Boring Outside

‘Know thyself’, was the motto of the ancient world. ‘Don’t be evil,’ is Google’s. ‘Don’t be boring’ is the adage of the high-tech consumer world.

Boredom kills. It’s killing Sony and Nokia and it may kill Intel. While you can get away with screwing your customers, short-changing your customers and fooling your customers – as Intel has demonstrated – you can’t get away with boring your customers.

While the CPU market is growing in single digit percentages, the applications processor market is growing at 40% a year and, by 2015, the value of the markets will match at around $33-34 billion, according to Nomura.


This will have important consequences.


ARM, of course, is dominant in the fastest growing processor end markets such as tablets and smartphones and Intel is dominant in slow-growing markets like notebooks, stagnant markets like desktops and declining markets like netbooks.


Intel’s iron grip on the netbook, notebook and desktop markets means that the products all look the same, and there’s been little innovation. Consumers are increasingly finding these product categories boring. ‘Intel Inside’ now equates with ‘Boring Outside’..


For excitement and innovation, consumers go to smartphones and tablets where volumes are soaring.

One interesting consequence of the rise of the apps processor, according to Nomura, is that, with ARM-based processors expected to move into computing, this gives Samsung its chance to overtake Intel in the processor market.


The gap is wide at the moment, with Intel’s chip revenue at $50 billion and Samsung’s at $30 billion, but a move to ARM-based computing could see Samsung closing the gap, says Nomura.


The battle lines are drawn between Intel’s lead in process geometry and ARM’s lead in power efficiency and lower production cost.


Is Intel’s 22nm finfet process really going to deliver chips as power-efficient and cost-competitive as ARM’s?


Forget Intel’s marketing BS – the proof of this particular pudding will be in the eating.



  1. Games companies have more experience in graphics than all other applications put together so I really don’t understand your point.
    The graphics drivers come from ATI and nVidia and from my experience the development cycle is Windows first, Linux second and Mac(where relevant) third.
    But the big problem with Macs is their attempt at alpha sorting appears to be far worse than those on Windows and Linux. No idea why this should be but it’s a major reason not to use a Mac. The music companies do have to use Macs of course and use all sorts of workarounds for this problem, whereas most games companies simply don’t bother.
    I agree on the Intel graphics point though. Although the hardware is getting better, Intel just don’t seem able to get the software right. Maybe once Haswell appears they will.
    @SecretEuroPatentAgentMan : I first met Piccolo on an EU FP4 project and it became obvious from early on that it was the wrong solution. ARM then began wotk on their next DSP attempt and I think the ARM sales force just quietly forgot about Piccolo altogether and waited for the next attempt from R&D.
    But you wouldn’t have liked it if they had told you about it. Whenever a Piccolo exception instruction was initiated the main core had to sit there asking “are you finished yet … are you finished yet …”

  2. SecretEuroPatentAgentMan

    The Piccolo embarrassment sure set in quickly, within weeks of seeing the announcement my employer contacted ASM for more information as it just fit a project we were starting. My memory is a little vague here but I believe part of the attraction was the extra wide accumulators of Piccolo. We never got a reply.
    I had hoped Dalvik would be an opportunity for the underdog. When I heard Google was buying Motorola I wondered if they were getting some CPU designers too but it seems what is left of that went to Freescale. Google had wed itself rather closely to ARM with its Android platform. A Dalvik core for ARM, Coldfire and others might change the landscape.

  3. I couldn’t care less what some games companies think, I prefer to trust actual experience.
    And my experience is twofold:
    1)Intel graphics chips in laptops are the most notorious for OpenGL quirks, no matter what OS is being run.
    2) Getting good drivers for Linux platforms is far, far more problematic than for Windows or Mac.

  4. After the embarrassment of Piccolo and then ARM’s next attempt at DSP, the limitation of ARM DSPs was discussed at the European Commission in the late 90s as it was recognised to be a major issue. A “Euro-DSP” project under FP5 was proposed but there was a lack of buy-in even from ARM itself.
    I agree on the Dalvik idea – hopefully it’s on ARM’s ‘to-do’ or preferably ‘being done’ list.

  5. “Well no, maybe you can’t get closed source apps to work on the Mac”
    That rules out 95% of all useful software then 🙂
    As for graphics, the Mac implementation of OpenGL is not 100% compatible with the Linux and Windows versions due to some Mac graphics quirks. Ask any games company what they think about OpenGL on the Mac and it isn’t polite !!

  6. SecretEuroPatentAgentMan

    > If ARM were a software IP company, one day, there would be an “open sourced ARM”
    There is one on OpenCores. There used to another one but it was pulled down for patent infringement. Not sure about speed though, but nevertheless it is an interesting project.
    > one day there might be an ARM++ coming from some licensee of ARM
    As mentioned elsewhere Qualcomm has an architecture license and makes fast ARM processors. What is more intriguing is that they were apparently not satisfied with the DSP extensions for ARM and instead went for their own design, the Hexagon. ARM has MAC units and NEON and they once had the ill fated Piccolo and more, but it seems their DSPs are found lacking both for audio as well as for baseband. Strangely very little has been written about this.
    > why they do not invest in Android
    ARM did provide Jazelle that should accellerate Java on handsets. They disclose so little about Jazelle it looks like they are embarassed by it. Java VM is stack oriented but Dalvik is register oriented. It seems the field is wide open for a core provider of Dalvik optimised cores or extensions.

  7. You can have a quite good multi-os setup on the quad core MacBooks. Simply install a virtualization application. Then run plenty of those VmWare/Parallels desktop VM’s, such as your favourite Linux distro and Win (7) for those pesky Win32 apps.
    In the (manufacturing) industry it’s fairly common to get a predefined virtual machine image with all the software’s and settings already done, it’s simply only to boot the machine image and you can code away in minutes.
    Great for legacy support as well. When a new environment is released, make a copy of your ‘old’ environment and proceed installing the latest and greatest.

  8. Well no, maybe you can’t get closed source apps to work on the Mac. But anything open source is possible.
    As for Mac 3D graphics incompatibilities, its not a problem if you use a decent graphics toolkit like Qt. My OpenGL code is exactly the same on Mac, Linux, Solaris and Windows.

  9. Unfortunately I think ARM shot themselves in the foot a little with the Architectural licences. For very high volumes it is already cheaper for Qualcomm and others to design their own ARM processor rather than licence cores individually from ARM. Putting up the licence fees would drive more people in this direction and possibly eventually fragment the ARM ecosystem.
    ARM is going for new markets and their drive for

  10. I somehow don’t think Xilinx, Synopsys, Mentor and co are going to give me their source code to recompile on a Mac … or anywhere else 🙂
    Mathworks do produce a Mac version of Matlab but it’s usually behind the other versions and often has inconsistencies.
    Also the Mac 3D graphics system has some major incompatibilities with the rest of the known universe so you have to totally re-write any code for that.
    You can use virtualisation on a Mac to run Windows or Linux but then you end up with less resources to run the actual code and it always ends up being easier/faster to just buy another PC and install Linux.

  11. “But there’s little point in buying a MacAir and then putting a real OS (Linux) on it so I can run anything useful, no matter how desirable it looks.”
    It already has a real OS, one that looks just like Linux! I can compile and run source code written for Linux with only minor changes.

  12. It seems to me to be such an unusual business model – the product of expediency rather than planning – that it’s better to let its customers decide what it’s revenues should be, Greg. Licenses aren’t cheap. Apart from designing new cores for new market segments, I can’t see how it could grow revenues without pissing off its customers.

  13. I like the Arm business model but I think if they are going to be successful (remain independent)in the long term they need to really raise revenues. 600MIL for a company with such market share and huge product numbers is not really up to snuff.

  14. Same here. I also have a Japanese car – built like a brick how’s your father and, so far, touch wood, as solid as a rock

  15. I thought American and European cars were on a par build wise – which is why I drive a Japanese built Scooby-doo 🙂
    In the past HP had two ranges – the well built top end to match Toshiba and the cheapo crap to match Dell. But HP lost their way big time and I agree my next one probably won’t be from them.
    But there’s little point in buying a MacAir and then putting a real OS (Linux) on it so I can run anything useful, no matter how desirable it looks.
    Also an interesting read I came across is at

  16. For me, Bitter, the iStuff wins because it’s less hassle to use.

  17. I suppose people rather shell out a slight premium to get that extra build quality and attention to detail that Apple seem to understand and get right, than go for that plasticky, cheaply built HP or Dell laptop. Even though its more or less the same dull electronics inside.
    The same could possibly be said about American vs. European cars? 🙂

  18. Yes … I agree 100%; complacency kills innovation, just as with Sony and Nokia. Every company needs an Andy Grove to kick arse and ruffle feathers … and at the first sign of success, tevery company need to set up a parallel company who’s sole job it is to keep them on their toes … or kill them!!

  19. True, mgp-1, but there’s a hardware issue as well. Did Intel come up with the idea of the tablet? No, they’re struggling to get into tablets. Did Intel come up with the idea of the MacBook Air? No they’re struggling to copy it with Ultrabooks. Did Intel come up with the idea of smartphones? No, they’re struggling to get into them late in the game. So in the three biggest changes in the computing environment in the last half decade, Intel was way way behind the innovation curve.

  20. It’s true Mike. For quite a while I thought I’d never buy an iPad. I couldn’t see what it would give me on top of the netbook and iPhone I already had. Then, on an impulse, I bought an iPad. Now I absolutely love it. I reach for it every morning when waking up, check mail and news. Get a good idea what I’m going to be doing that day. What’s more Terri, who used to moan about me reading my iPad, has got one too, and she loves it. It has to be magic dust – there’s no other explanation.

  21. ah but it’s got ‘magic dust’ on the outside.
    Members of HMP seem especially attracted to items covered in this.

  22. I think you can hold Microsoft to blame for PCs being so boring not Intel … they must be as frustrated as us all for the total lack of imagination from MS other than to keep loading us up with upgrades few users want of have need for whilst at the same time ignoring decade old frustrating ‘bugs’ and coming out with no decent new apps. Now MS IS boring … and that’s the kiss of death for their ‘tablets’ and Win 8 phones. Only the X-box scores OK and that only because they had real competition to live up to. Bottom line; after being screwed for years by the Redmond giant … anything with MS on it is a cue to run full speed ahead in the opposite direction … classic brand implosion 🙂

  23. Her Majesty’s Press, Mr C, rises above such sordid calculations.

  24. I see that your advertising manager is going to love you Mr M?
    Can’t see many Intel adverts in EW for awhile?

  25. Maybe the exception is being built right next door to yours David?
    The really awesome Intel based computers are based on the same home-brew philosophy that Apple started off with back in the day.
    Case-modders like that makes Apple hardware look boringly generic. Luckily it’s still possible to buy discrete computer components such as processors, memory, motherboards, etc.. Generic and standardized stuff isn’t always a bad thing, though.
    Then there is also the nutty overclocking crowd pushing the latest AMD’s/Intel’s to ridiculous clock frequencies, for example, using liquid nitrogen for cooling.
    Cool, isn’t it? If you don’t mind the ‘pun? 🙂

  26. Somehow, Bitter, I don’t think MacBook Pro was concocted from an Intel reference design. However you are, of course, correct. MacBooks are an exception to the ‘intel inside, boring outside’ principle. I wonder if there are other exceptions.

  27. The new ‘Retina’ macbook pro looks pretty desirable, and yep, it’s got ‘boring inside’.
    Blaming intel for the lack of interesting products in, from the most part, boring bean-counter run PC and mobile computing manufacturing companies, I believe, is quite off the mark.

  28. Thanks John much obliged

  29. I totally agree with you, David. I was mainly replying to another poster here.
    I like your slogan, though!

  30. My point, John, is that Intel’s domination of the PC industry for 20 years has led to a grey uniformity among PCs – like the grey uniformity of life in totalitarian states. A free market leads to diversity and innovation. Domination leads to boring conformity which is why, if it’s Intel inside, it’s boring outside.

  31. ARM’s revenues are closer to $600 million, Peter, but your point is absolutely valid. Different business models of course – but ARM come out with a seemingly never-ending stream of brilliant cores and Intel, with all their money, still don’t seem to be able to get close to them in power efficiency.

  32. The fact that ARM (revenues $200 million) can give Intel (revenues $50 billion) a run for its money, and out compete it in key markets is an outstanding achievement.

  33. Mike you and I will never agree about Intel. You see them as pioneers driving the industry forward, I see them as monopolists using monopoly profits to keep ahead of the game. You think it’s good they have a pot of dosh to keep process technology evolving, I think that dosh, like muck, does most good when spread widely.

  34. You have a point in saying that ARM is reliant on others for their success, but ARMs model is very much one of standards based dis-aggregation. By buying from ARM, all of these customers have eliminated expensive chip development programs of their own, and have created an eco-system around a common, increasingly standard, architecture.
    Intel have indeed woken up to the mobile market place, but without the MDF programs that have homogenized the PC/laptop insustry so heavily in the past, they will just be another player in a very different market to the one they competed in – and won – in the early 90’s. There was no incumbent then, but there most certainly is a cadre of very powerful players in the mobile market they are now trying to enter. It’ll be touch and go for them, but even ARM acknowledge they’ll get about a 10% share. ARM only claims a 90% share in smartphones, for example, so I suspect the displacement will come more from beating out the non-ARM 10%, as opposed to displacing a huge amount of ARM market share. We’ll see!
    It’ll be interesting to see who buys each of the two Surface tablets. The ARM-based tablet – looking identical to the X86 tablet – at the price that it is, will probably be the one favoured by consumers. With it appearing on the market before christmas, and 3 months earlier than X86 Surface, also gives it a surface advantage, if you’ll pardon my crummy pun. The X86 surface will undoubtedly be the preference for IT/Commerce, if tablets are desirable there.
    Some of your questions regarding ARM are intriguing, but there are practical answers to those. ARMs business is all about enablement – giving their customers what they need to succeed in their end markets. You rightly point out ARM don’t manufacture – their partners are the ones adding value to their final products. Why should ARM be called out as an ingredient, when it’s all about the platform – Snapdragon, Tegra, OMAP? All of these contain IP from elsewhere.
    ARM aren’t shit-scared of fabs. I’m sure you know how much fabs cost, and how much money ARM makes itself each year? In total, the ARM partnership sells more silicon both in terms of value and (by far) units shipped than Intel. But ARM itself takes a small percentage of that. They want to have a smaller piece of a bigger pie, says Warren East.
    It is doubtful that anyone can copy ARM’s designs. ARM licensees ARE allowed to develop their own CPUs through an architecture license, as do Qualcomm (Scorpion/Krait), as will NVIDIA (Project Denver), as do Marvell (PJK4 in the ARMADA XP etc.). But the thing they are not allowed to do is compete with ARM by licensing that IP to anyone. They may only use that for their own products.
    Hope all of that helps.

  35. And who is gaining from the rise of the smartphone/dumbpad ? Intel and Samsung !!
    All of these devices suck huge amounts of server power which of course almost universally use Intel devices and Samsung memories. And if you think notepads all look the same, you should try looking at servers 🙂
    ARM made some gestures towards the A15 being a ‘server’ chip but most people soon saw it was an order of magnitude too slow and that’s now becoming the next generation applications device processor.
    @Cheese : the ARM++ exists – Qualcomm design their own ARM processors and have most of the market. Intel aren’t actually trying to compete with ARM, beating Qualcomm is their target.
    Anyway Intel and ARM don’t make end user devices, they make processors/processor IP. It’s up to their customers to come up with innovative products and I fail to see how if there are 1, 2, 10 or 100 suppliers of processors how this would change. Indeed if there were even 5 suppliers of processors profit margins at Intel would not be high enough for them to drive the process research, equipment and materials industries forward and we would have had to give up on Moore’s Law at least a decade ago. In this scenario TSMC having even a 45nm process enabling ARM based Smartphones would still be distant dream.
    Moving back to the PC it’s gone through so much evolution that the ‘perfect’ PC has more or less been reached. Intel has been frustrated with the lack of innovation in PCs for years and keeps trying new initiatives, but the fact is most people like the 15″ notebook product as it does more or less what they want at a price they like, something that so far cannot be said for smartphones where the cost has to be hidden by the operator to make them palatable.

  36. Good Grief, R2D2, I can’t imagine where you got the idea for such a dynamic strategy.

  37. Well ARM seems to work as a player in an interdependent industry, whereas Intel wants to, and does, dominate the whole industry, cheese, to my mind it’s a domination which kills competition in the PC industry – just look at the rows of notebooks from 20 odd suppliers in a computer shop – they all look the same! When the computer industry can buy its CPUs from 20 different suppliers rather than 2, it seems to me quite obvious that we’ll see more differentiation, competition and innovation.

  38. I know Mike aren’t I a beast? Poor defenceless little Intel.

  39. Oh it must be Monday. David having another pop at Intel. Sigh.

  40. Yes, Intel is ripe for a shake up. Tell you what, let’s set up R2D2-Manners Private Equity with $1 each and buy out the Intel stock holders at a 50% premium for $200 billion. The $200 billion will be funded by 10% bonds requiring Intel to fork out $20 billion per annum in interest payments. With $50B in sales at 60% margin or more, that’s easy and will be further helped by mass layoffs, a 10% pay cut to all staff and closing all the fabs. In response to journalists pointing out that we are a pair of shysters, emphasise that we have developed proprietary ‘metrics’ to track and improve performance throughout Intel.

  41. Comparing ARM and Intel is unfair on a few counts – the most important one being that ARM doesn’t manufacture, and has nothing more to it’s credit than (the formidable!) IPs and customer base. The success of ARM based APUs depends on all the SoC and device makers using ARM IPs. And of course on Apple and Google (for iOS and Android respectively).
    While Intel has been a laggard in recognizing the mobile phenomenon, it is taking a few good steps – look at the Microsoft’s Surface tablets – the more powerful (professional edition) has Intel inside, whereas the entry product (targeting teens, housewives and the hoi-polloi in general) is ARM based. I wouldn’t hedge my bets on ARM architecture just yet – because Intel has momentum (mass * velocity) – where a non-zero velocity leads to a huge momentum!
    On the other hand, ARM as a company is positively boring. With the customer base as big and powerful, I fail to understand why they continue to play a “man in the gray shirt” role. Why for example, they donot have a branding program targeting consumers, or why they do not invest in Android or Win8 or whatever …or why they stay shit-scared of fabs. If ARM were a software IP company, one day, there would be an “open sourced ARM” beating the hell out of the closed source original (think of Java vs Dalvik). And who knows, one day there might be an ARM++ coming from some licensee of ARM, that can take the world by storm, by partnering with an Android or a Winx or what-have-you. Intel has got the warning siren from ARM, but ARM cannot rest – because what can get you to the top will surely not suffice to keep you there.

Leave a Reply

Your email address will not be published. Required fields are marked *