Holy Moorestown batman!

[yh125d]

Read the article on front page if you haven't


As it would seem, much better graphics performance (up to 2x max clock vs current powervr products, 1080p HW decoding), quite a bit better app performance (can do x86, which may be good, but means most if not all existing smartphone apps would need recoded), competitive battery life, but no WP7. or WP8.

Overall I think it will be a fairly good platform, but needs more integration (5 chips? :|). Moorestown successor single chip at 32nm is going to be a killer I tell ya


Though I wish Intel had teamed up with google and HTC instead of nokia, cause I dunno about moblin/meego. Moorestown powered Android running HTC device? Yes please.

 

[tatteredpotato]

I've been curious as to how long Intel was going to let ARM rule the smartphone market, seeing as it's definitely going to be a massive market for CPUs in the future. I'd be curious to see if they can get an x86 down to the power consumption required for mobile/embedded devices.

 

[Deeko]

Its impressive today - but how impressive will it be by the time there's a Moorestown-based device on the market? Will be interesting. Competition is good.

 

[theeedude]

Impressive as a chip, not all that impressive as a platform. Wrong architecture (not-ARM) for the market, and too many chips. Not going to see wide adoption, mark my words.
It's Itanic for phones, that's what it is

 

[Munky]

As the article states, the reason they went with Moblin is to have an OS capable of utilizing the HW to its full potential. Nothing wrong with Nokia either, I have one and would love to have a phone capable of running x86.

I've been speculating about something like this, good to see Intel moving forward in that direction.

 

[Deeko]

Impressive as a chip, not all that impressive as a platform. Wrong architecture (not-ARM) for the market, and too many chips. Not going to see wide adoption, mark my words.
It's Itanic for phones, that's what it is

Its success will depend on Nokia's success with Maemo/MeeGo phones. Nokia is still a massive name in the worldwide phone market - if they can actually create a solid platform that uses Moorestown, and the devices take off, then you might see the other OS/manufacturers open up to the idea.

 

[theeedude]

Why? X86 is an ancient ISA whose main appeal is that Windows apps were compiled to it. These phones won't be running windows anyways, so being X86 in ARM dominated space just means they aren't going to be compatible with most apps and OS's in their market.

 

[Munky]

Most apps and OS's aren't cross-compatible either. Sure, Apple, Nokia and HTC use ARM processors, but an app built for the Apple is not going to run on Nokia or HTC in any case.

 

[tatteredpotato]

Why? X86 is an ancient ISA whose main appeal is that Windows apps were compiled to it. These phones won't be running windows anyways, so being X86 in ARM dominated space just means they aren't going to be compatible with most apps and OS's in their market.

I suppose Intel should have just built an architecture from the ground up for the mobile space since there is not back-compatibility to maintain with existing devices.

Another, although very unlikely move, would be for Intel to pull a Qualcomm and license an ARM ISA and build their own SoC implementation... I'm sure Intel has the engineering expertise to differentiate their products from the competitors.

I guess I say unlikely, but it's no more absurd than AMD making x86 processors, however I'll concede that the smartphone market isn't near the size of the desktop market (and therefor an architecture change wouldn't be all that unlikely).

Now that I think about it, Android could probably pull an architecture change rather easily, seeing as the bulk of application code is written in Java, and only the apps that dive down into the NDK would really need any different binaries.

 

[Fox5]

I suppose Intel should have just built an architecture from the ground up for the mobile space since there is not back-compatibility to maintain with existing devices.

Another, although very unlikely move, would be for Intel to pull a Qualcomm and license an ARM ISA and build their own SoC implementation... I'm sure Intel has the engineering expertise to differentiate their products from the competitors.

I guess I say unlikely, but it's no more absurd than AMD making x86 processors, however I'll concede that the smartphone market isn't near the size of the desktop market (and therefor an architecture change wouldn't be all that unlikely).

Now that I think about it, Android could probably pull an architecture change rather easily, seeing as the bulk of application code is written in Java, and only the apps that dive down into the NDK would really need any different binaries.

Intel gets to leverage its strength in x86 designs by using x86. Not to mention the basic atom architecture isn't targeted at JUST smartphones, they're getting a lot of additional sales by basically owning the netbook market, as well as the low power market, and they're creating a reliance (in the smartphone market) on their already dominating architecture. Plus, they seem to have the performance advantage, and once the processes scale enough, they can start putting real x86 cpus into phones.

Intel had an ARM lineup before (xscale). It was nice, but not particularly outstanding. They had worse power consumption than competitors, slightly higher per mhz performance due to aggressive pursuit of faster memory buses, and clocked considerably above competing designs at the time. They also had wMMX before multimedia extensions/coprocessors were all that popular and were the first design to implement powervr graphics. Now Marvell owns the xscale lineup, and they're not even a competitor in the high performance market anymore. x86 development benefits from intel's considerable size, scale, and focus, but alternative architectures at Intel (ARM, Itanium) haven't seen the same benefits.
Even Atom, a much more significant product line up, isn't receiving the same kind of love as Intel's normal x86 processors. Atom could easily be higher performing (don't cripple 64 bit support, drop the FSB architecture, don't pair it with outdated chipsets, add SSE4), and it is still trailing on fabrication tech.

 

[Bateluer]

Anyone else catch the snippet in Anand's article about the OS being nonbranded? I, for one, don't want the carriers to hold the power again. Right now, the consumer holds the power, as it should be. If we start seeing Moorestown/Moblin devices from carriers that only allow custom applications from AT&T/Sprint/Verizon, disallow 'unapproved' content, etc, there is no way in hell I'd buy one. In that situation, I doubt enough intelligent modders and coders would pick the devices up to give it a healthy development community.

 

[Deeko]

Anyone else catch the snippet in Anand's article about the OS being nonbranded? I, for one, don't want the carriers to hold the power again. Right now, the consumer holds the power, as it should be. If we start seeing Moorestown/Moblin devices from carriers that only allow custom applications from AT&T/Sprint/Verizon, disallow 'unapproved' content, etc, there is no way in hell I'd buy one. In that situation, I doubt enough intelligent modders and coders would pick the devices up to give it a healthy development community.

Well, that's Moblin. A manufacturer needs to make the Moblin-based phone that the carrier can then screw with. I personally don't see Moblin really taking off on its own - I'd say MeeGo has a much better chance of being the "flagship" OS of Moorestown, and Nokia historically isn't a carrier whipping boy...hell, at least in the US they usually don't even sell their smartphones carrier branded.

 

[Pliablemoose]

Hmmm, good read, Anand seems impressed, and he's got a hell of a lot better understanding of CPU's than I do.

Apple is locking down an ARM solution, and it seems that the whole industry will be moving away from it for a number of reasons.

Looks like there's room for another player on the field in the very near future.

 

[theeedude]

IF there is going to be this hardware fragmentation on Android, where some are running IA, some are running ARM, and maybe even some running MIPS, middleware like Flash and Java is probably not going anywhere.