Is ARM the end of Intel's monopoly?

[AtenRa]

You know why they are pushing it?
It's really simple.
It's their only chance to compete...Intel has locked down all other sectors for them.

And just because they NEED it...dosn't mean it will happen.

And Intel is pushing x86 in to the mobile Phone market, just because they pushing it doesnt mean they will succeed.

 

[CHADBOGA]

And Intel is pushing x86 in to the mobile Phone market, just because they pushing it doesnt mean they will succeed.

What is your definition of success here?

 

[Scientist113]

Intel does not have a monopoly in the mobile device market, which is ARM's market. So no - Intel's large market-share will remain.

 

[GillyBillyDilly]

Intel does not have a monopoly in the mobile device market, which is ARM's market. So no - Intel's large market-share will remain.

It is about the future here ( 5 to 7 years ). In the future chances are that mobile and desktop chips will merge. PCs will become pocket sized, you stick them to a monitor and are ready to go. The question is If you think at that point Intel will have a monopoly.

 

[GillyBillyDilly]

What is your definition of success here?

In the context of this thread, it would mean Intel pushing ARM out of the market.

 

[Lonbjerg]

And Intel is pushing x86 in to the mobile Phone market, just because they pushing it doesnt mean they will succeed.

Intel has a plan B....do the others?

 

[dma0991]

It is about the future here ( 5 to 7 years ). In the future chances are that mobile and desktop chips will merge. PCs will become pocket sized, you stick them to a monitor and are ready to go. The question is If you think at that point Intel will have a monopoly.

Not every desktop section will be able to merge with mobile. There are people who still needs a powerful PC that gets their job done. As for the rest, they are not limited by the capabilities of the phone but the OS that it runs on.

Even right now, you could plug in your smartphone directly to a bigger screen but it has no real purpose other than entertainment. Android on a stick has yet to gain traction while Raspberry Pi only became a learning/development tool instead of a PC killer due to its limited performance which doesn't give it much legroom to fully emulate a PC's capabilities.

The game changer could be Ubuntu phone which has all of the features that you say of but as it is a completely separate ecosystem from iOS and Android, it has to go through the usual steps of building its popularity and apps. There's no certainty that Ubuntu mobile will be the next Android, it is untested and unproven with normal consumers.

 

[AtenRa]

What is your definition of success here?

Succeed to be competitive against ARM SoCs in Price, Performance, Power Consumption etc.

 

[CHADBOGA]

In the context of this thread, it would mean Intel pushing ARM out of the market.

Why wouldn't a 25% to 75% share of the premium smartphone market in 5 years time be considered success?

 

[King4x4]

Arm is low power stuff.
Intel is high power stuff.

Enuff said.

 

[greenhawk]

One may argue that mobile chips may someday become good enough for desktops too. If in 5 years from now I could get a cpu which has the power of a cheap i3 chip, runnig at 5 watts costing 20 dollars 😎, I'd get one for the desktop.

I might get one as a dedicated media player, but like all things, what you need to run in 5 years time might make that $20 / 5W chip fall to it's knees and beg for mercy.

or worse, it does not have the grunt to run the perfered OS of the time.

 

[AtenRa]

You seem to be missing the bigger picture. It doesn't really matter the sizes of the arm cpu core vs atom cpu core what matters is the size of the entire atom soc vs the entire arm soc.

The current generation atom soc is 5.60 mm^2 at 32nm assumming perfecting scaling that is 2.80mm^2 if it was designed on 22nm.

But it isn't just the cpu that gets smaller when you due a die shrink it is also all the other components of the soc. The l2 cache gets smaller, the graphics get smaller, everything gets smaller.

------------------

Thus assuming equal transistor density (which intel and tmsc do not have) intel 22nm will allow intel to have larger cpu cores then arm at 28nm and 32nm while keeping the soc roughly the same die size. All the while benefiting from 3d transistors/finet and lower power consumption that you are also able to achieve by being at a better process node.

Intels PenWell SoC is close to 62mm2, that is with a single ATOM core and PowerVR SGX 540 GPU at 32nm.

http://www.anandtech.com/show/5365/intels-medfield-atom-z2460-arrive-for-smartphones

Tegra 4 is close to 80mm2 with 5 Cortex A15 cores and 72 shaders at 28nm.

http://www.anandtech.com/show/6550/...00-5th-core-is-a15-28nm-hpm-ue-category-3-lte

Hondo is 75mm2 (+28mm for the FCH) with dual Cores and 80 Radeon cores at 40nm.

Kabini/Temash will be a Quad Core with 128(or more) Radeon cores with die size of ~110mm at 28nm (Integrated FCH in to the SoC).

http://semiaccurate.com/forums/showthread.php?t=6837&page=11

Now, Intels PenWell SoC is the smallest BUT it also has wayyyyyyyyyy lower performance than the rest. With 22nm Intel will be able to double the Transistor count (Dual Core + HT and 4 EUs for the iGPU) and significantly raise the performance keeping almost the same die size (60-80mm) but it will still trail behind in GPU performance.

A dual core Temash SoC with 64 Radeon Cores could be close to 60mm2. With almost half the TDP of current Hondo + FCH chip, it would be wayyyy faster both in CPU and iGPU at the same die size and power usage than Intels Z2460.

Smaller Core sizes help integrate more logic to something else, like the iGPU or add more CPU cores etc.

 

[djgandy]

Why is the Atom core so bloated? 😱

About 10x larger than a Cortex-A7 (even after having compensated for 32 nm vs 28 nm). Is there an inherent difference in the ARM vs x86 instruction set resulting in x86 cores becoming larger?

Also, I notice that the Atom core is about 50% larger than the Jaguar core (after having compensated for 32 nm vs 28 nm), despite the Jaguar core being much faster. 😱

Because the Atom core is 5 years old and the design probably started 7/8 years ago. Atom has barely changed since 2008. Expect 5 years of feedback and knowledge to appear in a new Atom CPU architecture soon.

edit: Clovertrail has 64-bit support which ARM currently don't have at all.

 

[GillyBillyDilly]

Succeed to be competitive against ARM SoCs in Price, Performance, Power Consumption etc.

Wrong. Success for Intel means no competition at all.

 

[djgandy]

Personally, I don't think so.

As a consumer of several mobile (ARM powered devices) I'm already starting to see stagnation in the mobile segment. Newer products are offering less and less breakthrough features and I'm often finding myself upgrading just because I'm bored and not so much because the next best thing is that much better, but I'm sure I'll be doing less and less of that. That said, the mobile segment is still strong because it has not yet been over-saturated like the PC market is, but with Samsung/Google flooding the market with products (both low and high cost) and Apple also shortening their product release timeline, saturation will set in pretty quick

+1. You get it. It is easy to have huge growth when you start from nothing. Do people not learn from history? Imagine what people in 2000 thought all these tech companies would be doing today (and how much profit they'd be making).

Mobile still has a lot of growth in terms of units, but the high end market is over saturated. I can't see $300+ profit per device carrying on much longer in strong volumes.

 

[GillyBillyDilly]

Why wouldn't a 25% to 75% share of the premium smartphone market in 5 years time be considered success?

For very big reasons.
Market share is not important as long as you cannot dictate prices. As long as there is serious competition, you can't dictate prices, and if you can't dictate prices, well, you know the rest.

 

[djgandy]

Succeed to be competitive against ARM SoCs in Price, Performance, Power Consumption etc.

And succeed to make $0.20 per chip?

 

[GillyBillyDilly]

I might get one as a dedicated media player, but like all things, what you need to run in 5 years time might make that $20 / 5W chip fall to it's knees and beg for mercy.

or worse, it does not have the grunt to run the perfered OS of the time.

For 99 percent of PC users the performance of an i3 ivy would be enough all the way. Of course there will also be a niche market for high performance.

 

[IntelUser2000]

A dual core Temash SoC with 64 Radeon Cores could be close to 60mm2. With almost half the TDP of current Hondo + FCH chip, it would be wayyyy faster both in CPU and iGPU at the same die size and power usage than Intels Z2460.

-Point is well made about the ARM cores

AMD's Bobcat is nowhere near advantageous enough over Atom for Hondo to make up for the clock speed difference of 80%. Considering how Temash is staying at the same clock, I bet on CPU Atom won't be disadvantaged.

Check out this benchmark: http://www.xbitlabs.com/articles/cpu/display/atom-cedartrtail_4.html#sect2

Atom D2500 is a Hyperthreading disabled Atom, meaning we can do close comparisons with E-350 per clock.

D2500 is clocked 16% higher than the E-350
E-350 is in average only 30.7% faster than the D2500. Meaning its clock per clock advantage is at 52%, at best. You'd need Temash to perform ~18% better per clock to equal the Atom Z2760. That can happen, but it won't result in a win for AMD. I also think you have been drinking too much of AMD's Kool-Aid. The lowest Temash TDP is at a still high 3.9W.

 

[Idontcare]

It is about the future here ( 5 to 7 years ). In the future chances are that mobile and desktop chips will merge. PCs will become pocket sized, you stick them to a monitor and are ready to go. The question is If you think at that point Intel will have a monopoly.

I doubt Intel would come to have a monopoly in the handset market because their business model ties their hands to only go after the premium portion of the TAM which yields high gross margins, they don't want to touch that bottom 25% of the TAM where margins are paltry.

So long as other companies are willing to live with 30-40% gross margins and near break-even profits, those companies will have a place to sell their chips because Intel has no interest in competing when it comes to those kinds of economics.

Intel could end up with an effective monopoly if everyone else basically walks away from the market...kinda like what has happened in the server segment...because that essentially turns the entire TAM into a margin-rich revenue opportunity where Intel has complete control over the supply/demand pricing curve.

 

[Genx87]

The server market and handset markets are different imo. I view Intel as the ARM of the server market. They were a high volume supplier competing with low volume high margin suppliers. They turned the server market into a commodity. I think going into the handset market vs ARM is a different scenario. Now Intel is looking at being the high margin supplier vs lower margin ARM suppliers. The handset market is already a commodity.

 

[Idontcare]

The server market and handset markets are different imo. I view Intel as the ARM of the server market. They were a high volume supplier competing with low volume high margin suppliers. They turned the server market into a commodity. I think going into the handset market vs ARM is a different scenario. Now Intel is looking at being the high margin supplier vs lower margin ARM suppliers. The handset market is already a commodity.

That was true 20yrs ago when Intel was upending the existing big-iron server market with their fledgling x86 chips, margins were less of a requirement to Intel as they were merely hoping to establish themselves as a competent hardware supplier in that market segment.

Now Intel has firmly placed the shoe on their other foot and critically depends on those juicy server segment gross margins to subsidize the intrinsically lower margin businesses that they are attempting to diversify into...but they will not push into any segment for which they do not foresee a way to someday arrive at a compelling gross margin story at some point.

The last time they pushed into the handset market they discovered a dismal gross margin situation and so they abandoned the effort. Same thing happened with their push into HDTV and discrete graphics cards.

They've come back around to the handset market once again but only because they foresee a path to extracting 60% margins from the smartphone segment. If they lose confidence in their ability to get to that gross margin level within a reasonable timeframe (2-3 yrs) then they will abandon it yet again, you can count on them to be calculating like that.

 

[Nemesis 1]

I doubt Intel would come to have a monopoly in the handset market because their business model ties their hands to only go after the premium portion of the TAM which yields high gross margins, they don't want to touch that bottom 25% of the TAM where margins are paltry.

So long as other companies are willing to live with 30-40% gross margins and near break-even profits, those companies will have a place to sell their chips because Intel has no interest in competing when it comes to those kinds of economics.

Intel could end up with an effective monopoly if everyone else basically walks away from the market...kinda like what has happened in the server segment...because that essentially turns the entire TAM into a margin-rich revenue opportunity where Intel has complete control over the supply/demand pricing curve.

Good reasoning . Except intel kinda threw a wrench into that sort of thinking with the Lexington Phone line for 3rd world . Its seemms intel wamts all the chips from top to bottom

 

[Axon]

I doubt intel doesn't have a team hard at work with the intent of taking over the ARM game.

 

[Idontcare]

Good reasoning . Except intel kinda threw a wrench into that sort of thinking with the Lexington Phone line for 3rd world . Its seemms intel wamts all the chips from top to bottom

(1) You/we don't know the margins on those products, they may well be acceptable even as low-ASP SKUs.

(2) Intel has no issue in taking on low-margins in the short-term in exchange for gaining critical mass provided they have confidence that someday in the mid-term that critical mass can be leveraged into a compelling gross margin story.

Look at their so-called "Intel Retail Edge" program where they are selling 3770K's for $105. You can't tell me the margins on those chips are 60% at that price point. They are probably closer to 10%.

So why would Intel offer them at that pricepoint in a mature market segment like the desktop segment? Its basically a marketing maneuver, ensuring the retail channel is seeded with plenty of happy-to-be-an-Intel-owner sales guys on the floor at Office Depot and so forth.

Why is Intel getting their name out there in the 3rd world portion of the industry?

Brand recognition for starters. A certain percentage of today's value-segment buyers are going to improve their income and become tomorrow's smartphone upgrader.

It is also about building confidence in one's competence. They need to build confidence within the handset manufacturer industry, Nokia and the like need to have confidence that Intel is competent when it comes to the technology (SoCs) used in making, selling and supporting handsets.