Why doesn't Intel use its superior manufacturing process and makes an ARM SOC?

[klansek]

I am just wondering...since Intel has a clear lead in manufacturing process, why does it not make an ARM SOC which would clearly be the most efficient and would thus eat market share from others manufacturers who rely on less efficient manufacturing process offered by TSMC and others. This way they would command all segments of the market - from mobile to server space.

 

[Vesku]

They had the XScale line of ARM cpus: http://en.wikipedia.org/wiki/XScale

Sold it off, Intel didn't like the margins in the ARM space. Intel has recently dropped some feelers about opening up their fabs a bit to other companies, but nothing as open as TSMU or GF fabs.

 

[exar333]

It would likely only make sense if Intel could not either (1) use all their existing fab capacity or (2) producing ARM was a higher-margin than existing production.

Either 1 or 2 is unlikely at this point. Look at SSDs, Intel is already exiting that market because of the shrinking margins.

 

[TuxDave]

Let's say Intel makes an ARM SoC and helps promote ARM's popularity so that majority of its business was ARM. It's now stuck in a place where:

1) Has to pay license fees to ARM
2) Is in a market with a lower barrier of entry (anyone can get an ARM license)

vs

1) Owns its own license
2) Barrier of entry is very high since no one except AMD can license x86


To me I still think it's in Intel's best interest to make sure x86 is the dominant instruction set by focusing on making x86 uber.

 

[sm625]

Intel is dumb for not pasting a cookie cutter ARM core into their designs. Tap it into the ring bus and let Windows use it to do background idle tasks, allowing the bigger cores more sleepy time. I said the same thing about AMD. These guys are way behind on what appears to be common sense stuff.

 

[greenhawk]

it's in Intel's best interest to make sure x86 is the dominant instruction set by focusing on making x86 uber.

it is just another example of not the best technology being used for a task, but who has the largest cheque book to make their technology the "perfered" option.

stuff what is best for the customer.

 

[aigomorla]

cuz its better shifted into controller cards... like intel's iop 348 and SSD's ie joint micron FAB.
http://www.intel.com/p/en_US/embedded/hwsw/hardware/iop

In theory ud sell a lot more SSD's + controller chips then u would a ARM processor.

Well not anymore... but back then... because controller cards are all going to HSBA which i dont like because it gets picky with Hard drives because of TLDR.

Intel shifted on SSD's... cuz they saw it as the future.

 

[waffleironhead]

The margins on the chip sales are not high enough to bother.

 

[greenhawk]

Tap it into the ring bus and let Windows use it to do background idle tasks, allowing the bigger cores more sleepy time.

There is the issue. Getting Microsoft and any / all other OS writers to change the underlieing code to take advantage of this new feature (espicially when the new hardware uses a different instruction set). Microsoft is more interested in making the front end look better for the masses than worrying about any improvement at the back end.

 

[sm625]

Intel should have a 16 channel SSD controller built right into its CPUs too. I been saying that for years also. I guess it just makes more $$ to charge an extra $50 for it. Just like with the northbridge and southbridge chips. Hell they'd have an east bridge and a west bridge if they could. It is so shortsighted because intel could have totally cornered the SSD market, as well as the motherboard market since everyone would want a motherboard with 4 SSD-DIMM slots.

 

[aigomorla]

Intel should have a 16 channel SSD controller built right into its CPUs too. I been saying that for years also. I guess it just makes more $$ to charge an extra $50 for it. Just like with the northbridge and southbridge chips. Hell they'd have an east bridge and a west bridge if they could. It is so shortsighted because intel could have totally cornered the SSD market, as well as the motherboard market since everyone would want a motherboard with 4 SSD-DIMM slots.

ud think that til IDC gave u a presentation on how difficult it is to change a node on a cpu and not have it laced with ERATA. :whiste:

Its like the 920 C1/C0... intel decided to move something over to another location for efficiency sake... and brought out the 920 D0.

The 920 D0 Spanked silly left and right the 920 C1/C0, and it was all down to a slight revision on the node.

 

[bryanW1995]

Intel is dumb for not pasting a cookie cutter ARM core into their designs. Tap it into the ring bus and let Windows use it to do background idle tasks, allowing the bigger cores more sleepy time. I said the same thing about AMD. These guys are way behind on what appears to be common sense stuff.

Huh? That would give software developers a very strong reason to develop more and more software for arm cpus. The whole point is for intel to push everybody to keep using x86/x64 b/c of the enormous barriers to entry. AMD has proven that even with a LOT of money it's nearly impossible to go toe-to-toe with intel, so there are added psychological barriers as well. Possibly the only people on the planet who would consider it are JHH and Pat Gelsinger, and it's unlikely that either of them could pull it off, anyway. So intel must maintain their x86/x64 dominance while at the same time ensuring the continued dominance of the x86/x64 brand/instruction set/whatever you wanna call it if they want to continue to keep their extremely high margins/stock price. If/when ARM takes over then intel can take nvidia's route into commercial, low volume/high profit segments, but eventually they'll start renting out fab space to arm supliers, and they might even get to a point where they ONLY build things for other companies ala TSMC. That doomsday scenario, as crazy as it sounds, might not be too far away, maybe 15-20 years if arm designs can continue to scale.

 

[klansek]

The margins on the chip sales are not high enough to bother.

I found this puzzling...if this is true, from what then are all this companies, who pay not only ARM for using their tech but also to manufacturing companies, make money?

 

[Lonyo]

I found this puzzling...if this is true, from what then are all this companies, who pay not only ARM for using their tech but also to manufacturing companies, make money?

Volume.
Small margins doesn't mean no margins.

 

[TuxDave]

it is just another example of not the best technology being used for a task, but who has the largest cheque book to make their technology the "perfered" option.

stuff what is best for the customer.

You're assuming ARM is best for the customer.

 

[aigomorla]

seriously guys.. intel is in enough trouble with monopoly...

just imagine if they took over the ARM sector too, that means ... anything which has a cpu that one uses either phone calculator, to PC, has an intel branded chip.

That would lead to a biblical law suit from every country in the world that intel would lose.

 

[Ferzerp]

You're assuming ARM is best for the customer.

This.

Like I said on the last "OMG ARM!!!!!11" thread. I want to take the ignorant journalist who started this whole misunderstanding and hit him with a stick.

Arm is slow in calculation in comparison to what we currently use, and an order of magnitude or more slower than what we use now for memory IO.

Arm is not magical. It's a low power, low performance solution.

 

[ElFenix]

I found this puzzling...if this is true, from what then are all this companies, who pay not only ARM for using their tech but also to manufacturing companies, make money?

in an economic sense the only company likely making money on the ARM segment of the processor is ARM.

every other company makes their money from the rest of the SoC (value add).

but an ARM license is cheaper than developing your own thing and gives you access to an increasingly larger group of developers.

Arm is not magical. It's a low power, low performance solution.

there shouldn't be any technical reason ARM can't produce high performance designs. problem is, if ARM were to design an arch that's similar in performance to sandy, it'd suck down nearly as much power as sandy. and that's assuming ARM designs could get manufacturing as good as intel's (which they can't), and intel's been designing these things long enough they have a lot more experience optimizing for the performance level (which ARM doesn't have). for proof i offer every other RISC arch that's tried to go up against x86 in the last 20 years.

you can't beat intel trying to copy intel unless intel screws up. the last time intel tried something radical in the x86 space amd came out with the decidedly P6-like K8 and beat intel for a while. when intel went back to a more P6-like arch they started whipping AMD. intel's been making evolutionary changes to the arch since. that's why AMD has decided to go with a radically different arch this time which is built to offload to a GPU (and will probably suck until software can really take advantage of that ability). unfortunately for AMD intel can just throw engineering resources at any good idea AMD has and beat them to the punch if it wants.

 

[soccerballtux]

it looks better if they can strong ARM them with x86

 

[soccerballtux]

in an economic sense the only company likely making money on the ARM segment of the processor is ARM.

every other company makes their money from the rest of the SoC (value add).

but an ARM license is cheaper than developing your own thing and gives you access to an increasingly larger group of developers.



there shouldn't be any technical reason ARM can't produce high performance designs. problem is, if ARM were to design an arch that's similar in performance to sandy, it'd suck down nearly as much power as sandy. and that's assuming ARM designs could get manufacturing as good as intel's (which they can't), and intel's been designing these things long enough they have a lot more experience optimizing for the performance level (which ARM doesn't have). for proof i offer every other RISC arch that's tried to go up against x86 in the last 20 years.

you can't beat intel trying to copy intel unless intel screws up. the last time intel tried something radical in the x86 space amd came out with the decidedly P6-like K8 and beat intel for a while. when intel went back to a more P6-like arch they started whipping AMD. intel's been making evolutionary changes to the arch since. that's why AMD has decided to go with a radically different arch this time which is built to offload to a GPU (and will probably suck until software can really take advantage of that ability). unfortunately for AMD intel can just throw engineering resources at any good idea AMD has and beat them to the punch if it wants.

but would it? I haven't looked super close at the x86 architecture, aren't there a lot of inefficiencies going on or...no? Something with it being made of RISC now or something.

 

[Arachnotronic]

If designing an Intel killer were so easy, then why can't even AMD -- a company that's been doing this x86 thing for a LONG time -- beat Intel?

This ARM hype sucks -- Intel and AMD aren't going anywhere for hardcore performance. Period.

 

[ViRGE]

but would it? I haven't looked super close at the x86 architecture, aren't there a lot of inefficiencies going on or...no? Something with it being made of RISC now or something.

The x86 ISA is for all practical purposes decoupled from the architecture of the chips themselves. Intel and AMD simply use a decoder frontend to combine and break-up operations as necessary. Intel can change the backend practically at will because they'll just decode instructions for whatever new architecture they use, and indeed the rumors are that Haswell will be that kind of a leap. In any case, a decoder combined with out of order execution means that they can deal with even terrible x86 code well.

The only place the ISA matters is for parallel operations (SSE, AVX), as those operations have to be explicitly bundled together. And no one is complaining about those instruction sets anyhow.

 

[gevorg]

Intel would be shooting itself in the foot by manufacturing ARM. Its goal right now is to crush it by new low power x86 CPUs for tablets and maybe even smartphones.

 

[Fox5]

The x86 ISA is for all practical purposes decoupled from the architecture of the chips themselves. Intel and AMD simply use a decoder frontend to combine and break-up operations as necessary. Intel can change the backend practically at will because they'll just decode instructions for whatever new architecture they use, and indeed the rumors are that Haswell will be that kind of a leap. In any case, a decoder combined with out of order execution means that they can deal with even terrible x86 code well.

The only place the ISA matters is for parallel operations (SSE, AVX), as those operations have to be explicitly bundled together. And no one is complaining about those instruction sets anyhow.

X86 outperforms or competes with almost all other high performance designs, even at a performance per watt level. It's a bit of a jack of all trades, master of none. Even in terms of low power or small die size designs, Atom and Zacate don't compare too poorly to other architectures. (I wouldn't be surprised to see Atom's successor actually make a play for a large portion of the laptop market as well)
Intel even made a decent showing of extending X86 to graphics with Larabee and Knight's Corner. It's not great, but considering how poor the x86 ISA SHOULD be for that work load, it's still within 1/2-1/3rd the performance of nvidia's hardware. And that's pretty much a worst case scenario for Intel, and they get 33% to 50% the performance of a comparable design with a better ISA for the task.
Intel's vastly superior fabrication processes pretty much mean they could force x86 into almost any market. Hypothetically, x86 could have 20% lower code density, 20% lower performance per clock, and 20% worse power consumption, yet with intel's manufacturing advantage they could give it 50% more cache, clock it 50% higher, and spend transistors on all sorts of power gating to reduce typical power consumption.

 

[SickBeast]

ARM pretty much owns the smartphone and tablet market at this point. It's got to be affecting Intel to a certain extent.

The thing is, the world of business runs on x86. There is just too much specialized software that is x86 exclusive. It leaves Intel in a nice position. They hold the x86 license, they are #1 in manufacturing, and they've also got a proven track record of consistently making the fastest x86 CPUs (aside from the few years when the K8 took over).

I would be surprised to see x86 phones and tablets suddenly taking over, but I suppose stranger things have happened.

It's going to be interesting to see what nVidia can come up with. They're going to be a threat to Intel going forward IMO. They have so many connections with not only game developers, but software developers, that it's not even funny. If they can convince the people who make software like AutoCAD and all the film industry's software to port their stuff over to ARM and Linux, it could be game over for Intel in a lot of respects.

The thing is, for a scenario like that to pan out, Intel is going to have to mess up pretty badly.

You never know. With enough engineers and architects wanting to run AutoCAD on their phones, Autodesk might have to make a port. I think they even ported CAD onto the Mac, so who knows, it's possible.