The Intel Developer Forum is currently in full swing, but it kicked off with a speech by Intel CEO Paul Otellini. Well, there’s bad news for those of us who long for a time where lots of different architectures compete with one another, ensuring that technology is moved forward. Otellini’s plans for Intel basically come down to one thing: x86 everywhere.
x86 or die
Paul Otellini’s keynote speech at the IDF was themeda “continuum of computing”, and in it, he laid out Intel’s plan to get an x86 chip in every device, from phones all the way up to supercomputers – but the focus was clearly the mobile space. The key development here is 32nm, which enables x86 (Intel’s prefers “Intel Architecture”) to be used in every device, big and small.
“Users and developers have to have the freedom to move seamlessly back and forth across this continuum,”, Otellini said. However, the focus lay on mobile devices, which is currently pretty much dominated by ARM. Intel has 12 System-on-Chip designs in the pipeline for this purpose.
We here at OSAlert – and maybe me specifically – really prefer a processor market which is not dominated by x86. Contrary to what you might think, this has nothing to do with this archaic idea that x86 is somehow inferior. No, it has to do with the simple competition: if everything is x86, and it becomes the norm on every type of device, competition in the processor market is pretty much dead not only in the desktop space (where it’s x86 all the way already), but also the mobile space, where ARM plays a major role, and even something like MIPS pops up every now and then.
I want a world where x86, SPARC, PowerPC, ARM, MIPS, and whatever else we can come up with compete with one another on all levels of devices, from big servers to small mobile phones. This keeps companies on their toes, and provides fertile ground for new and innovative ideas to take hold. It’s pretty much a pipe dream for now, though.
We’ve seen what a monoculture can lead to time and time again in our little geek world. Internet Explorer, Windows, and so on. It took long and hard work from mostly the open source community and Apple to force Microsoft to start improving their products.
Moblin & Atom
Intel also had quite a lot to say about Moblin, the company’s mobile operating system platform, based on Linux. They announced that both Adobe and Microsoft are supporting not only Windows, but also Moblin as platforms for their runtime environments, Flex and Silverlight. Does this mean Silverlight will find its way to Linux officially, from Microsoft?
On top of all this, Intel announced the Intel Atom Developer Program, a set of tools and SDKs aimed specifically towards developing applications for Atom-based netbooks and handhelds. “The netbook has become one of the most popular consumer devices in the market today, but its true potential has been limited by applications that are not optimized for its mobility and smaller screen size,” Intel said, “The Intel Atom Developer Program provides a great opportunity for developers to create useful and inventive applications that will unlock a netbook’s potential while opening a new sales and distribution channel.”
So Intel wants x86 to be everywhere. This is news?
Yes it is. Because AMD beat them to it.
I mean they beat them to a press release.
Look at the date. I could find AMD official press release.
2005 Baby!!!
http://blogs.zdnet.com/BTL/?p=1450
:biggrin:
Edited 2009-09-23 10:16 UTC
AMD? Intel = AMD = x86 monopoly
Obviously AMD just helps Intel in their crusades against infidels.
That’s sad. They have loads of money, but this is not enough. So this is their goal – grab all the money in the world and don’t let others breathe.
Anyway, computing was still fun 10 years ago or so.
you should see their new planned headquarters.
http://upload.wikimedia.org/wikipedia/en/f/f9/Death_star1.png
it isn’t finished yet, but I hear it’s running on their 32mn process.
also of note, the planet destroying laser is actually just stored up heat frmo the older P4 systems being released in 1 big push.
Yo, Intel, how is Itanium working out for you?
I am actaully really looking forward to Tukwila (intel’s next itanium). Looking forward to having one kick butt OpenVMS machine again.
my old itanium was a dual boot OpenVMS and MS server 2003 R2. it serverd me well (served, get it?) 
Nice. I hope Intel stop “overpromoting” x86 and cover a bit more on Itanium.
Edited 2009-09-23 02:56 UTC
At least it got mentioned on the roadmap that Tukwila and Poulson are both targeted for 2010.
it’s working quite well in the commercial industry, academic and industrial thanks.
Im im all down for it, x86 everywhere makes the life of the programmers a lot easier.
yeah, I’m all for it.
Some prefer easy life – some others: interesting life…
Writing a play makes life interesting.
Transcribing a play from Russian to Chinese is a chore.
But Writing a Play in Russian that speaks to Russian culture is much more interesting than writing a play in Chinese that tries to speak to Russian culture.
If the goal is to sell tickets and the audience is equally entertained by a universal theme then the preference of the playright becomes insignificant.
Edited 2009-09-23 06:41 UTC
The goal of art should be art, not commerce. Otherwise you end up with “art” like nickelback.
Edit: oh wait, we’re talking in metaphors. Well, You sort of have a point in this case. IT is not art. Good point, bad metaphor.
Edited 2009-09-23 07:00 UTC
PowerPC makes programmers’ lives easier.
ARM everywhere makes programmers’ lives easier.
68000 everywhere makes programmers’ lives easier, and we’d all get back to programming as God intended to boot: in 32 bits.
This much is sure: x86 will not be everywhere anytime soon, at least as long as there are embedded systems.
Aren’t 80386 and 80486 still used quite a bit in the embedded and machine control industries?
I was at the Embedded Systems Conference in Boston yesterday, and indeedy Intel Atoms really were everywhere, but so too were ARMs and a tiny scattering of PICs and pretty much nothing else. (Found out the PIC32 is really a MIPs ISA though). PowerPC and 68K-Coldfire no show even though IBM and Freescale were there.
As much I prefer almost any other 32b architecture to x86, the Atom pretty much nails it in the important features in industrial embedded, low cost, low power, fanless operation, and tools from the desktop on same ISA.
The ARM is far bigger into deeply embedded ASICs, not sure if Atom will ever do that. So the embedded market splits exclusively between not so cheap ($100 several) miniITX/PC104/custom type Atom boards in tight fanless boxes, and ARMs buried deep inside the custom ASICs in very cost sensitive stuff.
For Intel to be truly everywhere, they would have to allow TSMC and others to add the Atom to their core library for any ASIC customer to use.
How so? Unless you’re writing assembly code the CPU architecture doesn’t matter that much anymore.
Even if you were coding in asm, you would not want to touch x86 asm. x86 is a mess and is not making the life of anyone easier. It’s easier to learn 10 different clean architectures and not x86 than to learn x86.
Yeah that’s the same phrase what Intel marketoids are like to wash the brains of the crowd.
Easier? How and what for?
How many different platforms you programmed for?
Learning a new platform is very interesting and boost your experience.
Hiev mentioned …
“… x86 everywhere makes the life of the programmers a lot easier.”
This would be true for system programmers.
For application programmers (like myself), the CPU platform has little consequence in terms of producing code since we (the application programmers) can rely on “standard” libraries (POSIX, pthread, standard C library, standard C++ library,OpenGL, etc) that allows our domain-specific code to interface with any system (ie. a system that supports these “standard” libraries).
e.g.
I deal with C++ software development on OpenSolaris (using free SunStudio tools). One of my in-house research projects is optimisation of rendering of massive landscapes on a planetary level using a high-performance (in-house developed) persistent object database as the data-store for the landscape detail. The core of this system deals with standard C,C++,OpenGL, POSIX, pthread libraries. Higher-level libraries have been built from this core. The graphics engine (G3D Innovation Engine) has some optional inline assembly but the absolute amount of assembly code is so small that any extra time required to maintain the assembly code is a minor issue. While my hardware platform is 64-bit AMD-x86, the same code-base could instantly be ported to the scalable processor architecture (SPARC) platform with the only source code change being the optional inline assembly.
It stands to reason that there would be much more application developers than system (kernel, OS, device driver, etc.) developers and so the “average” developer would probably be more “application”-driven that “sysdem”-driven.
I think it’s foolish to promote the “x86 everywhere” fantasy since you have CPU platforms like SPARC with an “enterprise-ready” design goal that have traditionally contrast the design goals for x86 CPU. The enterprise-readiness, massive multi-threading ability, etc. of the SPARC platform has no use in the general home PC where the latter is the familiar and realistic territory of x86 platform.
Remember, Intel/AMD have used ideas from other platforms (e.g. DEC) to enhance the design of their CPUs.
How much creativity would exist if there was a monopoly ?
Edited 2009-09-24 00:07 UTC
MS/IBM/McDonalds Aims for World Domination, NT/Linux/BigMacs Everywhere!
Film at 11.
They’d probably do a better job than the current congress. I prefer the big mac to the whopper so I welcome our new corporate overlords.
We already can *only* buy viable x86 workstations. The last one to go (almost 2 years ago) was the Power based Intellistation from IBM, and it wasn’t anything to write home about.
It is already a monopoly in perhaps the largest sector of sold processors (I don’t have stats though).
Any idea on how the lowly geek who is not in IT can help?
The only thing that comes to mind would be
1) Buying a slightly overpriced desktop novelty computer that I don’t have much space or need for. Maybe an ARM computer that runs Haiku once it supports ARM better…
2) Getting my lab to buy a non-x86 server when we need one. But being as we will use Solaris, that limits it to SPARC for the foreseeable future, and most of the good SPARC chips these days are good at a type of multithreading not particularly useful for most types of scientific computing. (If only it looked like Rock would ever get here.)
Edited 2009-09-23 03:22 UTC
You are looking to support non-X86 cpus by buying devices that use them? Well you could start with a game console.
Anyways the best comment I heard about x86 was that it is like eating sausage. It tastes fine as long as you don’t think about what goes into it.
Apple gave the ppc a good run, and cell phones will continue to push non-x86 development. Who the hell cares if x86 takes over more of the world. It’s just sausage.
From TFA:
I don’t think so, they’ll probably keep supporting the Moonlight project, so they can always be one step ahead on Windows. But I must admit that the idea of MS looking for Linux developers to port Silverlight would be awesome
Intel’s greatest strength is their manufacturing. They sure can make the hell transistors and small ones at that. With x86 as a dominaning platform they’ve got a good base and they can manufacture it.
Intel’s greatest weakness is x86. “x86” processors don’t actually run x86 instructions, they translate them into a different instruction set that the cpu actually runs. This extra space requires a crapload of transistors and die space. That means an x86 core compared with an arm core has to have way more transistors to do the same amount of work the arm core does.
So to compete outside their market (the embedded space), their manufacturing technology has to kick ass over their competitor’s manufacturing to the extent that they are able to both overcome the issue with extra transistors and even beat their competition to overcome current inertia.
So the question is: how far behind are all the other non x86 manufacturers compared with intel ?
In that case, we can think of it as a compatibility layer. Those seldom makes for superior performance but can make life easier.
This arguments is a bit obsolete now, as most of the die is cache anyway.
I once read an analysis by a chip hardware guy, why the x86 architecture refuses to die.
His argument was like this: RISC-like architectures need no die space for that translation layer, but more instructions have to be stored in cache.
Imagine calculating a sinus: One instruction in x86, a whole algorithm in RISC. The translation layer of the x86 makes the one instruction to become the same algorithm as on RISK internally, but for caching and bus transfer purposes only one 32 bit command has to be considered, whereas for RISK it most often will be a much larger amount.
That leads to “more of the program can be stored in cache” for x86, and subsequently to faster execution. Loosing some speed in the translation layer is the disadvantage this architecture has, but as with all things tech: The best compromise is “The Solution”, no matter how outstanding one of the pieces may be.
I believe it depends on the software.
You don’t copy the sinus routine x times. You do it once and you call it with some kind of jump command. It should still take less space than the translation layer because your software will never use all routines.
Moreover, you can make better routines in software when you know what they are used for. You just need clever compilers.
This is only true if you compare the best CISC with the worst RISC (in terms of code density), say Motorola 86K against MIPS. Modern x86 code is not really very compact. ARM with the Thumb2 ISA has, generally, much better code density than x86. Sure, you can find examples of a single x86 instruction that needs a sequence of ARM instructions, but the converse is also true, but that is really besides the point: You need to look at the average code density over a large set of programs.
Wrong anology. FPU transcendent instructions are very slow. And it isn’t supported in 64 bit Windows anyway. You need to code the sinus in RISC way to be fast. And the code footprint here can be worse for x86 because of long SSE instructions encoding and lack of MAD instruction.
Maybe, but during that sinus calculation by x86 – the RISC will fill and empty its cache several times… there’s wrong assumption, that x86 and RISC are of the same speed; no, RISC is much faster.
The ONLY reason x86 is where it is, is because of Windows. The ONLY reason the Windows world is x86 only is because it’s a closed world.
You have open software on open operating systems, and the whole thing can be ported. You have software repositories and people don’t even have to know their processor architecture.
So ARM can go where Linux can go. So ARM smartbook and by the looks of it, ARM servers. I won’t be surprised to see the return of the ARM desktop, if the desktop still matters….. I think the laptop is taking much of its market, and I bet if ARM smartbooks work, full ARM laptops won’t be far behind.
This isn’t RISC vs CISC, ARM has quite a number of instructions now, and most x86 have RISC like micro code, but ARM gets much better performance per watt because of its RISC roots. But ARM is more than just RISC rooted, you could argue ARM vs RISC.
http://www.heyrick.co.uk/assembler/riscvcisc.html
I would like to see the death of x86 in hardware, but I don’t want an ARM monopoly either. I want competing hardware standards.
In a world freed from closed software, hardware is much less fixed. We all win. As a bonus, portable software tends to be better software as bugs have less places to hide.
Yeah, right. Have you by the way noticed that Intel is now one of the biggest contributors to the Linux kernel?
See, monopolies can win also in the world of free software.
Edited 2009-09-23 14:22 UTC
We’ll see what happens with the Linux ARM smartbooks.
The fact we are talking about the possibility of x86 netbooks being completely upstaged by ARM smartbooks shows the power of a good, free, portable software stack.
I don’t think brute force and loads of PR can win out here.
Intel have contributed to Linux when it has been in their interest. It’s not like there is a deliberate free software strategy to side line x86.
Monopolies are almost always bad for the consumer and the ecology. Customers are stuck paying more then they have to because there is only one place to get a chip they need. With only one chip running everything we throw out any new ideas or revolutions in design. We see from the Microsoft fiasco that one manufacturer is a recipe for disaster. We need to have competition so Intel is forced to come up with new ideas and new strategies.
Internet Explorer is an excellent example of what happens when a monopoly figures it has won. Why put money into innovation when there is no one you have to worry about. IE sat dormant for years while little Firefox started development. Firefox was then good enough to take on IE and Microsoft was so lost and confused they didn’t do anything against Firefox for years. After Firefox grabs 20% of the market then Microsoft decides maybe IE needs an update.
These kinds of things will happen if Intel is allowed to be the only game in town. We need to guarantee that innovation, security, and pricing has as much competition as possible. I sure hope we don’t have to deal with another IE type fiasco except in the microprocessor field. That will be much harder to fix than writing Firefox.
What do you mean by “will”?
When now looking back to the 1990s and 2000s, these things have happened.
If EU is to be believed, like I do in this case, Intel has been found guilty to economic crimes that your beloved Microsoft can only dream of.
http://news.cnet.com/8301-1001_3-10356876-92.html
I especially “love” this part!
‘PLEASE DO NOT… communicate to the regions, your team members or AMD that we are constrained to 5% AMD by pursuing the Intel agreement.’
x86 architecture might not be the best (in fact it isn’t), might haven’t the best design, and for sure is clumsy, bloated and due to legacy, today’s fastest processors must be compatible to the old 8086 manufactured decades ago.
But nobody can do anything. Even Intel tried to replace x86 with much better Itanium arhitecture. Even Microsoft released Windows for Itanium. And that was an obvious failure.
Even if we assume that MS will release a Windows version for -let-s say – ARM, the immense horde of x86 Windows developers will not follow that move and will not compile software for other platforms than x86.
But that might not be so important: even if x86 doesn’t have the best design and is a bit bloated, x86 processors are the best in terms of sheer computing power. That’s why Apple migrated to x86 and that’s why today’s supercomputers are manufactured with x86 chips.
And nothing will change in the foreseeable future.
That’s a pity because I think Intel has one of the best engineering team in the processor market and a ton of bright design ideas which applied to a smarter architecture would make the chips more faster than old x86.
Which is why free software is such a game changer. Everything can be ported.
It’s why ARM smartbooks will be really interesting, free software has reached the point it can offer an easy to use solid complete platform. If Windows compatibility doesn’t matter (and it doesn’t to me), you are free to use whatever processor does the job best. When power requirements count, you can’t fault ARM.
The economies of scale apply already to ARM chips as they are so common. It’s not a cold start.
Very interesting to watch what happens.
Please put an x86 in every device with optmized OS and I will be eagerly waiting in line to crunch SETI workunits and POVRay animations on anything you sell Intel!
My little ASUS 1005 HA has an overclock on it and is making a POVRay animation right now, which is only possible because it has an x86 engine. http://bit.ly/44CHFm