backtop


Print 25 comment(s) - last by Ammohunt.. on Sep 16 at 3:46 PM


  (Source: Jason Mick/DailyTech LLC)
Intel plans to use its aggressive process development to drive it to market dominance

On Tuesday morning Intel Corp. (INTC) CEO Brian Krzanich and Intel's equally new president Renée James showed off the company's ambitious plans to retain semiconductor leadership.

I. Looking Ahead to 7 nm

"Intel can deliver technologies no other company can [today] to the [semiconductor] market," Mr. Krzanich enthused to the audience at his morning 2013 Intel Developer Forum keynote.

Key to Intel's plans to gain ground in the mobile space (tablets, smartphones)  -- where it continues to struggle -- and to retain leadership in the traditional PC space is an aggressive path of die shrinks, which culminates with an unnamed 7 nanometer chip in 2017.

IDF 2013
[Image Source: Jason Mick/DailyTech LLC]

Having delivered a 22 nm architecture refresh earlier this year (Haswell) Mr. Krzanich revealed that the world's first 14 nm consumer chip will be shipping to partners before the end of the year.

Intel blasted analyst skepticism of sub 10 nm die shrinks, putting up a host of analyst comments decrying the technologies Intel is using in today's commercial offerings and financially or logistically unfeasible or limited.

Mr. Krzanich, who joined Intel in 1982 recalls, "[Analysts] have predicted the death of Moore's Law at least once a decade since I joined Intel, but today it is still alive and well."

II. 22 nm Smartphone Atoms are Shipping

On the mobile front, Intel's smartphone-aimed Atom variants continue to trail a bit behind the Core brand desktop/laptop line in feature size (and even the Atom tablet chip line which has been receiving die shrinks somewhat sooner).  

But Intel reveals that it is currently shipping its first 22 nm smartphone chips (core: Silvermont; SoC: ValleyView; chipset: Bay Trail).  It's unclear how soon this OEM stock will reach the consumer inside actual devices, but it's possible that Intel will have a feature size lead in smartphone chips for the first time in some time, given that the smallest smartphone processors are currently built on 28 nm processes.

22 nm Atom
[Image Source: Jason Mick/DailyTech LLC]

Mr. Krzanich claims that the jump from 32 nm to 22 nm cut smartphone Atom's power usage in half.

If that's true Intel -- who supports Google Inc.'s (GOOG) ubiquitous Android and long with numerous other lesser used mobile operating systems -- may have one mean performer on its hands. Intel's last generation Medfield Atoms were in many scenarios the fastest chip available, faster than anything Arm Holdings Plc (LON:ARM) and its coalition put out.  However, their battery life was only mediocre, and they struggled to sell OEMs to switch given (reportedly) higher pricing versus ARM chips.

III. Intel Has Big Plans for Mobile Modems

The new ValleyView SoC is the first Intel chip to feature an on-die LTE modem -- a major reason why the company was able to achive such substantial power savings.  Currently that LTE modem only supports so-called "first generation" (data only) LTE, but by the end of the year a new variant will be available that supports "advanced LTE", which carries both data and voice over high speed fourth generation wireless links.

Intel Atom w/ LTE
[Image Source: Jason Mick/DailyTech LLC]

Moving ahead in 2014 Intel's major announced mobile effort will be to roll out its carrier aggregation technology, which it's currently finalizing in the lab.  After some hiccups Mr. Krzanich showed the technology using signal optimization to transform a 35 mbps (average) link to 70 mbps.  Intel says that 100 mbps connections should be possible without substantially increasing network congestion or using more power.

All in all Intel's plans look to be enough to keep it competive, although it still must make a compelling case to OEMs to overlook compatibility concerns in the mobile space.


Comments     Threshold


This article is over a month old, voting and posting comments is disabled

7, really?
By melgross on 9/11/2013 1:57:35 PM , Rating: 2
I don't know. Most experts are doubtful that 10nm will be easy. They are doubtful that anything much smaller is even possible. 2017 is a long way from now in the world of electronics. Possibly Intel has found a way around the quantum tunneling problems, or maybe they are hoping they will.




RE: 7, really?
By Jeffk464 on 9/11/2013 2:56:09 PM , Rating: 2
It would be a bummer if we really hit the wall at 14nm. Ever since Sandy Bridge is seems performance gains have been getting less and less every generation.


RE: 7, really?
By Jeffk464 on 9/11/2013 3:09:55 PM , Rating: 2
Heck the performance gains have been so small it might take a second generation 7nm core i5 to get me to upgrade from my sandy bridge core i5.


RE: 7, really?
By TSS on 9/11/2013 3:18:57 PM , Rating: 2
It's entirely possible, i thought i read an article not too long ago (on here even, but too lazy to look for it tonight) about how they'd be able to make chips below 10 nm using traditional lithography. Using some kind of strange lens setup to focus the beam way below what it should be able to accomplish due to it's wavelenght.

Sub 7nm should even be possible using entirely new techniques, but below 4nm? That's going to be a serious stretch because of the quantum mechanical effects involved. I'm no expert on that but it's what i (vaguely) remember from reading articles over the years.

At some point physics will hit a wall. Moore's law just has to break down beyond 2020. It's not just technical problems that have to be overcome, it's theoretically not even possible to go smaller.

That's when the real interesting things start. Don't get me wrong, it blows my mind that they've managed to get 7nm to work, even though it's theoretically possible. But i've been hearing about dieshrinks all the way back from before there even where pentiums (...god i feel old. I'm not old, just feel like it). Whatever they're going to think up to increase computing power at the point they cannot go any smaller, will be closer to magic then actual science (well... beyond true 3D chips that is).

Technologically speaking atleast it's a great time to be alive.


RE: 7, really?
By CalaverasGrande on 9/11/2013 4:55:20 PM , Rating: 2
maybe by the time that comes around someone will have gotten the more exotic stuff working. Josephson junctions, quantum computing etc come to mind.


"We are going to continue to work with them to make sure they understand the reality of the Internet.  A lot of these people don't have Ph.Ds, and they don't have a degree in computer science." -- RIM co-CEO Michael Lazaridis














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki