x86 is history

[HeXen]

android? chrome os? ipad? heard of those? when you consider millions of units sold, windows IS going away.

thats why Windows 8 will also natively work for ARM too. So no, its not going away.

Yes, but where are consoles headed?

To basically be a gaming version of an HTPC. the center media hub for all your devices and entertainment needs.

 

[Seero]

Nvidia has gone off its rockers.... smart phones replaceing intire pc systems? maybe for web browseing? but for anything more demanding? like gameing... I mean theres a reason dedicated grafics cards have the size they do right? The 600 gram heavy air coolers ontop of our CPUs... Nvidia really want us to buy into this "smart phone" solution that ll replace our pcs?

Is gaming demanding? Or is the rendering of graphics of gaming that is demanding? Maybe a 16 cores CPU will rock your game, or maybe a dual core CPU with just a bit higher clock speed will beat a thousand dollars CPU in terms of gaming. What I can say for sure though is, if SB come without the GPU core which many of us here won't use, it can run at a higher clock speed. Maybe stripping all other craps bundled within the CPU will actually allow a bit more clock speed. What is ARM? Maybe it is a CPU that is without things that we don't need.

 

[SickBeast]

the only thing growing in the PC game market are Indie games. The rest are ported from consoles assuming they ever get ported...thats not really growing as it just keeping it afloat.
were already starting to see fewer desktops on store shelves and more lappy's with tablets getting ready to take up more shelf space.
by the time of next consoles release, depending on sales, there may not be as many ports, consoles will have all the AAA titles

As sad as it is, you could well be right.

That said, the PC does have a unique place in the gaming world. It's the only platform that you can completely customize to suit your particular needs. It also generally holds a performance advantage. For these reasons, IMO PC gaming is here to stay. Even if it moves primarily to tablets, regular old gaming PCs will run those games faster with more features like AA enabled, along with a higher resolution.

 

[Ben90]

What I can say for sure though is, if SB come without the GPU core which many of us here won't use, it can run at a higher clock speed.

The iGPU has nothing do with the x86 cores' clockspeed other than possibly lowering Turbo multipliers due to increased heat when using sub optimal cooling solutions.

 

[pw38]

android? chrome os? ipad? heard of those? when you consider millions of units sold, windows IS going away.

Well, I think it's more a case of the inevitable. If company A has 100% of the market and you introduce company B with 5% company A can never have 100% of the market again unless the competition goes away. Introduce company C with 10%, company D with 4%, etc. etc. and you can see where this is going. This doesn't speak to the declining dominance of Windows and more to the maturing of the market with competition finding ways to establish presence in an ever growing market with products the current version of Windows doesn't fit well into (hence the announcement of SoC Windows 8 support). I personally want to see how MS can adapt because the writing is clearly on the wall.

 

[Chaoticlusts]

Consoles aren't going away or being replaced with PC's...they never took off because they were more powerful or 'cheaper' or various other factors..they took off cause they were convenient...far less to deal with than a whole pc you plug it in throw a game in and the game runs...as for them not needing to upgrade in the next cycle..I'm willing to bet next gen consoles will launch with 4K resolution support since we're already seeing TV's demo'd that support it (TV's won't really hit retail till 2012-2013 but neither will next gen consoles) and just cause 4K won't be widely adopted then doesn't mean they won't support it (ah heck all people had 1080p TV's when the 360 launched)...the only way consoles won't be high end is if they switch to cloud run systems...which is possible but I doubt we'll see it happen until more countries pick up the 'broadband as a intrinsic human right rather than privilege' laws (yes this has happened) otherwise they'd shoot themselves in the foot in numerous countries

As for PC gaming being primarily indep/facebook/ports...your forgetting one very very important market....MMO's..so far very few have had real success on consoles and the main PC only products continue to rake in massive revenue...PC gaming is not going to die until they do and as long as new consoles come out there will be the drive for pc hardware to keep up (not to mention the console hardware is often provided by PC companies so the tech development is linked)

and as for windows going away...ain't going to happen anytime soon people said the same thing about IE when it had vastly superior competition launched and had anti-competitive lawsuits won against it....it's been almost a decade since then and IE has only just dropped below 50% market share (which still means it holds more than any other browser)...don't underestimate the stubbornness of people who are just used too one particular brand/product ...windows may die one day....but it'll take at least a decade of decline before it's really gone and it isn't dropping yet

as for the OP x86 dying...eventually yes...from nvidia, unlikely...in the next 10 years, possible it will lose a lot of market base to a successor but not truely die...the most likely thing I can see happening is Intel/AMD hitting a wall in die shrinks/architecture optimization and jumping ship to an entirely new architecture to keep up moore's law...and i'd bet whatever we see would have some inbuild cross-compatibility for a while (hell mebbe one x86 core on die to run old programs, if that's possible)

yey for long rambling posts i should say I'm not arrogent enough to believe i'm right in all this just my 2c+I'm bored at work so this is something to do

 

[HeXen]

..I'm willing to bet next gen consoles will launch with 4K resolution support since we're already seeing TV's demo'd that support it (TV's won't really hit retail till 2012-2013 but neither will next gen

i'll take that bet cause you already lost. Sony already announced that the PS4 will not offer higher than 1080p support. MS likely wont either, for starters it took over 10 years for 1080p to take off. I still remember HHgregg selling an HDTV back in HS (mid 90's) but it was stupid price.
Second, its too demanding on hardware..so upscaling at best is all it could do which isnt anything worth a crap.

.your forgetting one very very important market....MMO's..so far very few have had real success on consoles and the main PC only products continue to rake in massive revenue

no one forgot, just that no one cares. takes a certain breed to get into MMO's without careing about any other genre. I certainly don't want to buy an expensive gaming rig just to have nothing available other than MMO's.

 

[Dark Shroud]

android? chrome os? ipad? heard of those? when you consider millions of units sold, windows IS going away.

Android is for phones and it's already been shown that people don't pay for Android apps. The App store is anarchy, the OS is splitered, the best things going for it are flash support & price.

Chrome OS devices are not even for sale yet. It's also going to be a hard sell trying to explain to people it runs everything online and if their net connection hiccups the thing becomes usless for continueing what they were doing. Just try and explain to people they can't keep their pictures or music on it and their "iPod" won't really be supported.

iPads are mostly a toy at this point. Everything of importance can be done on cheaper eReaders or Android tablets that support flash.

 

[cotak13]

GF is losing money making AMD chips. Once the 2 truly separate, AMD will have to report actually buying chips from GF.

Even if you own a company you already have to report costs of doing business with it. Even in the same company interdepartmental costs are recorded. That's why admin assistants are so anal about who they hand stationary out to. They don't want another department's cost off loaded onto their dept.

When AMD and Globalfoundaries are fully separated there will be no difference.

Honestly, are most people posting here student who have never worked before?

 

[Dark Shroud]

Yes, but where are consoles headed?

In 2005, Xbox 360 was using high end hardware (triple core Power PC) and a discrete graphics chip.

But what if the next generation doesn't start off so "High end"? Cheap Console SOC already optimized for 1080p effectively cripples the discrete video card gaming market. (HPC would be a different story)

Microsoft & Sony can both dump in a new stronger more efficient Power PC quad core as the base and not cause too many problems for develpers while keeping everything cheap. Probably a bump up to 2GB of memory, maybe 4GB if we're lucky and they're smart and Blu-ray for both.

Microsoft could probably go with an upgrade to their current APU design with a new DX11 CPU core from AMD. The real questions are will Sony add a cell cluster again (I hope so now that Devs know how to use it)? And what will Sony use for a GPU since Microsoft is currently doing what Sony wanted to do last time with the PS3.

 

[Dark Shroud]

Is gaming demanding? Or is the rendering of graphics of gaming that is demanding? Maybe a 16 cores CPU will rock your game, or maybe a dual core CPU with just a bit higher clock speed will beat a thousand dollars CPU in terms of gaming. What I can say for sure though is, if SB come without the GPU core which many of us here won't use, it can run at a higher clock speed. Maybe stripping all other craps bundled within the CPU will actually allow a bit more clock speed. What is ARM? Maybe it is a CPU that is without things that we don't need.

LGA 2011 Sandy Bridge will come without GPU cores. ARM SoCs do have GPU cores as well. Gaming specifically on consoles uses PowerPC, most people don't even realize that the 360 & PS3 use the same core or "base" from IBM.

Right now Sandy Bridge & Power PC with or without GPU cores can run 2 or more full GHZ above ARM chips in the 5GHZ range as quad cores with the option of adding more cores or SPEs (Cell).

Intel & AMD both have SoC that are very small. Intel will be shrinking down to 22nm next year as well making another big jump with their IGPs. AMD has stronger GPU cores with Fusion technology and very low power usage. Both companies will have GPU cores with very low latency interconnects that can run hardware acceleration in applications.

 

[Bradtech519]

ARM is great for mobile low performance devices. I suppose it could be used to run the OS kernel and other tasks in some kind of SoC concoction if need be and then high performance CPU/GPUs do the rest. At the moment I don't see x86-x64 going away just for the hell of it..

 

[zephyrprime]

Intel is definitely losing out on the mobile market right now due to their lack of low end mobile parts. However, as mobile processors become more and more powerful, the mobile market will effectively grow *upwards* towards the realm of more powerful processors that Intel makes. Now, I don't mean that mobiles will be using desktop parts. I just mean that things will be easier for Intel the more powerful mobile processors get. Also, Intel is not sitting on its laurels. It's Atom has already had a lot of success. I expect that they will design more small procs in the future.

One of Intel's biggest achilles heels is that don't allow licensing of their designs. Arm is correct is claiming that licensing is a big advantage for them. Licensing allows their basic designs to get customized into different forms that are specialized for different device niches. Intel has more of a one size fits all approach which isn't working so well for the fast moving mobile market right now.

 

[Nemesis 1]

My bet is Intel is happy about this . If arm can run X86 code however. Than that should in time remove the monoply tag from intel . Intel unleashed! Arm running x86 is in fact a game changer . Its the results of Arm capable X86 cpus that many won't like. It would be nice to get that Monoply tag removed Intel can manuver than . If NV can capture 20% of the x86 market and AMD can hang on to their 20% thats it . Intel is no longer a monoply

 

[Vic Vega]

Doubtful. At least, not within a decade. Too much critical infrastructure, not to mention most datacenters, are entirely x86. Shoot, the change from 32bit to 64bit hasn't even fully completely yet.

Not saying ARM won't eat away at x86's market share, but its not like Intel is going to remain stationary either. They've got extremely low power versions of x86 chips coming. I'm still hoping they'll open up most of the x86 standard.

I agree with this.

 

[P4man]

x86 going away is about as likely as Windows going away

There is more truth to that than you intended I think.Neither windows nor x86 is about to go away, but what is going away is wintel. Intel is now pushing linux alternatives like Moblin/Meego, ChromeOS and Android. Its developing CPU's that are unable to run windows (moorestown).

Microsoft on the other hand is most likely developing ARM cpu's that will to some extend compete with intel (they bought an ARM architecture license as well), and they are porting windows to ARM.

Think about it.

We had a wintel duopoly working hand in glove, finally that got challenged from below by "LARM" (Linux+ARM). The result? Each half of the duopoly is now pushing one half of the alternative to wintel. Wintel is dead and buried. Both intel and MS will have to compete on merit. Neither is likely to get far on their own:

Intel because its tied its future to a proprietary and.. less than stellar ISA everyone else is trying to get away from, and its competing against a virtual gorilla that has a huge volume and a huge cost advantage. Its also a much more flexible business model. Dont be silly and look at ARM mobile phone chip performance to think ARM is somehow unable to be used in high performance chips. Its an ISA, like Power or MIPS. Looking at the Wii CPU doesnt tell you anything how Power 7 servers perform.

MS wont get very far either outside x86 either. Windows on ARM, who wants that? Not compatible with all windows apps, so who would pay for a windows OS if you can get free alternatives that are much faster, have far more software (for ARM) and are less prone to virus, bugs, etc. Its a non starter for all but a few niches (MS office on corporate ARM desktops, perhaps servers running MSSQL etc).

Because of the enormous amount of legacy, wintel wont disappear over night, it wont even go away in a decade, but it will go the same way as Unix Risc workstations once went: an ever shrinking highend niche. Wintel is the next workstation, an expensive dinosaur that will die a very very slow death.

 

[OBLAMA2009]

There is more truth to that than you intended I think.Neither windows nor x86 is about to go away, but what is going away is wintel. Intel is now pushing linux alternatives like Moblin/Meego, ChromeOS and Android. Its developing CPU's that are unable to run windows (moorestown).

Microsoft on the other hand is most likely developing ARM cpu's that will to some extend compete with intel (they bought an ARM architecture license as well), and they are porting windows to ARM.

Think about it.

We had a wintel duopoly working hand in glove, finally that got challenged from below by "LARM" (Linux+ARM). The result? Each half of the duopoly is now pushing one half of the alternative to wintel. Wintel is dead and buried. Both intel and MS will have to compete on merit. Neither is likely to get far on their own:

Intel because its tied its future to a proprietary and.. less than stellar ISA everyone else is trying to get away from, and its competing against a virtual gorilla that has a huge volume and a huge cost advantage. Its also a much more flexible business model. Dont be silly and look at ARM mobile phone chip performance to think ARM is somehow unable to be used in high performance chips. Its an ISA, like Power or MIPS. Looking at the Wii CPU doesnt tell you anything how Power 7 servers perform.

MS wont get very far either outside x86 either. Windows on ARM, who wants that? Not compatible with all windows apps, so who would pay for a windows OS if you can get free alternatives that are much faster, have far more software (for ARM) and are less prone to virus, bugs, etc. Its a non starter for all but a few niches (MS office on corporate ARM desktops, perhaps servers running MSSQL etc).

Because of the enormous amount of legacy, wintel wont disappear over night, it wont even go away in a decade, but it will go the same way as Unix Risc workstations once went: an ever shrinking highend niche. Wintel is the next workstation, an expensive dinosaur that will die a very very slow death.

i think i agree with this. the best way by far to surf the internet and do work is still a pc and that will always be the case. but in the future you wont necessarily need windows on a pc to do it.

 

[Cogman]

There is more truth to that than you intended I think.Neither windows nor x86 is about to go away, but what is going away is wintel. Intel is now pushing linux alternatives like Moblin/Meego, ChromeOS and Android. Its developing CPU's that are unable to run windows (moorestown).

Microsoft on the other hand is most likely developing ARM cpu's that will to some extend compete with intel (they bought an ARM architecture license as well), and they are porting windows to ARM.

Think about it.

We had a wintel duopoly working hand in glove, finally that got challenged from below by "LARM" (Linux+ARM). The result? Each half of the duopoly is now pushing one half of the alternative to wintel. Wintel is dead and buried. Both intel and MS will have to compete on merit. Neither is likely to get far on their own:

Intel because its tied its future to a proprietary and.. less than stellar ISA everyone else is trying to get away from, and its competing against a virtual gorilla that has a huge volume and a huge cost advantage. Its also a much more flexible business model. Dont be silly and look at ARM mobile phone chip performance to think ARM is somehow unable to be used in high performance chips. Its an ISA, like Power or MIPS. Looking at the Wii CPU doesnt tell you anything how Power 7 servers perform.

MS wont get very far either outside x86 either. Windows on ARM, who wants that? Not compatible with all windows apps, so who would pay for a windows OS if you can get free alternatives that are much faster, have far more software (for ARM) and are less prone to virus, bugs, etc. Its a non starter for all but a few niches (MS office on corporate ARM desktops, perhaps servers running MSSQL etc).

Because of the enormous amount of legacy, wintel wont disappear over night, it wont even go away in a decade, but it will go the same way as Unix Risc workstations once went: an ever shrinking highend niche. Wintel is the next workstation, an expensive dinosaur that will die a very very slow death.

You made one giant leap. You believe that Windows for ARM would be a non-stater, I disagree. Microsoft, if anything, has a lot of experience developing operating systems. If one day they decided to switch to ARM for their primary OS, you would see some pretty big changes to the way their OS currently looks. A large portion of microsoft's headaches come from legacy support. What happens when they don't have to deal with that any more? All the sudden, their OS becomes much much smaller and faster. They don't have to ship with 50000000000 drivers for devices that last run in the eighties or support interfaces that haven't been used for a long time. Huge sections of their code could be cut right out making them a very sleek OS.

Their biggest problem would be that they wouldn't be the big men around anymore. (see their attempts at windows Mobile.) that MIGHT kill them off as it was almost dumb luck that they became the giant that they are in the first place.

I wouldn't count intel out either. Love them or hate them, they have some of the best engineers and scientists around. If there was a major architecture switch, I wouldn't be surprised to see them leading the way. They have the resourced and the patents to make some pretty astounding architectures. Small time companies that licence the ARM architecture aren't going to take that away from them anytime soon. Heck, they could even licence portions of the arm architecture themselves an make some pretty powerful chips. If anything, Intel is in the best position they could be in right now. The only market they failed to get into was the mobile market.

AMD might go the way of the dinosaurs, but they might do that regardless of x86 dieing.

ARM is a great for the embedded and low power environments that it is put into. However, they don't have any design that even comes close to the power of a modern Intel CPU. IBM is pretty much the only competitor that fits that bill (with a different architecture)

 

[P4man]

You made one giant leap. You believe that Windows for ARM would be a non-stater, I disagree.

Here is the thing though, if MS succeeds with windows on ARM, its still the death of wintel. It will become "Warm" (but it wont). If intel succeeds in the android/chrome/meego/qnx/webos whatever space, its still the death of wintel, it will become lintel.

That said, MS will have some success with windows on arm, Im sure. Mostly in corporate (MS Office among other things) and maybe even servers. They will also sell to the clueless who bought windows machines for 10 years.

BUT. Windows on ARM will lose windows' by far biggest ace: a gazillion applications (and you make it even worse, by adding legacy hardware support to that). If you are going to switch to a platform that wont run all the software (and hardware peripherals) you own, know and are used to, you might as well pick the best OS, and for most, that wont be windows, even if it gets lean and fast. Or even free.

As for intel, they already made ARM chips. Look up xscale. They hardly out engineered their competitors, xscale was at best, an also-run. Intel has failed with every ISA that is not x86. Itanium, i960, iAXP462, xScale,.. heck even larabee, even though that is mostly x86. Moreover, they keep proving how unable they are to build a proper gpu and proper gpu drivers. Anything is possible, but in an ARM landscape, they would be an underdog again. But they arent going to try, they are going to continue trying to shoehorn the one ISA they know in to everything and it wont work. Not so much because the engineering challenges (which are very real with x86) as because of the business model.

ARM is a great for the embedded and low power environments that it is put into. However, they don't have any design that even comes close to the power of a modern Intel CPU.

See my other posts in this and related threads today. ARM is an ISA, like Power or x86. What you've seen so far is mostly cortex cores, designed by ARM themselves and licensed by apple, nvidia, samsung etc. Cores designed for mobile. But cortex is to ARM what Atom is to x86 or the Wii hollywood cpu to Power(PC) : low cost, low power implementation. Not a sign of potential performance.

nVidia and microsoft (and qualcomm and marvell) have architecture licenses, that dont limit them to ARM cores, and are free to design implementations that compete with high end x86 or highend Power designs like Nehalem, the xbox xeon cpu or even Power 7. But still remain compatible with ARM or borrow and adapt parts of the cores that they think suitable.

 

[Cogman]

Here is the thing though, if MS succeeds with windows on ARM, its still the death of wintel. It will become "Warm" (but it wont). If intel succeeds in the android/chrome/meego/qnx/webos whatever space, its still the death of wintel, it will become lintel.

That said, MS will have some success with windows on arm, Im sure. Mostly in corporate (MS Office among other things) and maybe even servers. They will also sell to the clueless who bought windows machines for 10 years.

BUT. Windows on ARM will lose windows' by far biggest ace: a gazillion applications (and you make it even worse, by adding legacy hardware support to that). If you are going to switch to a platform that wont run all the software (and hardware peripherals) you own, know and are used to, you might as well pick the best OS, and for most, that wont be windows, even if it gets lean and fast. Or even free.

Well, I just disagree. The "loads" of available userland apps available to an ARM environment are generally not all that great. Microsoft still possesses a large code base from which they can make some really nice applications. With their push on developers to use primarily C# for application development, the switch will be pretty dang painless as well. What they lose is legacy apps (which will hurt, but not kill IMO).

As for intel, they already made ARM chips. Look up xscale. They hardly out engineered their competitors, xscale was at best, an also-run. Intel has failed with every ISA that is not x86. Itanium, i960, iAXP462, xScale,.. heck even larabee, even though that is mostly x86. Moreover, they keep proving how unable they are to build a proper gpu and proper gpu drivers. Anything is possible, but in an ARM landscape, they would be an underdog again. But they arent going to try, they are going to continue trying to shoehorn the one ISA they know in to everything and it wont work. Not so much because the engineering challenges (which are very real with x86) as because of the business model.

See my other posts in this and related threads today. ARM is an ISA, like Power or x86. What you've seen so far is mostly cortex cores, designed by ARM themselves and licensed by apple, nvidia, samsung etc. Cores designed for mobile. But cortex is to ARM what Atom is to x86 or the Wii hollywood cpu to Power(PC) : low cost, low power implementation. Not a sign of potential performance.

nVidia and microsoft (and qualcomm and marvell) have architecture licenses, that dont limit them to ARM cores, and are free to design implementations that compete with high end x86 or highend Power designs like Nehalem, the xbox xeon cpu or even Power 7. But still remain compatible with ARM or borrow and adapt parts of the cores that they think suitable.

ARM is an ISA, and an ISA that doesn't have a company that is powerful enough to produce something competitive. I'm not saying it is impossible for the ARM architecture to be competitive, I'm saying that there isn't an ARM implementation that is competitive. I'm saying that there likely won't be one that is competitive as there just isn't a company that has the resources to make it such (or a company that is willing to).

As for the "failed" architectures. Itanium is still alive and kicking. It may not be a dominating force in the market, but it is there and it taught valuable lessons to the intel engineers.

The other failed attempts I would a credit to the fact that they were side projects, and not Intel's bread and butter. The more Intel sees the squeeze to move away from x86, the more money they are going to invest in the transition. The just haven't seen any sort of reason to move along.

Just because they haven't been successful in their new ISAs, doesn't mean they can't be. It is far more likely that they as a multibillion dollar company will be capable of adapting as opposed to joe's garage company coming in and taking them out.



As for the wintel stuff. I really don't know what you are trying to get on about there. The success of one going to a new platform doesn't mean the demise of the other. Hell, the success of the one may lead to the success of the other. Just because moving might help linux, doesn't mean it will kill windows.

 

[Any_Name_Does]

This may or may not be off topic, but I'd like to know if it is beneficial for the users, industry, and software developers to have many cpu architectures and operating systems.

 

[OBLAMA2009]

This may or may not be off topic, but I'd like to know if it is beneficial for the users, industry, and software developers to have many cpu architectures and operating systems.

so much is done on the web that it matters a lot less now i think

 

[Any_Name_Does]

so much is done on the web that it matters a lot less now i think

Can you explain that?

 

[P4man]

ARM is an ISA, and an ISA that doesn't have a company that is powerful enough to produce something competitive. I'm not saying it is impossible for the ARM architecture to be competitive, I'm saying that there isn't an ARM implementation that is competitive. I'm saying that there likely won't be one that is competitive as there just isn't a company that has the resources to make it such (or a company that is willing to).

Beating intel isnt impossible. The guys at VIA have delivered a better lowend chip than atom, despite a huge process handicap and despite an engineering team and R&D budget that is ridiculously small compared to what intel has.

But Samsung and Microsoft are not exactly small players, they certainly can afford the resources. I also think nVidia might well pull this off, they have assembled a team of very experienced CPU guys and nVidia arent new at designing complex highend chips.

Now beating intel in single threaded integer performance will be hard for anyone not called IBM, but you dont need that to be competitive. A few things are happening at once; the importance of raw cpu performance is dwindling, software is increasingly becoming parallel and much of the heavy stuff is being offloaded, either to the cloud or the gpu's.

Therefore nVidia (or MS or samsung or anyone) doesnt need a cpu that outperforms intels highest end cpu's in single threaded integer performance. For desktop, they need something that is faster than the current purely mobile ARM designs and that achieves something Core 2 class. That shouldnt be that hard, in fact a cortex A15 with a desktop class IO and memory interface would probably get you there already. Add to that a better GPU and GPGPU capabilities than intels, if needed throw more cores at it, while still retaining a smaller (therefore cheaper) chip that is probably a lot more energy efficient too, and you have a competitive chip. Not core i7 beating, but something that can compete with intels mainstream product range, selling in to notebooks, all in ones and desktops running ChromeOS or windows for ARM. And that is going to undermine intels fat monopoly margins.

For servers, competing with intel should be easier even. Many workloads, especially "cloud servers" are all about performance/W and performance/$. Small wonder Atom servers are getting so popular, you even see 512 cpu servers based on atom. That is, despite Atom currently not even having ECC support (ouch), despite having to chose between virtualisation, 64 bit support and dualcore. You cant have it all today in an atom, but thats not stopping big OEMs and startups alike to design servers around it.

Building an ARM chip that performs like an Atom on server loads, at similar or lower price points and power consumption is very doable. Id even say 'easy'. And you can do so while offering the features that intel deliberately disables on atom to help save its xeon margins. Marvell is working on this, Im sure nvidia is, and it doesnt seem to be something particularly unlikely to succeed.

Could intel react? Sure, they could drop some of the restrictions on atom. And allow $30 atoms to go in to server sockets where they used to sell $300 xeons. That will help their margins.

Then there is HPC; nvidia is already making some inroads with tesla. Why should it rely on x86 cpu's to feed their tesla's? You could cram a dozen cortex a9 class integer cores in a single fermi and it would barely have an impact on power consumption or die size. Now Cortex is probably too slow for this, project denver should fix that and offer nvidia a complete HPC building block.

As for the "failed" architectures. Itanium is still alive and kicking. It may not be a dominating force in the market, but it is there and it taught valuable lessons to the intel engineers.

Its not quite dead, their new ISA just failed to deliver a better cpu than the competition, including the x86 competition. So from a "merchant' 64 bit server cpu, it has become an HP only chip to replace HPs previous risc cpu's. In that sense, its an utter failure, it delivered nothing that couldnt and hasnt been done with existing ISAs. It just costed 5 or 10 billion dollars more to get there.

The other failed attempts I would a credit to the fact that they were side projects, and not Intel's bread and butter.

You mean they all failed to become intels bread and butter. They where all destined to replace x86, with the exception of their ARM chips.

The more Intel sees the squeeze to move away from x86, the more money they are going to invest in the transition.

Sorry, in what transition?

Just because they haven't been successful in their new ISAs, doesn't mean they can't be. It is far more likely that they as a multibillion dollar company will be capable of adapting as opposed to joe's garage company coming in and taking them out.

VIA's garage company did just that from a technical POV. As AMD is doing arguably with their fusion chips. Besides, if you look at the ARM ecosystem, thats not exactly "joe's garage". Its a sky scaper that totally dwars intel's shop in size and R&D expenditure. Dont forget the ARM business model, where 100s of companies essentially pool a big part of their R&D money in to ARM holdings thats doing a lot of the design for all of them. Combined, they outsell intel 10-1. Thats why a standard ARM core comes with a ~$0.2 core license. I suspect an Atom chip carries a per core R&D cost that is closer to $5 or $10. Its certainly more than the silicon cost of the chip. The same problem that AMD and VIA have competing with intel (volume) is going to be intels problem competing with ARM.

 

[sxr7171]

It will happen for sure, but not overnight. Give it 5 years and you'll see that people won't need big powerful computers. The future paradigm will be decent CPU, fast GPU and SSD. Coupled with lighter "cloud-based" operating systems it will easily handle all client PC functions.

Currently devices like the iPad are weak, but in 5 years why would you need a 2010 style PC? It could just as well have a keyboard and except for fringe elements like gamers it will work for everyone. All major processing will be done by a remote server, probably all your data will reside on a local or remote server. Home users need all of 3 or 4 things to be done well and businesses users generally just need a dumb terminal.