Intel Fights For Its Future (mondaynote.com)

An anonymous reader shares a post: The Smartphone 2.0 era has destroyed many companies: Nokia, Blackberry, Palm... Will Intel be another victim, either as a result of the proposed Broadcom-Qualcomm combination, or as a consequence of a suicidal defense move? Intel sees the Qualcomm+Broadcom combination as an existential threat, an urgent one. But rather than going to the Feds to try and scuttle the deal through a long and uncertain process, Intel is rumored to be "working with advisors" (in plainer English, the company's Investment Bankers) on a countermove: acquire Broadcom. Why the sudden sense of urgency? What is the existential threat? And wouldn't the always risky move of combining two cultures, employees, and physical plants introduce an even greater peril?

To begin with, the threat to Intel's business isn't new; the company has been at risk for more than a decade. By declining Steve Jobs' proposal to make the original iPhone CPU in 2005, Intel missed a huge opportunity. The company's disbelief in Apple's ambitious forecast is belied by the numbers: More than 1.8 billion iOS devices have been sold thus far. Intel passed on the biggest product wave the industry has seen, bigger than the PC. Samsung and now TSMC manufacture iPhone CPUs. Just as important, there are billions of Android-powered machines, as well. One doesn't have to assume 100% share in the smartphone CPU market to see Intel's gigantic loss.

Comments

Typical idiocy clickbait from the "editors"

[CajunArson]

As usual another stupid article full of line noise instead of anything intelligent to say.

Incidentally, if Intel is "fighting for its future" by making huge profits in a variety of areas then why the hell is Qualcomm -- the effective monopolist in smartphone wireless devices and also a huge player in smartphone SoCs -- even conceivably a target of a takeover? Why the hell isn't Qualcomm about to buy out Intel if Intel is so behind the curve and Qualcomm is supposedly so great?

Let's not even forget how this idiot "analysis" is somehow never applied to fanboy-favorite AMD who for some reason is destined for greatness without ever having made a single product that could be used in a mobile phone.

Re:Typical idiocy clickbait from the "editors"

[Anonymous Coward]

Now I think some old analog cell phones had some AMD discrete parts in them at one point. In early 90's or late 80's.

Re:Typical idiocy clickbait from the "editors"

[TWX]

The article (or at least the Slashdot synopsis) may be excessively editorializing it, but Intel is threatened by the reduction in the personal computer industry.

Consider that when PCs rose to prominence there were lots of architectures. Even after Wintel and Motorola/Apple dominated personal computers at home, business computing still had other architectures (MIPS, and Alpha immediately come to mind) to the extent that Microsoft felt the need to port their business OS to those platforms, rather than to force x86.

The end of the model that all software has to be compatible with x86/AMD64 and that the gatekeepers for software for new devices (Apple's and Google's respective repositories) require that the software work on their devices almost without respect to the underlying CPUs, plus the 'cloud' model and various other virtual machine models may abstract the software developer away from the physical hardware to the point that we might again see a proliferation of various architectures again. Intel has reigned supreme because it was difficult to port software or to write software to run on everything, but if that has changed then suddenly it doesn't matter what actual CPU is in the phone or tablet or even server, it'll just work when it's time for the software to run.

That's the threat to Intel's business-model, a loss of near-monopoly on processors because new devices don't need Intel's processors.

Re: Typical idiocy clickbait from the "editors"

[Anonymous Coward]

As usual another stupid article full of line noise instead of anything intelligent to say.

The article was surprisingly clear and full of detail.

Incidentally, if Intel is "fighting for its future" by making huge profits in a variety of areas then why the hell is Qualcomm -- the effective monopolist in smartphone wireless devices and also a huge player in smartphone SoCs -- even conceivably a target of a takeover? Why the hell isn't Qualcomm about to buy out Intel if Intel is so behind the curve and Qualcomm is supposedly so great?

Why, why, why? All your questions were addressed in the article.

Let's not even forget how this idiot "analysis" is somehow never applied to fanboy-favorite AMD who for some reason is destined for greatness without ever having made a single product that could be used in a mobile phone.

WTF has this got to do with AMD? They are not a dog in this fight. I thing l think your bias is showing.

Re:Typical idiocy clickbait from the "editors"

[Anonymous Coward]

I agree. I think that the author of this article has missed the point. Intel is not embracing mobile chipsets with this move - they won't need to. They are doubling down on the cloud (someone else's computer), and in particular the storage aspect of the cloud. With this move, Intel would get a virtual monopoly on RAID controllers. Broadcom makes proprietary controllers for Intel (that only work on Intel motherboards due to firmware restrictions on those boards), as well as for OEMs. LSI Logic, AMCC, 3ware, and Avago are a few of the companies that Broadcom has absorbed in the past decade. Most RAID boards (other than those by HP) can be cross-flashed back to generic LSI/Avago firmware (usually with more fixes) than stock firmware. If Intel buys Broadcom happens, I'm sure that cross-flashing and will stop, and all OEM - not just HP will have to lock their RAID cards to their motherboards.

One alternative I found to Broadcomm is Adaptec/MicroSemi. If they have enough of a market share, then regulators will likely go ahead with a Broadcom/Intel merger/buyout. Current president really likes when things are for national security - *sigh* - and Intel will likely make free backdoors in this stuff to any and all governments that ask.

Re:Typical idiocy clickbait from the "editors"

[Bert64]

Those other architectures were forced into the highend niche, and eventually died out...
The same is happening to intel now, they are forced into the highend niche where arm chips are suitable for an increasing amount of day to day tasks, and only people with specialised requirements currently require the higher performance intel chips.
Fast forward a few years and the increased volume of sales for the arm chips provides more development money, and arm starts overtaking intel in performance too.

Re:Typical idiocy clickbait from the "editors"

[Anonymous Coward]

Microsoft didn't feel the need to port to alpha or MIPS for customer reasons.
They did it for one of two reasons:
1) Digital and/or SGI/MIPS paid them to do so. (I know Digital had a big team working to port windows nt to alpha)
2) They needed NT on Alpha and/or MIPS because x86 couldn't do something. Memory/Speed/etc..

No customers were screaming for it. If you look at the sales numbers (at least for Digital), many Alpha NT boxes actually got used for linux and/or running OSF1 (digital unix). The NT boxes were at a lower price point than the Digital Unix ones.

Re:Typical idiocy clickbait from the "editors"

[Anonymous Coward]

Even after Wintel and Motorola/Apple dominated personal computers at home, business computing still had other architectures (MIPS, and Alpha immediately come to mind) to the extent that Microsoft felt the need to port their business OS to those platforms, rather than to force x86. .... Intel has reigned supreme because it was difficult to port software or to write software to run on everything

I used to believe this, but it's not really true; Linux compiles readily on multiple platforms, has enough backing to get sufficient driver support, and companies/people optimizing for a whole host of considerations very consistently have chosen x86/x86_64. Intel reigned supreme because they have the best price/performance ratio. They also have the highest per-CPU performance (AMD being really close and price competitive). This can easily be proven by pointing out that x86/x86_64 is still THE platform on servers. Repeatedly other chip makers have come along and tried to challenge either on outright performance, on outright cost, or outright power usage. Only in very small devices, like phones and tablets, has there been any sort of serious dent*. In the end, though, ARM design is over a decade behind and it's unclear when/if ARM will get to the level of today's PC processors, and of course at that point they'll likely be years behind.

The real point, though, is that in the past Intel was able to stay ahead of other processor makers with high R&D costs because of the basically steady guarantee of money because of Wintel. Microsoft's efforts to port to other processors was just Microsoft's usually backstabby behavior towards other companies, including its allies, to not be too dependent on Intel if necessary. Intel, of course, has done similar with their support of Linux. Regardless, ARM is in a similar situation now with a steady stream of money from phones/tablets, so in theory they could heavily invest in R&D and potentially license technology from AMD or otherwise develop their own technology to be competitive with Intel at a faster rate than Intel. They have the luxury of knowing what worked and what didn't and a vague idea of where to go to follow Intel's lead without directly copying them and tripping over one of their patents.

Having said that, I'm not exactly hopeful. Intel has spent a good deal of effort in optimizing their CPUs and being sure they're on the cutting edge, even though it comes at a great cost. ARM may have the drive to do this, but Intel has always been of the mindset that it needs to fight for its future. Perhaps that's the real core reason that x86_64 is the standard--excusing of course it took AMD to push Intel to accept a 64-bit x86 or risk losing their Wintel money machine.

* There's also a lot of auxiliary processors used in a lot of situations, but we're obviously focused on the dominant CPU for high performance applications, and that's very strongly now x86_64 and ARM (with some others still holding on in various niches where it makes sense or the sunk cost is too great).

Re:Typical idiocy clickbait from the "editors"

[Anonymous Coward]

Intel has reigned supreme because they had the best lawyers in the 90's and strong-armed Dell, Gateway, Compaq and everyone else to not use AMD.
^^^-- FTFY

AMD had the superior CPU (athlon) that was eating PPro/PII for lunch (beat them to 1GHz and to AMD64 extensions) and Nexgen had a great cpu as well that was clobbering Celeron. But Intel maintained its monopoloy power through lawyers, not engineering.

Source: I worked there for 24 years. Yeah, I'm an AC, but I worked in a perf lab in Santa Clara and the competition was destroying us in both price and performance. If it wasn't for deep pockets, Intel would not be a $100B monopoly and AMD would be sharing the wealth.

This would have given IBM and Motorola an inroads due to the shared x86 pool.

But doesn't matter now: all a PC needs to do is run a web browser. For 99% of the people that is sufficient. I'd say 99.9%.

Re:Typical idiocy clickbait from the "editors"

[Anonymous Coward]

But you can't quite separate the utility of the software from the hardware. Android and iOS are both fairly limited in what they can do. If I'm paying several hundred dollars for a machine, it better be able to run any software I give it and receive updates for a long time. That's actually Windows.

Apple charges way too much for its products, especially given how limited iOS actually is. Android has more utility, but has an update problem (to be fair, iOS devices don't really update beyond 5 years, compared to 10 years for Windows).

ARM was a good enough choice for smartphones, just like Android and iOS are good enough for what they do. Intel may not be able to break into the phone market, but ARM chips will have difficulty in Intel's markets, such as laptops (Apple's too expensive, the Android 2-in-1's have an update problem, and there are already cheap Windows laptops on the market).

Re:Typical idiocy clickbait from the "editors"

[TWX]

Android and iOS are limited in some senses, but that hasn't stopped massively widespread adoption of them either, and that includes on devices more than just phones. In some ways, what was old is new again. For the longest time the actual processing was done on a computer remote to the user, and the user's terminal was basically a dumb device that didn't really do that much. In some ways the Web has made that happen again. Admittedly not as thoroughly since there is client-side stuff going on, but by and large the content isn't kept locally and the local client device only has to be powerful enough to handle the end graphics and sound.

Re:Typical idiocy clickbait from the "editors"

[nitehawk214]

The same reason why AOL was able to buy TimeWarner in 2000. What matters is apparent value not actual value.

Smartphones are the hot product, so Qualcomm has more apparent value than Intel which has focus on boring (but profitable) pc and server memory and cpus.

Yeah, it is stupid, yeah it doesn't make sense; but that is how stock markets work. A company not making quite as much money as predicted is the same as that company losing money.

Re:Typical idiocy clickbait from the "editors"

[MachineShedFred]

Don't forget that it would also give Intel a massive market share on WiFi radios too. Every Mac uses a Broadcom, and many Windows PC OEMs have Intel PRO/Wireless for vPro, and a Broadcom alternative for non-vPro. There's a few out there using Atheros (including Dell) but the early versions of those were rather lumpy, and the fix was usually to throw a $30 Intel radio into the slot where the Atheros used to be.

But guess what? Atheros is Qualcomm. So if all of this comes to pass, it all becomes Intel.

Re:Typical idiocy clickbait from the "editors"

[Anonymous Coward]

I like how you say "Motorola/Apple" as if the Mac were the only thing using Motorola 680x0 CPUs and not the Amiga, the Atari ST520/1040 and numerous laser printers and arcade and game consoles.

You're also forgetting the MOS/Rockwell/Synertek 65x02 and 65816 processors in the Apple II, Comodore 64/128s, Microbees, etc. Motorola feared MOS so much they dropped the price of 6800 processors from $180/unit to $30/unit over a 2 year period and also tried to shut them down through court proceedings.

Re:Typical idiocy clickbait from the "editors"

[LinuxIsGarbage]

In some ways, what was old is new again.

I think we're returning to the idea of a "Home Computer" vs a "Personal Computer / Workstation" Android and iOS really simplify things down. This to me is like the return of "Home Computers" C64, Apple //, etc. These systems you either switch them on and load right into ROM based BASIC, or you pop a cartridge or floppy in and start it up your application. Very limited functionality, but very simple to use. "Personal Computers", which are mote like workstations, are significantly more powerful, and flexible, but harder to administer, which most people aren't capable of, and have no interest in.

For most people, they want their computer to operate like an appliance. Can they open their email? Can they listen to their music? Take a picture? Watch Netflix / Youtube? Go on Facebook / Instagram / Snapchat / Whatever new trending social media app? Android and iOS are good for this. PCs (running Windows, macOS, or GNU/Linux) still offer more power when it's required... usually for business or power users.

Re:Typical idiocy clickbait from the "editors"

[mikael]

It's not just ARM. There are lots of startups coming up with various chips to do face and emotion recognition, posture recognition, motion recognition and all sorts of basic vision processing that would form a single visual circuit in a mammalian brain. These are being designed using machine learning techniques and don't require the double/floating point precision processing that are typically supplied by Intel chips. Even the fluid dynamics people were realizing that machine learning techniques were helping to make simulations more accurate while not requiring double-point precision.

Re:Typical idiocy clickbait from the "editors"

[mikael]

I would say the categories can be extended to:
smart TV, smartphone, tablet, netbook, office workstation, gaming laptop/desktop PC, rack-mounted servers, PC server, engineering laptop/desktop PC.

Even the smartphones and tablets are more powerful than a early gaming console like an Ultra 64/Playstation 3. A gaming PC is more powerful than an office "workstation" with multi-screens, SLI cards. Engineering workstations can have dual-socket boards with quad-SLI boards and 40+ core XEON chips.

Re:Typical idiocy clickbait from the "editors"

[Anonymous Coward]

And each blew themselves up through inefficient instruction implementations. The 6809 had the stop-and-catch-fire instruction due to the way voltages could be applied to IO ports. It was known that it was quicker to subtract the value of one register from itself than it was to assign zero as a constant.

Re: Typical idiocy clickbait from the "editors"

[TWX]

The Amiga and the Atari ST lacked the market-share over the sum of decades that both Wintel and Apple enjoyed. They aren't unimportant in the history of computing but when discussing the totality of the proliferation of personal computers they don't matter.

Re:Typical idiocy clickbait from the "editors"

[Anonymous Coward]

That's the threat to Intel's business-model, a loss of near-monopoly on processors because new devices don't need Intel's processors.

No

The threat to Intel is Intel itself

Intel is stupid and lazy, too fucking stupid and too fucking lazy

Intel was one of the first licensee of the ARM processor - but they decided that the ARM processor was of no use and abandoned it altogether

And then Intel stuck to X86, totally ignoring the need to go 64-bits

Only after AMD came up with AMD64 would Intel do something

When Intel got both 32-bit and 64-bit covered, and AMD sank (due to its own stupidity), Intel again slacked off

Intel used to lead the world in shrinking node. It used to be the world leader - some 3 year ahead of its closet competitor

Then it slacked off, and got itself stuck in the 14-nm node

Now that the rest of the world (especially Samsung and TSMC) have caught up (and surpassed) Intel, only now that Intel decides to do something

Buying Broadcom?

What the fuck Intel wants in Broadcom?

What the fuck Intel can do with stuff it acquires from Broadcom?

Intel's biggest problem is Intel itself --- The combination of too fucking lazy, and too fucking stupid will one day kill Intel, and that fateful day is fast approaching.

Re: Typical idiocy clickbait from the "editors"

[Anonymous Coward]

AMD != ADI

Re:Typical idiocy clickbait from the "editors"

[toddestan]

AMD had the best CPU back then, and it was cheaper too. The problem is to actually use the Athlon you were stuck with garbage like the VIA's KT133 chipset. It's no wonder the OEM's fled in terror to boring but safe choices like Intel's 440BX chipset.

Strange article

[Anonymous Coward]

Full of sound and fury signifying nothing.

Because the PC market is dying. Someday. We do not know when but it will. But we will write articles of hyperbole for everyone in hopes of eyeballs and clicks.

Re:Strange article

[Anonymous Coward]

The PC market isn't dying, it's just changing architectures away from the one Intel has dominated.

However, thanks to the Internet, cellular companies wanting data transfer, and businesses wanting to offer everything "aaS" -- they still have a large market server side though they're now missing out from many typical consumers.

Re:Strange article

[omnichad]

Because the PC market is dying.

Only if you ask the stock market. Stable demand without growth is called a business (though shareholders tend not to care about that). Replacement cycles are long, but nothing has supplanted the PC.

Re:Strange article

[rtb61]

PC market is shrinking fast and what has replaced it is a plethora of smart products, the smart phone, the smart TV and tablet, for students in the family a cheap notebook. For business, smart terminals, simply easier to manager and no pesky USB or accessible disk drives of what ever type. See desktop replaced in by far the majority of instance and market shrinking back to power users, the core and they hate M$ only barely putting up with them and that is killing Intel because everyone is holding of upgrades because M$ are a privacy invasive big brother freaks who want to control people's digital lives and that is killing upgrades. Desktops are inevitably doomed but they can stretch out the next few decades as long as the OS improves substantially, privacy and security.

Re:Strange article

[Carewolf]

I am thinking this might be planted news by Intel to justify their acquisition as otherwise it would be rejected as a major monopoly already fined for abusing their monopoly expanding its monopoly further.

Re: Strange article

[Anonymous Coward]

Most businesses still use Windows based PCâ(TM)s and will do so for the foreseeable future. Why? Their line of business software. Consumers, on the other hand, can mostly go PC free. Iâ(TM)ve still got one at home but rarely use it. Still, there are some things that are just easier to do with a PC or laptop than on a phone or tablet and that wonâ(TM)t change anytime soon either. Intel may not be seeing the kind of growth that lets executives bring home the big bucks through stock options but the demand for their processors will still be there for decades to come.

Re:Strange article

[MightyYar]

but nothing has supplanted the PC

If you think about this, it is not a defensible statement. Sure, PCs still have a large collection of niches that nothing has supplanted - but you are ignoring the huge number of niches which have disappeared. You probably have a computer in your pocket right now with approximately the same power as a late-80s Cray. It has almost entirely wiped out the social aspect of the PC - email, IM, web forums, video and music sharing, etc. The PC games market is slowly losing ground to mobile.

Will there always be a market for powerful desktops (workstations)? Sure, I think so. Or maybe I lack imagination. But right now I think things like video editing, rendering, CAD, software development, and data analysis are safe. It's also not hard for me to imagine mobile processors eventually getting good enough for those fields, however. At that point, it will be more cost-effective to just do your work on a machine with souped-up mobile guts. Fast forward 10 years - why would anyone spring for a low-volume x86 chip when the ARM 10 in your pocket can be cheaply repurposed in a "desktop" for your video, CAD, or development work? Hell, even data analysis might be something better farmed out to a server somewhere... I already do this when I need more oomph than my workstation provides.

Intel should be worried - economies of scale are what won them the desktop, and x86 could easily be the next SPARC or Alpha or PA-RISC or MIPS or ...

Re:Strange article

[omnichad]

For business, smart terminals, simply easier to manager

Thin client vs thick client has been a tick-tock ever since the first mainframes entered the business market.

Desktops are inevitably doomed but they can stretch out the next few decades

That's as close to living as we've ever had... Decades are an eternity in computing.

Re:Strange article

[Carewolf]

PC market is shrinking fast

No, it isn't. The PC sales are dropping slowly and steadily, but the PC market penetration has not changed dramatically, people just upgrade more rarely.

Re:Strange article

[omnichad]

Since the argument was against the idea that the market for the PC is dying entirely, I wouldn't even consider all the people that never would have had a PC if they had an alternative. It's back to its original niche.

If the definition of PC is x86 compatibility, the market may eventually go away. If the definition is full-power, full-size personal computing device, I can't yet imagine a future like that.

Re:Strange article

[Anonymous Coward]

Perhaps one day the PC market will die, but the DIY market is increasing. https://www.statista.com/statistics/374093/global-diy-market-value/
     

Re:Strange article

[MightyMartian]

I think there's an argument to be made that it is ossifying. CPU and other architecture capabilities have risen to such a level in the last decade that it has disrupted the upgrade cycle. We're still running workstations we bought in 2009, and they even run Win10 (though not the latest creators update, but who cares). They do fine for browsing, document editing and the like, and now we simply replace them as they die, which doesn't actually happen all that often.

Now maybe my company is pushing the envelope a bit, but the refurb market, both for workstations and servers, is huge these days, because an off-lease three or four year old PC, server or laptop can be had for quite a bit less, and is going to do most tasks just as well as one I could go out and buy today. Yes, there are specialty and niche applications like gaming or CAD where you're going to want the latest and greatest, but for most people, if you browse and watch Netflix, you've pretty much covered the basis.

Where the frequent upgrades have been steady is in smartphones. That's where people seem to be giving their kids or Great Aunt Minny their three year old iPhone because the new shiny one is out. In part it's because portable devices are more likely to be damaged or destroyed, and in part because smartphone manufacturers have been as successful at using OS updates to push frequent upgrades as Microsoft was back in the day. And the fact is, in pure numbers, smartphones and feature phones are just much vaster markets than for PCs, and a great deal of the traditional casual computing that would have gone on on a PC now goes on on a phone.

Re:Strange article

[JMJimmy]

PC market is shrinking because they failed to solve the problems necessary to make them relevant in a portable device world.

A home PC should be like a furnace - rarely physically interacted with but fully integrated into the home. Every fixed screen in my home should be dumb, they should all run off a single PC. All "smart devices" should simply be interfaces that use the PC's hardware to execute/control their functions. Smartphones/tablets/portable devices should have a power saving mode that enables them to operate as a "dumb" screen or as a separate fully powered device.

The trouble is that they've left it to 3rd parties to solve these problems and write patchworks of underfunded software and "unique" hardware solutions. They should be developing standards, interfaces, and cohesive solutions to make a single, powerful PC relevant again for more than just hardcore gaming. I should need a multi-CPU system to run hundreds of small devices around the home without hiccups.

Re:Strange article

[MightyYar]

. It's back to its original niche.

The problem for us (and for Intel) is that niche is a small fraction of what they are sized for. Their multi-billion dollar fabs sitting idle is a financial disaster. Intel is already doing some contract manufacturing, but that's a tough game with many experienced competitors. Yes, I think you are right that there will always be a market for workstations, but I think we're going to see a slow drift towards what is becoming the new standard in commodity hardware. Most of us will use "PCs" with repurposed mobile guts, which is an interesting paradigm shift, and scary as hell if you are Intel.

Re:Strange article

[mysidia]

they should all run off a single PC. All "smart devices" should simply be interfaces that use the PC's hardware to execute/control their functions

For vendors... they keep more control if they use their own cloud, AND end users don't have to worry about replacing an expensive single PC to restore all those functions when it fails/dies

Re:Strange article

[Anonymous Coward]

PC market is shrinking fast and what has replaced it is a plethora of ..

Smaller, even more numerous PCs.

(The only catch being the new PCs are somewhat more user-hostile than the old ones, but way more portable so you can have one handily in your pocket at all times.)

Re:Strange article

[Anonymous Coward]

why would anyone spring for a low-volume x86 chip when the ARM 10 in your pocket can be cheaply repurposed in a "desktop" for your video, CAD, or development work?

For anything commercial, for this to work, we need to convince commercial developers to go from selling $1000 specialized vertical marker applications to selling $1 mobile applications.

Re: Strange article

[Anonymous Coward]

Smaller, even more numerous PCs.

(The only catch being the new PCs are somewhat more user-hostile than the old ones, but way more portable so you can have one handily in your pocket at all times.)

PC doesn't mean what you think it means.

Re: Strange article

[Anonymous Coward]

What do YOU think PC means?

Re:Strange article

[Anonymous Coward]

It has almost entirely wiped out the social aspect of the PC - email, IM, web forums, video and music sharing, etc.

Really? Why would I use a phone for that, with its tiny screen and lack of keyboard, except when being away from home means there's no alternative?

I do have a smartphone, but pretty much never write anything more than a few words on it as it's just too awkward.

Re:Strange article

[JMJimmy]

Vendors have an interest in not having to provide their own cloud. That is an on going expense for a one time purchase. Those who want it for control, are usually up to no good. Those who want it for data, really don't need the cloud service, they just need the software to phone home telemetry and perform updates, things that can be standardized for user privacy & security. The reality is that it's not a sustainable business model to run IoT devices without a subscription service.

On the expense issue, it's true that you don't want to replace an expensive PC to restore all those functions. I think the structure of PCs also needs to change along with it so they're externally modular. It would not be an easy problem to solve but I think it's solvable with common hardware interfaces.

By externally modular I mean that I don't need to open up the system to replace anything. I just order any X interface module knowing that it will be compatible with my X interface system. Need more processing power? Plug in another CPU module. One of the two RAM modules failed? Pop it out and plug in a fresh RAM module. The only time you would need to go to great expense is if you wanted/needed to change interfaces.

Obviously it's not that simple, there would be huge hurdles to overcome and such a system has potential drawbacks. As it stands now though I am relying on numerous 3rd parties for the continued functioning of my IoT devices and the retail price + diminishing returns on monetizing my data will not continue to sustain their cloud services. They also have more points of failure - not only what's in my home but also internet service at large and my ability to connect without heavy latency to their cloud service.

Re:Strange article

[mysidia]

Vendors have an interest in not having to provide their own cloud. That is an on going expense for a one time purchase.

No... it's probably a small expense for most products, and an eventual opportunity to get recurring revenue out of their customers -- either by starting to bill a new subscription (The Cloud excuse helps facilitate a "Rental model" for license to use the hardware and software --- Being cloud-based usually means additional revenue opportunities for the provider or more options to further exploit existing product owners for an extra revenue stream down the road), selling additional products or add-on services (Such as "plug-ins", "apps", or "music streaming"), or collecting data for marketing/advertising, OR after the end of the 3-year support period notifying their customers that they'll need to upgrade to $NEWER_HARDWARE_MODEL to keep using it.

they just need the software to phone home telemetry and perform updates,

Phoning home to a cloud infrastructure, so they need to build it anyways....
why not make it more integral to the product, to ensure the end user doesn't just firewall it off or disconnect it from the internet?

Re:Strange article

[Bert64]

Many don't get replaced at all...
A lot of people bought a PC to access the internet as there was little choice at the time, a lot of those people have moved onto tablets, gaming consoles and smartphones since then so while they still have a PC, it is probably gathering dust, won't be replaced if it dies and probably doesn't get used much if at all.

Re:Strange article

[Bert64]

Most smartphones support bluetooth keyboards...
Many can be docked to a larger screen, so you have the same device wether your mobile or in a fixed location - best of both.

For many use cases a smartphone is "good enough", typing may be slower but many people aren't very proficient typists anyway. There are also various speech to text options which have improved a lot in recent years.

Re:Strange article

[JMJimmy]

Subscription services sound nice, there's just not the income to support that very widely. People also can't afford the rate of obsolescence that would be required to sustain such a model either. We're already seeing that in the smartphone market and that's essentially what happened to the PC market. Buy small/cheaper devices that are "good enough" instead of an expensive/complicated/bulky PC.

I'm certainly not going to buy a "smart coffee maker" with a subscription & planned obsolescence when I can buy a dumb one without. I would buy one that is standardized and has a common interface though.

I think we also need to consider the massive shift that's coming with AI. If it turns out to be half as disruptive as experts claim there will be a serious shift in society. When AI/robotics can manufacture things cheaply and resources become more scarce (both physical and individual) will we still want the planned obsolescence model? We may just find society shifting the dynamic from profit based motives (obsolescence, disposability, etc) to longevity, interoperability, and customization.

Re:Strange article

[evil_aaronm]

I have a 2011 MacBook Pro as my primary workstation. With an upgraded SSD, this laptop is plenty fast enough for everything I do. Internet speed depends more on my home's connection, which isn't great but that's a product of where I live (Podunk). 3-D modeling with Blender is fast. Video editing isn't horrible. Xcode and compiling for Arduino / mcus is entirely reasonable. Until it completely gives up the ghost, I have no reason to upgrade this model. And from what I've read about soldered in memory and what not in newer models, I've no interest.

Re:Strange article

[Cro Magnon]

If the definition of PC is x86 compatibility, the market may eventually go away. If the definition is full-power, full-size personal computing device, I can't yet imagine a future like that.

This! My tablet isn't even close to replacing my x86 machine. I need a keyboard, mouse, multi-windows, file-system with user-created directories, etc. It might not have to be compatible with my current Wintel machines, but it has to have similar capabilities.

Re:Strange article

[omnichad]

No, we'd need to go from devices with walled gardens with a 30% cut on app purchases to a device with an open (but secure) market.

Re:Strange article

[MightyYar]

Whether you personally do it or not is beside the point - the market has gone that way.

Re:Strange article

[MightyYar]

For anything commercial, for this to work, we need to convince commercial developers to go from selling $1000 specialized vertical marker applications to selling $1 mobile applications.

I think they'll use the same OS and programs that they've always used, but the chips will change to whatever is currently offering up the best bang for the buck. Windows can run fine on an ARM now, and in 10 years I believe it will be a lot more common if Intel can't adjust. Autodesk, Microsoft, and Adobe can still sell their big commercial programs compiled for a different architecture. My Sony TV is already an all-in-one computer running Android TV... it's not a big stretch too see it being sold with upgraded storage, a keyboard, mouse, and Windows.

Re:Strange article

[omnichad]

Vendors have an interest in not having to provide their own cloud. That is an on going expense for a one time purchase

They love forcing their own cloud. They can force your devices to quit working so you'll buy the next new thing. If you had direct interface, you could use the thing pretty much forever.

Re:Strange article

[jd]

Ummm, that's a very stupid argument, I hope you realize.

Windows beat Linux because you can't play heavy games on Linux. And that's the only reason. You sure as hell can't play those games on a smartphone. First, it's running Linux, and second, it hasn't the power. It will never have the power because energy consumption goes up with the square of the CPU's abilities and batteries can't increase in capacity that fast.

Second, you can't do any decent wordprocessing in a phone or tablet, they're too prone to hanging, they don't have the screen and they still don't have the CPU to go beyond basics.

Third, the tablet market is collapsing because ebooks are crap and most people have now figured that out. Sales are way down and people are moving back to paper.

Fourth, the network speeds are too slow. It's fine in America, where 340 million suffer with barely faster than dial-up, but MOST OF THE WORLD is a damn sight faster and there are 6.7 billion of them. Rather a larger market. And they're expecting devices that can handle 10 gigabits per second to 400 gigabits per second, because to them that is NORMAL INTERNET SPEED. Smartphones and Tablets aren't capable.

Fifth, home users are used to 6-10 terabyte hard drives and SATA 3 speeds. That's NORMAL. When was the last time you bought a smartphone with a 10 terabyte SIM card?

Sixth, VR. You can't do CAVE or headsets on the USB port and, besides, it'll be needed for the charger to avoid setting the battery on fire.

Seventh, PCs don't explode. Smart phones and tablets DO. That's the problem with Li-Ion. And it can't be fixed, as "The Mouse" from Top Gear discovered.

Eighth, decent keyboard. A decent keyboard is faster and more effective than a touch screen, less vulnerable to false positives, less vulnerable to dirt and much more robust. Percussive maintenance on a PC keyboard won't shatter it. Try that on a cell phone and you end up with two half phones.

Ninth, HPC/Virtual Data Centre. Virtually all HPC these days is pile-of-pc style. That's why Intel's top Xeon processor is actually a good bet. As long as it doesn't have the same security bugs. When you're buying 100,000 PCs at a time, you ARE the market. The same goes for virtual data centres, such as Amazon's, or the system Google uses. Google buys vast numbers of commodity machines and plugs them together to form an incredibly high-end distributed system. Does anyone seriously think they're going to switch to a network of mobile phones? Higher cost for less power? That's not Google's approach.

Tenth, Programmers. Nobody programs phones on a phone. Nobody actually hacks the Android OS or the BIOS from the phone itself, simply because if you brick the phone you can't recover. They use a PC to do the programming and then upload, using JTAG where possible to unbrick. And most people program at least a little.

Eleventh, UI. The UIs for phones suck. They're badly designed, they're slow, they're unstable and they have to be that way because there aren't the resources to get them right and Android hasn't been ported to VST anyway.

Twelfth, real-time. Musicians, stage crew, roboticists, scientists, doctors, etc, need real-time. They either use an RTOS (FreeRTOS, VxWorks, RT-Linux) on a platform designed for heavy-duty real-time work (such as a VME crate) or they use a high-end PC that can get close enough. This is not something a Kindle can do.

Thirteenth, in honour of the new Doctor Who, there's SFX. How the hell does anyone think they could run Renderman, Maya or PoVRay on a cell phone? You can't stack up Kindles to create the effects for Titanic! For a start, those were high-end PCs running on a fast network. To get equivalent CPU power, you'd need enough Kindles to FILL the Titanic!

The total number of represented people in all these markets is so utterly overwhelming that if the world did abandon PCs, it would have to abandon civilization. You couldn't meet the requirements of any of them with any handheld device humanity will possess any time in the next hundred years. So unless you build an actual TARDIS, if you want a civilization, you're going to need the devices that keep it running.

Re:Strange article

[Anonymous Coward]

"Shrinking fast" may be hyperbole, but shrinking at all is an existential threat to Intel. It spells price competition twinned with growing technical competition.

At this point, consumers own their smart phone/tablets and will want them to talk to their big screen TV, and "Wintel" is something that only makes sense to grandpa. At the enterprise side, the growth is on the cloud, where its linux run on a VM and no one cares about the chips either. Intel is not necessarily doomed, but we can easily see plausible scenarios where the company withers to literally nothing.

Re:Strange article

[mysidia]

I'm certainly not going to buy a "smart coffee maker" with a subscription & planned obsolescence when I can buy a dumb one without.

The "Smart" coffee makers are called Keurig, and they're pretty popular --- some of the latest models use DRM (Digital Rights Management) technology in the form of a chip in their manufactured coffee pods to discourage/prevent using 3rd party or generic pods.

Re:Strange article

[Anonymous Coward]

And you'll still be able to have *all* that stuff! You'll just pay much more for it than you do today - that's the point.

Re: Strange article

[Anonymous Coward]

It has a BIOS and can boot DOS.

Re:Strange article

[Waccoon]

...people just upgrade more rarely.

Product Activation is what did me in. I used to rebuild my PC every 6 months, just because I could. After XP changed the game and I had my Windows license deactivated twice after upgrades (both times requiring me to call Microsoft and beg to use my PC again), I just stopped upgrading my hardware and learned to live with what I have.

No, I'm not going to switch to Linux (which I've been trying to do for 15 years). Yes, these days I can just use a pirated Windows if necessary. However, the point is that upgrading became more trouble that it's worth, and so I just upgraded when I needed to, not when I wanted.

iPhone CPUs?

[110010001000]

You are attaching too much importance to the iPhone CPUs (and Android) market. It is doubtful the margins are high on those, especially since Apple has multiple manufacturers. That is like saying Apple missed out on making Android phones because there were so many of them out there. You don't want to enter a cutthroat low-margin market.

Re:iPhone CPUs?

[StormReaver]

You don't want to enter a cutthroat low-margin market.

That would be fine if Intel's markets weren't shrinking. However, Intel can't maintain itself on its current markets, as they are all shrinking in favor of Mobile and, to a lesser extent, Cloud. Those are both areas where Intel is not terribly strong.

You would think that Intel and Cloud would go hand in hand, but that isn't necessarily true.

Re:iPhone CPUs?

[Anonymous Coward]

How are Intel and "Cloud" not associated hand-in-hand? Sure, AMD exists (providing very necessary competition) but other than that?

Virtualization in concentrated data centers can reduce physical hardware needs to a degree but this has a lower bound and doesn't continue indefinitely due to overhead and the need to dynamically meet demand. You can idle/power off CPUs to be more efficient when they're unneeded or continually re-purpose them as needed but the physical infrastructure requirement for computing still needs to be in place to meet any increases in demand unless Amazon, Microsoft, Google, etc. are now doing their own processor fabrication on-the-fly inside their data centers.

Re:iPhone CPUs?

[Anonymous Coward]

Even if profit margins are low, they would have been producing additional fab plants and funding additional research into technology.

Re:iPhone CPUs?

[tomhath]

No doubt they would like to be a bigger player in the phone market, but a good company looks for next year's opportunity, not last year's. They're in many things besides CPUs.

Re:iPhone CPUs?

[Anonymous Coward]

They're in many things besides CPUs.

yeah they make PCs, just like everyone else

they make SSDs, just like everyone else

but otherwise? not much!

Re:iPhone CPUs?

[Anonymous Coward]

However, Intel can't maintain itself on its current markets, as they are all shrinking in favor of Mobile and, to a lesser extent, Cloud.

I would say Intel can't maintain itself simply because the current trends indicates ARM will soon reach acceptable desktop level performance. Of course, i7 (i9?) processors may keep their high-margin market for a little more time... but I see i3 and i5s being run over by ARM processors in less than 5 years.

Re:iPhone CPUs?

[shess]

You are attaching too much importance to the iPhone CPUs (and Android) market. It is doubtful the margins are high on those, especially since Apple has multiple manufacturers. That is like saying Apple missed out on making Android phones because there were so many of them out there. You don't want to enter a cutthroat low-margin market.

Also, what if Apple had gone with Intel for their CPU, and then failed because Intel's CPUs sucked batteries dry? Or because having the same CPU in desktops, laptops, and mobile devices lead Apple to the obvious path of cross-platform compatibility, and that sucked batteries dry? Or if Apple wanted to gradually take over more of the system to customize it to better serve their needs, and Intel said "F. U."? I think Apple was probably lucky on Intel turning them down.

Re:iPhone CPUs?

[MightyYar]

You don't want to enter a cutthroat low-margin market.

In general, yes, I agree.

BUT. Semiconductor manufacturing is a very capital-intensive industry. If there isn't enough volume in the high-margin game to keep your multi-billion dollar factories occupied, you will not be able to justify the construction of another. And then you've lost the game, because your cheap-shit competitor will eventually surpass your manufacturing technology that allowed you to charge a premium in the first place. At that point, your best option is to contract out your manufacturing, a la Apple. In 10 years we could very well be buying high-margin, Intel-branded, x86-derived server chips that were produced in a Samsung or even Global Foundries fab. (More likely whatever company they spin their fabs off into.)

Re:iPhone CPUs?

[MightyYar]

Virtualization in concentrated data centers can reduce physical hardware needs to a degree but this has a lower bound and doesn't continue indefinitely

I think the worry is that the lower bound is too low to justify the expensive fabs Intel has invested in. There is a threshold below which Intel chips would cease to be a high-volume but lucrative cash cow and begin to be just another medium-volume chip that cannot justify its own fab and high R&D costs. Once any chip architecture comes rolling out of the same fab, where is the performance advantage? What will happen to their margins?

Re:iPhone CPUs?

[Anonymous Coward]

They asked Intel if Intel would make them CPU based on a non-intel design. Intel said no. This is the innovator's dilemma - they had an existing business that has a certain profitability that came from control of the design and partners (motherboard manufacturers, computer manufacturers and Microsoft). That control allowed them to extract unreasonable profits and run very inefficiently.

The disruptive forces to that business could come at the component level (AMD which they beat using underhanded business tactics and PowerPC which they beat due to economies of scale), at the system level (Chromebooks and Apple PCs and Palm Pilots which never had sizable market share until the iPad and iPhone came out), and at the ecosystem level (Chrome OS, OS X, and Palm OS which also did not have marketshare).

They got beat by a smartphone tsunami that disrupted their business at the ecosystem, system and component level. But as a component player, they could have just said yes to exploring a lower profitability, lower control opportunity with Apple. If the iPhone failed, there was a small impact. If the iPhone was successful, then they would have survived by participating instead of failing by not participating. And they could have always bailed if it didn't grow to be big enough.

Re:iPhone CPUs?

[tomhath]

"not much"? Seriously?

You don't count motherboard chipsets, network interface controllers and integrated circuits, flash memory, graphics chips, embedded processors and other devices related to communications and computing. as "much"?

Re:iPhone CPUs?

[UnknowingFool]

You are attaching too much importance to the iPhone CPUs (and Android) market. It is doubtful the margins are high on those, especially since Apple has multiple manufacturers. That is like saying Apple missed out on making Android phones because there were so many of them out there. You don't want to enter a cutthroat low-margin market.

That might be true if your only concern is margin and not survival. I would say that a case is being made that PCs are slowly dying and being replaced with smaller devices many of which do not and will not run Intel CPUs. I would say it's the same problem that Sun Microsystems faced. AMD, Intel, and Cyrix were all fighting on the x86 market with Intel coming out on top and AMD relegated to 2nd class. Sun stayed out of the consumer market completely and failed to innovate in the server market. These days lower cost Intel and AMD CPUs running Linux have largely replaced Sun.

Re:iPhone CPUs?

[Anonymous Coward]

"That would be fine if Intel's markets weren't shrinking."

Is this actually true? I've still yet to see any evidence that anyone is replacing desktops and laptops with tablets and phones, they're still universally in addition to, rather than instead of. Sure when mobile popped up we saw PC sales declining, but this is more to do with the fact that upgrade cycles have slowed down because no software is pushing PC's anymore. I only just last weekend upgrade my PC, it was 10 years old this year - a quad core 2.83ghz machine with 8gb of RAM and it could still run the majority of games at a decent detail level at 1080p at 30fps+. So it's less that the PC market was suffering to the mobile market and more because the old 3 year upgrade cycle has now extended out to 5, or even 7 years for most people and businesses, this was also largely driven by the recession - companies realised they could stop upgrading from Office 2007 to Office 2010 because it offered no real terms practical benefit to doing so, and in turn upgrade cycles could be reduced saving on hardware costs too. The only reason I ditched my 10 year old quad core 2.83ghz was because it was finally beginning to get slow when I was working with multiple Unity instances + Maya + 3DS Max + Photoshop + Chrome + Visual Studio all at the same time.

Furthermore, it's not as though mobile is growing that much anymore, there's been a lot of indicators lately that mobile is reaching (or at) saturation point with people slowing down their upgrade cycles, and not so many have nots left in the market to join the trend.

Intel had record revenues last year, I see no evidence to suggest there's any merit in the suggestion in the summary that Intel has been at risk for over a decade - it's hard to see how a company that's at risk for a decade can be growing it's revenues year on year to record levels.

So given that it's not losing revenue, because the desktop/laptop/server market really isn't shrinking, then surely a more sane view is that the mobile market is simply a growth opportunity, rather than a threat? Intel still has a rock solid base that isn't going anywhere, and even if Qualcomm and Broadcom try to get into that market it's far more likely that they'll just displace AMD, rather than destroy Intel.

Qualcomm and Broadcom are going to need at least a decade in the desktop/server/laptop market before any real momentum gets behind them, even if they release a chip now no one's going to trust their games and servers on a completely untested new entrant into the market, and with upgrade cycles at 5 years, it's going to take quite some time before any meaningful market share is achieved. Sure they could lower their prices against Intel to try and increase take up, but at best that just means again they'll take AMD's low profit marketshare, and at worst, end up with poor financials as AMD has faced over the last decade such that they're forced to pull back out of the market.

It's also not the first trend Intel has missed (GPUs) and it wont be the last. Anyone with any sense isn't focusing on mobile now, they're focusing on processors required in smart home devices, and IOT in general. The fact is that missing a trend doesn't change the fact Intel has a solid base that just isn't going anywhere, the trends they've missed are growth opportunities and not threats - the only threat to Intel would be entry into the classic CPU market, and if Intel can't break into mobile, why would anyone think mobile chip vendors can break into Desktop/Server/Laptop and take share from Intel? It's a two way street, and Intel has been doing this a lot longer.

Re:iPhone CPUs?

[DarkOx]

Right, Sun is probably the best example. They *thought* they were better off in the high margin, high end micro computer market. The problem is while the margins might have been better the market was shrinking and underwent a rapid shrink; to rapid for sun to effectively respond to once Intel ( and compatibles ) got good enough for a lot of those jobs. A nearly a century before you see the opposite with FOMOCO. There were 10s if not more little automakers around the great lakes region ( proximity to existing steel industry, and chemical industries which depended on the lakes for transport, I assume was why ). They were all building small number of expensive cars by hand. Ford comes along and creates a cutthroat low margin high volume business. They create a auto platform that was "good enough" and the market spoke. Even people who could afford a bespoke car often chose a model A because it had everything they really wanted.

I think we are getting there now, with CPUs at least in the consumer space. First in the high end desktop world you see traditional enthusiasts who are gamers, or video editors etc selecting mid-tier chips like the i5 or using legacy-generation i7 parts (the ticks, not the tocks). They are getting the performance they want without having to pull from the very top drawer. You see a bunch of other people running around with low end business laptops and who seem to be perfectly happy. Though they have better displays and expansion those i3 laptops probably are in terms of CPU performance as compared to a high end i7 machine what a netbook was to a core II duo back when those were introduced. Personally I think netbooks failed mostly due to the displays, and not the anemic CPUs.

Heck a RPi3 makes a reasonable desktop for anyone doing basic office work now; well okay a web browser is a bit slow but that is more a memory starvation issue than CPU for a lot of cases. There may not be much market for non-Xeon parts in a few years.

Re:iPhone CPUs?

[TemporalBeing]

You are attaching too much importance to the iPhone CPUs (and Android) market. It is doubtful the margins are high on those, especially since Apple has multiple manufacturers. That is like saying Apple missed out on making Android phones because there were so many of them out there. You don't want to enter a cutthroat low-margin market.

Same process happened in the 1980's when Intel overtook the Mainframe market, and folks said the same thing. Like it or not, laptops and (even more so) desktops are moving to areas of that generally require high performance stuff. As productivity (LibreOffice, MS Office, etc) and financial applications (Quicken) get better on mobile (Google Apps - Docs, Spreadsheet, Presentation; O365; Quicken is already on iOS and Android) then the need for even a laptop will go away entirely for the every-day-user will be able to 99% of things on a phone or tablet, and laptops and desktops will be left to specialized fields - CAD, A/V Editing and Production, etc - some of which is already starting to move to mobile platforms. (CAD will be among the last to truly go mobile, though viewers are already available so tablets can be used on the manufacturing floors.)

All said, time is *not* on Intel's side unless they figure out how to get back into the ARM market - a market they left shortly before the smart phone market took off.

Re: iPhone CPUs?

[Anonymous Coward]

The hole in your theory is that if Intel gave in and let Apple design the chips and only let Intel fab them, Apple could always move foundries (kind of like how Apple uses both TSMC and Samsung foundries for their current chips). That would mean Intel giving up their profit margin to participate in this new market (the innovators dilemma).

Interestingly Intel did this before. They started in the DRAM biz and successfully made the transition to microprocessors (eg 4004). However since then Intel has become orders of magnitude larger and their investments need higher gross profit margins to support thier existing corporate bulk.

Another Microsoft partner bites the dust?

[Anonymous Coward]

You really wonder why they would turn this down? There was no reason not to provide some mobile processors whilst getting cost coverage from Apple. Was Apple just negotiating too hard for them to justify? I would make a somewhat informed guess that they got told to back off by Microsoft believing that Windows Mobile could outrun Apple. There used to be lists of partners destroyed by Microsoft (like this list). I wonder if they will have to be updated with Intel?

The funny thing is that Intel has a strong history of involvement with both Linux and other alternative operating systems. They had plenty of RISK experience from which they could have built much lower power processors. They could have been very strong in this field. There must have been some strategic decision to avoid it, very like Nokia's strategic decision to avoid Android.

Re:Another Microsoft partner bites the dust?

[UnknowingFool]

You really wonder why they would turn this down?

From what I understand it would be a very different way of doing business. Apple wanted Intel to make CPUs that Intel didn't design. While Intel has made and makes ARM processors in small volumes, the numbers Apple projected would make Intel a chip foundry like TSMC, GlobalFoundries, and Samsung that is more invested in making other people's ICs. It would be like asking Ford to make GM and Honda cars. During WWII, all American automakers made Chrysler designed Jeeps for the war effort but they don't normally do so.

Intel lies, Intel spies...

[Quakeulf]

Intel dies.

Re:Intel lies, Intel spies...

[Anonymous Coward]

and intel can't create a power-efficient cpu without cheating the process and leaving exploitable holes in them.

Re:Intel lies, Intel spies...

[Anonymous Coward]

That's the most truth in all of this so far.

Re:Intel lies, Intel spies...

[Anonymous Coward]

Yeah, and good riddance after they got in bed with Microsoft to force users onto the spying nightmare that is Windows 10 and off of 7/8.1 by not releasing drivers for their 7th & 8th generation processor graphics. I wonder how much Microsoft paid them for that one. Modders have proven that there is zero reason the drivers wouldn't work on the older operating systems and have made them work flawlessly. Completely an artificial limitation.

Was that a hope?

[grasshoppa]

But rather than going to the Feds to try and scuttle the deal through a long and uncertain process, Intel is rumored to be "working with advisors" (in plainer English, the company's Investment Bankers) on a countermove: acquire Broadcom.

Is that what we're hoping for now? Or is it simply that we expect that to happen and are shocked when it doesn't?

Intel is getting nibbled on

[Anonymous Coward]

Intel chose not to worry about ARM.

Intel is getting its breakfast eaten by ARM.
Intel is getting its lunch nibbled on by Qualcomm/AMD
AMD is now tasting Intel's Dinner

Intel BOTCHED the Spectre and Meltdown patches. To the point that I will not apply those microcode pathes, and I will seriously consider AMD in the next build.

Intel seems to be doing the MSFT QA Principle, "let your customers be the beta-testers", except they have moved into the "Alpha" phase.

No thanks, I want a CPU that works and that is secure. It's AMD or ARM on the next build.

Re:Intel is getting nibbled on

[Anonymous Coward]

Might want to reconsider your stance then, AMD was affected by Spectre.

Intel relies on a monopoly

[gweihir]

That monopoly is ironically called the AMD64 architecture today. This comes with a number of problems. While Intel managed to keep AMD small after the last time AMD (not Intel, they did not have the skills) not only came up with the only viable 64 bit extension to the x86 architecture, for a while they also had the fastest CPUs. AMD engineering in the CPU space has basically always been significantly superior to Intel, except for raw speed. Meltdown and Spectre have now nicely illustrated what Intel did to get that speed. And AMDs weakness is over, with a brand-new architecture that is very well designed indeed while Intel has nothing. It helps to understand that Intel it not actually a CPU company, they are a memory company and have struggled with CPUs since they began making them. AMD, on the other hand, came from signal-processors to x86 and _is_ a CPU company. This nicely explains Intel's incompetence, incredible as it sounds. They do not have the right culture.

One other instance of that problem is also that while AMD can do extreme customization of their CPUs since the FX generation, Intel is completely incapable in this space. And just look how long it took Intel to get the memory controller into the CPU after AMD did it.

Now, Intel also did never manage to come up with anything x86 that was suitable for a smartphone. AMD did not even try, because they understand CPUs and knew this architecture is not suitable for that field. But they went one step farther: They have server processors that include ARM cores. So AMD has real experience in that field, but Intel is, again, lost. Yet AMD is far smaller and does not need the smartphone market to survive, while Intel likely does. And they messed it up.

My take is that finally Intel found out with much delay that they managed to screw themselves, in addition to their customers.

Re:Intel relies on a monopoly

[jeff4747]

AMD (not Intel, they did not have the skills) not only came up with the only viable 64 bit extension to the x86 architecture

You're leaving out a rather important detail: Intel didn't try to create a 64 bit extension to x86.

Instead, Intel tried to use Itanium and IA-64 to replace x86 and all the cruft in it that had built up over the years. Intel thought people would only buy Itanium for servers, since a 64-bit address space wasn't very useful for desktops at the time. So they priced their chips high.

AMD countered with 64 bit extensions to x86 and cheaper chips.

Cheaper won.

Re:Intel relies on a monopoly

[Bearhouse]

Yes, but it didn't help that Itanium was a honking disaster in the marketplace, (apart from HP, who tied themselves to that turkey...)

Re:Intel relies on a monopoly

[DCFusor]

MIPs/Joule is also AMD fail. Otherwise, I agree. It's also a big reason Intel failed in the portable market. Fast, but HOT. ARM beats both in that niche, which took over more market than Intel and AMD could cope with easily. Not all questions have only two possible answers.

Re:Intel relies on a monopoly

[Anonymous Coward]

AMD did try to create a low power x86 with DSP extensions appropriate for such a market. It was via a skunkworks company named Stexar.

Unfortunately, the ATI acquisition and contemporary price crash on x86 at the time made it look undesirable to continue development. In hindsight, a major lost opportunity as the smartphone market took off very shortly afterwards.

Re:Intel relies on a monopoly

[Eravnrekaree]

Intel didnt do a 64 bit extension at first because they were invested in Itanium. They did have the skills to extend x86 to 64 bit, its not a very difficult thing to do. Itanium never caught on, it provided too difficult for compilers to make code that was well optimized for it. Backwards compatibility won out.

You are correct that Intel played dangerous games and sold a defective product in order to give themselves a speed advantage over AMD, which did the right thing by its customers.

Re: Intel relies on a monopoly

[Anonymous Coward]

You're leaving out a rather important detail: Intel didn't try to create a 64 bit extension to x86.

Instead, Intel tried to use Itanium and IA-64 to replace x86 and all the cruft in it that had built up over the years. Intel thought people would only buy Itanium for servers, since a 64-bit address space wasn't very useful for desktops at the time. So they priced their chips high.

AMD countered with 64 bit extensions to x86 and cheaper chips.

Cheaper won.

You are leaving out the plethora of other reasons that caused Itanium to fail. Cost was one. The sheer incompetence of the project from architecture to silicon implementation were bigger reasons.

Re:Intel relies on a monopoly

[Eravnrekaree]

You contradicted yourself here. On one hand you said x86 is "cruft", on the other hand, you admit its cheap to make x86 chips. There really isnt cruft in x86, its a perfectly useable design. Its not hard or expensive to implement any more than ARM. x86 instruction encodings can be weird, but not being aesthetic doesn't make them a performance problem or hard to implement on chip. Instruction encodings are things compilers need to be concerned with, not app programmers, anyway.

Re:Intel relies on a monopoly

[AlwinBarni]

... AMD engineering in the CPU space has basically always been significantly superior to Intel, except for raw speed...

Some time ago I did basic performance tests for AMD (PC), Intel (PC) and PowerPC (Mac) CPUs of similar class. The test was designed to check floating-point unit (povray) and overall performance (compiling a large project), and AMD beat Intel significantly in povray, whilst lost in compilation. Later tests showed that compilation was strongly correlated to the cache size (my testing machine with AMD had half the cache of the one with Intel CPU). PowerPC lost in both categories.

Re:Intel relies on a monopoly

[Anonymous Coward]

"Now, Intel also did never manage to come up with anything x86 that was suitable for a smartphone. AMD did not even try"

Not quite true, AMD had Elan processors, x86. Although they bought them from some company and developped them from that. They were used in Nokia Communicator for example.

Re:Intel relies on a monopoly

[Anonymous Coward]

You contradicted yourself here. On one hand you said x86 is "cruft", on the other hand, you admit its cheap to make x86 chips.

(1) Intel charged a premium on the ia64 line so it wasn't about "cost" per se. (2) Cruft doesn't mean something is substantially more expensive to make. It does tend to limit potential performance in the future and complicate the design substantially.

There really isnt cruft in x86, its a perfectly useable design. Its not hard or expensive to implement any more than ARM.

That's really apples and oranges. Modern amd86 processors are more expensive to develop and produce even with the small manufacturing size countering a lot of that. This is because amd64 are performance based and ARM processors can be built with older, cheaper fabs because they're more focused on performance/watt. There's also the point that ARM licenses out production and so their position is different from Intel and AMD as far as chip production.

x86 instruction encodings can be weird, but not being aesthetic doesn't make them a performance problem or hard to implement on chip. Instruction encodings are things compilers need to be concerned with, not app programmers, anyway.

And micro op decoders. The real saving grace for Intel was amd64 doubling the registry count to combat stalls for a lack of registers--register renaming can only do so much. What Intel has pushed because of its clusterfuck of an instruction set is a very disjoint execution unit cluster where there are many missed opportunities for parallel execution of instructions. Ie, they've spent a lot of effort overengineering and working around the limitation of the instruction set instead of a much simpler design that would have very consistent EUs and for which the main effort would be merely to increase cache hits and have the compiler possibly do a better job interleaving instruction blocks or pushing more hyperthreading like designs.

It's really impressive what Intel has done, and it's possibly that ARM won't be able to compete on the high end ever because Intel is in the lead and is constantly pushing to be on top, but let's not pretend that they're not handicapped with the instruction set in their designs.

Re:Intel relies on a monopoly

[gweihir]

AMD (not Intel, they did not have the skills) not only came up with the only viable 64 bit extension to the x86 architecture

And if you actually believe that, then you are stupid. Of course they tried. They just never had anything good enough to go public with.

Re:Intel relies on a monopoly

[swb]

I think they thought they were going to be sitting in the catbird seat. I think Itanium was supposed to replace PA-RISC *and* PC server processors. So they would have a long-haul future with a new CPU for both PC server and workstation/midrange markets and would be able to start picking off Sun's business, too.

I'm sure there was some MBA math involved that also took into account getting a share of licensing revenue for the chip patents, too.

You have to admit that looking back it didn't seem like a terrible idea.

Re:Intel relies on a monopoly

[Anonymous Coward]

Intel knew shortly after the release of Itanium that Itanium would not be the future, and that a more conservative 64 bit extension was needed. They saw that AMD was already doing pretty much the same thing they would have done, and they knew that getting Windows ported and validated on a new instruction set costs about $2 billion, since they had just paid that for Itanium Windows. So, they decided to let AMD win on purpose, because then AMD has to pay the $2 billion bleed, which is a much larger sum for them. After AMD won, Intel already had an compatible implementation that was faster than AMD ready and filled their supply chain with it in about a year. AMD's R&D budget was still taxed by the $2 billion Windows porting expense, and they weren't able to spend fast enough to counter Intel's unexpectedly strong execution.

Re:Intel relies on a monopoly

[Kjella]

Intel didn't try to create a 64 bit extension to x86.

Yeah, Intel was trying to pull an IBM and introduce the MCA architecture to get rid of the competition and over on their shiny new platform where they held all the essential patents.

Instead, Intel tried to use Itanium and IA-64 to replace x86 and all the cruft in it that had built up over the years.

That's one way of putting it. Somebody at Intel managed to do a huge sell-in of compiler optimization and profiling as the future and so they created "Explicitly parallel instruction computing (EPIC)" which gave extremely fine detailed control to the compiler. The problem is that there's a balance between run-time optimization based on the data and static compiler optimization so without a lot of iteration on realistic workloads performance was not that great and sometimes "let's just cross that bridge when we get there" was a superior strategy. And they discovered alternatives like hyper-threading where you let some low-intensity threads use those in between moments, better branch predictors etc. so the ideas of Itanium pretty much died and never came back.

The AMD64 instruction set was more the "let's keep the house but clear the cruft", particularly they added more general registers which was a really big deal. AMD had done real experiments with more addressable registers but found that it was better to only increase it from 8x32 to 16x64 bit and instead use additional registers for run-time register renaming. That offset most of the performance penalty of doubling pointer sizes by itself, so x64 mode was almost free (though with a big higher memory usage for pointer-intensive work) whereas EPIC required a full do-over and struggled massively to emulate x86. Itanium was already struggling when AMD kicked it in the nuts.

Re:Intel relies on a monopoly

[Anonymous Coward]

AMD also one the 1GHz race with Athlon.

Re:Intel relies on a monopoly

[Junta]

Cheaper was not the only, or even necessarily the primary reason.

Intel forgot the whole reason they ultimately dominated the market: backwards compatibility. Every product launch, no matter how revolutionary had a *massive* catalog of functional software to go with it.

Intel had enough hubris to think they could spin up an ecosystem from scratch. AMD proved that to be an incorrect strategy.

Also, the chips were plagued with performance issues, promiment among them was placing more demand on memory bandwidth than x86 counterparts. AMD's opteron sported massive memory bandwidth *and* less relative demand on the memory subsystem.

This was also about the same time Intel was screwing over the consumer market with the Netburst architecture. It's probably a good thing that Intel had their eyes on Itanium and let AMD define the archtecture. *At the time* Intel's engineering was a mess. Their engineering got better and now are very formidable in the CPU, but around 2000, Intel's engineering was woefully bad.

Re:Intel relies on a monopoly

[Anonymous Coward]

Which PowerPC? The G4 head decent cores, but completely lost due to slow front side bus, coupled with relatively small caches. G5 were certainly competitive up to Core2Duo. And current Power9 are real beasts (up to 96 threads in a socket with ~100MB of cache).

Re:Intel relies on a monopoly

[jeff4747]

There really isnt cruft in x86, its a perfectly useable design

We've figured out better ways to make CPUs than x86, and one of the ways is better instruction sets. Those instructions sets reduce the complexity in creating a chip. x86 can't do that because it has to remain compatible.

In other words, x86 is carrying along things that are not as useful as they could be, and maintaining them is a limitation. Also known as cruft.

Re:Intel relies on a monopoly

[Anonymous Coward]

the main idea was to shift the responsibility of instruction scheduling to compiler. that was the 'hp' part of the job, as they had experience with c and c++ compilers (and of course cpu design experience). intel had almost none.
so in order to achieve the performance you had to rewrite your application quite hard and it was as much exotic as ps4 (cell spu) i'd say, so the learning curve was quite steep.
in general dynamic workloads it wasn't as excellent as they thought, so they tried to save the performance with adding OOO in itanium2. too little, too late.

i think intel suffocated itself here. in desire to solve the cpu problem once and for all they wanted all the 'bestest' ideas and designs in the chip and that was the cause of final complexity of the chip (and of the compiler). no wonder the crowd went crazy for 'the evil you already know' in form of x86 extended for 64 bits.

btw: the vliw architecture is not dead, Ivan Goddard uses it in his Mill cpu design. worth checking.

Re:Intel relies on a monopoly

[Anonymous Coward]

And AMDs weakness is over, with a brand-new architecture that is very well designed indeed while Intel has nothing.

Holy shitballz man. I like AMD as much as the next man. Both of my main desktops at home are AMD, but to claim Intel has nothing is ridiculous.
The top of the line i9 handily beats the top of the line threadripper.
I'm looking to replace my FX 8350 with an i7 8700 instead of a Ryzen 1800X because it's a comparable performer, cheaper, and only uses 65 watts. AMD have kept Intel honest and don't need my support anymore so fuck it, I'll give Intel a run next most likely.
I'll probably wait till they fix Meltdown and Spectre in silicon first though.

Re:Intel relies on a monopoly

[Prien715]

AMD countered with 64 bit extensions to x86 and cheaper chips.
 
...and faster.

  Working in the cash rich Oil and Gas industry in the early 2000's, SGI tried to make a machine using the processors. A couple customers bought them, but realized after just a little bit of testing that they were horrendous for single threaded tasks and even crippled for larger ones. We ran some benchmarks in-house and couldn't find anything the machines were good at -- other than the huge memory footprint (3 TB of RAM is still massive even a decade later) -- compared with commodity AMD64 boxes.

Re:Intel relies on a monopoly

[Anonymous Coward]

I'm not terribly impressed with Intel. Their latest CPU's aren't much faster than Sandy Bridge which is now close to 7 years old. Their only real recent innovation is increasing the core count. The lower power usage of course is also nice, but the increases in single thread performance is not impressive.

Of course, if you stagnate for 7 years, that gives AMD plenty of time to play catch-up.

It's funny

[ausekilis]

Intel had a Wintel phone some years ago that was actually really quick and responsive. Plenty of power to multitask and do whatever on your phone. However, ARM continues to utterly destroy x86 on power consumption.

Now it may be too little too late, unless they are somehow able to get that consumption better and maybe move toward tablets/phablets.

Re: It's funny

[kenh]

Intel CPUs are in$60 Windows tablets - they are in the tablet market, just not the Android tablet market.

Re:It's funny

[Anonymous Coward]

However, ARM continues to utterly destroy x86 on power consumption.

ARM has cheaper idle cycles. In every other way, x86 chips are more power efficient. This means ARM works great on things like phones and smartwatches, because over 95% of the time you won't be doing anything (at modern ARM CPU speeds). The architecture is arguably superior for certain types of server task as well, web hosting seems like a viable low-logic use with lots of demand.

Intel should forget about Mobile

[pablo_max]

Seriously, Intel needs to get out of the mobile chipset game because they are pretty shit at it.
I have been in the mobile certification business for a long time. We do 10's of thousands of tests on protocol stack and hardware layers of modules integrating these chipsets. Intel based products are always a pain. Their support is crap too. Most say, ok.. never again with Intel. We will use Marvell or something. QC tends to be 4 or 5 times the price, so it often doesnt make sense for high volume, low cost stuff.

Anyhow.... they started way too late in this game and missed the boat.

Re:Intel should forget about Mobile

[Anonymous Coward]

We will use Marvell or something.

Maybe DC?

How is Nokia destroyed, exactly?

[Anonymous Coward]

The bulk of the Nokia engineers are now working at HMD Global, a.k.a "the new Nokia", still in Finland, and their new line of Nokia Android phones and feature phones have generated ~80 million sales in their first 12 months alone.

Far stretch from having been "destroyed".

Re:How is Nokia destroyed, exactly?

[Dog-Cow]

People who were employees of Nokia are not Nokia.

Re:How is Nokia destroyed, exactly?

[Anonymous Coward]

How come they did not bring with them that awesome piece of technology called Symbian?
That was so GREAT!

Re:How is Nokia destroyed, exactly?

[Anonymous Coward]

People who were employees of Nokia are not Nokia.

Whilst you are of course correct, in many ways they can be,

The folks who make up a company, are far more the company than anything else. If you have a sufficient percentage of employees to also transfer the culture, then you effectively have the company. I would argue that this is far more important than product lines (which can be recreated), or intellectual property (equivalents of which can be found). I note that HMD Global seem to have retained the Nokia brand name for phones, which is good for communicating the continuity of culture. Good luck to them.

Re:How is Nokia destroyed, exactly?

[Anonymous Coward]

Nokia is still very much around, just not in the very visible mobile phone sector anymore, but still leading the mobile phone infrastructure behind the scene. I don't see them losing that position any time soon.

Re:How is Nokia destroyed, exactly?

[Anonymous Coward]

The bulk of Nokia was Nokia Networks and other enterprise businesses, which are still very much alive. They just sold off Mobile, a few patents and some time-limited trademark rights.

Lazy analysis

[kenh]

Seriously, because Intel declined to build CPUs for Apple in 2005 they are ever so slowly going out of business?

That's just stupid on the face of it - it displays a level of ignorance of the original poster, boing the entire industry down to his personal purchasing decisions and a random fact (billions of iPhones with non-Intel CPUs!).

History

[DrYak]

Now, Intel also did never manage to come up with anything x86 that was suitable for a smartphone.

Worse.
They did never manage to come up with anything specifically running the x86 instruction set that was suitable for a smartphone.
They used to have a decent Intel-manufactured CPU running ARM instruction set, but somehow managed to abandon the market and sell it off, just at the time when ARM is getting even more relevant thanks to smartphones, routers and IoT.

Search for "Intel StrongArm" and "Intel XScale".

Note that, according to Wikipedia, Intel is still in possession of ARM license that they acquired when bought StrongArm.
So even after selling XScale out to Marvell, they could still start a new line of ARM core *now*, after having come to realization that the Atom doesn't scale down as much as they would have liked (isn't that well suited for smartphones and routers) and its x86 compatibility makes absolutely not sense in those markets (Seriously, nobody is going to run legacy Windows code on a smartphone)

AMD did not even try, because they understand CPUs and knew this architecture is not suitable for that field. But they went one step farther: They have server processors that include ARM cores.

I'm still hoping that, next to the ARM light-weight servers that they are targeting, these ARM cores will eventually also evolve to some high range phablets and dev boards.

Re:History

[Zorpheus]

(Seriously, nobody is going to run legacy Windows code on a smartphone)

Wait, that sounds like a great idea to me. Why not run all our windows apps and games on a smartphone? There are still things that haven't been ported. Sure, it is not good for regular use. But if I could for example quickly check something in a Visual Studio project or make a small edit to an image in Photoshop on my phone it would be really great.

Re: History

[Anonymous Coward]

The few people that need to do that have TeamViewer on their phone...

Re: History

[Anonymous Coward]

Ok, also a good idea, maybe I should set that up ... just need a pc running all the time then

Re: History

[Anonymous Coward]

Then you will have to deal with an interface that's completely inadequate for a mobile device. Or you could use a graphics editor that's properly designed for a mobile system, like ArtFlow, Ibis Paint, Sketchbook, PaperOne and so on.

Re:History

[epine]

Now, Intel also did never manage to come up with anything x86 that was suitable for a smartphone.

I went though a short, thirty-year obsession with all things microarchitecture. The appalling stupidity of accepted memes in this space I'll surely carry to my lonely grave.

Crufty x86: here's how it broke down.

First, about 50% of the original cruft drank the shrink-me fluid, and shrank down so small you can barely see it now (e.g. some extra microcode entries in a rarely used, unpopular spiral annex of the instruction decode table for misbegotten 286-era CISC call gates.) Jesus, people, exponential happens.

Second, about 25% of the cruft turned out to not nearly be so crufty as legend would have it. The RISC camp soils itself over the read-modify-write instruction group. But generating a complex address once (yes, x86 does complex address generation within the context of a single instruction) rather than twice alleviates substantial pressure on the address look-aside unit. It's also a very handy and compact addressing mode for minor stack spill (e.g. function variables that don't quite manage to stay in registers all the time). With a 30% instruction encoding density advantage over the original ARM32, you need many fewer transistors in your i-cache to achieve the same i-cache hit ratio. The bigger your caches, the more free transistors to apply elsewhere. x86 is still a bit short on registers despite rmw, but you gain a bunch of this back on lighter context switches, so it's not a complete write-off.

The other 25% is an eternal pain in the ass. Here's how the PITA component breaks down. The majority of it has little impact on peak throughput at all, but it comes at a thermal efficiency cost. The thermal cost is mostly irrelevant if you are sucking juice from a wall socket, and your processor is not hitting the thermal wall. The other side of this is a hideous sunk-cost in the engineering trickery required to pull this off (for a company the size of Intel, however, hideous is mostly peanuts, and nice barrier to entry you've got there, shame if a different device category became prominent).

A minority of the PITA aspects of the instruction set are just permanently a PITA. Deep OOO requires extensive hazard detection, and x86 has hazards up the wazoo (many partial register writes, and seventeen different flavours of flag register update subsets). This costs silicon, this costs power, this costs cycle time, this costs pipeline stages. Lose, lose, lose.

Considering the architecture is now 40-years old, that's not exactly a resounding F on the old report card, by a sane grader.

Because of aspects like instruction decode alignment (with those blasted variable prefix bytes) and extremely complex hazard detection x86 is just always going to produce twice as much heat arriving at the same result 20% faster than any reasonable design that was originally power conscious.

I suspect most of this fixed thermal inefficiency resides in the front end and not the back end. Meaning that an alternative x86 instruction set could be devised (somewhat more drastically different than Thumb-2 vs. Thumb) with vastly more efficient instruction decode (thermally) and vastly fewer implicit scheduling hazards. Caches, register sets, dispatch pipelines, retirement unit, memory ports, execution units, these could all remain the same. Perhaps the only register you'd want to muck with is the flags register, and maybe you'd trash the ability to write to AH (though you'd probably keep partial register writes to AL to handle common byte operations).

[*] Fifteen years ago, the ugly details of this stuff was more in my head, so my examples predate AMD64, but mutatis mutandis.

Maybe by doing so you'd even close the gap enough to compete with ARM. But: a huge redevelopment and validation cost (what, me validate?), another substantially different code generation mode for every major compiler, another by

Re:History

[Anonymous Coward]

How could all these be usable on a screen so small? That would require more than just CPU compatibility.

Re:History

[TJ_Phazerhacki]

Long, well informed comment with references, jokes, and asides? 5 digit UID? Brings me back to the glory days of /. when I was browsing using my dad's account...

Shed no tears for them.

[Gravis Zero]

Intel has a long history of anti-competitive behavior. One needs only search "Intel anti-competitive behavior" or see their Wikipedia page to recognize that it's a persistent and ongoing. Yes, they have brought advances to the semiconductor field but they have always behaved in the most unethical manner possible to subvert the competition.

I look forward to the rise of AMD.

Re:Shed no tears for them.

[Nemyst]

If you think Intel's demise is at the hands of AMD, you're deluding yourself. Intel's scared because the entire x86 market is shrinking and they don't have a presence in other markets. AMD is in the exact same spot while being substantially smaller and dragging along a seriously hurt GPU division. I'd be delighted to see more competition in the x86 space and Ryzen will certainly help, but that's not what Intel's concerned about here.

Re: Shed no tears for them.

[Anonymous Coward]

Like I lol forward to the year of Linux on the desktop!

Re:Shed no tears for them.

[Anonymous Coward]

AMD's GPU division isn't hurt. It's basically the only player in home consoles and they can't make discrete units fast enough to meet demand.

Operate standalone

[sjbe]

And wouldn't the always risky move of combining two cultures, employees, and physical plants introduce an even greater peril?

Depends on how they handle it. If they operate the acquired company as a stand alone entity (sort of like how Berkshire Hathaway operates) then the cultures don't really have to mix much at all and that can work fine. Mixing company cultures is a serious challenge but it's not always required.

I think Intel's biggest challenge is that they've been a de-facto monopoly for so long that they seem to have forgotten how to compete in areas where they don't dominate. It's always a risk for company that has one big cash cow that they just milk it to the exclusion of all else. The biggest risk to Intel is software makers leaving the X86 platform which is where the vast majority of their revenue comes from. They make some money from IoT and flash memory and security but these are about 12% of their revenue and 7% of their profit combined.

Most revenue from X86

[sjbe]

However, Intel can't maintain itself on its current markets, as they are all shrinking in favor of Mobile and, to a lesser extent, Cloud.

Well Intel supplies an awful lot of those CPUs for the cloud so I don't think that worries them so much. Mobile is an issue for them because that is definitely where the growth is. The biggest threat to Intel is that they have so much of their revenue and profit tied up in the X86 platform. If software and PC makers continue to migrate away from X86 it's going to hurt Intel badly sooner or later.

Re:Most revenue from X86

[Anonymous Coward]

I don't think ARM or AMD are going to be the next big threat to the Intel monopoly. Even they will be supplanted. It's going to come from left field, specifically from a company that is researching the next generation of computing. It will come from researchers who find a way to make processors from inexpensive materials that are easy to manufacture. Silicon is cheap, but the manufacturing process is expensive and prone to defects from manufacturing. What if, for instance, a researcher finds a way for a biological process (bacteria, for instance) to lay down traces according to a pattern and "grow" a CPU? It's not entirely beyond the pale, and with CRISPR an other GM techniques, it wouldn't be impossible to imagine that the next generation of processing could come out of a petri dish with a component cost of sugar water, silica and trace minerals.

Re:Most revenue from X86

[TemporalBeing]

However, Intel can't maintain itself on its current markets, as they are all shrinking in favor of Mobile and, to a lesser extent, Cloud.

Well Intel supplies an awful lot of those CPUs for the cloud so I don't think that worries them so much. Mobile is an issue for them because that is definitely where the growth is. The biggest threat to Intel is that they have so much of their revenue and profit tied up in the X86 platform. If software and PC makers continue to migrate away from X86 it's going to hurt Intel badly sooner or later.

Even Cloud providers are looking to move to a different platform and many are involved in OpenPower (https://openpowerfoundation.org/membership/current-members/) to find better hardware in terms of cost per watt for Cload loads where floating point or integer computational performance may not matter as much.

This article is nonsense

[lamer01]

Intel has a stranglehold on server CPUs. AMD is making a comeback there but the ARM camp does not have compelling enough solutions in that space. Low power is great but it's not everything.

Re:This article is nonsense

[njahnke]

ARM camp does not have compelling enough solutions in that space.

Check out these benchmarks.

https://blog.cloudflare.com/arm-takes-wing/

Looks like the situation is rapidly deteriorating for Intel.

Intel did not turn down Apple

[locketine]

Steve Jobs wanted Intel CPUs in his iPhone but his engineers did not. Anyone who's familiar with the differences between ARM and x86 would know that an Intel powered smartphone was not a good idea.

  Intel tried to enter the smartphone market five years ago and failed due to glitches, power consumption and incompatibility with existing apps. If Steve Jobs had gotten his way, the smartphone revolution may never have happened.

Re:Intel did not turn down Apple

[Eravnrekaree]

From what I've heard from engineers is that there is very little or no overhead to support x86 on a chip. Its an infinitesimal part of the design. Its basically one of those urban legends that was cooked up in the 80s and isn't relevant any more. We've been in this situation many times before, people hear something and they keep on repeating it even though things have changed. I've looked at both x86 and ARM ISAs and its not much difference in complexity. We are in a post-RISC era. Even the so called RISC chips are complex with extensions for things such as SIMD instructions. ARM is not RISC, its a complex set with many extensions. With the engineering problems of increasing clock frequency, the chip has to be sped up by having a larger instruction set with more functionality loaded instructions which can allow for smaller code, allowing more to be done with each clock cycle. The worst kind of ISA to implement on a chip is one with a very small instruction set, such as stack based ones such as WebAssembly would be a very poor choice.

Re:Intel did not turn down Apple

[UnknowingFool]

Steve Jobs wanted Intel CPUs in his iPhone but his engineers did not. Anyone who's familiar with the differences between ARM and x86 would know that an Intel powered smartphone was not a good idea.

I don't think that's what the proposal was. Apple has tried different device prototypes with Intel CPUs but they didn't work out. I think the Apple proposal was that Intel manufacture ARM CPUs for Apple.

Re:Intel did not turn down Apple

[Anonymous Coward]

You are of course correct that the difference in power consumption between ARM and Intel x86 chips isn't primarily due to instruction set. But ARM chips, which are used in phones and watches, do typically have better power consumption than Intel chips, which typically aren't. Intel did have some chips which were almost competitive in power/performance but it wasn't enough to overcome ARM's other advantages.

I agree a stack machine wouldn't be a good choice of ISA for a modern phone. WebAssembly is close to being an ISA (well not really, it's a serialised AST) but it's an ISA for a register machine, not a stack machine. It uses static single assignment form (SSA) - what most compilers use as an intermediate form - which assumes an unlimited supply of registers. WASM implementations have to allocate registers during codegen.

Re:Intel did not turn down Apple

[Anonymous Coward]

This states that WASM is not SSA, though it can be decoded to a compiler's internal SSA form.

While it's not exactly a competitor, the Mill architecture similarly provides an abstraction of the physical ISA that is more amenable to compilers. However, the Mill hardware itself is fundamentally SSA in nature, and naturally extends to a generalized form, presenting a genuine SSA target for compilers.

20 years later...

[Anonymous Coward]

Intel has a long history of buying companies that succeeded where they couldn't. 20 years ago they had a 'mergers/acquisitions' team (M&A). They weren't getting the market share they wanted, so they purchased several networking companies, including Level One Communications (at the time big Broadcom competitor.)

Was it successful? Not long after the M&A team added a 'D' for divestitures...

Whatever Broadcom is doing differently that makes them successful, that is the first thing Intel will change.

Intel needs Windows 10 to exist

[Anonymous Coward]

Until the developers at GNOME and KDE figure out how to make a desktop that works for the 99% and on an architecture other than x86 (raspberry pis don't count), Windows 10 and Intel will dominate in the corporate world forever.

As the largest player, that's where you should go

[SuperKendall]

You don't want to enter a cutthroat low-margin market.

As one of the largest players in the CPU space, you absolutely do want to do that.

First of all, low-margin does not mean NO margin, and a billion of anything at low margin is still a lot of money.

Secondly, that is a lot of great R&D opportunity in a challenging space you are giving up to sone other company. You can sit around all day designing new processors or features but until it comes into contact with real world uses and needs, your design will lack the coherence it needs to attract other buyers. ARM has had tremendous uptake because its evolution has been guided by the fiery furnace of having to provide real working CPU's for so many mobile devices across multiple generations.

Lastly - if you are the high volume producer you are SUPPOSED to be able to go for low volume business because you have built to work at low margins already. The production capacity you have is supposed to give you higher margins than other smaller companies could manage when selling a ton of whatever, so it shouldn't even be that low a margin if you are one of the few companies that can produce billions of something.

Acquire?

[AlanObject]

Wasn't Broadcom already eated by Avago?

Re:Acquire?

[radarskiy]

After Avago bought Broadcom, it changed its name to Broadcom.

"fighting"? really?

[ravrazor]

Intel stock is up 50% in the last 12 months (to $50) and they made about $63 billion dollars in 2017.

I think they're doing okay.

Re:"fighting"? really?

[Nemyst]

The headline is obviously sensationalized, but Intel's always been pretty forward looking in their planning. They have to act now if they want to face challenges 5 years down the line as ARM takes more and more marketshare. When you're dealing with CPU designs and fabs, you can't turn on a dime.

Re:"fighting"? really?

[roc97007]

And "now" seems almost like a desperate attempt at catch-up. I think Intel should have made their move a few years ago.

Re:"fighting"? really?

[iggymanz]

ARM is low performance, there is no threat to intel's main markets.

amd epyc has the pci-e for storage without

[Joe_Dragon]

amd epyc has the pci-e for storage without needing to cross-flash / reflash to IT mode or lot's of pci-e switches.

Changing market

[Anonymous Coward]

Intel did succeed in winning the Mac market. But Intel did not win a large portion of the UNIX market, because it eschewed UNIX. HPUX, AIX, Sun SPARC, Motorola, all these UNiX platforms were non-Intel. Intel lost out on the RISC market to a large extent. But remember that Intel, along with DEC, contributed to the demise of the IBM mainframe market.

Intel also lost out on the low heat, low current, low cost, portable, small device market. The logic goes: it doesn't make sound business sense to sell a processor for pennies, when you can sell a more powerful processor for $250.

Intel chose its niche in the marketplace, partnering with Microsoft. Now that market has changed and Intel is holding the bag. Some free clone of UNIX is driving those billions of small devices.

IBM lost out on typewriters and adding machines, Kodak lost out on film, Zenith lost out on TVs, GE and Westinghouse lost out on the incandescent bulb. The maelstrñom in the marketplace is forever changing.

My CPU history:

[Anonymous Coward]

MOS(6502-2Mhz),
MOTOROLA 68000 7Mhz
INTEL 486DX2 66Mhz
INTEL 486DX4 100Mhz
INTEL Pentium 166Mhz
INTEL Celeron 300A OC 450Mhz
INTEL Pentium 3 700Mhz
AMD XP1600-1.4Ghz
AMD XP32000-2,2Ghz
INTEL Core 2 Duo 2,66Ghz 2 CORES
INTEL I5 3470-3,2Ghz 4 CORES
AMD Ryzen 7 1800X- 3,6Ghz 8 CORES - 2 weeks old

For me they already lost it. I will be looking at them again in 5 years.

5G and "always online" netbooks as a threat?

[boa]

5G is coming soon with big promises about speed and availability. Always-online netbooks is a thing already. Maybe the next generation netbooks will use a non-Intel CPU to save both production costs and power?

If people replace their PCs with a new Internet-enabled device(netbook, glorified cell phone, or something entirely new), sales of Intel CPUs will drop. A lot. It may be the death of both Intel CPUs and Windows OS.

All hail Android? All hail ARM?

Re:5G and "always online" netbooks as a threat?

[Anonymous Coward]

Highly unlikely since data caps are still a thing in the US. It's also the same problem I have with Chromebooks (before they could run Android apps): I don't want to depend on an Internet connection that could potential fail or be spotty. When the device's primary purpose is to watch online video, it's okay, but anything that can be done offline should have the capability to do so.

Re:5G and "always online" netbooks as a threat?

[boa]

I feel the same way, but people seem to be happy with their cell phones, data cap or not. BTW, in some countries, like Finland, there's no data cap. If 5G delivers what it promises, throughput should not be an issue. And Wi-Fi isn't dead yet either :)

Huh?

[Anonymous Coward]

So Intel has half of the worldâ(TM)s FPGA market, they have a massive part of the networking component market, they have massive parts of the Flash market, they have massive parts of the chipset market, they have massive parts of the SoC market...

Oh... Iâ(TM)m pretty sure Intel also has some fabrication abilities as well.

In fact, Intelâ(TM)s CPU market is certainly considerable and if Broadcom purchased Qualcomm, they would make a mess of it as they have of so many other acquisitions. Broadcom has never managed to build a proper developer infrastructure around any of their products. The whole Raspberry Pi thing happened almost despite of Broadcom. Hell they donâ(TM)t even manage driver downloads worth a damn.

I think the best thing that could possibly happen to Intel would be Broadcom absorbing Qualcomm. It would be like HPE buying someone out.

Intel's Future is pretty solid

[foxalopex]

Intel isn't likely to go away anytime soon. They have some of the most advanced chip fabrication facilities in the world. Even if they didn't make x86 Processors, other manufacturers would be lining up to get their chips fabricated at their plants. They've also expanded into other areas although their x86 market remains to be their cash cow.

Now the only grudge I've got against Intel is their massive anti-competitive behaviour against AMD in the past but nowadays they're one of the few companies that provide full opensource access to their GPUs and they generally do produce excellent mobile laptop chips.

Re:Intel's Future is pretty solid

[Anonymous Coward]

Perhaps you should add the IME to your list of grudges, because it weights heavily against any positive behavior by Intel, and is potentially even worse than their anti-competitive behavior. The day when Intel is reduced to a mere chip foundry can't come soon enough.

The world desperately needs secure hardware that isn't riddled with black boxes and potential back doors. With respect to this, Qualcomm+Broadcom are almost as bad.

Can Intel even play in this market?

[roc97007]

This is a real question. I don't have anything against Intel, and my current workstation has Intel Inside.

Does Intel have anything that plays well in the phone/tablet market? My understanding is that Qualcom and/or Samsung don't own the market just because they were there first, but because their products are designed specifically for the application, whereas Intel's offerings in that arena all appear to be relatively low power x86 chips. Key term being "relatively". Like Microsoft's early struggles with hand held devices, trying to shoehorn a desktop OS into something with a 4 inch screen, Intel appeared to be trying to leverage existing designs in a market where they weren't appropriate.

I could be missing something, but it seems like Intel's largest current issue is that they make the best possible processor for an increasingly smaller market, and don't make anything particularly appropriate for the most aggressively expanding markets. An issue they share to a certain extent with Microsoft.

It'll be interesting to see what happens should Intel acquire Broadcom. I think there's a good chance -- maybe 40% that after acquisition Intel will drop or severely de-emphasize Broadcom's SoC products in favor of one of their lower power laptop x86 processors. And fail miserably at it.

I don't know, we've seen RISC vs CISC before

[lamer01]

ARM is RISC with some extensions...RISC was always touted as better but CISC always won out. Intel has putzed around for the last 5-7 years as they did not have a real competitor in the CISC world. I think with Epyc being as good as it is, it will become a race again.

Rockchip Intel Atom with ARM Mali Graphisc cores.

[Anonymous Coward]

Yes they actually have them.

Rockchip is the only company with both licenses manufacturing something.
VIA has access to the Intel/AMD ISA patents AFAIK, but hasn't leveraged them for anything outside embedded in 10ish years, which is also when they stopped having access to Intel system busses to base their motherboard/cpu interfaces off of. AFAIK they were still using the P4/C2D bus up until the last generation or two when they went IMC like everyone else. Someone correct me if that last assumption is wrong.

I would say here is hoping to VIA getting back in the game, but I assume they are just going to follow in Intel, AMD, and the ARM manufacturers footsteps and make a signing mandatory locked down chip that can spy on us from behind the curtain. Fuck that shit.

No truth to iPhone cpu myth

[Darkness Of Course]

There was exactly zero chance of Intel making iPhone CPUs. It was never on the table. Intel wasn't in the business of fab to mobile market.

Otellini was smoking something when he made that claim. Absolutely nobody else in the company believed it. The market was too small, the IP was wrong (as in Intel was on the wrong end of it). Just like everybody else Intel/Otellini didn't think apple could cook up enough business to change their business model. Which was complete verticle slice of IP/Process/CPU/MB/Servers and it was making quite a bit of coin doing it. Becoming just the company that makes apple designed CPUs for phones, no viable business model for that.

Maybe a fever dream left a vague unease behind, but it wasn't even a possibility. Intel never made the short list.

stay hungry, stay foolish

[Anonymous Coward]

"By declining Steve Jobs' proposal to make the original iPhone CPU in 2005, Intel missed a huge opportunity. "
This is news to me.
At that time, Intel just won the battle with AMD, and it is dominating the field of CPU of PCs and servers.

Threats to Intel

[sjbe]

I don't think ARM or AMD are going to be the next big threat to the Intel monopoly.

AMD isn't and likely won't be a threat. Intel makes more in profit than AMD does in revenue. AMD is unfortunately a rival in name only and they operate at a significant cost disadvantage to Intel because Intel is vertically integrated and AMD isn't. I'd like to see AMD doing better but if you look at the financial statements of both companies (and I have) you'll quickly conclude that AMD is trying to diversify away from competing with Intel because it's a game they cannot win. They've been trying and failing for 30+ years and the only reason they continue to exist is because Intel needs them around to keep the anti-trust authorities off their back. Intel could put AMD out of the CPU business in a blink if they had a free hand to do so. (They would simply drop the price of their CPUs to less than AMDs cost until AMD exits the market - Intel wouldn't even have to take a loss to do it)

ARM is a threat but an indirect one. ARM can't (and likely won't) compete with Intel in PCs but ARM is kicking Intel's ass in mobile. As mobile grows the traditional computer market shrinks to some degree which hurts Intel. The biggest threats are usually the indirect ones. Nobody is going to compete and win head on with Microsoft in desktop operating systems but linux is winning through the back door by eroding the market through mobile. Intel faces a similar threat. ARM's threat to Intel is by eroding their existing markets.

Even they will be supplanted. It's going to come from left field, specifically from a company that is researching the next generation of computing.

One can always imagine some advance in technology that can kill the incumbents but there are more immediate threats to Intel that don't require invoking some as yet undeveloped new technology. The biggest one I think is that the large companies that operate data centers (Apple, Google, Amazon, Facebook, Microsoft, etc) start to roll their own CPUs. They appear to be working on it. Power management is a serious issue with data centers and Intel has never been very good at that. (which is a big reason they struggle in mobile) Plus there is significant margin leakage when you buy as many computers as these companies do. Vertical integration can provide serious cost savings. And these big companies definitely can bankroll the move to a different CPU architecture if they are so inclined. Intel makes about 30-40% of their revenue from this sector so it's not trivial to them.

Better battery

[DrYak]

Well, running the software on dedicated hardware (or on a full blown VM somewhere on the cloud - as several gaming solutions do) and streaming to your phone still beats everything in term of battery life vs performance.

(Also, you can turn the PC on and off remotely no need to have it run all the time. That's the whole point of Etherwake or newer technologies for lights-out management like IntelME, IPMI, etc. those even provide the VNC remote access. But saddly often also provide tons of exploitable bugs.).

Enables you to also leave the complex tax running on a power-grid machine and completely disconnect it from the phone (Zorpheus mentionned Photoshop. But such a setup would make even movie rendering possible).
Enables to switch device (quickly do a few manipulation over the smartphone, then switch the remote controlling to the tablet when in a more comfortable train, then finally switch to physicall keyboard and screen once you're back home).

Not architecture

[DrYak]

I wasn't arguing wether the x86 or the ARM micro architecture is the "ultimate best one ever".

The parent was just point that intel never had a good CPU for smartphone.

I'm point that it's worse: they used to have one (an ARM based one) but managed to sell it out at the wrong time.

the micro-architecture is only relevant to make the "never had a good cpu for smartphone" sentence true.
They never had a x86 one.
They actually has a good CPU for smartphone (which happens to be an ARM, but that's completely orthogonal to my point. it could have been MIPS, it could even had been a motorola-compatible, etc.) but they managed to throw it away.

---

Though I partially agree with that the "RISC is always better no matter what" mantra is a bit overrated.
(Mainly, the "cruft" often attributed to x86 doesn't really play a role on the scale of CPU currently made by Intel and AMD for workstations/servers/laptops)

On the other hand, the RISC is still relevant in some extremely small form factors (embed). And as the smartphones progressively grew out of the embed world (there are the precusors of IoT before IoT was a thing), ARM eventually ended up sticking.

Yes, Intel could have poured ressource to make a smartphone-grade lesser-atom-like embed cpu. Probably.
But as you point out, it's definitely not worth the cost.
The single main advantage selling point of x86 (running unmodified binary code) is moot in this form factor (again, 3D Studio on a 5 inch screen ?!)

What would make sense, instead of re-inventing a whole new wheel (a hypothetical smartphone-grade intel cpu) intel did the right thing and acquire a company and a core that is smartphone-ready (StrongArm. - it happens to be an ARM but that's orthogonal to the whole debate. The key point here isn't x86 vs ARM. The key point is "core designed to work in sub-notebooks and settop-box" (i.e.: relatively power hungry) vs. "core designed to work in smartphone, pda, etc." (i.e.: even less ressource consuming)
But somehow Intel threw it away at the wrong time.

That's also why we aren't seeing that many practical real-world use of ARM in servers.
Most of the existing ARM cores happens to be optimized for the ultra-low ressurce stuff such as smartphone. Not that many ARM cores happen to be good for servers (e.g.: not many do feature SATA bus).

And finally note that AMD is planning to eventually go the reverse route : acquire a 3rd party ARM core (Cortex core in Opteron A), and then progressively build a server-grade ARM CPU out of it (upcoming K12, eventually one day, when AMD has enough left-over ressource after the current focus on Zen).

Intel inside, matters no more

[Anonymous Coward]

This.

Intel , brilliantly marketed an unseen, barley understood by most , component.

Fast forward two decades, and mostly no one cares what chips run in mobile, the market Otellini ceeded to others, nor does anyone care what chips run on their cloud servers (save when the chip causes millions to be spent in time when Intel chips causes a security hole)

So this article is click bait, yep. Intel will no longer be the growth engine that they've been. Their market is now commoditized and will be a nice business, but the days of Intel under Grove are long gone. But they'll survive , after wave after wave of layoffs, they'll become a steady business.

The world's tiniest violin is playing

[Anonymous Coward]

I am SOOOOOOOO sorry for Intel. It's not like it ever used it's WinTel monopoly to destroy competition back in the 90s.

And, it is just SOOOOOOO consumer-friendly, as exemplified by the way it handled this latest design bugs and enabled people to live a slower, less stressful life by slowing down their devices.

If only there was something I, as a common man, could do to aussage the grief these poor millionaires and billionaires are suffering. Oh, the humanity!

You live by the sword, you die by the sword. Good riddance.

And that goes for the rest of you greedy, sociopathic Silicon Valley types. We look forward to dancing on your graves, too.