Should Intel openly licence x86?

maxius

2[H]4U
Joined
Dec 17, 2001
Messages
3,376
Obviously arm is knocking at x86's door the proof is in the pudding
http://www.microsoft.com/presspass/press/2011/jan11/01-05SOCsupport.mspx

i honestly feel intel needs to change its licensing plan for the x86 instruction set. not taking anything from via or amd or intel they are great x86 innovators but the threat of arm replacing x86 due to Microsoft's move is very high. who would of thought Microsoft supporting a non-intel processor i bet this has people at the x86 chip makers extremely worried
 
Obviously arm is knocking at x86's door the proof is in the pudding http://www.microsoft.com/presspass/press/2011/jan11/01-05SOCsupport.mspx i honestly feel intel needs to change its licensing plan for the x86 instruction set. not taking anything from via or amd or intel they are great x86 innovators but the threat of arm replacing x86 due to Microsoft new move is very high. who would of thought microsoft supporting a non-intel processor i bet this has people at the x86 chip makers very worried

Arm has a long ways to go to replace x86...

Also, Microsoft has supported many other processor platforms over the years. Nothing special about it supporting another one that would benefit them.
 
Arm has a long ways to go to replace x86...

Also, Microsoft has supported many other processor platforms over the years. Nothing special about it supporting another one that would benefit them.

1 year this drops http://androidandme.com/2010/09/news/arm-reveals-quad-core-2-5-ghz-cortex-a-15-slated-for-2012/

i am sure this does not hurt arms chances either http://www.channelregister.co.uk/2011/01/05/nvidia_arm_pc_server_chip/

you are correct but how many had support in its flagship os?

all i am saying is it looks like arm is coming up fast and with it openly licensing the designs to world + dog perhaps intel should consider the same
 
When I see ARM powering enterprise PCs - game design, code compiling....servers....then I'll say it's moving in to replace x86. But it's not there yet, and afaik, even a quad core A-15 won't do that. These are meant for phones, so we're talking low power; x86 has never been low power, so it's no surprise it has no penetration there.
 
lets try windows os's

so i can find windows xp/7 for PowerPC, MIPS and Alpha. the itanic is intel based after-all and does not qualify :p
 
Last edited:
There is no point, all it would do is slow the move to x64. In fact windows supporting ARM is almost late with AMD's low power APUs that will be able to power tablets in a year's time. And hopefully Atom will be competent with Ivy Bridge.
 
Since you edited:

so i can find windows xp/7 for PowerPC, MIPS and Alpha. the itanic is intel based after-all and does not qualify :p

Since Intel is a licensee of ARM, doesn't that mean that ARM shouldn't count either?

I never said that Microsoft has always supported other platforms for all their products. Just that over the years, when they considered it to be worth the effort, they did so. ARM has 90-98% of the mobile market. MS wants to become as common in the mobile market as they are on the desktop, so it only makes sense that they support the architecture that's got it on lock-down.
 
There is no point, all it would do is slow the move to x64. In fact windows supporting ARM is almost late with AMD's low power APUs that will be able to power tablets in a year's time. And hopefully Atom will be competent with Ivy Bridge.

i agree completely but right now arm has its plans on the table to take over x86. apple is thinking about dropping x86 for arm and microsoft is offering its support to get a piece of the market.

if intel adopts arms licensing plan would be the best way to fight arm on arms turf this will lead to more money for intel and thus bringing even more devices that are x86 based
 
i agree completely but right now arm is a knocking. apple is thinking about dropping x86 for arm and microsoft is offering its support to get a piece of the market.

if intel adopts arms licensing plan would be the best way to fight arm on arms turf this will lead to more money for intel and thus bringing even more devices that are x86 based

I'd like to see a source for Apple dropping x86 for ARM, because honestly that would be the stupidest thing has done in a long time.
 
Since you edited:



Since Intel is a licensee of ARM, doesn't that mean that ARM shouldn't count either?

I never said that Microsoft has always supported other platforms for all their products. Just that over the years, when they considered it to be worth the effort, they did so. ARM has 90-98% of the mobile market. MS wants to become as common in the mobile market as they are on the desktop, so it only makes sense that they support the architecture that's got it on lock-down.

intel dopped its arm xcale processor and team ages ago
 
oh i have no doubt when they had that team they where very intimate with arm architecture. i can not find anything recently that states intel still holds an arm license

From the co-founder of arm its old

“The reason why ARM is going to kill the microprocessor is not because Intel will not eventually produce an Atom [Intel's low-power microprocessor] that might be as good as an ARM, but because Intel has the wrong business model,” said Dr. Hauser. “People in the mobile phone architecture do not buy microprocessors. So if you sell microprocessors you have the wrong model. They license them. So it’s not Intel vs. ARM, it is Intel vs. every single semiconductor company in the world.”

http://blogs.wsj.com/tech-europe/20...cessor-business-doomed-claims-arm-co-founder/

what he says makes sense as to why we are not seeing x86 traction in areas where arm is currently flourishing
 
Last edited:
oh i have no doubt when they had that team they where very intimate with arm architecture. i can not find anything recently that states intel still holds an arm license

From the co-founder of arm its old

“The reason why ARM is going to kill the microprocessor is not because Intel will not eventually produce an Atom [Intel's low-power microprocessor] that might be as good as an ARM, but because Intel has the wrong business model,” said Dr. Hauser. “People in the mobile phone architecture do not buy microprocessors. So if you sell microprocessors you have the wrong model. They license them. So it’s not Intel vs. ARM, it is Intel vs. every single semiconductor company in the world.”

http://blogs.wsj.com/tech-europe/20...cessor-business-doomed-claims-arm-co-founder/

what he says makes sense as to why we are not seeing x86 traction in areas where arm is currently flourishing

X86 not making traction into new areas is no reason to believe that ARM would be able to move in where x86 reigns. The premise of the thread was that it would. ARM is simply not in any position to move into the desktop, workstation, and server markets where x86 chip makers get their bread and butter.
 
Itanium, and Itanium had a way, way smaller install base than ARM.

WHat do people have against Itanium? It has 26% of the revenue of the market right now if not more, against Sparc and Power...they're not the top, but they're still VERY much in the market.
 
WHat do people have against Itanium? It has 26% of the revenue of the market right now if not more, against Sparc and Power...they're not the top, but they're still VERY much in the market.
It was a failure in scope if nothing else.

That said, my comment only reflected on the fact that there are many, many more devices out their running ARM processors than IA-64 processors, and yet Windows has a specific version for IA-64.
 
Not knowing much about the architecture, what pro's does ARM or any other architecture bring to the table over X86? I'm under the impression X86 has remained as long as it has due to legacy support?

I guess my question is, would a new architecture allow vast improvements over X86? Would processing power clock for clock be significantly increased? Less power usage? What's the real benefit besides being free of licensing?
 
Not knowing much about the architecture, what pro's does ARM or any other architecture bring to the table over X86? I'm under the impression X86 has remained as long as it has due to legacy support?

I guess my question is, would a new architecture allow vast improvements over X86? Would processing power clock for clock be significantly increased? Less power usage? What's the real benefit besides being free of licensing?
Three big things ARM brings to the table is: cost, power consumption, and (custom) system-on-a-chip. Without a cheap system-on-a-chip, using x86 in smaller than netbook-sized devices is not particularly attractive.

While Intel processors eclipse the computational power most users need, ARM devices are heading toward being feasible 'main rig' chips for the just-checking-facebook crowd. Licensing x86 would allow Intel to get their fingers in that market without needing to produce their own hardware.
 
Not knowing much about the architecture, what pro's does ARM or any other architecture bring to the table over X86? I'm under the impression X86 has remained as long as it has due to legacy support?

I guess my question is, would a new architecture allow vast improvements over X86? Would processing power clock for clock be significantly increased? Less power usage? What's the real benefit besides being free of licensing?

well firstly you're not free of licensing: Qualcomm pays a fee to ARM to build the all too popular snapdragon, which is why, as quoted above, ARM is being smug to Intel about chip design: ARM doesn't have to get their hands dirty with foundries.

ARM, MIPS, and most other modern architectures do not take the same approach that x86 did in that they have much much smaller instruction sets --in the case of MIPS, less than would fill-out an 8.5" x 5.5" cheat-sheet :p--. Any optimization of instructions is then differed to the compiler, rather than actual microcode hardware. In 2010, with nearly every chip ever made running a level of software in C, this allows compilers to be incredibly sophisticated as there is a huge amount of research into C compiler design.

And in terms of benefits nobody really knows because only those under NDA know exactly how good modern micro-code implementations are. [Brief review of x86: last time I checked it had about a thousand instructions, if dedicated hardware had to be built for each instruction the chip would be massive; so it isn't. Chip designers use whats called "microcode" to split apart big instructions into little ones.] Realistically there are instructions in x86 that your CPU will literally never, in its quintillions of execution cycles, execute.

So, in short, to support these thousand instructions x86 CPU's must have dedicated hardware devoted to microcode. What percentage of the chip is given to microcode? Again, that's NDA'd information, but recently I heard that roughly 80% of the instructions in x86 were done using at least one clock cycle of microcode, meaning ~1000 instructions are widdled down to ~200. If I had to guess how much space that hardware consumed, as an aspiring hardware engineer, on these huge billion-transistor chips... maybe a couple million transistors? Its tough to draw the line between what in Microcode is specifically microcode and what is just regular decode.

So the gains from dropping such hardware aren't enormous, but they do mean that this architecture is suited for large 40W+ TDP chips, as if you try to scale down, by percentage, those few million transistors become a larger and larger part of the die. Atom, as an example,(keeping in mind atom doesn't really support all of what we consider modern x86, as it does not have all the SSE instructions, or the x64 extensions) seems to be about as low as x86 will go (and even then, as most of you are aware I'm sure, its not exactly a great product).

And as for Microsoft porting Windows to other architectures, keep in mind it is all in C, so the vast majority of code isn't going to need re-writing, and if the documentation's good and the code is properly object oriented, I cant imagine there is really all that much to NT which is specific to x86.

my .02.
 
Per the overall scenario,

X86 is not the primary reason, so licensing or not does not help.

The main reason is ARM alliance is establishing a viable global environment with iOS and Android dual pathing into many new and existing markets. This is not theory but have some published results to show.

Thus, if you do not have viable response, and cannot easily alter the course, you may have a situation developing a large installed base that not only cannot run Windows, but also cannot run Office.

And not only that, you may have new Enterprise entries per Office Productivity Tools, which may enjoy vigorous growth due to non-availability of Windows/Office, which may then lead to new situations even back to the home field.

Therefore, Windows must be running on Qualcomm/TI/nVidia/Samsung/etc...
 
So, in short, to support these thousand instructions x86 CPU's must have dedicated hardware devoted to microcode. What percentage of the chip is given to microcode? Again, that's NDA'd information,
I have also heard it is quite small. This is one area of x86 CPUs which probably hasn't grown a lot in transistor count, since it is largely there to support legacy CISC instructions.

Newer instructions usually have a concise to a uOp(s).
 
Thus, if you do not have viable response, and cannot easily alter the course, you may have a situation developing a large installed base that not only cannot run Windows, but also cannot run Office.

And not only that, you may have new Enterprise entries per Office Productivity Tools, which may enjoy vigorous growth due to non-availability of Windows/Office, which may then lead to new situations even back to the home field.

Therefore, Windows must be running on Qualcomm/TI/nVidia/Samsung/etc...

It's a good thing MS Office currently has web versions and hosted versions. So you don't need windows to use MS Office.
 
It's a good thing MS Office currently has web versions and hosted versions. So you don't need windows to use MS Office.

It is correct there is a web version, but that requires MS to give up its current favorable consideration and instead drawn into the strength and playground of Google for a lot of uncertainties.

And that still leave the scenario of fertile ground of hundred of millions non-Windows devices opening to App-stores, multiple.

With this suggested new environment, MS extends its reach into the domain of several enterprising friends, plus opening new paths for its software service division.

In a way this is interesting for both architectures. As it is extremely low-power/low performance oriented, there's plenty of room and plenty of time for all parties before we reached the "good enough" status, thus everything old is new again.

As for existing x86 users with different needs, there's no need to follow if those are not your focus. The existing x86 parts are very good and nobody is being forced to change architecture.
 
Back
Top