Intel may accelerate 14-nanometer Atom production

Intel may accelerate 14-nanometer Atom production

Summary: Intel is rumored to be getting ready to slash six months off the Atom timeline, bringing low-energy parts to market faster than ever before, and putting it in a strong position to compete against ARM.

SHARE:
TOPICS: Intel, Processors
20
(Source: Intel)

While Intel was undoubtedly the king when it came to PC processors, the company is having a tough time making the transition from the PC to post-PC devices such as smartphones and tablets.

But if the rumors are true, the company is to make an assault on mobile with new 14-nanometer Atom processors much earlier than expected.

The problem facing Intel at the moment is the gap in time between releasing new Core parts and Atom parts built using the same architecture. For example, 22-nanometer Core processors have been available for PCs and servers for over a year, but the company has only just started production of 22-nanometer Atom parts, and devices powered by these parts aren't expected to be in stores before the holidays.

That's a massive lag, and one that is hurting Intel given the current affinity for mobile devices.

But according to Barron's, Intel will announce a much more aggressive timeline for it's 14-nanometer parts — codenamed Bay Trail-T — at its annual conference for developers, called IDF, scheduled to kick off September 10.

Here it is expected that Intel will unveil 14-nanometer Core parts that will enter production during the second quarter of 2014, and 14-nanometer Atom parts which will enter production six months later, cutting the lag by six months.

"This is rather like in chess," analyst Tom Halfhill of the Linley Group told Barron's, "setting up the board before trying to take the king."

Thee 14-nanometer Atom parts are expected to give Intel an advantage over ARM's mobile offerings, and put Intel in the forefront of mobile chips for the first time.

The downside of Atom is that they bring in far less cash than the corresponding Core parts, which means that Intel needs to figure a way to exist in a leaner ecosystem. 

Topics: Intel, Processors

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

20 comments
Log in or register to join the discussion
  • Thats not actually correct

    Intel didn't moved current Atoms to 32nm because they didn't want the "trouble" . The current atom architecture (from 2008) is dead and they didn't waste the money of moving it to 22nm , they are moving Atom to a arquitecture based on the i core.

    And they had anounced a long time time it would keep up with Core manufacturing process , now to 22nm and move to 14nm alongside Broadwell and so on...
    brrunopt
  • "Competing Against ARM" For The Last Five Years ... And Failing

    Intel has been promising an ARM-killer since the inception of the Atom line. Just you wait, they said: the next generation of Atom chips will finally fix the performance problems, finally improve the battery life, finally become a worthy alternative to ARM, you'll see.

    Only each time, it has fallen short in some way. What's different this time? And the scandal over fixing the AnTuTu benchmark hasn't helped Intel's credibility any.
    ldo17
    • Are you still in 2011 ??

      Razr I , that singlecore Z2460 smoked most dual-cores...
      And the Clovertrail is on pair with high-end ARM SOC's (from the same time-frame)
      brrunopt
      • Re: Razr I , that singlecore Z2460 smoked most dual-cores...

        Unfortunately, no http://forums.anandtech.com/showthread.php?t=2330027
        ldo17
    • They have not.

      Intel has made no such promise of each Atom being an ARM killer. I've heard nothing like that even now. They want to compete but I don't think any of them thinks they can kill ARM. Some of the Atom line wasn't even intended to compete with ARM. That said, they are reducing the power requirements in some of the lines to the point that they are getting closer to competing with ARM while at the same time having performance advantages. They do not perform very well compared to Intel's mainstream processors but they do perform quite well against the ARM core.
      DevGuy_z
      • Re: Intel has made no such promise of each Atom being an ARM killer

        So you don't remember Intel's "full internet experience" campaign? http://www.zdnet.com/news/intel-slams-arm-for-slow-iphone-cpu/243127
        ldo17
      • True.

        The two platforms have never actually competed in the same mobile markets. So, neither platform is killing the other. If they get the new architecture Atom out at 14nm soon enough, they -might- be able to legitimately compete in the same mobile markets for the first time ever. Even then, it will come down to the popularity of the OS more than the hardware it runs on.
        BillDem
    • How about 2.4W TDP vs 5W for ARM

      And better performance at least on the processor side. The graphics side may be a problem. The only other criteria would be price.

      At these power levels I don't think they are an ARM killer but they are an ARM alternative.
      DevGuy_z
      • Not to mention,

        The Tegra 4 seems to be more like a 10W chip, peaking at around 15W under load, that needs to be actively cooled, as Nvidia themselves have done in the Shield. So it seems that the much vaunted A15 may not be the silver bullet that kills the x86 monster after all.
        Daishi83
      • Re: How about 2.4W TDP vs 5W for ARM

        The same old Intel tale....

        "look, we lowered out CPU TDP!" ... while at the same time the supporting "bridges" continue to suck power. Well, while this might have deluded someone in the desktop, where you just need to figure out what fan to use for that 140W CPU --- this is not the case with mobile SoCs.

        So the game is for actually lower TDP of the whole SoC *and* supporting components, not just the "CPU part". In that area, Intel are just too far away, because they never ever considered this an important metric.
        danbi
    • i bet amd said the same thing

      never bet against intel in the long haul. they may have senior moments at times (cough, netburst) but once they figure out where their head is at, look out.
      dosmastr@...
  • Still hindered by graphics

    The 14 nanometer atom will only do well if Intel Can manage to put a decent integrated graphics processor in it. They are still behind ARM based soc's when it comes to graphics. Maybe their acquisition of Power VR will change that.
    tommcd64
    • Power VR ??

      Power VR are the GPU's on the atom so far , see where it lead them.....
      they are ditching them and putting Intel GPU on them...
      brrunopt
    • Re: Graphic

      Well theoretically, Baytrail with its slimmed down HD4000 GPU should be giving a 200% improvement in graphics performance. Admittedly this will probably take its real world performance from 'woeful' to 'tolerable', but hopefully physical and thermal space freed up by the move to 14nm production will allow Intel to go further.

      Given the consistency of manufacturing process between Broadwell and Airmont hopefully we will see a slimmed down version of a HD6000 (or whatever they call it) which, making some assumptions based on what they have done with the Bay Trail gpu and HD5000, has something like 10-15 execution cores, compared to the 4 we are seeing this year, that will really drive performance to the 'good enough' level that seems to be their target.
      Daishi83
  • Windows RT (ARM) is the future for the masses.

    For the masses, RT is much better then Atom based tablets. Most of poeple do not need to install PC apps on tablets. They just browse, check e-mails, work with office. And RT is perfect for that. RT devices are thinner, cheaper, and with better battery life. And the most important part is, that it is safer for normal users. They do not install any crap on those machines which will make them slow after a year!!! I think this is the biggest benefit of RT. I would recommend it to any person who is not IT geek and just need a device to consume the net.
    RT is the best os, that MS has made, but poeple will need time to find out. Specially, when "IT gurus" will spread that it is already dead. :)
    Dijkstra()
    • Re: Windows RT (ARM) is the future for the masses

      The masses are buying Android instead.
      ldo17
      • true sadly

        my wife has a surface RT. other than the occasional glitch where a text field doesn't trigger the on screen keyboard to appear, she's been very happy with it. And he's right, she hasn't wanted to install anything not in the app store (or whatever they call it) as its an email/browsing/netflix device for her.

        I've had issues with 2 android devices, one /ICS one GB, (the ICS on a samsung is pretty good though) wife has a RT and a WP 7.5... she's has next to no issues with both (aside from the one above)

        RMillers: Never thought I'd say this but somehow.... uh.....well.... the MS OS's are more stable and trouble-free.... wow I feel like I just threw up saying that.....I'm in shock in fact....
        dosmastr@...
  • I make my living on Microsoft platforms

    If you take backward compatibility and office apps out of the mix, and are not running thin or thick clients as part of an enterprise ecosystem, there is no reason to run Microsoft on your tablet. That's why I run an Nexus 7. Intel MAY end of catching and passing ARM. If fact, I think it will happen. But Windows RT has a LONG way to go match the ease of use and intuitive nature of Android. I think they may get there too. But Microsoft, even with they're army of developers, is 100,000s of man hours behind. it'll take a while to catch up. I guess it depends on Apple and Google and how fast an far they push Android and iOS in the meantime.
    rmillersbs
  • look at you, figured out how to copy-paste?

    now try ctrl+z -- it's even more fun
    vpupkin
  • Intel could outperform ARM

    Of course, technically, Intel can produce an Atom chip that outperforms ARM in any possible way, by manufacturing it to the latest fabs they purchased.

    Will they do this? Very unlikely, because it makes absolutely no sense.

    Intel has built their current business model around their famous tick tock model (http://www.intel.com/content/www/us/en/silicon-innovations/intel-tick-tock-model-general.html).
    This in essence requires that they manufacture their highest profit products (server and desktop CPUs) using the most advanced (and very expensive) tech they can buy. This is in order to keep competitors such as AMD at bay, because otherwise those will eat significant portion of the profitable market, as AMD did with Opteron.
    Only when these fabs are no longer used for the profitable parts, will Intel start manufacturing their "cheap" parts (aka Atom) there.

    In the past, Intel would refuse to build parts for certain customer (segment) and instead insist that whatever they do is for general mass market use. However, the economic realities of their business and competition made them realize this is not very smart thing to do.
    We see the first cracks with examples such as the (almost unique) use of specific Haswell versions in Apple's latest MacBook Air, that lets Intel successfully compete with the AMDs integrated solutions -- while at the same time continuing to sell their mainstream chips to less demanding customers.

    So in general, I do not expect Intel to change their strategy just yet. At least, not until they feel serious lowering in the demand of their expensive chips.
    danbi