TH: Where do we need to go next? I mean, from everything you’re describing and I’ve seen, it seems like we’re “there,” with all of the performance and power anyone could need. So what do we do next?
TT: For a handheld device, there are two things that are always going to be very pertinent. The first is how small you can make the device. Can we make it smaller than an iPhone? The second thing is going to be whether you can make it even higher power. Batteries increase in capacity by about 7% to 10%—10 % is an outlier, so really 7%—year to year in terms of density. Some of the bigger phones are going to have instead of 1,000 mAh, maybe 1,500 mAh batteries. So the pressure to get power lower is always going to be there. Also, I’d add a third challenge: how can we continue to drive higher capabilities and functionality in this class of device? We’re going to continue to push on those three angles as hard as possible.
TH: We’ve already profiled how Atom has evolved from its first appearance to Linfield. How do you see that evolution continuing in the future?
TT: Obviously, we will continue the architecture. We clearly want to go push the CPU performance higher, but getting higher performance at lower power is going to probably be the biggest thing we’re going to focus on. But also remember our 2x performance increase in graphics, the addition of full HD video recording, and dual-camera capabilities. If you look at those trends, it’s going to be very important to make sure that they continue.
TH: Speaking of graphics, Intel obviously has not yet made a name for itself in performance graphics. Now, we’re suddenly talking about powerful graphics and video from an integrated Intel chip, which is essentially an oxymoron in the desktop world. So can you tell us more about exactly how Intel is ramping up this graphics capability and doing this acceleration?
TT: We’ve added graphics capability that includes both vertex and floating point, as well as the rendering capabilities increasing. But you have to really look at system-level performance. The key thing there is whether we provide enough bandwidth capability to the graphics accelerator. Because if you don’t, most of these things will get choked right where you need the bandwidth. Balancing system throughput is key. Look at our memory subsystem and design internally. We put in a lot of energy in there to make sure we have a very effective bandwidth, not just for the CPU but also for the accelerator. Unfortunately, I can’t go into more details than that at this point.
TH: If these are the PCs of tomorrow...well, we’ve seen netbooks start to cannibalize notebooks. I’ve got to believe that these ultramobile devices, if they really are PC-like as you say, will take a pretty sizable chunk out of your existing mobile product base—only now at lower price points. Is that a problem or are you anticipating that volume is going to make up the lost ground?
TT: If that were the case, yes, it would be a problem. But that is not the case. Netbooks have not cannibalized notebooks.* They have added to top line and bottom line growth for us. If I go to the phone space, we don’t play there today, right? With Moorestown and Medfield and beyond, the intent is to become a player in this space because these devices are coming to where Intel is strong. So, I don’t buy the premise of your question. This all growth for us.
* Ed.: Apparently, prior sales analyses are open to interpretation. See http://www.reuters.com/article/idINB97620320090527?rpc=44.