Slashdot Mirror
Will Vanderpool Make Linux More Popular?
Digitaldonkey writes "New Scientist is reporting that Intel's forthcoming multi-core processor architecture, codenamed "Vanderpool", could undermine Microsoft's dominance by letting other operating systems run simultaneously more easily. From the article: 'The chip will allow future machines to run, say, Windows XP together with Linux or the Apple operating system as easily as today's Windows computers run Word and Internet Explorer simultaneously.'"
Yeah, it's that super secret X86 version Dvorak keeps calling for. Don't you know anything, silly?
Taking guns away from the 99% gives the 1% 100% of the power.
Seems like another case of technology history repeating itself. Still, the idea is fantastic although i don't see how a company like microsoft in the article can really benifit from it.
Clustering, servers, dual-booting, emulation, dual-OS, blah-blah, and other geeky stuff isn't going to make Linux popular. It is just going to keep it alive in specialised applications. To become popular, geeks need to stop looking down on the average user and start treating them as a customer and design things even an idiot can use.
Does anyone think their mother or father would switch because of what this article discusses?
Conversely, is this going to get businesses to try a new OS? No. If a business wants to try a new OS, in general, they can afford a machine dedicated to the new OS to try it out.
This will have interesting implications with Microsofts licensing mechanisms. All the virutal machines *should* look the same, and Microsoft shouldn't really care if I run multiple copies of their workstation version on the same desktop. That way, I can clone the OS, apply the latest patches, see if they work without blue screening the system, and then put that system into "production". Just like how they use VM on mainframes.
From March of this year:
Dvorak Thinks Apple Will Switch to Intel
Dogma - "let's just say we'd like to avoid any empirical entanglements."
There have been faster, cooler, better designed processors for years.. PPC/POWER, MIPS, SPARC, ALPHA, PA-RISC, ARM etc... in the case of alpha atleast, cleanly designed from the ground up as a 64bit chip, the others were all designed as 32bit chips with 64bit extensions added later, but the transition went smoother because 64bit extensions were always planned... Compare this to x86, which was originally designed as an 8bit architecture and noone in the original design team considered making a 32bit, let alone a 64bit design.
Unfortunately, new innovative designs, while far superior, dont sell chips for a number of reasons, mostly compatibility and cost, chips dont become cheap until they can be mass produced, no company will mass produce chips unless theyre sure of selling them, people wont buy chips which won`t run the apps they need, and chips with few users wont recieve much or any commercial support from software vendors...
If opensource were dominant nowadays, this would aid the development of new and innovative processors, since the porting of applications would be somewhat easier, and initially could be done by the processor designers. Unfortunately, commercial vendors release binaries which are tied to a particular processor architecture and third party software, and its not commercially viable for them to support anything until theres a userbase (and a userbase wont grow unless theres a software base)
So, you have people tied to existing systems, most notably windows/x86, but in many companies theres legacy apps tied to solaris/sparc or various ibm or other old mainframe systems etc. And, vendors arent going to change this behavior any time soon, because lock-in suits them.. if people are forced to use their products no matter how crap they are, simply because they have no other choice... then the vendor doesn`t need to spend so much money making improvements.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
- It is developed faster than Windows (it has 9 years, Windows has 18) Mozilla has new releases every 3-4 months, IE had its last release in 2001 next should be in 2006
- It is free and computers are getting cheaper, so Windows price will relatively grow making it unattractive
- It is more secure, so it is easier for newbies
- Programmers get the source code, so they can better integrate their programs with Linux
- Linux is being developed mainly by technical people
- Linux has plenty advertising in Linux magazines and Linux portals and they are lots of them
- It is relatively cheap to write portable apps - see Unreal Tournament 2003 and 4, so migration should be easy
See - I can do it too!Really, you cannot predict what will happen, unless these are obvious things (like I predict pigs won't fly upside down under water wearing steel smokings), but then such predictions are useless. Linux is kind of phenomena, because no one ever predicted such a thing would occur. I think it will be decided in the next few years, whether Linux has the potential to replace MS or not.
Reluctant Doc files?
Use either
Crossover office to run Microsoft Office on linux
or
OpenOffice 1.1, now much faster, with improved Doc support. It seems to handle everything I chuck at it!
Games?
Use wine(x)! Severeal Game companies are contributing to it already to make it work with their games!
There is NO reason to use Windows these days, and Microsoft knows this, thats why its FUDDING linux to death!
This could really cut out the relevance of application support behind an OS. Any application not supported by your current OS could be built in with the app and booted separately almost like a Knoppix CD.
Ah, someone else who sees the bigger picture, rather than the trivial idea of making virtualization just a bit faster...
Rather than viewing a program build as bound to a particular OS, each program can act as its own OS. I see this as a fairly logical extension to the idea of multitasking in general - The OS no longer needs to multitask, because the CPU does it explicitly.
This wouldn't benefit every program, and in fact would hurt programs (like IE and Word, to use the parent article's own example) that already work well together. For any application that doesn't need to interact with other programs on the system, however, this would increase both reliability and ease of programming. Sick of Windows' hideous system calls needed to do seemingly simple tasks? Use an open source microkernel and let your program run as its own OS. Sick of requiring OS support for certain hardware features (such as MMX on the PII)? You no longer need it.
This will do a lot to improve PC stability in general, and I look forward to it. To all those who ask "why", or only see it as faster virtualization, I say, "look beyond Windows vs Linux".
Would anyone in the audience be willing to start a GNU version of a virtual machine monitor (VMM)? Writing a VMM only takes tens of thousands of lines of code as opposed to tens of millions of lines for an operating system. The project could be done within a year. Take a cue from the work done on Disco, the VMM developed at Stanford University.
Then, we in the open-source community could feedback to Intel what we want in terms of support for VMM. We could even get help from our uncle, IBM. IBM invented VMs and VMMs back in the 1960s. Unlike Sun Microsystems, IBM has been a strong supporter of the open-source movement and Linux and would surely be willing to help in building a GNU version of VMM.
This is a golden opportunity for the open-source community to impact the future direction of processor development. Is anyone up to the challenge? Would anyone like to accrue the same fame that Linus Torvalds has?
Quickly googled up a link from eWeek:
... providing Apple releases/creates its rumored-but-horribly-unlikely (imo) x86 build of OS X. ... Darwin, which is an OS, just not a very popular one and not much of story. Though AbiWord does run fairly well there with X11 installed. :^)
As Apple Computer Inc. draws up its game plan for the CPUs that will power its future generations of Mac hardware, the company is holding an ace in the hole: a feature-complete version of Mac OS X running atop the x86 architecture.
There have been rumors of the move to x86 for a while. I'm not sure if I buy them -- that's a ton of QA overhead for a potential move down the line, and hopefully the G5 negates any reason for them to move. Not to mention if Apple swapped processors, all the AltiVec-optimized code would be worth creee-ap without having multiple processor *types* in each new, partially x86 powered Mac. And any way you cut it, Apple would still, I'd assume, stick some hardware dongle in there to do what Open Firmware does now: stop cheap generic hardware (or expensive hardware when you talk Pegasos) from running OS X easily. Apple is a hardware company too, you know. Solutions, not just software, etc.
But the point of the article stands, even if the author was overhyping. Anil (the author) really has two outs:
Due for launch within five years, the chip will allow future machines to run, say, Windows XP together with Linux or the Apple operating system...
1.)
2.)
It's all 0s and 1s. Or it's not.
"Sounds a little like how those big iron mainframes run - virtualized os's playing together managed by hardware to a certain degree.
Seems like another case of technology history repeating itself."
Pretty much the entire personal computer history is "history repeating itself". The word for this is trickle down.
This has nothing to do with multi core cpus.. This is for running multiple OS`s simultaneously on a single machine, wether that machine has 1 or 512 processors.
And yes, Intel do own the X86 ISA.. look at recent history...
AMD tried to improve the x86 isa with 3dnow, intel didnt care...
Intel added sse, AMD followed with their own implementation, intel created the x86 isa and amd have to remain compatible. And yes, Intel did create the basic architecture and they still dictate the development of it. AMD might change that with x86-64, but we`l have to wait and see there
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Linux used TSS to do task switching in older version like 2.2. The switch to software has been made because : * you can check the data between task switching (E.G. segmentation registers), which cannot be done with far jmp * the time to switch task is about the same, but I it easier to optimize software than hardware...