Let's talk about Mobile CPUs

There's interesting stuff happening in mobility these days, so I thought I'd rattle off a few thoughts.

AMD quietly announced their Turion brand at CES. Intel put $2B of marketing dollars into Centrino and AMD combats with a nearly silent launch at CES of nothing more than a name?

When I first heard about Turion I thought - well, maybe AMD had spun off a design team to work on a mobile core from the ground up. However, given their announcement strategy and the fact that all they are saying about Turion is that it's based on the "AMD64" architecture, I get the feeling that it's little more than a power binned, low voltage, 90nm Athlon 64. I'm hoping there's more to it than just that, but at this point I don't have very high expectations.

If Turion does end up being just a rebranded Athlon 64, I can't help but feel that AMD will be competing with Centrino on little more than a marketing basis. The architecture of Dothan, from a mobile standpoint, has yet to see a rival from any of its competition - and it's going to take a lot more than a new brand to be competitive.

On the flip side, AMD does have a much more power consumption friendly basis to derive a mobile chip from than Intel did with the Pentium 4. It is possible, if AMD had been working on it for the past 3 years, for Turion to be a separate rev of the K8 core with a handful of mobile specific optimizations. In terms of pipeline depth I'd expect that K8 and Banias/Dothan are quite similar and with its on-die memory controller AMD could, in theory, do some pretty interesting power management that would offer lower power consumption compared to an external north bridge. This is all very speculative and may be a bit of wishful thinking, but I guess we'll find out soon enough.

I'm working on the delayed desktop Pentium M article right now, which is partially responsible for the subject of today's post. From my discussions with motherboard manufacturers it sounds like they are all interested in making desktop Pentium M motherboards, but no one is actually going to commit to it other than AOpen and DFI, who are both currently shipping. Shuttle is toying with the idea as well, but I'm afraid that outside of the AOpen/DFI solutions, Dothan's life will be over before the rest of the motherboard market gives it a chance on the desktop. That may or may not be a good thing (as you'll see from the article), but it may be that we have to wait for Yonah/Jonah (65nm dual core Dothan successor) for another chance at Pentium M on the desktop.

Because of some HTML mixup everyone seems to think that Apple is on the verge of releasing a PowerBook G5. Looking at the power consumption of IBM's PowerPC 970FX, it is quite possible for Apple to do a 1.8GHz G5 in a notebook, especially given how aggressively the CPU will scale itself down when idle or not fully tasked. I wouldn't be too surprised to see Apple offer a G5 based PowerBook sooner rather than later simply because it seems that their own marketing has cast a heavy shadow over the G4 processor, despite the fact that the G4 is a very capable contender - especially at its higher clock speed versions. I think more than offering a G5 based PowerBook, I'd like to see Apple try their hands at an ultraportable (or something at least a little lighter than the 12" PowerBook).

That's all for now, more ramblings as they pop into my head.