They also usually have the smallest process technology of the time. If necessary, Intel could fab ARM chips and dominate the ARM market from the inside.
Precisely what I was thinking. Does anyone know if there is a reason Intel hasn't just acquired ARM? I assume Intel believes that they might still be able to compete and win in the market, but that seems to be less true every day.
(I assume the answer is that it would raise a bunch of anti-trust flags everywhere, but I'm not positive about that.)
Because the Intel leadership signed off on selling their ARM interest (XScale) to Marvell. (The very group that is now turning out a product aimed directly at Intel's lucrative server products)
So it would take a gigantic admission of error for Intel to start up another ARM group. Someone would probably have to resign and I can't imagine anyone's eager.
Also, the profit on being an also-ran ARM supplier is nothing compared to what they could do as a market-dominating mobile-x86 supplier. It's longer odds, but higher payoff.
And the way ARM licenses get passed around (as I understand it) Intel isn't really falling behind by pursuing x86. If they reverse themselves five years down the line, they can quickly catch back up, close enough to the state of the best ARM processors that their superior fab tech can make competitive chips.
In any case, they've always said mobile x86 was a long-term bet. I don't think they honestly expect to be competitive for at least a few more years. So any evaluation and possible surrender is still down the line a bit. Particularly as they're still hitting their targets and estimates for improvements in the ATOM line.
Intel makes several 'families' of ARM (Xscale) based parts: IXP, IXC, IOP, PXA and CE. Intel sold the PXA family to Marvell Technology Group in June 2006.
The simple fact is that Intel still holds an ARM license even after the sale of the PXA family to Marvell.
Plus they have a profit margin of 25% on revenue of 40B. So they could be selling everything 20% cheaper and still earn money (and lots of them). And that's including all the settlements Intel had to do this year. It's just a big Chipzilla
Also, considering the sheer inertia in the economics of the larger fabs like TSMC's and Intel's (and by "large" I'm including both quality of product and production capacity), having an excellent silicon design in no way guarantees you market leverage. Having an excellent design and optimized arrangements to fab it quickly in large numbers do.
In the last 2 years we have seen many ARM netbooks prototypes showing better battery performance and cheaper price than Intel netbooks. But I still can't buy an ARM netbook. Where are they? What is holding them back?
heh, just got mine delivered today, and fired it up just to post this reply :-)
It's a Toshiba AC 100 with an ARM-based Nvidia Tegra processor and Android operating system in a standard netbook form factor. I don't believe it's available in the US though, if that's where you are.
First impressions: it's very light, to the point of wobbling around when I type with it on my lap. Many of the peripheral keys are very small for my Western fingers, although I feel I'm getting used to them already. It's fanless, and has only flash storage, so it's eerily silent and apparently gets up to 8 hours of battery life. The trackpad doesn't have enough vertical space to be comfortable, and there are neither function keys nor a delete key, although there are a whole row of special keys which presumably can be xmodmappped.
The online reviews mention that Android provides a less than elegant interface with the lack of a touchscreen, and I agree; I was planning to install a Linux or BSD OS on it all along. As I understand it, that's pretty tricky; should make for a nice weekend project :-)
As for why there aren't more of them around already, my best guess is that OEMs don't want to piss off Wintel by selling what is effectively a PC which doesn't use their platform, for fear of losing the discounts and rebates they get when they do. I hope I'm just cynical and there's a better reason; for casual use the system resources definitely seem adequate. Acutally, as soon as I get a proper OS and WM installed on this thing, I can easily see it taking up half of my computer time, and it will be brilliant for having a quick hack on the go...
[edit]
After the first half hour or so of usage, it's clear that the Android as currently installed is definitely not a good fit for a non-touchscreen, full-keyboard device: After 5 seconds of idleness, the screen dims, then blacks out completely, just like a phone would; navigating through screenfulls of icons using keyboard or mouse becomes tedious very quickly, and to top it off I've run into a few bugs as well..
This thing is begging for a linux/bsd makeover, hopefully I will uncover a real beauty...
Netbook makers are getting too good a deal on Windows to switch back to shipping flavors of Linux, at least for the US market. There is far more demand for "cheap Windows laptop" than just "cheapest laptop", if that makes sense.
That said, the rumor is that Chrome OS is only going to ship on ARM, and will do so Real Soon Now. I'm not convinced that Chrome OS is going to be a hit, but I can see more people being willing to give up Windows for it than for a typical netbook Linux distro.
I assume the Windows OEM prices have been tuned so that it's more expensive to develop and manufacture ARM-based netbooks that run Linux than it is to stick with x86 for the current generation.
Also, the processor is not responsible for a huge part of the power consumed on a netbook. Disk, screen and memory seem to be a big part. If you make a netbook that can run Gnome and Firefox decently, it will have a hard-drive (or SSD) and a couple gigabytes of memory. The LCD is the same.
If you sell an ARM powered netbook that means no windows. So you can wave goodbye to being able to use OEM copies of windows on any other products that you make.
So anyone like Asus or Acer that have a PC business and launch a non-windows netbook soon learn the error of their ways.
They can sell one line of Windows-proof highly capable, futuristic computers while keeping and developing their legacy-ish, kludgy, virus-infested, Windows-bundled, 8080-descendant-based inefficient lines.
Of course, Microsoft will probably find a way to increase their OEM pricing, specially if the Windows-proof line ends up making a dent on the Windows-bundled lines. And they don't even need to increase, say, Asus's licensing prices. All they have to do is to lower Acer's.
MSFT tend not to increase pricing they tend to say - ship even one computer without Windows and you wont be buying OEM Windows for any of your other machines.
They've gotten in rather a lot of trouble for doing that in the past. Do you have any evidence that they have been doing it since the emergence of the netbook market?
I think it is very workload dependent how big of a power consumer the hard-drive is.
When I only do word processing or browsing simple web pages, I never have to read data (everything is cached). Laptop-mode also delays writing data back.
I prefer to think of this in a Christiansen disruptive technology context. Technologies grow upmarket beyond the needs of its user base and create a disruptive opportunity for downmarket technologies.
I recently made a comment about this on an Intel blog article (SC10: Jeff’s Notebook - Innovation and Disruption: How a Successful Company can be gone in 10 Years): http://intel.ly/dZAxKk
"People in the mobile phone architecture do not buy microprocessors," says Hauser, "so, if you sell microprocessors you have the wrong model. They license them."
Someone is building them. In iPhone's case its Samsung. The phone's Apple-designed A4 processor, which was built by Samsung, costs $10.75 each. Most of that is going to Samsung, part of that goes to ARM. The ASP of ARM 10-20x lower than that of Intel.
Someone seems to have forgotten that ARM means "Advanced RISC Machines, Ltd" and that Acorn, Apple and VLSI were partners in A(dvanced) R(isc) M(achines).
Of the three (Acorn, Apple and VLSI), only Apple is really still around. Apple must already have significant equity position in ARM being a founder.
VLSI was bought by Philips Semi, which was later spun out as NXT. NXT is a user/licensee of the ARM architecture but not even the biggest one by a long shot.
Intel's revenue from ARM (and yes, they still make some Xscale parts) is probably bigger than all of NXT's revenue.
Acorn went belly-up as a going concern in 2000 with the remains bought by Morgan Stanley.
The ARM "JV" IPOed (as ARM Holdings).
In any case, I doubt that Apple pays much, if anything for its ARM license.
And I doubt we really know what Apple pays for A4 parts.
In the server markets, you look at performance per watt and performance per rack unit and ease of setup - for general purpose computation, Intel still dominates those metrics. (with some exceptions [1]). Being ahead in fab technology, it can take the performance hits due to the legacy x86 architecture.
You can replace 1 beefy server with 8 small nodes, but you can not replace the convenience of administering that one server.
Also, retargeting performance-sensitive JITs/compilers/maths libraries such as GCC, LLVM, Java, V8, *BLAS is no small feat either - x86 gets support first, because it is the market leader, and also gets tuned the most, since it's the oldest backend. A feedback loop that's hard to break, so don't expect Intel to go out of business any time soon.
I don't see how intel is even a target. PowerPC owns gaming. There are still tons of MIPS chips in networking applications as well as some other verticals. ARM owns phones, largely because every phone has a custom chip in it. It's not clear if it will grow beyond that.
Why not suggest ARM will defeat PowerPC or MIPS? That sort of seems plausible, Intel is so insanely good at what they do and they backed out of ARM a few years ago. I can't imagine they did that without understanding how they'd compete.
i think you're talking about console gaming perhaps, xbox and wii use powerpc variants, but ps3 uses the cell processor and nds uses an arm processor. in terms of mobile gaming, powerpc is nowhere to be found, this is owned by arm. in terms of pc gaming, intel and their we're-not-totally-a-monopoly division, amd, own that realm. though arm is starting to make inroads into low powered servers so it'd be interesting to see where this goes in the long run.
ARM is not really friendly with competitors especially those not paying the high license fee. Do you remember the nnARM project at OpenCores? this has been removed on request by ARM limited. If ARM wants to "own the market", they should allow in some ways competitors to build compatible CPUs.
"Merging" the desktop and mobile-touch universes is a fool's errand. They shouldn't be unified for the very same, very good reasons that the iPhone was served by a fork in the first place.
Frankly: you're looking for a fundamentally different experience at the desk top than you are on a mobile. And ARM is a bad choice for the types of tasks that people are still buying traditional PCs and Macs to do.
It may be a fool's errand but that's the way Apple is going. And I do not see them having any lack of market success with their approach.
I don't think it's really a matter of desktops becoming like mobile devices. I think it's "desktops as we know them will be going away (except in specific use cases)".
My wife's Macbook has seen better days. When it finally dies, we're going to replace it with an iPad, a bluetooth or docked keyboard and a camera connection kit. It does everything she needs. I imagine this will be the same when my kids get old enough.
And why is ARM a bad choice? Have you heard about performance complaints from iPad users? Have you seen Rage HD on iOS yet?
> "I think it's "desktops as we know them will be going away (except in specific use cases)"."
I tend to agree with that. But that's a wholly-different argument than suggesting Apple would make their desktop OS ARM-only.
OS X is Apple's solution for those use cases where desktops do and will persist for the next decade, at least. As such, making it ARM-only would require the use of currently-inferior chips, chipsets and tools. It would murder Apple's steadily-growing share of the very market OS X exists to serve.
iOS is already serving their mobile devices, and doing it very well. It stands to gain very little from a hypothetical OS-merge and stands to lose quite a lot by becoming more PC-like. (simplicity, stability, performance, etc)
"Mac OS 11" (sic) is a marketing term. It doesn't necessarily have to imply anything relating to the existing product. I'm not trying to be pedantic, just trying to illustrate that I see Apple going strongly towards the iOS route as their main computing platform and I think the facts illustrate the genesis of that.
It may be true that the current ARM offerings are inferior in terms of benchmarks, but as far as general computing use they're probably overpowered. I don't really need an i7, a Core 2 Duo is fine. Nearly everything seems generally overpowered these days.
To you maybe. Some of us also do stuff like Video and Audio editing, processing and mixing on our machines, not to mention photoshop and stuff. Remember, the industries that made Mac cool in the first place? yeah, we're still here, just getting addressed less and less. i everything is nice, but there's still millions of people who use this shit professionally.
Apples attitude - i.e. killing the xserve when there's hundreds of the things in usage in render farms and video asset serving etc. - is, while typical, somewhat frustrating.
Yeah, you didn't get my point. I said general use, which video and audio editing isn't.
Apple's "attitude" is called meeting market demands and eliminating the cruft from their product lines. They're here to make money not do you (or me) favors.
iPads certainly perform well. With hardware acceleration for h.264, the ARM chip may not be seeing the toughest loads it would otherwise. I believe there's an ARM build of VLC.
I've been meaning to check it out to see how video playback is with other codecs. (It's easy to use a USB tuner to record HDTV on a laptop or desktop, but the MPEG2 used in the U.S. is a bit of a chore to transcode to 720P h.264 which is the Apple default for iPad/iPhone/iPod-touch/ and the current Apple-TV (all on ARM now).
I'm not sure how far towards the high end ARM will go, but it sure makes sense to have things like Apple TV on ARM.
It uses less than 6 watts.
Try running Photoshop CS5, Starcraft 2, AutoCad on ARM. I am not saying it's not going to happen. ARM is focusing on the embedded market. They are not going to be able to compete with the likes of a Quad Core Nehalem or Sandy Bridge.
ARM processors have, so far, been targeted to mobile, low-power applications for a very long time. When it was launched, powering a desktop computer (the Acorn Archimedes), its speed was comparable to then current 386s with FPUs.
There is nothing that prevents building an ARM processor that performs as well as a current x86. In fact, due to its simpler implementation logic, an ARM implemented with the same technology as a x86 would probably run faster or leave more silicon for caches, inter-core communication, more cores and specialized logic.
You can bet Microsoft will make Windows 8 ARM compatible. They don't want to lose the tablet market and since Intel is having such a hard time getting into that market, they can't simply rely on them.
And if you're going to suggest WP7 for tablets, don't. Microsoft will only do that in worst case scenario if Windows 8 will prove to not be good enough for tablets. They don't want to get $10 per license when they could get $50, and that's especially true if touch/mobile computing will replace desktop computing.
The way Microsoft has handled the mobile market for the past 15 years, it seems likely that they'll go through at least two or three more mobile operating systems before the market can decide whether x86 belongs in handheld devices.
When you talk about Apple, and you talk about ARM, remember that Apple bought two companies: PA Semi and Intrinsity (to get chip design talent). The PA Semi acquisition came with IP for a POWER chip, the PA6T. Thus far Apple has announced nothing using that chip (and possibly never will).
Apple certainly has the ability to make OS X run on POWER, but doing that might people fall over from dizziness.
Daniel W. Dobberpuhl (the founder of PA Semi) was also the architect for StrongARM (when he was at Digital).
Intel only got involved because DEC sued Intel for patent infringement (over the Pentium processor), and lost. Intel grabbed a lot of DEC's technology in the settlement, including StrongARM (which became XScale).
from the article, "stifled innovation", "monopoly", i wonder how true that is, where would we be if any of this was true, would we have a more awesome chip in our laptops today?
i'm pretty excited about gpu's, but they're only good for parallel instruction processing, some tasks still require the usage of a highly spec'd cpu.
Has ARM's design licensing model ever been demonstrated in the high end (server/workstation) CPU market? It seems to me that the amount of engineering expertise required to push that sort of high performance, fully custom design through to fabrication would make it difficult.
More accurately, they did build ARM chips -- the XScale line. But they sold that division to Marvell in 2006, although they still retain an ARM license.
ARM could probably kill x86 given some time, if that was what the writer implied - and let's just wait a few years and see how things currently brewing will develop before we start shooting off our downvotes, shall we? I am willing to bet that an often-blamed-for-being-fruity computer manufacturer will shortly migrate from x86 to ARM. Just wait.
Killing Intel themselves, though, just won't happen. If anyone can dominate the CPU market wielding ARM in their hand (haha!), it's Intel. The rest of the world has a good few years to go before catching up to their level in terms of manufacturing.
Intel has so much FAB capacity that they can't easily be killed even by superior competing products.