I can distinctly remember the day when Intel Corp. (s INTC) launched the Pentium processor. It was the day the desktop computing changed for me and for a lot of others. It was also the day when Intel started to put a gap between itself and all its wannabe processor rivals. I bring up that day because I feel that we are about to see a similar shift in the world of mobile, thanks to ARM Holdings (s ARMH), a company that develops and licenses chip technologies to others like Texas Instruments (s txn), Samsung and Qualcomm (s qcom).
ARM, today is introducing a new chip architecture called the Cortex-A15 MPCore. This architecture will form the underpinning of the newest (and perhaps the beefiest) members of the Cortex family of mobile chips that power our iPhones, Samsung Galaxys and the iPads. Thanks to this new architecture, companies such as TI and Samsung will make chips that will come in dual and quad core configurations and will run at clock speeds of up to 2.5 GHz. Don’t be surprised that by 2012 our tablets and smart phones on average be about five times as powerful, with no detrimental impact on power consumption.
And while Apple is nowhere to be found in official ARM’s literature, it goes without saying that many of its products are going to get a major boost because of the new generation of Cortex-A15 architecture-based chips. Why – because Apple is a major license of ARM’s technology.
The Power Principle
So why is this new new chip architecture important? The answer is pretty simple. As we have often explained in the past, the computing is going through a transition akin to the shift from fixed line phones to cellular telephones. Computing is becoming portable and pocketable. It is omnipresent and at our finger tips. It is making us rethink all current notions about the Internet. Mobile connectivity is also bringing the power of the cloud to our palms.
Soon we are going to have even faster networks at our disposal, thanks to the rise of next generation wireless broadband technologies such as Long-Term Evolution or LTE. These faster networks will bring data to our devices at much higher speeds, which mean we will need faster chips to process that information. Just as the growth of faster broadband sparked the sales of ever-more-powerful Pentium chips, a similar trend is going to take hold in the wireless world.
This new world needs a new kind of architecture – one that marries power with very little power consumption so as to give long battery life to our portable devices. “Even with a lot of bandwidth, we are still going to need processing power in the devices,” explained ARM’s director of marketing, Nandan Nayampally. Think of this chip as a heavyweight boxer with the stamina of a long distance runner.
Augmented Reality Gets a Boost
Playing games in 3-D, running work and home environments on the same machine, conducting videoconferences along with dozens of other activities are going to be a breeze for devices powered by this new chip technology. But that is not all since it will be able to equally at home inside a new generation of web servers and personal home devices that need beefy yet power efficient processors. The low power requirement eliminates the need for fans and makes these device cool and quiet. Running on these new chips are a slew of operating systems including the fast-growing Android, Ubuntu Linux and Symbian.
One of the mobile technologies that would likely to get a big boost from this new chip –- augmented reality. Sure you have heard of companies like Layar, but the fact is that AR is going to remain a curiosity unless the chips can take all the visual and other information and turn it into something magic instantly. We are not there yet, but a chip built on the Cortex-A15 architecture can help.
Intel’s Problem
The new ARM architecture is likely to cause further heartache for Intel which has been trying to position itself in the mobile world through various efforts including its low-power Atom processors, and more recently via a $1.4 billion acquisition of Infineon’s wireless chip business. The company has made some strides with its new mobile oriented chips, but the folks at ARM aren’t really sweating it for now.
Nayampally pointed out that slightly older Cortex-A8 chips are enough to take on net books and the newer Cortex-A9 chips are leaving Intel in the dust. Intel, he said is trying to focus on lowering power consumption, a problem ARM has licked. For now, he said the Cambridge, UK-based company is pretty comfortable with its Cortex roadmap. And why not –- more than 20 billion ARM-based processors have found way into our lives. And the company is only just getting started.
Related content from GigaOM Pro (sub req’d):
As Devices Converge, Chip Vendors Girding for a Fight
For Phones, the Future Is Multiple Cores
“As we have often explained in the past, the computing is going through a transition akin to the shift from fixed line phones to telephones.”
What? cellphones.
Awe shucks. Missed cellular in there. Fixed and added!
This make me feel that by 2012 we will have client side applications for all heavyweight web applications to boost up the speed/execution.
Yup, that would make sense.
It is a great news to hear about innovation in chips makers other than Intel.
Intel has become too big already and it is starting to affect innovation in the Industry.
Thanks ARM!
Arm has been pushing the envelope for quite a bit now. Enough to worry intel. By the way more than 20 billion processors based on ARM have been shipped so far.
Proofread. Please.
“It was the day the desktop computing changed for me and for a lot of others.”
“As we have often explained in the past, the computing is going through…”
“But that is not all since it will be able to equally at home inside…”
Love the article, interesting as always.. But ye gods… Grammar check disabled? I winced each time I ran across and extra “the” and wondered how much was missing in the 3rd booboo.
The Future of Digital Life:
Augmented Reality should change everything in computing.
http://www.committedexpertise.com/2/post/2010/06/mobile-augmented-reality-prime-time.html
/Malick.
D’oh, I typed my comment on an ipad so I am not responsible for the extra “d” in my own spelling.. Seriously though, what do you think the upper limit on cellphone processing power is; will we eventually see water-cooled cellphones hitting 5GHz?
Ronald
Thanks for the grammar related comments. Late night man!
Secondly I think we should see some more speed bumps, but not much. All mobile chip related innovation has to combine power+Power efficiency + bandwidth and that wouldn’t necessarily mean mega chips like we are used to in the pc world.
9:01 PM is late night? MAN!
yes for some of us who start their day at 4 am it is late night Don. thanks for your comment
It truly is.
Isn’t there something missing on the SW site. If we have 4core systems where’s the SW layer which allows apps to interchange data easily and safely? Apple cut back on inter apps data exchange for a reason.
Context processing is massively parallel, as smarter ones apps have to get as more context one needs as more parallel processing is required. Making it faster doesn’t necessarily make it “smarter”.
A bit skeptical on the performance comparison of the current Cortex line to the Atom
(see http://www.engadget.com/2009/09/16/arms-cortex-a9-beats-atom-n270-too-bad-its-not-2008/ )
Now that being said, Intel is still in trouble as the Atom cores can’t compare power consumption-wise. There will still be a market for servers which will drive the content to mobile devices, but as mobile devices continue to take on more tasks (AutoCAD just got released for the iPad) from the desktop world, Intel’s market share will continue to shrink.
Maybe selling off their X-Scale line was a bad idea 🙂
Om – a couple questions;
1.) Given the support for virtualization and higher memory addressability of this processor, two deficiencies in prior ARM cores, is this new processor the key to penetrating the server market? Is this the core that SmoothStone is using for their designs?
2.) Since this core will support virtualization, is it possible to run Windows in the abstraction layer through the hypervisor and therefore this becomes a back-door way to run windows on ARM? seems too simple of a solution to actually work…
thanks in advance
best,
gt
Hey
1. Yes I think this will help them get further penetration in the web server market.
2. On Virtualization, I think it has less to do with running different OSes and more about running different “use case environments.”
agree, even early adopters will not be willing to run their mission critical applications on ARM-based servers initially. But we’ll need OS compatibility w/ ARM first. Windows is the obvious OS that hasn’t ported over to ARM yet.
Don’t forget that virtualisation does not mean it can run many oses, for instance, Windows is built to run on X86 architecture which if I remember correctly is not part of ARM’s featureset
correct – windows is built to run on x86 and not ARM. my question centered around the ability to use cortex a15’s VMWare-compatibility as a work-around to get Windows to work on ARM…using the hypervisor as the interface between the two. not sure if it would work, i’m speculating it wouldn’t.
i am fairly certain that Windows v8 will be optimized for ARM though. if this version is released in ~18 months, it will likely coincide w/ the production runs of these new Cortex A15 chips, could be interesting to see what happens.
Power and more power. The argument made is well put together but probably unrealistic. More bandwith? Streaming? Greater power needs? It may cost an arm and a leg to do what is written. Service Providers will drive growth in this area, not chip manufacturers nor device assemblers nor software makers and nor anyone else. The tariffs servie providers will place per download and the way wireless web will develop will dictate this.
Not the same old game in the PCs as big difference between smartphones and netbooks is that one of the two cannot really be used without access to the Internet and bandwidth.
This will be great for home devices, but OM, come on … not even you can believe this paradigm shift is going to make any real dent or difference in the mobile space with carrier greed running rampant like it is today.
Sure (sharks grin) you will be able to use your devices full capability for a measly $299 a month for unlimited everything* (*up to 7 GB total bandwidth per 31 day period.)
Hah! Major shifts worth using in the mobile space will only ever come when someone somehow neutralizes the carriers. Its going to be funny how the big telecos corporations will be vilified in the future for all they did to kill innovation. Like some bad scifi story.
Hopefully, the corp. hate will create change after the dreaded Telecos riots of 2056 leave only the scared small providers behind.
“Intel, he said is trying to focus on lowering power consumption, a problem ARM has licked”.
What the hell does that mean?
Well, the Atom processor ‘success’ (which helped spawn a new market segment – netbooks) was reportedly a failed attempt to develop a cellphone-worthy chip… Intel has had a hard time making its chips fit a cellphone/smartphone/tablet power budget. Reasons are too numerous and complex to fit here, but suffice it to say that there is still a lot of over-design inherent in Intel’s design flow, notably to meet timing and power integrity constraints.
One of the reasons we are seeing multi-core proliferation is that ‘overclocking’ existing architectures is not easy. With clocks and memory buses running at GHz speeds, maintaining ‘signal integrity’ in the chip becomes a nightmare. ARM has been relying on Synopsys to solve such problems in their design flow, while Intel has been heavier on home-brew tools…
I’m game for any device that Windows desktop doesn’t run on. And I hope that ARM devices become the consumer’s favorite computer platform. It’ll save a lot of space and lots of electricity. It’ll turn computing into a greener environment.