SAN FRANCISCO--Intel Corp. may have missed the smartphone boat the first time around, but the chip maker is determined not to let the opportunity set sail without it again, making its first big moves in ultra-mobility at last week’s CES 2012.
Showing off a reference design smartphone and announcing partnerships with both Motorola Mobility Inc. and Lenovo Ltd, Intel is determined to make its mark on mobile this coming year, starting with the Chinese market and then expanding out.
Spearheading Intel’s efforts in the space is Mike Bell, a former Palm executive who also spent time at Apple working on the first iteration of the iPhone. Bell’s official title is vice president and GM of Ultra Mobility, a role formed after the previous head of Intel’s Ultra Mobility segment, Anand Chandrasekher left the firm last year.
Bell hopes to succeed where Chandrasekher failed by honing Intel’s efforts with a singular focus on Google’s Android operating system, a strategy he believes will catapult the chipmaker forward and even provide it with an edge on its previously untouchable competition – the ARM ecosystem.
Choosing Android as Intel’s platform of choice for smartphones was a decision made owing to the sheer number of people in the ecosystem working on Android enablement, said Bell, noting that he believed Intel had the hardware to make Android “really shine.”
Bell acknowledged that in the past Intel had focused on making chips more geared towards speed than power efficiency, but said this had been a choice rather than anything more fundamental.
Medfield, he said, would change all that, with the 32-nm SoC making huge strides in power efficiency, with Bell claiming it was as power efficient as any of its competition.
“We have chips that are just as power efficient as everyone else. At the same time, they’re faster,” he said adding that from a performance standpoint, Intel was really pushing the envelope.
“On standby time we’re well within shouting distance of best in class, on some of the web benchmarks we smoke the competition, in some cases we’re two, three, four times faster on some benchmarks. There’s some we don’t do as well on but honestly, with a Medfield based phone in your hand, the user experience is phenomenal. There’s no downside, there’s only upside,” he said.
In order to succeed in smartphones this time around, Intel has seemingly gone all out in its effort to throw significant weight behind Android as a platform. Back in September the firm announced it had officially partnered with Google to create a highly optimized port of Android to its x86 chips, and Bell said Intel had also let loose its group of software engineers on the Android ecosystem to help developers make their apps run more seamlessly on Intel’s platform.
“We’ve found people whose apps work fine, but in our testing we’ve found ways to make them faster or better, so we’re helping to raise all the boats in the Android ecosystem by enabling those people to have a better product,” he said.
Bell went on to note that Intel would be shipping out technology that would even allow applications developed for other hardware run on Intel’s x86 platform without modification.
“The user shouldn’t have to care what the app was written for. For the most part, it will just run,” he said, demoing a non-optimized, but smooth running version of Angry Birds to prove his point.
When Intel is claiming 2-5x performance, what are they comparing with? ARM7 or Cortex-A15? There are few fundamental reasons I don't believe Intel is going to fly there:
1. ARM is encumbrance in mobile world and there is no fundamental incentive from Intel processor to convince cell manufacturer to replace it;
2. Intel margin is high and fab is expensive. That is why Xscale division never made money before sold to Marvell.
3. Comparing many ARM SoC choices from different vendors such as TI, Qualcomm, Marvell, etc, no one wants to stuck into a single vendor x86 solution;
4. It might be the power that help ARM to win initial position in Mobile world. But there are many other reasons help ARM hold its helm.
5. At last, has Intel had good tracking record to score major design wins outside their PC/desktop world.
My recommendation to Intel: Don't be obsessed with Mobile world. You don't have to control everything to do with CPU. Focus on your core business before ARM could steal from it.
It seems that the giant has suddenly waked after a very sound sleep in the mobile processors segment. But it is better late then never. Android required a strong interfaceablity with a stable, up-gradable and consistent hardware platform, that only a dedicated processor manufacturer can provide as compared to IP vendors.
Apple, Samsung and other design their own mobile processors and few more chips. When other mobile vendor gets them from Intel with their hefty profit margin, their product may becomes costy. Intel will have difficult time in this price sensitve market.
Don't know about that, although possibly it would be a good business model.
The reason I'm skeptical is, the vertical integration of Apple is exactly why I prefer Wintel and Android. I'd hate to see Intel becoming a closed shop too.
Well, Mike Bell did use to work for Apple... and he's in charge of Ultra-Mobility now. I think Intel is starting to get fed up of waiting for others to adopt their chips for phones... and this reference design is almost at the point where Intel has made one itself and is basically delivering it on a silver platter to anyone who will sell it. That's a pretty sweet deal for OEMs... Intel makes all the design investment and they just have to slap their name on it and sell it? Intel couldn't be making it much easier!
This is all extremely interesting, IMO.
So, we have seen speculations that future PCs may run ARM instead of x86 architectures, and now it looks like Intel is working hard and evidently succeeding to support Android over x86 architecture, without the power penalty.
Things are evolving in the most interesting ways. The possibility of Win8 running over ARM or x86, and Android running over ARM or x86, and x86 no longer being a non-started for power-challenged devices, can only be fun to follow.
The article pretty much quotes the significant parts of the video clip.
I think the most significant point being, any apps written to run over Android should run over an Android-Intel system as easily as they run over Android-ARM. That's what you'd expect of an OS, after all.
But Bell says that Android over Intel will be faster, and that they've invested a lot of resources to improve power efficiency, to the point that battery life should last all day.
Good for Intel, is my comment. Go for it.
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.