News Windows 7 on ARM a possibility?

http://www.bit-tech.net/news/bits/2009/05/05/window-7-on-arm-a-possibility/1

Comments from Microsoft's CEO point the finger at an ARM-compatible build of Windows 7 in the pipeline, with the company unwilling to give Linux a foothold in the market.

 

if MS does get Windows 7 to work on ARM, would it not be a big kick in the pants for Intel? a CPU that does not need their license but can run all of today's programs (with all of ARM limitations), might just be what the future of computing needs.

 

Will all windows applications work just like it does now? What about all those applications that need x86, emulation?

 

This means Nvidia will be able to make a CPU for Windows after all since Nvidia Tegra is a system on a chip on ARM architecture!! This will hurt Intel more than AMD tbh!

If this is indeed true then great move Microsoft!

 

Interesting. I'm currently doing some work on a Cortex A8 based CPU, running Linux. While this would be an interesting development, I have a few reservations regarding Windows 7 on the ARM.

These mainly revolve around Compatibility with existing software, and OS Size / Speed. Linux has already been pared down and runs very well indeed on these low power CPUs. I have a build here on my ARM A8 using only a fraction of the resources of any recent Microsoft OS, running very nippily on a system with only a 512Mb flash drive and 256Mb of DDR1 RAM. On the Driver / software compatibility front, the ARM is a totally different architecture to the x86. There will be a lot of hardware interface problems to overcome, which Linux has already had years to sort out. All the Windows API calls will need to interface to a completely new underlying architecture, meaning that they can't be cross compiled, they will surely have to be re-written. When doing this, MS will have to make certain that they maintain the common interface required for program compatibility with the x86 machines. This will essentially give Microsoft another code base to maintain, and even more trouble than they already have performing bug fixes and keeping support across the two very different architectures.

That said, if they pull it off, MS may well get a foothold in the growing ultra portable market. Sadly there seem to be many people who lack the patience to try something different, and would happily take the machine with the MS Windows install on it just because it means they don't need to spend time learning their way around it. Furthermore, if MS did port full Windows 7 to the ARM, they could potentially do away with Windows Mobile; I can imagine that companies that have to make Windows x86 PC compatible utilities to complement their hand held smartphones / similar products may welcome this - code recycling would be much easier if their portable hand held devices ran the same OS as the machines they usually talk to. Ultimately I feel it would come down to OS stability and how steep Microsofts Royalties / installation fees will be for use of "Windows 7 ARM edition", and what, if any limitations MS impose upon the ported version of their OS.

 

Definitely worth watching.

 

as Bodger said, i don't think it's going to happen within 1 or 2 year of Win7 official release, there's simply too much needs to be re-written.

but great news, more reason to study RISC architecture!

 

The Bodger: unless WinMO devices improve a lot in speed then they are no space for a full blown version of Windows.

Otherwise I think this sound great, but I don't want to see it working but unusable - putting too much of a strain on the limited resources available and bogging everything down.

edit: actually, if my assumptions are correct, which they probably aren't, why would they need two version of Windows? Once they have made one that will run on a reduced instruction set processor, like the relevant ARMs, then will that not work on an x86 CPU too? Or do reduced instruction processors have some different ones to x86?

 

Remember that apple made the leap to intel without a huge amount of (perceived) trouble...

 

If they do it in a reasonable timeframe, they'll just slap an x86 virtual layer over the base hardware and have the arm transparent to most installed apps. Tweaking the .net frameworks with an optimised JIT compiler will restore performance to the apps which play things the MS way.

The more hardware choices the better!

 

I'll certainly be watching this with interest, in the mobile device market this would make application portability and functionality a much more achievable goal.

 

Oooh! Me likey..

 

Ironic that after all these years Apple redesigned their OS from RISC to X86, and now Microsoft might go from X86 to RISC. Does Windows Mobile CE run natively on RISC processors?

 

Windows CE (which Windows Mobile is based on) runs on several instruction set platforms like ARM, MIPS and x86. Microsoft has enough knowledge from WinCE to be porting Win7 to ARM. And they can in fact cross-compile a lot of stuff. They only need to re-write the kernel and Win32 subsystem. All userland code will just have to be recompiled for a new ARM-target just like WinCE apps.

This will in many ways be no different than the current situation with AMD64 (x86_64, x64) apps vs. old x86 code.

 

OSX is unix

 

Also dont forget Microsoft use to make WIndows NT and 2000 for alpha chips, so it would not suprise me if there is a code tree for 7 that would work with a different cpu arch

Kimbie

 

This holds far more appeal to me than a desktop version of ARM-based Windows 7. I've long been a fan of Windows Mobile devices (my latest being the Touch Diamond), but it just feels like there's something lacking.

One thing I have noticed is that the centralised application distribution platforms of the iPhone and Google Android have both helped to spread adoption of what is essentially a high-power smartphone. Windows Mobile-based devices have been around for much longer, but haven't had the same level of consumer adoption. With their "cloud computing" projects and Live, it does seem as if they are attempting to moving to a completely seamless service. It's not hard to imagine a world where I can seamlessly move between my mobile phone, desktop PC and netbook, while retaining all my applications, data & services. That's possible now, but by using a more familiar platform for developers, it could make the process a whole hell of a lot easier.

Not to mention the difference it can make to actually using the device itself; Windows Mobile has made great strides over the years, but it's still not quite there. At it's heart, it's still an overhauled version of WinCE. It takes third parties, such as HTC, to make vast improvements on the interface and improve the user experience (see HTC TouchFlo).

I do recall reading a while ago that Microsoft have made progress in overhauling the kernel - now called "MinWin" - of the operating system and reducing it's dependencies; a stripped down version was demonstrated which only consumed a measly 25mb of disk space and 40mb of RAM. It's not hard to imagine that miniature core kernel forming the basis of an ARM-based version. (linkies: http://arstechnica.com/microsoft/ne...ows-7-taking-shape-meet-the-minwin-kernel.ars , http://blogs.zdnet.com/microsoft/?p=842 and more from google: http://www.google.co.uk/search?rlz=1C1CHNH_en-GBGB326&aq=f&sourceid=chrome&ie=UTF-8&q=minwin )

 

WinMin seems to be very interesting. So does Live Mesh.
I think we'll see a lot of development and change in both the mobile devices and cloud computing/thin client sectors in the next couple of years.

 

While Eric Traut's now infamous demonstration is nothing more than an academic exercise, it is a proof of concept. If we're talking about Win7 on mobile devices, then the first steps have already been taken - notwithstanding the task of porting it to a different architecture .

It won't be the desktop version of Win7 on a mobile device, but the very core of the platform could be the same.

Although instead of philosophising on the future of Windows on ARM, I think for now I'll be content with waiting for Windows Mobile 7

 

This will open up ION2 devices that would currently be limited by the lack of a compatible windows too. Finally it might push intel to do something about the shocking chipset they force to go with atom cpu's.