Reports of Intel's Death are Wildly Exaggerated
If you've ever struggled through a skiing lesson, you know exactly what the expression "too far over your skis" means: You're headed in the right direction, but you're leaning so far forward you're going to take a tumble. And that's what's going to happen to my colleague Galen Gruman, who along with other pundits is falling all over himself to bury Intel under the oncoming ARM tsunami.
Yes, as many of us have written, the old PC-centric computing model is running out of gas as tablets and smartphone become more powerful and more central to how we live, work, and play. That's the right direction. But timing is everything -- there's lots of evidence that Wintel (the Microsoft-Intel partnership) still has many years left in it -- which is why Gruman and others are seriously ahead of themselves.
[ Read Galen Gruman's case for why Intel will struggle against ARM given Microsoft's and Apple's support. ]
This is a big topic, so I'm going to stick to one part of it: Intel's ability to compete with ARM in delivering high-performance, low-power chips in packages tailored for mobile devices. One name to keep in mind in this discussion is Zolo, the Android-based smartphone from Lava now selling in India.
"People who have seen it say that it is comparable in performance to [smartphones based on] Qualcomm's Snapdragon and Nvidia's Tegra," the ones used in practically every Android smartphone sold in the United States, says Nathan Brookwood, the principal analyst of Insight64 and a longtime observer of the semiconductor industry. The smartphone's energy efficiency is likewise comparable to these established ARM chips -- except Zolo uses Intel's x86-based Atom chip. That's why counting Intel out, Brookwood says, is a big mistake.
Intel versus Qualcomm and Nvidia: An uneven match
The chip inside Lava's Zolo is based on Intel's new Medfield design. Without getting overly geeky, it's important to note that the chip used in the Zolo is still being produced on the old 32-nanometer process. A year from now, Intel will shift its mobile chips to the same 22-nanometer process used in the just-shipped Ivy Bridge laptop and desktop chips.
Why is that important? The smaller process means far more transistors and far less heat. If Intel is competitive at 32 nanometers, what happens when it shifts to a more aggressive process?
It's also important to remember that Intel owns its own fabs and controls its own process technology, while its competitors generally use foundries owned by companies like TSMC that do not have the same mastery of process technology, notes Brookwood. Don't forget the addition of trigate, or 3D, transistor technology to increase both computational power and efficiency.
That's not to say Nvidia and Qualcomm won't be improving their own ARM chips. Ditto for Apple, which designs its own ARM chips for its iOS devices. Of course they will, Brookwood says. But now that Intel has decided to focus on performance per watt, as opposed to pure computational performance, it's a very different ball game.
A pair of announcements during last September's Intel Developer Forum demonstrated Intel's power in the mobile industry. Intel CEO Paul Otellini and Google's Android development boss Andy Rubin appeared on stage together, and Rubin promised, "We're going to collaborate very closely to make sure that Android is optimized the best it possibly can be for the Intel architecture."
He added, "Going forward, all future releases of Android will be optimized [for Intel]."
Moreover, Intel is one of the largest suppliers of code to the Android code base, says Dean McCarron, principal analyst of Mercury Research. If Intel didn't believe it was going to get something significant out of Android, would it make those contributions just to be a good citizen? Of course not.
So why does Gruman believe that "Intel is kept at the back of the line in terms of support and access to the pseudo-open source Android code"? I don't see any evidence of that, but maybe I'm missing something. Gruman tells me it's the utter lack of Android 4 "Ice Cream Sandwich" devices using Intel chips, though Android 4 first shipped for ARM devices six months ago. Intel says to expect Intel-based Android 4 devices soon and notes the first Android 3.3 "Gingerbread" Intel-based devices shipped last month.
The other key announcement at the Intel Developer Forum came from Motorola Mobility, which signed an extensive "multiyear, multidevice strategic partnership" with Intel to produce Android-powered smartphones and tablets. Motorola wouldn't have done that without looking very closely at Medfield, and it must have concluded that the Intel platform is indeed competitive with ARM in power consumption and raw performance.
Once these devices are in the market, we'll see something new: serious competition for the Qualcomm and Nvidia ARM chips used in most Android devices.
Will you get work done on ARM?
Another exaggerated story line is the threat ARM chips pose to Intel in the productivity market. Sure, Qualcomm showed off a PC-like device earlier this year, but what anyone will be able to use it for is hardly clear.
"I'm quite skeptical," says Mercury Research's McCarron. The software infrastructure is dominated by the classic x86 architecture of Intel and AMD. In particular, Windows 8 on ARM will not support binary translation, so any apps written to run on it will have to be rewritten from the ground up, he says.
Here, too, there is some nuance. McCarron notes that the rise of the app stores will make it easier for independent developers to write and sell ARM-specific apps that could get some traction with the public. Plans on the drawing board for new chips by Qualcomm and Nvidia, expected to debut around 2015, could change the game, he says.
A report by Flurry, a mobile analytics company, found that an astonishing 1.2 billion apps were downloaded between December 25 and December 31, 2011. About 20 percent of those were tablet-specific, estimates Flurry's Peter Farago. However, only 1.4 percent of those downloads were productivity-related, meaning that most people are not relying on their mobile devices as the primary means of getting their work done.
So with all due respect, Galen: Your pronouncement of Intel's death is greatly exaggerated. I hope you're wearing a crash helmet.
This article, "Reports of Intel's death are wildly exaggerated," was originally published by InfoWorld.com. Read more of Bill Snyder's Tech's Bottom Line blog and follow the latest technology business developments at InfoWorld.com. For the latest business technology news, follow InfoWorld.com on Twitter.