it take AMD nine months to decide and announce, “Let’s slice up a Sempron?” These cut-down Semprons should be around now, not next year.
More importantly, though, tinyCPUs are a whole new ballgame, with entirely different rules. A few extra watts means nothing on the desktop, and not all that much for a full-sized notebooks, but will hurt much more with Eee-class mininotebooks and will be deadly for anything smaller than that. Power saving has to be built into these chips from the ground up, and I’m sorry, but just using SOI isn’t going to be good enough.
You are simply not going to be able to compete for any length of time in this new market if your idea of making a tinyCPU is to cut down a desktop chip. You need an entirely different kind of beast.
You may say that AMD’s competition looks like half-man/half-beast, and I wouldn’t disagree, but at least the transformation has begun, and the involved parties know they have to do a lot more and plan to do so.
In contrast, is Hector and Company even aware that they’re going to need a truly new CPU line to compete in this market past the first flush of products? If they are, can they spend the money necessary to compete?
I greatly doubt either.