MS, Intel "Goofed Up" Win 7 XP Virtualization

clang_jangle writes "Ars Technica has a short article up describing how Microsoft and Intel have 'goofed up' Windows 7's XP Mode by ensuring many PCs will not be able to use it. (And it won't be easy to figure out in advance if your PC is one of them.) Meanwhile, over at Infoworld, Redmond is criticized for having the 'right idea, wrong technology' with their latest compatibility scheme, and PC World says 'great idea, on paper.' With Windows 7 due to be released in 2010, and Redmond apparently eager to move on from XP, perhaps this is not really a 'goof' at all?"

Comments

2010?

[swaq]

I thought it was leaked that it would be released this year?

Re:2010?

[Anonymous Coward]

Based on the 3 year rule, most people say 2010.

All the postings from Microsoft developers, combined with their probably-intentional leaks are hinting strongly at a late-fall release.

Re:2010?

[postbigbang]

No, most people don't say the "three year rule". Where have you been?

Microsoft's suffering revenue shortfalls because Vista was a bad idea. You'll see the tip in September, real McCoy in October. That's not 'late fall'.

Re:2010?

[JoelisHere]

from the currently available Windows 7 release candidate info page (http://www.microsoft.com/windows/windows-7/installation-instructions.aspx)

IMPORTANT: The RC will expire on June 1, 2010. Starting on March 1, 2010, your PC will begin shutting down every two hours. Windows will notify you two weeks before the bi-hourly shutdowns start. To avoid interruption, you'll need to rebuild your test machine using a valid version of Windows before the software expires. You'll need to rebuild your test PC to replace the OS and reinstall all your programs and data.

Re:2010?

[maxume]

Don't get too excited. Microsoft still had decent growth in 2008 and is selling into a terrible economy right now (making it a little silly to attribute the entire drop in revenues to Vista).

Hardware Virtualization needed.

[mister_playboy]

AMD has placed this support in almost all of their recent chips, but Intel has been more stingy with it.

It's necessary to use 64-bit guests in Virtualbox, but VMWare can make due without it.

Re:Hardware Virtualization needed.

[AndrewNeo]

It surprised the heck out of me when I found I could run 64-bit guests on a (32-bit host OS, 64-bit hardware) with hardware virtualization, at least on my AMD.

Re:Hardware Virtualization needed.

[pegr]

Apple did this, not once but twice. Why is Redmond so afraid of trading out the basic underpinnings? I guess they married the concept of permenant backwards compatibility when they used that very stick to beat OS/2 into the ground.

Is Rosetta Stone a good technology? No, but it got users over the hump. (It was, however, a great hack...)
How about Fat Binaries? Good lord, Win binaries are fat enough already!

There's no good solution, so Redmond has to go with "good enough" to get users over to "the other side". Hey Bill! Maybe they don't want to go...

Re:Hardware Virtualization needed.

[SatanicPuppy]

This is actually very similar to what Apple did...They kept OS 9 support on all systems that had the old power PC processors. Once you bought intel, however, no more OS 9, no matter what version of OS X you were using.

Like anything else, users will have to decide for themselves if there is anything that is good enough to make them upgrade.

Re:Hardware Virtualization needed.

[EvanED]

Apple did this, not once but twice. Why is Redmond so afraid of trading out the basic underpinnings?

And what's Apple's market share? What's Microsoft's?

You say it yourself: "they used that very stick to beat OS/2 into the ground."

I would say that the single biggest reason that Windows is as prevalent as it is today is that to a very large extent, MS has maintained backwards compatibility at almost any cost. I can only think of a couple exceptions: transitions to the NT line stopped some old DOS programs that access sound cards and stuff directly from working, XP SP2 made a few similar strides (I don't know details), Vista makes a couple more, and x64-based Windows drops support for 16-bit programs (but this is largely the fault of AMD/Intel rather than MS, who would have had to work around processor limitations since 16-bit instructions aren't available in 64-bit mode).

But even with Vista 32-bit, my experience is that each of the three or four DOS programs from the mid-80s still ran. There are few systems that can claim this lineage. So it's no wonder to me that MS doesn't want to give it up.

And it's only recently that the pile of compatibility hacks and inability to make fundamental design decisions has caught up to MS and been harming them from the market's point of view.

Re:Hardware Virtualization needed.

[arth1]

AMD has placed this support in almost all of their recent chips, but Intel has been more stingy with it.

Also, with Intel, it's not enough that the CPU and chipset supports VT-x, it also has to be enabled in the BIOS. Some manufacturers disable it, and some (most notably Sony) often won't even show the option in the BIOS set-up, making it permanently turned off. All to save a few bucks in support costs.

Re:Hardware Virtualization needed.

[digitalunity]

I'd be happy if Vista included virtualization technology for DOS6.2 on a 386. That would allow much smoother operation of very old programs that some of us still use, or want to use at least.

Re:Hardware Virtualization needed.

[srmalloy]

Eternal backward compatibility is an inevitable result of Microsoft's "Every current Windows user needs to buy a new OS from us every 3 years" cash-flow model. Look at how lackluster the adoption of Vista has been, when it's just the lack of any clear advantage from upgrading. Now consider what purchasing decisions will be made when the cost analysis is choosing between "continue using our existing OS and applications" and "Upgrade to the new version of Windows and replace every single application we're using" -- the benefits to be achieved from upgrading will have to be very significant to justify a complete replacement of all applications, with the attendant problems associated with legacy data and applications. Not to mention that the lead time on releasing a new OS that broke backward compatibility would need to be much higher to allow application developers to get the development tools for the new OS and build compatible applications.

Re:Hardware Virtualization needed.

[arth1]

Indeed they don't. Apart from the BZ-series, all Sony Vaio PCs have VT-x disabled, and no way to turn it on in the BIOS setup. Even if you buy a brand new Core 2 Duo P9600 Vaio Z with 4 GB RAM for 2-3000 bucks, it won't be an option.

Re:Hardware Virtualization needed.

[Runaway1956]

I see this as something like poetic justice. I have never forgiven Intel for their bright idea of making processors that identified themselves to the world at large. Anonymity has it's place in this world, and that chip was a monumental boner. More, I see most of Intel's business decisions as based on greed. AMD is greedy, to be sure, but their decisions seem designed to move computer science forward, while capitalizing on that move. Intel? Money first and foremost, with computer science a side benefit.

How about an automobile analogy? Detroit is folding today, because they've been focused on the almighty dollar, rather than producing better transportation.

Lesson learned? Money is important, but it should never be "the goal". Improving the world should be the goal.

Re:Hardware Virtualization needed.

[Thuktun]

I would say that the single biggest reason that Windows is as prevalent as it is today is that to a very large extent, MS has maintained backwards compatibility at almost any cost.

Curious. At home, I've got a large collection of children's educational and gaming software written for DOS through Windows 98 that utterly fail to run properly in Windows XP or Vista. I haven't experienced this compatibility of which you speak.

Re:Hardware Virtualization needed.

[Belial6]

Exactly. Windows has always had backward compatibility problems, and virtualization is obviously the future. By emulating the old environments, they can clean out the old cruft while keeping a better level of compatibility than they could ever hope for with their various 'compatibility modes'.

Re:Hardware Virtualization needed.

[earlymon]

I would say that the single biggest reason that Windows is as prevalent as it is today is that to a very large extent, MS has maintained backwards compatibility at almost any cost.

I have one pgm from OS X 10.0 that no longer worked at some point (10.2? Not sure.) It was freeware. I had to update OroborOSX a few times with my OS X update. I have a platform running Tiger (10.4 - current rev now is 10.5, Leopard) - and it is running my Microsoft Office X that I bought when that was a 10.1 machine. At some point, IE wouldn't run and MS said they would no longer support it under OS X (I'm just rounding out the list - I could care less if that was Apple's or MS's doing.)

None of my user files are affected by upgrades.

Like you, I can point to a few - very few - examples of OS X not bending over for backwards compatibility.

I can't say what backwards compatibility OS X has with OS 9 and prior (supported via Classic mode for a while on OS X) - other than to say - pretty much none.

By the time OS X came out, Apple had lost all sorts of market share - are you suggesting that that was because they weren't providing backwards compatibilities?

I just cannot believe that a large extent of the reason for MS's market share is their over-the-years backwards compatibility.

You are correct in that they did have that, and I've seen Win users over the years tout it as important, and then brag to me that OS X didn't have that (info source: see flying monkeys). Then those same users would get slammed when technology moved on and not complain, because whatever it was had a good run before being obsolesced.

My take on it is much simpler: MS saved money over the years by not upgrading fundamental parts of their OS until/unless absolutely forced to do so (see: Win32) or sometimes never. Now, their own technical inertia may kill them.

OS X is gaining market share. I'd like to believe that Linux is as well, but I don't know (maybe it's dropping). But quite simply, that has nothing to do with backwards compatibility.

Re:Hardware Virtualization needed.

[PitaBred]

Try Dosbox. But if you did that, you could get off Windows entirely ;)

Re:Hardware Virtualization needed.

[CarpetShark]

I haven't experienced this compatibility of which you speak.

That's not what you said earlier:

software[...]that utterly fail to run properly

Re:Hardware Virtualization needed.

[Penguinoflight]

This has to do with supporting virtualization only when it makes sense. By not running XP-mode on systems without VT support Microsoft is ensuring that the experience is acceptable. Nobody wants the vista fiasco again where Microsoft fell to the pressure of OEMs allowing incapable hardware to be labeled as "Vista Capable." If Microsoft can educate users to the stupidity that intel applies to their processors, all the better.

Difficult?

[blueg3]

I suppose it depends on your definition of "difficult" -- it's not particularly hard to find out if your processor supports virtualization extensions.

The Ars Technica is terrible -- it implies that it's a complete mystery why a virtualization system would require processor virtualization extensions to function.

I'm also not entirely sure it's reasonable to call a logical design decision you disagree with a "goof". I would hazard a guess that requiring virtualization extensions is intentional, not a mistake.

Re:Difficult?

[nlawalker]

Previous virtualization systems did not require processor virtualization extensions to function.

They're calling it a "goof" because it would have made more sense *not* to require the extensions and use them only on an as-present basis to enhance performance. This is especially appropriate given that that many of Intel's offerings are lacking VT, and Virtual PC 2007 (the foundation for XPM) does not require extensions, but can use them if they are present.

I imagine that the reason Microsoft requires them is that they wanted to have an excellent baseline for performance of XPM on all machines. They made the mistake of assuming that many/most machines have virtualization extensions, when the article states that that is not the case.

Re:Difficult?

[not already in use]

I love now how slashdotters are faulting Microsoft for going the safe, secure, performance-minded route.

Re:Difficult?

[je ne sais quoi]

I went to Intel's Core 2 Duo list of chips and counted. 18/60 chips don't have the VT extension. That's 30%. I have no idea on the relative popularity of these chips but if 30% of the chips out there can't run the XP emulation mode, I'd say that's pretty significant, wouldn't you? It's not like these chips are ancient or anything either, I'm guessing their Core 2 Duo is their most popular chip for desktops right now and maybe even laptops.

Re:Difficult?

[Blakey Rat]

Wait, you mean this isn't going to be offered on all/most versions of Win 7?

You mean exactly like Microsoft's been saying since they first announced it? Yes.

It'll go on Windows 7 Business/Enterprise and Windows 7 Ultimate (which has everything.) No home versions will get it.

Re:Difficult?

[EvanED]

Actually, yes they kind of did. Before VT extensions the guest OS needed to be aware of the fact that it was not running on bare metal.

No it didn't. In fact, that was one of the primary benefits of Vmware over Xen for a long time -- that the guest did NOT need to know that.

Various hacks by VMWare and Xen (NT kernel hacks, funny drivers, etc)

Vmware didn't require NT kernel hacks or funny drivers at all. They provided some for the video and mouse, but they were only to (in the case of video) improve performance and (in the case of the mouse) increase functionality a tiny bit, and weren't strictly required. In fact, the only way to install them was once the guest was already up and running.

Virtual PC -- a Microsoft product (at least after they bought it) -- also supported running unmodified guests.

(I can't speak to anything Xen did.)

I hate to ask the obvious

[fermion]

Why does MS need a separate XP mode? Why are the two so different that one needs a separate product, virtual PC to run the code? Why do they want XP to run on a virtual machine at all? It this decision based on the way Windows work, or does MS just not want such an ability integrated into the OS.

The reason I am confused is because this would have been great for the Vista transition, and seems to be old technology. Over ten years ago Apple included this capability in OS X, allowing OS 9 application to run in the classic environment. Apple also included bundles to allow the transition from 68K to PPC, and later PPC to Intel. Why did MS not do the same, and why are the including a hack solution at the last minute.

Re:I hate to ask the obvious

[blueg3]

The solution Microsoft is adding to Windows 7 is the same solution Apple used for the OS 9 - X transition. Classic was a second operating system that ran essentially as a virtual machine.

Re:I hate to ask the obvious

[blueg3]

An interesting fact, yes -- although Windows 7's compatibility with Vista isn't the issue, it's its compatibility with XP. Try suggesting that Vista and XP are compatible.

Interesting though it may be, I don't see the point. The solution Microsoft is adding to Windows 7 for XP compatibility is, in fact, the same solution Apple used for OS 9 -> X.

Original commenter wanted to know why Microsoft isn't using the same approach Apple used, instead of this "hack solution". The answer is that Microsoft is.

Who Cares?

[siuengr]

I don't understand why people are making such a big deal about XP Mode. It is meant for enterprise systems that have millions invested in software that is difficult to convert. 99.9% of people are not going to be using XP mode.

Re:Who Cares?

[PPH]

Because enterprises that have millions invested in software have the clout to beat vendors over the head to make them support the latest technology du jour. Its all us little people that, individually will get told to fuck off when we call to complain that our apps don't run on Vista or W7 native mode.

How to figure it out

[bflong]

When running Linux, open up a terminal and run this:
echo -n "Does my cpu support virtualizaiton? "; if `cat /proc/cpuinfo | grep -q svm || cat /proc/cpuinfo | grep -q vmx`; then echo Yes; else echo No; fi

Another issue you may have is if your system has the virtualization functions disabled in BIOS. Some laptops do this, and have no way to turn them on. My Dell D820 works fine.

Re:How to figure it out

[Mad Merlin]

echo -n "Does my cpu support virtualizaiton? "; if `cat /proc/cpuinfo | grep -q svm || cat /proc/cpuinfo | grep -q vmx`; then echo Yes; else echo No; fi

This is even easier:

echo -n "Does my cpu support virtualizaiton? "; if `cat /proc/cpuinfo | egrep -q '(svm|vmx)'`; then echo Yes; else echo No; fi

Re:How to figure it out

[entrigant]

cat foo | grep ...

seriously?

Re:How to figure it out

[bu1137]

This is even easier:

egrep -q "(svm|vmx)" /proc/cpuinfo && echo yes || echo no

Re:How to figure it out

[pthreadunixman]

You both win the useless use of cat award.

Re:How to figure it out

[DaleGlass]

Actually I think that's a common artifact of a process of gradual refinement.

You start with:

cat /proc/cpuinfo

To get an idea what the contents of the file are like.

Then you figure out something to search for, or to exclude. A quick way of doing that is to recall the previous history line (up arrow), and tack on grep on it:

cat /proc/cpuinfo | grep '^processor'

Then you refine it further, with another history recall and tack another command at the end:

cat /proc/cpuinfo | grep '^processor' | wc -l

Then sometimes you copy/paste that into a shell script and earn another "useless use of cat" award.

Starting with grep is uncomfortable because you have to grep for something, and you often don't know what that should be before seeing the file's contents. Unless you realize you can grep for '.*'.

Oh, it's a goof all right

[localroger]

The only reason MS is putting this silly scheme into 7 is the large number of corporate interests who have apps that will not run on Vista or 7 natively, and which they do not want to rewrite. The virtual machine was supposed to get them to stop demanding XP from their vendors since there would be a solution. Only it might not be such a reliable solution, particularly on those millions of boxes which won't be quite new but also won't be quite old enough to discard which are in use today.

This is a very critical problem for Microsoft. I have heard people who would never have even looked at a non-MS solution two years ago whispering about Macs and Ubuntu. If migrating is going to involve a vast amount of unscheduled pain, reinstallation, down time, and retraining, do you migrate to the next level of the company which is screwing you or look for an alternative?

Seven has to solve the problem of legacy apps that don't run. If it doesn't, the Mexican standoff will continue with Seven in Vista's place, and one or two Fortune 100 shops throwing their hands in the air and switching FOSS could start a stampede. The unlikeliness of that, while high, decreases just a bit for every day the current situation persists.

Re:A minor update

[WaXHeLL]

That suite of applications that you're testing doesn't accurately represent the target population for XP emulation.

XP Emulation is primarily geared towards businesses with legacy/custom business applications which have not been re-written for Windows Vista/7. Not to mention, some of those vendors for those business applications will charge large hefty upgrade fees to get their software compatible with the newest versions of Windows.

Why ARE there new Intel CPUs without VT-x?

[ZorinLynx]

What I wonder is why there's still NEW processors coming out from Intel WITHOUT the VT-x extensions? These extensions have been around at least since the original Core Duo days; shouldn't they be standard on all Intel CPUs by now?

I boggled when I learned there were still new CPUs being sold without the extensions. It's not like it's something that's hard to implement; the work is already done.

Re:Why ARE there new Intel CPUs without VT-x?

[fuzzyfuzzyfungus]

Price discrimination/market segmentation.

Disabling that feature is an easy way to make an otherwise adequate product unsuitable for corporate buyers(IIRC, some of Intel's "Active Management" sauce either depends on, or is bundled with, VT extension support). By disabling the feature on some chips, they can capture more of the surplus value. Pretty much the same reason that all versions of Vista aren't Ultimate.

Re:Why ARE there new Intel CPUs without VT-x?

[lukas84]

Only the low-end CPUs don't have them.

So if you bought a machine aimed at corporations, it will have the VT extensions. For example, all current ThinkPad R series and Tseries have them, and all Lenovo M series and ThinkStation products also have them.

All bets are off if you bought one of those shitty 699$ 17" laptops with horrible screen resolution ;)

XP Mode can't play games, it's a lot of work to maintain (seperate domain account, seperate users, seperate AV, seperative firewall). It's aimed at small businesses - not at home users (they don't need it), and not at big corporate environments (There's MED-V for them).

Re:Why ARE there new Intel CPUs without VT-x?

[Nimey]

Objection, m'lud! The Q8600 is not a "low-end CPU". It's not as high-end as a Q9600, but it's no Celeron, and it doesn't have VTX instructions.

Re:what about the common denominator?

[argent]

They will have to either replace legacy software (which is difficult for some clients) or buy a high-end computer just to get a decent upgrade for Vista/XP.

Or just stick with XP.

Re:Ask Slashdot

[flappinbooger]

Ummm, compared to where you're at, just about any dual core proc. Are you really concerned about price? Just get a C2D, ASUS, 4GB ram. That's bang for the buck. Going with 8GB of ram you must really be planning to go all out with the vm's. Most people will tell you that you only really need enough ram so you don't swap with the vm's, otherwise it's overkill.

But, if you're going 64bit linux and plan to keep this one as long as you've kept your last one, maybe 8GB is the way to go.... I am running 64bit linux with XP, Vista, and win7 in VMware right now, and it's just fiiiiiiiiine......

Re:I've tried it

[arth1]

The parent isn't insightful, because he erroneously believes that a new machine guarantees that virtualization will be available. Not so.
Buy a $3000 Sony Vaio-Z, for example, and despite it having a chipset and a brand spanking new Core 2 Duo CPU that both supports VT-x, the manufacturer has chosen not to give the users the ability to turn it on in the BIOS, so it doesn't help one bit.

From the...

[zarmanto]

From the "Department of redundant redundancy department".

(For those of you who actually read all three linked articles... or is that, all two?)

Are the slashdot editors getting desperate?

[Sycraft-fu]

Seriously, this Windows 7 stuff is getting silly. That Slashdot isn't a pro MS site is fine. How about more Linux news, less MS news then?

It gets tiresome to see all these bullshit "OMG Windoze sux!!!1111one" stories any time a new version is coming out. Just leave off it already. If you don't like Windows 7 that is totally fine, but that isn't any reason to try and spread FUD about it. Make no mistake, that's what all this is too. They are trying to find minor things to pick on and make them out to be major problems. They are trying to say "Oh this will be a horrible OS!" They are trying to seed fear, uncertainty, and doubt.

Also I kinda think it shows the opposite: If all you can find to report bad are extremely minor things, then I guess it really isn't that bad, is it? I mean let's see what criticisms they've tried to blow up lately"

1) Windows doesn't show extensions by default.
2) Windows 7 isn't all that much faster than Vista SP1, and the release version of Vista was slower than XP.
3) An optional Windows 7 addon, that most people will never download, requires a CPU addon that not all CPUs have.

Oh gee wow, what a problematic list. I mean really, if that's all you can come up with, if that's the worst of the worst, the stuff that's headline worthy, I think really that shows that 7 is a good OS, not a bad one, because it's all a bunch of BS. As a quick example for each point:

1) So what, every version of Windows since 95 has been like this, and in Linux, anything can be an executable. You can have any extension or no extension and run it.
2) This is a fake comparison. Vista at release was slower than Vista now, a better comparison is Windows 7 to XP directly, in which case 7 does pretty well. Also, new OSes are usually a bit slower, due to new features, what else is new? DOS is screaming fast, but rather worthless.
3) Very few people will ever get this, because it just isn't needed. Native compatibility is extremely high in Windows 7. This is for businesses who have some odd old apps. It is just a nice, free, addon is people want it.

So please, can we stop with the FUD? If there's real news worthy 7 stuff, post it. If not, then just ignore it, because right now it seems like they are grasping at straws to try and find things wrong with 7.

Re:A minor update

[Ephemeriis]

That suite of applications that you're testing doesn't accurately represent the target population for XP emulation.

XP Emulation is primarily geared towards businesses with legacy/custom business applications which have not been re-written for Windows Vista/7.

We run an ancient version of Televantage here.

The Televantage server itself is still running NT4. The client software refuses to run on anything newer than Windows XP SP1.

The solution has been to go ahead and update our machines to SP2/SP3/Vista/whatever and run Televantage inside a small virtual machine running Windows 2000 SP4 - it works great.

This is the kind of problem the XP-mode is intended to address.

No Virtualization on some new Intel

[clarkn0va]

AC obviously didn't read the article, which states clearly that Intel uses VT availbility as a market segregation tool.

A recent example would be the new Core 2 Quad Q8400, now with less VT!

I don't think the city is at fault

[Tarlus]

Meanwhile, over at Infoworld, Redmond is criticized...

Whenever people refer to Microsoft as "Redmond" it sounds so condescending and ignorant. They're called Microsoft. They're located in Redmond, WA. But Microsoft != Redmond.

Re:Ask Slashdot

[phantomcircuit]

Most people will tell you that you only really need enough ram so you don't swap with the vm's, otherwise it's overkill.

And they would be wrong. Disk caching in memory has a significant impact on performance. If you are utilizing a significant percentage of your memory then there will be little caching.

I have experienced this myself on a system with 2GB of ram. While running a single VM utilizing 1.1 GB of memory the system was responsive and snappy. Running two VMs using 1.8 GB of memory the system would crawl to a stop waiting for disk operations to complete.

Fortune has the answer

[kybred]

Apple did this, not once but twice. Why is Redmond so afraid of trading out the basic underpinnings? I guess they married the concept of permenant backwards compatibility when they used that very stick to beat OS/2 into the ground.

The fortune program describe this very well...

"I've finally learned what `upward compatible' means.

It means we get to keep all our old mistakes."

-- Dennie van Tassel

Why MS made this decision

[steveha]

The ArsTechnica article makes an interesting point: could Microsoft have done an XP mode that works on all CPUs?

PC virtualization has been around for years, and predates the special instructions. There is a hack called Binary Translation (BT) where a VM system patches the memory image of the guest program to cause a trap where the guest program uses any difficult-to-virtualize instruction.

http://en.wikipedia.org/wiki/Full_virtualization

The new Windows Virtual PC feature is based on the old Microsoft Virtual PC. Microsoft Virtual PC does not require the virtualization instructions; it can run using BT. So, the point ArsTechnica asked is: why did Microsoft require the virtualization instructions?

I'll try to answer that question. But first, I'm going to rant.

Microsoft has made this much weirder and more confusing than necessary. The new feature is "Windows Virtual PC" and the old, rather different feature, is "Microsoft Virtual PC". In three years, will we have some new third thing that is completely different and is called "Microsoft Windows Virtual PC"? I'll use some abbreviations: I'll call the shiny new Windows 7 virtualization solution, Windows Virtual PC, "W7V" (Windows 7 Virtualization). I'll call the old Microsoft Virtual PC "VPC" (Virtual PC). My first draft of this article was full of "Microsoft Virtual PC" and "Windows Virtual PC" and it was hard to keep track of which was which. Also, Microsoft has broken their web site: links that used to go to VPC are now redirected to W7V. If you are trying to get information on VPC, ha ha! You lose. I was able to find the download page for VPC 2007, but all the links for information now redirect to the W7V page. <end_rant>

So, why did Microsoft require virtualization instructions for W7V? I'm just guessing here, but I think it's pretty obvious.

Take a look at the comparision page for Windows Virtual PC:

http://www.microsoft.com/windows/virtual-pc/features/compare.aspx

W7V adds some new features over the old VPC. Smart cards work, USB devices work, storage drives can be shared. This means that Microsoft did a nontrivial amount of work for W7V. I'll guarantee you that it was easier to just require the virtualization instructions than to try to use BT hacks across the whole Windows XP infrastructure; and this requirement slices away a whole bunch of old computers that now don't need to be tested for compatibility with the new W7V features.

So, the work to create W7V was easier, and testing and support costs reduced, by this decision. Since only the very cheapest new CPUs don't have the virtualization instructions, and this feature was chiefly aimed at corporate customers (who usually don't buy bargain-bin hardware), this decision was likely viewed as a no-brainer.

VPC is still available; customers who have old hardware and don't need the full features of W7V can just use VPC. And VPC remains a free download. (Of course, those customers could also switch to Ubuntu and run their old apps in VirtualBox. I'm just sayin'.)

steveha

Re:Ask Slashdot

[EvanED]

And they would be wrong. Disk caching in memory has a significant impact on performance.

This is even more true in a virtualized environment. (My data is a little old (based on pre-VT Vmware) and I'm doing a lot of inferring of how things should work, but I don't have much reason to believe that what I describe changes much with a VT-based hypervisor.)

If your OS is on the bare metal, when a program makes an I/O request, there's a syscall trap that goes to the OS which issues an I/O instruction to the hardware; when it returns, the OS sees the interrupt, schedules a task to copy the data to userspace, and then upcalls to the userspace.

If your OS is virtualized, when a program makes an I/O request, there's a syscall trap which goes to the guest OS which issues an I/O instruction; but then that I/O instruction is trapped by the VMM, and the instruction is emulated. Similarly, when the interrupt comes back, I think it should be the hypervisor that receives it (though maybe not), which then simulates an interrupt in the guest OS, which then upcalls to the originating process.

So basically, the VMM adds an extra layer to each I/O operation. If you're doing a lot of I/O, this can become significant. (It's worse under old (non-VT) Vmware, because the syscall the user application issues is then caught by Vmware which simulates a syscall in the guest OS which issues an I/O instruction which goes back to Vmware, so there are TWO extra layers compared to the bare metal.)

I have a dream...

[NotInTheBox]

Well Microsoft, that is what you get when you don't build your own hardware like Apple does.

Microsoft should get out of the software business and start designing their own computers together with their own software. Want Windows? Buy that WinBox from Microsoft!

That way you know what kind of hardware is needed and you can drop support for all kinds of crap.

Let Dell, HP, etc try selling computers without an usable OS.

Re:Ask Slashdot

[asdf7890]

You'll need a 64-bit OS to see the extra RAM in your virtual hosts. (I think. As stated, you can run a 64-bit guest on a 32-bit host if you have hardware virtualization acceleration, but I assume you still need a 64-bit OS to have the guests see all the RAM.)

Not quite. You need a 64-bit kernel to use RAM above 4Gb (sometimes above 3.25Gb - it depends on your hardware and OS). Though in Windows this does mean you need a complete 64-bit system, with Linux you can run a 32-bit userspace on a 64-bit kernel. No individual process (so no individual virtual machine in most cases

will be able to use more then 3Gb but over-all you can use the lot (and any not used the host OS will use for cache/buffers as needed).

I run my home server this way, as when I built it the CPU was not 64-bit. A Motherboard+CPU+RAM upgrade later I had 64-bitness and 4Gb RAM of which 3.5Gb was seen - so I moved to a 64-bit kernel to use the other 0.5 (as easy as "aptitude install [name-of-64-bit-kernel-image-package]" under Debian/Etch) but didn't fancy messing around converting my userland setup.

When I next upgrade the OS (to Debian/Lenny, in the next few months) I'm going to do a fresh install full 64-bit, but until then the this arrangement seems to work just fine. The machine now has 6Gb RAM (a recent upgrade as things were getting crowded at times and RAM is cheap), shared between the host and a couple of permanent VMs (currently 32-bit) and occasionally some small other VMs (also 32-bit).

Though to be honest, if you are installing Linux as your base OS, there is no reason to hold yourself at 32-bit - I am only still there due to the history of the rig.

Re:I've tried it

[geekboy642]

You're not running legacy code written by a contracting team who disbanded 4 years ago. You're not running the latest crufty hack on a codebase that was originally ported from CP/M by a group of five coders with 2 GEDs among them ll. You're not running billing software written in COBOL on a custom interpreter that was first launched on an 80286 with 512KB of RAM and a greenscreen. XP Mode is not for you. XP Mode is for companies that can't afford to rewrite all of their business software in six months. Rest assured, (almost) all of your games and shareware crap will still work natively in windows 7.