X

Intel has ARM in its crosshairs

Chip giant isn't content dominating the PC and server markets. It wants to make gains in mobile phones, where competition is stiffer.

Tom Krazit Former Staff writer, CNET News
Tom Krazit writes about the ever-expanding world of Google, as the most prominent company on the Internet defends its search juggernaut while expanding into nearly anything it thinks possible. He has previously written about Apple, the traditional PC industry, and chip companies. E-mail Tom.
Tom Krazit
8 min read
The world's largest chipmaker didn't make a whole lot of news at the latest edition of IDF, but it did send a message to the legion of chipmakers that build chips based on the ARM architecture: We're coming for you.
A correction was made to this story. Read below for details.

That didn't exactly send shivers down the spine of executives at Texas Instruments, Qualcomm, Samsung, STMicroelectronics and others that build chips for mobile phones. They've seen this coming for a long time, an inevitable consequence of Intel finding itself with reams of chipmaking capacity and a maturing PC market. And Intel has already tried this once, spending billions trying to develop a combination of chips for the cell phone market but failing miserably.

Following Intel's show in San Francisco last week, ARM developers will be meeting next week in Santa Clara, Calif.--Intel's hometown--for its annual developers' conference to discuss new applications and techniques for extracting more performance out of ARM's processor cores. The collective effort of both camps should do wonders to jump-start a market for mobile devices built for real people, not just coffee-toting executives rushing through O'Hare trying to get the 7:42 flight to San Francisco.

Intel's definitely the challenger, not the favorite, when handicapping the looming showdown between the two companies. ARM is very well-known inside the tech industry, although its lack of an "ARM Inside" marketing campaign means that few mobile phone owners know much about the 1,700-employee design company based in Cambridge, England.

But there's an ARM core chip of some type or another inside almost every mobile phone on the planet, and in many cases, there are a couple. Chip companies like TI and Samsung license cores from ARM, and then build the processor and assemble the rest of the chips needed for modern mobile phones and their more powerful cousins, smart phones.

We must have a comparable power envelope.
--Gadi Singer, assistant
general manager,
Intel Ultra Mobility Group

Despite its earlier struggles, Intel thinks it can be the company that vastly improves the mobile Internet experience by creating chips for a whole new category of devices. We're talking about mobile devices that are more powerful and nimble than current smart phones, and sleeker and with better battery life than today's mobile minitablets.

This is where the next battleground will be fought in the chip industry. ARM and its partners have a long history of working in environments where battery life is often the most important consideration. They also have the benefit of having gotten there first. Intel, however, brings a ton of experience developing products for advanced computing, and has challenged itself to redouble its efforts to harness power consumption.

A few years ago, Intel tried to get into mobile phones with a product called Manitoba. That project was an attempt to waltz into the mobile phone industry by integrating one of its XScale applications processor (based on ARM's instruction set, coincidentally) a communications processor, and some flash memory into a single package. It was a flop. Intel was never able to convince phone makers to take a risk on its products, and it sold the business to Marvell last year for $600 million.

However, the trends that made Intel take interest in the mobile market aren't going away. PC sales growth is slowing in the U.S. and Western Europe (PC vendors shipped 239 million units during 2006, according to Gartner, up 9.5 percent from 2005), and while PCs are still hot in emerging markets, it's only a matter of time before growth there settles in at a respectable 10 percent clip.

Smart phones, on the other hand, are taking off. Mobile phones as a whole already sell more than a billion units a year, and Gartner thinks smart-phone shipments (defined as phones that can run sophisticated operating systems and access the Internet) are set to grow 52 percent from 2007 to 2008, from 102 million units in 2007 to 156 million by the end of next year.

So after failing to dial into the phone market, Intel now wants to invent an entirely new product in anticipation of a mobile computing world. The concept--for now--is called the MID (mobile Internet device), and was once known as the UMPC (ultramobile PC).

Next year, Intel plans to introduce a chip called Silverthorne that uses the x86 instruction set found in every PC chip from Intel and Advanced Micro Devices. The design goal for Silverthorne is to consume 10 times less power than the first Banias Pentium M chip Intel introduced in 2003.

"We must have a comparable power envelope," said Gadi Singer, assistant general manager of Intel's Ultra Mobility Group. Many ARM chips run comfortably while consuming less than a watt of power, and that's the goal for Silverthorne. Beyond Silverthorne, Intel wants to cut idle power consumption by another factor of 10 with Moorestown, a future ultra-low power project introduced last week that CEO Paul Otellini said would arrive by 2010.

As its silicon design teams pursue these power consumption goals, Intel's busy selling the merits of the x86 instruction set as the blueprint for future mobile computing. The idea is that there are a lot of application developers with considerable experience in the PC-based x86 world just itching for a chance to get into different types of devices. By extension, x86 developers are in the best position to improve the software experience on today's mobile phones, or so the pitch goes.

Intel is determined to drive that point home. Last Wednesday, during a speech delivered by Intel's Anand Chandrasekher, the company's marketing team fired the first salvo against ARM. Intel attempted to draw a direct comparison between the two architectures by counting the numbers of errors encountered while browsing the Internet on notebook PCs with Intel's chips, versus mobile phones and Internet tablets with ARM's processor cores.

The implication was that ARM-based devices aren't suitable for reading anything beyond stripped-down mobile Web pages because of compatibility problems, and that anyone thinking about buying a mobile device or developing one should choose an Intel-based device to receive the best possible experience.

Intel refused multiple requests to publicly identify the ARM-based devices it used in making its comparisons, citing the need to avoid embarrassing the very partners it might be courting with chips like Silverthorne. Walter Grayeski, who works in the competitive marketing segment of Intel's Ultra Mobility Group, said the company chose phones that were marketed as smart phones, not the basic phones that you can get for free with a two-year subscription.

"Who will decide this will be the consumers. We have to provide the infrastructure and the products, and if we both don't do it, the (industry) won't grow."
--Ian Drew, vice president of segment marketing, ARM

Many of the browsers on those phones weren't necessarily designed to work with sites designed for PC browsers, and Intel thinks that's part of the reason why it will one day have an advantage. "It's fragmented," Grayeski said, referring to the number of players in ARM's world and their competing interests. Not all of the mobile browsers on ARM phones support Web standards like Flash, which automatically excludes a fair number of Web pages from a smart phone's browser.

There's no doubt as to the ubiquity of x86 software. But ARM thinks the more appropriate question is whether x86 software is simply too large for a power-constrained environment like a smart phone.

"The memory footprint, for a start, dictates what runs on some of these devices. Do you need a hard drive in there to go run an operating system?" said Ian Drew, vice president of segment marketing for ARM. "More memory consumes more power."

Intel's stance is understandable, because the company is just now turning its attention to the ultra-low power market and it has to sell the advantages it already has, said Will Strauss, an analyst with Forward Concepts.

"ARM's way ahead in low-power processors," Strauss said. "The PC (industry) says, 'we can bring x86 to the market,' but do you really need the x86 architecture for this wonderful new thing?"

ARM is aware that it has work to do to head off Intel's challenge, especially as more and more applications are accessed via the Internet, not on the device itself. Another ARM executive, Kerry McGuire, director of strategic alliances for mobile computing, noted that users of Nokia's N800 Internet tablet can download Firefox to run on ARM. But that's a much more powerful device compared with the breadth of what is available running ARM's instruction set.

"I think you'll always want a better browsing experience, and I think plug-ins become more important in this space," ARM's Drew said. "That's where I think the software initiatives should be."

And that's one area where Intel thinks it has an advantage, given the powerful software already available for Windows and Linux PCs. Linux, especially, will be an important part of Intel's strategy.

The first UMPCs were designed in cooperation with Microsoft, with plans to run Windows XP and Windows Vista. But since the Spring IDF in Beijing, Intel has been gradually backing away from its long-time partner to tout Linux-based operating systems for MIDs.

Last week, it invited Canonical CEO Mark Shuttleworth on stage during Chandrasekher's keynote to discuss implementing Canonical's Ubuntu Linux in future MIDs. Chandrasekher later said that Microsoft remains an important partner for Intel in the mobile space.

"But we're getting clear requests (for Linux) from customers for a variety of reasons--lower power, and smaller footprint is another," Chandrasekher said. And by extension, there are a lot of Linux enthusiasts who would love to get their hands on an open x86 mobile computer like an MID.

A solid argument can be made on behalf of either company. Intel, with a lot of money and plenty of smart engineers looking for The Next Big Thing, isn't a company to be taken lightly. ARM has the home field advantage--and the help of some heavyweights who would love to keep Intel out of one of the fastest-growing markets on the planet. So who has the inside edge?

"I won't decide that, and nor will Intel," Drew said. "Who will decide this will be the consumers. We have to provide the infrastructure and the products, and if we both don't do it, the (industry) won't grow."

 

Correction: This story incorrectly stated the site of ARM's headquarters. It is based in Cambridge, England.