Twin-track development plan for Intel's expansion into smartphones
Android-optimization is key to beating Apple
Intel is planning a two-pronged attack on the smartphone and tablet markets, with dual Atom lines going down to 14 nanometers and Android providing the special sauce to spur sales.
Mobile and communications group VPs Mike Bell and Hermann Eul laid out Chipzilla's plans for slurping up market share in the sector. Intel has speeded up the process technology shift for this sector and its Atom chips will shift from 32nm today, then 22nm next year. 14nm hardware is scheduled for 2014.
Later this year Intel will release the Atom Z2580 chip, which promises 2x processing and graphics performance within a good power envelope for smartphones and tablets. The next stage will be the 22nm Merrifield chip, which Bell predicted would change the game for Intel in the smartphone market.
"This is a really big deal for us," he said. "It's not just a technology shrink to 22nm, it's a fundamental change. There's a brand new processor core, it has state of the art imaging and graphics and is a new part from the ground up."
Merrifield will ship next year in high-end smartphones and tablets, and will be augmented with some intellectual property gained through acquisition. Bell mentioned imaging patents that allow smartphones to take multiple high-resolution camera shots and power-saving gains that could give 14 days standby time.
But there's also plenty of money to be made in old rope. The 2G and basic handset market is still growing and could be the backbone of mobile telephony for the next ten years, Eul commented, and will be the entrance point for the billions of humans on the planet who currently lack a mobile phone.
To address the more basic market the Z2000 series of Atom processors, which operate at around 1GHz, will ship later this year and be aimed at the low-end handset market, along with 2G and 3G chipsets and HSPA+ connectivity.
In line with the announcement of the close partnership with Google revealed at last year's IDF, Intel is betting the farm on Android to give it leverage in the smartphone and tablet market. Any hopes in Redmond that the old Wintel alliance could be extended to smartphones look forlorn at best, although it's clear Intel is optimizing some tablets for Windows 8 use.
"Our phone efforts right now are concentrated on Android and we have thousands of engineers right now optimizing Android to be the best version on Intel architecture," Bell said. "This is a fundamental advantage that not many other people have." Eul said the same for Intel's tablet range.
Intel is ploughing a lot of time and money into this market, with Chipzilla CEO Paul Otellini promising a development cycle of twice its beloved Moore's Law. Part of that is due to rapid die shrinkage, but it seems clear that Intel has had enough of losing in the mobile market and is willing to put time and money into carving out its niche.
Both Eul and Bell are unusual, in that neither is an Intel man – both are relatively recent hires. Eul came over from the Infinion buy, so has iPhone experience. Bell is an ex-Apple VP and manager at Palm for many years. Maybe with Intel's lack of success in the field new blood was needed. ®
Re: X86 designs totally suck for Mobile?
Yup - my summary of the Medfield reviews is that performance is there as long as you don't have to emulate ARM for your apps (i.e. games), battery life is about average for a smartphone and I suspect ARM will destroy Intel on SoC price.
While moving to smaller process sizes will give Intel an advantage in increasing performance/battey life and the requirement to emulate ARM will be greatly reduced/disapper if they are successful in the marketplace, their ARM competitors are at least half a process node (i.e. 40nm vs 32nm for Medfield) behind Intel allowing them a significant cost advantage that they will retain regardless of the process node used by Intel.
Remind me what the upside of X86 in mobile is again?
They know full well that x86 based designs totally suck for mobile. Okay, I had a phone with a 486 in it once, but it was Nokia communicator brick and the 486 was pretty old and relatively efficient by then.
If you like open source and freedom then don't let Intel take over the mobile market, you don't want another monopoly. While you think ARM is a monopoly in mobile, they don't actually make any chips and anyone can licence their designs. ARM's income isn't vast but they aren't a huge company in terms of employee count.
I fail to see why they talk about greater standby times for a smart phones, If I only wanted to use it for calls I would have got a "dumb" phone. I want a greater in use time please!