Intel has the Fabs, so it think just going down the nm scale they will eventually get inside the power envelope, duh. So, they fired the guy at the helm who was their architect for low power designs, what are they trying to tell us?
Umm, wait we have a chipset christened Tapioca which is going to come in 2015 with 100 Angstroms geometry which will run under 0.25w and run Windows!
As has already been pointed out, Intel is pretty much out of the running in the cell and tablet spaces. The last paragraph of the article talks about how this new Atom enables larger displays, which is a little silly considering that it was Intel contractual limitations that restricted netbooks to 10.1" to avoid cannibalizing full notebooks. I am finding myself less and less surprised that they dumped their former mobile division head.
It is possible, or maybe even likely, that the Atom can get down to 1W at 22nm but where will ARM be then or soon after. It's not just a question of power. It's also a question of what can be done with that power. A smaller geometry ARM will be able to integrate more cores and/or have lower price and/or even less power required for some applications.
It could be argued that at some point the processor power becomes a relatively insignificant proportion of total power budget but there will be pressure in many or most applications to maximize performance at whatever the power budget might happen to be. There's also die size / cost issues. The x86 carries unnecessary and costly baggage in all these areas.
I would have to agree, 4W seems to be an awful lot of power for a portable. Intel will have to focus on a much lower power device with great power savings modes if it wants to expand into the tablet and smartphones worlds.
@mike: Injtel said for more than a year the 32nm Medfield would e the one to make it into smartphones, but I think we are about to find it is not. A 22nm version will be better but will still have to compete with nex-gen ARMs and its lack of a mature software environment embraced by developers
Every Atom up until 22nm is just a stepping stone for Intel into smartphones -- a foot in the door to wring out any design issues before 22nm hits production. Don't look at today's 45nm beast and think that it's the same as you'll see later.
When Intel first introduced the first Pentium II cartridge-based Celeron in the late '90s, it was pretty much panned as being too underperforming -- Intel had simply taken the L2 cache chips off the standard PII cartridge for a cheaper, tho' lower performing, entry level product. It was a stepping stone. With the next cycle, they came out with the real deal, a single chip product with integrated L2 cache and established Celeron in a new lower power, lower performance, cheaper market segment that ate up the entry-level market.
Look for Atom's 22nm offspring with dramatically lower power to be the real contender. That's what Intel is pointing to.
Intel continues its downward slide in the mobile market. There is no way x86 is going to hack it.
Intel has three choices, stick with x86 and whither away, me-too another ARM or come up with a new architecture better than ARM.
x86 beat RISC by using the x86 as a cash cow to fund development at the same time that RISC was having to get more complicated to keep up. That was before speed-power efficiency became king.
At this point Intel should analyze performance-power efficiency from a system perspective and develop the necessary computation, memory and I/O elements to meet optimum user expectations.
"The 45nm Intel Atom Z670, aka Oak Trail, is a two chip set that consumes up to 4W and costs about $75." !! 75$ for 2 chips in tablet formfactor! Intel must be dreaming. Hoping to get the same x86 margins in mobile cpu business. :)
Replay available now: A handful of emerging network technologies are competing to be the preferred wide-area connection for the Internet of Things. All claim lower costs and power use than cellular but none have wide deployment yet. Listen in as proponents of leading contenders make their case to be the metro or national IoT network of the future. Rick Merritt, EE Times Silicon Valley Bureau Chief, moderators this discussion. Join in and ask his guests questions.