ARM Attacks Intel's Netbook Stranglehold

Barence writes "British chip designer ARM is launching an outright attack on Intel with the launch of a 2GHz processor aimed at everything from netbooks to servers. ARM claims the 40nm Cortex A9 MPCore processor represents a shift in strategy for the company, which has until now concentrated on low-power processors for mobile devices. In the consumer market, ARM is pitching the Cortex A9 directly against Intel's Atom, claiming the processor offers five times the power while drawing comparable amounts of energy. 'It's head and shoulders above anything Intel can deliver today,' ARM VP of marketing Eric Schom claims. However, it has one major hurdle to overcome: it doesn't support Windows. 'We've had conversations with Microsoft and you can imagine what they entail,' says Schom."

Comments

Goody

[Hadlock]

Broken, first gen/beta ARM drivers for all my hardware!

Re:Goody

[Albanach]

Looking at most the Atom devices around, they tend to be in small devices with a limited amount of hardware. Looking at my eeebox, ir has nothing other than a keyboard, mouse and hdtv attached. For netbooks, you know pretty well exactly what hardware you need to support.

If they can make sure there's an HD supporting graphics chipset with drivers, this will be an interesting chip.

Re:Goody

[Microlith]

I know, there's nothing like a lack of attention to hinder the pace of driver development. Therefore we should never adopt the alternative platform, as the drivers will obviously not improve.

On the other hand, I would like to see someone give Intel a run for their money since it seems AMD is being kneecapped. If ARM does it from the low/embedded end and moves up (leveraging their huge number of licensees) then all the more power to them.

Re:Goody

[Andy Dodd]

Yeah. With a few exceptions, about the only variation between most netbooks out there in terms of required drivers are the following:
1) WiFi chipset
2) Card reader chipset (newer ones all seem to be USB mass storage, older ones tended to be a bit less standardized)
3) Bluetooth chipset (Bluetooth chipsets are basically standardized - While I know nonstandard ones exist, Bluetooth adapters that aren't a USB device compliant with a particular USB class are extremely rare.)

This is because the Intel Atom platform is EXTREMELY standardized. With a few rare exceptions, if you use an N-series Atom processor, it'll be paired with one of two variants of the Intel 945G chipset with GMA950 graphics.

Atom Z-series are a different story - they are all paired with a particular chipset with "GMA500" graphics, which unlike most Intel chipsets has basically nonexistent Linux support. So never buy an Atom Z-series based machine if you want to run Linux, they are nearly always paired with unsupported graphics.

Re:Goody

[Hadlock]

Great! You go ahead and be an early adopter, suffer through first gen/beta headaches, buggy drivers, random system crashes. Call me and let me know when it's stable enough for "mom". I don't know about you, but I've grown used to stable hardware, and I'm not about to go back to pre-XP SP1 crashyness for an extra hour of battery life, maybe even two. 5 hrs is plenty enough for me.

Re:Goody

[Anonymous Coward]

Dell is selling Atom D520s with Ubuntu. Surprisingly, you don't even have to search for Ubuntu now. Some pages don't show the link, but this one does:

http://www.dell.com/us/en/home/notebooks/laptop-inspiron-10/pd.aspx?refid=laptop-inspiron-10

"Customize with Ubuntu"

Re:Goody

[TheRaven64]

Any half-decent OS (I think this even include Linux these days) uses the same drivers on multiple architectures with just an abstraction layer for dealing with the different busses. OpenBSD on ARM, for example, supports exactly the same set of USB devices as OpenBSD on x86, including things like USB video cameras. If anything, supporting multiple architectures improves the quality of the code. NetBSD and OpenBSD both recommend testing all drivers on x86 and SPARCv9 and this has helped find a lot of bugs that are not obvious on x86 but crash on SPARC, which has improved the drivers and benefitted x86 users.

Re:Goody

[Rasperin]

it's only 5hrs better because it spends more time down then up :D

Re:Goody

[Anonymous Coward]

I like you you casually throw "hdtv" into the list as though it were as simple as a keyboard or mouse, instead of the massive driver clusterfuck that it historically is.

Re:Goody

[nxtw]

If they can make sure there's an HD supporting graphics chipset with drivers, this will be an interesting chip.

The GMA 950 found in Atom systems is capable of outputting at 1080p60, and with a hardware decoder, can play HD H.264/VC1/MPEG2 content (such as Blu-ray). Broadcom makes such a decoder in PCI Express Mini Card form.

Re:Goody

[Ant P.]

Oh you can keep your XP SP1. You're not even the target audience for ARM.

Re:Goody

[RazorBlade99]

There is no real abstraction layer at all. All you have to do is ensure the correctness of endianess with the appropriate calls and everything else is exactly the same. Most of the consumer ARM CPUs are little endien anyhow. Most of the issues I've encountered are usually sloppy driver writing such as taking x86 hardware based caching ops for granted and not call the appropriate cache flush/invalidate functions when operating on other ARCHs that do those things in software for DMAs.

Re:Goody

[BikeHelmet]

So... no different than Intel? ;)

Atom Z-series. Frak!

Re:Goody

[Anonymous Coward]

Don't fall into that trap. The future of that driver is very much an unknown at this point in time. It is unclear if Intel intends to support it, and "clusterfuck" is the word xorg developers have used to describe the project handling so far...

Summary: Do not buy Z series before a proper driver appears, unless you're happy running a specific version of Ubuntu forever, or vesa is good enough for you.

Re:Goody

[hairyfeet]

While that may be true for...say a geek with IT experience, you really have to think like Joe and Sally average. You know what my customers call Netbooks? They call them "baby laptops" which is VERY important. You see they expect their "baby laptops" to be able to do most of the things a big laptop would, only slower...well because they are babies and babies are little. Intel was VERY smart in that respect, by pairing the Atom with WinXP it runs the apps folks are used to, and because of the "baby laptop" aspect they even expect it to be kinda sucky, because babies are little and aren't strong like the "big" laptops.

So while I have no doubt these things will find a niche, if for no other reason battery life, the real question to me is how big of a niche, and whether that niche will be big enough to sustain it. Because never underestimate how much folks love their little Windows apps. In my 15 years I have seen everything from cheesy photo software that came with a camera to some 10 year old graphic arts program (Xres) labeled a "must have-no matter what" and there is nothing Joe and Sally hate more than change. Hell even now I am building two brand new XP boxes because the customers were willing to shell out the dough just to NOT have Vista/Win7.

Trying to switch folks over to X86 Linux is one thing, where you can at least give them Crossover Office which will help cover the "must haves" but with ARM you are expecting the user to not only throw away everything they know, but every program that they like. And whether Joe and Sally will go for that is a BIG if my friend.

Re:Goody

[Hurricane78]

You mean other than broken first gen / beta Windows 7 drivers running at a fifth of the speed for the same price? ^^

Re:Goody

[node 3]

If they can make sure there's an HD supporting graphics chipset with drivers, this will be an interesting chip.

The GMA 950 found in Atom systems is capable of outputting at 1080p60, and with a hardware decoder, can play HD H.264/VC1/MPEG2 content (such as Blu-ray). Broadcom makes such a decoder in PCI Express Mini Card form.

The GMA950, and the associated chipset to go with it, found in atom systems is capable of using up significantly more power than the Atom processor itself. Adding in yet another chip is not the answer.

Ever wonder why there aren't any Atom smartphones, but tons of ARM based ones? As a platform, Atom is pretty crappy. It just happens to have the best mix of performance, low cost, and low power, while still being able to run Windows.

The Cortex A9 looks very promising, and may trounce the Atom in all ways, except being able to run Windows (which, depending on your point of view, may not be a negative).

Atom + Ion looks like a good combination for those who must have Windows, but the new ARMs look like a more perfect fit for the netbook hardware itself.

Re:Goody

[node 3]

While that may be true for...say a geek with IT experience, you really have to think like Joe and Sally average. You know what my customers call Netbooks? They call them "baby laptops" which is VERY important. You see they expect their "baby laptops" to be able to do most of the things a big laptop would, only slower...well because they are babies and babies are little.

Except they suck as "baby laptops". Trying to run a netbook the same way you run your desktop (let's be fair, laptops are already "baby desktops"), is doomed to frustration. Just as using a netbook as a primary computer is misguided (sometimes circumstances require it, but a netbook is, at best, a temporary hold-me-over until you can replace it with a real computer).

Netbooks should really be seen as auxiliary computers, like the iPhone. Windows is a poor fit for netbooks, but probably the best there is at the moment. Linux, especially Android, has a lot of potential, but has yet to convert that potential into something useful and usable.

For a netbook as a "baby PC", Intel and Windows is pretty much a requirement. But for a netbook as a netbook, ARM is *much* better suited for this than Intel is. It's just lacking the proper OS (which, unfortunately, is the hardest part to get right).

Trying to switch folks over to X86 Linux is one thing, where you can at least give them Crossover Office which will help cover the "must haves" but with ARM you are expecting the user to not only throw away everything they know, but every program that they like. And whether Joe and Sally will go for that is a BIG if my friend.

Switching people to Linux on a netbook implies they are replacing or foregoing have a full-sized system (desktop or notebook).

Re:Goody

[node 3]

Great! You go ahead and be an early adopter, suffer through first gen/beta headaches, buggy drivers, random system crashes.

I think you're operating on a flawed assumption. These systems won't be running Windows.

Re:Goody

[nxtw]

The GMA950, and the associated chipset to go with it, found in atom systems is capable of using up significantly more power than the Atom processor itself. Adding in yet another chip is not the answer.

And the chipset combination in Atom systems provides: Serial ATA, support for graphics output for two displays up to 2048x1536 each (with 3D acceleration), many USB 2.0 host ports, PCI Express, support for up to 2GB DDR2, and a 533 MHz FSB.

Ever wonder why there aren't any Atom smartphones, but tons of ARM based ones?

No, because Atom+945GSE+ICH7M has plenty of features not needed on smartphones. And this combination is significantly faster and more powerful than any smartphone I've used.

The Cortex A9 looks very promising, and may trounce the Atom in all ways

How much power will an ARM-based netbook need to provide a similar feature set and similar performance?

Re:Goody

[eav]

I don't mind not running Windows.

Re:Goody

[DRACO-]

Got a windows mobile cell phone? Half of them are ARM based. My current cell carries two Arm chips. Works well enough. Linux is already into ARM processors, though I have barely tinkered with that.

My old HP jornada had an ARM processor and there was a linux distro based off debian arm that could be installed in it. I almost got it running before I got bored of it. The kernel would boot but would halt when trying to load the file system from the memory card. The kernel didnt have the right file system type that I wrote to the card in order to mount and move on so it just halted. I never got back around to figuring out the why.

Personally, I'd love to have an ARM based linux laptop that potentially could last hours on batteries. I don't need the fastest processor out there, just enough to run web browser, flash and local video/audio. I used to use my jornada for a month before putting it on charge.

Re:Goody

[hairyfeet]

While I would say you would be right if we are talking about somebody trying to game on the little fellers, remember we are talking Joe and Sally here. A LOT of the crap they label as "must have-no matter what" frankly would run fine on HALF the power of an average Netbook.

Take my ex-NASA engineer down the hall. His "Must have-no matter what" is a graphics arts program called XRes. I had to build him a NOS(new-old stock) because this damned thing is so old it would only run on a CPU less than 2Ghz, NO SATA, and more than 1Gb of RAM and it would get flaky. But for him, he says he can get his graphics work down in 1/5th the time with XRes than PhotoShop, as he says the most common features he uses are buried in Photoshop and only take one or two clicks in Xres.

And I could go on all day listing folks that are just like that. They have one or two programs they really like, rarely use more than one at a time, and as I said because they are "babies" they expect it to be slower than what they are used to. And lets be honest-WinXP really don't need much. I am typing this on a 733MHz Office box with XP Sp3 that was given to me because it is storming here and I don't want to risk my bad boy. Maxed out at 384Mb of PC100 along with an old FX5200 GPU I had lying around, it is actually a quite nice Internet experience. I have half a dozen tabs in Firefox, plus my favorite extensions, hell it even plays flash just fine as long as I don't try to multitask and play videos at the same time.

As I was telling a customer just the other day that was buying another off lease Office Box (2.8GHz,1Gb of RAM) after picking up a 2.5Ghz and found it was perfect for his office work, depending on the job computers passed "good enough" quite awhile back. Will you play crysis on the thing? Nope, but the folks using these things Ain't TRYING to play Crysis. They just got one or two little apps they are happy with and want to keep. And as for comparing Netbooks to phones? Uuuhhh...that is what the phone is for bud. I've found the Joe and Sally type just want to surf, watch vids, and run their two or three little apps. for some it is their little camera software, for others some little office app, for my mom who no matter how many nice boxes I offer her refuses to let go of her 700MHz Win98 box, it is Bounce Out and Age of Empires 1. I don't know what it is about AOE 1, but my last boss used to sell older boxes by loading a copy of that (he bought a crate full of that one game) and set it out front where the females could see it.

But just like the other "must have-no matter what" that ain't gonna run on ARM either. So while I have no doubt they'll find a nice niche I doubt they make a dent in Atom. Sorry hoss, but folks are just creatures of habit, that's all.

Re:Goody

[Hadlock]

SP1 was the first "non-crashy" consumer OS. SP3 is pretty solid as a consumer OS - I can understand why Microsoft is trying to move away from it!

Re:Goody

[Spexius]

Just a sidenote: The 3D (and 2D?) acceleration core used in the GMA500 (ImgTech SGX core) is build into a good number of recent ARM based SoCs aswell (for example Beagleboard with the TI OMAP SoC). As far as I followed the whole thing, ImgTech itself doesn't provide any drivers (contrary to NVidia, for example), but seems to put that task on the shoulders of these who make the chips or sell the computer hardware. That doesn't mean there aren't any drivers for Linux (i.e. there should be some video footage of demos on YouTube showing off the SGX core in combination with Linux), but digging these up ... good luck.

The driver for the GMA500 for Atom based boards, running under the "psb" label (coming from "Poulsbo", which is the low power chipset for the Atom CPUs, also check US15W), is truly a "clusterfuck", because the patch contains not only the driver itself, but the whole drm driver directory (and things got changed a good deal in the kernel since then to top it off). It relies on some binary blob(s) to handle the SGX part of the things, whereas the handling of the displays etc. is done similar to existing Intel stuff (which means given some interest and time it should be at least possible to salvage this for proper display device handling, mode setting and otherwise non-accelerated output, no idea about 2D, haven't looked at this all in more detail).

What worries me most is what happens in terms of graphics in the embedded world. It's a big battlefield, and propietary solutions are not the exception. While it's easy to run Linux on a multitude of devices, this 2D and especially 3D acceleration stuff is going to bite Linux sooner or later in the rear end, at least with the way the hardware manufacturers/IP-holders handle the things. But that's exactly what drives the development. Devices like the iPhone wouldn't be possible without 3D hardware acceleration. Add to this things like Flash or the different approaches to bring 3D to the browser, or just simple things like "fancy" UIs (clutter anyone?) ... a company like Microsoft is more likely to get what they want, because they can actually throw cash at the problem.

Re:Goody

[VoltageX]

Shameless plug, but I'm working on tracking down what's happened to the drivers on my website. (my username).org

Re:Goody

[Darinbob]

With smaller CPUs, very often devices are packages in a system-on-chip as well. Not sure if this ARM, or the Atom, bit into this category or not. There are a lot of "drivers" that a lot of end users don't think about since they're somewhat standard on PC motherboards - PCI control, interrupt controller, SDRAM controller, SPI/I2C drivers, etc.

Linux and OpenBSD are good bases to use, because they've dealt with a wide variety of CPU types. Ie, they know that not everything uses a BIOS (no INT 21), many things use only memory mapped I/O, the caching models are all radically different, and different endianness (when will little-endian finally die? :-).

Re:Goody

[arminw]

....while still being able to run Windows....

Which if it is not, relegates the DRM processor to only running Linux. That is likely to be a very small market, that most manufacturers will not bother with. Of course, a computer manufacturer could do what Apple does, and write their own operating system, that is perhaps compatible with Windows programs or OSX programs. However, that is quite expensive and you have to have a lot of money to do.

Re:Goody

[Xero]

I wouldn't quite call gma500 linux support non-existent. I have a dell mini 10 that includes the gma500 "poulsbo" chip and run ubuntu happily on it. While it didn't work out of the box, with a bit of fussing I got it working to my satisfaction. I can even play fullscreen 720p video on the 1336x768 display. The drivers aren't as polished as other chips but definitely far from non-existent.

Re:Goody

[smoker2]

There are no Atoms in phones because they're way too hungry for a phone cpu. ARM have stuff that runs at around 85mW per GHz. Compare that to the lowest Atom at 0.65W for 800MHz.

Re:Goody

[nxtw]

ARM have stuff that runs at around 85mW per GHz

Did you read the entire page, or just a single sentence? That core is much less complex than the ARM cores found in mobile phones. Nor does it claim operation of 85 mW per GHz - this implies that the CPU can actually run at 1 GHz, and there is no such claim other than .085 mW per MHz on the page.

From the press release:

The new processor extends the company's MCU roadmap into ultra low-power MCU and SoC applications such as medical devices, e-metering, lighting, smart control, gaming accessories, compact power supply, power and motor control, precision analog and IEEE 802.15.4 (ZigBee) and Z-Wave systems.

Re:Goody

[Hal_Porter]

It's a shame VESA didn't invent some way to handle accelerated graphics really. I remember reading that adding acceleration to line drawing and BitBlts gave you most of the speed up for graphics. Now admittedly this was back in the Windows 2000 days, and modern GUIs probably need much more hadrware support.

Still you could imagine a scheme where the graphics card had could say "I have line drawing, write the start and end position here and write a command to this register" and "I have BitBlt, write the parameters here and write a command here". That way non Windows OSs could at least have used these operations. Now the problem is that you really don't want to make a Bios call. Still you could imagine an ACPI like mixture of tables that describe the hardware and byte code which controls it. You'd need to interpret the byte code in a native driver of course, so it would have a heavy speed penalty. If you used tables, I think that would work fine. Of course that means that all cards would need to do things in a very similar way.

Now unfortunately VESA didn't do this. So the now all graphics cards support the Windows kernel API and that is the extent of the standardization for accelerated graphics.

Re:Goody

[bhtooefr]

EFI has provisions for having the graphics driver provided by the card, and having the OS just make calls to that, IIRC...

Re:Goody

[node 3]

But just like the other "must have-no matter what" that ain't gonna run on ARM either. So while I have no doubt they'll find a nice niche I doubt they make a dent in Atom. Sorry hoss, but folks are just creatures of habit, that's all.

You're still making the assumption that netbooks are going to replace the main computer. They're not. They are better suited as auxiliary computers, and as the iPhone has proven, your auxiliary computer doesn't have to run the same apps as your desktop.

Re:Goody

[hairyfeet]

Uhhh....hoss? You are looking at the post through your own opinions and not what I wrote. Where does it say ANYTHING about it being used as their main PC? It don't, because they ain't used for that. I'm saying that when they use these "baby" laptops the customers like the fact that it allows them to take their one or two "must have-no matter what" apps with them. And like I said the side really helps Intel in this regard, because folks naturally expect it to run nowhere near as good as their home boxes, because it is a "baby" and babies just ain't as strong and powerful as their bigger desktop/laptop "daddies". here let me give an actual example from an actual customer-

I have a customer that goes back quite a few years name of Todd. Todd and I know each other from way back in HS, so Todd knows I'll always give him a square deal, just as he always does when i need heat & air work. Todd is having me price him one of those "little laptops" as he calls them to replace his old (like Win9x old) setup in his work van. Now he has NO intentions of it being his "main" computer, as he recently bought a 2.5Ghz for the office and a 2.8Ghz for the house, but what it "must have-no matter what" is the ability to run these old invoice and billing programs that he has used for quite awhile.

By being able to run those he is able to simply have me import all the data from his customers via flash drive, and for him this is VERY important, as he has quite a few long term customers and this allows him to simply begin typing their names and all of their purchasing history as well as all form info is filled in. He simply then fills out the data for the new job, prints a reciept for his customer, and voila! Stick a fork cause he is done!

Now for the job he is wanting to do could he theoretically do it on an ARM or as you have suggested, by treating the Netbooks as a mobile phone? yes, he could, but it would be a royal PITA for him. He knows this software by heart, can fill in the forms without even looking, and it already has data from many clients that he would have to painstakingly recreate if he moved to a new platform, so instead he simply has me get him something that will run what he has, and since it will be only doing that single job a good 90% of the time the speed of the Atom means nothing. Hell I might even see if i can't catch him one of those first gen EEE $150 refurb specials, as even the 900Mhz Celeron should be enough for what he is doing.

And THAT was the point I was trying to make hoss. It isn't that they are gonna try to REPLACE their main boxes with these things, it is that over the years folks have gotten attached to programs they have been running quite a while and are loathe to give them up. Because these laptops are "babies" they have NO expectation that they will run anything like their big boxes/laptops at home, but they don't care. And frankly for the way Joe and Sally average use devices-one app at a time, rarely any multitasking at all-these Atoms are "good enough" to run the vast majority of the apps these folks have gotten attached to. So for them ARM would equal throwing away years worth of familiarity and data just to have to start over on a phone like device that for them will be worse than the Atom, simply because they ARE different, and folks want to stick with the apps that they've got.

Folks just ain't big on change hoss, that's just good old human nature. And as long as they are willing to cut me checks for building a new XP box, or finding a "baby laptop" that will let them run their favorite apps when they are away from home, well I'll just keep making sure they get what they ask for. Customer is always right you know.

Re:Goody

[Haeleth]

Except they suck as "baby laptops".

On what grounds do you base this claim?

Trying to run a netbook the same way you run your desktop (let's be fair, laptops are already "baby desktops"), is doomed to frustration.

Works fine for me.

Just as using a netbook as a primary computer is misguided (sometimes circumstances require it, but a netbook is, at best, a temporary hold-me-over until you can replace it with a real computer).

I have two real computers. I also have a netbook. I use the netbook as my primary computer because it's the most convenient.

It's great having a machine so light it can be safely picked up with two fingers, so small it fits in any bag, and with such long battery life I don't need to spend half my time hunting for power sockets. The keyboard is fine for typing, the screen is large enough for working, and the processor is fast enough for anything short of games or HD video. What's the problem supposed to be?

For a netbook as a "baby PC", Intel and Windows is pretty much a requirement.

Damn, I never realised I required Windows! Thank you for telling me. I should have realised something was wrong when I managed to install a driver for my wireless card and connect to my router immediately without even needing to reboot.

Re:Goody

[rphenix]

Actually. Debian runs very very well on my Kurobox Pro (ARM CPU based) only downtime is for power cuts! The only problem you could potentially have is binary blob's (like Adobe's Flash Player) but everything else will be fine.

Re:Goody

[DaVince21]

OMG the Intel i5 is new, therefore it's full of bugs and crashes tons?

ARM has been around for almost as long as the x86 processor structure. I'm sure they know what they are doing.

Re:Goody

[Hadlock]

The i5 uses the x86 instruction set, and will happily run windows using x86 and x86-64 drivers. Drivers for windows ARM are going to be new and extremely buggy, if windows ARM ever surfaces.

Re:Goody

[Anonymous Coward]

WALLOFTEXT
A
L
L
O
F
T
E
X
T

Re:Goody

[DaVince21]

Oh, you mean for Windows. That doesn't have anything to do with all the software that's already ARM-compatible (or even written for just ARM), which works just great.

It's not easy to add ARM capability to Windows, as you'd end up with executables that may or may not run on your system. Kind of the "universal binaries" dilemma Apple had to cope with when they went from PPC to x86. Plus, existing games and apps wouldn't just run unless Microsoft implements a translation layer between ARM x86. Fat chance of seeing that happen (though it'd be cool).

Re:Goody

[PiSkyHi]

What is it with Intel graphics support ?

Its not just the GMA500 and the other newer Intel chipsets like the 4500, the 3d driver support for the 945 has gone unless you want to run Intrepid, which I don't - I had considered it, but I am really getting used to KDE4 now on my desktop and I think Nokia are seeing this path as the future for mobile devices that, in a few years, will be fully fledged PCs.

If I want to have 3d support on a an MID or UMPC right now, my only real option is an ION based machine, with this, I could run KDE4 on an SSD based machine - possibly a little slow, but with at least 1Gb it would be very usable and run everything I need in a PC.

I like this OMAP idea, but I think the drivers for Ubuntu are still lagging - I am not getting any MID until I can run KDE4.

Re:Goody

[node 3]

You keep talking about "Joe and Sally", but your examples are of people with very specific needs (a NASA guy, someone who needs billing software).

For Joe and Sally, the only "must have" software is the web browser, unless you're talking about the main computer, which has other needs. But for their auxiliary computer? Web browser covers it.

And, I ain't a horse, chief.

Re:Goody

[hairyfeet]

Wow, thanks, as I so rarely get to use this in a sentence...WHOOOOSH! Kinda missed the point there bud! I said every Joe seems to have one thing that is a "must have no matter what" that they want on whatever they use.

Want some more run of the mill examples? Well there is little velma, who always wants to have this cheesy camera software, even though she has gone through two or three cameras since then, because for her it is 'easy to fix stuff with", there are the kids that seem to think music began and ended with iTunes, which they want on anything they touch so they can tweak their tuneage, my GF has this lame match three style game she has to have, can't switch that with another, because she likes that one because "the other ones are cheaters!", hell I could go on all day.

It is kinda like that old saying about why folks will pay for MS Office when OO.o is free. It is because everybody only use 10% of MS Office, but everybody uses a different 10%, so you really need to be a 100% match to keep up. Like I said, these things WILL have a niche, the question is how big of one. And finally don't underestimate how much folks prefer the familiar. Just look at how quickly Windows XP was able to take over the netbook market. Folks are willing to put up with different with cell phones ( the thing you keep bringing up) because they are used to that particular interface changing completely every time they change phones. But those "baby laptops" LOOK like "big daddy" laptops, only smaller, so folks naturally expect them to behave the same, only slower.

But time will tell, and I'm willing to bet you can mark this post and in 3 years you'll see that ARM ended up just another one that was supposed to unseat Wintel, only to end up on the dustbin of history. If I was gonna bet on what will be big then, it will either be Ion or the new ULV AMD platform. Both have the muscle to run videos real smooth and give a nice experience, probably more than nice enough for Windows 7. Folks will slowly but surely switch over as XP nears EOL, and when they do they want their "baby laptops" to run what they are used to. Both the new AMD and Nvidia platforms have more than enough juice to do so, and I predict that the line between the netbook and low cost laptops will become even more blurred. In the end the netbook will probably be a full featured laptop, simply minus a DVD drive. But you come back in three years, and we'll see whether you're cell phone paradigm or my familiarity wins out. I'm betting folks will stick with what they know, unless they cutthroat the ARM down to sub $99 territory. Then all bets are off.

No windows support?

[Publikwerks]

I guess it doesn't do EVERYTHING the Intel Atom does.

Re:No windows support?

[Anonymous Coward]

If the processor is really five times as fast as Atom, then it should be able to beat Atom at running x86 Windows with a good enough emulator. Modern binary recompilation techniques are often able to reach 50 % of native speed - already in the 1990's Digital's FX32 for Alpha ran x86 code at 70 % of native performance.

Re:No windows support?

[Big Hairy Ian]

Actually it's pretty much the other way around the fact that M$ doesn't yet support the processor is down to M$ not Arm. I suspect there are a couple of factors here firstly M$ PPC software supports older ARM processors & secondly it's just a case of re-tooling some of M$'s compilers to support the new processor and recompiling. However how it's going to support non .net/java windows software is another matter.

Re:No windows support?

[TheRaven64]

Clock for clock, the Cortex A8 is a bit faster than the Atom on most workloads (in about 10% of the power envelope). The A8, however, typically ships at about half the clock speed of an Atom (they go up to 1GHz, but 600MHz is the most common speed). The A9 is slightly faster than the A8 clock-for-clock, but goes to twice the clock speed and scales to four cores, so it's not a stretch to imagine that it's more than five times the speed of a single-core Atom. I've not seen any figures for the A9's power consumption yet though...

It's worth noting that ARM doesn't make chips, they are an IP-only company. ARM licenses designs to other companies who combine their cores with other stuff and ship them. One of the more high-profile Cortex A9 licensees is nVidia, who are using it in their Tegra line. Other existing ARM licensees, like Qualcomm, TI, Samsung and Freescale have already signed up for the A9 as well.

It's also worth noting that the A9 isn't really news. The designs have been available from ARM for a while now. I don't know of any shipping chips including A9 cores yet (being mass-produced, anyway; there are a few being sampled), but TI announced the OMAP4 series a little while ago which is based around the A9 and looks like a very nice chip for handheld machines.

Re:No windows support?

[DigitalPasture]

I'm guessing that means it doesn't (can't) do most of what an ATOM can do. No x86 support is kind of a dealbreaker.

Re:No windows support?

[Perp Atuitie]

Windows has pretty much a lock on the desktop, so the new chip won't have much market there. The desktop is also the declining market, so the new chip won't be missing that much. The big growth will remain in servers, where windows is optional at best, and netbooks/mobile devices where windows is a minority player. ARM may have made a rather astute decision to concede the dying segment to Wintel and make a big footprint in the markets that will continue to grow, and which also happen to do just fine without Windows. If they make sure to brilliantly showcase the not-windows OSs, ARM could come roaring back as a force to be reckoned with in consumer-level computing.

Re:No windows support?

[Anonymous Coward]

Oh yeah it clearly is Microsofts fault that ARM didn't bother to deliver a platform up until now.

Re:No windows support?

[Albanach]

Oh yeah it clearly is Microsofts fault that ARM didn't bother to deliver a platform up until now.

That'll be news to the folk that have been using computers with ARM processors since the very early 1990s.

Re:No windows support?

[DragonWriter]

Clock for clock, the Cortex A8 is a bit faster than the Atom on most workloads (in about 10% of the power envelope). The A8, however, typically ships at about half the clock speed of an Atom (they go up to 1GHz, but 600MHz is the most common speed). The A9 is slightly faster than the A8 clock-for-clock, but goes to twice the clock speed and scales to four cores, so it's not a stretch to imagine that it's more than five times the speed of a single-core Atom.

Of course, that's kind of a silly comparison to make, since there are dual-core Atom chips like the N330. I also wonder if a 32-bit processor can really challenge a 64-bit one "everywhere from netbooks to servers".

Re:No windows support?

[Albanach]

No x86 support is kind of a dealbreaker.

Well, you wouldn't necessarily expect x86 support on a non x86 architecture, would you.

It need not, and should not, be a deal breaker though. Windows has run on other architectures in the past - Windows NT and its successors have variously run on PowerPC, Alpha and MIPS and Itanium.

Re:No windows support?

[Bigjeff5]

It's not as easy as all that, Windows is built on the core x86 instruction set. x64 changed things a bit, but that was still built on x86, so it was not all that big of a deal.

ARM is a completely different architecture altogether, and porting isn't so easy. Not to anywhere near the edge of their capability, but I imagine they will run into quite a few more problems than just re-writing the compilers will fix.

In any case it should be interesting, if ARM can gain some ground and create another alternative in the processor market it will be better for everybody. For MS, this could end up being the big push they need to move their stuff over to ARM, and then they could truly take over the world, as 99% of PC processors would work with windows. It would also be a big push to produce software in .NET, since that's the only way you'll get full Windows compatibility. Then compatibility to other platforms is even easier for Windows software.

Re:No windows support?

[Anonymous Coward]

This might be a good place for ReactOS to step in. They introduced the beginnings of ARM support way back when. Sure x86 software still wouldn't run on it, but it could at least still present an interface familiar to Windows users.

Re:No windows support?

[e3m4n]

Do you remember transmetta? They have a VLIW architecture chip to make things like tablets etc. They hired Linus T. to create the only software unique to the chip. Code morphing software. It morphed x86 code to run the instructions into the VLIW chip, and vice versus. Instant compatibility with x86 software.

Re:No windows support?

[QuantumRiff]

It doesn't look like we'll have to wait too long to see these implementations in action either. Schorn reckons we'll be seeing ARM ecosystem products containing Cortex A9 designs in the first half of 2009 and then Osprey related silicon to appear later that year.

From the Hexus Article..

It sure would be nice to have an update to that linked article that was written a year ago. I've seen lots of info on ATOM since then, but not much on the A9 systems that should already be out.

Re:No windows support?

[Junks Jerzey]

Remember Windows CE? One of the standard supported processors was ARM, along with SH4 and MIPS. And this was in the mid-1990s.

Re:No windows support?

[Anonymous Coward]

And Microsoft has an OS for these platforms: Windows CE.

Don't pretend you can just recompile Ubuntu and run that on these processors. Sure it will compile, but it won't perform very well.

Re:No windows support?

[Yokaze]

Depends on what you understand under "being built on":
- "Is only compiled for": Windows also is built and sold for the Itanium.
- "Has been only developed for": It also has been developed architecture independent. There were versions for MIPS, Alpha and PowerPC up to NT 4.0. In fact, initial development started on non-x86 processors. The X Box 360 is supposed to use a modified NT 2000 kernel.

Re:No windows support?

[TheRaven64]

64-bit is a buzzword rather than a useful feature for most people. People with 64-bit SPARC or PowerPC machines generally run very little 64-bit software, because doubling the word size doubles the cache usage for no benefit (especially on SPARCv9, where integers registers are all 64-bit even when running 32-bit code). It's only important on x86 because 64-bit also means twice as many general-purpose registers, SSE required (no x87 ugliness) and a few other improvements. If someone defined an ILP32 profile for x86-64, where pointers were 32-bits but the chip ran in 64-bit mode then pretty much all code would be faster than the standard LP64 profile. On non-x86 architectures, 64-bit support is only important when you need more than 4GB of address space and are willing to pay a speed penalty for it.

32-bit only becomes a limitation on NetBooks when you start to get applications that can't fit comfortable in less than 4GB of RAM. This is not likely to be a problem for a few years. NetBooks may start getting more than 4GB of RAM in the next couple of years, but that doesn't require major changes, as long as the OS can address it and map it into processes' 32-bit address spaces (we still aren't getting many machine shipping with more RAM than a Pentium Pro with PAE can address).

Re:No windows support?

[larry bagina]

Most of the time, the only advantage x64 has over i386 is the larger register set. ARM had a larger register set to begin with. If you need 64-bit integers or > 4GB address space, ARM isn't an option, but many servers would be happy with 32-bit cpu (especially a low power, low heat one)

Re:No windows support?

[RightSaidFred99]

You keep saying this in this thread and other threads. Do you have links to support that a RISC architecture (even OOO) beats the Atom in a clock for clock contest? On what workloads? SIMD workloads - I think not... Other workloads? Not sure, I can't really find any good data. On what workloads exactly is ARM faster, I'm genuinely curious for some data.

Re:No windows support?

[larry bagina]

I run Debian on my ARM server (@ 500Mhz). It performs very well. Thanks for pretending!

Re:No windows support?

[MBGMorden]

Windows is coded mostly in C - it's not tied to x86 at all. There is a version for Itanium which is completely different than x86, and there was for a long time a version for DEC Alpha chips which was also a non-x86 architecture.

Porting Windows to a new architecture is largely the same process as porting Linux or any other non-assembly OS to a new architecture - it's just that being a commercial company, Microsoft has to make a business case and see a RoI for the maintenance on a new platform. I'm sure they also don't want to crowd things too much as they could tick off customers who may inadvertently buy the wrong version and want refunds.

Comparatively, an open source OS needs only the desire of a few developers to undertake the project, no RoI is needed, and given that most all of them are available for free,the worse that happens if a user gets the wrong version is they redownload it.

Re:No windows support?

[SBrach]

70% of the netbook market is not a minority.

Re:No windows support?

[MBGMorden]

I think you're spot on for netbooks, but the article specifically said "from netbooks to servers". Most servers that I'm seeing these days are coming with a lot more than 4GB of RAM (and it's needed for most of them). I can see smaller very specifically tasked servers though (say, an email spam filter) getting along with the ARM and lower RAM, but in reality to gain a foothold in that market they'll need a lot more than a few niches.

For netbooks though (and maybe even desktops if Linux ever does become mainstream, or if MS ports Windows to ARM), it's a very nice option.

Re:No windows support?

[TheRaven64]

Look through the Slashdot archives for the article containing benchmarks - I am too lazy to dig it out. It is a gross mistake to regard ARM as a RISC architecture. It is in the sense that the instruction set is orthogonal, but it is incredibly dense (much denser than x86). Almost every instruction can be predicated on one of the condition codes, which eliminates the need for a lot of branching (and, therefore, reduces the overhead from superscalar designs) and every instruction gets free use of a barrel shift on the result. Added to that, most ARM chips from the last decade support one or more of the Thumb instruction sets, which are 16-bit versions of the ARM instruction set, and most ABIs let you switch between these on a per-function basis, so you can compile functions that don't touch more than 64KB of RAM into thumb code and get even better cache usage.

You'd also be surprised at SIMD performance. The Cortex A8 and A9 support both Neon and VFP vector instruction sets. They are not so fast for double-precision vector floating point workloads, but on single-precision and integer SIMD loads they do reasonably well. For very FPU-intensive workloads you are generally better off using the DSP that comes with most ARM SoCs.

Re:No windows support?

[Anonymous Coward]

No x86 support is kind of a dealbreaker.

Well, you wouldn't necessarily expect x86 support on a non x86 architecture, would you.

It need not, and should not, be a deal breaker though.

If your goal is to run x86-Windows apps, then yes it is.

Windows has run on other architectures in the past - Windows NT and its successors have variously run on PowerPC, Alpha and MIPS and Itanium.

Emphasis on "in the past". They stopped making those other versions because it was too much work, and too few customers were buying. Turns out, most people buy Windows only to run their x86-Windows applications.

It's like a three-way chicken-egg scenerio. MS won't make Windows for alternate chips because there's no demand. Customers don't demand because there's no applications. Developers don't port their apps because there's no OS or toolchain.

Any ARM netbook would be a Linux netbook, which is why Schom says "We've had conversations with Microsoft and you can imagine what they entail." Microsoft wants built-in x86 compatibility so their OS and apps will run without modification.

Re:No windows support?

[TheRaven64]

It takes a while for an ARM core to get to market. First ARM releases the IP core. Then chip makers license it and add their own tweaks. Then they start selling the chips. Then OEMs start building devices. We are currently at the stage where chip makers have announced the chips, such as TI's OMAP4. The next step is for people to build devices around them. It's worth noting that TI is trying hard to push ARM chips into the open source community, with things like the BeagleBoard and the OMAPzoom serving as cheap (ish) development and prototyping platforms.

Re:No windows support?

[Disgruntled Goats]

That's because significantly more work was put in to it then a simple recompile. Which was the whole point of the GP.

Re:No windows support?

[LWATCDR]

Yes. 64 bit only gives you a performance boost if any one task needs more than four gigabytes of ram. Most applications on the desktop do not and will not for a while.
In fact using 64 bit code for small programs will actually slow them down. Why waste the memory on 64 bit pointer and integers for a program that doesn't need them.
Only in X86 land will you see a performance increase using 64bit mode. The reason is that the x86 is starved for registers and AMD added more registers when they created the 64 bit extension.
ARM on the other hand does not have that issue.

Re:No windows support?

[LWATCDR]

You can add more ram to a 32 bit CPU than 4GB Intel created PAE to solve that issue a while ago. Most servers probably don't run tasks that require more than a 32 bit address space. Some do but most mail servers, NAS, webservers and even database servers will work just fine with a 32 bit address space.
So since the ARM uses only 10% the power of an Atom imagine a server that instead of say 16 or 24 Intel cores vs one with 160 or 240 ARM cores?
Of course it isn't that simple but you get the idea. Plus for all I know there is a 64bit ARM core already available.

Re:No windows support?

[joeyblades]

People with 64-bit SPARC or PowerPC machines generally run very little 64-bit software...

I'm not an IT professional, so I don't know the technical details. However, we have a lot of SPARCs running 64-bit and only a few running 32-bit. The 32-bit machines are only there to support a handful of applications that can't run in 64-bit mode. However, the applications that run on both run much, much faster on the 64-bit machines... and just to be clear, we're talking about the same physical hardware, just a different OS. Therefore I challenge your claim about the limited use of 64-bit software and your claim that addressability is the only concern (at least for SPARC-based users).

Re:No windows support?

[DragonWriter]

Yes. 64 bit only gives you a performance boost if any one task needs more than four gigabytes of ram. Most applications on the desktop do not and will not for a while.

That "most applications on the desktop" don't need 64-bit isn't an argument that the lack of 64-bit isn't relevant to the ability of this ARM design to compete with Intel's offerings "everwhere from netbooks to servers".

(Which isn't to say that some of the other responses haven't raised some good points on that score.)

Re:No windows support?

[TheRaven64]

However, the applications that run on both run much, much faster on the 64-bit machines.

Of course they will; the 64-bit machines are much newer. If you have a machine to hand, run file on the binary. Tell me what it reports. Solaris supports three ABIs on SPARCv9. One is the old v8-compatible one, which is identical to the old 32-bit SPARC ABI. One is the pure 64-bit ABI. Most software uses the third one, which uses the new SPARCv9 stack layout and instructions but keeps pointers as 32-bit. This is faster than the pure 64-bit ABI on SPARCv9 chips (and doesn't work on SPARCv8 chips), but is restricted to only 4GB of RAM per process. The kernel runs in 64-bit mode so that it can address all of the physical memory. If you are running a 32-bit version of the OS, then you are also running the v8 ABI which had a number of limitations (for example, v9 introduced a much faster way of saving register windows, so it made context switches faster). You can find the ABI documentation for SPARCv9 on Sun's web site, along with recommendations for when to use the different ABIs supported by Solaris.

Re:No windows support?

[hattig]

You could find some answers to your post by reading the linked articles.

Dual-core 2GHz Cortex A9 1.9W
Dual-core 0.8GHz Cortex A9 0.5W

(I'm assuming dual-core for this metric, as all the other metrics mention this).

Dual-core 800MHz Cortex A9 outperforms 1.6GHz Atom (hyper-threaded, one presumes).

Also this is a hard-core, not a soft-core. It's a macro for a particular process, allowing ARM to provide something a lot easier for smaller clients to integrate into their designs. This is clearly a response to (but probably initiated far earlier than) Intel's TSMC Atom SoC plans.

Re:No windows support?

[TheRaven64]

It depends. If you look at most servers, they have less than 4GB of RAM per core and they generally use a NUMA architecture so often have less than 4GB of RAM per chip too. For a while, I've been suggesting that you could make very nice blade servers with 64 Cortex A8 chips on a single board. These, in a typical SoC configuration, would be paired with 512MB of Flash and 256MB of RAM per package. Add a fast interconnect between them and some iSCSI storage and you've got a decent server. Run a web server on one chip, FastCGI web apps on a dozen of them, a clustered RDBMS on another dozen, and so on. The flash can be mapped directly into the chip's address space, so you can get very fast access to it, and you can use something like iSCSI if you need more space.

Re:No windows support?

[BosstonesOwn]

there is a world of difference on the same os, patch levels, and same hardware with Solaris.

64 bit Solaris has been tweaked beyond belief because it was pushed to the forefront. 32 bit Solaris has been sort of a red headed stepchild that has only been maintained not really tweaked as much as the 64 bit version has. And while they do share the same code base , the code has been optimized for 64 bit

Re:No windows support?

[asdf7890]

That's because significantly more work was put in to it then a simple recompile. Which was the whole point of the GP.

Try again. The Debian build system probably creates the Arm architecture releases from much the same sources as x64, i386, mips, ..., ..., ...

Ubuntu could do similar, to go back the grand-grandparent posts point, and you might find they already have: http://www.ubuntu.com/products/whatisubuntu/arm

Re:No windows support?

[ckaminski]

The reason that DEC/MIPS/PPC NT failed had absolutely nothing to do with the lack of apps, but the lack of the value-proposition those companies provided to end-users. MIPS/ALPHA/POWER platforms were two-four times more expensive than commodity Intel platforms. No one forsaw the power-house that Intel was to become with the Pentium Pro/Pentium II, and that single factor alone killed those platforms.

The apps issue was only secondary. When NT 3.1 first came out, boatloads of apps were ported to MIPS/Alpha as well as Intel. BOATLOADS. I supported a network with all three (four, we had some pre-release IBM PowerPC's in 1997). But that cost differential is what killed any chance for a multi-platform windows.

This is a different case. A9 can be as cost efficient as Atom.

Re:No windows support?

[gerddie]

However, the applications that run on both run much, much faster on the 64-bit machines.

Of course they will; the 64-bit machines are much newer.

Nope, grandparent wrote:

and just to be clear, we're talking about the same physical hardware, just a different OS.

Re:No windows support?

[jmauro]

The WOW64 subsystem in Windows x64 for running 32-bit programs in 64-bit Windows is actually based on the older WOW system that let Windows on Alpha run x86 programs without recompling. Windows on Alpha didn't suffer from lack of programs, it suffered from being really, really expensive.

And yes, any vendor wants compatiblitiy with what they already have. It's easier to maximize profits if you can externalize the costs of upgrades on others instead of having to spend R&D money yourself.

Re:No windows support?

[amorsen]

Yes. 64 bit only gives you a performance boost if any one task needs more than four gigabytes of ram. Most applications on the desktop do not and will not for a while.

64-bit kernel gives you a performance boost (or lack of a performance loss, if you prefer) on 1GB and up. Sure, the applications can easily stay 32-bit, but you can't do the classic 32-bit userland 64-bit kernel if the CPU is 32-bit.

Re:No windows support?

[Disgruntled Goats]

Well, you wouldn't necessarily expect x86 support on a non x86 architecture, would you.

It need not, and should not, be a deal breaker though.

Except for all the people running apps that only run on x86 that aren't going to up and ditch those programs to run an OS ported to an ARM chip. Yes, other than that huge barrier to switching there is no deal breaker.

Re:No windows support?

[jimicus]

That'll be news to the folk that have been using computers with ARM processors since the very early 1990s.

Correction: Late 1980's.

http://en.wikipedia.org/wiki/Acorn_Archimedes#Early_models

Re:No windows support?

[node 3]

I guess it doesn't do EVERYTHING the Intel Atom does.

I think you've got that backwards. This chip has one sublime feature that the Atom lacks...

Re:No windows support?

[larry bagina]

I've compiled plenty of software and only encountered one error: v8 (from google) assuming linux was x86 (it also supports ARM). Fact is, recompiling "linux" software on ARM/Linux is easier than recompiling on x86/BSD.

Re:No windows support?

[Bert64]

Servers with lots of ram are generally used for either virtualization or databases...
If you have an extremely low power server, you could put many more of them in the same space using the same power. Power is expensive these days, and a lot of places use virtualization to run a large number of virtual machines with far less than 4gb of ram.

Re:No windows support?

[Bert64]

PAE is an x86 thing, i'm not sure if ARM has anything like it, or how suited the architecture is for a 64bit transition.

Re:No windows support?

[Bert64]

I didn't realise it was even possible to install a 32bit Solaris kernel on modern SPARC hardware, i thought that options went away with Solaris 8, and 32bit support was dropped completely with 10.
Most Solaris machines use a 64bit kernel, combined with a mostly 32bit userland, and i have never encountered compatibility problems using a 64bit kernel except with kernel drivers, and since 64bit Solaris has been around for so many years now, i doubt any Solaris kernel drivers written in the last 10 years are 32bit only.

Re:No windows support?

[EXrider]

The X Box 360 is supposed to use a modified NT 2000 kernel.

...and the XBox 360 runs on a tripple-core PowerPC processor, similar to the G5's that Apple used to use, and the PPE in the Cell processor

Re:No windows support?

[Bert64]

Pity that these ARM based developer systems are so difficult to get in the UK, despite ARM being a UK company...

Re:No windows support?

[Bert64]

Porting windows is not the issue, there is no question that current NT based versions of windows would be relatively easy to port to ARM. The problem is applications, the vast majority of windows users run precompiled binary applications which currently only exist for x86...
While it's true there have been various non x86 NT versions, they have either failed and been dropped (alpha, ppc, mips) or are in the process of dying (ia64), largely because there were little or no third party applications available for them.

Re:No windows support?

[Anonymous Coward]

On your server... so it has no desktop. How does that compare to Windows?

Maybe you don't agree with what Windows is and what comes with it. Maybe you call all that GUI stuff and all these default services bloat. I am not going to argue that. That still does not change that they are included in the standard Windows.

If you want a slim version of Windows there is Windows CE and it works very well on these ARMs. But stop these dishonest bullshit, comparing a headless Debian server with a standard Windows installation.

At work we have a lot of ARM devices. They all have a screen smaller than my hand. I just can't imagine that a merely recompiled Windows would be usable with these without a lot of customizing. The same is true for a recompiled KDE or Gnome.

Bottom line: stop this bullshit and use your brain.

Re:No windows support?

[MBGMorden]

That's a chicken and the egg problem though. Third party application support didn't materialize because it wasn't picked up by a large number of users (and none of those chips were ever targeted towards home users). If Microsoft put in the leg work to port the OS, most actual application companies could port their apps to ARM merely be recompiling - and if there was consumer demand, they certainly would. Look how eagerly Mac developers transitioned to x86 when the PPC architecture was phased out.

Re:No windows support?

[Darinbob]

This is all back to the same terminology confusion when the micro switch from 8 to 16 bit happened; and then from 16 to 32 bit. Ie, you can have a processor with a 32 bit data bus, 32 bit registers, etc, but a 36 bit address bus (a full 64 bit address bus would be overkill at this point). You can even have a completely 32 bit processor attached to an MMU that understands more than 4GB of RAM, which would adapt to the ARM architecture which works with coprocessors.

Re:No windows support?

[Darinbob]

As far as application code goes, 64-bit is only going to be faster if either there is a need for larger integer sizes, or if there are additional benefits to the 64-bit mode other than the register/bus size (more registers, additional instructions can be used, etc).

Re:No windows support?

[joeyblades]

No. I said exactly the same hardware, just different OSes loaded. Different versions of linux, to be precise.

Re:No windows support?

[joeyblades]

We're running linux, but the same probably applies to some smaller degree.

Re:No windows support?

[joeyblades]

Sorry, forgot to mention... linux

Re:No windows support?

[Darinbob]

Microsoft did support the DEC Alpha with an early NT version. Of course since then I'm sure a lot of non portable code has crept back in :-)

PowerPC used to be the "alternative", not sure why it seems to have dropped out except for mid-range embedded systems. Even more registers than ARM, no wonky alternate instruction sets, 64-bit family available, etc.

Apple dropped PPC not because it was an insufficient architecture, but because Intel was a better supplier (economies of scale). ARM may have similar issues down the road. ARM may set a new bar, but Intel can just do the same trick of over-engineering a crappy architecture to get the desired performance/watt ratio, once the netbook market it big enough for them to focus on.

Re:No windows support?

[rohan972]

70% of the netbook market is not a minority.

True, but he did specify "netbook/mobile devices". iphones are pretty popular as are nokia devices. N900 looks pretty promising as far as usability goes and is linux on arm. Mobile devices still need to connect to windows machines because of MS market share on the desktop but as more functionality is moved to the internet and non-windows mobile devices then MS lock-in is becoming less of a problem.

Re:No windows support?

[larry bagina]

Maybe you should imagine the story title: "ARM Attacks Intel's Netbook Stranglehold." Or Maybe you should imagine RiscOS and Acorn. ARM was on the desktop long before it went embedded.

Re:No windows support?

[Albanach]

Now I feel positively behind the curve, not having owned one until 1993!

Re:No windows support?

[arminw]

...Look how eagerly Mac developers transitioned to x86....

Could that be because 90% of the world's computers were using x86? For a profit-making corporation like Microsoft, what advantage would there be in putting out a version of Windows that nobody would buy, because there is no application software. The whole point of the computer is not its operating system, but what actual work you can do with it. What good is a computer with a 24 or 30 hour battery life, if there are no programs available for it, or at least very few. It seems that most of the techies here on /. are enamored with one operating system or another, and are totally forgetting that an OS is only a means to an end, not in and of itself. The reason that Macs are selling rather well, despite their cost, is that it is easier to get work done without futzing around with the computer a lot.

Re:No windows support?

[countach]

Yes, but it isn't terribly efficient. If the total VM (including swap) is > 4GB, its generally best to have a 64bit machine.

Re:No windows support?

[countach]

Come now, you can't say it has "absolutely nothing" to do with apps.

Re:No windows support?

[Haeleth]

It's really not as bad as you make it sound.

(a) Apple managed to switch essentially all their customers from PowerPC to x86, despite the fact that when they started the process, 100% of the most popular Mac apps were only available for PowerPC.

(b) Increasingly, people are tending to use apps that aren't architecture-dependent at all. As more and more computing moves to the web, it matters less and less what processor you use.

The same goes for apps written in Java, .Net, Python, etc. Unless they go out of their way to call native code, they will be portable to anything that can run the correct VM.

(c) Most of the programs that still require native code running locally -- things like Photoshop, HD video editing, and graphically-intense 3D games -- aren't the kind of things people are going to want to run on a netbook in any case.

Re:No windows support?

[ggeens]

And Microsoft has an OS for these platforms: Windows CE.

Indeed. And that's what these Netbooks will be running when (if?) they arrive in the stores.

FWIW, I have already seen advertisements for mini-laptops running WinCE. I didn't check the CPU type but I assume it would be x86.

Re:No windows support?

[Big Hairy Ian]

That got re-branded PPC (which I referred to) thanks though

Re:No windows support?

[Hal_Porter]

Actually this points to another problem with ARM competing with Intel. Intel have their own, very good fabs. For desktop and server CPUs this is absolutely critical - you need a very high end process to compete. Most ARMs are built on a much less hi tech process, e.g. TSMC's. That means that even if ARM had a desktop class CPU, the OEMs don't have a desktop class chip process to make it on.

Fabless CPU companies can compete only if they have some sort of architectural advantage, e.g. MIPS 4000 vs the 486. If they don't they will end up shipping at too low a clock rate to be competitive. In the long run MIPS gave up trying to take over the desktop and concentrated on embedded stuff. ARM has always been about embedded. I think the chances of the architecture taking over the desktop is minimal.

Re:No windows support?

[Hal_Porter]

...Look how eagerly Mac developers transitioned to x86....

Could that be because 90% of the world's computers were using x86? For a profit-making corporation like Microsoft, what advantage would there be in putting out a version of Windows that nobody would buy, because there is no application software.

Not having any stupid user mode codes means the CPU can switch down to a low power state more often. Turn off the LCD and put the DRAM into self refresh and you should get an awesome battery life.

Re:No windows support?

[bhtooefr]

Well, PPC is just IBM and Motorola.

ARM isn't even an ordinary fabless company, they sell their designs to other companies (sometimes even other fabless companies) to integrate into SoCs.

From TFA, they've got 6 licensees listed publicly, and another 6 confidential, for this particular design. Plus, they license the ISA - Qualcomm has their own ARMv7 implementation.

Re:No windows support?

[bhtooefr]

WOW was for 16-bit applications on 32-bit OSes, IIRC.

You're thinking of FX!32, and I don't think it is all that related.

Re:No windows support?

[LWATCDR]

Actually not that many servers need 64-bit right now.
Some big database servers do but many do not.
The real benefit of ARM will be just how many cores you could pack in to a server. If the ARM cores have 10x the power to performance ratio of the Atom then how many ARM cores can you pack in to server in the same thermal envelope as say a 4S Intel box? One or two hundred maybe? I see memory bandwidth get to be an issue at that point but the idea of many small cores is interesting to say the least.
 

Re:No windows support?

[DaVince21]

Direct x86 support on an ARM processor would probably make the chip a lot more complicated and expensive. I'd say it's better if software developers focus on making a software layer/solution to translate instructions if required instead.

Re:No windows support?

[petermgreen]

Look how eagerly Mac developers transitioned to x86 when the PPC architecture was phased out.
Funny, I remember many of the big software vendors taking thier sweet time doing it, and that was with the only OS-X hardware vendor making it clear that they were not only introducing intel machines but they were dropping PPC ones.

Lukilly for apple the combination of a good emulator and the fact that they made quite a jump up in CPU power when they switched smoothed over the transition.

Re:No windows support?

[WebCowboy]

I have already seen advertisements for mini-laptops running WinCE.

CE based mobile computers are very old news. Heck, NETBOOKS are very old news. The term netbook is just a buzzword for what we all called ultraporatbles or subnotebooks a decade ago.

Look up the "Comapq Areo 8000". It is a 10-inch netbook from TEN YEARS AGO based on a low-power orthogonal/RISC processor, has 12 hours of operational battery life, no moving parts (used a CF card instead of HDD, no cooling fan). It used WinCE in firmware as the OS. The machine is the precursor to and architcturally nearly identical to the HP iPAQ palm computers/PDAs, just in notebook format.

Windows CE based OSes (WinCE/WinMobile, etC) are a NON-STARTER. MSFT has had ten years to make CE successful in subnotebooks/netbooks and they've gone over like a lead balloon. Corporate customers knew what they were getting and there was a little bit of uptake there but by and large consumers viewed them with mild hostility. With the little palm-style PDAs there wasn't the perception it was a "real computer". They were very small, they had limited capabilities and people had limited expectations. But when that technology was scaled upwards to the subnotebook form factor they looked a little too much like "real" notebooks. Ten years ago there were still a number of people using 800x600 screens on their desktops and "real" notebooks and laptops. The Aero subnotebooks had that resolution on small-but-serviceable 10 inch colour screen, and it had a proper keyboard too, and when it booted up there was the Microsoft Windows logo and start button and everything just like in the contemporary Win9x machines. Sure the machine was very capable BUT IT STILL WASN'T COMPATIBLE! You couldn't go out and buy apps off the store shelves and load them in (no floppy or CD to begin with--and even if you docked the machine to connect it to a "real PC" you couldn't install normal apps to it).

So FWIW, the fact CE-kernel OSes from MSFT are already there for ARM it really ISN'T worth anything. In fact, the "customer confusion" factor that people claim hapered Linux would be EVEN WORSE for WinCE! At least with Linux it wasn't emblazoned with MSFT and Windows logos on it and it didn't boot up into a nearly identical LOOKING environment. People could sense a difference. Put WinCE on netbooks and they'll be REALLY confused, because it'll LOOK familiar but in fact CE is really no more compatible with NT-based Windows (XP/Vista/7) than Linux is. It'd be more consumer hostility just like 10+ years ago when they tried the first time.

Besides that, Linux-kernel OSes seem to be where it's at. Google is on its third release of Android in the time it took MSFT to more up one release. Palm's WebOS looks absolutely stellar and the company looks like it may have actually pulled its arse out of the fire. Moblin v2 is ready to debut and developers of many moblin components are actively working on ARM ports, so by the time WinMo 7/next-generation CE would be ready Linux could already have a lot of traction.

And the prospect of full-blown Windows 7 on ARM? I think winning the lottery is more likely. Even if Windows 7 were to be ported, what about the apps? They'd need to port REAL MS Office, etc. and get all the vendors to rebuild too, or spend a long time on emulation/virtualisation and take a performance hit. It was a huge undertaking for Apple that happened over years (starting by a massive OS shift from OS9 to UNIX-based, multi-platform capable OS X with all sorts of frameworks to support legacy baggage, and once that was done the hardware shift to x86, etc...). It would take MSFT a decade of true dedication to herd all their cats and pull that off, and no matter what words Ballmer bellows about Mobile/CE, MSFTs actions have demonstrated they totally lack the dedication it would take to seriously support NT on non-Intel architectures.

I didn't check the CPU type but I assume it would be x86.

You

Re:No windows support?

[jmauro]

You are correct. I was confusing FX!32 with WoW64 on Itanium which does do instruction translation to let x86 programs run on the Itanium arch. My bad.

Re:No windows support?

[DigitalPasture]

Well, you wouldn't necessarily expect x86 support on a non x86 architecture, would you.

Only when they are comparing to Intel (as they are). Can't see how it could slay Intel's hold if it's not competing in their market.

Re:No windows support?

[DigitalPasture]

I understand what you are saying, but I think you are missing my point. Without the (end) user base, which is using x86 and x64 programs... It will not gain support. Windows is what a majority of users use whether we like it or not. If Windows doesn't run on it, how will it get market share? We've already found through Dell's efforts (among others) that Ubuntu doesn't work out when users are expecting the same ole desktop they've always had.

What does it support?

[BadAnalogyGuy]

I suppose Ubuntu Linux is just chopped liver.

C'mon people. Wake up! There are tons of operating systems out there. Some are even better than Windows! *gasp*

Re:What does it support?

[Anonymous Coward]

It doesn't matter if Ubuntu is many times better.

Most people are garbage, that's the problem.

Re:What does it support?

[Publikwerks]

I like Ubuntu, but to ignore a large percentage(albielt shrinking as linux netbooks gain popularity) is kinda a big deal. It will be intresting to see if they can get hardware support, or if they will just end up like Transmeta

Re:What does it support?

[kav2k]

Mod parent up, please..

Re:What does it support?

[QuantumRiff]

Just out of curiosity, does the ARM version of Ubuntu take advantage of some of the stuff in ARM for doing HD video at low power? Or is it just ubuntu, recompiled for the architecture? There are several advantages to each different CPU. Do things like Flash (or even Gnash) work on ARM? Or VLC, or anything?

Re:What does it support?

[BeardedChimp]

The PcPro article showed their ignorance when they wrote:
"Nevertheless, netbook manufacturers running the ARM processor will be forced to adopt an alternative such as Google's Android, Windows CE or even Windows Mobile."

They mentioned running windows mobile over the ability to run fully fledged distributions such as debian with huge repositories.
I can't wait for these laptop to start coming out, microsoft can't even pay people to put XP on them this time.

Re:What does it support?

[jellomizer]

Pretty much... Yea.

Linux for the mainstream desktop/notebook/netbook is just a dream. Linux is a Server OS and it is good as a Workstation OS too... However for Netbooks to sell. They need to be familiar to the end user. Ubuntu is neither familiar like Windows XP to the average user. nor is it Sexy and well polished like OS X. All the pieces are there however they are just not glued together correctly. For Netbooks the fact that it has Windows running on it is a big thing, and a selling point. Having Linux running on it or some other OS brings to mind the old PDA where there is a very few functions that work. And whenever there is a change you are stuck until the next upgrade. Yes this isn't true for linux you have a slew of stuff. But the mental image of using a linux system is just that espectially on a netbook.

Re:What does it support?

[Bigjeff5]

...(albielt shrinking as linux netbooks gain popularity)...

I don't know where you've been seeing the growth, but linux has held pretty steadily at sub-1% desktop market share for years. Netbooks gave it a slight boost when first released, but MS quickly squashed that and now dominates the netbook market. It's true that Windows has been losing ground, but it's OSX that has been gaining, they are up to almost 10% share last time I looked, just a few years ago they were at less than 5%, so that's pretty darn good.

Linux? Not so much. As for the popularity, ARM is pretty popular as is on small devices, one could say they dominate, and MS already has some software that runs on ARM processors, so if this new breed of ARM is popular then we could see MS make the jump. But it will have to work in that order, the ARM will need to be popular and THEN MS will jump on it, it won't magically happen the other way around (unless MS has a major stake in ARM, which I don't think they do).

Re:What does it support?

[H0p313ss]

I suppose Ubuntu Linux is just chopped liver.

C'mon people. Wake up! There are tons of operating systems out there. Some are even better than Windows! *gasp*

There are also other planets than Earth, which is equally relevant.

Re:What does it support?

[pizzach]

Inside the iPhone: Mac OS X, ARM, and iPod OS X

Re:What does it support?

[Anonymous Coward]

I didn't know Apple had an OSX that worked with ARM

Oh really, ever seen iPhone? Regarding your other rant above about Windows built on "x86", false as well. windows has been available on other architectures before, such as PowerPC, DEC Alpha and MIPS R4000. Today it's also available on Itanium, which is by Intel, but shares nothing with the x86 architecture. They can port it, if they decide they want to.

Re:What does it support?

[TheRaven64]

Most people won't be using something like a (cheap) ARM portable as their only computer. For those few apps that depend on Windows, they still have their other computer.

Re:What does it support?

[Anonymous Coward]

Does it matter if you can get Windows for ARM, when there is not a single 3rd party win32 application compiled for ARM?

Re:What does it support?

[cb88]

Actually Apple is an arm licensee and designs their own modified ARM chips for the iPhone ... etc the iPhone O/S IS OS X Also no linux isn't soo bad seriously windows has just as many flaws people are just used to them

Re:What does it support?

[Ephemeriis]

I suppose Ubuntu Linux is just chopped liver.

C'mon people. Wake up! There are tons of operating systems out there. Some are even better than Windows! *gasp*

Yes, there are tons of operating systems out there...

But "better" is completely subjective. If the app you need only runs on Windows, then Ubuntu (even with WINE/VirtualBox/whatever) is not "better". If your users are terrified of technology and change and have only ever used Windows, Ubuntu is not better. If the hardware you need to interface only supports Windows, Ubuntu is not better.

The fact of the matter is that there are an awful lot of people out there who will only use Windows - whether they actually need to or not. And if these new ARM chips can't run Windows, then they're going to have a much smaller market than Intel.

Re:What does it support?

[RightSaidFred99]

You people are like the Beta vs. VHS people claiming Beta was "better". In what way? In what way is Ubuntu "better" than Windows? "Different" I'll give you. Better in some dimensions, I'll even give you. Overall? In what way? Because it's "teh Unixez"?

Re:What does it support?

[Anonymous Coward]

I suppose Ubuntu Linux is just chopped liver.

C'mon people. Wake up! There are tons of operating systems out there. Some are even better than Windows! *gasp*

Linux is a liberal myth. The very idea of another operating system besides Windows is simply ridiculous. Windows is the pinnacle of mankind's achievement in the field of computer science. To make the claim that there is another operating system (and that it is superior!) is laughable.

Re:What does it support?

[larry bagina]

The iPhone is ARM-based, so yes they have an OS X that works with ARM.

Re:What does it support?

[the_womble]

linux has held pretty steadily at sub-1%

Steve Ballmer says otherwise

Re:What does it support?

[onefriedrice]

It's true that Windows has been losing ground, but it's OSX that has been gaining, they are up to almost 10% share last time I looked, just a few years ago they were at less than 5%, so that's pretty darn good.

Not sure where the 10% is coming from, but your supposed 5% stat from a few years ago seems to be more accurate for today's market share.

Re:What does it support?

[Bigjeff5]

So now Steve Ballmer can't be wrong? Now that's a twist.

Granted, Linux is notoriously difficult to track in overall market share, since Linux PCs often spoof themselves as Windows machines, and sales figures are unreliable for obvious reasons, but even if 2/3 of Linux users spoofed their user agent ID, it would still only put Linux at about 1.5% according to recent (August '09, not Feb '09) figures.

In Feb, it certainly looked like Linux would take over the Netbook scene, but MS quickly responded and now most Netbooks come in a Windows flavor, rather than a Linux flavor. So much for that. Seriously, go to a big box computer store and see how many netbooks are of the Linux variety, it is not many.

Linux certainly has the most potential, but honestly I don't see the Linux community getting their act together and creating a truly competitive "product" any time soon. Ubuntu doesn't cut it, not as it stands now.

Re:What does it support?

[Bigjeff5]

You're right I was a off on the Mac's overall market share, I was thinking of recent sales figures - Mac sales make up about 14% of all PC sales today, which is an even bigger growth than I thought it was, but they will need a sustained growth at that rate to make a sizeable dent in the overall market share. Still, their total market share is around 5% and growing, while total Windows market share is around 93% and shrinking for the time being. Linux makes up part of the leftovers.

It's interesting to note that Win2000 still has more market share than Linux.

Re:What does it support?

[Cro Magnon]

Windows == Earth? Somehow, I always thought of it as Uranus.

Re:What does it support?

[Anonymous Coward]

1% is plenty. It means more than enough developers to code everything you need and more than enough users that run into problems so you don't have to.
I care more for that 1% than the other 99%.

Re:What does it support?

[koiransuklaa]

You're right I was a off on the Mac's overall market share, I was thinking of recent sales figures - Mac sales make up about 14% of all PC sales today, which is an even bigger growth than I thought it was

You really need to start citing your numbers, because I'm not buying that one either ;)

14% sounds insanely high compared to what I'm seeing (my experience is in several European countries). Are you sure your numbers aren't for the US? It's very common for US websites to quote figures without mentioning that the research only included sales the US.

Re:What does it support?

[Lennie]

I think the biggest problem for Microsoft with Linux is, people don't need to buy any other hardware. If Linux is good enough instead of Windows people can just install over it. If you want to switch to Apple you would also need to buy hardware.

Re:What does it support?

[Lennie]

If I'm not mistaken Dell sells about 1/3 of their netbooks with Linux and 2/3 with Windows XP in the US. Just look it up.

Also if you want to talk about it: the return rate is about the same, because on their website they actually make it pretty clear it's not Windows.

Even better at times I've seen the Linux/Ubuntu version is cheaper, never the other way around.

Re:What does it support?

[Bigjeff5]

The 14% is worldwide growth in the home market, I have not been able to find where I saw those figures, much to my chagrin.

However, here are some similar figures, though not the study I was quoting, and they aren't looking at the exact same segments either. I may have mis-quoted the article I came accross as well, I can't tell, since I can't find it again.

You are certainly right to question.

Re:What does it support?

[Narishma]

You need to start taking those numbers (wherever they come from) with large pints of salt. How do you even track Linux's market share when everyone can distribute it? How do you count those machines sold with Windows that get promptly formated and loaded with Linux?

Re:What does it support?

[leromarinvit]

and MS already has some software that runs on ARM processors, so if this new breed of ARM is popular then we could see MS make the jump.

So what? Nobody will care. Windows Mobile (Win CE) has existed for years, and it's barely usable on PDAs. It has just about nothing in common with normal x86 Windows, so there is no benefit whatsoever in using it. No proprietary Windows app will run on it, and porting isn't all that easy either. It's a different OS. There's a decent amount of software for it around, but just about all of it is designed for PDAs, i.e. touchscreen, no keyboard, and a specific display aspect ratio (QVGA portrait) which is completely unsuitable for a netbook.

Linux, on the other hand, already has loads of software that runs fine on existing laptops. Look at Debian's repository - almost all of it doesn't care which architecture it is compiled for, including ARM. You get basically the same software as on any x86 laptop.

Re:What does it support?

[snadrus]

Every program with source should work on ARM [except things like Wine's pass thru x86 instructions (Windows) expecting an x86 processor]. That includes VLC, Gnash, Firefox, Webkit, etc.

Additionally, There are nice Adobe-supported ARM Flash players for Linux in the Maemo stack. Google's working to get a better one than it into Android (Linux on ARM).

As for using decoder chips, it would be a matter of a kernel driver & a player that could interact with it. Anyone know?

Re:What does it support?

[smoker2]

Yet windows doesn't run on ARM. Your point is ?

Re:What does it support?

[MichaelSmith]

You need to start taking those numbers (wherever they come from) with large pints of salt. How do you even track Linux's market share when everyone can distribute it? How do you count those machines sold with Windows that get promptly formated and loaded with Linux?

Web server stats are pretty good for that.

Re:What does it support?

[fwarren]

Do you think Steve Balmer would say that Microsoft sees Linux as more of a threat than Mac OS X if it was not true? They list Linux as the biggest threat to their business behind "Pirated Copies of Windows".

I would remember that Microsoft funds lots of studies. Anytime one of them turns up good Linux info and bad Windows information it is just killed and never sees the light of day. Microsoft has a good idea of how many Linux machines out there, no matter how they publicly spin the numbers. They know what the studies say, they know what their customers are running. They know when their large customers drop Windows because they have moved services over to Linux.

If things are bad enough that Microsoft running their own numbers say that there are more desktop Linux installs out there than Mac OS and that they are a bigger threat to their business, I believe them. I have known for years this would come. No monopoly can stand forever and the computer industry is still to young and disruptive for anyone to stay on top for to long.

Re:What does it support?

[koiransuklaa]

The 14% is worldwide growth in the home market

I assume you didn't actually mean "growth" but "sales". Anyway, picking a specific market segment (home in this case) is just as important as the geographical area... I'm sure Apple does a lot better at home than in the workplace.

The only figures I've seen for PC sales are the yearly IGD and Gartner studies (or rather, reports based on those). In those, I've never seen Apple over the 10% mark even in the US (example). In fact, Apple is only ever mentioned in the US numbers because it's not in the top 5 anywhere else...

Your appleinsider link didn't contain any relevant numbers as far as I can see (images don't load so maybe there was something there?).

Re:What does it support?

[Bert64]

Depends how you market them...
If you market them as "small laptops", people will expect familiarity with big laptops, but they're also likely to be disappointed with the performance (many existing netbook users have been disappointed with how they perform)..
Instead, you need to target them as something new....

Imagine an ARM based machine running ubuntu which instead of being marketed as a small laptop, is marketed as a big ipod...
There is already competition among phones and music players, so a new unfamiliar interface won't be unexpected so long as it looks pretty and is simple/intuitive enough.
Such a machine would do everything an ipod does, but have a bigger screen and a qwerty keyboard, so typing anything, browsing, video playing etc would be easier. Sure it won't be pocket sized, but it's portable enough.
You could also market the ubuntu package management application as "the app store", linux has had convenient package management for years, and people would love it if they only knew it existed... Apple did effectively the same thing but wrapped it up in a friendlier interface and marketed it.

Re:What does it support?

[Haeleth]

It's hard to make any firm conclusions based on sales figures.

I've seen a number of people mention buying the cheaper Linux model and then installing Windows on it. And I myself have done the reverse: even though there was a Linux version of my netbook available, I bought a more expensive model with Windows on it, and immediately replaced it with Linux. (The Windows version had a higher resolution display, is why.)

Re:What does it support?

[Haeleth]

Planets are a very poor analogy. It is literally utterly impossible, as of today, for me to switch to a different planet, which is not true of operating systems.

A better analogy would be countries. There are many different countries in the world. Most people prefer the one they're used to. It's possible to switch to a different country, but it involves some work, and there can be compatibility problems, so most people don't think of it as a serious option. But people do it, even enduring severe hardships in the process, if it offers them real benefits. That's why the Pilgrim Fathers aren't still sitting around on Plymouth docks bitching about how the local religion only gives them 3 hours of battery life.

Re:What does it support?

[Anonymous Coward]

desktop share is not the same as notebook much less subnotebook. And where is your citation for linux being less than 1% market share? Good numbers are hard to come by, and it isn't just a matter of source for numbers, but how do you measure it? I run linux on my workstation, windows in a vm. At home I run linux on a workstation, windows in a vm, windows on a separate hardware and have a third box that dual boots windows and linux. I also have a mac laptop running OSX, but has windows installs via fusion and parallels. Now, how do you count that up? The truth is, the vm's (except at work) are hardly ever fired up and I work almost exclusively in linux. But how do you measure that? Uptime? The one windows exclusive box (used about thirty minutes a week) is on 24/7. I'm lazy, stay logged in and have applications running.

 

At the university I work at we're collecting statistics on systems. We've made a serious push to eliminate non-Windows from university owned machines and have driven OS X from ~10% to ~2.5% in the last five years. Despite this, linux use has managed to climb from non-existent to ~4% (with ~1% undetermined). Then if you consider student systems the breakdown is even more interesting. Windows is 32%, linux 12% and OSX 11%. About 12% for miscellaneous (printers, routers, BSD, etc.). By the way, the windows number includes Win2k, WinXP, Vista and Win7. Now, that doesn't add up to 100% because 34% are "undetermined" -- this happens when it is locked down *tight*. Could be any operating system, really. When I've checked it has been some form of BSD, linux or OSX. It *could* be a locked down windows, but given the rest of the numbers (and remember, detection of Vista and Win7) and failure of spot checks to find them, it seems most probable that *at most* Windows only accounts for 1/2 of them (the same as it gets compared to the identified operating systems as a whole).

 

What that means is that *probably* there is less than 50% Windows (Win2k, WinXP, Vista, Win7) among the student systems, with (probably) at least 16% linux, 15% OSX and 16% miscellaneous. Sure, Windows is still a majority, possibly even more Windows than all others combined. But nowhere near the 90%+ figure that usually gets cited.

 

How meaningful are these numbers? I can characterize the "undetermined" based on spot checks, but I can't make any truly strong arguments about how they are distributed. This university has a very technical student body and isn't that large (a few thousand students) and is not representative of students at large. Extrapolating these numbers is not a good idea. But remember, there simply *aren't* any good numbers. They are all guesses. Some /horrors/ are based on browser identification from a self-selecting subset of Internet users.

 

Oh, yeah, MS will jump on ARM if it starts being used for subnotebooks. Why? To stop linux from further eroding their hold on the desktop. ARM doesn't need to be popular. Intel-based subnotebooks have WinXP given to the vendors to prevent them being loaded with linux when MS has ended support for WinXP (or did you miss the memo from MS where despite claim of providing critical security patches for WinXP until 2014 and Win2k until 2010 they have stated they won't fix the tcp/ip stack in either one? Arguments about time required aside, they clearly state and claim to provide critical security patches -- which these vulnerabilities are critical -- until those dates. And MS are *still* providing new issues of WinXP...)

Re:What does it support?

[marcosdumay]

MPlayer already takes advantaje of several different processors characteristics, requiring a simple recopile. If it doesn't aready, it doesn't take a lot to take full advantaje of this chip.

Also, flash does run on ARM, but I guess it doesn't optimize for each processor. If we are luck, that will make Google start streaming Youtube videos on a way that uses mplayer. They can even keep the flv format.

Re:What does it support?

[H0p313ss]

A better analogy would be countries.

That was what occurred to me first. But I really didn't want to engage in (yet another) anti-nationalistic rant. Sadly for the majority of the business world a migration away from Windows is still a pipe dream.

Re:What does it support?

[DRACO-]

Funny, my siblings did not even notice they were using linux the last time they borrowed my computer. My brother said he preferred my computer over my father's winxp laptop since the linux never crashes.

The simple problem with everything is familiarity. They expect internet to be under the blue e icon or the red and blue fox. Another problem is word of mouth new 'hot game' is not produced on anything except windows.

Re:What does it support?

[Anonymous Coward]

Well, no manufacturer will put vanilla Debian on a netbook (or any other device), though Debian variants may still be done, and the article did miss the obvious, Ubuntu, which will certainly be on some.

There is a cheap little underpowered MIPS-based (I think) netbook around which originally came with Linux and is now available with Windows CE, so I won't be surprised to see Windows CE on these ARM-based netbooks.

NO WINDOWS ARM APPS SO -- SO WHAT?

[Anonymous Coward]

It's dead, Jim (Schon). When the lights, go down in the city, does anyone really care? There are ZERO APPS for ARM Windows (if there ever were such a thing, and Windows CE is not Windows) so ARM is better off trying to get people to compiler Linux ARM apps (good luck with that !!).

Re:NO WINDOWS ARM APPS SO -- SO WHAT?

[hattig]

There are 75,000 apps for ARM iPhone OS X.
There are 10,000+ apps for ARM Android OS.
There are loads of apps for ARM Maemo.
There are loads of apps for ARM Symbian.
There are loads of apps for ARM Windows CE and derivatives.
There are loads of apps for ARM Linux and derivatives.

Re:NO WINDOWS ARM APPS SO -- SO WHAT?

[Anonymous Coward]

Why would people need to compile? Debian repositories work ok.

Re:NO WINDOWS ARM APPS SO -- SO WHAT?

[Bert64]

There are already ARM based laptops, and most of them come with windows ce... I know a handful of people who bought them, and all of them saw "windows" and assumed it would be the same as they have on their desktop, boy were they disappointed. None of those people still uses their windows ce machine and several people tried to return them.

Re:NO WINDOWS ARM APPS SO -- SO WHAT?

[Hal_Porter]

I have an ARM in my Windows Mobile phone. There are a lot of applications and some of them are very useful on a phone. I can't imagine using any of them on a laptop.

A compelling Linux on ARM netbook will worry MS

[mikeabbott420]

Linux already made MS drop their price, allwing cheap windows netbooks because of linux. It's not out of the question that a really compelling ARM netbook would scare them into ARM support. I would be surprised if they didn't have something similar to the x86 apple builds in the powerPC era. Of course windows is mainly valuable for its 3rd party software so people who buy these putative ARM/windows machines may be better off with linux anyway.

Re:A compelling Linux on ARM netbook will worry MS

[sunderland56]

A Linux-based netbook won't worry MS if it only does what a MS netbook does. It needs to do more.

For example - they brag that the ARM "offers five times the power while drawing comparable amounts of energy". But, netbooks rarely use all of the processing power they have right now. If the ARM had equal processing power, but five times the battery life, they'd have a compelling product. The current standard of eight hours on a XP-based netbook is barely enough; a netbook that lasted forty hours would be a market breakthrough, and would be compelling enough to get people to switch to Linux.

Re:A compelling Linux on ARM netbook will worry MS

[Microlith]

The current standard of eight hours on a XP-based netbook

Having owned an XP netbook (aspire one) I must say that an eight-hour standard is optimistic beyond belief, and likely only possible if you leave it sitting there. The Atom processor is power hungry and once you start actually using it the battery life plummets considerably.

ARM already has an advantage on power consumption, if they can match the Atom on performance I suspect they'll win on battery life by default.

Re:A compelling Linux on ARM netbook will worry MS

[timeOday]

Whether the ARM chip performance is even adequate for normal netbook applications (e.g. watching youtube) is an open question until somebody tries it. Sure, ARM threw out this number of 5x, which is a meaningless number until we get a better overall idea of how fast and slow it is on different tasks.

Second, even cutting the CPU power consumption to zero wouldn't give you anywhere near 40 hours of battery life in a netbook. The CPU is just one piece of it.

Re:A compelling Linux on ARM netbook will worry MS

[Patch86]

Correct me if I'm wrong, but TFA is about an ARM chip with a a 2GHz clock and a low power footprint.

Or in other words, a chip with the same processing power as an Atom, but with better battery life.

Or in other words- is that what you were after?

Re:A compelling Linux on ARM netbook will worry MS

[AlecC]

There are already Arm based netbooks out there, using the current low-perofmance chips, so presumably Arm has a reasonable reference on how fast their new chip will run a Linux netbook.

Re:A compelling Linux on ARM netbook will worry MS

[UnknownSoldier]

> (aspire one) I must say that an eight-hour standard is optimistic beyond belief

It depends on the _battery_. With my 9-cell battery, my aspire one gets 8 hours of continuous usage.

http://www.amazon.com/HQRP-Replacement-Lithium-Ion-AOA150-1777-AOA150-1840/dp/B001OXRTVU/

Re:A compelling Linux on ARM netbook will worry MS

[Blakey Rat]

Far too little, far too late, I'm afraid.

Two years ago, I'd be all behind this. Now, Intel and Microsoft have such a lead in the market, it's going to be a much harder market for ARM to enter.

Re:A compelling Linux on ARM netbook will worry MS

[hackerjoe]

If the ARM had equal processing power, but five times the battery life, they'd have a compelling product.

Well, it sort of does. Battery life and CPU power are actually somewhat convertible.

When the CPU isn't doing work, its power consumption drops considerably -- if you have two CPUs with the same designed maximum consumption, but one has twice the computing power available, then for the same workload that processor will use (a little bit more than) half the energy.

Of course the real picture is not so rosy, because a CPU that uses that little power to start with is probably accounting for less than half of the total power consumption of the system, and of course the workload is likely to increase if you have more CPU available (people watch video fullscreen instead of windowed, games will generally render as fast as they can and use all available CPU, etc.).

Re:A compelling Linux on ARM netbook will worry MS

[Anonymous Coward]

You are wrong. Clock speed says very little about the actual performance of a CPU.

Re:A compelling Linux on ARM netbook will worry MS

[IGnatius T Foobar]

Linux already made MS drop their price, allwing cheap windows netbooks because of linux. It's not out of the question that a really compelling ARM netbook would scare them into ARM support.

And Microsoft would still lose. The only thing Windows really has going for it is the existing library of PC software. That's the network effect that keeps Windows out front, otherwise the market would have dumped Windows ages ago. Windows on ARM runs existing Windows x86 software about as well as Linux does: not at all.

In fact, ARM netbooks running Windows might actually be at a disadvantage relative to Linux. People would see the Windows logo on the box and take it home, assuming that they could run PC-Windows software. When that software fails to load, the netbook gets returned to the store.

Netbooks running Linux on an ARM processor with insanely long battery life and a true dedicated mobile operating system may be what it takes to get people to realize that netbooks were not intended to be merely smaller laptops.

Re:A compelling Linux on ARM netbook will worry MS

[Anonymous Coward]

Hold on now. A couple years back the rant was that Windows was too much bloat for the light weight processor market and that netbooks would be the first step to smashing Windows. It was 40-60 Linux/Windows on the netbook market. Ever since then it's been nothing but decline for Linux and a lot of complaints about trying to run different distros on netbooks. MS currently has 90-95% of the netbook market and we're hearing of a new MS killer?

Wake me up when the numbers are in. All of the Linux predictions I've heard over the last 15 years always seemed like there was going to be a revolution in computing that has never happened. This one is no different. People who pay for this technology have already proven the case that they would rather have Windows over Linux on their netbooks. Unless there is some real sweet reason to cross over they simply will not do it. Battery life is not likely to do it but may be an in-road but that can only go so far.

A more likely prediction? ARM will lack Windows support and be a faint memory to anyone outside of the embedded systems realm in less than 5 years.

Re:A compelling Linux on ARM netbook will worry MS

[TheRaven64]

Note that watching YouTube is actually not a relevant benchmark for the ARM core. On most existing ARM SoCs, video decoding is offloaded to the DSP, ISP, or GPU and most of the A8 versions can decode 720p H.264 without any problems (and without touching the ARM core). SoC manufacturers like Freescale have partnered with Adobe to ship custom versions of Flash that take advantage of the extra hardware on the chip for exactly this. This means that an ARM chip will generally do a lot better, in terms of power usage, than Atom when watching YouTube because it's using dedicated hardware for the video decoding, while the Atom is doing it on the CPU.

Re:A compelling Linux on ARM netbook will worry MS

[sirinek]

I don't agree. I mean what was IE's market share before Mozilla/Firefox started to gain share? What was IBM's share in the PC market at one time?

Re:A compelling Linux on ARM netbook will worry MS

[omnichad]

Especially with completely different instruction sets. In other words, they do totally different things during a clock cycle.

Re:A compelling Linux on ARM netbook will worry MS

[rolfwind]

Whether the ARM chip performance is even adequate for normal netbook applications (e.g. watching youtube) is an open question until somebody tries it. Sure, ARM threw out this number of 5x, which is a meaningless number until we get a better overall idea of how fast and slow it is on different tasks.

One already out based on the beagle board:
http://www.alwaysinnovating.com/home/index.htm

Re:A compelling Linux on ARM netbook will worry MS

[Anonymous Coward]

And with an ARM processor, that would last days. Eight hours is definitely not the standard, you're lucky if you get six with some Eee's, and around three hours otherwise.

Re:A compelling Linux on ARM netbook will worry MS

[Blakey Rat]

I said it would be hard, not impossible.

The simple fact is that now people *expect* netbooks to run Windows. Selling one that doesn't, even if it's cheaper, is going to be a challenge.

(Also your second question depends on the meaning of "PC." If you mean "PC-compatible" then, of course, IBM once had a 100% share. If you mean "personal computer", then IBM's share was high, but not dominating.)

Re:A compelling Linux on ARM netbook will worry MS

[shutdown -p now]

It's not out of the question that a really compelling ARM netbook would scare them into ARM support. I would be surprised if they didn't have something similar to the x86 apple builds in the powerPC era.

Microsoft already has a lot of things relying on ARM - XBox for one, most WinMobile devices for another. So compilers are already there, a lot of other things like Direct3D are, too. And Windows was originally designed with several abstraction layers specifically so that most code in it would be platform-independent, even in the kernel (well, pretty much like any other modern OS). So if ARM netbooks start becoming popular, I would expect to see Win7 ARM edition very soon afterwards.

Applications? It's a problem, but not as big as you might think. Windows software developers already had to learn to deal with quirks of cross-architecture development when x64 and especially ia64 were introduced. Microsoft C++ compilers spew out warnings on any kind of code that is likely to be architecture-specific - like storing a pointer in an int - since 2003. So, most Windows applications written in the last few years are a recompile away from working on ARM. If Windows starts supporting ARM officially, you bet that recompile will happen soon enough. Of course, this doesn't solve the problem of legacy software and abandonware, but who needs that on netbooks?

Re:A compelling Linux on ARM netbook will worry MS

[TheRaven64]

The BeagleBoard uses an OMAP3430, which has a Cortex A8 core. TFA is about the Cortex A9, which will be used in the OMAP4 series.

Re:A compelling Linux on ARM netbook will worry MS

[narrowhouse]

Have you noticed how Linux is used as a club to bring Microsoft around in so many situations? Governments threaten to switch to Linux/a large discount offer from Microsoft appears. MS tries to kill XP sales, Linux notebooks appear/XP sales are extended and a leaner replacement for Vista gets fast tracked. If ARM finds some manufacturers who are willing to create and sell ARM based devices that are not too expensive and outperform comparable X86/64 devices I think the "no ARM version of Windows" would disappear.

X86 compatibility is holding everyone, including MS and Intel, back. Intel wanted to kill it with Itanium but the momentum was too much. I think Intel had an ulterior motive (too many x86 licensees to compete with them) but I think they still know they could do better, and in many ways the competition has if you just judge processor performance, price, and efficiency without regard to software base.

Re:A compelling Linux on ARM netbook will worry MS

[narrowhouse]

Linux predictions always depend on Microsoft bungling the response to a new threat. If MS had stuck with its plans to mothball XP Linux would be considerably more common on Netbooks and there would be more active interest from manufactures in support a Linux version for their machines the way the currently do for Windows. As it was MS did what they had to do to give customers a familiar Microsoft product to choose over Linux.

People will most often choose familiar products if they can, switching to something that is different but "just as good" ,or even better, isn't worth it to most people unless they are really unhappy with what they have.

The key is that Linux keeps creating more challenges for MS, forcing them to lower prices, increase performance and think of what the customers really want. Without the competitive threats Linux poses, Microsoft would normally charge more for bloated products that are upgraded when Microsoft WANTS you to upgrade. It's not their fault, that is what a complacent monopoly does. They should thank Linux, it has saved them from 2 or 3 really stupid moves.

Re:A compelling Linux on ARM netbook will worry MS

[Megane]

Microsoft already has a lot of things relying on ARM - XBox for one

BUH? Original Xbox used an x86 Celeron variant, Xbox 360 uses a 3-core PowerPC variant.

Re:A compelling Linux on ARM netbook will worry MS

[hattig]

No, the 2GHz ARM had twice the power of the 1.6GHz Atom, if you had read the article. That's 1.9W.

The 800MHz dual-core ARM matched the 1.6GHz Atom, at 0.5W.

That's one benchmark (third party, backed by both ARM and Intel), quoted by ARM. Salt, lemon, tequila, take.

Re:A compelling Linux on ARM netbook will worry MS

[Locutus]

wasn't there recently a story saying how Windows Mobile was pretty much dead at v6.5 and that Windows 7 was to be its replacement in 2 years? I saw this once and only once and was surprised to even see it because it does not sound like something Microsoft would want out there yet. Two years is a long time and pre-announcing the end of Windows Mobile would cause many OEMs to look for something else right away. Maybe this is why we're hearing that HTC is moving Android up the stack and Windows Mobile is getting pushed aside for newer models?

And don't forget that the ARM guy said they had people working at Microsoft. I would have a hard time believing it had to do with WinCE. So it appears they do not have anything other than WinCE on ARM in anyway ready for general consumption, if you consider WinCE ready. Microsoft is 100% reactionary in their tactics and rely 100% on FUD and PR tactics in the initial attack on a segment they are entering. From what I heard, Windows 7 still has a larger footprint than Windows XP but still much smaller of a footprint than Vista. Because it is still larger than WinXP, you have to wonder how they'll compete with Linux which is already smaller than WinXP and far far more capable than WinCE. But two years is a very long time.

LoB

Re:A compelling Linux on ARM netbook will worry MS

[Anonymous Coward]

What?

My Eee reaches 8 hours while I am hacking and debugging code. I turned off wireless and bluetooth and lowered the brightness to the lowest level. And I was using a USB UMTS modem, occasionally browsing the web. Other from that I just used the preinstalled Windows XP Home.

Re:A compelling Linux on ARM netbook will worry MS

[rolfwind]

I meant an Arm-based notebook is out, not this new chip, because my previous parent post said:

Whether the ARM chip performance is even adequate for normal netbook applications (e.g. watching youtube) is an open question until somebody tries it. Sure, ARM threw out this number of 5x, which is a meaningless number until we get a better overall idea of how fast and slow it is on different tasks.

Which may not give us an idea what the new chip can't do, but it will give us a good idea of the minimum the new chip can do (i.e. if the old chip can handle youtube, so can the new).

Re:A compelling Linux on ARM netbook will worry MS

[BikeHelmet]

I agree. I don't know why no company is designing an ultimate traveller's netbook.

The closest I've found is the Touchbook, which gets about 10 hours battery life while in use with screen brightness ramped up. The Pandora apparently gets between 9 and 15 hours depending on what you're doing, and whether you're willing to lower screen brightness at all.

Both of these are OMAP3 devices. Couldn't they just stick the same SoC in a netbook frame with a huge battery for awesome battery life?

I guess for now the only option is a Pandora with a bunch of extra $20 batteries.

Re:A compelling Linux on ARM netbook will worry MS

[BikeHelmet]

Well, it sort of does. Battery life and CPU power are actually somewhat convertible.

Not at all.

Keep in mind that Cortex A8 CPUs @ 600mhz consume as little as 400mw when under full load.

Even adding in the SGX(GPU) and DSP won't bump it up much. Things like LCD brightness, or presence of an HDD, have a far bigger impact.

Re:A compelling Linux on ARM netbook will worry MS

[DarkOx]

That's why the marketing people are already calling these "Smart books" so that its something new and progressive in the publics eye. Netbooks will with any luck be last years toy.

Re:A compelling Linux on ARM netbook will worry MS

[myrdos2]

Desktop Windows has become such a bloated mess that I simply don't believe Microsoft (or anyone else for that matter) is capable of porting it to another architecture. Maybe they could adapt their Windows Mobile line?

Re:A compelling Linux on ARM netbook will worry MS

[Hurricane78]

That's exactly what's the case in these new smartbooks!

They have much much longer battery life, because the ARM CPU plus the GPU (to play HD movies without stuttering and accelerate Flash) use only 1-2 watt. Instead of the 20 of an ATOM CPU.
They don't get hot, and run very very long.

Forty hours is a bit unrealistic, because eight hours is too. Thing more like 2-3 hours. Now take this times five, to get to 10-15 hours. Reduce it by the usage of the added GPU, and I think real 10 hours are actually to be expected. That's also what the people who sell those state. (Could ARM smartbook manufacturers be honest? We'll see. ^^)

They'll be avaliable in a month. For $100-$200. They will only run Linux, but you can bet that I'll get one as soon as I find one for $100.

Re:A compelling Linux on ARM netbook will worry MS

[Hurricane78]

Parent is exactly right. The nVidia Tegra does this. And I've seen it running live. HD and Flash are no problem. The manufacturers actually recognized those as must-have features, and I don't think someone will be that dumb to sell one without them, because when even cheap Chinese knock-offs can achieve that, everybody without those features will sell exactly nothing, and they know that. :)

Re:A compelling Linux on ARM netbook will worry MS

[JAlexoi]

All I see it that all is to ARM's advantage, since it allows more discrete processing units with low power consumption to be integrated closely with it.

Re:A compelling Linux on ARM netbook will worry MS

[jimicus]

Purely out of curiosity, what applications will this recompiled version of Windows run?

Re:A compelling Linux on ARM netbook will worry MS

[Anonymous Coward]

They should thank Linux, it has saved them from 2 or 3 really stupid moves.

If anything has challenged MS in this market it's Apple. Let's get real about the numbers here. Apple did in the past 2 years for their own marketshare what Linux couldn't do in 2 decades. For all the chants of free software and open source goodness there really hasn't been a move in the market to Linux on desktop. As was pointed out earlier, Linux has lost shares in the netbook market and there's no reason to think that this product won't suffer for it. Market trends are so much more insightful than any ten thousand posts by fanbois of any particular product. When the market starts to shift is when I'll be interested. Until then it's more pipe dreams of Goliath falling to David.

Technological superiority rarely makes a difference in the mass market. If you ever stopped and looked at the 80s computers wars you'd know this as well as you know anything else. I've run Linux, I've run OSX, I've run Solaris, I've run OS/2. To date I can't see any reason why to leave Windows. Not to say some others shouldn't and not to say that I even care if other OSs are out there. I don't need my OS to be the only one to feel vindicated that I'm making the right choice. I know I've made the right choice when I can shut down my PC at the end of the day and got what I wanted to do on it done without a ton of hassle. 99.9% of all users feel the same way. Playing with a system on the hobbyist level is a ton of fun but when I need to get something done it's a pain. I still run multiple OSs around my house but I go back to the same ones when I feel like doing something and not playing with a PC to get it to work the way I want.

So, sure, the competition is a great thing but this doesn't look like a competition product. It might be big for Linux but for those not in the Linux arena it's not going to be significant.

Re:A compelling Linux on ARM netbook will worry MS

[MaraDNS]

netbooks rarely use all of the processing power they have right now

Not in my experience. When I used a netbook this summer to do video calls on Skype, the Netbook's cpu could barely keep a 2-way video conversation going, and only when I closed all other applications. Trying to open up a browser window while having a 2-way video call on Skype would grind things to a halt (the browser window would never open and the video call would pause).

Netbooks need about 20-50% more horsepower than they do right now before they can comfortably do two-way video on Skype.

Re:A compelling Linux on ARM netbook will worry MS

[Anonymous Coward]

The Always Innovating Touchbook has been measured at 4 watts, while not recharging simultaneously, at the wall. That's the whole package running.

Re:A compelling Linux on ARM netbook will worry MS

[billakay]

Remember, Windows NT wasn't even developed on x86 (it was done on MIPS and Intel i860 and backported to x86). In fact, up until Windows 2000, NT had support for x86, Alpha, MIPS and PowerPC right out of the box. NT was originally designed to be very portable and its kernel design facilitates this. It's just a matter of will.

Re:A compelling Linux on ARM netbook will worry MS

[slashqwerty]

Windows on ARM runs existing Windows x86 software about as well as Linux does: not at all.

It's worth noting .Net apps would work just fine. Much like Java, they are compiled to byte code and converted to native code by the runtime environment. As more and more apps are rewritten in .Net this becomes less of an issue for Microsoft.

Re:A compelling Linux on ARM netbook will worry MS

[stephows]

If MS wanted to make an ARM Win7 they would do what Apple did when they shifted from the 68k to the PowerPC. The new Mac OS was compiled for the PPC but old 68k applications ran in a 68k emulator. Likewise, MS would compile Windows itself as ARM code (they already have hardware abstraction layers designed for MIPS, PPC, etc back in the 1990's). Old applications would run in an x86 emulator. Most application spend a lot of time waiting for events or inside system calls (which is in native ARM code), so performance would still be okay (except maybe for video apps).

Even better, I read an article about 8 years ago where DEC Alpha code was being run on an x86 system. Each small piece of Alpha code currently being run was compiled JIT and then saved back into the binary file. Over time the entire application was converted piecemeal into an x86 binary. The final binary ran almost as fast as a natively compiled application.

Re:A compelling Linux on ARM netbook will worry MS

[DigiShaman]

SoC manufacturers like Freescale have partnered with Adobe to ship custom versions of Flash that take advantage of the extra hardware on the chip for exactly this.

So when Flash player 11 is released sometime in the future, which version does the Flash updater get? The generic or one for custom hardware?

It's rather a moot point IMHO. Flash interfaces with DirectX. So if you have proper hardware and drivers, Flash will get an accelerated boost anyways. So why bother with custom Flash compilations? Is the performance improvement over DirectX really that substantial?

Re:A compelling Linux on ARM netbook will worry MS

[Anonymous Coward]

You can reach 9:30 or 10:00 hours with EeePC. Tried it myself. Yes, it involves putting Linux (Arch) on it and leaving it just standing there, not doing anything.

Re:A compelling Linux on ARM netbook will worry MS

[Bert64]

I have the eee 901 which was advertised as having 8 hours with xp, and 5 hours with linux... The reality seems to be around 5 hours with either.

Re:A compelling Linux on ARM netbook will worry MS

[Bert64]

The Nokia N800 has a 300mhz ARM, and a linux port of flash which can play youtube videos, it's fast enough for the standard videos, i doubt the HD stuff would work but the screen isnt big enough anyway.

Re:A compelling Linux on ARM netbook will worry MS

[BikeHelmet]

And the Pandora has been measured at under half that. ;)

http://www.gp32x.com/board/index.php?/topic/49385-current-drains-by-speed-and-device/page__p__751966&#entry751966

Here was an idling kernel with various backlight settings:
100% - 507mA
88% - 484mA
70% - 456mA
53% - 424mA
35% - 390mA
18% - 364mA
0% - 337mA

They use a separate chip which is accurate to single digit mA. At 3.6v, 500mA is 1.8 watts?

Re:A compelling Linux on ARM netbook will worry MS

[Haeleth]

Netbooks running Linux on an ARM processor with insanely long battery life and a true dedicated mobile operating system may be what it takes to get people to realize that netbooks were not intended to be merely smaller laptops.

Maybe the people who designed the first netbook didn't expect it to be a smaller laptop. But the fact is that the vast majority of people who buy a netbook want a smaller laptop.

Selling people what they want is the most reliable way of making money. Convincing people that they ought to want what you're selling generally only works for "lifestyle" products like carbonated sugar water or clothes with expensive logos printed on them.

Re:A compelling Linux on ARM netbook will worry MS

[hitmark]

one thing to note is that even on a desktop, flash makes use of the GPU to help speed things along.

just right-click on any flash object, select settings, and note the content of the leftmost tab, a single checkbox talking about hardware acceleration...

Re:A compelling Linux on ARM netbook will worry MS

[Anonymous Coward]

Absolute crap. RTFM son. Where M = Spec sheet. The Atom itself that you will be using (N270) draws 2.5W. The chipset is the major draw but... ... why am I even arguing with you? The Aspires (and the clone Advent machines, one of which I am using now) uses a shocking 25Wh battery. Get a decent netbook like the Asus 1000HE and you're golden for 8hrs.

Re:A compelling Linux on ARM netbook will worry MS

[TheRaven64]

That depends rather on how well DirectX runs on ARM Linux...

On Windows, Flash does not use DirectX for video decoding. It uses its own codecs. This is why playing H.264 in Flash uses a lot more CPU than playing it back with WMP (which typically uses codecs that use hardware acceleration).

Re:A compelling Linux on ARM netbook will worry MS

[Hal_Porter]

If you want more battery life, why not just buy yourself a couple of 9 cell batteries? It'll be cheaper than replacing all you applications, most likely with "work alike" ones since the originals never get ported to ARM.

Re:A compelling Linux on ARM netbook will worry MS

[petermgreen]

Code bloat is largely orhtogonal to portability. Code messiness can have an impact if it leads to assumptions that are valid on one system but not another. Generally the first port is hard and subsequent ones are easier as the platform specific assumptions are either eliminated or at least moved to platform specific code sections..

Afaict NT has run on five architectures (alpha, mips, x86, x64 and ia64) over it's lifetime and current versions are offerend for three (though for ia64 MS only offers server editions not desktop ones). This makes me suspect that it could indeed be ported given the time and resources.

A bigger issue is running x86 windows apps, an emulator would have to be included (as was done for alpha and ia64 afaict) and that emulator would need to perform very well to make windows on arm a viable platform.

Another issue is FPUs, afaict there are a load of different FPUs for arm. So if you want to make a general purpose arm OS you either have to use softfloat (as debian armel does), choose one and screw everyone else even worse than if you had used softfloat (as debian arm did) or find some clever soloution to supporting multiple FPUs without too much overhead (maybe putting floating point in a library and swapping that library out based on the FPU could work, not sure how much overhead that would add).

Re:A compelling Linux on ARM netbook will worry MS

[ckaminski]

Most applications written on Windows since the release of Windows95 are more than capable of being ported to any other Windows platform in relative ease. It's too bad the only other Windows platform available is Itanium (that PoS).

Re:A compelling Linux on ARM netbook will worry MS

[shutdown -p now]

Most applications written on Windows since the release of Windows95 are more than capable of being ported to any other Windows platform in relative ease.

I wish. Win32 API for a long time was rather 32-bit centric - for example, it was very commonplace to use GetWindowLong/SetWindowLong to store pointers associated with a window, something that breaks on x64/ia64 as pointers no longer fit into longs in LLP64. It wasn't until VS2003 (and the Windows SDK that was released about a year earlier, IIRC) that SDK headers got typedefs like LONG_PTR, and GetWindowLong was deprecated in favor of GetWindowLongPtr, so that this could be fixed.

Re:A compelling Linux on ARM netbook will worry MS

[Anonymous Coward]

The 9-cell battery makes my Aspire One significantly heavier and less comfortable to hold, I use it when I need the battery life and use the 3-cell when I don't. As for the apps, I use Linux pretty much exclusively, so what I use will most likely be ported anyway.

I love my Aspire One, but the battery life isn't great with the 3-cell battery and it gets too warm for my liking. I'm looking forward to be able to replace it with an ARM-based netbook which runs cooler and for longer.

Re:A compelling Linux on ARM netbook will worry MS

[Hal_Porter]

Well an ARM netbook would be good for Linux. The problems is the vast majority of the market is x86/x64 Windows. Even if Windows were ported, the application will not be. And frankly even if they were they'd probably run slower on a 2Ghz ARM than they would on an 2Ghz Atom.

Now ARM based phones are a great idea. I've got a Windows Mobile phone and there are loads of applications for it and performance is fine. However applications for Windows Mobile are much, much less bloated than for desktop Windows. From what I've seen desktop Linux applications have suffered the same bloating over the years. E.g. Firefox is not quick on an Atom and will be worse on an ARM.

Actually I don't really care about battery life - all my machines are laptops but I always use them plugged in. Basically I don't really buy power consumption as something which will drive a switch away from x86. Even if it started to happen, Intel could just work a bit harder on power consumption - in most Atom machines the chipset uses more power than the processor. Actually I'm not even sure if x86 is more power consuming than ARM and even if it is the power consumption of a laptop is mostly things like the display and storage. Those will consume the same power regardless of CPU architecture.

E.g. if you look at the pie chart in the Base Hardware Platform section here -

http://blogs.msdn.com/e7/archive/2009/01/06/windows-7-energy-efficiency.aspx

The CPU is only consuming about 9% of the total average power.

"Windows CE or even Windows Mobile"

[argent]

That's like saying "Linux or even Ubuntu". :)

Microsoft used to have a laptop/netbook-friendly Windows CE version back in the late '90s, but dumped it in favor of the "Tablet PC" build of Windows NT around 2000-2001. It would be interesting to see them bring that back.

Re:"Windows CE or even Windows Mobile"

[Anonymous Coward]

That's like saying "Linux or even Ubuntu". :)

Microsoft used to have a laptop/netbook-friendly Windows CE version back in the late '90s, but dumped it in favor of the "Tablet PC" build of Windows NT around 2000-2001. It would be interesting to see them bring that back.

They still do, the problem is it's shit and it won't run any off-the-shelf applications. It's used in a number of industrial PDAs, particularly ruggedized, intrinsically-safe ones.
The way I see it, using CE on a laptop is far worse than Ubuntu because it looks like windows (95), behaves (mostly) like Windows, but won't run any Windows apps. In some ways it's the perfect combination - you get all the 'It-won't-run-Outlook/Oblivion/Photoshop' problems of Linux, all the 'It-won't-work-with-my-USB-doodad' problems of OpenBSD and all of the bugginess of Windows.

And unless it's CE6 (WM and most devices are still CE5), it will have that abysmal 32MB-per-application limit, so good luck porting any substantial win32 apps to it.

Much as I'd like a linux ARM netbook, I am a little worried that they don't seem to have 64-bit addressing in that architecture yet. It won't be so many years before it becomes a needed feature for a netbook too.

Re:"Windows CE or even Windows Mobile"

[argent]

They still do, the problem is it's shit and it won't run any off-the-shelf applications. It's used in a number of industrial PDAs, particularly ruggedized, intrinsically-safe ones.

Is it? I haven't seen any devices using the Handheld PC user interface since around 2001. I'm not talking about the Pocket PC fork that Windows Mobile is based on and Symbol and others use in their PDA-form-factor devices. The Handheld PC fork required a larger screen and a keyboard, and disappeared when Microsoft decided that the way to beat Palm was to copy Palm OS (of course, with 20-20 hindsight, we know they didn't have to do anything... Palm disintegrated on their own due to corporate ADHD).

Re:"Windows CE or even Windows Mobile"

[BadAnalogyGuy]

I haven't seen any devices using the Handheld PC user interface since around 2001.

But then again, why would you? WinCE is designed to allow for fully custom shells.

That said, I would love to see a device like the old Vadem Clio that had wifi (or even 3G GPRS) and could RDP into my home PC.

Re:"Windows CE or even Windows Mobile"

[Colourspace]

You should consider changing your ID to BetterAnalogyGuy.

Re:"Windows CE or even Windows Mobile"

[argent]

WinCE is designed to allow for fully custom shells.

Indeed, and this particular shell hasn't been used on a new product in about 8 years. Which is all I was sayin', eh?

I would love to see a device like the old Vadem Clio that had wifi (or even 3G GPRS) and could RDP into my home PC.

Me too, though I'd probably load something other than WinCE on it.

I did flash my iPaq with Familiar, but it just wasn't the same without a keyboard.

Re:"Windows CE or even Windows Mobile"

[Anonymous Coward]

Ugh, been away a few days, sorry for the late reply. Still for thems that cares:

Like I said - it's used in industrial PDAs. If you're interested, look up the DAPTECH CE5000B, MaxID IDL, Psion Teklogix Workabout Pro, Casio IT3100 etc. Those are just the ones I've used. There's also similar units like the Fleetwood Blueport, and the Belgravium Atlanta.
These are all CE5.0 devices - the CASIO IT600 shipped with CE6 I think, but they seem to have discontinued it.

I program these devices for a living (unfortunately). And I swear, one of these days I'll get an account.

Re:"Windows CE or even Windows Mobile"

[argent]

I think we're talking at cross purposes.

Those devices have a PDA form factor, they're all using either a custom shell or something based on the Pocket PC shell. They're not using the Handheld PC shell that was used in things like the Thinkpad z50 or Jornada 720.

Re:"Windows CE or even Windows Mobile"

[Tapewolf]

I think we're talking at cross purposes.

Those devices have a PDA form factor, they're all using either a custom shell or something based on the Pocket PC shell. They're not using the Handheld PC shell that was used in things like the Thinkpad z50 or Jornada 720.

Okay, it took 10 years but this conversation has finally piqued my interest enough to finally get an account ;) We may possibly be talking cross-purposes, but these manufacturers aren't using a per-vendor custom shell because they're all the same. They most certainly aren't shipping Windows Mobile either. I've never used Platform Builder so I don't know for sure, but as I understand it, a stock CE build (branded as CE.Net since v4) ships with a default shell based on the old Handheld PC one and only slightly updated. WM6 is just the CE.Net core with a prettier shell. FWIW, you might want to look up the CNMbook, since that actually is a CE5.NET netbook (based on a MIPS core, I think). Smartphones and some industrial devices like the Motorola MCxx generally ship with WM5/6, but in many cases (for the Industrial units) they give you a choice of the CE.Net shell and Windows Mobile. Intermec actually sent me a survey asking if their customers preferred CE.Net or Windows Mobile. So at the end of the day, yes, HandheldPC itself is dead and buried, but there are still a number of devices shipping with the same basic shell, albeit on a CE5 kernel rather than CE2.11.

Re:"Windows CE or even Windows Mobile"

[argent]

We may possibly be talking cross-purposes, but these manufacturers aren't using a per-vendor custom shell because they're all the same.

OK, it was hard to see that from the pictures... some of them have the bar hidden. I'll take your word for it.

a stock CE build (branded as CE.Net since v4) ships with a default shell based on the old Handheld PC one and only slightly updated.

It sure looks like the Palm-PC/Palm-sized PC/Pocket PC shell, not the Handheld-PC one.

FWIW, you might want to look up the CNMbook, since that actually is a CE5.NET netbook (based on a MIPS core, I think).

Interesting. I really thought Microsoft had killed the Windows-CE-based Handheld PC. I've seen devices described that way since, but they've all been NT-based. It's an ARM, BTW. Does anyone still use MIPS in handhelds?

Re:"Windows CE or even Windows Mobile"

[Tapewolf]

It sure looks like the Palm-PC/Palm-sized PC/Pocket PC shell, not the Handheld-PC one.

Here's an example, if you're curious:

http://mylostblog.altervista.org/i/windows-ce-6-mlb.png

...start menu at the bottom in the left corner, and a taskbar. I think this shot is running in an emulator rather than a physical device, though.

Interesting. I really thought Microsoft had killed the Windows-CE-based Handheld PC. I've seen devices described that way since, but they've all been NT-based. It's an ARM, BTW. Does anyone still use MIPS in handhelds?

Ah. There is conflicting information about the CNMbook CPU. Some say ARM, some say MIPS.

My guess would be that they switched in later models. They have Linux and CE.NET versions of the thing, so they may have had an earlier MIPS/Linux one or something. Just a guess, though.

As for MIPS in CE handhelds, that certainly seems to be dead. WM2003 and onwards are only for ARM, though there are also the occasional THUMB-only devices like the Casio IT500 (which yes, used the Win95-alike shell).
I don't honestly know if MIPS is supported for CE.NET itself or not.

Re:"Windows CE or even Windows Mobile"

[argent]

Your link sends me to a placeholder (1x1 "void.gif").

To me the Palm[sized] PC shell appeared vaguely similar to the Handheld PC shell, but it was stripped down with all the root windows stuck at full screen and no widgets to detach or minimize them. That was the distinction I was getting at. Anyway, if there's still a viable Handheld PC code base in active development at Microsoft it seems really strange that they're not pushing it on linux-friendly netbook makers.

Ah. There is conflicting information about the CNMbook CPU. Some say ARM, some say MIPS.

Their web page says the Windows CE version is ARM, and the older Linux one is MIPS. Good guess, confirmed by Google.

Re:"Windows CE or even Windows Mobile"

[Tapewolf]

Your link sends me to a placeholder (1x1 "void.gif").

To me the Palm[sized] PC shell appeared vaguely similar to the Handheld PC shell, but it was stripped down with all the root windows stuck at full screen and no widgets to detach or minimize them. That was the distinction I was getting at.

Try this: http://www.bluechiptechnology.co.uk/files/CE6UProgrammersUserGuide.pdf

...it has a few screenshots of the thing. See around pages 5 and 18.

Windows Mobile definitely does the fullscreen-only thing, and that seems to be one of the key differences from the bare CE.Net shell.

Anyway, if there's still a viable Handheld PC code base in active development at Microsoft it seems really strange that they're not pushing it on linux-friendly netbook makers.

I imagine they will if the ARM/Linux platform takes off, but like I said earlier, it looks like Windows 95 and doesn't work with any off-the-shelf apps or USB devices more complex than a memory stick (if that).
If Linux is enough to make people return the devices, even with OpenOffice etc installed, Windows CE certainly will. And that might be one reason they aren't actively pushing it.

Re:"Windows CE or even Windows Mobile"

[argent]

Windows Mobile definitely does the fullscreen-only thing, and that seems to be one of the key differences from the bare CE.Net shell.

Goes all the way back to '98. The Cassiopeia E-10 I inherited from my disgruntled (at Casio, not me) boss at Ferranti had no way to switch apps from full screen to windowed mode. Seems to be a standard thing with the QVGA "Palm[-Sized] PC" devices.

Re:"Windows CE or even Windows Mobile"

[DRACO-]

Perhaps it is time for microsoft to lose the 'windows' moniker and start something new on ARM.

All anyone needs now a days is a terminal applications to get into whatever else they need to do work on. Half of the company I work for use remote desktop to process reports and dont need hefty processors.

But, does it run DOS?

[Locke2005]

If it ain't X86 compatible, it ain't shit. 99% of those netbooks are running Windows XP; will this new ARM processor do that?

Re:But, does it run DOS?

[Anonymous Coward]

It's a feature. This time microsoft is going to have a bit harder time destroying the netbook concept.

Re:But, does it run DOS?

[TheRaven64]

People are willing to pay $15 more for XP (the cost of an XP Netbook license), but are they willing to pay $100 more for Windows (the difference between the cost of the announced ARM-based netbooks and a typical x86 model)?

Re:But, does it run DOS?

[poetmatt]

uh what? I didn't read anything saying it was x86 incompatible, merely "can't run windows right now". This is more a licensing issue - MS will have to make a case as to why they don't want to support ARM, and as the AC says will affect that.

Can't say I saw coming that ARM might be an Intel/AMD competitor, but depending on how things shape up that could be the case in a solid fashion.

Re:But, does it run DOS?

[Dog-Cow]

You are not all that bright. Some might even call you an idiot.

The ARM instruction set is not x86 compatible. End of story.

Re:But, does it run DOS?

[Anonymous Coward]

There has been an X86/MS-DOS emulator for ARM since 1987, I used it then to make my Modula-2 assignments :)

http://www.cbronline.com/news/ms_dos_for_acorn_archimedes

Re:But, does it run DOS?

[Anonymous Coward]

ARM processors (including Cortex series) have an entirely different instruction set from 80x86's. That's pretty much the definition of x86-incompatible. Whether it runs windows is an entirely different matter (but don't hold your breath on that either).

Re:But, does it run DOS?

[poetmatt]

ARM is not x86, I know that. The issue here is more about the emulation. Nothing says ARM can't be emulated on X86, except microsoft currently. Save the ad hominem.

Re:But, does it run DOS?

[WEqR0lDRR6I]

Remember FX!32 ?

Re:But, does it run DOS?

[Anonymous Coward]

uh what? I didn't read anything saying it was x86 incompatible,

Uh yeah it did. It said ARM. It's only a licensing issue in as far as the fact that there is no product that they could license.

Re:But, does it run DOS?

[AlecC]

We are already wondering how fast a native ARM system would be. Emulation on a netbook is an absolute no-go: you lose ten times over the gains you made by putting in an ARM. They are very, very differnt architectures: I would expect a 20:1 or worse slowdown by running in emulation.

Re:But, does it run DOS?

[poetmatt]

sorry, my own error, I meant interpreter - isn't there some way to do this at a hardware level underlying the operating systems themselves?

Re:But, does it run DOS?

[AlecC]

No, not on conventional hardware such as ARM and the Atom. The (failed) Transmeta system was intended to do something like this - but they didn't pull it off. Each CPU has its own very hard-wired instruction set, and the one thing it does well is executing that instruction set. Every layer you add on top of that consumes power and nanoseconds: the art of OS design is to consume that to the best effect.

Re:But, does it run DOS?

[Anonymous Coward]

ARM processors already ran DOS emulated 20 years ago. There was even a PC emulator. If I recall correctly the ARM2 running at 8 MHz was able to emulate a PC running at about 4 MHz

Re:But, does it run DOS?

[poetmatt]

So would emulation be worse or would having some kind of northbridge interpreter or equivalent to a math coprocessor to translate from arm->x86 and vice versa be more efficient?

Re:But, does it run DOS?

[james.m.henderson]

Dog-Cow: Despite the fact that you are right, there really is no need for adding insult to the correction. Simply stating that the parent is incorrect and giving a correction is enough for rational discourse. Furthermore, you really don't have enough evidence to assert that the parent isn't bright; you only know that they were talking about something they didn't know about. Maybe they are very knowledgeable in other areas. Maybe they aren't very old and haven't had time to learn a lot of things. Adding insult only hurts someone else and does not add anything of value to the discussion. Perhaps this is a needlessly long post on something most don't care about. However some people love to toss in ad hominem attacks, and they only go to weaken the argument being presented (if one reads it critically) and increase hostility between parties.

Re:But, does it run DOS?

[Anonymous Coward]

Would the ARM processor still outperform the Atom if it had to emulate the x86 instruction set? Probably not, so Intel would still win out even if the processor wasn't as powerful.

Re:But, does it run DOS?

[julesh]

are they willing to pay $100 more for Windows (the difference between the cost of the announced ARM-based netbooks and a typical x86 model)?

Not to mention probably sacrificing half the battery life.

Re:But, does it run DOS?

[Locke2005]

the ARM2 running at 8 MHz was able to emulate a PC running at about 4 MHz Takes twice as many cycles to do anything? That should do wonders for battery life!

Re:But, does it run DOS?

[jimicus]

By the time you put in hardware to do that (and your idea has existed for years - you used to be able to buy a 486 chip on a card to stick in your Acorn Archimedes and you could run Windows 3.1 in a window on RISC OS. Software emulation at the time was only able to run DOS. Slowly.), you'll find that it's cheaper simply to build a straight x86-based netbook.

Re:But, does it run DOS?

[Kumiorava]

Why not use Intel/AMD processor to run x86 code then? Alternatively Microsoft could port Windows to run on Java or some other virtual language and then just build virtual machines on all target platforms... but the performance still wouldn't be there.

It's just easier to do the straight forward thing and port the code to ARM. I don't think that majority of Windows code is in any way dependent on underlaying platform and therefore doesn't need to be ported just compiled.

Re:But, does it run DOS?

[Anonymous Coward]

ARM *can* be emulated on x86, for example using qemu. Microsoft hasn't done much to try to stop that.

Re:But, does it run DOS?

[Anonymous Coward]

>> Nothing says ARM can't be emulated on X86, except microsoft currently.

Wtf? None of your posts make any sense. First off why would Microsoft be able to say ARM can't be emulated? Second, why would you want to emulate ARM on x86? Did you mean emulate x86 on ARM, to run Windows?

And most importantly, while it's possible to emulate x86 on ARM, the issue is speed. We're talking hundreds of times slower here. You can't run an operating system or an office suite like that. Here's an interesting tidbit towards that actually: if you've ever done any mobile development on ARM (e.g. iPhone, BREW, Windows Mobile) you'll have used a device simulator. Well turns out all these simulators are not emulating ARM; it's just too damn slow. You actually need a different build for simulators because you have to compile them to x86; the simulators run x86 builds instead of emulating ARM.

Re:But, does it run DOS?

[laederkeps]

And do you think Microsoft will allow Acer, Asus, etc. to sell their ARM netbooks that much cheaper? Seriously?

Re:But, does it run DOS?

[EXrider]

If it ain't X86 compatible, it ain't shit. 99% of those netbooks are running Windows XP

I know I don't represent the masses of benevolent Windows netbook users, but I could personally give two shits about a netbook running Windows. As long as it has a decent mainstream browser, full Flash plugin support, and has the balls to play full-screen video smoothly with the latest codecs, I don't care if it's running Linux. I don't want to run Quicken or Outlook on the damn thing, that's why you have VPN and an RDP client on it to connect to your real computers as needed.

Re:But, does it run DOS?

[DaVince21]

If it ain't X86 compatible, it ain't shit.

You're right, it's not shit. Getting Windows to work on a different CPU architecture is up to Microsoft, not ARM.

Re:But, does it run DOS?

[Locke2005]

You're correct in that if all you want the netbook for is web access, Linux should work fairly well. However, aren't the latest version of Flash and most other Browser plugins both X86 and Windows specific? And aren't some #$%! webpages still pretty much IE only? (Yes, I'm still upset that my Android G1 doesn't do Flash!)

Re:But, does it run DOS?

[EXrider]

aren't the latest version of Flash and most other Browser plugins both X86 and Windows specific

Nope, there're also Flash plugins for Nokia n-series (which I believe is Symbian OS running on ARM), x86 Linux, PPC Mac OS, x86 Mac OS... someone even mentioned their Archos media tablet running an ARM CPU with a full-featured Flash plugin.

And aren't some #$%! webpages still pretty much IE only?

Most of these are things like Intranet sites, and other poorly supported corporate garbage. I can only think of one website I use regularly off the top of my head: hp.com B2B is awful, it works in other browsers, but it is buggy as HELL. I haven't encountered any other websites in years that are IE only.

Porting code to a new architecture

[Xocet_00]

What is involved in porting code to a new chip? I've done some programming in my life, but it has mostly been limited to personal interest and school projects. I imagine it can't be as simple as just recompiling. So what does it take to port code?What are the hurdles? Assume (accurately) that I'm a total noob.

Re:Porting code to a new architecture

[BadAnalogyGuy]

Let's say 30 years ago I gave you one end of an infinitely long piece of yarn and told you to start knitting a sweater. At first, it's not too bad. The yarn has a pretty standard consistency, although it sucks compared to some other yarn on the market. Then I start changing things up. Adding some knots and tangles in the yarn I hand to you. You do your best to accomodate and actually come up with a pretty nice sweater. Then you start re-designing the sweater to take advantage of the knots and tangles, and I just keep putting more and more complex knots in there since you seem to be doing great with the ones I've sent so far. Your sweater grows thick with piles of yarn and by the time 30 years rolls around, you've got yourself a pretty great sweater. Of course, you had some massive screwups like sweater ME and sweater Vista.

Now let's say I ask you to knit the same sweater using a beautifully crafted roll of thread.

I think you can see how hard that would be.

Re:Porting code to a new architecture

[bezenek]

What is involved in porting code to a new chip? I've done some programming in my life, but it has mostly been limited to personal interest and school projects. I imagine it can't be as simple as just recompiling. So what does it take to port code?What are the hurdles? Assume (accurately) that I'm a total noob.

The main issue will be handling of virtual page tables in the OS. The code for x86 will not work for ARM.

There will be some other issues with the boot sequence (BIOS), and of course the need to be able to drive the devices attached to the ARM, as some people have noted with respect to getting Linux to run on the ARM-based netbooks.

I expect the code base for Windows has been hacked for so many years by so many different people that moving it to another architecture would not be as easy as it is for Linux. But, a good team should be able to do it. Perhaps Microsoft took this possible future into consideration and cleaned up a lot of the code when they did Windows 7.

-Todd

Re:Porting code to a new architecture

[pikine]

It depends on how the original code is written. In a well-structured OS like Linux and NetBSD, they isolated the idiosyncrasies of the CPU and focused on using common high-level features across most CPUs (memory paging and interrupt handling among the chief of them), and can optionally adapt when a particular feature is not available on some architecture (e.g. high resolution timer, atomic instructions). In such case, porting to a new architecture just entails writing the assembly language glue that bridges high-level hardware feature with the machine instructions that does the real work.

But among the worst things you can do is to hard-code low-level hardware handling and scatter that throughout the source code. Or some important code may be overly dependent on CPU specific feature (e.g. task gate for intel x86) which makes it essentially non-portable. I've never seen the source code of Windows, but I suspect this is the case with them. According to some Windows NT Internals book I read many years ago, it started out well-structured, with a nice hardware abstraction layer and all that. But since Windows dropped Alpha processor support, I think the abstraction started to suffer bit-rot and made things much worse than if they had no abstraction at all.

Re:Porting code to a new architecture

[ianare]

It all depends on how close you're getting to the hardware, and to what extent you're using specific hardware features of a chip. In some cases it really is as simple as tweaking a few things and recompiling, if the software was designed to be portable. The original windows NT (from which 2000, XP, Vista, and 7 are all evolved from) was designed to be portable, in the early days it ran on RISC also.

Re:Porting code to a new architecture

[Anonymous Coward]

I build embedded code for a living. And we almost exclusively use Linux on our platforms. And with the exception of drivers everything I build compiles for x86 ARM and the older PowerPC we use. If you use C (or C++) then there are just a few things you have to watch out for.

Main thing you have to watch out for is endianness. This page explains it: http://www.ibm.com/developerworks/aix/library/au-endianc/index.html?ca=drs-

Other then that it's just a recompile and run. As soon as you port to Windows you have more problems.

Re:Porting code to a new architecture

[TheRaven64]

It is either a complete rewrite, a simple recompile, or something in the middle depending on how different the architectures are. In terms of programmer-visible features (ignoring things only visible to compiler and OS writers), ARM and x86 are very similar; same word size, almost identical alignment constraints, same byte order. If you wrote your program in a high-level language, it is just a recompile. If you used any assembly language, then you will need to rewrite it. If you used a language somewhere in the middle, like C, then it will probably be a straight recompile. This is unlike porting, for example, from x86 to SPARC64, where you suddenly have very strict alignment, opposite byte order, and different

Of course, this is assuming the operating system interfaces are the same. If you're on something like OpenBSD, for example, then the OS does a good job of isolating the userspace code from having to know anything about the underlying architecture. Linux, on contrast, exposes a lot of architecture-specific details to programmers (and that's ignoring the fact that embedded Linux often ships with a non-GNU libc, which lacks a lot of features). Wince is about the worse at this, where every single device implements some subset of the Win32 APIs and so you end up having to do some tweaking for every device.

Re:Porting code to a new architecture

[psbrogna]

I've done a token amount of cross-platform work and will say that the recompiling your own code for a different target CPU is the "easy" part (relatively)- it's the retooling the code for different desktop widgets that can be a back-breaker. Cross-platform widget libraries exist but aren't always used, used well, or received well.

Re:Porting code to a new architecture

[AceJohnny]

In an ideal world, you just recompile your C code.

In the real world, your code (indirectly) uses low-level libraries and system calls that only work on a given chip because they make use of specific hardware, either through assembly or through hardware-mapped structures.

More rarely, some basic C operations don't work as expected. A common gotcha on previous ARM architectures were that all memory accesses had to be 32-bit aligned (it saved transistors and power). That meant that you couldn't use a char[] array, you necessarily had to use an int32_t[] from which you'd then extract each byte.

Re:Porting code to a new architecture

[oodaloop]

Once again, you live up to your name. Well done.

Re:Porting code to a new architecture

[FourthAge]

Another ARM gotcha is that "char" is "unsigned" unless you specifically make it "signed", because "unsigned char" can be manipulated more efficiently by the instruction set. This is not what C programmers usually expect, although it is permitted by ANSI C. It can cause some interesting bugs.

Re:Porting code to a new architecture

[timeOday]

Porting Windows itself is almost irrelevant. The tens of thousands of apps in the Windows ecosystem still wouldn't work.

Re:Porting code to a new architecture

[Knitebane]

I'm not sure I understand your analogy.

Does the sweater go on a car?

Re:Porting code to a new architecture

[Sabathius]

*wipes a tear from his eye* They should have sent a poet.

Re:Porting code to a new architecture

[MrMr]

I imagine it can't be as simple as just recompiling
It is exactly that simple. If you don't get trapped by a particular vendor to use a deliberately unportable C dialect.

Re:Porting code to a new architecture

[Anonymous Coward]

> ARM and x86 are very similar; same word size, almost identical alignment constraints

x86 doesn't have alignment constraints, ARM does (as do most RISC chips). If you want to do an unaligned read on an ARM, the compiler has to load two words then shift/mask/or. Because this is slower than an aligned read, it only does this if it has reason to believe that the data may be unaligned (e.g. if the pointer has the "packed" qualifier).

Re:Porting code to a new architecture

[Anonymous Coward]

This has nothing to do with arm and everything to do with the compiler.

Re:Porting code to a new architecture

[xilun]

That meant that you couldn't use a char[] array, you necessarily had to use an int32_t[] from which you'd then extract each byte.

uh. That rather seems to mean that you were not using a valid C compiler.

Re:Porting code to a new architecture

[TheTurtlesMoves]

Intel is not that bad. Oh wait... yes they are... ;) I still can't believe the market is still stuck on x86 crap....

Re:Porting code to a new architecture

[Anonymous Coward]

Another ARM gotcha is that "char" is "unsigned" unless you specifically make it "signed", because "unsigned char" can be manipulated more efficiently by the instruction set. This is not what C programmers usually expect, although it is permitted by ANSI C. It can cause some interesting bugs.

I fail to see why unsigned is manipulated more efficiently by the ARM instruction set, as byte loads are available in signed and unsigned variants (LDRB / LDRSB). SIMD instructions are also available in both signed and unsigned variants.

Implementation wise, some cores may add an extra pipeline stage for "Sign Extension" processing (or endianism reshuffling) which will increase your load-use-penalty by 1 for signed variants. But I doubt that's what you are referring to.

Re:Porting code to a new architecture

[Anonymous Coward]

on previous ARM architectures were that all memory accesses had to be 32-bit aligned (it saved transistors and power). That meant that you couldn't use a char[] array, you necessarily had to use an int32_t[] from which you'd then extract each byte.

As far as I remember you were always allowed to access bytes at "non 32-bit aligned' memory addresses.
I don't believe you have to manually break 32-bit data when accessing bytes.

Regarding word accesses: on earlier processors, issuing an unaligned load would just... rotate your 32-bit word. But it was legit.
Starting v6 cores, unaligned accesses are supported, but clearly suffer from a performance and power hit (imagine going twice to the cache and/or on the external bus ?)

Re:Porting code to a new architecture

[Simetrical]

But among the worst things you can do is to hard-code low-level hardware handling and scatter that throughout the source code. Or some important code may be overly dependent on CPU specific feature (e.g. task gate for intel x86) which makes it essentially non-portable. I've never seen the source code of Windows, but I suspect this is the case with them. According to some Windows NT Internals book I read many years ago, it started out well-structured, with a nice hardware abstraction layer and all that. But since Windows dropped Alpha processor support, I think the abstraction started to suffer bit-rot and made things much worse than if they had no abstraction at all.

Windows still supports three architectures: x86, x64, and IA64. IA64 is probably different enough from the others to keep them at least somewhat honest, no? Windows Internals, Fifth Edition definitely says there's still a wall between the HAL and the rest of the OS.

So I imagine Windows wouldn't be too hard to port to ARM, given Microsoft's resources. But without any of the applications, it would be at an incredible disadvantage.

Re:Porting code to a new architecture

[FourthAge]

I seem to recall it's not because of the load instructions at all. It's because of arithmetic instructions.

If you use arithmetic operations on a signed char, you have to sign-extend the result using two operations, e.g.:

mov r3, r3, lsl #24
mov r3, r3, asr #24

If, however, you use an unsigned char (the default), then you only need one operation to mask out the high 24 bits.

and r3, r3, #255

That's the difference. If all chars were assumed to be signed, programs would be (unnecessarily) slower.

Re:Porting code to a new architecture

[FourthAge]

That's true. It is also the behaviour of both "gcc" and ARM's reference C compiler, so I think it's fair to mention it here.

Re:Porting code to a new architecture

[Scott Francis[Mecham]

As hard as..pulling this "thread", while you walk away?

Re:Porting code to a new architecture

[dkf]

If you're on something like OpenBSD, for example, then the OS does a good job of isolating the userspace code from having to know anything about the underlying architecture.

For the real architecture differences, such as fundamentally different ways of laying out data in memory? Nothing isolates that. Going between endiannesses can cause a lot of problems for poor code, and a lot of otherwise good code doesn't do to well on the 32-bit/64-bit transition. Byte ordering of IEEE floating point is another issue again (like it is for integers, but actually worse). The issue is that fundamental architectural features tend to end up as assumptions in the program.

Re:Porting code to a new architecture

[Kumiorava]

Most of the code you would come across as an application developer will be just recompiled to support new architecture if all other things stay the same. In case of operating systems things get a bit more complex as you need to think about hardware specific optimizations, new hardware abstraction layer code and fix any silly platform specific assumptions. It's not as bad as it sounds, bulk of the code is not related to the underlaying hardware at all. That's why hobbyist have been able to port operating systems / drivers (porting to new hardware) and other applications (porting to new OS) on various new platforms.

Re:Porting code to a new architecture

[TheRaven64]

Modern x86 chips also can't do unaligned loads / stores quickly; you receive a performance penalty if you use packed structures. This is exactly the same situation on ARM if you are not writing assembly code; the compiler will sort out the details for you, and it will be slower if you are requiring unaligned accesses. Unaligned loads / stores also have some very nasty gotchas on x86; they can be split across cache lines, which means that they have very different atomicity (and, on at least some chips, the LOCK prefix doesn't work correctly with words that are spread across cache lines).

Re:Porting code to a new architecture

[Curate]

ARM doesn't understand C code. A C compiler can treat char however it likes. I suspect you are talking about a particular C compiler you've encountered that happens to treat chars as unsigned by default when targetting the ARM platform. That's saying something about that particular C compiler, not the ARM platform in general.

Re:Porting code to a new architecture

[wvmarle]

I suppose it will partly depend on the language used. I've been programming mainly in Python, which means you're quite far from the hardware. In the documentation you can see here and there some OS dependencies, but those are generally minor and quite rare. I don't recall having ever seen processor-dependencies. So most python software will run on OS-X, Windows, Linux, etc on basically any processor supported by the OS and the Python interpreter. Which is really cool, this allows me to write programs on my Linux/x86 system and on my OSX/ppc system without having to worry really about the OS let alone the underlying hardware.

If you've been using something like C then you're closer to the hardware so you may run into hardware specific issues. And you will certainly run into OS specific issues as your code talks directly to the OS. I haven't used C myself so don't know too much about that really.

Re:Porting code to a new architecture

[DoctorBit]

This is why I read slashdot. *wipes tears from eyes and gets up off the floor*

Re:Porting code to a new architecture

[Anonymous Coward]

That's the "all the world is x86 attitude", but even Kernighan (or Ritchie, can't remember) claimed that they should have made
char unsigned by default. And unsigned chars are also more efficient on x86. BTW the only machine instruction which uses
a char as an index on x86 (xlat, really only useful in 16 bit mode) does not perform sign extension. The fact that char is signed
on x86 is a mistake, I don't where it originated, but the standard argument that char should be signed because short, and int and long are signed is flawed: if you call it char it means the intention is to use it as an enumeration to represent symbols and enum
by default start from 0 and go up to, they do not jump from 127 to -128! Things would be different if it had been called a "short short int"; naming is important.

Really all code should be tested on machines of both endianness: PPC is big endian with unsigned char, and even then I have few bugs to report on my Debian on a PPC machine.

Re:Porting code to a new architecture

[XDirtypunkX]

The NT kernel (including the virtual memory subsystem) was designed to be abstracted enough for easy porting by re-writing the device specific parts, so I don't think handling virtual page tables will be that hard (although, code above that relies on 4k page assumptions etc is not going to be happy if you use different page sizes etc). You're not easily going to be able to port to ARM cores that don't have an MMU, but I don't think anyone is talking about targeting Windows to somewhere without an MMU.

 

Re:Porting code to a new architecture

[Zoxed]

> It is either a complete rewrite, a simple recompile, or something in the middle depending on how different the architectures are.

Unfortunately porting can also reveal hide bugs in existing code. Recent experience of porting a C/ADA app from Solaris 6 to 8 (i.e. no arch. change) that was rock solid on Solaris 6 revealed: use of non-thread safe libraries, functions with undefined return values, multiple buffer overruns, etc. :-( All fixed and working now, but a lot more work than expected (re-compile and test).

Re:Porting code to a new architecture

[FourthAge]

The two compilers I am thinking of are "gcc" and "armcc".

Re:Porting code to a new architecture

[Bert64]

Intel are not the problem... I doubt they like x86 very much, but they are stuck supporting all that legacy cruft... They tried to move to IA64 remember.

It is closed source software which is the problem...
You have old apps which are no longer developed, good luck getting those ported.
The catch-22 situation with commercial apps, no users = not viable to port apps, no apps = no users...

If you want to run open source, non x86 architectures are great, if you want commercial apps then non x86 is pretty useless these days.

Re:Porting code to a new architecture

[Thundersnatch]

According to some Windows NT Internals book I read many years ago, it started out well-structured, with a nice hardware abstraction layer and all that. But since Windows dropped Alpha processor support, I think the abstraction started to suffer bit-rot and made things much worse than if they had no abstraction at all.

You don't know what you're talking about. Windows still runs on three separate architectures: x86-32bit, x86-64bit, and IA64 (Itanium). The hardware-specific stuff in Windows is generally limited to NTOSKERNEL.EXE, HAL.DLL, and device driver files. That's about it.

So, maybe 0.001% of the code in Windows would need to be touched to port Windows to ARM. Microsoft is not stupid. The real issue is with 3rd-arty applications and drivers - almost none of these are compiled for anything but x86-32bit. Some server applications and OS utilities have x86-64 or IA-64 versions.

Finally, supporting another hardware platform means dealing with a whole new set of device drivers, and buggy hardware, which is where most of the stability "bugs" people encounter in windows actually reside. So the market has to look fairly profitable for MSFT to jump in. The reason they support Itanium is because x86-64 wasn't quite there yet and it enabled the sale of very expensive ($25K/socket x 8+ sockets) SQL Server Enterprise edition licenses.

Re:Porting code to a new architecture

[uninformedLuddite]

It's been many years since I was a programmer. Think PDP-11, TSX, RT11, D, RPL, Filetab, C and you get an idea of how out of date I am. I would have hoped by now that if the OS ran on the hardware then the majority of apps written for that OS would just work. Have we come that far in 30 years?

Re:Porting code to a new architecture

[pikine]

Actually, you're the one who doesn't know what you're talking about. The NTOSKRNL.EXE (you misspelled it) and HAL.DLL you mentioned are relics of the old hardware abstraction, and these are no indicators of well-structured code.

When you write code, scattering #ifdef throughout your code is not a good abstraction. If I were Microsoft, and the number of platforms I need to support are miniscule, I could get away with #ifdefs everywhere else because it's the path to least resistance to get a product out. Also, some the device drivers that have no business being platform dependent (e.g. USB stack, network stack, etc) might still be just because I could get away with it. Again, without access to Windows source code, I can't really tell if that's the case with them. Didn't I make it clear that the key point is, it depends on your software engineering? What part don't you understand?

You can defend Microsoft from the cost perspective as much as you want, and you must acknowledge that this is no more than a speculation than what I said about their source code. I think it's pretty clear to anyone that the trend is to have one operating system to run on multiple classes of devices like Linux and Mac OS X, which both run on any device from mobile phones through servers. It's not about porting anymore. It's about engineering software that can easily adapt to different requirements, and this must be done as a continuous effort, not only when you need it.

Being profit driven as Microsoft is, this might actually be their grave.

Re:Porting code to a new architecture

[tepples]

The sweater goes on the car's driver.

Re:Porting code to a new architecture

[petermgreen]

This is exactly the same situation on ARM if you are not writing assembly code; the compiler will sort out the details for you
Wrong, if you have a pointer to a type the compiler will assume it follows the alignment rules for that type. If you are lucky (high error settings on the compiler and direct conversions) the compiler may error at the pointer conversion that produced the troublesome pointer conversion but in many cases it won't and you usually get silent corruption (though I believe on some chips it is also possible to trap unaligned access attempts from the kernel).

Re:Porting code to a new architecture

[Ben Hutchings]

MS still supports Windows Server on Itanium, and the X-Box 360 runs a somewhat modified NT kernel on PowerPC (even big-endian!). So I think the HAL is still useful to MS.

no windows?

[uncreativeslashnick]

This won't end well. I have an ARM device (nokia n810) and it's great. But Wintel monopoly will kill this just like it did Sparc and IBM Power. I'm sure if it's as good as they claim it'll carve out a niche, but it won't directly compete in numbers or presence with intel CPUs.

Re:no windows?

[PaintyThePirate]

I wouldn't be so sure about that. There are significantly more ARM devices out there than x86, Sparc, and Power combined.

Phone like devices are getting larger and more powerful, and laptops/tablets are getting smaller and lower power. It is converging on a market space where ARM has no competition, and is exactly where the A9 would thrive. Microsoft is even entering the game with the Zune HD packing an Nvidia Tegra. This is not a low volume niche either. Think of the iPhone, Android devices, PSP, DS/DSi, Windows Mobile phones, etc.

That is just on the mobile end too. It makes no sense to stick Windows Embedded and a Celeron in a router, network storage, or a printer when Linux/A9 is cheaper and as powerful.

Re:no windows?

[jabjoe]

I think your wrong. It's different this time round. First the market is normal consumers, secondly, much of what people do now is on the web, so doesn't matter what platform you on (bar the whole Flash issue, but there is a ARM Flash). If it's cheap, has a long battery life, plays music and videos, some games, has a web browser so they can get on facebook, they will be more than happy. Windows big draw is it's software base, but that matters less and less as the free stuff is so good now, and so much is web anyway. If anything Windows biggest weakness is it's software base, much of it is fat, people install crap from all other the place, and before you know it a new machine is uselessly slow.

Re:no windows?

[FourthAge]

But ARM does have competition. Atom. Currently, Atom's power consumption is higher, but that will change. There is nothing special about the ARM ISA that makes it require less power.

Intel and ARM are going for the same market space: powerful smartphones, netbooks with long battery life. ARM has a head start in the "low power consumption" business. But Intel has a huge advantage in the "application compatibility" business. Intel will catch up with low power consumption. What then for ARM?

In my view, ARM have crippled themselves by refusing to adopt hardware support for the x86 ISA, which will be necessary in order to compete in this area of the market. Without that, and the application compatibility that it brings, ARM will be relegated to low-end phones, embedded CPUs and microcontrollers.

Re:no windows?

[TheRaven64]

I wouldn't be so sure about that. There are significantly more ARM devices out there than x86, Sparc, and Power combined.

This is not true. PowerPC is doing very well; it is in every current-generation console and most new cars. If you buy a BMW, you are getting something like 40 PowerPC chips for the various control functions. In automotive and industrial control applications, PowerPC is the dominant player. SPARC is doing less well, although it has, I believe, the highest market share once you leave the atmosphere (radiation-hardened SPARC chips are very popular on satellites, helped a lot by the fact that ESA funded the development of open source SPARCv7 designs).

Re:no windows?

[TheRaven64]

There is nothing special about the ARM ISA that makes it require less power.

Yes there is. The ARM instruction set is simpler to decode than x86, which means that the (fixed) power cost of the instruction decoder is higher on x86 chips (you can't turn off the decoder as you can, say, the FPU or an adder while not in use because it's always in use unless the whole CPU is in power-saving mode). The Core 2 has to do a lot of clever stuff with the x86 instruction set because it doesn't match up at all well to a modern microarchitecture; not only does it split complex instructions into smaller operations, it also has to combine sequences of micro-ops into things that can be executed. Atom doesn't do any of these things, so it is a lot slower (per clock) in an effort to save power. ARM also gets to cheat a lot with things like Thumb code. This is a simpler, 16-bit ISA, which achieves very good cache density at the cost of some flexibility. You can switch ISA on a call with ARM chips, so you can have some routines in Thumb code and some in the full instruction set. Unlike the compression that Intel gets from a variable-length instruction set, this helps power saving because you can turn off the thumb decoder when it's not in use (and turn off the other instruction decoders when in thumb mode).

And that's ignoring things like the predicate instruction and barrel shifter that make ARM code denser and more cache-friendly than x86 code (which has the same advantage over something like SPARC). This means that ARM chips can get away with smaller instruction caches, which saves power.

If you want a more detailed explanation, Jon Stokes does a good job of explaining the advantages ARM has over x86 in his analysis of the Atom.

Re:no windows?

[xilun]

Power is alive thank you very much for inquiring. It is used in more or less critical systems you probably don't want to hear about, because if you do that would probably mean something wrong happened, with various bad consequences. Also used in the Wii, PS3 and XBox 360 (the last running a NT kernel derivative IIRC)

Re:no windows?

[serviscope_minor]

If Wintel killed Power, then why do Microsoft (you know the makers of Windows) use a chip made by IBM in the XBox360. Guess which instruction set it uses...

Re:no windows?

[JAlexoi]

SPARC and IBM's Power are alive and well. You must mean the desktop variants of those? And last time I checked IBM PowerPC processors power Microsoft XBox360, Nintendo Wii and Sony PalyStation(PS3 processor is PowerPC at it's core). And SPARC processors are used in the server market. And PowerPC on desktop was killed by Apptel, not Wintel.

Re:no windows?

[Hurricane78]

Well, if what you stat comes true, it's because of naysayers like you. Get some balls, and try again. ^^

Wait, I'll write it just like you, but the other way around. Let's see if it then looks realistic to you. (If not, then yours doesn't too.)

This will end great. I have an ARM device (nokia n810) and it's horrible. Wintel monopoly has got no chance, unlike with Sparc and IBM Power. I'm sure if it's as good as they claim it'll only leave a niche for Wintel, because they can't directly compete in numbers or presence with ARM CPUs. (Note: This is actually the case right now. The most sold CPUs by far, are ARM ones.)

Re:no windows?

[FourthAge]

Yes.. and the article you refer to is very interesting, but I'm still not convinced by the ISA argument. In Atom, a frontend is needed to translate x86 operations to uops. However, something similar is needed in ARM, because the ARM ISA doesn't really fit a high-speed pipeline very well either.

Not all ARM instructions map directly to a single uop. The LDM/STM family are the most obvious examples. But there are also instructions with exotic addressing modes, instructions that *might* be branches depending on whether they write to r15 or not, and several ways to do common operations such as "return". The reference manual documents all sorts of special cases for various operations, apparently just to retain compatibility with legacy code. It's highly reminiscent of x86.

So, yes, there is a difference, but it's more like a move from a 1975-vintage CPU design to a 1985 vintage. Not such a huge change. ARM is not a modern ISA. I think the differences in power consumption have more to do with the fabrication process than the ISA, and that's something that Intel can easily change.

Re:no windows?

[Narishma]

PSP doesn't belong in that list. It has two MIPS CPUs, not ARM.

Re:no windows?

[RzUpAnmsCwrds]

The question is whether any of this is relevant in netbooks. The reality is Atom is crippled right now by an outdated power-hungry chipset. Typical idle power consumption on an Atom-based netbook is around 8W, which breaks out (approximately) as follows:

- 0.5W CPU
- 3.5W Northbridge (i945)
- 0.5W Southbridge (ICH7)
- 0.5W DDR2
- 0.5W HDD
- 0.5W 802.11
- 2W Display

The problem with these ARM netbooks is that they only address the power consumption of the CPU and the chipset. Once you consider the total platform power consumption, it's more like 4W, which is half as much, not 5x less.

OK, half as much power consumption is still great. But Intel is launching a die-shrunk, lower power, integrated die-shrunk NB (it's an MCM) Atom platform this fall. Expect that to shave 2-3W off of total platform power consumption. So now ARM-based netbooks have a 1-2W advantage.

Architect weenies like to comment on how complex and awful the x86 decoder is. And you're entirely right - Intel is going to have a hell of a time trying to cram x86 into anything cellphone-sized. But in a netbook, where a 30Wh battery is perfectly reasonable given the size of the device (which is limited by the need to have a big keyboard and screen), and where other elements (storage, memory, display) are more power-hungry, the half-watt advantage you have from having a simpler ISA isn't going to matter as much.

Re:no windows?

[Anonymous Coward]

The ARM instruction set is much simpler than the x86 set, and it lends itself to cheaper and more efficient designs. ARM chips basically stop drawing power when they're idle.

BlackBerries and iPhones are not low-end phones.

Count how many CPUs are in your home. Don't forget the fridge! Now compare that to the number of PCs. I'm sure ARM doesn't mind being relegated to the bigger part of the pie rather than the slice that keeps shrinking.

Application compatibility matters, but only in the short term. In the long term, people want cheap. Just ask Microsoft! I really believe that if you give someone a $200 machine that runs beautifully, lasts all day, has the ability to run things like Firefox and OO.org, and they'll start to question what's so great about Windows. This may not happen in a large amount in the next couple years, but it *is* happening.

What would you have said if someone had explained the netbook market to you five years ago? People want something cheap they can browse on. And the browser itself gets more interesting as a platform every year.

Re:no windows?

[iYk6]

SPARC ... has ... the highest market share once you leave the atmosphere

I see what you did there.

Re:no windows?

[Alioth]

I'll add to what the other poster said - just the bit in an Intel x86 chip that figures out how long the instructions needs as much die space as an *entire* ARM core. So yes, there is something about the ISA that makes ARM inherently low power.

Re:no windows?

[Anonymous Coward]

I don't buy your argument. Code density is actually a feature of x86. Because it contains complex operations that need multiple operations on RISC architectures, and because of the variable length of the codes.

I also don't think the decoding of the instructions is such a big issue as you display it. The Atom really lacks a lot of the clever mechanisms a Core 2 has to reduce size and power consumption. But the vast majority of that is reduced number of execution units, instruction reordering and especially speculative execution.

Additionally I read somewhere that half of the die is just cache. So if I think of all the stuff that has to fit in the remaining space (the execution units, hyper threading, the more modern SSE instruction sets, etc) I just can't imagine the x86 instruction decoding and micro architecture are a that big part of it.

Re:no windows?

[MtViewGuy]

The biggest problem with the ARM CPU--if you want it for a netbook--is that you MUST write the OS from close to scratch to take advantage of the CPU, a similar "chicken and egg" problem that plagued the PowerPC CPU.

What made the Intel Atom CPU so popular was that operating systems already designed for x86 CPUs will run on the Atom CPU with very little to no modifications. This means Windows, MacOS X versions designed for Intel CPU's, and many Linux distributions will run on the Atom CPU "out of the box."

Re:no windows?

[BhaKi]

In Atom, a frontend is needed to translate x86 operations to uops. However, something similar is needed in ARM...

similar, but simpler. That's the point.

Re:no windows?

[Anonymous Coward]

Both PowerPC architecture and Sparc are still around. The only thing they managed to kill was HP's PA-RISC. I don't want to know how much money Intel put into that, but it defenitely was a loss performance-wise.

It'll be very interesting to see ARM entering the netbook market. If it manages to get it's share there it might take the route on to the business notebooks and desktops. Interesting times ahead!

Re:no windows?

[EXrider]

You can also add the majority of Cisco network hardware to the list of things utilizing PowerPC CPUs.

Re:no windows?

[EXrider]

The only thing they managed to kill was HP's PA-RISC

Don't forget Alpha, a lot of the same engineers went to work on Itanium at Intel, but we all know how well that's turned out so far.

Re:no windows?

[petermgreen]

And PowerPC on desktop was killed by Apptel, not Wintel.
Afaict powerpc on the desktop/laptop was finally killed by the fact that the CPU manufacturers who made it either couldn't or wouldn't deliver the performance per watt to make their laptops and all-in-ones competitive and the top-end performance to make their powermacs competitive (they tried to get around this to some degree by throwing CPUs at the problem but that strategy has it's limits).

And a large part of the reason for that was that intel had the massive resources from the dominance of the wintel PC to fund thier development.

Re:no windows?

[Cochonou]

European satellite designs are working with SPARC or ERC32 processors, but I believe that most of the american satellites are working with PowerPC processors.

It's time for apple to step in

[bugs2squash]

They have supported non-Intel before.

Re:It's time for apple to step in

[BlackSnake112]

Actually, microsoft supported non intel before. Anyone remember the DEC Alpha chips? There was an NT flavor for that. It ran faster then the intel chips of the day.

It would not surprise me that in an microsoft lab there was windows for power PC, windows for ARM, windows for . It would be in microsoft's best interests to have them.

Re:It's time for apple to step in

[TheRaven64]

Apple already ships a lot of ARM devices (iPhones, iPods). They bought PA Semi to begin designing their own chips, so it's unclear whether they will be shipping ARM-designed cores with their own extensions or ARM-compatible cores in future devices.

Re:It's time for apple to step in

[Anonymous Coward]

iPhones are ARM-based. They run Mac OS X (with a different UI).

Re:It's time for apple to step in

[Anonymous Coward]

They have supported non-Intel before.

Yes, they could do it easily since they already have an ARM based OSX on their iPhone and iPod Touch. But would they want to be limited to app store apps on netbooks? I don't know.

Re:It's time for apple to step in

[Anonymous Coward]

Windows NT ran on x86, PowerPC, MIPS, and Alpha chips through version 3.51. NT 4.0 only ran on x86 and Alpha. Alpha support was dropped sometime during NT 4.0's lifetime and now Microsoft only writes NT (i.e. XP, Vista, Windows 7) for the x86 platform.

Re:It's time for apple to step in

[rilister]

Hmmm, so what do you think that new Apple tablet will be based on? ;)
The A9 has been available for a little while already (according to other comments here), and I see a sentence on their presentation "Hard macro already has it's first licensee".

Apple also uses ARM chips everywhere (iPhone) and so is really familiar with getting OSX to work on it...

(health warning: totally uninformed random guess.)

Re:It's time for apple to step in

[rickb928]

Hopefully it will run Flash sometime soon... Though I hate Flash and everything it stands for...

Re:It's time for apple to step in

[Jason Earl]

Yes, there have been non-Intel versions of Windows before, and they sold exceptionally poorly. The old Alpha version of NT at least had the advantage of having a high profit margin. If you were running Windows NT on a Dec Alpha chances were very good that you were also going to be running SQL Server and paying enough in license fees that it was worth Microsoft's time.

Or apparently not, as the Alpha port has gone away, and even the Intel-friendly Itanium support is a generation behind.

Arm9 is basically an entirely different beast. Instead of a high end offering it is an extremely low end offering. Arm9 will be debuting in the netbook arena--a market where pressure from Linux has already driven Microsoft to continue support for Windows XP, and at a drastically reduced price to boot.

Microsoft is going to stay as far away from Arm9 as possible. The last thing that Microsoft wants to do is to actively encourage people to move towards netbooks. Even if Microsoft did want people to move towards netbooks it wouldn't want to spend time and energy on a non-x86 architecture unless it was absolutely clear that Intel was getting creamed. After all, getting Windows to run on Arm9 is only part of the battle. The tricky bit would be getting the quadzillions of other Windows applications to run on Arm9.

Windows without its assortment of Windows applications is not likely to compete well against Linux.

Microsoft is simply going to ignore Arm and hope it goes away, which, without Windows support, is likely to happen. For now Intel's Atom processor is likely to be good enough, and Microsoft would just as soon the entire netbook category went away anyhow.

Re:It's time for apple to step in

[Anonymous Coward]

Actually a version of Windows for powerpc ships in the xbox360. I would not put it past Microsoft to look at a way to hurt Intel, they were shafted by Intel on the processor for the first Xbox and that'w why they chose IBM for the 360 (actually IBM did the design but
Microsoft can have another foundry build the chip, Intel would never allow that). On the other hand, I suspect Sony to be disappointed
with the Cell and the PS4 might well have Intel designed chips (depending on other factors, like the aforementioned possibility of
having the chip built not only in Intel's foundries).

So m bet for the next generation of consoles is that Microsoft and Nintendo will stay with IBM (they are satisfied and have good reason to stay, like easier compatibility with existing games), while Sony is an unknown and migh well jump ship.

Re:It's time for apple to step in

[XDirtypunkX]

I heard from a source who worked for DEC at the time that Windows NT was ported to Alpha as part of an under the table settlement due to patent infringements (mainly on the virtual memory subsystem) and was never expected to make a profit.

Will ARM compete?

[TheBilgeRat]

Does ARM plan on integrated video along the lines of Nvidia and ION? http://www.nvidia.com/object/sff_ion.html

Re:Will ARM compete?

[BadAnalogyGuy]

No, but NVidia has gone ahead and integrated ARM.

Re:Will ARM compete?

[TheBilgeRat]

ooh, that's sexy....

Re:Will ARM compete?

[TheRaven64]

You realise that nVidia has licensed the ARM Cortex A9 core that TFA is about and is shipping chips based on it, right?

Re:Will ARM compete?

[Jesus_666]

ARM doesn't do that as ARM doesn't make chips. They make chip designs, which they then license to other companies. Look at the TI OMAP series and you'll find that their chip comes with a built-in PowerVR GPU theoretically rated for DX 10.1 as well as a built-in DSP. Other manufacturers will most likely have similar offerings.

Just to put things into perspective, the Pandora ships with an OMAP3530 and will have as one launch "title" a PlayStation emulator, which has already been demonstrated to run smoothly. We're talking about something the size of a Nintendo DS that has a projected battery life of 9+ hours playing games off a 4200 mAh battery, capable of emulating the PSX. The Cortex A8-based OMAP3 is a seriously powerful little beast. I don't expect the Cortex A9-based OMAP4 to be any worse.

Re:Will ARM compete?

[TheBilgeRat]

I do now :).

Although now I am curious-if Nvidia is integrating with the Arm, will this spur better driver integration into the Linux kernel?

Re:Will ARM compete?

[Adam Hazzlebank]

So it's pretty cool that ARM now have a design capable of going to 2GHz but has anyone signed up to make the silicon? The A9 design has been around for a while, but I've not seen any implementations yet.

Re:Will ARM compete?

[Jesus_666]

As said, TI has announced the OMAP4 series, which is based around the A9. Only 720 MHz and "1+ GHz" so far, though. (Source: Wikipedia)

What does Linux on ARM support?

[Kjella]

I've only been running Linux on x86 hardware... so would Linux on ARM:

a) Run a typical distro only recompiled or is a lot of software x86-specific?
b) Run wine?
c) Run virtualbox w/windows?
d) Be able to use w32codecs so everything plays?

I'm sure there's a few that's removed all traces of Windows, but I'm not one of them...

Re:What does Linux on ARM support?

[Anonymous Coward]

a) Run a typical distro only recompiled or is a lot of software x86-specific?

Yes

b) Run wine?

No, Wine is designed to run x86 windows applications. It doesn't emulate a different CPU.

c) Run virtualbox w/windows?

No, Virtualbox is a purely x86/amd64 solution. It uses hardware virtualization in your CPU, it doesn't emulate a different CPU. You're looking for QEMU.

d) Be able to use w32codecs so everything plays?

Maybe with QEMU, it has some application level CPU emulation, interested parties might make it work in that case.

Re:What does Linux on ARM support?

[Microlith]

Let me rephrase your questions, and condense them:

would Linux on ARM run x86 software

And the answer is no, not without emulation.

Re:What does Linux on ARM support?

[GungaDan]

My Archos V media tablet has an ARM processor and runs linux. It plays .avis, .wmvs and lots of other video (some codecs sold separately, dammit), hooks up to wifi for full-featured web browsing, plays flash video/apps, displays pictures, plays back every format of music I've thrown at it, and has the most flawless touch-screen interface I've yet used. The only things I want it to do that it can't are VPN and SSH.

Re:What does Linux on ARM support?

[queazocotal]

A) much simply needs recompiled, if it doesn't - with an app with the source - it's usually a bug.
B) No - wine is simply a conversion layer between the windows and linux calls - the windows program is never emulated.
C) No - again - not without emulation.
D) I think you can probably guess this one - but again no.

Emulation may be _lots_ slower than the host processor - slowdowns of ten times or more are not uncommon.

Re:What does Linux on ARM support?

[ianare]

a) Almost all of the open source stuff will run on ARM (GTK/Qt stuff, interpreted language apps, Firefox, etc).
b) No, Wine Is Not (an) Emulator.
c) No, because again you would need an emulator.
d) Not the MS ones. OSS equivalents might.

Re:What does Linux on ARM support?

[fuzzyfuzzyfungus]

a) Yeah, pretty much. Most desktop oriented distros only bother with x86/x86-64(though I think Ubuntu either has or is coming out with an ARM version); but with something like Debian running on ARM is pretty similar indeed. Obviously, less popular architectures get less attention; but the fundamentals are already well in place, and the introduction of ARM based netbooks would presumably increase attention pretty quickly.

b) In principle it could; but there wouldn't be much point. Wine on ARM would allow you to run win32 ARM applications. There aren't any.

c) Possibly. Common virtualization packages try to do as much as possible directly on the host CPU, which makes things much more efficient; but also means that x86 guests require x86 hosts. If you use full virtualization setups, like QEMU, you can run an x86 guest on another host, at the cost of performance. Virtualbox has recompilation support based on QEMU; but I don't know if that is for non x86/x86 setups, or just for dealing with troublesome x86 instructions.

d) Probably not.

Re:What does Linux on ARM support?

[AlXtreme]

a) Run a typical distro only recompiled or is a lot of software x86-specific?

It depends on the distro. Debian has a complete ARM-port, Ubuntu was working on one last time I checked. Maemo is an ARM-only distro.

b) Run wine?

Nope.

c) Run virtualbox w/windows?

Nope.

d) Be able to use w32codecs so everything plays?

Not likely (assuming these are binary blobs). Flash video, avi/mpeg's and various other formats shouldn't be a problem though.

An ARM netbook wouldn't be someones only PC, just like current netbooks aren't. If it can do 90% of the things you're used to you're set.

Re:What does Linux on ARM support?

[mpapet]

Most things actually.

a) Run a typical distro only recompiled or is a lot of software x86-specific?
Debian has an ARM version of their distro. From personal experience, I found everything to run an nslu2 server without exception. http://www.nslu2-linux.org/ Very, very efficient platform. The nslu2 had no crypto coprocessor, so ssl stuff was slow, but still, the nslu2 was one of the most useful devices ever.
b) Run wine?
No. Wine isn't an emulator, so all of those x86 Windows compiled apps won't work.
c) Run virtualbox w/windows?
Depends. If virtualbox is an emulator, then you will be good to go. If it isn't, then you run into the wine problem.
d) Be able to use w32codecs so everything plays?
Probably not. Again, the whole x86 thing gets in the way.

What's more important to keep in mind are the binary blobs buried in some Linux drivers/applications. Epson's Linux driver is the perfect example. Not only is it a bear to build because it needs an ancient tool chain, but there are x86 blobs in the sources as provided by Epson. So, Epson printers are a total waste of money on arm.

HP's drivers build perfectly on ARM though. So, not only do you get an easily networked printer, but an easily networked scanner too.

Re:What does Linux on ARM support?

[TheRaven64]

a) Run a typical distro only recompiled or is a lot of software x86-specific?

Some is, but most *NIX software is portable.

b) Run wine?

Maybe. There is a project that integrates WINE with QEMU, but I don't know its status. This would (in theory) run the app in QEMU, but every call to WINE stubs would be proxied to the native WINE libraries so code inside WINE would run the native versions (including things like DirectX). This would be fast enough for all but the most CPU-intensive Windows applications.

c) Run virtualbox w/windows?

No, VirtualBox is x86 virtualization software so obviously won't work on a non-x86 chip. You can run QEMU in full-system emulation mode but it will be a lot slower.

d) Be able to use w32codecs so everything plays?

No. Do people really still do this?

Re:What does Linux on ARM support?

[Anonymous Coward]

a) Yes

b) No

c) According to the QEMU website there's some support for running x86 software on an ARM host.

d) If QEMU can manage c), then yes.

Re:What does Linux on ARM support?

[Compholio]

b) Run wine?

No, Wine is designed to run x86 windows applications. It doesn't emulate a different CPU.

There was a build of Wine at one time that used a special version of QEMU to translate instructions. Seems to me like it might be time for someone to try again, as the ability to run Win32 applications on ARM would be a huge boon to the platform. I haven't kept up with things in a while, so it's possible that there's already a stable way to use Wine with ARM.

Re:What does Linux on ARM support?

[redhog]

a) Most likely. Some programs, like the Linux kernel, needs some hacks (but they already exists for ARM :)

b) Nope, but see c

b) Nope, but qemu would work. However, as the x86 instruction set would be emulated on ARM, there would be a performance hit

d) Nope, but see c

For b and d it is at least theoretically possible to write some kind of wrapper code that run the x86 binaries on qemu but had them interface with the surrounding ARM-based Linux directly. I don't know if anyone has done this, but it would be sort of similar to misc binary support + java vm...

Re:What does Linux on ARM support?

[jimicus]

There is a project that integrates WINE with QEMU, but I don't know its status. This would (in theory) run the app in QEMU, but every call to WINE stubs would be proxied to the native WINE libraries so code inside WINE would run the native versions (including things like DirectX)....This would be fast enough for all but the most CPU-intensive Windows applications.

That I very strongly doubt.

The only reason modern PCs do such a good job of emulating older computers is that they are literally 100 times faster. The 5-10x speed loss you typically see through emulation suddenly becomes a non-issue. But in this case you wouldn't be trying to emulate something that used to run 100 times slower than your native CPU, you're trying to emulate something that runs at the same speed.

Re:What does Linux on ARM support?

[Anonymous Coward]

What's more important to keep in mind are the binary blobs buried in some Linux drivers/applications. Epson's Linux driver is the perfect example. Not only is it a bear to build because it needs an ancient tool chain, but there are x86 blobs in the sources as provided by Epson. So, Epson printers are a total waste of money on arm.

What is Epson’s driver needed for? I have many Epson printers and scanners, old and new ones, and they all work with cups/gutenprint and sane with no binary blobs required.

Re:What does Linux on ARM support?

[Pecisk]

With serious improvements within ffmpeg stack, w32codecs as mandatory package is already gone for some time. Most of newest netbook oriented distros (Moblin, Maemo, Ubuntu Netbook Remix) uses Gstreamer as multimedia engine, which has serious developers working for speeding up things for ARM platform. Also I bet ffmpeg guys already have been working on this.

Re:What does Linux on ARM support?

[Haeleth]

Apple managed to make PowerPC emulation work on x86 chips. It wasn't fast, but it ran acceptably. It only required a processor with about 2x the performance -- the same ratio as these ARMs are to Atoms.

Bear in mind that most applications actually spend most of their time sitting around waiting for user input, and a lot of the rest of their code is simply API calls for which native implementations can be provided. We're talking about things like Quicken here, not the latest games.

It could well be feasible to provide a solution that worked adequately for a significant subset of programs -- particularly for the kinds of domain-specific application where it's particularly unlikely that an alternative program will exist.

Re:What does Linux on ARM support?

[snadrus]

1. Everything Ubuntu can offer is available for ARM: http://www.ubuntu.com/products/whatisubuntu/arm

4. W32Codecs is obsolete since FFMPEG does WMV & Quicktime. Real player is in Helix. All these are distributed by source and should work on ARM.

Does it really need to support windows?

[longfalcon]

netbooks are a great place to quietly slip in non-windows OS's that meet customer needs. the mobile phone/smart phone market has shown that customers aren't slavishly devoted to Windows. they will buy what works.

Re:Does it really need to support windows?

[Ant P.]

the mobile phone/smart phone market has shown that customers aren't slavishly devoted to Windows. they will buy what works.

But it's also proven they'll buy what doesn't. The practice of selling crippled and sabotaged phones in the US hasn't slowed down one bit in spite of the iPhone. Put a shitty ARM port of windows in front of these people and they'll mindlessly slurp it up.

Re:Does it really need to support windows?

[Cro Magnon]

Depends on what people do with their netbooks. If they just surf the web and check MyFace, Firefox will work just as well on ARM Linux as it does on their Win desktop. If they plan to use it as a pint-sized laptop, they might be SOL.

Re:Does it really need to support windows?

[longfalcon]

different use cases will dictate what the consumer will look for. we don't even know what kind of consumer the supposed ARM netbook would target.

if it is targeting Macbook Air customers, they need to run windows well, and get app devs on board as well. but if they target the ultra-low-end, those consumers will buy whatever OS, no matter how crippled, if the price is right - and not having a windows licence fee in the price could make the difference.

The Windows era

[PriceIke]

From the looks of the past five-seven years, the tea leaves seem to be saying that Microsoft's star is on the decline.

chip supports OS? Hmmm, backwards...

[kharris312002]

This may be the first time I've ever heard it said that a processor doesn't support an OS... Usually it's the other way around.

5x slow

[Anonymous Coward]

"the processor offers five times the power while drawing comparable amounts of energy."

So it is five times as slow?

YES! Please bring one to market

[Anonymous Coward]

I have been waiting to buy a netbook, specifically for an ARM-powered one because I want good battery life. If I can get something that's small, cheap and runs 10+ hours on a charge I'll gladly do without Windows.

Why do they even bother?

[Anonymous Coward]

According to this thread ARM has no chance on the high end.

A call to ARMs!

[MarkvW]

A Microsoft refusal to support a really cool netbook technology would be a good opening for Linux.

Re:A call to ARMs!

[lordandmaker]

It'd be good for Linux users, as all the devices are sold cheap.

Windows users will still pay the premium (in cost or power) to stick with what they know, unless this netbook is incredibly good. Apple is practically the definition of a polished product, they've fantastic brand awareness, they're commonly held to be superior, and they're still not really knocking Microsoft off their perch. I don't see why this is likely to make linux netbooks any more successful than the x86 ones have been.

Re:A call to ARMs!

[Anonymous Coward]

Microsoft won't refuse to support these devices, they aren't that stupid. You'll at the very least get WinCE versions of these ARM netbooks coming out. Whether they port Windows 7 over is a different matter, but even if they did they can't port the third-party apps which are so important to the Windows ecosystem.

But, yeah, it is a good opening for Linux, because the third-party support Windows gets on the x86 platform is its real advantage over Linux, and it won't have that here.

I am rather curious as to exactly how Microsoft will handle this situation and what exactly they said to ARM when they talked to them about it. I wouldn't be surprised if Microsoft uses its influence over Manufacturers to dissuade them from building these though, so the success of this market really depends on the Manufacturers that have the balls to stand up to Microsoft, and smaller ones that don't sell Windows PCs whom Microsoft is unable to pressure.

ARM is not a chipmaker

[sigxcpu]

A small correction.
ARM is not a chipmaker. They only liscense the design to others, don't do their own chips.
Can't buy an ARM from ARM only the IP to incorporate it into your own chip.

Interesting

[zephead99]

I think it's interesting that ARM has decided to try and break into Intel's current marketshare around the same time Intel is revealing plan to push it's 32nm chips into ARM's.

real solution

[99BottlesOfBeerInMyF]

There are a lot of barriers to Windows adoption on the ARM processor that go beyond MS not really wanting it. If they really want to gain market share above and beyond cell phones and PDA's, ARM needs a strong partner to create a real, integrated, polished solution. And by solution I don't mean a device. They need to do something akin to the iPhone, in creating a nice device or set of devices with a consistent polished operating system and with an integrated ecosystem of solutions. The project is large in scope and they need a partner that preferably has an existing position to leverage, experience, money, and which is not beholden to Microsoft. A cell phone service company might be a viable partner or Canonical and someone, or RIM or Google or an appliance maker that has not entered the netbook market yet.

If they really want to sell netbooks with ARM processors in them they have to think big. They need to better than hope MS is scared. They need to commit to building a system that bypasses MS's core monopolies through vertical integration. This is no small task. They need the hardware, which has to be cheap and hit a sweet spot. They need an OS and applications. They need dev tools for applications and services. They need Web and network services integrated with the device. More than all those pieces which are out there, they need someone to put it all together in a nice package and usability test the whole user experience from buying to opening the box right up through using it for all the common tasks: Web surfing, E-mail, chat, word processing, potentially phone calls and videophone, playing games, playing music and video, and adding new applications. The problem with a lot attempts at this sort of thing is the assumption that someone else will take care of parts or that blaming someone else somehow makes a failure better.

Re:real solution

[jabjoe]

The iPhone is a ARM processor.....

Re:real solution

[99BottlesOfBeerInMyF]

The iPhone is a ARM processor.....

Yes, but what does that have to do with what we're talking about... getting ARM processors implemented in netbooks? Unless someone takes it and provides an entire solution for netbooks like Apple did for smartphones, it is unlikely to overcome the barriers introduced by Windows domination of the desktop OS market and subsequent influence on current netbook manufacturers.

Re:real solution

[TJamieson]

The iPhone is a ARM processor.....

Exactly. What if Apple chooses this opportunity to make their ARM variant of OSX available to the general public? I know, never gonna happen, but I'd bet they're working on an A9 chip for either their tablet or a MacBook Mini (hint: rumors are that MacBook and iMac lines will see refreshes in next month or so).

Re:real solution

[Anonymous Coward]

Then Apple has a potential foot in the door for netbook market. Really, iPhoneOS is a version of OSX. That means Apple really doesn't have that much to do to create a version for ARM-based netbooks. All it need to is to distribute OSX like Windows. What could go wrong?

Re:real solution

[Idayen]

Cortex A9 MPCore processor + Iphone OS = Good ITablet?

Re:real solution

[SpinyNorman]

ARM isn't exactly struggling to find a niche you know... they sell BILLIONS of cores per year... Most smart phones (iPhone included), Windows CE devices and digital cameras are based on ARM cores. Linux based netbooks is a very obvious market for them to go after with their low power and low cost (low transistor count - it's a RISC not CISC CPU) advantages.

Re:real solution

[Anonymous Coward]

The iPhone is a ARM processor.....

Actually this might surprise you, but the iPhone is a phone.

Re:real solution

[99BottlesOfBeerInMyF]

ARM isn't exactly struggling to find a niche you know...

Who said they were? This article and discussion was about ARM attempting to get a foothold on netbooks. It doesn't have much to do with their other sales in other markets except in the very limited ways they can leverage those other markets. MS's monopoly position provides a significant barrier to entry for anything non-Windows, and supporting ARM chips in Windows doesn't seem to make long term business sense for MS unless there is a credible threat from alternative OS's on the platform.

Re:real solution

[tbuskey]

My niece in high school rarely uses a computer. She does most of what she needs on her iPhone.

That's what scares MS....

Re:real solution

[petermgreen]

Two things

1: it's a smartphone not a netbook. Windows netbooks (particularly the HDD based 10 inch ones) are building the expectation that you can run most PC apps on them just slower and with a bit more cramped screen. Expectations for a smartphone are lower.
2: Apple has a pretty large base of developers from the mac who they have managed to convert to iphone developers.
3: I strongly suspect apple poured a LOT of resources into the iphone.

Re:real solution

[Anonymous Coward]

No, the iPhone *uses* an ARM processor.

Is this what will undo the x86 ISA

[Anonymous Coward]

The DEC Alpha could not do it, the Itanium could not do it. (Or has not yet for economic + technical reasons)

Surly there must be a market that simplyt does not have to worry about legacy x86 code, or if they do, the performance hit of emulation is not a problem to them.

* Not undo entirely of course. COBOL is still around and so will the x86 be for sometime.

ooh

[nomadic]

As a product of British manufacture, is it safe to assume it will spend most of its lifetime at the computer repair shop?

Re:ooh

[oldhack]

It's a really odd feeling to loot for a British tech company. A British tech company. That's so... unnatural.

Re:ooh

[H0p313ss]

It's a really odd feeling to loot for a British tech company. A British tech company. That's so... unnatural.

Indeed, my grandfather wrote an article about the demise of British technology about 50 years ago when his employer closed up shop and he had to take a job with Monsanto.

Re:ooh

[H0p313ss]

As a product of British manufacture, is it safe to assume it will spend most of its lifetime at the computer repair shop?

Or on strike.

Re:ooh

[gbjbaanb]

yeah, we usually invent all the cool technology that is used around the world, and then give it away :(

I guess Britain was the original open source provider :).

Re:ooh

[TheRaven64]

In the smartphone market, most of the market share is owned by British-designed chips on a British-designed operating system. I think a lot of people forget that Symbian was originally developed by Psion (who also created the netBook term), a British company that made very nice handheld computers in the '90s. Both ARM and Symbian had similar starts and both have outlasted the companies that originally spun them out.

Re:ooh

[Hurricane78]

You mean unlike the other billions (literally) of ARM chips in phones, TVs, stereos, etc, etc, etc, that are so "bad" that they're the most sold CPUs on the planet? ^^

Seems like for once, a car analogy is not the best one to make.

Re:ooh

[CodeBuster]

Yes, but that's part of the charm, right mate?

Re:ooh

[oldhack]

Btw, where do all the Cambridge and Imperial engieering graduates end up?

Re:ooh

[Anonymous Coward]

It's alright ARM don't actually make the chips, just license the IP, so it doesn't count as British manufacture.

I will buy one

[cyberthanasis12]

Imagine a computer that does not run Windows. One that is not able to run Windows!
I want one. Now. (I assume that it runs a full Linux distro of course).

Re:I will buy one

[lordandmaker]

Yeah, those used to be commonly available.

Re:I will buy one

[flyingfsck]

Get a Touchbook: http://www.crunchgear.com/2009/03/02/touch-book-tablet-netbook-with-arm-cpu-10-hour-battery-detachable-screen/ It is already available and quite cheap.

Re:I will buy one

[werfu]

Sparc, Amiga, PPC Mac... just to name a few. Most of them used to be available to general public. Now they're either gone or not available to commoner.

Re:I will buy one

[robbrit]

Grab it straight from their site: http://www.alwaysinnovating.com/touchbook/

No Windows? Great! No Microsoft tax!

[MoxFulder]

However, it has one major hurdle to overcome: it doesn't support Windows.

Fuck Windows. Seriously.

I've been unwillingly paying the Microsoft tax for TEN YEARS. All I ever do is wipe Windows and install Linux. If my new computer can't run Windows then... great!! Maybe I won't have to pay the tax.

I'd love a low-power, high-performance ARM notebook. I'd be happy with MIPS or Loongson (Chinese MIPS clone) as well. Debian already has a full-blown ARM port and I'll bet they could get it working on an ARM netbook in a day. Ubuntu would undoubtedly be soon-to-follow.

As a side benefit, having multiple widely-used architectures for desktop systems (x86 and ARM) would be a support nightmare for hardware companies that still keep their drivers proprietary and undocumented. Yeah, I'm looking at you, Broadcom and NVidia. This would just be another nail in the coffin for their obstructionist attitudes towards free/open-source operating systems.

Free transistors anyone?

[Anonymous Coward]

Where are all those guys who say:
"ISA doesn't matter"
"overhead of x86 is negligible"
"all x86 CPUs are RISC inside anyway"
etc.

its the screen and wifi, stupid

[Gothmolly]

I don't care how much the CPU uses. Get me a screen and WiFi adapter that use 1W of power and I'll be interested.

Just for kicks

[Yvan256]

If anyone ever starts a new CPU-related company, can you please call it LEG for the sake of "it cost an ARM and a LEG" jokes?

Thank you.

Re:Just for kicks

[Anonymous Coward]

Not to mention the dual cpu motherboard jokes of spreading you LEGs apart.

Re:Just for kicks

[TheRaven64]

LEG should make DSPs or GPUs, so SoC manufacturers can include an ARM and a LEG on the same chip.

Re:Just for kicks

[Anonymous Coward]

Re:Just for kicks

If anyone ever starts a new CPU-related company, can you please call it LEG

I see what you did there.

Re:Just for kicks

[Zoxed]

> so SoC manufacturers can include an ARM and a LEG on the same chip.

And then install Emacs and claim that your product *includes* the kitchen sink :-)

Re:No Windows? Great! No Microsoft tax!

[Anonymous Coward]

There is No Microsoft Tax Except on Lazy Twats. I enjoy Microsoft operating systems, but I've never been taxed. My first two PCs were built from scratch, then I purchased a Windows OS separately. My last 2 PC purchases came with Windows already installed as this was my choice, but I had other options.

If you were buying your PCs with a Windows OS you never wanted and never needed, you are only a victim of your own ignorance or laziness. You are the one at fault, not Microsoft or anyone else you'd rather blame.

Two things jump to mind. . .

[Fantastic Lad]

The first is. . .

Wow! --After looking at the difficulties lined up in the marketing of this new architecture, it seems to me that if it isn't backwards compatible with Windows software, then it's going to have to offer some kind of serious advantage in order to win in the game of software evolution. There's a very good chance this won't happen because the dominant approach to OS's remains full-steam ahead.

Example from biology: Humans have dozens of idiotic design parameters, features and various do-dads which despite their idiocy remain embedded in the primary design simply because the over-all template was good enough to win at the evolution game. Thus we have appendixes, wisdom teeth, crappy eyes, redundant body hair, and inefficient muscle levers, (to name a few).

We might be stuck with silly chip designs on the Enterprise.

Second thought. . .

Maybe not! --Remember the Y2K bug? How the estimates of how long it would take to re-code all the software so that it could manage an extra two digits? Disaster-scenario lovers predicted that it would take twenty years of sustained, super-expensive effort to fix the thing. As it happened, we got it done with time to spare. So maybe porting all our beloved software so that it runs on this more efficient architecture would be similarly within our grasp.

While I and every other geek in the world would love portable computers to run many times longer on a single charge, I'm not sure that this force is quite the same as, "Planes will fall out of the sky and all the escalators will fail on the eve of 2000!" We'll have to wait and see.

But good luck to ARM; I always like to see clever innovation, and my wisdom teeth were horrible.

-FL

Re:Two things jump to mind. . .

[A12m0v]

my wisdom teeth were good to me you insensitive clod!

Re:Two things jump to mind. . .

[Fantastic Lad]

my wisdom teeth were good to me you insensitive clod!

Oh, that just means you're further along the evolutionary path than I am, and your large-jawed seed will prosper while my kind vanishes down some dead end branch of Darwin's tree.

Big-mouths inherit the world. Ain't that always the way?

-FL

Re:chip supports OS? Hmmm, backwards...

[Geekner]

This isn't your standard x86-compatible chip. There are many non-x86 chips out there, but they are usually segregated from "PC Computing". They are used in phones, PDAs, embedded devices, consoles (Nintendo DS and most other handhelds), servers, and much more. PowerPC was the last common "Personal Computer" processor that wasn't x86 compatible.

This is simply a case where a specialized processor designed for highly integrated and mobile uses is trying to break into the mainstream Personal Computing market. The primary limitation is that the instruction set on this processor does not support the Intel x86 standard, which is used by most mainstream Operating Systems (Windows, OSX). However, it is possble to port these OSs to ARM, as most open source operating systems (Linux, BSD) have been ported successfully.

The question is whether or not Microsoft or Apple will port and take advantage of these new devices, or will Linux gain a lead in this emerging market?

Mod Parent Up!

[Cassini2]

You nailed the comment in 6 words.

ARM Holdings is a closed company

[Anonymous Coward]

the arm architecture is own by them and they have a history of using patent threats and strategic headhunting to kill competition are related open source projects.

other applications

[Theodore]

I'd love to see an ARM netbook.
With a 9-cell battery,,, granted, the screen and mechanicals are the biggest drain, but still, that'd waste the Atom.
(especially since MS insists producers use inferior versions of the Atom. Read my history, know that I think using anything less that the 330 is a horrible sin).

I'd also like to see....
These things use HOW much power? (or better yet, how little?)
How many can you fit on a PCI card?
How much power would it draw?
Think it could give Larrabee a run for it's money?
I'm not concerned about video, just processing.
Give us a good API for it, similar to CUDA/OpenCL, then sell it as a co-processor
(if it's doable, I do confess a decent amount of ignorance of the basic ability of this chip, but no shortage of ideas).

win

[hyperion2010]

RISC: fuck yeah! Time to invade the Western United States. ROLLIN

Intel am doomed

[A12m0v]

the death of x86 is imminent!

I wonder...

[enriquein]

I wonder if it will go as "quick and smooth" as the move towards 64-bit x86. I mean, seriously, we're talking a whole new instruction set, not just an extension to an existing one. Can't imagine how long it'll take for ATi and Nvidia to provide drivers.

If it's giving 5X the processing power of the Atom

[tbetz]

... as 2GHz clock speed, how much could overall power consumption be reduced by underclocking it by 50%?

If ARM isn't selling wolf tickets, that would still yield 250% of the Atom's processing power in that circumstance.

Go for Linux and let Microsoft+Intel rot.

[miffo.swe]

The only way the ARM manufacturers stand any chance is to either go for a completely new OS or jump onto the Linux bandwagon. Microsoft wont endanger their cooperation with Intel and AMD until ARM has secured a sufficient enough marketshare. This makes ARM and Microsoft a catch 22 happening. Any support will be superficial with lots and lots of fot dragging.

On the other hand Asus has shown just how successfull cheap small devices can be with Linux on them. If the ARM companies goes ahead full steam pushing devices with Linux Microsoft will be forced to jump aboard no matter what they really want to do. By then Microsoft wont be calling the shots and ARM will have a much better bargaining position.

I also think the ARM manufacturers should take a long hard look at the Wintel OEMs and think about their situation. Do they really want ot find themselves in a position where all their revenue is taken by a third party like Microsoft who doesnt contribute anything at all to the platform? Are they comfortable to be totally in the hands of a company that cant manage to turn out a new version of their OS in almost ten years?

Re:No Windows? Great! No Microsoft tax!

[Smivs]

However, it has one major hurdle to overcome: it doesn't support Windows.

Fuck Windows. Seriously.

I've been unwillingly paying the Microsoft tax for TEN YEARS. All I ever do is wipe Windows and install Linux. If my new computer can't run Windows then... great!! Maybe I won't have to pay the tax.

On a serious note, why not get your computer built for you (or DIY if you can). I had mine built by a small local company (Intel core2 quad, 4Gig RAM and 250Gig hard drive so a decent spec) and it cost well under £300. It came 'empty' - no OS - so I could install Ubuntu with NO Windoze contamination. It works geat. It's never given me any trouble at all and it does everything I want, quickly and very well.

Re:No Windows? Great! No Microsoft tax!

[Hadlock]

Dunno why the other comment got modded as a troll; most companies will happily refund you the cost of Windows; it's been documented many times. Dell sells a bunch of Linux laptops now, and you can find no-MS laptops from smaller vendors too. Perhaps you aren't trying very hard.

The best thing about it

[the_womble]

it doesn't support Windows.

That's not a bug, its a feature.

Re:The best thing about it

[Corson]

No problem, Ubuntu for ARM will be out soon (it's already available for specific platforms).

Re:The best thing about it

[jonaskoelker]

Or you could put Debian on it, which has supported ARM since... roughly forever?

Re:The best thing about it

[Erikderzweite]

In the meantime -- stick with Debian.

Re:real solution: Google

[Informative]

TFA mentions that Google Chrome OS should support ARM, so since we already see Google Phones with Android and Google Apps, I don't think it's overly optimistic to hope to see a "GoogleBook" or Google Tablet.

nVidia makes ARM chips

[Anonymous Coward]

Don't expect nVidia to open anything up: they're pushing ARM too.

Re:nVidia makes ARM chips

[MoxFulder]

Don't expect nVidia to open anything up: they're pushing ARM too.

True. But they're pushing ARM (in the form of their Tegra systems-on-a-chip) for totally closed and locked-down portable platforms, like cell phones and media players and the Zune HD.

Open and extensible non-x86 platforms, such as cheap ATX motherboards with ARM rather than x86 processors, would totally throw nVidia for a loop. They'd have to scramble to support their proprietary drivers on such systems, or risk losing market share very quickly.

The Linux kernel already has the code to make a lot of ATI video cards work fully on an ARM system, should anyone manage to acquire an ARM board with a functioning PCI-Express slot. The Linux kernel already has the code to make an Atheros wireless card work fully on a MIPS system... the same driver code can run your desktop PCI wifi card, or the one in your router. The Linux kernel has no open-source code to do 3D acceleration with nVidia cards. Why not? Because nVidia won't open up any specs...

Basically, all the companies that have done The Right Thing and opened up hardware specs have gotten free cross-platform code written by the Linux kernel developers. Those, like nVidia, that haven't, they're at risk of being left in the dust.

Re:nVidia makes ARM chips

[bhtooefr]

Technically, some of the Intel IOP processors (XScale-based, so they're left in the dust on performance) have PCIe controllers. They're meant for storage applications, but the IOP321 was used in a desktop. (You're stuck with the nVidia card in that thing, though, IIRC.)

Re:No Windows? Great! No Microsoft tax!

[MoxFulder]

On a serious note, why not get your computer built for you (or DIY if you can). I had mine built by a small local company (Intel core2 quad, 4Gig RAM and 250Gig hard drive so a decent spec) and it cost well under £300. It came 'empty' - no OS - so I could install Ubuntu with NO Windoze contamination. It works geat. It's never given me any trouble at all and it does everything I want, quickly and very well.

It's a reasonable point, and it's true that some companies will actually refund the Microsoft Tax if you demand it... although it inevitably turns out to take so much time and effort that it's not worth it.

The problem is that most of the best prices on pre-built computers come with Windows pre-installed. Sure, Dell may sell some systems running Ubuntu... but the selection is limited. For example, none of them use AMD processors. So, invariably, for me to get the best price on what I want, I have to accept Windows bundled with it. I am getting the best price, but it would be even lower if the seller weren't passing along the cost (perhaps $10-100, hard to say) of a Windows OEM license.

I imagine many manufacturers would willingly offer a choice of Operating System ("none: subtract $40" or "Windows: $0") if Microsoft didn't strong-arm them.

Also, I have built a number of DIY computers. However, often you can get a better deal buying a pre-built system from a big box store.

SheevaPlug

[DamonHD]

Hi,

I'm in the process of moving my main Internet-facing server onto an ARM-based SheevaPlug (1.2GHz, 512MB memory) that consumes 3W--5W (pegged at 5W right now doing a large Java build/obfuscation).

http://plugcomputer.org/

http://www.earth.org.uk/note-on-SheevaPlug-setup.html

Rgds

Damon

Re:SheevaPlug

[Bert64]

I looked at those devices a while ago, but was unable to get hold of one in the UK...
You can order from the US, but the shipping costs as much as the unit itself, and there doesn't seem to be any european distributor.

Re:SheevaPlug

[DamonHD]

The total including shipping and duty was ~£100 for me in the UK (with UK plug, FWIW).

Rgds

Damon

Re:SheevaPlug

[calidoscope]

The OpenRD system (SheevaPlug with more I/O) looks pretty tasty as well. Do wish there was an OpenBSD port for the OpenRD as it would make for a darn good firewall/router.

Re:SheevaPlug

[petermgreen]

Looking at the manufacturers reccomeneded supplier (globalspec) the sheevaplug is just over £60 and the shipping just over £40. So either something has changed since you bought it or you bought it somewhere else.

Re:SheevaPlug

[sucati]

I've been running OpenJDK on mine but I'm disappointed with the performance. It turns out that hotspot was written in assembly and hasn't been ported. They are working on something called Shark, built on LLVM, to overcome this issue.

Re:SheevaPlug

[DamonHD]

Hi,

Shark should be ready by December so Xerxes Ranby tells me.

For now I'm evaluating Sun's embedded SE JRE for ARMv5 and though it only has a C1 (client) compiler it works pretty well.

http://www.earth.org.uk/note-on-SheevaPlug-setup.html#Java

Rgds

Damon

Re:SheevaPlug

[Anonymous Coward]

Does the OpenBSD armish port not work on the OpenRD board?

Re:No Windows? Great! No Microsoft tax!

[mattcasters]

Perhaps MoxFulder is not living in the US like me.
"Dell" has different OS policies in different countries. In Belgium, I checked with Dell, IBM, Lenovo, Acer, Asus, etc.

Desktops: No Linux
Laptops: No Linux
Netbooks: no Linux

It's only for server platforms that you can find boxes without software, with Windows software, RedHat, SuSE etc.

That leaves white-box, unsupported, un-serviced boxes that leave me in the dust when they fail. It's better for me to buy an XP box from say Dell, dump Windows & install Linux and get a decent support contract with on-site support.

Re:No Windows? Great! No Microsoft tax!

[chicago_scott]

So I guess it all comes down to what the definition of the word "unwillingly" is. :-)

Re:No Windows? Great! No Microsoft tax!

[aminorex]

They should take a tip from the DEC Alpha, and compile x86 code to ARM, run windows natively without any support from uSoft. This would be vastly easier today due to the paravirtualization work of the past decade.

Re:chip supports OS? Hmmm, backwards...

[Locutus]

in the Windows world, you'll hear that the processor runs on Windows all over the place. They've been trained that Windows is the end all, be all, and center of the universe so the concept of "it runs on Windows" is their world. Talk about a CPU and _it_ runs on Windows is the norm. They really don't know how to think about it without Windows at the center or in a hierarchy of the hardware->OS->applications. They can't imagine a world without Windows. Combine that with software people and marketing people with no clue of hardware and you get "processor X doesn't run on Windows"

LoB

Re:No Windows? Great! No Microsoft tax!

[BikeHelmet]

In the last decade x86 code got a lot more complicated too, though.

Re:real solution: Google

[99BottlesOfBeerInMyF]

TFA mentions that Google Chrome OS should support ARM, so since we already see Google Phones with Android and Google Apps, I don't think it's overly optimistic to hope to see a "GoogleBook" or Google Tablet.

I actually have little hope for such an endeavor. Google is developing an OS and they have apps and services. They don't have hardware and they have been reluctant to move into such things. All the Android phones to date have been iffy in their implementation and polish. Someone will undoubtably slap Google ChromeOS on a NetBook, but I doubt it will be well integrated with services, well supported, properly polished for usability, or widely marketed. So far devices with Google OS's attached have been missing exactly the kind of integration and overall end user experience usability that is needed to overcome the drawbacks of working in an ecosystem that has grown around the Windows hegemony.

Re:If it's giving 5X the processing power of the A

[BikeHelmet]

... as 2GHz clock speed, how much could overall power consumption be reduced by underclocking it by 50%?

If ARM isn't selling wolf tickets, that would still yield 250% of the Atom's processing power in that circumstance.

Probably not much. Keep in mind that with aggressive power management, most of those cores will be sitting unpowered, so your consumption is likely under a watt.

If you downclock them, then maybe it only uses 500mw per core, but now you need more cores running. That's a big jump in power consumption. Also, everything feels slower.

There's a reason a C2D @ 4.0ghz feels faster than a C2Q @ 2.0ghz. ;)

doesn't support windows??

[Anonymous Coward]

Doesn't support Windows? Ha! You've got it backwards...

it is Windows that doesn't support any processor but X86.

Re:chip supports OS? Hmmm, backwards...

[Anonymous Coward]

In Soviet Russia...

doesn't support Windows

[amoeba1911]

However, it has one major hurdle to overcome: it doesn't support Windows.

Well, duh: CPUs aren't programmed for operating systems, operating systems are programmed for CPUs. You can make a x86 compatible CPU and have Windows run on it, but if you have a different CPU architecture you can scream until you're blue in the face, it will never support Windows. So, really: Windows doesn't support it and Windows will never be able to support anything other than x86 CPUs. There's no way Microsoft can port all that code to work on a different CPU.

The reason Atom is power hungry and slow by comparison isn't because Intel sucks at making CPUs, it's because they felt the need to stick to the x86 architecture to make the CPU compatible with Windows. The x86 instruction set was designed over three decades ago, the only reason we are still using it is because of Microsoft Windows. Why did Apple jump onto the x86 bandwagon I have no idea, maybe because everyone's doing it? So much for "think different".

Re:No Windows? Great! No Microsoft tax!

[MoxFulder]

So I guess it all comes down to what the definition of the word "unwillingly" is. :-)

Perhaps... I don't think a monopoly or product-tying has to be absolutely unavoidable to be coercive. It just has to be sufficiently pervasive as to make it unfavorable, in terms of cost and effort, to get around it.

I would call my acceptance of the Microsoft Tax "unwilling" since the only way to escape from it is to buy from a low-volume manufacturer or choose from a very reduced selection (such as Dell's Linux systems, presumably negotiated with Microsoft??), thereby negating the benefits of not paying for Windows.

It's the "least bad option" in some cases, but there's a much better option (buy the same computer with no OS) that's artificially suppressed by Microsoft.

Attack?

[Anonymous Coward]

Since when is competition considered an "attack"?

ARM has been making processors for a very long time. If anything, the Atom processor was an attack on ARM's marketspace, since Intel had always, until very recently, been primarily concerned with the high-end consumer/workstation processor market.

But seeing how ARM doesn't currently have Windows support, they won't really get anywhere in the marketplace. Even the netbook market didn't boom until Windows support became standard- it's what customers want.

Re:No Windows? Great! No Microsoft tax!

[node 3]

Dunno why the other comment got modded as a troll; most companies will happily refund you the cost of Windows; it's been documented many times.

Where to even start?

most companies

Not even remotely true.

will happily

"Happily" is a somewhat meaningless term here. Let's assume you mean this as a synonym for "readily and easily". Not true. Can you point me to the web form or phone number or even postal address where I request my refund from any major PC maker?

it's been documented many times

This is extremely misleading. It's been documented many times because each time it happens, it took tenacity and is not the normal outcome. Sort of like how natural cancer remission has been documented many times.

What it doesn't mean is that it happens so much, it's normal and simple. Nothing could be further from the truth.

Unless, like I mentioned, you could provide the requisite contact info for the refund? A web form seems like the most obvious choice for a company that will do this "happily" and does it so routinely that "it's been documented many times".

Re:No Windows? Great! No Microsoft tax!

[Hurricane78]

Uuum, why exactly didn't you build it yourself, out of off-the-shelf components? It's cheaper for what you need, almost all the time.
I'd never ever buy a whole "computer" in one piece again. I mean half of the stuff in there is usually just crap that you have to replace anyway.

Or did you mean laptops?
Luckily I never had to dive into that business.

ARM == Hype

[Erich]

I'm continually unimpressed with ARM implementations.

They claim that they are low power and high performance, but at the end of the day, on real benchmarks, they always disappoint.

And it's not surprising. It's a serial architecture with only 15 registers. That's the same as x86-64. x86 decode isn't trivial, but Intel has gotten good at it. And ARM isn't without its instruction set issues; the encoding isn't incredibly straightforward as they've gone back and crammed things in like Thumb2.

So the architectures are roughly a draw. Maybe ARM has a slight advantage theoretically in slightly smaller code size, and Intel likely has an advantage in practice that there is a bigger ecosystem.

So then it all comes down to implementation and process. Both of which Intel does very well. Atom is a full-custom, well floorplanned, very nice chip. It was their first try as a "low power" chip, and it performs quite well for the energy on real-world things you need to do. ARM, which traditionally has sold soft macros, hopes to compete? For those of you software folks, "full custom" == hand-optimized, "soft macro" == source code. Sometimes the compiler can do a decent job, but Intel does fully custom designs for a reason.

I think the main thing ARM has going for it is hype. TrustZone(TM) is the most highly-marketed physical address bit I've ever seen. They lucked into being the default control processor at the low end, and have made a decent business out of selling soft cores to folks needing small amounts of processing power.

But they have no special magic. They have a ho-hum architecture, certainly inferior to something like Alpha. They have a set of simple, low-performance designs that get decent energy efficiency. And they have a great marketing engine and speak loudly.

The real reason you might see netbooks with ARM cores is that the manufacturers are just reusing smartphone chips, or have some other special SOC they want to reuse. You can license an ARM and put around it all kinds of interfaces and accelerators. But now you can get an Atom core from TSMC, too, so maybe the advantage is waning....

Re:ARM == Hype

[FourthAge]

Your post goes against the accepted wisdom, but I think it's accurate. The future of netbooks and smartphones is x86-compatible CPUs: no reason they can't be low power too.

I'd like to add that ARM is also an extremely litigious company whose main focus is not on creating the best technology but on "protecting ARM IP". While I am sure that Intel or IBM would go after anyone making money by infringing their patents, ARM appears to believe that anyone reimplementing their ISA should also be threatened and sued, even if it's a free project being done by some students or hobbyists. Legal threats have closed down a couple of free open-source projects on Opencores.org, merely for attempting a clean-room implementation of the ARM ISA. This is, I think, quite nasty behaviour.

Re:chip supports OS? Hmmm, backwards...

[jimicus]

This is simply a case where a specialized processor designed for highly integrated and mobile uses is trying to break into the mainstream Personal Computing market.

ITYM break back into the mainstream Personal Computing market. Desktop personal computers (in the personal sense, rather than the IBM PC Compatible sense) based on ARM chips have existed before:

http://en.wikipedia.org/wiki/Acorn_Archimedes

(I do like the way the photograph shows a computer from about 1988 with an LCD monitor on top. It looks completely out of place - I'm pretty sure the computer that it's sat on didn't have a 15-pin VGA output)

Re:No Windows? Great! No Microsoft tax!

[Anonymous Coward]

Not to mention that the folks writing viruses would have a lot more work. Granted, this isn't currently a problem on *nix variants, but could one more hurdle hurt anything?

Re:No Windows? Great! No Microsoft tax!

[ak3ldama]

Silly me, replying to an AC Troll... Here you go.

I don't think you know anything about Linux portab

[anti-NAT]

ility. It runs on more architectures than OpenBSD, and you're saying it's far less portable, and that the architectual differences are exposed? Way back in 2000 I ran (Debian) Linux on a Sun Ultra 5, and it just worked. The only issue I had was nmap, and that was likely due to a missing htonX() calls. OpenBSD wouldn't have magically put those instructions in the nmap code if they didn't exist either.

I've written networking kernel code for Linux, and never encountered any CPU specific requirements - it's all abstracted behind function calls.

Re:No Windows? Great! No Microsoft tax!

[Bert64]

Even in countries where Dell sell the Ubuntu based machines, they are only available on a very small subset of systems, for instance in the UK there are a handful of lowend laptops which offer it, no higher end laptops or desktops of any kind.

Ofcourse as like anywhere else these days, all of the servers have the option of coming with linux.

Re:No Windows? Great! No Microsoft tax!

[Bert64]

Doing so incurs a performance hit, quite a significant one, as well as using extra memory.... When DEC did it with the Alpha, current Alpha processors were hugely faster than any available x86 so even with the performance hit you got comparable or better performance than using a real x86 system. ARM processors are not as fast as current x86, and the performance would be so poor as to eliminate the benefits ARM has over Atom.

Re:No Windows? Great! No Microsoft tax!

[Hadlock]

The short and sweet solution is to buy it on a credit card, print out the EULA, highlight the pertinent part, mail it to your CC company, and then have them charge back the OEM cost of Windows. You'll have to prove you contacted customer service, but that's typically not hard, and just notify them you're going to do a chargeback for the cost of the OEM windows install. I doubt they'll balk much about it; there's not much they can do once the laptop has delivered.

ARM vs. Atom

[AnnonUSA]

That is like inventing a new car that won't run on current roads.

Re:No Windows? Great! No Microsoft tax!

[Anonymous Coward]

That doesn't really disprove the GP.

If you buy a PC with Windows it isn't really a scam that you pay for Windows. Just do what I and the GP did: buy a PC without Windows.

Re:Is this what will undo the x86 ISA; who cares?

[bitemykarma]

Re. Alpha, as might be apparent from other posts there is more than one market for chips. According to http://h18002.www1.hp.com/alphaserver/?jumpid=reg_R1002_USEN there were over 1 million Alpha chips sold between 1992 and 2007.
That's OK, considering that it was derailed by HP in favor of the Itanium.
Maybe to some people, every chip has to run Windows to have any importance, but that is a parochial attitude. Those Alphas are out there doing some serious work, rather than spreadsheets and games, and they were doing so when Intel chips were good only for spreadsheets and games.

Re:No Windows? Great! No Microsoft tax!

[hedwards]

But, it's not really about performance, sure you do need a certain amount in order to be competitive, but ARM is far more about efficiency and battery life than it is about speed. That being said, it's not necessarily going to be too slow unless you're talking about having to emulate non-existent instructions and similar.

Re:I don't think you know anything about Linux por

[TheRaven64]

Okay, how do you enumerate the amount of system RAM and set the power saving state on Linux, in a way that is portable across all platforms on which Linux runs, and doesn't depend on a third-party (i.e. not part of Linux) library?

Re:No Windows? Great! No Microsoft tax!

[Darinbob]

Forget free/open-source. I'd be happy if the hardware was just accurately documented so that others can write a driver without having to reverse engineer things. Even a redistributable HAL object file would be great. I've never understood just what secrets they think they're protecting, as the important intellectual property is almost always in the hardware and not the driver.

Re:No Windows? Great! No Microsoft tax!

[Eil]

I've been unwillingly paying the Microsoft tax for TEN YEARS.

Sweet Lord Baby Jesus, why? It's been possible to get barebones computers with no OS for much longer than a decade. It takes close to zero effort or knowledge to put in a hard disk, memory, and CPU, and fire the system up for OS install.

If you're talking about laptops, there have been vendors that specialize in Linux laptops for years. They are almost a moot point now, however, with most major OEMs starting to preload Linux on their x86 laptops and netbooks despite licensing threats from Microsoft.

Close enough?

[Anonymous Coward]

How about an ARM and a Lego?

http://en.wikipedia.org/wiki/Lego_NXT

Re:No Windows? Great! No Microsoft tax!

[mikevdg]

If you want a ARM-based notebook, you could get a Touchbook. It contains a Cortex A8-based ARM CPU: http://www.alwaysinnovating.com/

No Windows support is a feature

[gig]

No Windows means no PC monopoly issues, no legacy technology, no malware, and the device you make an ship will not be yet another yawn-inducing generic PC.

Re:No Windows? Great! No Microsoft tax!

[MoxFulder]

The short and sweet solution is to buy it on a credit card, print out the EULA, highlight the pertinent part, mail it to your CC company, and then have them charge back the OEM cost of Windows. You'll have to prove you contacted customer service, but that's typically not hard, and just notify them you're going to do a chargeback for the cost of the OEM windows install. I doubt they'll balk much about it; there's not much they can do once the laptop has delivered.

Have you personally done this? Just because you think the law or some contract or agreement allows you to do something, does not mean it will be easy, worthwhile, or even possible in practice.

I have dealt with chargebacks for credit card companies several times. While they are generally very happy to chargeback an obviously fraudulent or unfulfilled process, I highly doubt they would happily agree to do the legal and technical legwork to determine if chargeback for the Windows license (price undisclosed) of a system is justified.

Re:No Windows? Great! No Microsoft tax!

[MoxFulder]

If you're talking about laptops, there have been vendors that specialize in Linux laptops for years.

Yes, and their prices are uniformly higher than those of the big-volume laptop vendors, enough to negate any savings from the Windows OEM license. They have fewer models, sell less volume, and offer many fewer sales.

If you go looking for coupon codes and deals for a new laptop on Slickdeals or eDealinfo or Fatwallet, you'll never come across a system pre-loaded with Linux (with the still very rare exception of some netbooks in the last couple of years).

They are almost a moot point now, however, with most major OEMs starting to preload Linux on their x86 laptops and netbooks despite licensing threats from Microsoft.

A change that I whole-heartedly welcome, and eagerly await!

But still, most OEMs offer preloaded Linux with only a fairly restricted subset of their products, and offer no OS-less options for the rest. While part of this may be due to difficulties adjusting to the support and driver distribution models prevalent in the Linux world, I suspect part of it is still due to threats and sanctions from Microsoft.

Re:No Windows? Great! No Microsoft tax!

[smoker2]

Back when win95 was new, I must have reinstalled that sucker 2000 times in 2 or 3 years. Every time after the first was "unwillingly". Same again with win98. I've never had to reinstall XP, and I've never used vista. But I had learned my lessons by the time XP came about and I didn't even install it once until SP2 came out. A laptop came with it on in 2001, but that was my only exposure for about 3 years. I now keep a legacy machine running XP (AMD 2200XP) (plus the laptop still works) but everything else is linux. In fact I was running linux before I ever installed XP, come to think of it. (SuSe 5.3 -on a Cyrix 333, 384MB RAM. You had to write your own modem init strings and my modem used the same IRQ as my mouse - great fun :( My phone has a faster processor now, and more storage. Not bad in 10 or 11 years)

Re:No Windows? Great! No Microsoft tax!

[smoker2]

Somebody really should mod you funny. It's all about the drivers.

Re:No Windows? Great! No Microsoft tax!

[MichaelSmith]

Well they were porting only from VAX but the argument may still stand.

Please. Do tell.

[xactuary]

'We've had conversations with Microsoft and you can imagine what they entail,' says Schom."

Flying chairs?

Re:No Windows? Great! No Microsoft tax!

[siloko]

While part of this may be due to difficulties adjusting to the support and driver distribution models prevalent in the Linux world, I suspect part of it is still due to threats and sanctions from Microsoft.

You missed out the main reason - lack of a market!

Re:No Windows? Great! No Microsoft tax!

[bhtooefr]

FX!32 is what was being referred to, and it was most definitely emulation of x86 on Alpha.

That said... when Alpha had rough IPC parity, and double the clock speed or more of the best x86 CPUs... of course it's gonna be competitive even in emulation, even with as much as a 50% hit.

But, a 50% hit on a CPU that has IPC parity with the worst IPC modern x86 CPU that Intel still sells, and not much faster clock (and the Intel CPU in question is considered slow...) it's gonna be ugly.

netbooks + webapps + ARM = PC wintel disruption ?

[simplerThanPossible]

The only thing Windows really has going for it is the existing library of PC software.

If netbooks are used primarily for webapps (hence the "net"), the OS and its libraries don't matter so much.

Re:No Windows? Great! No Microsoft tax!

[anss123]

In the last decade x86 code got a lot more complicated too, though.

Not really. The trick DEC used was that FX!32 only emulated usermode code (just like Apple's Rosetta) so it could get by without having to emulate the memory management unit. But you'll need a native version of the operation system to handle kernel mode and the memory management.

Contact!

[Slur]

...in case anyone missed it...

What about a Beowulf Cluster of these...

[Anonymous Coward]

No, this is not a joke. One trend in cluster computing is using a greater number of lower powered CPUs. If I remember correctly, some of the biggest IBM PowerPC based systems are using simpler cores with less pipeline depth, because of electrical power and cooling requirements. It's s cost/benefit ratio effect.

  This A9 ARM could be a real winner for large scale cluster computing, even if it's double precision performance is not so great.

Re:No Windows? Great! No Microsoft tax!

[MichaelSmith]

Oh okay. I thought you meant the VAX to Alpha migration stuff.

Re:No Windows? Great! No Microsoft tax!

[Haeleth]

They should take a tip from the DEC Alpha

That's like saying anyone developing a new video recording system should take a tip from Betamax.

The Alpha was a great chip. But running Windows didn't help it. It still lost to x86 in the end, despite being a superior product in almost every respect.

That doesn't sound right to me...

[Anonymous Coward]

Just a sidenote: The 3D (and 2D?) acceleration core used in the GMA500 (ImgTech SGX core) is build into a good number of recent ARM based SoCs aswell (for example Beagleboard with the TI OMAP SoC). As far as I followed the whole thing, ImgTech itself doesn't provide any drivers (contrary to NVidia, for example), but seems to put that task on the shoulders of these who make the chips or sell the computer hardware.

That's not strictly correct. ImgTech supply reference drivers to the technology customer who is then at liberty to adjust it to fit their SoC. (They'd have to anyway as registers/memory/interrupts etc etc will be mapped differently on each device). A customer is probably also at perfectly liberty to write their own driver.

Since every device would be different, I doubt ImgTech could directly supply drivers to an end user. (shrug)

Apple Tablet

[wood_dude]

I'd be very interested in the rumored Apple tablet with iPhone OS running on one of these chips. That would be a great device to have on the coffe table. ! Chris

Re:No Windows? Great! No Microsoft tax!

[ak3ldama]

The person I replied to was saying there was no tax. I gave an instance of a tax. Thus I disproved him. I wouldn't have even cared to comment except for the colorful language.

There is No Microsoft Tax Except on Lazy Twats

Re:No Windows? Great! No Microsoft tax!

[Hadlock]

Nope, I haven't done this. I buy my laptops used (or Ubuntu Netbooks), and build my desktops from scratch. Note I'm not the whiny guy bitching about the MS "tax". IANAL, but I don't doubt the credit card company would do this, especially if you pay your bills on time.

It doesn't support Windows?

[DaVince21]

I think you mean "Windows doesn't support it" and not "It doesn't support Windows".

Virtual PC

[tepples]

Windows on ARM runs existing Windows x86 software about as well as Linux does: not at all.

Unless Microsoft can manage something like the Rosetta emulator in Mac OS X, which lets PowerPC apps run on x86 machines. Specifically, I seem to remember Microsoft bought Connectix for its Virtual PC product, which back then was an x86-on-PowerPC emulator.

Re:Virtual PC

[IGnatius T Foobar]

Unless Microsoft can manage something like the Rosetta emulator in Mac OS X, which lets PowerPC apps run on x86 machines.

Come on here, we're talking about a company that is putting a virtualized Windows XP inside Vista 7, both running on x86, because their API's are so broken they can't even map them to themselves cleanly. (sup dawg, I heard you like computer crashes so I put a Windows in your Windows, so you can bluescreen while you bluescreen!)

Rosetta is an emulator

[tepples]

Wine Is Not (an) Emulator.

Rosetta is an emulator. What stops Microsoft from making a counterpart that can dynamically recompile userspace x86 code to ARM code, especially since it bought Connectix for the x86-emulator code base?

Re:No Windows? Great! No Microsoft tax!

[ak3ldama]

"First batch is very limited. Reserver yours now" They're not out yet. More-so they are more of an up-market item. Though it appears they offer a model for only $299.

I know of no ARM Netbook that is actually on the market yet. It has been a media illusion for the past year or so. But then, if you have a link to one that is actually on sale please let me know. I have in the past submitted an entry for Ask Slashdot wondering what barriers the ARM Netbook makers are facing but it was not accepted.

Re:chip supports OS? Hmmm, backwards...

[petermgreen]

I'm pretty sure the computer that it's sat on didn't have a 15-pin VGA output
At least some of them could drive a VGA monitor though, you just needed a wiring adaptor to connect them up and I think to configure them to the right (which could be "fun" to do with your monitor not operating) speeds.

Re:chip supports OS? Hmmm, backwards...

[jimicus]

Oh yes, I know that. I'm pretty sure the A5000 and later machines did natively.

But the machine in the photo looks more like an A400 series to me - not sure which vidc they used but I'm not sure it was a huge improvement over that in the A300 series and that drove a TV-resolution monitor.

Re:chip supports OS? Hmmm, backwards...

[petermgreen]

But the machine in the photo looks more like an A400 series to me - not sure which vidc they used but I'm not sure it was a huge improvement over that in the A300 series and that drove a TV-resolution monitor.
Basic VGA resoloution isn't that much higher than TV resoloution. I know i've seen a VGA monitor driven off an A3000, I don't have any experiance with anything lower than that.

Also I saw some mention online of vidc upgrade podules.

Re:No Windows? Great! No Microsoft tax!

[Ben Hutchings]

I'd love a low-power, high-performance ARM notebook. I'd be happy with MIPS or Loongson (Chinese MIPS clone) as well. Debian already has a full-blown ARM port and I'll bet they could get it working on an ARM netbook in a day. Ubuntu would undoubtedly be soon-to-follow.

There was an ARM employee at DebConf with a protoype ARM-based netbook. I forget whether it was running Debian or Ubuntu, but in any case the software is basically ready.

As a side benefit, having multiple widely-used architectures for desktop systems (x86 and ARM) would be a support nightmare for hardware companies that still keep their drivers proprietary and undocumented. Yeah, I'm looking at you, Broadcom and NVidia.

I'm pretty sure Broadcom is already providing drivers for ARM as many home routers run on ARM. As for Nvidia, do they have any chips that fit in the price and power brackets of netbooks?

I see a significant barrier to ARM adoption in GCC's code generator, which is poor and apparently getting worse as the optimisations are more tuned for x86 than RISC processors. ARM, like Intel, has its own commercial compiler but it would probably make business sense for them to contribute to GCC as well, just as Intel does.

Re:No Windows? Great! No Microsoft tax!

[MoxFulder]

I see a significant barrier to ARM adoption in GCC's code generator, which is poor and apparently getting worse as the optimisations are more tuned for x86 than RISC processors. ARM, like Intel, has its own commercial compiler but it would probably make business sense for them to contribute to GCC as well, just as Intel does.

Interesting... I hadn't heard much about the GCC code generator. Do you have any sites with more info?

I know that GCC is pretty darn good for some other RISC processors, like MIPS.

I imagine GCC's ARM code generator would get more attention in a hurry if it got more use on the desktop. That's a great thing about free software :-) And I agree that it would make sense for ARM to fund GCC development, or simply hire coders to work on it.

Re:No Windows? Great! No Microsoft tax!

[Ben Hutchings]

As for Nvidia, do they have any chips that fit in the price and power brackets of netbooks?

Hmm, apparently they do, in the form of their own ARM SoC...

Re:No Windows? Great! No Microsoft tax!

[Ben Hutchings]

No references; it's just what I've heard (backed up by examples) from friends who develop ARM-based embedded systems.

Re:No Windows? Great! No Microsoft tax!

[MoxFulder]

Gotcha. Did they mention any specific compilers that do generate good code for ARM? Presumably ARM's own is the best, but any others...?