For years, Mike Rayfield had a lonely job as the top mobile exec at graphics chipmaker Nvidia. Nvidia has always been a beast in graphics chips for PCs, but its processors for mobile phones haven’t been huge sellers.
But at the recent Consumer Electronics Show in Las Vegas, Rayfield (pictured here) finally got some bragging rights as general manager of Nvidia’s mobile business unit, as it announced a number of customers for its Tegra 2 processors. And in a big industry earthquake, Microsoft showed ARM-based microprocessors — including Nvidia’s Tegra 2 — running a future version of Windows. That version is expected to ship in 2012 and, for the first time, it will run on chips other than Intel-compatible x86 microprocessors.
At the same time, Nvidia unveiled Project Denver, a high-end microprocessor design based on the ARM chip architecture. Project Denver is expected to be Nvidia’s mainstay chip for running Windows and its wedge into the PC microprocessor market. As such, it’s one of the key components that could open up the market for new innovations in blurring the lines between phones, tablets and computers. We talked to Rayfield about the details on the announcement and Nvidia’s push into superphones, tablets, and computers.
VB: Can you talk about Project Denver, how long that has been in the works, and how you are describing it right now?
MR: We are collaborating with ARM on their next generation architecture. We have an architecture license and our objective is to build an extremely high performance CPU (central processing unit), and put it with our GPU (graphics processing unit). We will have a solution that basically goes from personal computers up through supercomputers.
VB: Is that replacing Tegra or will it become part of the Tegra series?
MR: No. The relation is obviously that it uses the ARM instruction set architecture. But in the same announcement we talked about how we licensed A15, which will continue the series for Tegra. So we are going to be ARM-based from the smallest devices to tablets (pictured right) up to supercomputers. We will have a much higher performance offering at the high-end and then a Tegra for these sort of affordable, pocketable computers like super phones and tablets.
VB: And how long was that like in the works?
MR: So we’ve been working on it for years. We’ve been in collaboration with ARM for a shorter time because it uses their future generation chip architecture. We’ve had hundreds of people working on it for a very long time. It’s been rumored to be everything, like an Intel-compatible product. We thought the best thing is to tell people what it was.
VB: At the Microsoft keynote at CES, you demonstrated that you could run the next version of Windows on a Tegra 2?
MR: Yes, that was a Tegra 2.
VB: So any processor from Nvidia should be able to run Windows?
MR: They were running Tegra pretty smoothly with great performance. We can run Windows. We will develop future generations of Tegra and be able to do whatever Microsoft is trying to accomplish with Windows. I would just add one thing there real quick. The applications running on Tegra were more demanding than the ones running on other ARM processors.
VB: How did you work with them to get into the program to make chips to run on Windows?
MR: We’ve worked with Microsoft since the beginning of Nvidia. We’ve got a great relationship. Clearly we understand these PC graphics, workstation graphics, and what’s necessary to be accelerated in the software. We have naturally been partners for a long time. When we do Tegra chips, we think about what has run on the PC in the past. We can very quickly get those kinds of things up and running. I think it showed in terms of the performance you saw in the demos. They highlighted graphics in our demo.
VB: So they used your chip to highlight graphics in the next version of Windows?
MR: I think the work with us is more about, we just worked together for a long time, we’ve grown up and built a PC graphic stuff and we understand this. We have a good working relationship with them in both mobile and Windows.
VB: Will Project Denver be a family of chips?
MR: Well no so Project Denver will sort of in the highest level now will go from Tegra on up.
VB: Tegra represents more of a low-end in the mobile device.
MR: It is a mobile device. It is a portable device. Will the line get blurred at some point in the future? Sure. I take GPU technology from the other side of the PC business and integrate it. There will be a blurring in the middle, but for now, Tegra is in automobiles and mobile devices.
VB: Will you have a multiple operating system strategy for Project Denver? Like the Chrome OS too?
MR: Stay tuned. If you look at personal computers up through supercomputers, you can imagine that this probably does not all run the same operating system. If you run different apps, you care about different things. So as that comes out, we can start to talk about it.
VB: If you had a choice way back when between x86 (Intel-compatible) and ARM, why did you choose ARM?
MR: So clearly we believe ARM has become the most dominant instruction set architecture. It is an open architecture that people can innovate upon. It just makes sense to go off and do this. I wasn’t involved in our business decisions from years ago. But it’s clear that the world is now built around ARM.
VB: And it looks like the legal path so its clear on the ARM side, right?
MR: They are collaborative partner. We have a great relationship, we like working with them, and that is an easy way to go.
VB: You have hundreds of people on Project Denver now. Does that mean you still have a lot more working on graphics chips?
MR: It’s a big team. We will leverage GPU technology. The time frame is a handful of years from now. It is not a set time frame yet. Denver is our processor strategy for the high-end personal computer and up.
VB: Have you said how many Tegra 2 design wins you have?
MR: We have announced the LG 2X super phone, the Motorola Atrix 4G (pictured) and the Droid Bionic, and there were three Honeycomb devices announced. There was the LG tablet, the Verizon-Motorola Xoom tablet and the AT&T tablet. There were also models from Acer, Asus and Dell. We will see more large companies announcing later in the year.
VB: Why is Tegra 2 doing better than Tegra?
MR: Maturity. People observed us on Tegra 1. They didn’t know if we would stick with it. Then we announced Tegra 2 and they said we’re in. These designs are coming out a year from when they got the first prototype.
VB: You are also making progress in cars with Tegra 2?
MR: Audi cars will have them in 2012. In 2012, BMWs will have them too. The Tesla Model S will have it too.
VB: A year ago, Jen-Hsun [Huang, chief executive of Nvidia], said it would be the year of the tablet computer. It turned out to be the year of the iPad. Will this be the year of the Tegra 2?
MR: No one would argue that it wasn’t the year of the iPad. This year will be a great year for Tegra. People figured out where they want to go. We’ll see some amazing Android tablets this year. Honeycomb, 4G, and tablets are all coming together. Apple did a great job convincing people who a tablet is something they want to use. That gave the other computer makers courage to go off and design something interesting. Apple enabled the market to grow a lot quicker. During that time, Android has established itself as an amazing new operating system. It went from a standing start to 300,000 new units activated a day. The new tablets being shown here are a significant departure from the smartphones we’re used to.
VB: It seems like Project Denver and ARM are a good escape path for Nvidia, given the pressure Intel is putting on graphics. And Tegra 2 is gaining a lot of momentum. You don’t have to compete with Apple’s A4 chip, since they aren’t selling that beyond Apple devices. The question is whether 2011 will be the year of the Apple iPad 2, or if there is room for other tablets.
MR: I don’t think they are standing still. I don’t expect them to release anything but very nice devices. We are going to have to just keep working harder and better.
VB: There isn’t much talk about Atom right now.
MR: I see that Intel didn’t talk about it much. There is just not a compelling solution. For four years in a row, they have said they will have something good next year.
VB: Meego [the joint mobile operating system from Intel and Nokia] hasn’t made much progress. Are you in a good position?
MR: It feels good now. We have a lot of work to do. We’ve got the plane in the air. We can make a lot of mistakes that make the plane come down. We have to take what we’ve got and build it.
VB: With Project Denver, you now finally have an answer for all those analysts who said that Intel and Advanced Micro Devices were going to squeeze you out when they put graphics and a CPU on the same chip.
MR: The question is what will we do to be successful in the mobile business. We’ve been getting that question for a long time. We finally have some answers. These smartphones are becoming computers. They are super phones.
VB: How do you define super phones?
MR: They have to be four-inch displays or larger. They have to be able to do Flash animation acceleration. They have to have high-resolution cameras and be very intuitive to operate. They should have great graphics and play great high-definition videos and output them to bigger monitors. You will want to do most of your computing on them. Pretty soon, I’m getting rid of all of my mobile devices and I will use just one.
VB's research team is studying web-personalization... Chime in here, and we’ll share the results.