The ARM v.s. Intel Thing - Let's Discuss It

[Arachnotronic]

I constantly hear from folks that "Intel can't build mobile chips" because of some inherent flaw in X86. Does anybody else here find that supremely annoying?

Am I the only one who thinks that Intel's only "behind" because they haven't really (up until now, anyway) decided to focus on that segment of the market? Further, am I the only one who seems to think that when Intel leverages its process lead in mobile and finally releases a new microarchitecture (the company just announced its 22nm SoC process), the other vendors don't really stand a chance? In fact, the mere notion that the old-and-crappy "Bonnell" microarchitecture is competitive with the Krait clock-for-clock and core-for-core should have gotten everyone's attention.

Maybe I'm just used to talking to investor types who generally don't know too much about technology and think Intel will be bankrupt within 5 years.

Thoughts?

 

[Yuriman]

Intel doesn't want the kind of margins other chipmakers are getting. They want to be able to build a competitive chip for half the cost, or be able to charge a significant premium for phones with Intel chips inside.

 

[ShintaiDK]

Are you surprised after how it is with everything the last 5-10 years? Remember when Core 2 was "overclocked Opterons" for example?

ARM makers are already canibalizing in their own ranks. And its not funny when the gorilla gets into the chimps camp to wreck further havok.

 

[Genx87]

I find it interesting both x86 manufactuers on that list saw decrease in revenues. x86 devices sold 2.3 billion less in 2012. Qualcomm is kicking major ass.

 

[Fjodor2001]

On the other hand the ARM stock price has been going very well the last 5 years. Up more than 600% during that period despite that we have gone through a financial crisis, see:

http://www.google.com/finance?chdnp...edLine&q=LON:ARM&ntsp=0&ei=7lHbUOC4NoOPwAPzZw

The ARM stock also has a P/E of 70.94 as of 24 December 2012. That is very high, indicating that the market believes ARM to have a bright future ahead.

For comparison, Intel only has a P/E of 9.04.

So things may change in the years to come...

 

[Arachnotronic]

On the other hand the ARM stock price has been going very well the last 5 years. Up more than 600% during that period despite that we have gone through a financial crisis, see:

http://www.google.com/finance?chdnp...edLine&q=LON:ARM&ntsp=0&ei=7lHbUOC4NoOPwAPzZw

The ARM stock also has a P/E of 70.94 as of 24 December 2012. That is very high, indicating that the market believes ARM to have a bright future ahead.

So things may change in the years to come...

It's a hype-bubble. Do you really think that a company that has net income of ~$200M is really worth $17.5B?

Intel is "worth" $103B with net income of ~$13.5B.

 

[Idontcare]

I constantly hear from folks that "Intel can't build mobile chips" because of some inherent flaw in X86.

We'll never know the truth because we aren't in any of the rooms where the big decisions are being made, but we can guess as to the topics they will have had to wrestle with in the journey to get to where they are at.

Is x86 superior to ARM when designed to be manufactured on the same process node? Or does x86's advantage merely a masked disadvantage that has been more than over-compensated by billions spent on developing a process node that is far superior to that available at the foundry for other ARM designers?

And what of the design itself? Is x86 superior to ARM because it has benefited from decades of outsized R&D budgets to enhance, optimize, and tweak the design for maximal benefit on any given node - at the expense of billions upon billions in R&D bucks - whereas ARM is more like a scattered collection of nomads who wander the wilderness using stone tools to eek out a living year over year?

If Intel threw as many dollars into ARM as they have invested into their existing mobile x86 effort how would such an ARM chip fare against the current x86 offerings? If Nvidia had access to Intel's leading edge process tech would Tegra 3 be a much more potent competitor?

We can't hope to ever know the answers to these questions. And as consumers, investors, competitors, or OEMs, the answers to these questions are irrelevant anyways.

All that is relevant is what reality has made available, sans the resource and investment normalization efforts which are at best topics of relevance for the academically inclined.

Is there an "x86 tax"? Perhaps, or perhaps there is an "ARM tax" instead.

 

[frozentundra123456]

On the other hand the ARM stock price has been going very well the last 5 years. Up more than 600% during that period despite that we have gone through a financial crisis, see:

http://www.google.com/finance?chdnp...edLine&q=LON:ARM&ntsp=0&ei=7lHbUOC4NoOPwAPzZw

The ARM stock also has a P/E of 70.94 as of 24 December 2012. That is very high, indicating that the market believes ARM to have a bright future ahead.

So things may change in the years to come...

AMD had a very high P/E in its heyday as did RIM and a bunch of other tech stocks. I am not saying something like that will happen to ARM, but a high P/E can indicate a product with great growth prospects or something grossly overvalued, or possibly elements of both. Eventually a company has to make money, it cant go up on future expectations forever.

ARM is in the growth spurt now, and partly because each new generation of product improves on the previous one, while x86 is pretty much "good enough" for several years. Remains to be seen how the ARM market will shake out as the consumer market becomes more mature and they have to move into server/enterprise markets.

 

[ShintaiDK]

I find it interesting both x86 manufactuers on that list saw decrease in revenues. x86 devices sold 2.3 billion less in 2012. Qualcomm is kicking major ass.

ARM went down as well as a combined segment. There was just more dead corpses to feed on.

 

[Arachnotronic]

I find it interesting both x86 manufactuers on that list saw decrease in revenues. x86 devices sold 2.3 billion less in 2012. Qualcomm is kicking major ass.

Qualcomm also sells a boatload modems, so keep that in mind. It's not all $20 Snapdragons.

 

[Fox5]

We'll never know the truth because we aren't in any of the rooms where the big decisions are being made, but we can guess as to the topics they will have had to wrestle with in the journey to get to where they are at.

Is x86 superior to ARM when designed to be manufactured on the same process node? Or does x86's advantage merely a masked disadvantage that has been more than over-compensated by billions spent on developing a process node that is far superior to that available at the foundry for other ARM designers?

And what of the design itself? Is x86 superior to ARM because it has benefited from decades of outsized R&D budgets to enhance, optimize, and tweak the design for maximal benefit on any given node - at the expense of billions upon billions in R&D bucks - whereas ARM is more like a scattered collection of nomads who wander the wilderness using stone tools to eek out a living year over year?

If Intel threw as many dollars into ARM as they have invested into their existing mobile x86 effort how would such an ARM chip fare against the current x86 offerings? If Nvidia had access to Intel's leading edge process tech would Tegra 3 be a much more potent competitor?

We can't hope to ever know the answers to these questions. And as consumers, investors, competitors, or OEMs, the answers to these questions are irrelevant anyways.

All that is relevant is what reality has made available, sans the resource and investment normalization efforts which are at best topics of relevance for the academically inclined.

Is there an "x86 tax"? Perhaps, or perhaps there is an "ARM tax" instead.

I would think the only real tax for x86 is in backwards compatibility. Up until now, all X86 have had to run all prior x86 instructions, but performance could (and has been) degraded on instructions as the internal design of the chips has changed. Now that Intel's Atom chips for phone no longer require backwards compatibility, Intel is just as free as ARM to do things that break backwards compatibility if it provides an advantage.

X86 comes from a world not too dissimilar from embedded systems, where memory was slow and limited. The x86 instruction set is very compact, with implicit operands and the variable length decoding means it makes better use of memory bandwidth and cache size than just about any other cpu architecture out there, both of which are big wins for an embedded system. The decoding is probably comparatively expensive in terms of power usage, since that was never a limitation of when the ISA was originally designed, but even still just about every other aspect of it has evolved significantly over the years.

The comparable ARM trait to achieve the same memory/bandwidth efficiency is Thumb mode, which involves a context switch to make the processor behave as if it were a 16 bit processor (and I think 8 bit may be an option as well). I'd imagine there's some penalty to making the switch to thumb mode and back, so it could be more difficult for ARM compilers to output code that maximizes memory efficiency while also getting the job done in the minimum compute time. Still, it's a pretty neat trick and ARM achieves code density within ~25% of x86 (with x86-64 being a bit better than x86, due to additional registers available despite larger pointer sizes), without the penalty of an expensive decode step, but probably pays for it with an expensive context switch.

 

[Stoneburner]

Why doesn't intel do what other company's do when they want to protect their cachet but still explore budget markets? SPecifically, create a shell sub-entity. To humor us, they should call it "Larrabee Enterprises".

 

[cbn]

Am I the only one who thinks that Intel's only "behind" because they haven't really (up until now, anyway) decided to focus on that segment of the market? Further, am I the only one who seems to think that when Intel leverages its process lead in mobile and finally releases a new microarchitecture (the company just announced its 22nm SoC process), the other vendors don't really stand a chance? In fact, the mere notion that the old-and-crappy "Bonnell" microarchitecture is competitive with the Krait clock-for-clock and core-for-core should have gotten everyone's attention.

One question I have:

At what point does Intel put the core processors on the low power/leakage "SOC" process? (As currently stands even the ULV ultrabook chips are on the same process as the high power desktop, just binned differently)

A dual core "core processor" on the leading edge node, but with low leakage xtors, would be so much more interesting than a quad core atom chip for tablets.

 

[lopri]

I constantly hear from folks that "Intel can't build mobile chips" because of some inherent flaw in X86.

Please provide such an assertion from [semi-]credible source because I have not encountered one yet. I constantly hear from folks that 'Intel CAN build mobile chips!' because of x86 power inefficiency is a myth. See: The x86 Power Myth Busted: In-Depth Clover Trail Power Analysis (Quite reminiscent to the article - which was taken down voluntarily - that attacked Sony and MS for not using Intel CPUs for PS3 and XBox360)

Am I the only one who thinks that Intel's only "behind" because they haven't really (up until now, anyway) decided to focus on that segment of the market? Further, am I the only one who seems to think that when Intel leverages its process lead in mobile and finally releases a new microarchitecture (the company just announced its 22nm SoC process), the other vendors don't really stand a chance?

No, there are hordes of financially (or otherwise) motivated Intel "supporters" as you can see in this very thread. Usual names, with harmonious opinions. It's hilarious. Nothing much has changed here, which is a good thing to be reminded of. ^_^

On the other thread, I saw your pointless attack upon AMDZone. I thought it's quite akin to a Britney Spears fan posting at Britney Spears fan forum, claiming how delusional Rhiana's fans are. "Rhiana's ass is uglier than Britney's, those Rhiana fans are warped into an alternate reality!"

In case you haven't realized, there are many delusional people on the Internet. You see, go to any fan-oriented sites. Be that sports, singers/actors, games, energy drinks, cosmetics, and indeed, computers and gadgets. What's unfortunate is that these days it becomes harder to discern delusions from other nefarious motivations. (Money = Speech, SCOTUS agrees)

When I went to AMDZone a couple of years ago because I could not learn anything about my 1090T here at AT Forums, I did not see many delusional people there, no more than I see those elsewhere. And I owe half my knowledge of K10 to some very intelligent folks at AMDZone. I could actually learn something there, unlike somewhere else.. where people have no interest in tech, but in corporations' well-being and their Internet creeds.

Anyway, back on topic:

Maybe I'm just used to talking to investor types who generally don't know too much about technology and think Intel will be bankrupt within 5 years.

I think you are onto something here. But as far as I can see, it's Intel propaganda machine screaming how x86 is not disadvantaged to ARM in mobile space, not the other way around. (see the article linked above again) I haven't heard OEMs that sell ARM devices taking time to claim how shitty Intel is in press conferences. :biggrin: They are, as far as I can tell, too busy making new stuff.

Happy new year, everyone.

 

[Chiropteran]

I feel as though part of the "flaw" in Intel chips is the licensing. I get the impression that cell phone manufacturers like to have control over what they produce, and with ARM licensing they can do that, while being reliant on Intel for CPU, no matter how good Intel's supply is, isn't the same thing.

This is just the impression I get, I don't really have anything to back it up, just a blind opinion.

 

[ShintaiDK]

I feel as though part of the "flaw" in Intel chips is the licensing. I get the impression that cell phone manufacturers like to have control over what they produce, and with ARM licensing they can do that, while being reliant on Intel for CPU, no matter how good Intel's supply is, isn't the same thing.

This is just the impression I get, I don't really have anything to back it up, just a blind opinion.

Besides Samsung and Apple. I dont think any of the rest got any control, they could get exactly the same as they get today from Intel so to speak.

 

[jvroig]

If it is "ARM vs Intel" within the context of "who will end up dominating the smartphone/tablet (non-PC, so no laptops) market" (and foregoing all technical "which arch is better" discussions), the only real limitations I have in mind are those of Intel's own making.

Bluntly, they are not approaching every fight with both hands untied and ready to do "whatever it takes". We can imagine that with the tech at their disposal and the billions in budget, they can pretty much kick ass in every related field. And this would be very true, if Intel had no reservations about dominating every related field.

Unfortunately, they are hamstrung by the gross margin requirements. It is completely within the boundaries of tech-feasibility and market-feasibility for Intel to produce a chip that will go into every smartphone/tablet, be superior in all tech metrics, and be cheaper than those of any competitor. Some of us would think of that as an absolute good thing for Intel, an obvious course of action. But for Intel, if it would end up severely damaging their their accustomed >60% gross margins, it simply wouldn't be worth the investment. Their history shows this. The board wants it. CEO's job depend on it. It is not a cycle that someone in Intel can just break by saying "Screw the gross margins! We must dominate smartphones no matter what!".

This is the short and quick of it. If the ARM / smartphone / tablet makers make it an industry with margins at 30% (completely pulled out of air, just for the sake of exposition), Intel may simply leave the market, as long as they feel that the company is in no danger of extinction, if it ends up affecting their overall gross margins significantly in a negative way. If it were not for this particular limitation that has Intel quite literally fighting with one hand tied behind its back, I don't see how ARM (the design company, not the arch) can actually compete against Intel if Intel wanted to take ARM's cake.

 

[cbn]

Unfortunately, they are hamstrung by the gross margin requirements. It is completely within the boundaries of tech-feasibility and market-feasibility for Intel to produce a chip that will go into every smartphone/tablet, be superior in all tech metrics, and be cheaper than those of any competitor. Some of us would think of that as an absolute good thing for Intel, an obvious course of action. But for Intel, if it would end up severely damaging their their accustomed >60% gross margins, it simply wouldn't be worth the investment. Their history shows this. The board wants it. CEO's jobs depend on it. It is not a cycle that someone in Intel can just break by saying "Screw the gross margins! We must dominate smartphones no matter what!". This is the short and quick of it. If the ARM / smartphone / tablet makers make it an industry with margins at 30% (completely pulled out of air, just for the sake of exposition), Intel may simply leave the market, as long as they feel that the company is in no danger of extinction, if it ends up affecting their overall gross margins significantly in a negative way. If it were not for this particular limitation that has Intel quite literally fighting with one hand tied behind its back, I don't see how ARM (the design company, not the arch) can actually compete against Intel if Intel wanted to take ARM's cake.

Margins for chips going into smartphones can be quite good:

http://ycharts.com/companies/QCOM/c...l&recessions=false&zoom=5&startDate=&endDate=

 

[Arachnotronic]

Computer Bottleneck,

The majority of Qualcomm's operating profit is from patent licensing revenue. And then a good part of the chips it sells isn't Snapdragons but, in fact, modems. This significantly helps Qualcomm's gross margin profile.

 

[jpiniero]

Intel may simply leave the market, as long as they feel that the company is in no danger of extinction

Extinction isn't the word I would use, but I don't think Intel would be doing what they are doing now with Atom unless they thought there was a real threat to a large chunk of their revenue.

 

[Arachnotronic]

Extinction isn't the word I would use, but I don't think Intel would be doing what they are doing now with Atom unless they thought there was a real threat to a large chunk of their revenue.

Agreed. Intel cannot afford to have an inferior product at the low power end of things. With the growth happening in Windows RT/Android, there's a lot of room for lower end market share gains for the ARM chips. The Qualcomm Snapdragon S4, the Nvidia Tegra, and Samsung Exynos all are proving to be quite viable in tablets and even low-end clamshells.

Putting out crap Atoms is a bad, bad plan.

 

[Arachnotronic]

Please provide such an assertion from [semi-]credible source because I have not encountered one yet. I constantly hear from folks that 'Intel CAN build mobile chips!' because of x86 power inefficiency is a myth. See: The x86 Power Myth Busted: In-Depth Clover Trail Power Analysis (Quite reminiscent to the article - which was taken down voluntarily - that attacked Sony and MS for not using Intel CPUs for PS3 and XBox360)



No, there are hordes of financially (or otherwise) motivated Intel "supporters" as you can see in this very thread. Usual names, with harmonious opinions. It's hilarious. Nothing much has changed here, which is a good thing to be reminded of. ^_^

On the other thread, I saw your pointless attack upon AMDZone. I thought it's quite akin to a Britney Spears fan posting at Britney Spears fan forum, claiming how delusional Rhiana's fans are. "Rhiana's ass is uglier than Britney's, those Rhiana fans are warped into an alternate reality!"

In case you haven't realized, there are many delusional people on the Internet. You see, go to any fan-oriented sites. Be that sports, singers/actors, games, energy drinks, cosmetics, and indeed, computers and gadgets. What's unfortunate is that these days it becomes harder to discern delusions from other nefarious motivations. (Money = Speech, SCOTUS agrees)

When I went to AMDZone a couple of years ago because I could not learn anything about my 1090T here at AT Forums, I did not see many delusional people there, no more than I see those elsewhere. And I owe half my knowledge of K10 to some very intelligent folks at AMDZone. I could actually learn something there, unlike somewhere else.. where people have no interest in tech, but in corporations' well-being and their Internet creeds.

Anyway, back on topic:



I think you are onto something here. But as far as I can see, it's Intel propaganda machine screaming how x86 is not disadvantaged to ARM in mobile space, not the other way around. (see the article linked above again) I haven't heard OEMs that sell ARM devices taking time to claim how shitty Intel is in press conferences. :biggrin: They are, as far as I can tell, too busy making new stuff.

Happy new year, everyone.

I wasn't attacking AMDZone, I was simply pointing out how...interesting it was over there. Maybe I just don't really understand blind fans. I own quite a few AMD products because they were the best I could get for the money (Llano is a pretty great laptop processor, especially in a $400 machine, and my Athlon II X4 635 is a champ), and I recommend AMD processors to friends and family where it makes sense. But the claims that the 3770K or 3930K aren't, in the majority of cases, better than anything AMD's got are just...well, I can't understand it.

 

[Acanthus]

It's a hype-bubble. Do you really think that a company that has net income of ~$200M is really worth $17.5B?

Intel is "worth" $103B with net income of ~$13.5B.

Not to mention Intel has billions and billions of assets in fabs, which is real, tangible, worth.

I love ARM, but the sleeping dragon has awakened, and they will meet the same fate as AMD.

They will be playing a distant second to Intel in the long run.

Intel just has too much talent and resources to push behind any market they want to take over. The fab superiority alone will be enough to tip things in their favor.

 

[Vesku]

I think ARM has some pretty good business allies though. Google, Apple, Qualcomm, Samsung it's not like Intel is pushing into virgin territory.

IMO, Intel hurt themselves quite a bit relaunching various mobile Linux projects while Android and iOS were grabbing tons of developer mindshare - Moblin > Meego > Tizen.

 

[IntelUser2000]

I think you are onto something here. But as far as I can see, it's Intel propaganda machine screaming how x86 is not disadvantaged to ARM in mobile space, not the other way around. (see the article linked above again) I haven't heard OEMs that sell ARM devices taking time to claim how shitty Intel is in press conferences. :biggrin: They are, as far as I can tell, too busy making new stuff.

Happy new year, everyone.

Happy new year and holidays!

Sorry lopri, but I don't think you are doing anything different from the very people you criticize. There are always two different sides to a discussion(or an argument), and for those who are completely outside, both look equally delusional.

It's not until one side is "proven" to be right that opinion changes. And when that happens, the opposite side is ridiculed and treated like psychos.

Intel claimed that they aren't disadvantaged compared to ARM. What they never claimed was they aren't disadvantaged when everything is equal.

More than a decade ago that they proved for high end servers ISA differences are minor in the big picture of things. Process shrinks and increasingly complicated MPU designs have relegated ISA differences to be negligible.*

(*And there are far more important factors like execution. The ability for the specific company to drive projects the way they want it)

While we are talking about small chips like Atom today, the fact is no doubt still true. We went from having a "CPU-focus" to a "Platform-focus", with integration being greater than ever! GPUs, North AND Southbridges, even the communication chips! The ISA is again looking like the smaller part in the big picture. Look how AMD and Nvidia achieve similar results on the GPU side with vastly different architectures. SIMD vs VLIW. Single issue vs Multi issue.

Of course it does not matter if the chip is superior if the rest of the device is not. I am not saying their success is guaranteed, but I think most people(including you) are overstating ISA differences when purely comparing at CPU vs CPU.