For year now, the x86 microprocessor market has been dominated by Intel and AMD, and the rivalry between the two companies forced both to be innovative in order to gain a competitive advantage over the other - benefiting customers. With the rise of 'mobile internet devices' and low-power budget notebooks, this new market will be enriched by not only Via, but also nVIDIA.
Post a Comment
Member since:
2005-07-06
One advantage that Linux and the BSD's have in this arena is portability. There are already ARM ports. Windows currently is tied to X86, even though NT was ported to Alpha and MIPS, so perhaps it could again be ported (I doubt it). Instead, only Windows Mobile competes in this realm. I afraid I'd much rather have a nice Ubuntu or FreeBSD with Firefox or Konqueror than WM with IE. Go Free OS's!
Member since:2006-01-12
Microsoft engineers have said that it wouldn't be hard to port Windows NT kernel to new platform, they did it on IA-64. But truth is that there is very little markets for such a product. It's intresting to see if Nvidia based machines will use Windows Mobile 7 like Qualcomm new ones.
Member since:2006-06-11
I was thinking that I read a long time ago that NT wasn't originally developed on x86 hardware. This bit from Wikipedia confirms it:
In order to prevent Intel x86-specific code from slipping into the operating system by developers used to developing on x86 chips, Windows NT 3.1 was initially developed using non-x86 development systems and then ported to the x86 architecture.
NT runs on 3 architechtures right now: x86, x64, and IA64. Portability doesn't seem to be an issue. Userland may be a whole different issue though.
Edited 2008-06-06 04:44 UTC
Member since:2005-07-22
Member since:2006-01-11
Member since:2007-02-11
Yes, but porting code from x86 to ia64 or x86-64 is fairly easy, since the two architectures are reasonably similar. As long as these small-ish devices keep an architecture that has 86 in its name, I hardly think it will lack third party software. Porting Windows to something that has nothing to do with x86, like PPC, ARM and so on -- now that's something.
If we were to talk just about the operating system, Windows CE/Mobile is a fair solution, especially since the development tools are quite good.
Member since:2005-08-11
"Porting Windows to something that has nothing to do with x86, like PPC, ARM and so on -- now that's something. "
Windows used, and still, runs on the PPC, I doubt that it would take a massive effort to re-port it. A version of Windows runs on the xbox 360, which is a 3 core PPC design, it just runs a different userspace.
Edited 2008-06-06 20:06 UTC
Member since:2007-09-08
The Xbox 360 is really not using Windows. It has a kernel based on the Xbox OS, which in turn was loosely based on the lowest-level basic parts of the Windows NT 5 kernel. By now, it's less closely related to desktop Windows than Windows CE is.
Anyway, the problem isn't that the kernel isn't portable. The problem is that the drivers and userland are not portable. Without it's huge base of existing software, Windows is effectively worthless. An ARM version of Windows could do nothing more than run the default included apps. An ARM version of Linux can do everything an x86 version can, except run the Flash plugin, Wine, and commercial games.
Even if you're selling the thing as an appliance, so third-party software doesn't matter, Microsoft are the only ones who could legally offer an appliance-style version of Windows. They don't - you'd have to use one of their desktop systems. At least with Linux, hardware manufacturers can roll their own.
That's ignoring WinCE, of course, but there's not a lot in the way of pre-existing Windows CE software to work with. A hardware manufacturer might write their own UI, but I doubt they'd write their own office suite or web browser.
Edited 2008-06-07 12:37 UTC
Member since:2005-06-30
Sure, NT4 ran on X86, Alpha, MIPS and possibly other architectures (PPC?). I don't think portability for the OS itself is an issue. However, demand for backward compatibility mixed with lack of portability in legacy code led to the domination of the x86 architecture.
Anyway, recent processor architectures have so little to do with the original 8086... The x86 instructions are internally translated to micro-ops. Therefore, the ISA doesn't seem to stifle innovation.
Member since:2007-08-22
Except with regards to what Intel's Atom and similar processors (such as this one from nVidia) - Linux and the BSDs would likely be competing with WinCE, which is already on ARM, probably MIPS too. (If I recall correctly, most WinCE devices are ARM-based, not x86-based; at least historically.)
BTW, when a vendor wants to support WinCE, they get the source code to modify too - at least for device manufacturers. So they can port it to whatever platform they choose to.
Member since:
2005-07-15
Member since:
2007-03-14
Member since:
2006-12-27
Member since:
2005-07-06
Member since:2005-06-30
Not sure about this so-called monopoly. The microprocessor market goes well beyond the PCs. Desktop processors are merely sharing about 3% of the whole market in volume!
We don't hear much about other architectures in the PC market because the two major commercial OSes (Windows and OS X) are now running exclusively on the x86. Furthermore, other markets have different needs (power consumption, cost, etc) than the PC market.
Member since:2006-05-26
Correction: OS X is still running on PPC (admittedly a dying platform in the consumer desktop space, though strong in consoles and many embedded spaces), and is currently running on ARM in the form of iPhone and iPod Touch platform (and I think AppleTV, but I'm not sure about that one, but it makes sense), and who knows what Monday will bring to light, exactly? The iPhone OSX AFAIK is not truly comparable to Windows Mobile: it's a very compatible kernel codebase, with a lot of userspace compatibility as well, with the biggest difference being the GUI SDK, which (unlike Windows Mobile) is well-suited to the portable realm.
Member since:2005-06-30
Member since:2005-07-06
Member since:
2006-01-08
First of all, I really wish that articles would stop comparing Nano to Atom. In reality, Atom is Intel moving into Via's territory (low cost, low power, in-order x86) and Nano is Via moving into Intel and AMD's territory (faster, out-of-order x86) while retaining most of their cost and power attributes.
The Atom, at 1.6Ghz, has compared similarly to 900Mhz Celeron M's in benchmarks (drawing much less power, of course), while the Nano just outperforms later Pentium M designs clock-for-clock while drawing 1/3 of the power. This is pretty big for Via, which is now performance competitive with a desktop chip Intel still actively sells (PM 1.6Ghz) while besting it at cost and power consumption.
They simply don't compete on a technology level, though there may be some overlap in netbook devices due to similar cost.
There will be Nano-based Mini-ITX 2.0 boards and netbooks before too long I'm sure, which is exciting.
I would absolutely love to see a Linux-based Tegra netbook or small PC though... Arm Linux is, what, the third largest and fastest growing arch now? Debian ARM is pretty good from what I hear. If nVidia commits to supporting Linux on the Tegra with specs and/or drivers for hardware accelerated Video, 3D and 2D graphics and all the other nice bits it offers, it would be huge.
Edited 2008-06-05 23:47 UTC
Member since:
2007-07-27
watch: http://www.youtube.com/watch?v=XXYshhuJzh4
they sure seem to be mighty proud of their chip. would be interesting to know the technical details of the systems.