A lot of people, myself included, had high hopes for Bobcat. It's the fabled Atom-killer, the next challenger to stand and face the titan of the ultra mobile market in a long line of would-be usurpers. Reading up on the architecture, it seemed to be a pretty robust design for its market, owing more to VIA's Nano than anything AMD had previously. This was not a shrunken, slimmed-down version of the K8 architecture, like the old AMD might have done (and did). The new AMD is more nimble, aggressive, and cunning, and doesn't recycle old CPUs in shiny new packaging. They have less to worry about in terms of fabricating their chips, so now they focus more on their design, and where their products fit in the strategic view of things.
ARM has yet to live up to the netbook promises they've made in the past, likely scaring off OEMs with their lack of major OS support. Netbooks are just a smidge too large to satisfy people with a small linux-based OS such as Android, and so ARM processors will likely find better success in the newly invigorated tablet market where Apple currently holds reign. So we're back to x86 solutions, where only two hopefuls reside alongside Intel.
AnandTech recently did a preview for the new dual-core VIA Nano, a refresh hoping to rekindle interest in the mobile architecture that failed to gain much ground in its first incarnation. Nano held a lot of promise, particularly in its early years when Atom was still very new and the market was perhaps a little more open to alternatives. But now Atom has slowly refined into a real battery life monster, getting to the level of some of the better CULV systems out there and distancing itself from similarly positioned architectures. The problem is Nano is still built on the now downright ancient 65nm process, and much like everything else VIA does, a die shrink will likely be too little too late. The dual-core Nano simply hits an entirely different power envelope, higher even than CULV systems, and with a merely adequate IGP platform, it just isn't going to turn a lot of heads in the ultra mobile field, less so than even the single-core variant (which at least found itself in a reasonably-built Samsung system). I feel for VIA, because I'd like them to succeed in a market I feel they have the best chance at, but unfortunately as big as Intel and AMD are, they're simply more agile and capable than VIA could ever hope of being.
So that leaves Bobcat. For the first time, AMD actually has a die size advantage at least when compared to CULV platforms. AMD knows it's the mobile market that their past architectures have been hurting the most. Desktops don't care much for TDP considerations so AMD can simply price to match as they see necessary. In the mobile market, that shit don't fly, and merely adequate to mediocre performance coupled with pretty lousy battery life mean it doesn't have a toehold anymore as far as notebooks are concerned. But try to remember, that was the old AMD, and starting next year they're gonna try their best to make you forget that image and focus on how they're doing things now. They've got more specialized CPUs and platforms designed specifically to address those markets, thanks to the long-in-waiting ace up their sleeve in the form of their new GPU branch (must...not...say...."ATI"...). Bobcat is the first time they've gone so far as to draw up an entirely new design separate from any of their desktop lines exclusively for mobile platforms, and this not long after AMD's CEO poo-pooed the whole netbook category.
Bobcat's performance generally outpaces the top-of-the-line Atom by a fair margin, as expected. Obviously the real bread and butter is the equal-parts GPU component of the chip, which trounces anything within sight of the Mississippi. What's concerning, however, is how close they are when the Atom compared to the particular Bobcat model that was tested boasts about half the TDP. This Bobcat SKU, called the E-350, is meant more for entry CULV platforms, where it fares less well against, and to get to a level competitive with the Atom power envelope, you have to knock down the core clock by 600MHz, or 37.5%, and cut the GPU clock nearly in half, which gets you the C-50. The performance picture might get a lot fuzzier at that point, and even more unclear is how well systems will do on the whole compared to some of the most popular Atom netbooks. Will there actually be a substantial uptake of OEM solutions made by popular brands such as Asus, Acer, HP, and Samsung? Or will we get one or two experimental machines and then back to business as usual with the Atom offerings? My hope is that these companies take Bobcat seriously, and that it gets the treatment it deserves in systems that have the quality and battery life you'd expect from a real Atom killer. One thing's for sure, Atom, and Intel that makes it, needs a nice firm slap on the behind to get moving so we actually get some decent performance out of the netbook category. It won't be til the middle of next year that Intel will finally supply a fully integrated HD video crunching Atom platform, and the advances won't get much more elaborate after that, according to their roadmaps. Bobcat will be ready-built with HD media in mind, and aggressive performance ramping going forward with future iterations, so it really all depends on the reception it receives in the market come release time.
All I know is my 1000HE is getting a little long in the tooth.
