The partnership with Google, said Bell, also makes Intel a “first class citizen” in the Android ecosystem, meaning that as new iterations of Android emerge, the corresponding Intel optimized versions will be released simultaneously.
“There shouldn’t be a time lag between the availability of the various flavors of Android. We will be doing all the enablement work, so our partners can almost have a seamless experience getting their products to market,” explained Bell.
To further help push its chips out into the mobile market, Intel is going an extra mile by building its own fully featured reference design smartphone to offer customers interested in becoming an Intel mobile partner. Partners can either adapt that reference design or use it as is, said Bell.
“It’s almost doing it a disservice, calling it a reference design,” Bell noted, saying the phone had fully passed Google’s compliance tests and was undergoing carrier certification. “For all intents and purposes, you could sell this as a phone tomorrow,” Bell noted.
The 10mm phone does indeed boast every bell and whistle available on the market, from NFC to HDMI, to support for full 1080p video in a chassis just 10mm thick. “This is a no excuses smartphone,” said Bell, showing off the phone in a demo.
The reference design shown by Intel at CES sported an 8mp camera, but Bell said the chipset could support up to 16mp cameras, as well as a bevy of other features and a battery capable of all-day use. Similarly, the demo model was using Android Gingerbread, because it was slightly more optimized than Google’s newer Ice Cream Sandwich at the moment.
“It’s up to the vendor, up to our partners, to pick what parts go into their design. We’ll help them with that. But we’re not dictating what people have to ship,” he said, noting that Intel had no interest in controlling the specs of the phones their chips were eventually used in. “We think people will largely adopt what we’ve done, but it’s really up to them,” he added.
At the end of the day, said Bell, Intel’s foray into smartphones would be successful based on positive customer experience. “We believe that the user experience is really what it’s all about,” he said, positing that the only way Intel could really differentiate its product in the market was to delivers omething better or different than its competition.
“Different is not just a spinny, 3D globe UI or some sort of tacky eye candy,” he noted, “it’s something that fundamentally makes the user experience better.”
Intel, he said was investing heavily to help its partners come up with that differentiation, building on the already powerful Android ecosystem.
“The nice thing about Android is that the ecosystem already exists and it’s an open platform, we have the full source code. There’s no reason why we can’t add capabilities to our flavor of it and maintain Google compatibility at the same time,” he said.
Indeed, much of the work that went into Intel’s failed MeeGo experiment is now purportedly being moved over to what’s being called Tizen, the firm’s latest open source effort.
“There’s no reason we can’t take the best pieces of that and integrate it in with our Android offerings,” Bell added.
Despite the clear, streamlined vision, however, it will take more than a reference design and a partnership with Lenovo in China to make Intel a success in mobile.
Analysts at CES were quick to note it was still early days for the chip giant in a market it has been attempting to penetrate for some 20 years. Mobile World Congress in Barcelona, many said, would be the true test of Intel’s success in persuading partners it was a serious player, and with that show just over a month away, we may not have long to wait to see if Bell and his team can deliver on their vision.
The reference design phone reminds me of the way Intel releases motherboards for their desktop processors. Could this be the start of the PCification of the smartphone? May we be able to one day swap cases, boards, displays (betteries) with the same ease we do now with the PC.
He said angry birds was an ARM native application. ARMulators have been around for a while for x86, I wonder how they're doing the ARM simulation. The best silicon process, dragged down by a legacy x86 design, dragged further by emulating the competition, and yet it's still competitive. Remarkable, but there's got to a be a simpler way.
The main reason people need real computers these days is for media and information creation. It's been said many times in the past and the recent improvements in smart phones and tablets have not changed anything. Have you tried rendering video, modifying a weekend's worth of pics or even doing a term paper on a tablet? It doesn't make sense. We all enjoy our smartphones and are amazed how much they do. But the gap in capability and practicality is still enormous.
I'm not so sure this is a fight that Intel can afford to step away from.
Much as I would enjoy ARM using the same sort of FUD Intel used to market the 386 against the 68K. "Will you risk not using the industry standard".
Superphones and Tablets are eating PC sales. My niece does all her Facebook on her iPhone, she doesn't have a PC. Superphone + Internet TV, why have a PC?
Most ARMs are going into featurephones, the Cortex-A7 is fine for that and operates at 10% of the power of this new Atom. So ARM's core business is safe. I can't see the upcoming 'tock' of architectural redesign of Atom dropping power by more than half.
Intel's still enjoying good profits, I suspect that's at the expense of AMD. They haven't landed a blow on ARM yet.
I think the ultimate shortcoming that Intel in their ARM war is that their whole business is tuned to high margin chips. Will they be able to change gears and tackle the low-margin business effectively?
Intel also is one chipmaker. ARM, on the other hand, is a whole ecosystem of hundreds of different parts from scores of different manufacturers with different capabilities.
ARM's single core chip (stripped down to achieve power numbers) is going up against ARM parts which are going the other way - multi-core chips with DSP and SIMD processing on-chip.
Sure Intel have deep pockets and some clever people on board, but they have a lot of challenges if they want to play this game.
I predict that Intel will do what they've always done with their non-core business (8051, i960, ...) - do something interesting for a bit and then just walk away leaving their customers high and dry.
The problem with that recommendation is that sooner or later the CPU business itself will shrink to mostly mobile CPUs, as they get enough performance to run major apps. ie. Mobile CPU will disrupt the entire CPU business. So Intel is forced to play now.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.