Intel and ARM collaboration?!


#1

Yes, you read that right.
The possibilities are countless, and a powerful ARM chip capable of running full blown Windows is not just a dream anymore…
This could create a new category in the smartphone industry and take us to new heights of computing.
The tablet that can replace your laptop?
How about a phone that can replace everything?
unnamed
Thoughts?
Source:


#2

Intel has made ARM chips themselves before, so no surprise. But keep in mind this collaboration here has nothing to do with Intel’s instruction set. It’s only about the manufacturing technology.


#3

I’m sure MS has brought them together. This will help MS most…


#4

Eheh, you’re talking about Intel changing ARM so that it’s no longer ARM, but something that windows will run on? I don’t think that’s gonna fly anytime soon with the ARM community. Doubt it’s gonna happen.


#5

Yeah, he’s probably referring to this emulation thing that Microsoft is bragging about. But then again, this collaboration has nothing to do with chip design, so it won’t make the emulation better in any way. It’s only about the manufacturing process, Intel has been said to potentially hint about suing in case of a breach of one or many of their x86 patents and that still stands.
(thanks @Mike for correction)


#6

This. I’m pretty skeptical too, but I can imagine this being easier to implement then say, an Intel x86 CPU inside a smartphone( with adequate cooling and decent battery life).
Easier to implement does not mean it’s better, I agree, but this is just the beginning…


#7

And I can’t… I mean look at Asus Zenfone 2 for example, it has good enough cooling and really good battery life. Now replace all that plastic with metal… You see, that top of the line Snapdragon that Microsoft has showed off has pretty much the same power draw as a Core M. I don’t imagine it performing better, lol…


#8

What do you suppose are the chances of Intel emulating their own x86 instruction set on their own ARM processors? Would they still be able to maintain it as a proprietary IP and not allow other ARM manufacturers to emulate as well? It seems like if they could keep it in-house and manufacture without sacrificing ownership of the emulation ability, they would be able to control a significant market of ARM-based x86 devices. It still won’t be easy to implement, but if I were to place bets, it would be on a collaboration between Intel and ARM to get it done.


#9

Their own ARM processors? Why? They won’t make ARM processors. They will just share the manufacturing technology with ARM.


#10

In order to monopolize on ARM-based x86 emulation.


#11

Yeah but they aren’t gonna make ARM processors.


#12

ARM chips are a logistical nightmare of epic proportions. The major issue with it is compatibility and portability issues. I used to own a 2-in-1 android TF701 laptop that was running an ARM chip and Android OS. I tried to install any flavor of Linux which failed spectacularly because of the difference in architecture. I even attempted to do some major coding to get a flavor of Debian working on it and the results, were mildly disappointing.

There is also not a wide range of ARM based applications, at least not as diverse as what the Intel devices can have access to. Legacy applications would of course, cease to exist and good luck trying to play any existing or future games on an ARM chip computer.

The only saving grace to ARM chips is that they are lightweight and power efficient, which is good for battery runtime. They are good at decoding instructions but Intel is better at executing those instructions and putting them into system memory.

Essentially, ARM chips are still very slow and not suitable for thousands of tasks we do on our computers or laptops despite the gains in power-efficiency.

Would you trade, three decades of optimisations and industrial raw power of a highly developed chip like Intel for an immature line-up of ARM that has yet to find its ground?


#13

As @pauliunas says. The big problem for MS introducing Windows on Arm is definitely Intels rights of the x86 command set. However MS is calling the thing. In my opinion it’ll ever be a emulation of x86 to get an Arm Processor working with “the real Windows” Intel won’t be that stupid to give this clincher out of their hands. And this kind of “Joint Venture” has really nothing to do with new hardware coming up there at all.


#14

Well, the “solution” that Microsoft is proposing is essentially emulation. They would probably port the “under the hood” Windows API stuff to run natively, but software using it would run in an emulated x86 context.

Actually it’s also quite mature, it just isn’t optimized for desktop usage. It would take a long time for it to scale up so much to be usable for desktop computers, if that’s at all possible. But in general, it’s also a very old and advanced architecture. As you can see from the Wikipedia article, it was introduced in 1985:


That’s just 7 years “younger” than x86.
They’re not “slow” compared to Intel’s laptop processors, but they’re boggled down by the fact that they need emulation to run any useful software. There are also some quite fast ARM processors used in servers.


#15

In the context of Linux this is, basically, not true. The vast majority of software available for Linux is open-source, and all(most all) of that can be compiled to run on ARM.


#16

He probably meant pre-compiled software that can be downloaded straight from a package manager. Not everyone can/wants to set up a build environment for every app they want to install. I mean the same environment might work for several apps, but then you need to find the right compiler settings for example… and what about different languages…


#17

Let’s sharpen this statement a bit, how would you like;

“Intel has been said to potentially hint about suing in case of a breach of one or many of their x86 patents”

?

All the things these companies do (of this size) have to be considered in context of what they can do.

Way back, the US economy was basically consolidated by a few giant trusts, that shared the markets between each company inside the trust (most famous was probably Rockefeller’s Standard Oil).

These trusts were broken by congress with Anti-Trust laws and so huge conglomerates like standard oil were broken down to smaller companies. This was ofc done to introduce competition.

Naturally all these huge (effectively monopoly) companies like MS or Intel have tremendous resources to affect the decision making (even lae making should they choose to do so) but clear abuses will be penalized harshly. As such, they have to be really careful in order not to break these laws, and this is why it would seem unlikely that the scenario you describe (which makes sense economically and for any company though) would realize.


#18

Yeah that indeed makes the statement clearer :wink: thanks


#19

Depends. I wouldn’t do that on my desktop. Not by a long shot. But if I lose 30% of the Photoshop and SolidWorks performance for 300% battery life on my Eve V m3, while the vast majority of things that I do (browsing, watching video, etc.) aren’t affected, I think its a no brainer

Im not sure if I could agree with this. Qualcomm Snapdragon performance has been stagnating for the past couple of years, and while Intel isn’t making an exceptional progress either in that regard, Intel still have the head start as we discussed earlier, while ARM needs to catch up. Server chips aren’t directly comparable to consumer chips as they often perform poorly in single-threaded tasks, the same reason why AMD chips have been epic fails in the consumer market for 3 generations in a row. But I digress.

The emulation solution wouldn’t be that bad for a start. If ARM gets enough track on the market, then Adobe and other pro applications will start supporting it. Meanwhile, I imagine the customer of the first batch of Windows-on-ARM would only a few CPU-intensive x86 apps. Programs like Notepad++ wont have any problem running on emulation as theyre not CPU intensive


#20

To be honest, I think ARM vs. x86 will always be sort of apples to oranges comparison. They are two different architectures… Intel has the advantage of a powerful core, while ARM has the advantage of fitting multiple times more cores in the same thermal envelope. We have octa-core mobile processors with a measy 2W TDP, while Intel has struggled to fit 2 cores into 4.5W (now they’ve done it, good job). Of course those 8 cores don’t run all at the same time, but 4 of them do and that’s already very impressive. ARM will always be more suited for multi core workloads, just imagine how many of those cores you can fit with desktop TDP… But as I said, it’s an apples to oranges comparison.

You also said Snapdragon is not improving… But where should it improve? What’s the point? Those flagship smartphones are already way overpowered for what they do. Don’t forget that there are no applications on Android that would require more powerful processors than we currently have. That’s why Snapdragon stopped improving - nobody would notice the difference anyway. Intel, on the other hand, is always pushed by ultramobile users for efficiency and by enthusiasts for raw power. Currently, many people still don’t buy 2in1s to replace their laptops/desktops because the processors are still too weak, when they’re already used to and have software that basically requires a desktop processor. Or because 2in1s are expensive, but that also boils down to improving the manufacturing process, making CPUs better and cheaper. As for enthusiasts, I guess I don’t need to explain their need for performance…

I don’t see anything like that in the smartphone market. There is nothing that would push Snapdragon to improve, except competition from Mediatek, but they seem to hve successfully divided the market in two parts and be happy with it.