View Full Version : Nokia + Intel - Does that mean we're going X86?
Will Nokia's next generation of computer-phones be powered by an X86 architecture?
Is this possible considering the complexity of X86 vs. ARM?
daemonfin
2010-04-22, 07:41
http://meego.com/about/faq
Ok, of course it will support X86 since it will be also released on PCs and laptops.
But as far as I know the next N### device will use an Intel CPU, will it be X86?
cashclientel
2010-04-22, 08:52
No x86 on mobile handsets for at least a few generations.
jcompagner
2010-04-22, 09:02
Ok, of course it will support X86 since it will be also released on PCs and laptops.
But as far as I know the next N### device will use an Intel CPU, will it be X86?
if it is intel then it will be x86
intel dropped all arm support for quite some time ago (they sold it)
But i can't believe that we will get an atom based N900 successor. That will be still a ARM based hopefully a Cortext A9 (dual that would be really great) but it seems that the successor is already pretty soon so i guess it will be a A8 then i do hope that they choose the 3630 instead of the 3430 of TI OMAP's
un-named_user
2010-04-22, 09:08
The next Maemo6/Harmattan/Meego1.0 by Nokia is going to use an ARM chipset only, specifically the OMAP3/Cortex A8 series.
This was confirmed here.
http://www.slideshare.net/peterschneider/maemo-6-technology-highlights
What the generation after that will use, nobody can say.
Okay, so MeeGo supports BOTH the x86 and the ARM architecture officially.
And we already know that the next Maemo device (powered by Harmattan, which will also be called MeeGo at the time) will use an OMAP 3 chip very similar to the N900.
No x86 on mobile handsets for at least a few generations.
Although ARM is improving, I still believe it's dying a slow death. Everyone knows that Intel and AMD have monopolized the CPU patent world.
Also regarding nanometers, ARM design always seems to be one step behind.
Cortex A8 has 45 (so does my core 2 duo but it's anything but new) while i7 is built on 32 nanometers and the technology for 20 nanometers has already been announced.
The fewer the nanometers, the lower the power usage.
Also this would be a great step for cross platform applications and we could install the same .rpm on our PC as in our mobile device.
IMHO this is the future, I can only hope Intel and nokia will make it happen a lot faster :D
If the rumor that ARM will be bought out by Apple plays out, I'm sure the Nokia-Intel coalition will go 100% x86.
WOW let them buy it and go down the way they did with motorola CPUs! :D:D:D:D:D
More X86 for me :p
jcompagner
2010-04-22, 09:48
Cortex A8 has 45 (so does my core 2 duo but it's anything but new) while i7 is built on 32 nanometers and the technology for 20 nanometers has already been announced.
The fewer the nanometers, the lower the power usage.
its not just about the nanometers, its the whole design,
ARM always is optimized for mobile usage, coming from the mobile parts.
Intel just tries to go from desktop/laptop chips to mobile, so the other way around.
And they are far of yet in the lower power area. Besides that the current atoms (that where released in Jan 2010) are still on 45nm! not 32...
So that far ahead isn't intel for there atom parts.
The only thing is that the current cpu in the N900 is made in 65... (34xx) so yes thats why i hope that the next NXXX will be a 36xx because that one is in 45nm (and a bit higher clocked by default)
But still get a current Atom and run it on 900mhz then get the CPU of the N900 and run that one on 900mhz. Then compare the power consumption i think it is still way way apart, completely different league. And i do believe that the atom is even faster..
and for some info about the A9:
http://www.arm.com/products/processors/cortex-a/cortex-a9.php?tab=Performance
so a dual core at 2GHZ the total power is 1.9W
Perhaps not the atom. I remember reading about a new CPU type specially designed for handsets but I have no idea where.
Anyway the atom is not doing that bad either: http://www.notebookcheck.net/Mobile-Processors-Benchmarklist.2436.0.html?&3dmark06cpu=0&archive=0&benchmark_values=&cinebench10_m=0&cinebench10_s=0&deskornote=0&dhrystone=0&month=&or=0&search=&sort=tdp&superpi1m=0&superpi2m=0&superpi32m=0&sysmark2004se=0&type=specs&whetstone=0&wprime_1024=0&wprime_32=0
The Z550 @ 2GHZ 45nm and 2W but again... single core.
In what concerns resourcefulness I trust intel.
On the other hand I have never considered an ARM as a CPU - more like a chip if you ask me.
I remember there was a time when I had some black intel chip that used to be cooled by two mini fans placed sideways (pentium1 perhaps?) put that was a long time ago.
Now I'm rather used with tiny shiny silicon capsules that have "diffused in france" written on them.
Hence diffusing (some complicate process involving a silicon dye and a nuclear reactor) and other complicated processes I read about make me believe that X86 must be something far more advanced than ARM.
I'm willing to bet on intel since they've always had the monetary resources to come up with the best CPU there was - as far as power management goes, we can just wait and see.
http://flowtv.org/wp-content/uploads/2009/10/rockstars.png
It is only natural that in some point, mobile phones CPUs will be based on x86/x64 architecture - in fact, I think the only thing that separates us from having a Core i3 or i5 cell-phone is probably batteries...
As soon as there will be a breakthrough with mobile power sources (which - as far as I know - is a field which progresses VERY slow), I'm more than certain that we'll see a major switch from ARM to x86 based CPUs.
sjgadsby
2010-04-22, 10:50
I'm willing to bet on intel since they've always had the monetary resources to come up with the best CPU there was...
By what criteria are you ranking procssors?
I'm taking your "Intel CPU" to equal "x86", as that's the common usage, but ARM processors outnumber x86 processors in terms of units shipped and units in active use. Furthermore, as architecture goes, there have been a number of competitors to x86 that have been hailed as more modern, more advanced, faster, etc. In fact, it seems most competing CPU architectures are considered "better" than x86...in everything but running legacy x86 code. And even there, it was AMD who developed x86-64, not Intel.
True, and it was AMD that broke the 1GHZ barrier while intel tagged along - but even if later they always managed so far to come up with the superior technology (however power was not as constraining as it is now).
I can't really say If 1ghz ARM = 1ghz X86 since we have no objective testing methods.
But what I'm sure of is that ARM will never best the "i" series and if they'll ever do, there's already gonna be i2.
Regarding power management, I think it's CPUs that should adapt and not energy sources. You always have to think extreme and extreme means doing a lot with little power.
chowdahhead
2010-04-22, 11:20
Perhaps not the atom. I remember reading about a new CPU type specially designed for handsets but I have no idea where.
Anyway the atom is not doing that bad either: http://www.notebookcheck.net/Mobile-Processors-Benchmarklist.2436.0.html?&3dmark06cpu=0&archive=0&benchmark_values=&cinebench10_m=0&cinebench10_s=0&deskornote=0&dhrystone=0&month=&or=0&search=&sort=tdp&superpi1m=0&superpi2m=0&superpi32m=0&sysmark2004se=0&type=specs&whetstone=0&wprime_1024=0&wprime_32=0
The Z550 @ 2GHZ 45nm and 2W but again... single core.
In what concerns resourcefulness I trust intel.
On the other hand I have never considered an ARM as a CPU - more like a chip if you ask me.
I remember there was a time when I had some black intel chip that used to be cooled by two mini fans placed sideways (pentium1 perhaps?) put that was a long time ago.
Now I'm rather used with tiny shiny silicon capsules that have "diffused in france" written on them.
Hence diffusing (some complicate process involving a silicon dye and a nuclear reactor) and other complicated processes I read about make me believe that X86 must be something far more advanced than ARM.
I'm willing to bet on intel since they've always had the monetary resources to come up with the best CPU there was - as far as power management goes, we can just wait and see.
http://flowtv.org/wp-content/uploads/2009/10/rockstars.png
That number is for the cpu only and doesn't include the gpu, memory controller, and so forth. The whole beagleboard consumes about the same as that atom cpu alone. If the 28nm SOI deal between ARM Holdings and Global Foundries happens, then Atom will lose its's fabrication process advantage and possibly it's performance lead as well. It's hard to envision Atom catching up to ARM in efficiency, but ARM is closing the gap in software performance; and when you acount for the hardware DSP's, one can argue that the arch is already there.
attila77
2010-04-22, 11:25
Guys. There is much more to embedded devices than just lowering CPU wattage, slapping it in a Phone and boot. Take for example system-on-chip designs. What made ARM ubiquitous in the embedded arena exactly is that it was so modular - every manufacturer rolled their own combo of CPU/GPU/DRAM/Flash/whatever which allowed for a very cost efficient solution for the particular task. Intel doesn't do that yet, they are currently only in the phase of pushing the GPU in, but a modern ARM chip is really a SYSTEM on a chip, so when you compare wattage and cost with an OMAP3 it is a total of north+southbridge+DRAM+FLASH+video stuff+etc.
Second, Intel needs to address idle consumption. Again, this is something historical, the original PC design (as that's actually what people actually mean with X86) was never meant to really be idle, and the hacks/tricks to save energy are just that - hacks and tricks as they had to preserve a good deal of backward compatibility. This is also the (hardware part of) the reason why a N810 can spend more time idling AWAKE than most notebooks in stand-by mode. X86 *can* be made more efficient, but at the cost of losing compatibility. But if you lose that, what is the point of X86 in the first place ?
Third, the ARM core is, even after decades of improvement, pretty small and this reflects well on die sizes. Due to historical baggage, this cannot be said about the X86, it's huge compared to an ARM chip (plus, the reliance on bigger L1/2 caches and long(er) pipes does not help). So while the benchmarks look good for Atoms, ARM based solutions are still pretty much ahead when it comes to performance/watt (or even performance/cost). You're basically sticking a revamped truck engine in a sportscar and hoping the brute HP will pull you through. This sounds pretty much like a Dodge Viper, cool in demos, but keep near gas stations and a healthy reserve on your credit card ;)
@sjgadsby
Are you judging this from an objective view point or rather by the fact that microsoft is X86 based and further development of ARM technology would give Linux a fair chance regarding OS share in a world dominated by microsoft?
Also the question to answer is in fact RISC(ARM) vs. CISC(x86) if I'm right?
By my readings so far it is obvious CISC got a great head start because of the early adoption of X86 and cheaper manufacturing and especially cheaper manufacturing line upgrades.
Regarding performance it's really hard to draw the line, this is a little something I found: http://www.pocketables.net/2008/10/mid-battle-aigo.html
Icyseanfitz
2010-04-22, 11:37
ive heard from a couple of different people that there is breakthrough coming in the battery division has anyone here heard similar things or have any info on this, this would seem to be the holy grail for mobile devices
If the rumor that ARM will be bought out by Apple plays out, I'm sure the Nokia-Intel coalition will go 100% x86.
Suddenly this doesn't sound that good anymore...
Sure, then it will be Intel, ATI-AMD, Nvidia, Nokia and Microsoft which is enough to dominate the market.
But if I'm to look at the cash X86 swallowed vs ARM I believe Atoms performance should not be as questionable as it is right now.
I wonder if RISC (ARM), was developed on so hard as Amd and Intel developed X86, could compete in the high end high performance segment (PCs, servers)
ive heard from a couple of different people that there is breakthrough coming in the battery division has anyone here heard similar things or have any info on this, this would seem to be the holy grail for mobile devices
Breakthrough in battery is not the solution.
It's who does more with the same amount of mAh.
Suddenly this doesn't sound that good anymore...
Lol, took you awhile to realize that ;)
Sure, then it will be Intel, ATI-AMD, Nvidia, Nokia and Microsoft which is enough to dominate the market.
Dominate what market?
But if I'm to look at the cash X86 swallowed vs ARM I believe Atoms performance should not be as questionable as it is right now.
I wonder if RISC (ARM), was developed on so hard as Amd and Intel developed X86, could compete in the high end high performance segment (PCs, servers)
I don't think Intel/AMD can outrace ARM (in the race-to-idle) anytime soon, so if Apple buys them out, then they'll be collecting (or increasing) the taxes while getting a bit of privy information of their competitor's potential devices roadmap.
It's not all about cash. Look at what Microsoft has been doing with all their money lately.
ive heard from a couple of different people that there is breakthrough coming in the battery division has anyone here heard similar things or have any info on this, this would seem to be the holy grail for mobile devices
There's always a breakthrough coming in the battery division. I'm in materials and you can't throw a stone here without hitting someone who is working on some cutting-edge battery technology. The media is always catching wind of some laboratory-scale miracle device that for technical or practical reasons never sees the end of an assembly line.
There's no reason to believe there will be a revolutionary improvement in battery technology anytime soon. We are already using one of the most efficient chemistries possible (lithium *), and the possible improvements to overcome materials limitations are generally evolutionary rather than revolutionary.
Back on topic: i sure hope not. Even more than capacitive screens, it would represent a change based on management decisions rather than good sense.
Dominate what market?
The smartphone/computer phone market the same way X86 has dominated the PC market.
Also as far as I know has ARM SoC been used so far anywhere else than in the iPad? (dumb question maybe but I was curious to see what CPU they re-branded this time so I read about the big SoC bla bla)
Intel for an example also entered the SoC race.
It's not all about cash.
IMHO cash is the only thing that allowed Intel to outrun AMD (just consider the innovation amd brought - X64 and breaking the ghz barrier - how else could intel have prevailed if not my investing billions into research?).
And in the end it's not about the race to idle or about the one with the greatest potential. It's about who makes himself popular first.
Just as RISC vs CISC.
Probably if RISC would have been adopted first, our computers would be twice as powerful and twice as green...? Who knows...
It seems to me that people (I am) are more likely to embrace X86 since it's wider spread largely adopted and has a huge application base (makes it significantly easier for developers or do I get the wrong idea?) rather than switch to a whole different standard.
Mobiles are regarded as non standard systems incompatible with PCs so ARM will never make an impact.
But the second I copy the Pidgin.rpm (I just installed on my pc) and install the same package on my phone, there's a whole different story.
I know linux is ARM compatible but that won't change the fact that the world is 90% X86-X64 and reluctant to change.
I don't think the best technology will decide.
I only hope Intel will make a huge step forward to become the best technology.
This thread is revelatory(to me):rolleyes:
i sure hope not. Even more than capacitive screens, it would represent a change based on management decisions rather than good sense.
I don't pretend to fully grasp the essence of CPU architecture.
But leaving aside the power factor, Wouldn't X86-X64 represent a revolution for developers and enable them to create applications for PC and mobile devices at the same time?
Wouldn't it ease the development process and create an unified application ecosystem?
attila77
2010-04-22, 14:17
But leaving aside the power factor, Wouldn't X86-X64 represent a revolution for developers and enable them to create applications for PC and mobile devices at the same time?
Wouldn't it ease the development process and create an unified application ecosystem?
No. It's WAY easier to go from Linux X86 to Linux ARM than Linux X86 to Windows X86. It's the OS/tools that define the ecosystem, not the HW.
But isn't it even harder to go from Linux ARM to windows X86?
And it looks like LG GW 990 will use some SoC Atom and meego...
Also if ARM is "buyable" why doesn't nokia make the move?
attila77
2010-04-22, 14:27
Actually, no, it's about the same difficulty - you start from scratch in both cases :)
As for LG, nobody can 'prevent' you from doing it, maybe it will be real snappy, but will still be a comparably power hungry brick.
why x86 when we could have x64? :<
not that we need that necessarily on a mobile device, but why not...?! ^^
@msa
X86 X64 same thing at the moment, I don't think you can have one without the other ;)
Actually I think it's that long (GW 990)(seriously it's the longest phone I've ever seen, you could probably wrap it around your chin and it would still reach your ear:p) just to have room for the battery.
So from a linux point of view the architecture is not as important as the power/performance ratio of the CPU and since microsoft doesn't have an ARM solution, ARM notebooks and laptops would allow linux to be more widely spread.
But... wouldn't it have been more important for nokia to build a stronger connection with ARM than with intel? I know that this partnership has unified (will unify) the mobile linux into one powerful distribution, but is that more important for nokia than releasing powerful devices (power/cpu/formfactor?)
Maybe they made this choice knowing that intel has a ace up their sleeve?
Laughing Man
2010-04-22, 14:45
Lol, took you awhile to realize that ;)
Dominate what market?
I don't think Intel/AMD can outrace ARM (in the race-to-idle) anytime soon, so if Apple buys them out, then they'll be collecting (or increasing) the taxes while getting a bit of privy information of their competitor's potential devices roadmap.
It's not all about cash. Look at what Microsoft has been doing with all their money lately.
Sounds to me if Apple bought ARM the DOJ would soon get involved to break it apart.
Also as far as I know has ARM SoC been used so far anywhere else than in the iPad? (dumb question maybe but I was curious to see what CPU they re-branded this time so I read about the big SoC bla bla)
Yes, in every modern smartphone including the N900.
IMHO cash is the only thing that allowed Intel to outrun AMD (just consider the innovation amd brought - X64 and breaking the ghz barrier - how else could intel have prevailed if not my investing billions into research?).
IMHO it's Intel's anticompetitive practices that went on for years and for which they were fined 1,2 billion euros by the EU that allowed them to outrun AMD. :)
And in the end it's not about the race to idle or about the one with the greatest potential. It's about who makes himself popular first.
An ARM processor is used in pretty much every single phone(smart and dumb) and other electronic gadget you can ever think of that isn't a PC. ARM already utterly dominates the embedded market, it's Intel that's trying to squeeze in.
An ARM processor is used in pretty much every single phone(smart and dumb) and other electronic gadget you can ever think of that isn't a PC. ARM already utterly dominates the embedded market, it's Intel that's trying to squeeze in. .
Yes but most users don't see the "ARM inside" logo so what do they care? The intel logo makes an impact - it's advertising and it's on every second PC.
So still I ask my question:
Why did Nokia choose intel? - I don't think that unifying the mobile linux distributions was THAT important to them or that Moblin represented any competition.
Sounds to me if Apple bought ARM the DOJ would soon get involved to break it apart.
It wouldn't even get that far.
So still I ask my question:
Why did Nokia choose intel? - I don't think that unifying the mobile linux distributions was THAT important to them or that Moblin represented any competition.
Joining with Moblin wasn't about reducing compeition-- it's about increasing scale. Of course unifying the distributions was important.
Yes but most users don't see the "ARM inside" logo so what do they care? The intel logo makes an impact - it's advertising and it's on every second PC.
The brand doesn't run your devices, the CPU does.
So still I ask my question:
Why did Nokia choose intel? - I don't think that unifying the mobile linux distributions was THAT important to them or that Moblin represented any competition.
1) That way they don't have all their eggs in one basket. Their partnering with Intel doesn't mean they'll stop using ARM.
2) Intel needs to get into embedded/mobile badly and is willing to put considerable effort in software and hardware platform development and has a long history in open source. Moblin and Maemo were very similar so it makes a lot of sense to get together. Both win.
3) ARM only designs CPU's, then licence them for manufacturers to build and that's it. What would Nokia get from partnering with them directly?
Yes but most users don't see the "ARM inside" logo so what do they care? The intel logo makes an impact - it's advertising and it's on every second PC.
Don't make the mistake of equating the mobile market with commodity PC machines.
Don't make the mistake of underestimating ARM in the embedded market... they cover some 70+% of the marketshare.
So still I ask my question:
Why did Nokia choose intel? - I don't think that unifying the mobile linux distributions was THAT important to them or that Moblin represented any competition.
Some say they're the odd couple out without a significant marketshare in the smartphone\highend phone segment... plus, I guess they have some similar strategies that can be aligned (both use open source community to further their business, etc.)
3) ARM only designs CPU's, then licence them for manufacturers to build and that's it. What would Nokia get from partnering with them directly?
Well... that in the context of apple buying them - that would certainly create a loss.
Anyway the rumors of ARM being bought by apple freak me out.
Brand is very important. Most basic users have no idea of what ARM is. If you'd present them a high end ARM notebook and a low end Intel, most of them would choose the one they've heard of.
But I guess you are right about not comparing the PC market with the mobile market and manufacturers will choose X86 when/if it will be "ripe".
I'm glad to be using linux and being able those whenever I want from whatever I want (ARM/X86) :D
tumblebobm
2010-04-22, 22:57
Personally I don't think x86 could replace ARM/RISC now or in the future for mobile devices.I won't say can't or never thats just asking to be proved wrong. CISC is just complete overkill imho.
ARM works well, considering its in the iPad and is standard for mobile device manufacturers (people don't know or care for the most part) its dominance isn't going to change for a time and of course there'll be plenty of improvements.
This is being typed from a netbook with a VIA-C7 which isn't perticularly great in my opinion. My sisters netbook with an Atom isn't much better really but thankfully using a form of linux means a lot of these factors can b ignored. it would take way too much effort i think to modify these for a smart phone.
For some extra information what most people call x64 is really supposed to be called x86-64 meaning that its like an extended version of x86, details are a bit tricky. We haven't moved over to pure x64 in the most part (cept for servers and supercomputers i think) as all the old OS and programs would need remade (don't know if completely rewrote or just modified and compiled - this is didn't need to know for my exams so I don't remember exactly) Pure x64 supports massive amounts of ram etc which isn't possible yet.
Wether the the OS used is 8, 16, 32, 64 bit depends purely on the cpu used. I hope I'm not talking down to any one here, and i'm new, but it seems like this is stuff that folk don't know and i think is interesting and useful.
The brand doesn't run your devices, the CPU does.
3) ARM only designs CPU's, then licence them for manufacturers to build and that's it. What would Nokia get from partnering with them directly?
The Nokia Rapido processor is an ARM. So, Nokia has a partnering with ARM.
http://pdadb.net/index.php?m=cpu and find "Nokia".
At this moment, the only x86 device from Nokia is the booklet.
Also regarding nanometers, ARM design always seems to be one step behind.
Cortex A8 has 45 (so does my core 2 duo but it's anything but new) while i7 is built on 32 nanometers and the technology for 20 nanometers has already been announced.
The fewer the nanometers, the lower the power usage.
That has nothing to do with the architecture -- Intel's superior process scaling just means that Intel is great at process scaling, but they could do equally well shrinking ARM chips if they still made them. Not that this matters, in the end -- Intel doesn't make ARMs, and TI doesn't make Atom SoCs, so the question is how their products actually compare, not how they ended up that way.
And ARM already wins big in performance/watt, regardless of whether it's "a step behind". (Can you imagine what a 20nm ARM would do? :D)
Also this would be a great step for cross platform applications and we could install the same .rpm on our PC as in our mobile device.Except that no corporation wants desktop software running on mobiles, because the UX is horrible with unoptimized software on touchscreens, QWERTY thumbpads, and the like. They just don't see the added functionality as a selling point. So as nice as it might be to power users, it doesn't enter into any of the decision-making that matters (barring some Pandora-like team building a hobbyist mobile based on x86).
Moreover, I've only ever had occasion to use a handful of x86 binary-only software on Linux -- Skype, Flash player, and Google Earth are the only apps that come to mind, along with a few closed hardware drivers (which are platform-dependent anyway). Right now my main desktop is a pure x86_64 system, with no 32-bit compatibility layer, and I'm doing fine with none of those. So I don't really see the big advantage for cross-platform binaries -- source-compatibility is enough.
@sjgadsby
Are you judging this from an objective view point or rather by the fact that microsoft is X86 based and further development of ARM technology would give Linux a fair chance regarding OS share in a world dominated by microsoft?Not to answer for sjgadsby, but his statements are objectively based. Everyone who knows anything about CPU architecture (including CISC fans -- x86 is quite possibly the ugliest CISC in the world) agrees that x86 is a huge bag of fail, but since backwards compatibility rules the market, it keeps going.
Also the question to answer is in fact RISC(ARM) vs. CISC(x86) if I'm right?No, that's the question to argue endlessly about. We will always have RISC and CISC chips, they will always be reasonably competitive, and the argument will never be settled. :(
Certainly, if the RISC/CISC war were to be won (by RISC, obviously (me? biased? :cool:)), that would settle the x86/ARM battle for high-powered mobiles, and indeed everywhere else. But since that won't happen, x86/ARM will happen on its own -- since nobody codes assembler, and compilers optimize code beyond recognition anyway, ideology will prove subservient to practical effects (which is faster? which uses less power? which is compatible with relevant binaries?)
By my readings so far it is obvious CISC got a great head start because of the early adoption of X86 and cheaper manufacturing and especially cheaper manufacturing line upgrades.Then your reading is woefully inadequate. x86 is one specific CISC family, and it has a head start for certain markets (starting with desktops (i.e. the IBM PC) and scaling upwards to servers), sure. It has almost no share in the handheld market -- an incursion here or there (HP LX100 anyone?), including Intel's current push, but as yet no long-term presence. Their success on the desktop was directly tied to the success of the IBM PC architecture, which in turn succeeded not due to any particular technical win, but because of the PC clones -- other systems using different CPUs never got that level of commoditization, so they got squeezed out.
But that says nothing to the overall CISC/RISC war, and RISC (and more specifically and relevantly, ARM) has an equally big head start in the handheld market. Which just might say more about which one will win the handheld market.
The last bit is just plain wrong -- RISC are simpler, therefore (all else being equal) cheaper to manufacture.
attila77
2010-04-23, 08:53
Also regarding nanometers, ARM design always seems to be one step behind.
Cortex A8 has 45 (so does my core 2 duo but it's anything but new) while i7 is built on 32 nanometers and the technology for 20 nanometers has already been announced.
Ah, that one is slightly misleading, but easy (though you have to think business). The biggest gain is on the biggest chips, that's why Intel moves mid-to-high end chips there first (yeah, lowering Atom power consumption by 10-20% is cool, but making 15% more i7-s for the same amount of money is a LOT better business-wise as the profit margin on the i7 is significantly higher than on an Atom).
Take a look at the Atoms - not even the newest spanking N4xx/D5xx series is 32nm, they're still good ole 45nm (and even the next-gen, Moorestown, is 45nm). In fact, with OMAP4 and it's Cortex-A9 kin coming this year, ARM will actually be AHEAD even in manufacturing (as the first 32nm Medfield Atoms will arrive only in 2011)
Also this would be a great step for cross platform applications and we could install the same .rpm on our PC as in our mobile device.
Say, something like what Easy Debian (http://wiki.maemo.org/Easy_Debian) already does ? ;) But seriously, Linux already does that pretty painlessly.
IsaacDFP
2010-05-09, 19:31
I really don't know much about cpu architectures... but I don't understand why ARM is still alive. It's almost orgasmic to me thinking an Intel chipset could be use inside a mobile device (such as the N920+), lol, so why isn't it happening? Is the only reason ARM is still being used is because corporations can still squeeze a few profits out of it?!
I really don't know much about cpu architectures... but I don't understand why ARM is still alive. It's almost orgasmic to me thinking an Intel chipset could be use inside a mobile device (such as the N920+), lol, so why isn't it happening? Is the only reason ARM is still being used is because corporations can still squeeze a few profits out of it?!
A very rough example. You have a laptop and an N900? Let's assume you browse the web on your laptop and your N900 until the batteries run flat. You got maybe six hours of use on both of them. Now take out the batteries and compare them side by side. Imagine your laptop's battery being the size of the N900's battery, how long would it last? That's why ARM is used.
IsaacDFP
2010-05-10, 00:14
A very rough example. You have a laptop and an N900? Let's assume you browse the web on your laptop and your N900 until the batteries run flat. You got maybe six hours of use on both of them. Now take out the batteries and compare them side by side. Imagine your laptop's battery being the size of the N900's battery, how long would it last? That's why ARM is used.
Very good exemple of an answer, thanks. So by that anology, I understand Power Consumption is one of the main different points. So what about the day we will see fuel cell batteries, would you say ARM would still be used?
To be competitive, ARM RISC is increasing instruction sets while x86 is becoming more and more efficient, effectively leaning against each other. At some point, they will be roughly equivalent, and then x86 will kick asterisk because the compatibility base for x86 is huge.
It's not like ARM will ever die because they do small chips by design while x86 is large by design. ARM will always power small devices, routers, PNAs, etc. x86 will keep on powering behemoth desktops and slim down, taking the anything-larger-than-a-pad market. There's a limit between them, and that limit is moving down. If it reaches the size of slightly-larger-than-N900, we're golden.
Finally, no "app base".
Oh, and it's entirely possible that batteries will never be large enough to accommodate this. We may move forward (laterally?) to supercapacitors, which, even though less than a Li-Ion, can be charged at about any rate. The fact the battery dies in 6 hours is not going to matter if it can be recharged in 10 seconds.
Already proof-of-concept hybrid batteries can be charged in a minute or two. In which case, charge all the time is going to become as common and annoying as the bluetooth headset. Also, wireless power would be nice.
attila77
2010-05-10, 01:56
To be competitive, ARM RISC is increasing instruction sets while x86 is becoming more and more efficient, effectively leaning against each other. At some point, they will be roughly equivalent, and then x86 will kick asterisk because the compatibility base for x86 is huge.
Note that they will never be equivalent. ARM is far more modular and simpler in design, so unless something radical happens in X86 land (which is very unlikely), it will be cheaper and more efficient for years to come. The compatibility base is IMO overrated (what *exactly* do you expect to be able to do on a X86 phone that you weren't able to do on ARM, considering we're talking about Linux ?). The only place where there is a slight gain is the development, but with tools like the Qt SDK, even that advantage is fading.
Don't you guy's think the new atom z6xx is good for larger 7-10 inch tablets?
Equivalent, no but similar. And when similar enough it makes little sense to not go full monty.
Also, WE are not talking about Linux. We talk about x86, meaning I get to choose the platform, meaning *I* will go with Windows for the app base and my ability to build for the platform with a RAD IDE on a known language.
Might not be a big deal for Linux, but for other OSs it will be gold. Especially open/free Windows.
attila77
2010-05-10, 11:57
I still don't get what the advantage is, even with Windows ? Today, you have WinMo on which you already use a standard RAD IDE (VisualStudio et al) and it supports multiple architectures. X86 compatibility will not solve any of the UI problems Windows apps face when running on small/touchscreens, and stock win X86 libs/GUIs are not written with mobile (power constrained) use in mind, so, what exactly is the gain ?
nwerneck
2010-05-10, 12:16
I like the idea of using non-x86 processors, but it's sad that there are no good tools available. I still don't understand well how much is maemo using the capabilities of the OMAP processors, which have an extra DSP, and also if we are using the ARM cores thelves to their full extent (the NEON stuff, etc). And the worst is I don't feel like pushing these boundaries with the current development tools.
I bought my N800 knowing it was something for crazy bit-brushers, and I thought "hey, that's my thing!" But today I have the feeling one has to be a true uber geek to have his fun with it... And there's too many people wanting to make it a desktop in their pockets. That's wrong!
If you ask for a different processor, you are literally going against the _core_ of the tablets!
Also, WE are not talking about Linux. We talk about x86, meaning I get to choose the platform
That's assuming the vendor doesn't take action to lock the platform down. Remember, this is the mobile space which is still extremely end-user hostile.
meaning *I* will go with Windows for the app base and my ability to build for the platform with a RAD IDE on a known language.
Different is scary, I know.
Might not be a big deal for Linux, but for other OSs it will be gold. Especially open/free Windows.
There is no such thing as "open" or "free" Windows. Open and Free software on Windows, yes, but that defeats the point.
Having a non-x86 architecture in the mobile world frees us from having to constantly be backwards compatible, and an open source (preferrably Free Software) OS and applications negates the need to be.
That's assuming the vendor doesn't take action to lock the platform down. Remember, this is the mobile space which is still extremely end-user hostile.
It would make little sense to lock down a hardware platform, limiting your user base.
Different is scary, I know.
There's no reason to blindly poke. Allow me to clear things up. First, there aren't many mechanics that used to be brain surgeons. After decades (yes, I'm moving to plural in a couple of years) of developing for a platform using a tool, you never get the same proficiency again, lest I screw up my health again. Careers don't grow on trees.
Also, I see little reason to go "different" as long as my native platform is an option. Heck, if it's so fun and easy to just switch, come on over.
Throwing it up there that getting me out of my hole is going to make me wet myself is little more than flamebait.
A thanked post nonetheless.
There is no such thing as "open" or "free" Windows. Open and Free software on Windows, yes, but that defeats the point.
Hello. (http://www.reactos.org/en/index.html)
It would make little sense to lock down a hardware platform, limiting your user base.
Why not? Apple does it with their devices, every Android vendor does as well (to some extent.) Motorola goes the extra mile with every device but the Droid to the extent that unless you hack the bootloader you won't be able to load a 3rd party ROM.
if it's so fun and easy to just switch, come on over.
Perhaps that was too snarky. I don't advocate switching, I advocate expanding ones understanding to platforms and ways of doing things outside the normal purview. It's how creativity thrives. Tying ones career to a single platform would make me nervous, personally.
Hello. (http://www.reactos.org/en/index.html)
That's not Windows. It's trying to be Windows compatible, but the fundamental nature of Windows itself neither open nor free, and cannot be so long as it is wholly under the control of MS.
quipper8
2010-05-10, 17:26
I think the nokia + Intel partnership is great as it MIGHT be a step towards the days when we can just buy hardware and run whatever OS we want on it and connect to whatever wireless provider we want to.
I just bought a toshiba nb-305 with some new qualcomm chipset in it and it is an x86 atom processor and it has some qualcomm gobi chipset in it and I can connect to wifi or cdma or gsm networks. I think ti is f****ing brilliant. I mean, why not have something that can just connect to whatever?
Why not? Apple does it with their devices, every Android vendor does as well (to some extent.) Motorola goes the extra mile with every device but the Droid to the extent that unless you hack the bootloader you won't be able to load a 3rd party ROM.
a) Locking a platform people hack afterward isn't really locking
b) None of those are X86. It's easy to lock down a proprietary hardware with a proprietary OS. It's virtually impossible to do so with X86 lest you risk platform incompatibility. I very much doubt Intel cares who wins the OS wars, as long as they supply the ammo.
I don't advocate switching, I advocate expanding ones understanding to platforms and ways of doing things outside the normal purview. It's how creativity thrives.
That's not Windows. It's trying to be Windows compatible, but the fundamental nature of Windows itself neither open nor free, and cannot be so long as it is wholly under the control of MS.
Fundamental shmundamental. It is compatible with Windows binaries. It is compatible with Windows UI. Save for the logo on the back, it's open Windows. It walks like a duck, it quacks like a duck. It has the Windows Application base, and compiles and runs Windows IDEs. If I don't like Nokia's messaging app, I either switch or develop my own.
Also, it matters not if it's open Windows or Microsoft Windows. You should expand your understanding to platforms and ways of doing things outside the normal purview.
I hear that's how creativity thrives.
a) Locking a platform people hack afterward isn't really locking
But should they have to? I would argue that they should not and vendors should strive to give people the options, like Nokia and Google (at least with the Nexus One, which is basically flipping a switch.)
b) None of those are X86. It's easy to lock down a proprietary hardware with a proprietary OS. It's virtually impossible to do so with X86 lest you risk platform incompatibility. I very much doubt Intel cares who wins the OS wars, as long as they supply the ammo.
Intel, or other SoC vendors if they start integrating Atom, will happily provide security modules like ARM does and list it as a bullet point on their spec sheets. There is nothing about x86 that makes it immune to lockdown and as you said, Intel does not care who wins the OS wars so long as it runs on their chip.
Fundamental shmundamental. It is compatible with Windows binaries. It is compatible with Windows UI. Save for the logo on the back, it's open Windows. It walks like a duck, it quacks like a duck. It has the Windows Application base, and compiles and runs Windows IDEs. If I don't like Nokia's messaging app, I either switch or develop my own.
I think you miss the point, namely pointing out that no matter what you do, MS can throw tacks in your path at which point you fix the flat or you diverge from the MS path.
Also, it matters not if it's open Windows or Microsoft Windows. You should expand your understanding to platforms and ways of doing things outside the normal purview.
With respect to keeping Windows, as a platform, in my sights I will stick with the core as run by MS as they are the sole leading force. Not to say that what ReactOS is doing is wrong or bad, it is an admirable effort; it is not where Windows is going but where it has been.
attila77
2010-05-10, 21:11
b) None of those are X86. It's easy to lock down a proprietary hardware with a proprietary OS. It's virtually impossible to do so with X86 lest you risk platform incompatibility.
This is completely bogus. Intel (like all major chipmakers) has a very serious investment in 'trusted/secure computing', has even worked with ARM on TrustZone in the XScale days, and is shipping with a lot of security enabled chipsets for years. The fact that this tech has not yet been widely used for lockdown is most certainly not the result of Intel fighting tooth and nail for your hack-rights.
http://en.wikipedia.org/wiki/Trusted_Execution_Technology
This is completely bogus. Intel (like all major chipmakers) has a very serious investment in 'trusted/secure computing', has even worked with ARM on TrustZone in the XScale days, and is shipping with a lot of security enabled chipsets for years. The fact that this tech has not yet been widely used for lockdown is most certainly not the result of Intel fighting tooth and nail for your hack-rights.
http://en.wikipedia.org/wiki/Trusted_Execution_Technology
I read that to the letter. I also added what I know about trusted execution. I see nothing keeping me from running Linux on such a chip if Intel intended it for Windows. It is, therefore, not "completely bogus".
The chip allows for a secure key storage in a way that the key never needs to be decrypted in-memory, the age old weakness of everything. What it does is store the keys for you.
In later implementations, it allows for separation of memory and registers per-core so that a process can't access another process' memory.
Thus far, I note the following:
a) It's an optional feature, allowing for better security
b) It prevents nobody from running nothing. It helps people run sensitive code without interference.
c) While it has its uses in DRM, such as allowing a player to decrypt data without someone stealing the keys or modifying the runtime, replacing a JNZ with a JZ, it does not disallow decrypting off-chip because the CPU knows not what you run.
d) I'll not even discuss DRM. It has been tried before a million times and it failed a million times. Intel is simply collecting on Hollywood and their wet dreams in the process of improving a platform.
e) The number of people running Linux, "modified" Windows, live systems, etc is huge. Even people with bought licenses need to run a live now and then. Microsoft patches stolen OSs. They know they're stolen and the leave it at that, because of several resons. You can't lock them out lest you enrage the whole lot of them. Or do you expect Joe Average to still buy your laptop knowing he can't run "THAT"? AMD would love to hear that. The maker offering an open system will be the preferred vendor.
I don't expect this technology to hinder anything major. And it definitely has no bearing on my OS of choice.
Also, I point out at this moment that a whole shovel of OS runners and precisely those who DO buy OSs are corporate users. And corporate users will NOT allow complex procedures on thousands of units. That's why a corporate disk has no serial number.
You don't lock these people out.
Overall, not worried in the slightest. I don't even expect to see a laptop or a laptop-smartphone hybrid that disallows an OS (save for lack of drivers). And if it does lock one out, by whatever reason, it won't be anything MS put out.
But should they have to? I would argue that they should
You would, if there were any takers. I'm not one of them. x86 was open this far and will be open IMO from now on. That's what pushed Intel forward, nice docs and heavy developer support. It's also what pushed ATI forward.
Remember SLI being the top dog? People bought nwhatitsname chipsets like nuts, even though they made the worse chips ever. You needed 2 nVIDIAs, a nSomething board, some drivers and (initially) a dongle of sorts. Specs were closed and only nVIDIA made the boards. Profitville!
ATI devised Crossfire (X) and opened the specs to the planet. Any Joe could build a board that supported it, moreover, the extended configuration doesn't even care what the chipset is. ATI, a division of AMD, opened specs to Intel. Now Intel makes the boards, we buy. Guess who's swimming in cash. (If you said Intel you get credit too).
Right now, PC/x86 architecture is the major platform because it's extensible and anyone can make any hardware for anything they want. I don't think anyone is that nuts as to strangle it.
I've been an Intel faithful client for many, many years, after Cyrix died. But if the idiots restrict my OS I'm jumping to AMD in 48 hours. I hear they overclock really well.
Intel, or other SoC vendors if they start integrating Atom, will happily provide security modules like ARM does and list it as a bullet point on their spec sheets. There is nothing about x86 that makes it immune to lockdown and as you said, Intel does not care who wins the OS wars so long as it runs on their chip.
Intel doesn't care who wins as long as someone does.
They may make locked an unlocked chips and charge for the premium. In the end, you'd have N900 and N900 unlocked, for a little more, as is the case with provider lock on phones now. No more.
If they are all locked, they limit purchases, because I might just not buy an N900 if I can't have Windows on it, and maybe you might just not buy it if it has no Linux. So, really, it's more profitable to make a flexible device than to lock it.
I have no doubt there's going to be nutcases like the fruit people. I doubt they'll ever win. If Nokia locks it, I'm switching the next day at 8 AM.
I think you miss the point, namely pointing out that no matter what you do, MS can throw tacks in your path at which point you fix the flat or you diverge from the MS path.
They always throw tacks behind them as a matter of course. But the point that ReactOS makes is that they can't throw tacks in the past. Once it's out there, it's out there. It takes time and effort to build an OS by clean room and specs, but MS can't really do much in a legal sense, and they can't back-revoke the documentation.
The only way for them to tack it is to keep pushing the envelope and make the next OS so much better that the old one will seem old. That will keep ROS back for a long time.
What I first started using ROS, it was an NT4 project. Then they moved to 2000, and now they have an XP/2003 target.
XP is 2002-2003 technology. ROS is behind, but XP is still popular, flexible, stable (with proper care) and widely used. It has the biggest ever driver base, and, as a platform, can be extended by 3rd party just as well as the original.
It will be a long time before ROS will be Windows 7. I know that. Do I think 300E for W7 is worth the luxury? Well, I do, because I bought one for my baby. I can't have water pipes and crossfire and go around with a stolen OS. But for other applications, heck yeah.
it is not where Windows is going but where it has been.
Indeed. And cheap old technology is always going to be a contender on the market. All really cool corporations still have a few 300 MHz PCs somewhere. They all have some Linux somewhere for a PC that does nothing but act as a router, or an internal HTTP, or a SMB, and simply doesn't justify price. Joking aside, no matter what you do, there's always going to be a terminal just barely worth the hardware, the occasional daemon, etc. These could go well with a simpler, older OS that just works.
And as soon as ROS is going to be close to XP, it's going to be a viable alternative.
NT4 still runs on a ton of machines, servers nobody dares touch, workstations, ATMs, etc. And if the darned thing still had a decent driver base it'd still be used.
I have no problem with a cheap router that is now designated a Linux machine actually running an OS I can administrate with zero effort. Plus, less gadgetry and less eye candy makes for a slimmer, faster OS. It's the best of both worlds.
x86 brings with it the ability to use wine. I don't see how you can downplay the usefulness of being able to access win32 programs.
What i find worrisome is AMD's reluctance to get into the SoC market.
@ndi
I'm not sure I'm really getting you here. You seem to be mixing a lot of unrelated things to the same discussion. What I still haven't figured out is why exactly would X86 be better on smartphones than ARM. You seem to be equating ARM = closed, X86 = open, while it's not the architecture that makes it or breaks it, it's the manufacturer who makes the product. Now mobile products have traditionally been more closed than PC products but it's not because they use ARM instead of X86.
Now let's assume there was an open X86 handheld and you could just drop Windows on it. What then? Desktop Windows isn't actually designed with mobility in mind, neither is the UI suitable for tiny screens that you could possibly only drive with your finger. Then drivers, what about touchscreens, accelerometers, proximity sensors, cellular and other wireless HW, cameras, DSP stuff etc. etc? Let alone software that would make use of those. It often seems to be the hardware drivers that make porting mobile OSes into other devices unfeasible, why would it be different with Windows? Why would anyone write such drivers for WinXP in the first place?
It just seems to me that there's much more to lose than to gain by running any desktop OS on a handheld device which again has nothing to do with X86 or ARM.
Why would anyone write such drivers for WinXP in the first place?
The logical route, to follow ndi (I think) is that if you were to use ReactOS you could apply existing windows drivers or simply write windows-style drivers for it, since the platform is open source.
In a sense, however, this route is flawed since you end up reinventing the wheel to chase an OS architecture you have no real control over for the sake of backwards compatibility, something even Microsoft is having trouble with to the point of virtualizing win32.
It'd still work, unless you went the Motorola route and signed your root file system and kernel, then used a chain of trust to lock the whole thing down (see everything but DROID.)
Give it a read to this:
http://talk.maemo.org/showthread.php?t=52074
What I got from both articles is that the new chips from Intel can beat without problems ARM devices in raw processing power, however, they will kill the battery very quickly (unless the systems is kept under control by the OS).
If they have a OS optimized for hand held use (like MeeGo is supposed to be) I for one would be very interested on this, because there would be a chance to unlock this "overdrive mode" for faster apps (and I would have the hand held plugged to a nearby power socket).
attila77
2010-05-11, 07:33
I read that to the letter. I also added what I know about trusted execution. I see nothing keeping me from running Linux on such a chip if Intel intended it for Windows. It is, therefore, not "completely bogus".
?
Trusted Boot (tboot) is a pre- kernel/VMM module that uses Intel(R) Trusted Execution Technology (Intel(R) TXT) to perform a measured and verified launch of an OS kernel/VMM.
Stick that in your BIOS and I can see it very effectively keeping you from running ’unwanted’ OSes. Besides, we don’t need to be theoretical, the XBOX was a widespread locked down X86 based platform way before these sophisticated security technologies.
The bottom line is that a functional equivalent of security features is just as present on X86 as it is on ARM, it’s just the MB/Computer manufacturer’s current business practice that keeps your freedom.
david.hicks
2010-05-11, 09:03
I have a question for the people that want mobile X86.
My background - I'm a software engineer that writes in C, targetting Windows, Linux on x86 and S390, HP-UX on PA-RISC, Solaris on Sparc and X86 and AIX on Power. At home I run windows, mac osx and Linux on intel, ARM and MIPS.
My question - why do you want a monoculture?
Each of those platforms has its own peculiarities but each one has a processor, memory, networking etc etc and to the programmer is much the same.
Having a thriving ecosystem in which there is competition is how we advance. X86 would not be where it is now without AMD nipping at intel's heels in the consumer space, or without Sun and IBM to compete with at thw high end.
ARM getting mor popular is, IMHO, a very good thing. It will drive competition and massive investment in the mobile market. Even if they don't go fully mainstream (and by most measures they already are), we end up with a marketplace full of better products because of their innovation AND others feeling the need to compete with them.
I will fully support them and love the small, cheap, low power consumption but decent performance devices that they enable.
@david.hicks: great post.
But just one thing... I thought ARM is the popular one (in mobile space) and Intel is the one trying to nip at ARM's heels.
I'm all for ARM though.
david.hicks
2010-05-11, 13:46
@david.hicks: great post.
But just one thing... I thought ARM is the popular one (in mobile space) and Intel is the one trying to nip at ARM's heels.
I'm all for ARM though.
I may have slightly misworded my post if I gave the opposite impression, yes they are far ahead, ans as such intel entering their market may be a good thing too. I just wanted to warn off the "one platform to rule them all" meme.
(I missed a platform earlier, and I know I'm just bragging now, but I also work with HPUX on Itanium :)
Also, there is AAVA Mobile (http://www.aavamobile.com/downloads.php) : it is a startup in finland that has done a prototype smartphone with Atom processor + Android or Meego. There is several people coming from Nokia in their investor :www.nexitventures.com (http://www.nexitventures.com/portfolio/aavamobile.html). I wonder if that company could be a dummy corporation that serves to intel and Nokia to prepare discretly a MeeGo+Atom Z600 Nokia smartphone.
quipper8
2010-05-11, 19:17
aava looks cool. ilove that they are just selling the hardware and not trying to get you on a specific network/service/app ecosystem etc.
i think(hope) this is where the mobile industry is going. you buy a device just like you do an x86 pc nowadays and then do what you want with it.
aava looks cool. ilove that they are just selling the hardware and not trying to get you on a specific network/service/app ecosystem etc.
i think(hope) this is where the mobile industry is going. you buy a device just like you do an x86 pc nowadays and then do what you want with it.
Maybe. I'll believe it's going that way when HTC starts offering devices to end-users directly in the US, instead of forcing them to go through a carrier.
This is one reason that I like Nokia, they actually sell to their customers, instead of to 3rd parties.
admiral0
2010-05-11, 19:47
X86 is a very old platform, full of "patchy" stuff that is ugly. Arm is relatively new, and by being RISC it will never be very complex (at least as X86). This is why i consider arm better than X86.
Arm processors are not only for mobile use. I've seen some interesting arm powered netbooks lately. I really hope in an arm-powered desktop, because it would be less noisy, less power hungry(yes, it's important even on desktops) and have a longer life.
@ndi
React OS is a bomb. I don't know when it'll explode, but i'm really happy i'm not tied to it.
david.hicks
2010-05-12, 03:39
@ndi
React OS is a bomb. I don't know when it'll explode, but i'm really happy i'm not tied to it.
How a bomb? Do you mean they're setting themselves up to be sued into the ground?
I've been watching Reactos for about ten years, and whilst they seem to make some progress, it's been a long time now and they've never had anything much ready for day to day use. I wish them luck, but where free windows was interesting a decade ago, linux is now my thing.
The nearest you'll find to an ARM desktop right now is a device called the OpenRD. As far as I know, anyway.
It's the big brother of the sheevaplug and contains a 1.2 GHz chip.
vBulletin® v3.8.8, Copyright ©2000-2024, vBulletin Solutions, Inc.