CCI completes probe into Intel malpractice [The Times of India]

[SiliconWars]

This is true to a certain extent, but the wildcard is x86. It matters to a lot of people.

~10 billion ARM devices are sold every year compared to ~300 million x86 devices. x86 becomes more and more irrelevant with every passing year. Even if Intel could take 10% of that 10 billion device market it would be an incredibly huge undertaking. They are trying to move up into a scale that is unprecedented.

It's much more likely that ARM will take over that 300 million x86 market, or at least 200 million of it, on price.

 

[frozentundra123456]

No because the $1.25 billion Intel paid AMD in the 2009 settlement brought an end to all outstanding legal disputes between them.

I believe he was referring to the new link posted in this thread, not the settlement with amd. That is old news, people need to move on.

 

[SOFTengCOMPelec]

This is true to a certain extent, but the wildcard is x86. It matters to a lot of people.

The problem is (for Intel and x86), the market place is changing very rapidly, and all sorts of new markets have opened up.

How many small hand held devices, phones, tablets, iphones, ipads, SmartTV's, massive embedded electronics cpu's, HDD (with controller cpu's), cars with huge numbers of embedded cpu's, lots and lots of other things, have got or use x86 ?

x86 kind of = Microsoft Windows

I think x86 is gradually being phased out, a bit like vinyl records, audio tapes, video tapes and many other things. x86 = desktops = disappearing.

Sure, the widespread use of x86 for such a long time, means it will still be available for ages yet. But, how long can it continue being the mainstream computing platform ?

 

[Charles Kozierok]

~10 billion ARM devices are sold every year compared to ~300 million x86 devices. x86 becomes more and more irrelevant with every passing year.

You're completely missing the point of why x86 is attractive and remains so. It has to do with the benefits to business of having unified platforms; they are related but distinct markets, and x86 isn't going away because of how many ARM toys are sold every year.

Intel can make x86 do what ARM does. I am not seeing any indication that ARM can move towards what x86 can do.

 

[Khato]

~10 billion ARM devices are sold every year compared to ~300 million x86 devices. x86 becomes more and more irrelevant with every passing year. Even if Intel could take 10% of that 10 billion device market it would be an incredibly huge undertaking. They are trying to move up into a scale that is unprecedented.

It's much more likely that ARM will take over that 300 million x86 market, or at least 200 million of it, on price.

According to ARM's own page there are "more than 30 billion processors sold with more than 16M sold every day ARM". Even rounding that 16M number up to 17M only arrives at 6.2 billion ARM processors sold per year. Other interesting point being that according to Wikipedia's article on ARM Architecture they shipped 6.1 billion ARM-based processors back in 2010. Not exactly growing fast are they?

I wonder how many billions of those ARM-based processors are slower than a 486? A Pentium?

Edit: Very nice - Anand's ARM article actually specifies some of these numbers.

 

[aigomorla]

India has a GDP of $2 trillion USD. This case will probably go on for 5 years or more, and Intel will probably be fined less than $50 million USD.

In terms of "making India money", it's about the same as changing to a slightly less luxurious make of toilet roll.

and how much of that GDP is based on people using intel computers?
and how much of an impact would it take if they boycotted Intel.

and less then 50 million you say?
i think at least 10 times that amount sir... at least... 500 million.

 

[frozentundra123456]

You're completely missing the point of why x86 is attractive and remains so. It has to do with the benefits to business of having unified platforms; they are related but distinct markets, and x86 isn't going away because of how many ARM toys are sold every year.

Intel can make x86 do what ARM does. I am not seeing any indication that ARM can move towards what x86 can do.

You said it well. Those posters who want to push ARM dont consider the business, education, research markets. Sure consumers love their Android toys, and they do have their place, but their is still a huge market for x86 computing.

I work in a research environment and we are doing a large clinical study. We schedule our patients for visits using excel and tabulate progress in access data base. Even the chromatography system I use to analyze the samples and the data handling of the results is done by software running on x86. Every lab I am familiar with in our building has at least 3 or 4 computers, and none of them use anything running Android. So I dont see a place for ARM in this kind of environment at all.

 

[SOFTengCOMPelec]

Deleted

 

[SOFTengCOMPelec]

You said it well. Those posters who want to push ARM dont consider the business, education, research markets. Sure consumers love their Android toys, and they do have their place, but their is still a huge market for x86 computing.

I work in a research environment and we are doing a large clinical study. We schedule our patients for visits using excel and tabulate progress in access data base. Even the chromatography system I use to analyze the samples and the data handling of the results is done by software running on x86. Every lab I am familiar with in our building has at least 3 or 4 computers, and none of them use anything running Android. So I dont see a place for ARM in this kind of environment at all.

Good point!

Yes, I see what you (and other) posters mean.

The x86 platform, is still the "main" serious use PC 'environment', and would take a long time to replace, as even now, there are not much in the way of "Arm" desktops, especially NOT powerful ones, anyway.

It's going to be very interesting, watching what happens in these various market places.

 

[Torn Mind]

You said it well. Those posters who want to push ARM dont consider the business, education, research markets. Sure consumers love their Android toys, and they do have their place, but their is still a huge market for x86 computing.

I work in a research environment and we are doing a large clinical study. We schedule our patients for visits using excel and tabulate progress in access data base. Even the chromatography system I use to analyze the samples and the data handling of the results is done by software running on x86. Every lab I am familiar with in our building has at least 3 or 4 computers, and none of them use anything running Android. So I dont see a place for ARM in this kind of environment at all.

Software like Matlab, Stata, SAS, etc are not ARM-friendly because they need to do a lot of math, and quickly. I don't think Red Hat or Windows 7 Enterprise would play nice on an ARM CPU performance-wise.

 

[ShintaiDK]

~10 billion ARM devices are sold every year compared to ~300 million x86 devices. x86 becomes more and more irrelevant with every passing year. Even if Intel could take 10% of that 10 billion device market it would be an incredibly huge undertaking. They are trying to move up into a scale that is unprecedented.

It's much more likely that ARM will take over that 300 million x86 market, or at least 200 million of it, on price.

As already stated your ARM numbers are overrated and your x86 numbers are underrated as well. Lets look at revenue shall we? x86 commands about 70% of the entire MPU revenue.

As it looks now, x86 will take all higher revenue parts away from ARM. Its x86 gaining on ARM, not ARM gaining on x86.

 

[SOFTengCOMPelec]

As already stated your ARM numbers are overrated and your x86 numbers are underrated as well. Lets look at revenue shall we? x86 commands about 70% of the entire MPU revenue.

As it looks now, x86 will take all higher revenue parts away from ARM. Its x86 gaining on ARM, not ARM gaining on x86.

Can I check my understanding of what you are saying ?

Analogy.
If we go back a number of years, there was a time when Vinyl Records had 70% of the market, and CD's had 30% of the market.
So, from what you are saying, everyone should have Vinyl Records, and CD's should be obsolete now (since 70% is > 30%) ?

There are a number of signs that Arm stuff are gaining and x86 are declining, e.g. Disappearing desktops

Anyway, what exactly has the 70% .. 30% revenue got to do with anything ?
Since it will change in coming years, as Arm sales increase (in value) and x86 gradually declines (SOURCE: = My opinion).

EDIT: But it is not clear who is going to win, because of things like Silvermont, Intels significant process advantage, the inertia and resiliance of x86 and many other factors.
So, we may well have to all wait and see what happens.

 

[ShintaiDK]

Can I check my understanding of what you are saying ?

Analogy.
If we go back a number of years, there was a time when Vinyl Records had 70% of the market, and CD's had 30% of the market.
So, from what you are saying, everyone should have Vinyl Records, and CD's should be obsolete now (since 70% is > 30%) ?

There are a number of signs that Arm stuff are gaining and x86 are declining, e.g. Disappearing desktops

Anyway, what exactly has the 70% .. 30% revenue got to do with anything ?
Since it will change in coming years, as Arm sales increase (in value) and x86 gradually declines (SOURCE: = My opinion).

EDIT: But it is not clear who is going to win, because of things like Silvermont, Intels significant process advantage, the inertia and resiliance of x86 and many other factors.
So, we may well have to all wait and see what happens.

ARM is not something new, its over 25 years old. And ARM never went anywhere outside its segments. And money got everything to do with it. Without money you cant do anything.

When you remove LTE modems from Qualcomm, its an entirely different picture of that company revenue wise. Samsung will also drop down to around 1B$ revenue after losing Apple.

 

[SOFTengCOMPelec]

ARM is not something new, its over 25 years old.

I agree with you, I've actually got a "PC" or similar, from or approaching that time, with a fairly early Arm processor.
But those early processors, are quite a different story (but similar as well), from the "modern" Arm processors, because the company had lots of changes over the years.

And ARM never went anywhere outside its segments.

True.
They were hoping to move into 'new' PC desktop sales, when Windows 8 (Arm version) came out, but that has probably not turned out too well (my own opinion).

And money got everything to do with it. Without money you cant do anything.

Money has got lots to do with it, if you don't believe me, give me all your money as an experiment, and we can see if it makes any difference.

But the thing is, although Intel is doing well (money-wise) today, things may be completely different in 2, 5, or 10 years time.

If I had a magical time machine, and time-jumped to 2023, right next to a computer shop, rushed into the store, and bought a computer, what processor would be at the heart of it ?

Intel (Possibly), Arm based (Possibly), AMD (possible, but not the 1st choice that comes to mind), something else (maybe, but not necessarily likely).

I'm actually torn between Intel and Arm, because some signs say Intel are going to win, and others say Arm, so I'm really not sure.

Arm is sort of climbing up from the bottom, and Intel is sort of reaching down from the top, because Arm holds much of the bottom end of the market (hand held devices, embedded etc), and Intel (and to an extent AMD) holds the top, with PC's and the server market.
In the middle ground, all of Intel, AMD and Arm have their sales successes.

So, I really DON'T know, who will win out.

 

[SiliconWars]

8.7 billion ARM devices were sold - not shipped, sold - in 2012

phx.corporate-ir.net/External.File?item=UGFyZW50SUQ9MTczODc5fENoaWxkSUQ9LTF8VHlwZT0z&t=1

That will be ~10 billion this year.

 

[MisterMac]

There's a reason intel is dropping Desktop line focus(hell even laptop\ultrabook slowly).


If everything becomes thin client based - who do you think will power the cloud\DC infrastructure ....ARM?


It seems to Intel is more than happy to let 2-3 desktop sales go to an expensive 2k-4k USD server equivelant.

(+ all the extra revenue from chipsets\connectors\NIC business'es).

 

[jvroig]

If everything becomes thin client based - who do you think will power the cloud\DC infrastructure ....ARM?

Yes, there will always be workstations and servers using high-powered Xeons/Opterons and even high-powered compute cards (Teslas/Quadros/Firepros), so that market for Intel/AMD/NV will remain, but that's not to say they will corner it completely. There will be a small niche for microservers too, which ARM can happily fulfill, being targeted by Calxeda and Seamicro/AMD. Powering the web (web-serving), for example, is a perfect fit for microservers while leaving database crunching for the bigger servers.

It seems to Intel is more than happy to let 2-3 desktop sales go to an expensive 2k-4k USD server equivelant.

The ratio certainly isn't 2-3 desktops : 1 server. Actual figures depend on what the server is supposed to do. For LTSP-like arrangements, you can have 10+ terminals/thin-clients connecting to a single powerful server. This would be the low-end of the scale for estimating how many desktops to trade against a server. To support computing as it is now for the masses / non-enthusiasts / non-content-creators (pure consumption; Facebook/Twitter/Instagram/IM/email+general web browsing), you can have one powerful server serve a couple hundred users at least, so that's quite a few more desktops that don't need to be bought. As desktops/laptops last longer and longer before needing to be replaced (more performance than most consumers really need), and as more and more devices get "PC-like" capabilities (smartphones, tablets, even consoles), desktop demand will be negatively affected and Intel cannot count on desktop sales&growth to hit its gross margin targets.

So Intel isn't dropping desktop because it can recoup everything through increased server demands. Rather, Intel is slowly changing focus from desktops to the next wave of computing for the masses: mobile devices, particularly non-laptop devices. This is what needs to be done, because this market, unlike the desktop market, is expected to keep growing. This is why Otellini regretted the Apple decision (i.e., not jumping on Apple's iphone when given the chance). Intel's fixing it now, in full force, and Servers + Mobile is the way to the future. Desktops won't be left completely in the cold, because it can always use hand-me-downs from server tech, and also benefit from scaled-up laptop parts (and if the atom celeron/atom pentium rumor(?) is true, then perhaps even scaled-up mobile parts). Desktops just won't be a main focus, because focusing on desktops is a losing proposition.

We enthusiasts will most likely always have high-end parts to play with no matter what happens. The worst case is that they might just cost a bit more than we are used to as the TAM for our toys shrink.

 

[MisterMac]

Yes, there will always be workstations and servers using high-powered Xeons/Opterons and even high-powered compute cards (Teslas/Quadros/Firepros), so that market for Intel/AMD/NV will remain, but that's not to say they will corner it completely. There will be a small niche for microservers too, which ARM can happily fulfill, being targeted by Calxeda and Seamicro/AMD. Powering the web (web-serving), for example, is a perfect fit for microservers while leaving database crunching for the bigger servers.


The ratio certainly isn't 2-3 desktops : 1 server. Actual figures depend on what the server is supposed to do. For LTSP-like arrangements, you can have 10+ terminals/thin-clients connecting to a single powerful server. This would be the low-end of the scale for estimating how many desktops to trade against a server. To support computing as it is now for the masses / non-enthusiasts / non-content-creators (pure consumption; Facebook/Twitter/Instagram/IM/email+general web browsing), you can have one powerful server serve a couple hundred users at least, so that's quite a few more desktops that don't need to be bought. As desktops/laptops last longer and longer before needing to be replaced (more performance than most consumers really need), and as more and more devices get "PC-like" capabilities (smartphones, tablets, even consoles), desktop demand will be negatively affected and Intel cannot count on desktop sales&growth to hit its gross margin targets.

So Intel isn't dropping desktop because it can recoup everything through increased server demands. Rather, Intel is slowly changing focus from desktops to the next wave of computing for the masses: mobile devices, particularly non-laptop devices. This is what needs to be done, because this market, unlike the desktop market, is expected to keep growing. This is why Otellini regretted the Apple decision (i.e., not jumping on Apple's iphone when given the chance). Intel's fixing it now, in full force, and Servers + Mobile is the way to the future. Desktops won't be left completely in the cold, because it can always use hand-me-downs from server tech, and also benefit from scaled-up laptop parts (and if the atom celeron/atom pentium rumor(?) is true, then perhaps even scaled-up mobile parts). Desktops just won't be a main focus, because focusing on desktops is a losing proposition.

We enthusiasts will most likely always have high-end parts to play with no matter what happens. The worst case is that they might just cost a bit more than we are used to as the TAM for our toys shrink.

I didn't mean it directly.

But from a greater perspective in a 10 year range - a Server portfolio and a ATOM only portfolio, with less focus on mobile\desktop - it would make sense to see Intel trying to "segment" the market this way for margins.

 

[galego]

If that were true then you would need a single world-government and identical laws and rules enforced in every single country regardless of national borders.

The world does not operate the way in which you feel it ought to.

I.e. there is reality, and then there are your philosophical ideals.

The two currently do not overlap.

I don't know how and why you confound the real world, with its countries and laws, with an imaginary world without them.

Intel is being investigated by violating the laws of a particular country, playing dirty against the competence in that country.

I would have thought that in the real long term, companies may be harming themselves, by "playing dirty".

E.g. Maybe Intel would have been better geared up to fight "Arm", if they had NOT played dirty.
This is because they used "business tricks" (dirty ones at that), to win sales, rather than genuine excellent products and marketing.

E.g. If Atoms had been significantly better, faster, cheaper, easier to use etc etc, than 'Arms', the small hand held device market place, may have looked completely different to what it looks now.

EDIT: My explanation, is not very good, or clear. I was trying to illustrate what they could have done better, 10 years ago, which was when they really needed to get their foot in the door, as regards 'Arm'.
Now, Intel may well be trying to close the barn door after the Horse has escaped, as they are VERY late to the market. Time will tell.

Intel has a track record of playing dirty when cannot beat the competence in the lab or in the fabric. If Intel really knew how to design and fabricate products that were good enough to beat the competence, it would not need to cheat, lie... play dirty.

 

[SocketF]

You said it well. Those posters who want to push ARM dont consider the business, education, research markets. Sure consumers love their Android toys, and they do have their place, but their is still a huge market for x86 computing.

I work in a research environment and we are doing a large clinical study. We schedule our patients for visits using excel and tabulate progress in access data base. Even the chromatography system I use to analyze the samples and the data handling of the results is done by software running on x86. Every lab I am familiar with in our building has at least 3 or 4 computers, and none of them use anything running Android. So I dont see a place for ARM in this kind of environment at all.

Well you answered to " I am not seeing any indication that ARM can move towards what x86 can do."

But your case sounds just like the usual office stuff.. get Linux and Openoffice for ARM and then you are done.

Of course some kind of special software might be a critical problem. The market is small and the producers probably dont have time and money to make a Linux version. In that case it would be a problem. A bit comparable with MacOS-users. Sometimes they also cant find their favorite programs.

Anyways I am curiously waiting for the first 64bit ARM chips. THere will be several server-versions and I guess such SoCs would also fit nicely in Office PCs. Maybe they are already too much (nobody needs 8 cores, but if it is cheap ... ^^).

 

[SOFTengCOMPelec]

Intel has a track record of playing dirty when cannot beat the competence in the lab or in the fabric. If Intel really knew how to design and fabricate products that were good enough to beat the competence, it would not need to cheat, lie... play dirty.

Without knowing much more details, ideally from reference sources with minimum risk of being biased and/or involved enough to potentially invalidate their information. I would prefer to stick with the "innocent, until proven guiulty".

At the end of the day, Intel are a big thriving, international company, who have been around for a long time.

So, it is not necessarily too unreasonable, if they do something on the following lines :-

1. Get sales by making the best, fastest, cheapest, best value, most relaible, etc stuff
2. Sell their stuff, which is boadly similar to the competition, by using, advanced (legitimate) sales techniques
3. Secretly, internally know that their stuff is inferiour, but give their sales team a huge push and authorization to "GET THEIR STUFF SOLD" and beat the competition, including reducing the price as necessary

Ideally, they would/should be in (1), or even (2) sometimes.

But, for various real life practical reasons, they won't always be in that situation (e.g. Their latest process does not work out very well, due to completely unforeseen factors).

The answer is NOT for Intel to stand in a corner, start crying and waiting, until the company goes bankrupt, until all the employees lose their jobs.

They should attempt (3).
Differing companies, have different internal principles, as to how far down (3), they will go.
But, (unfortunately), it is a fact of life, that some companies will go TOO far down the (3) route.

Time and time again, I have heard that being "the nice guy", and running a business, are somewhat mutually exclusive.

 

[galego]

Without knowing much more details, ideally from reference sources with minimum risk of being biased and/or involved enough to potentially invalidate their information. I would prefer to stick with the "innocent, until proven guiulty".

At the end of the day, Intel are a big thriving, international company, who have been around for a long time.

As said there is a track record

http://en.wikipedia.org/wiki/Intel#Lawsuits

http://en.wikipedia.org/wiki/Intel#Anti-competitive_allegations

 

[blackened23]

As said there is a track record

http://en.wikipedia.org/wiki/Intel#Lawsuits

http://en.wikipedia.org/wiki/Intel#Anti-competitive_allegations

Pretty much any corporation the size of intel will have litigation to match. It's just the nature of the business, there is no corporation free of litigation, period. I'm willing to bet that even your AMD has a large share of litigation to deal with.

 

[SOFTengCOMPelec]

Pretty much any corporation the size of intel will have litigation to match. It's just the nature of the business, there is no corporation free of litigation, period. I'm willing to bet that even your AMD has a large share of litigation to deal with.

Well put!

As said there is a track record

http://en.wikipedia.org/wiki/Intel#Lawsuits

http://en.wikipedia.org/wiki/Intel#Anti-competitive_allegations

In my mind, they (Intel) are "Innocent, until proven guilty".

This seems to be (your OP quoted link) a new case against Intel, which HAS NOT BEEN PROVEN yet, so we will have to wait and see.

Are you trying to say that someone who committed a crime ten years ago (speeding, 35 MPH in 30 MPH zone), can be convicted of murder, with ABSOLUTELY no evidence, or court trial, or jury, and given the death penalty ?

Or, if we hear of such a trial, we can 100% assume THEY ARE GUILTY, even before the court case starts ?

 

[galego]

In my mind, they (Intel) are "Innocent, until proven guilty".

This seems to be (your OP quoted link) a new case against Intel, which HAS NOT BEEN PROVEN yet, so we will have to wait and see.

Are you trying to say that someone who committed a crime ten years ago (speeding, 35 MPH in 30 MPH zone), can be convicted of murder, with ABSOLUTELY no evidence, or court trial, or jury, and given the death penalty ?

Or, if we hear of such a trial, we can 100% assume THEY ARE GUILTY, even before the court case starts ?

Sorry I misunderstood you. I believed that with your "Innocent, until proven guilty" you were negating Intel history, as some others like to do.

I agree with you we must wait and see in this new case. Of course, Intel is innocent whereas the contrary is not proved.

Said that, history (previous links) shows that Intel has been playing dirty for decades in many countries, which means it is highly probable that this has been the case now in India.