We've heard a similar message from Nokia dating all the way back to
MeeGo's introduction at
MWC back in February, so it comes as little surprise that Espoo is apparently trumpeting the virtues of ARM for its first MeeGo-powered device that's still targeted for the tail end of 2010. What might make this particularly interesting is the fact that
MeeGo 1.0 is clearly further along for Atom devices than it is for the Cortex A8-based
N900, not to mention that Nokia has already warmed up to Intel thanks to its
Booklet 3G -- but regardless of the silicon, getting the platform solid enough for any sort of retail device by the end of 2010 still seems like a tricky proposition when you figure that the ARM build doesn't even have a proper user interface yet. Ultimately, it might come down to a question of size; Intel still hasn't proven that it can scale Atom down far enough to tackle the smartphone market head-on, so if Nokia wants to go small with its first MeeGo hardware, that alone could be impetus enough to go ARM.
posted May 30th 2010 9:58PM
Is this just gonna become another niche product like the N900?
@Almo
Probably not, given the apparent ease of MeeGo 1.0 UI.
The question now is: are they gonna produce a smarter phone, or will they officially ride the wave of the slates and the tablets?
Btw Engadget Nokia's first MeeGo OS is Harmattan and Harmattan was announced in 2008 already and release date was end of 2010.
It will be seen if it's left as a niche but Nokia have made it clear that Harmattan/Meego is their new high end OS and they are expecting 30% of their whole revenue to come from there already next year.
That would probally mean something like +20 million sales next year.
@Almo Define niche. iPhone/Android might also be considered niche products. If by niche you mean, high end? Yes MeeGo will only be delivered for high end devices (for now) since Symbian S40 will be the bulk of the sales.
One correction to chris "the ARM build doesn't even have a proper user interface yet." The build you are referring too is the merged branch of MeeGo from both Intel and Nokia, and no, it does not have UI, yet. Having said that MeeGo Harmattan does have a UI and it is going to be the version of MeeGo released by Nokia in late 2010.
So to sum up, and for you to update the article (ha! I wish), Nokia does have UI for MeeGo Haramattan.
ps. And please Chris, dont say that MeeGo does not have UI if you dont even know/understand how things are going. I have seen working UI for non-Harmattan MeeGo.
Seems that my "sources" are better than yours.
@Mr w00t
Did you perhaps mean S60? S40 isn't Symbian.
@Mr w00t
By niche I meant a couple of ten thousand sold and not the mainstream bulk-seller than Nokia is looking for in the smartphone market. Sorta like the N900 really. Nice device but didn't sell much, mostly to a select demographic.
@Almo
I could agree on the select demographics part, but couple a ten thousand is ranking it way low. From a recent Engadget article, we learn that N900 sold 100,000 in five weeks, and Nokia still can't match demand (or don't need to, since this was just supposed to be a concept).
That's how big Nokia is, they can churn out a concept set before real deal, and even that gets sold out. :D
The real deal, according to rumours, will be a supposed "N9", and given the extremely positive response Nokia got from the test run of N900, and seeing the hype that is following the upcoming N8, I think the N9 will most definitely not be "niche".
@Almo
N900 outsold Nexus 1 by a factor of 3.
@JFH
"(Reuters) - Nokia sold less than 100,000 top-of-the-range N900 smartphones in its first five months on the market, researcher Gartner said, indicating it has yet to mount a serious challenge to the iPhone and Blackberry."
"According to analytics firm Flurry, 135,000 units of the Nexus One were sold after the first 74 days. "
I mean it doesn't really matter we're talking margins of error, but I'm pretty sure N1 has sold more than N900...
@hfm
Yeah well NO genius. Look at the entire updated article. Gartner got it wrong, and Engadget corrected it in the article themselves. 5 weeks to hit 100k, with major markets not served when they hit the 100k. Ergo, it is virtually impossible the N900 sold less than the Nexus 1, and very likely that the device sales are around 350 to 400k, based on downloads in maemo.org, usage patterns in flickr, and various other metrics. Industry experts agree. I do not have an exact number, but the multiple is between 2 and 4, so I went with 3.
Is something wrong with ARM on smartphones? I really don't see the benefits of using x86 instead, even if intel could give the same performance/watt.
@MaTdg
"Is something wrong with x86 on smartphones? I really don't see the benefits of using ARM instead, even if Nvidia could give the same performance/watt." Fixed
Read the article below for some perspective on why Tegra 2 on A9 isn't going to be the latest-and-greatest, and why Jen-Hsun Huang is unnecessarily trashing Intel.
http://www.anandtech.com/show/3696/intel-unveils-moorestown-and-the-atom-z600-series-the-fastest-smartphone-processor/
http://www.engadget.com/2010/05/24/nvidia-intels-moorestown-is-like-an-elephant-on-a-diet-ipad-s/
Probably the only thing wrong is that Nokia hasn't coded MeeGo for operation on ARM architecture.
For a company with no real history in developing computing platforms and only limited skills in making cross-cpu compatible systems this is probably a monster issue.
Nokia has gotten away with dumb phones and simple menu systems for too long
@MaTdg
I could ask the same question about ARM.
It all comes down to who was first. ARM clearly has a more mature architecture in the mobile world, and their licensing methods allows developement by MANY different companies and foundaries...Also, being basically the only architecture in smartphones right now, it also means most mobile operating systems are developed for ARM chips...
On the other hand, Intel, and x86 in general, has the upper hand, when it comes to performance. Intel is trying to cut down the power consumption, and Moorestown is a start. They are at the 32nm level already, so once the new Medfield SoC comes out from Intel in 2011, they become a serious threat to ARM processors. The irony of Intel's woes comes from the software side of things. Intel had to develope Moblin/MeeGo and port Android for x86 just to make their chips attractive to other manufacturers, which puts quite a burden, both financially and time-wise on Intel.
In the end, the consumers win, since they will get better prices and better performance due to competition.
@Cy Starkman
Um, Symbian? Nokia has been coding mobile OS's for longer than anyone else in the business. You are clearly misinformed.
@MaTdg lol at the people who want x86 on smartphones, obviously they don't know that arm processors and other RISC processors are a sh*t load more efficient at processing information than x86 processors. intel will never gain a foothole in the smartphone arena.
http://arstechnica.com/gadgets/news/2010/05/intel-fires-opening-salvo-in-x86-vs-arm-smartphone-wars.ars
@Cy Starkman Have you ever heard of Qt? Qt is cross platform. In principle anything that is open source is cross platform because it can be easily compiled to target architecture. Nokia is doing a LOT of open source. Symbian in GPL-ed in fact.
@saturnblackhole
For the most part, I love Arstechnica (despite the fact that their name does not mean what they think it means in Latin). But Stokes is misguided and stubborn in his view here.
"Nowhere does Intel reveal how much power Moorestown draws when running Quake 3 at 100FPS, or when executing the SunSpider JavaScript benchmark in two seconds. That's because the power draw in those situations is __certain__ to be astronomic by smartphone standards."
I'm surprised to read that Jon Stokes is so intimately acquainted with Intel's R&D team. I'm even more surprised the Stokes would have the gall to present the absence of evidence as certain proof, on which he bases such a powerful dismissal. Stokes's article, which borrows from Anand's much more qualified and substantiated piece, entirely misses the point Intel reiterates. The power consumption issue in the mobile arena is most noticeable during idle use. Yes, if software video playback kills a 1.5 Ah battery in 1 hour, then we have a problem (even though no one outside of Intel can know this for sure). But if x86 is as powerful as Intel claims, you will NOT tax the processor for something so menial. Like Speedstep in the Atom family, power states allow for unparalleled high performance in necessary situations, while lowering power usage in idle states, as mentioned.
To clarify: Intel's philosophy is drastically different from ARM's. The former wants lower power usage all the time, while idle power constrains top-end performance. Intel wants high performance at a reasonably efficient power cost, so that you can get what you need done quickly and return to idle. The latter is much more flexible, but as should be obvious, the lack of evidence is hardly a promise of anything. It's possible that Moorestown will offer the same performance of A9-based chipsets with the same power draw. The difference is that Intel's solution can scale up, using more power to increase performance at a lower performance/power cost, then scale down to reduce performance and allow the touted low absolute power cost.
Yes, this is heavily reliant on software, but consider Intel's $400 million acquisition of Wind River and development of a Moorestown version of Android. This is not a one-bunned effort. As seen with the iPhone and Pre, when you optimize only for one product and ignore the need for flexibility, you get better result comparing equal products. What HVGA phone can perform the iPhone/Pre's level? Have you even compared Maemo to Android (ignoring the number of apps from the latter)? Android has to run on QVGA as well as WVGA, which is fine and dandy for cross-platform development, but absolute performance suffers compared to a platform that ONLY runs one or the other. But this rigidness is why x86 is important. While Intel develops power management software that is device specific (regarding low and high performance), the flexibility of x86 can emulate ARM's ubiquity in the mobile space while allowing for optimized products. In other words, you can get a 32-bit platform that can run extremely intensive programs at its best, or you can get a much less favorable experience but at very high efficiency, but they BOTH run it.
Intel is also gambling, in assuming that software development will progress greatly and unpredictably. The only analogy I think you'll understand is from Iron Man 2 (SPOILER ALERT!): Tony Stark's dad had an idea for a certain element that would better fuel the reactor, but the technology had not come along yet to synthesize it. Intel has technology (well, they plan to), but the level of software involvement is currently a little higher than production cost efficient.
I'm not claiming that I KNOW Moorestown will be good. But if you take Stokes's perspective, then why the hell would Intel be investing so much if Medfield is going to be so much better? Why would Intel half-ass something as economically promising as mobile computing?Remember the Silverthorne MIDs? Do you know you don't see any today? A lazy product doesn't sell. Would Intel rather develop a temp solution first, waste production resources, and continue developing? They already have a cash-cow, so Stokes's business logic is at best faulty.
Please come into this company war with an open-mind. Preliminary numbers mean nothing to consumers. When Intel has a Z6 proc in your hand, THEN you can tell me it's bad.
@stoffer
Symbian is released under the Eclipse Public License, not the GPL. Qt is available under GPLv3, however, in addition to LGPLv2 and some custom commercial license.
that's kind of a slap in the face to intel
@JeremyBenthem
Atom is not good enough for the highly mobile netbooks Intel envisioned with Moorestown. We're going to have to wait until Q3 to see any x86 architecture that will viably compete with (and likely out-do) Tegra 2 and other A9-based chipsets. Q4 should bring the devices toting them. Medfield should be the dreamy mobile processor that we've all been hoping for, but there's no accurate timeline.
@Hung
Well, clearly the reply system should have added my reply as a thread to JeremyBenthem
@Hung
what I mean is that MeeGo was a colaboration between Nokia and intel, and then Nokia just goes ahead and skips on the intel product
No surprises here. Intel's Moorestown is a joke when it comes to power management. Their graphics solution is the same as any other typical ARM vendor, IMG. Intel has been struggling to just improve their chips against the previous Atom iteration. Meanwhile, ARM is eating their lunch.
I am wondering if Intel is regretting their sale of XScale to Marvell? Because it sure seems like ARM is doing a much better job scaling their architecture UP than Intel is at scaling their architecture down.
@duine
Again, an uneducated, trollish response. Have you actually used a Moorestown fitted device? What about the previous generation? With the exception of the correct fact regarding the graphics technology, can you back up ANYTHING you just said? ARM is doing well, due to their architecture, and their licensing practices, as well as being the first to the mobile game. This battle to mobile chip domination is in its infancy. If you want to compare timelines with the desktop chips, imagine we are just before Win95 launch, but instead of x86, RISC/ARM is dominating...
I think that Nokia has been making some poor decisions lately. First creating a partnership with yahoo (who still uses yahoo?) and now using x86 in its smartphones when we have seen that it is not ready. I have ben defending Nokia all this time, and am a fan of their phones; but if this continues then i'll be switching to something else.
@geegee
Who still uses Yahoo? Around 40% of Asia and Middle East, that's who.
In North America, Yahoo still holds a good percentage (Around 25-30%) of search. And they have a huge number of users for their Email. Yahoo has a lot of parts. If they started selling some of them, as they have done so in the past year, and focus on some main sections of their business, they are a force to be reckoned with, but then again, Probably not~!
@geegee
Did you read the article?
Nokia will use ARM on their next Meego smartphone, not x86.
Seriously?
@kapanak Not to mention that Asia is the biggest market for Nokia.
Is this really news? Hasn't it been known since the Meego announcement that the mobile half of Meego would run on ARM and the computer half run on Intel Atom chips?
Well I believe that Nokia is talking about their MeeGo, meaning Maemo 6. Not the MeeGo 1.0 that was released few days ago which was based on Moblin. I'm not good with any of this but I rememebr hearing that Moblin and Maemo will be completely joined around MeeGo 2.2. Untill then they will be different.
@1q2w3e4r I think you are right. There is no way they would completely merge two operating systems in such a short time.
Interesting interview about the Nokia side of Meego project by the big boss.
http://mynokiablog.com/2010/05/29/video-meego-qa-with-alberto-torres-nokia-evp-meego-computers-at-openmobile/
So now we've narrowed it down to a dozen's of possible chips eh'. Let's get more specific, will it be another Omap processor or what?
@Edobe, OMAP is a SoC, not a processor. OMAP SoC uses various ARM CPUs at its core.
@incognito k Yes technically it is a series of microprocessors, however, the word processor is widely substituted for soc. Any many "soc's" use arm based cpus. I'm saying if Nokia is saying there first Meego phone will be arm based, they must no wich soc they'll be using. No?
@Edobe, last year they've announced that their Harmattan (Maemo 6) device, that is now rebranded as MeeGo, will sport the same core platform as the N900, so yes - their first MeeGo device will be running on the OMAP3430 (or maybe they size it up a notch and go for the OMAP3530), hence it will be running on ARM Cortex-A8 at its base. What will be the base of their next years' device, which should be a true MeeGo device, remains to be seen and I guess it mostly depends on will Intel be able to make their CPUs more power efficient. Well, Intel is surely able to do that, but then it wouldn't be an x86-base CPU - and I sincerely doubt they could scale down a CISC processor to be as efficient as a RISC...
Thanks Nokia, but MeeStay.
Be nice to see Nokia make a 9" 3G/4G, WiFi tablet with Meego on it, can't have the Sth Koreans and Apple having all the fun.
I was talking to a Nokia insider a few days ago and I was told that Nokia will use ARM on their smartphones regardless of the platform and that they will use Moorestown on an upcoming Meego tablet. Though don't quote me on that just yet.
Good grief! Silicon is the last thing that matters in Nokias' transition to Meego. Engadget proves once again that it understands (or cares to understand) almost nothing about the task at hand.
Existing Maemo has full support for the underlying hardware on N900 - when it comes to UI, drivers for the SoC-integrated SGX GPU are most relevant, but also there are the companion chips that make the device usable - and most of these are such as power management at a very explicit level. Situation is hardly worse than in the case of Intels' porting effort: they have hardware essentials covered, into the level of supporting 3G modem and other components peripheral to the SoC.
What is the real issue on Meego effort is that it's actually being approached from two code bases towards sufficiently common target, not yet reached a common target by any reasonable measure. Meego Netbook User Experience is basically slightly modified Moblin in disguise - and "Meego Core Software Platform" is Maemo with all the parts under radical transition stripped away. Eventually they will share same software platform, but at this point, "Meego" is mostly a moniker for this process, not a product ready for use on all intended environments.
The fact that Maemo-derived release is so much more limited at the point has very little to do with the underlying silicon, instead it has lots to do with software platform transition where Nokia has chosen to make bigger changes to its' own codebase, or rather, their software platform architecture. Radical operations always have moments where whole operation seems like a glorified stabbing - but with advance planning and consistent task execution, results are better than that.
And talking about ARM vs Intel silicon and the pondering why Nokia doesn't choose Intel: Intel is *still* about as far from taking over ARM's home market turf as ARM is in taking Intel on high-end desktops and servers. ARM has been properly designed for the tasks in energy efficient/low power computing from the start. Why would Nokia choose Intel silicon in a situation where Intel is *still* struggling to prove that any of its' solutions can beat energy efficiency of any ARM solution on the market, and certainly lose in the lower end? Unless Nokia would move to Meego Netbook UI/UX on the phones and build them with the philosophy of ultra-small netbooks, there would be no benefit from current state of Meego software architecture either.
Design phase of a Nokia phone takes at least an year, and real-world consumer appeal of pocketable PCs has been disappointing at the best - so, why Engadget would think jumping to Intel bandwagon would be the solution in a situation that's probably going to change in at most two months..?
@foobaz2 Someone mode the parent post "Informative"
@stoffer
I finally have mod points! Yes!!
Oh wait... this isn't slashdot ;-P
I almost bought the N900 until I read about the quirks associated with the UI because of the fact that this is the first time Maemo has made it to a phone. I was going to hold out for the N900's successor. Now I'm wondering if the change to Meego platform is going to be a set back for the UI.
Fingers crossed it won't be.
@Nate Dogg
I just got my N900 a few days ago, could install 1.2 the next day, and the only quirk I see is around portrait texting / portrait virtual keyboard. Outside of that, it does everything any modern OS does, plus a bunch more. I have not overclocked it but this thing already flies. I would say go for it!
Nokia MeeGo device will run on A9.
@JFH I think you meant A8 (MeeGo is currently using n900 hardware as the base for the next MeeGo device).
As far as I know the next Nokia's MeeGo device will be a capacitive[-] & based on A8 hardware.
Feel free to correct me , if this is wrong.
@Engadget: You seems you don't know what you are talking about (About UI/UX , ...)
@Sp4mer
Harmattan MeeGo on A8, true MeeGo on U8500 by ST Ericsson... that is what I heard.
@Sp4mer
You're right, it's not A9. It's on a OMAP3630