Intel is outlining plans for its next-gen chips for smartphones and tablets… as well as the generation after that. Starting in the second half of 2014, Intel plans to start offering 64-bit mobile chips based on 14nm designs and featuring next-gen graphics.
By 2015, Intel plans to move onto an entirely new versatile chip called “Broxton” whcih is designed for easy upgrades and reconfigurations.
Merrifield
In the first half of 2014 we should see next-gen dual-core Atom chips based on the same Silvermont technology currently used in Bay Trail processors.
These 22nm chips will offers up to twice the graphics performance of the company’s current smartphone chips, and up to 1.7 times the overall performance. They’ll be aimed at high-end phones and mid-range tablets.
Moorefield
Why stop at 2 cores when you can have 4? Moorfield is basically a quad-core variant of the 22nm Merrifield chips.
SoFIA
While high performance chips tend to grab the headlines, Intel is also making a play for the low-end smartphone space with a new system-on-a-chip code-named SoFIA. It’ll feature integrated 3G capabilities when it launches in late 2014, and by the following Year Intel plans to have a version with built-in support for 4G LTE networks.
Cherry Trail
By the second half of 2014, Silvermont will be old hat and Intel will move onto a new 14nm architecture code-named “Airmont.”
The first Atom chips using this architecture will be code-named Cherry Trail, and they’ll be designed to offer higher performance and lower power consumption than they’re predecessors. The’ll also feature next-gen Intel graphics.
Broxton
By 2015, Intel plans to roll out chips based on a new “Goldmont” architecture. Broxton chips will have a chassis which lets Intel connect other parts, allowing Broxton chips to be changed or customized. This’ll let Intel roll out new variations of the chips for different markets more quickly.
Intel also describes these chips as offering “converged cores” for both phones and tablets,
via PC World, X-bit Labs, and VR Zone
At 14nm it will probably cost an ARM and a leg. Any words from Intel about pricing?
Pricing should be the same as the current phone chips.
If Intel can keep up with their excellent open source Linux drivers for their smartphone based chips as in their other chips (ie. GPU, WiFi NICs, Ethernet NICs, etc.) and can keep a competitive performance per watt ratio compared to ARM then I’d definitely would buy an Intel based smartphone over anything ARM at that time.
With open source drivers and maybe even mainlined into the Linux driver, an Intel phone will likely receive official Android updates longer than ARM based ones. If not, then custom ROMs with newer versions of Android will actually be stable and likely would have more working hardware due to the open drivers.
I hope Intel maintains the excellent open source Linux support for their in-house chips. With that, I hope ARM Holdings and/or ARM vendors get together and start standardizing the platform and opening things up like on x86. Win-win for consumers.
We live in interesting times. Can Intel scale down faster than ARM can scale up?
Long term, things look to favor Intel Intel has several things in
its favor: Not such a bad thing as Intel’s an American company.
Many of the companies involved in ARM chips are foreign, such as
Samsung, Meditek, Taiwan Semiconductor (TSMC), and ARM Holdings itself is a UK company American companies involved in ARM include Motorola, Qualcomm, AMD, Apple, and Google.
– a big pile of cash to fund coop marketing, develop reference
designs (which offloads R&D from device manufacturers, freeing
them to quickly bring new devices to market), and do its own
R&D, fund OS development, and build new fabs
– unmatched process technology that’s 1-2 generations ahead of
anything ARM’s foundry partners can come up with. The advantages Intel holds here can make up for architectural disadvantages that x86 has vs. ARM.
– Intel has oodles of fab capacity and can crank out chips to meet any surge in demand.
– a cash cow and fat (obscene?) margins in its Core product line,
compared to the slim margins ARM chip makers endure. Texas Instruments had to pull out of the brutal merchant ARM chip market So Intel can get down and dirty on price if it needs to.
It is all about the ecosystem. Intel will scale down in less than two years and that will mean having the entire x86 ecosystem to work with. ARM will take decades to reach that ecosystem level, even if it can upscale and compete on processor performance. The ecosystem is what matters most!