Half of Windows 7 Machines Running 64-Bit Version

nk497 writes "Microsoft has said that nearly half of machines running Windows 7 are using the 64-bit version, up from just 11% of PCs running Vista. The 32-bit version is limited to 4GB RAM, while the 64-bit version allows 192GB, as well as added security and virtualization capabilities. While Microsoft is pushing 64-bit as a way to gain performance in the OS, it earlier this year advised users to install the 32-bit version of Office 2010, 'because currently many common add-ins for Office will not function in the 64-bit edition.'"

Comments

64-bits

[Anonymous Coward]

Twice as fast! Thus my Second First Post since I went all 64-bits!!

Statistics, statistics

[mstefan]

The reality is though that 10% of Windows systems are 64-bit (there's actually still more systems running Vista than Windows 7 out there, although the gap is shrinking). The vast majority of Windows desktops are still running the 32-bit version of Windows XP, and that's not going to change until businesses decide they have a compelling reason to upgrade.

Re:Statistics, statistics

[Monkeedude1212]

The vast majority of Windows desktops are still running the 32-bit version of Windows XP, and that's not going to change until businesses decide they have a compelling reason to upgrade.

And my guess is that'll happen when they stop supporting XP P3 - which if my memory serves correct is 2014? Can someone back me up on that?

64 bit isn't too far off. As a developer you'd be better off getting a copy soon and work on merging your projects over to work on 64 bit now, rather than wait for crunch time.

Re:Statistics, statistics

[Z00L00K]

And I'm running XP-64, so if you love XP and want 64 the possibility exists.

Re:Statistics, statistics

[Anonymous Coward]

64 bit isn't too far off. As a developer you'd be better off getting a copy soon and work on merging your projects over to work on 64 bit now, rather than wait for crunch time.

You never needed a 64-bit system to compile for the architecture. Anyone who waited this long to port is stunningly ignorant and about 5 years behind the curve.

BTW, one thing not reflected in pure breakdown of user distribution, is that those "50%" of 64-bit users are the FAR more important ones when it comes to marketing. Users stuck behind the technology train are simply not worth catering to, generally.

Re:Statistics, statistics

[alta]

And considering so many companies are moving towards web based interfaces for their internal applications, this is going to take a long time. Sure, MS can stop making IE for XP, but get SP3 on the machines and it's pretty damn secure. Add to that Chrome or FF to run the business applications and you have a machine that's going to last for many years to come. Want to go faster? Get faster/more servers! XP can essentially become a dumb terminal as for as those enterprises are concerned. I think that about the only thing they could do is to make new versions of office not run on XP. That'll make a few companies switch because they can't be without outlook+exchange/word/excel.

Re:Statistics, statistics

[Zixaphir]

Unless, of course, they are the majority and it doesn't look like that is going to change anytime soon. I mean, you're talking about being behind the curve, but I also bet you're running an x86 derivative. A 64-bit instruction set doesn't fix the gaping problems of x86. Yet, x86 remains the defacto standard.

Re:Statistics, statistics

[MobileTatsu-NJG]

The vast majority of Windows desktops are still running the 32-bit version of Windows XP, and that's not going to change until businesses decide they have a compelling reason to upgrade.

Ah, you mean in a year or two when the machine is upgraded.

Re:Statistics, statistics

[drsmithy]

64 bit isn't too far off. As a developer you'd be better off getting a copy soon and work on merging your projects over to work on 64 bit now, rather than wait for crunch time.

Pro-active developers ? You've got to be kidding. It took the "annoyance" of Vista's UAC before developers finally started changing their Windows applications not to needlessly require admin privileges. They're not going to be implementing 64-bit support one second before "crunch time" arrived.

Re:Statistics, statistics

[VGPowerlord]

A 64-bit instruction set doesn't fix the gaping problems of x86.

...but it does fix some of them, such as increasing the number of registers and support for SSE2 (I think) extensions on all x86-64 chips.

Re:Statistics, statistics

[commodore64_love]

I thought most people say XP-64 is crap?

I just bought a Windows 7 PC used for $150. I never even thought about the RAM limit. It's just a 32-bit CPU so that means I can never go larger than 3 gigabyte. :-( On the other hand maybe I'll never need to. My current XP-PC is still on just half-a-gig and works fine. (shrug)

Re:Statistics, statistics

[Hadlock]

What's the driver situation like on XP-64 these days? Do you get security updates at about the same time as your 32 bit brethren? I remember in ~2004 when my buddy got XP he had to downgrade to 32 bit due to driver avalibility issues but I'm sure a lot has changed in six years.
 
I recently upgraded from XP-32 to Win7-64 and was amazed that everything I own had recent and fully functional 64 bit drivers... except my netgear brand wifi card, which is 64 bit supported, unless you have more than 3.5GB of ram. The workaround of which involves unsigned Spanish language drivers...

Re:Statistics, statistics

[ultranova]

64 bit isn't too far off. As a developer you'd be better off getting a copy soon and work on merging your projects over to work on 64 bit now, rather than wait for crunch time.

Every 32-bit program I've tried to run on 64-bit Win7 have worked perfectly. Some have even benefited, since a 32-bit program in 64-bit Windows can use the entire 4GB virtual memory space for itself, assuming that the correct executive headers are set. And of course having more physical memory makes multitasking easier and allows for more disk cache.

Re:Statistics, statistics

[ultranova]

A 64-bit instruction set doesn't fix the gaping problems of x86.

Which are?...

Re:Statistics, statistics

[adonoman]

For the most part, there's very little reason to port apps over to 64-bit. As the OS level, it's great, allows more memory, and allows you to run 64-bit apps. But at the app level, unless your app can make use of the expanded address space, it's not really that useful to port. The few apps that can really take advantage of it were ported long ago.

Re:Statistics, statistics

[Anonymous Coward]

Or they will implement something after the axe falls, but alpha or at best beta quality. When users call in griping how their product doesn't work, the company will say it was MS who did this. Vista got a lot of flack from lazy development houses because they would not bother making their stuff UAC compatible, or even writing solid drivers for Vista's driver model, blaming any crashes and blue screens on MS.

It is funny how on every other platform but Windows, should a major shift happen, devs gripe, but they deal with it. On Windows, just getting companies to separate user/superuser code causes a major trainwreck because the software companies (or the offshore code sweatshops) are too lazy to deal with it.

Re:Statistics, statistics

[mstefan]

Businesses don't tend to have that short of an upgrade cycle when it comes to operating systems. They typically prefer to stay on the trailing-edge of technology as long as possible -- "if ain't broke, don't fix it" is the mantra of most IT departments, particularly in larger companies. If you look at a lot of the "droneware" business desktops out there today, they're sold with 2-4GB RAM and downgrade rights to XP 32-bit. So while the system may be "sold" with a Win 7 64-bit license, that doesn't mean it's ultimately how it's being used.

Re:Statistics, statistics

[Anonymous Coward]

Unless, of course, they are the majority and it doesn't look like that is going to change anytime soon.

'Majority' is almost without meaning. A 64-bit (OS) user is vastly more likely to buy your product than a 32-bit user. He has dispensable funds. Many 32-bit users (often stuck on 5-10 y/o hardware) do not. Furthermore, if you are (still!) writing desktop applications today, you are likely targetting specific hardware capability, which means recent HW is even more relevant.

This doesn't usually apply if you write software for businesses, of course. IE6 is important to corporate aiming web developers for largely the same reason.

Re:Statistics, statistics

[Cornelius the Great]

Running XP-64 here for work (not by choice- some legacy support needed for projects in Visual Studio 2003, which doesn't work on Vista/7). I've been using it for 3 years.

The desktop/taskbar will occasionally freeze for no reason. That's my only major annoyance.

The driver situation is better than it used to be. Nvidia's drivers have come a long way- up until last year, I couldn't enable antialiasing while multiple monitors were in use (now I can). Coworkers using SLI still have issues, but then again, even 32-bit drivers aren't perfect at SLI. I no longer get blue screen crashes when hot-swapping USB-drives.

Still doesn't hold a candle to Win7-64 though- it seems more stable than 32-bit XP, at least on the 2 systems that have been upgraded from XP to 7 at my home.

Re:Statistics, statistics

[mstefan]

As of right now, yes, Windows XP SP3 reaches end-of-life in 2014. Of course, that doesn't mean that companies will automagically stop using it on that date, but I'm sure companies are using that as part of their migration plans.

If you think about it, it took about a decade to complete the transition from 16-bit to 32-bit, if you consider the release of Windows NT in 1993 as the starting point for the platform. Windows 95 is when it started becoming mainstream, and for a long time after that you continued to have businesses use 16-bit legacy applications. With the release of Windows 7, I think we're at the "Windows 95" point of that transition curve for 64-bit.

One other thing to keep in mind is that Y2K provided a reason for businesses to upgrade their hardware and software at once; so you saw a large shift in the late 90s as they prepared for the transition (or the apocalypse, depending on who you listened to). That was a significant driver towards phasing out their legacy 16-bit applications; there really is no corresponding justification for businesses to dump working 32-bit desktop applications in favor of 64-bit. For the most part, Suzy the Secretary and Bob the Shipping Clerk aren't going to get their work done any faster or better on a 64-bit system. And when there's no concrete, tangible benefit to their bottom line, companies aren't very interested in making technology investments "just because they can" (particularly in this economy).

Re:Statistics, statistics

[h4rr4r]

Nope, We are at another upgrade cycle here and the desktop folks have already bought machines that will come with Win7 x64 and will be promptly imaged to 32bit XP. Our customers apps require this. Business apps are the last to upgrade, so do not expect those to move until at least 2012.

Re:Statistics, statistics

[RobertM1968]

I thought most people say XP-64 is crap?

I just bought a Windows 7 PC used for $150. I never even thought about the RAM limit. It's just a 32-bit CPU so that means I can never go larger than 3 gigabyte. :-( On the other hand maybe I'll never need to. My current XP-PC is still on just half-a-gig and works fine. (shrug)

I didnt realize that anyone made 32bit CPUs anymore... what machine is it?

Re:Statistics, statistics

[TheRedShirt]

The vast majority of Windows desktops are still running the 32-bit version of Windows XP, and that's not going to change until businesses decide they have a compelling reason to upgrade.

And my guess is that'll happen when they stop supporting XP P3 - which if my memory serves correct is 2014? Can someone back me up on that?

64 bit isn't too far off. As a developer you'd be better off getting a copy soon and work on merging your projects over to work on 64 bit now, rather than wait for crunch time.

Yup, April 8th, 2014.

I think that in many cases though, that the "compelling reason" may be sooner for some due to issues with their own system. I have a secondary machine that I "inherited" that was poorly maintained. I attempted to install SP3 a while back and it failed for some reason I can't recall why. I can't be bothered with reinstalling because it is only a secondary system and it isn't worth the effort.

However, I think that the state of this machine may be indicative of many computers currently in use by "John Q. Everyday Citizen." I may be over generalizing with that, but it seems apparent to me that the well informed and able users are in the minority. I have the feeling that the end of SP2 support tomorrow and not the 2014 end for SP3 will be the large scale death knell for XP.

I think that by and large, the majority of those users (ie those less able) still using XP will feel compelled to upgrade. Probably by the end of the year, I'd wager.

Re:Statistics, statistics

[Sycraft-fu]

It truly amazes me how lazy developers are when it comes to supporting new things. They whine and bitch and drag their feet and blame MS, rather than just admitting they have to learn something new and doing it.

For example back in the Windows 2000 days I had a professional sound card. It had Windows NT drivers, but had some limits since NT wasn't good at sound. 98 was the preferred OS. Well when 2000 came out, the refused to release 2000 drivers. They claimed that the kernel mixer introduces 30ms of delay that you couldn't get around and that was unacceptable for pro work, etc, etc. Just use the NT drivers even though that caused some problems.

I (and probably many others) found the MSDN page on kernel streaming and sent it to them, showing them they were full of shit. Finally, many months later, they released a WDM driver... That supported only 2 of the interfaces 10 inputs and outputs. They claimed that WDM could only support one stereo set, that's it. A built in limitation by MS, nothing you can do about it. In frustration I e-mailed MS and I think they were sufficiently surprised by the stupidity of the statement that a developer actually responded and showed me where to find the docs. Turns out that WDM support lots of audio devices, either enumerated as multiple stereo pairs (as old style drivers did) or as a single multi-channel output. in fact you could do both at once.

So that went off to them and they ignored it for a long time and finally got out a real, full, WDM driver that was buggy as shit. The proceeded to work on the bugs and eventually had a nice driver. They decided they really liked WDM at that point, and quickly stopped supporting the older formats. It went from "We can't do it," to "This is the only way to do it.

But it took like a year and a half.

Many, perhaps most, developers are extremely, EXTREMELY lazy at updating to new technologies and fixing up their code. They want to keep doing shit the same way they always have, no matter how outdated that is.

Re:Statistics, statistics

[Bruiser80]

It is crap because it was weird and drivers were hard to come by. It is based on Windows Server, if I remember correctly...

Re:Statistics, statistics

[commodore64_love]

I feel like I just took a step back in time to the Nintendo 64 days. "There's very little reason to write games in 64 bit, so virtually all of them use the CPU's 32 bit mode."

Ya know, I just realized how long I've been using 32 bit applications. First I had the 8 bit Commodore 64/128 machines, and then I jumped to 32 bit with the Commodore Amiga in 1987. I've never used anything but 32 bit applications since then. That's a loooong time.

Re:Statistics, statistics

[Firethorn]

Ah, you mean in a year or two when the machine is upgraded.

Around 6 years ago we went from a 4 year replacement policy to a 6 year one. Some of our machines are even older - we wait until somebody complains before replacing, it's just that 'over 4/6 years' was/is a valid excuse for replacement. If they don't complain, they don't get a new computer.

Fact of the matter is, except for bloat modern machines are typically vastly overpowered for what we use them for. Heck, even the bloat hasn't kept up all that well.

The new machines? I think the latest lot cost $300 each.

Re:Statistics, statistics

[TheLink]

I've had better luck getting Outlook for web to work in some scenarios. The sysadmin or someone at work wanted Outlook to use NTLM auth. That doesn't seem to interact well with Microsoft Communicator on my Win 7 64 bit machine.

So people might stick to XP + Web Apps. FWIW, I've tried Kingsoft Office before and if they come out with an Outlook replacement, I wouldn't be surprised if Microsoft sues them just to slow them down ;).

Re:Statistics, statistics

[TheLink]

Too bad food prices don't go down like that ;).

Re:Statistics, statistics

[mcgrew]

there's actually still more systems running Vista than Windows 7 out there

That's incredibly surprising; I thought Vista was a disastrous flop while 7 was a big success. I don't doubt there are a lot more machines running XP than 7, since its shipped on almost every new PC for almost ten years.

Re:Statistics, statistics

[mstefan]

The latest stats that I read (which was for June) was that in terms of marketshare, Vista is at around 14.5% (down from almost 20% when Win7 was released) and right now Win7 is at around 13.5%. So if you factor in the percentage of 64-bit systems all combined, it puts 64-bit Windows platforms at roughly 8.25% give or take. These are all rough numbers, and there's varying estimates out there (some higher, some lower in terms of overall marketshare) but I think it's safe to say that 64-bit Windows systems are around 10% overall.

Windows XP usage has dropped as well since Win7 was released (by about 7%, it's now at about 62% of systems overall). So for the most part, Win7 has cannibalized both Vista and XP installations fairly equally, but the effect has been much more profound for Vista because there were far fewer systems out there running it. I would exepct that by the end of the year, Win7 will clearly surpass Vista. Long term, I'm looking for when XP drops below 50% marketshare; when that happens, I think that'll be the indicator that the transition is accelerating.

Re:Statistics, statistics

[Lorkki]

It truly amazes me how lazy developers are when it comes to supporting new things. They whine and bitch and drag their feet and blame MS, rather than just admitting they have to learn something new and doing it.

I'd expect that a lot of developers just don't see the significance, as long as they can get a product out that works in a reasonable number of real cases. That's where their bread comes from, after all.

Re:Statistics, statistics

[aztracker1]

Most applications simply don't need the extra memory... 64-bit versions of a given app will usually use more memory as-is... With office, aside for very few users, a 64-bit version isn't really necessary. I think mac got 64-bit about as right as anyone... I do wish running 32-bit and 64-bit in linux was a bit less painful.

Re:Statistics, statistics

[Jeremy Erwin]

netgear brand wifi card, which is 64 bit supported, unless you have more than 3.5GB of ram

Then what's the point?

Re:Statistics, statistics

[Anonymous Coward]

It isn't crap. I used XP x64 for years before and had no major problems. My biggest issue was I couldn't get AT&T Natural Voices working on it, but that applies to all 64-bit versions of Windows.

The only people who say it's crap, like the only people who say that post-RTM Vista is crap, are the people who haven't actually used it and are simply regurgitating something that they read someone else say or are basing their opinion on their own ignorance from ten minutes worth of use.

Re:Statistics, statistics

[pnutjam]

As a Sysadmin type, I am not surprised by lazy developers.

Re:Statistics, statistics

[LinuxIsGarbage]

BTW, one thing not reflected in pure breakdown of user distribution, is that those "50%" of 64-bit users are the FAR more important ones when it comes to marketing. Users stuck behind the technology train are simply not worth catering to, generally.

Interesting point. Microsoft is cutting XP off of "Windows Live" and IE9, however they released Office 2010 for XP as well. Presumably because of the glacial uptake of Vista or newer in the business world, but generally good uptake of Office 2007.

Re:Statistics, statistics

[LinuxIsGarbage]

And of course this doesn't necessarily mean compiling 64 bit versions, but at least making sure your 32 bit version runs happily on 64 bit.

Re:Statistics, statistics

[ShadowRangerRIT]

Intel briefly revived 32-bit only processors with the Core (Solo and Duo), before releasing the Core 2 (Solo, Duo, Quad, etc.) line that used a similar architecture but supported x86_64. The 32 bit Core was first released in January 2006, and for about half a year it was the only form Core-series processors were available in. Starting in July 2006, they started selling Core 2 processors, but Core laptops continued to be sold for another year or so, though Core 2's displaced them quickly.

Re:Statistics, statistics

[LinuxIsGarbage]

Or they will implement something after the axe falls, but alpha or at best beta quality. When users call in griping how their product doesn't work, the company will say it was MS who did this. Vista got a lot of flack from lazy development houses because they would not bother making their stuff UAC compatible, or even writing solid drivers for Vista's driver model, blaming any crashes and blue screens on MS.

It is funny how on every other platform but Windows, should a major shift happen, devs gripe, but they deal with it. On Windows, just getting companies to separate user/superuser code causes a major trainwreck because the software companies (or the offshore code sweatshops) are too lazy to deal with it.

I agree that's a reason Vista got a lot of flak. Another reason is OEM preloads. The performance hogging junk the likes of HP preinstall is amazing. NIS 2007/2008 were probably the worst, and right during Vista's prime. An unactivated NIS trial with no tray icons or indication it was running doubled the boot time. I also saw a recent HP i7 laptop with Windows 7 preloaded and it took 5 minutes to boot. Ridiculous. Of course when these computers are downgraded to a fresh blank version of XP, of course it will run better.

Re:Statistics, statistics

[Man On Pink Corner]

Vista got a lot of flack from lazy development houses because they would not bother making their stuff UAC compatible

UAC wasn't the problem. Goofy, spurious changes with no security-related justifications, like breaking .INI files shared between different .exes in an application, were bigger compatibility concerns.

Re:Statistics, statistics

[LinuxIsGarbage]

Atom N270 and N280 are 32 bit.

Re:Statistics, statistics

[LinuxIsGarbage]

The vast majority of Windows desktops are still running the 32-bit version of Windows XP, and that's not going to change until businesses decide they have a compelling reason to upgrade.

Ah, you mean in a year or two when the machine is upgraded.

My brand new work PC from December with 4GB of RAM has a Windows 7 sticker and CoA key. Around me are coworkers with slightly older models and Vista CoA keys. Yet, the company still runs XP, and I haven't heard of any migration plans.

Re:Statistics, statistics

[Kjella]

It truly amazes me how lazy developers are when it comes to supporting new things. They whine and bitch and drag their feet and blame MS, rather than just admitting they have to learn something new and doing it.

Probably because of the large number of developers that throw new interfaces out there all the time as a work in progress, where the interfaces themselves are buggy, incomplete and constantly changing. Sure it goes both ways, but I'm sure some are very conservative with good reason based on past experience.

Re:Statistics, statistics

[ShadowRangerRIT]

There are some other minor benefits that come with 64 bit, but very few of them are marketable. For example, at least on x86_64 with Windows (and I believe all current OSes), the address space is still limited substantially (to 48 bits or less). One of the advantages to this is that every memory address has at least two bytes of zeroes in it. Most buffer overruns require at least one jump instruction to work, but with two nulls in every memory address, typical buffer overruns (triggered by an unbounded strcpy) don't work, because any address you use will terminate the string prematurely. Granted, not using strcpy would provide the same benefits, but you can limit your attack surface simply by recompiling for 64 bit, no need to scan every line of code for strcpy or other code with the same behavior that isn't so easy to find with grep.

Other benefits include taking advantage of memory mapping for performance and simplified coding. Memory mapping is much easier to work with, but on 32 bit architectures you have to be much more careful with it, since each mapping eats a contiguous chunk of your virtual memory space. A single 2.001 GB file, or a couple large files mapped to non-optimal locations can cause you to run out of your process's virtual address space and crash. To handle that case, you'd have to write the non-mmapped version of your I/O anyway, which means you aren't saving any complexity, even if you get the small performance boost. But on a 64 bit system, you can mmap files willy-nilly, no need to worry about running out of address space.

Re:Statistics, statistics

[Anonymous Coward]

...like breaking .INI files shared between different .exes in an application, were bigger compatibility concerns.

Would you elaborate or maybe provide a KB article as a reference? I've done what it seems like you're saying is broken on both Vista and 7 without problems. Of course, the INI files are under either ProgramData or Users where they belong - not under Program Files.

- T

Re:Statistics, statistics

[mdielmann]

First, the largest pool of developers who do stuff on their own time, Linux OS developers, doubtless had a 64-bit version of their software out before Win7 (or Vista) was even thought of. Feel free to look up the kick-off date for Vista (was it called Longhorn back then?) and when Linux support for the DEC Alpha was around.

Second, you want to know where the problem starts, look at management. How do you convince your boss to look at this radical new 64-bit environment when currently only 10% of the world would even notice the difference (and most of those will work fine with 32-bit)?

This reminds me of the old "Silly programmers, using 2-digit years" story. Sure, the fact that years were stored as characters was a problem, and got taken care of in time, but those extra two bytes of data cost a substantial amount of money (especially when multiplied by over 100k records), and there are many accounts of software purchasers/application managers being told this problem would raise it's head in some 40 years. But if you don't decide to hire programmers to fix it...

Re:Statistics, statistics

[Bruiser80]

I should correct myself. The OS itself wasn't crap. The fact that NOBODY developed for it made the experience with the OS crap.

Before Vista came out, we had one machine in our office running XP64. We use primarily Solidworks and they came out with a 64-bit version of their software. Completely buggy and some files weren't compatible with the 32-bit versions of them. Complete fiasco when you're staring down a deadline.
Eventually, they fixed it and now things are seamless. I have more machines running 64bit, but am limited because our stupid VPN can't handle 64 yet...

Re:Statistics, statistics

[Man On Pink Corner]

Of course, the INI files are under either ProgramData or Users where they belong - not under Program Files.

You can say they "belong" wherever, and maybe they do, but if an application that was built in the Win98 era that would otherwise have run perfectly on Vista/Win7 fails because of a misguided attempt to turn Windows into a strict multiuser OS, that's more the OS vendor's fault than the app developer's.

The problem happens when one executable, like a hardware configuration program, writes to an .INI file that other apps in the same directory are expected to be able to read. A virtualized copy of the .INI file is written under \Users\username\AppData\Local\VirtualStore\Program Files\[application name]\whatever.ini. The mistake Microsoft made was including [application name] in that path, because it means that an .INI file written by one .exe won't be accessible to another .exe in the same software package. This was an extremely common practice at one time, especially among developers (such as myself) who never drunk the Registry kool-aid.

Since .INI files are not executable and have thus been responsible for precisely zero OS exploits, it can be concluded that this was just a case of Microsoft fixing something that wasn't broken. This particular change sounds obscure and trivial, but it broke a large number of legacy apps for no good reason.

Re:Statistics, statistics

[Bigjeff5]

It truly amazes me how little the average person understands about economics.

A business exists to make money. A company doesn't sell a sound card out the kindness of their heart, and they don't write drivers for said sound card out of the kindness of their heart either. They can't survive if they do. If being pro-active does not generate more income than it costs, then it is economic suicide to be pro-active.

It's called return on investment, and any halfway decent developer is going to do at least a rudimentary ROI analysis before beginning even a small upgrade, let alone a complete overhaul of their code.

Basically what you are asking developers to do is exactly the same as your boss telling you you need to work evenings and weekends for no pay. You'll do it if it means losing your job, but otherwise you'll tell your boss to either pay you for it or fuck off. Getting an updated driver is no different. If it's free, expect it to take a while - they'll have to spread the cost over a long period of time to make the investment feasible. If it means losing customers because they didn't write the driver, well then you'll get it pretty quickly.

Re:Statistics, statistics

[leptons]

In my experience, it isn't the developers who are lazy, it is management who won't allocate resources.

Re:Statistics, statistics

[quanticle]

Windows XP 64-bit earned its bad reputation because of the lack of stable 64-bit drivers. Hardware manufacturers (correctly) assumed that very few Windows XP users would go to 64-bit, and reserved their 64-bit driver efforts for Vista and 7. As such, the 64-bit XP drivers were half-assed ports of their 32-bit cousins, and often never worked properly.

To be honest, I think the first 64-bit iteration of any OS (be it Windows, Mac, or Linux) is bound to have problems as software and hardware manufacturers make the transition.

Re:Statistics, statistics

[Anonymous Coward]

Many, perhaps most, developers are extremely, EXTREMELY lazy at updating to new technologies and fixing up their code. They want to keep doing shit the same way they always have, no matter how outdated that is.

Then there's the cases of needing to update undocumented and uncommented code by some unexperienced first-year student kind-of horror scenarios..

Usually though, unless there's a need to support new tech it can save a big penny.

Re:Statistics, statistics

[quanticle]

True that. I just got a new desktop at work. Its got a sticker on it with the Windows 7 logo. What OS is it running? Windows XP SP 3.

Re:Statistics, statistics

[Anonymous Coward]

And my guess is that'll happen when they stop supporting XP P3 - which if my memory serves correct is 2014? Can someone back me up on that?

I back you up on that, and although I'm not behind times, it doesn't mean I have to be stupid. I have vast amounts of hardware which run the latest and greatest flavors of software on the good old trusty XP SP3.

By 2014 we'll have Windows 8 64bit and some service packs out, so I can simply change OS the hardware is phased out. I applaud Microsoft for supporting XP for full 14 years (even if you have to have the latest XP SP) after release.

You can imagine that they could get some quick bucks by means of force-upgrading their users when they cut support early. But they didn't do this.

Re:Statistics, statistics

[Kalriath]

Your entire post is unadulterated wrong.

The Virtual Store is an exact replica of the folder structure of Program Files. The only reason "Application Name" would be in the path, is if the application put it there. Windows didn't.

Second, whether it's an INI file is beyond irrelevant. Windows redirects all writes to the Virtual Store, because no user-mode application should be writing to program files except installers.

Re:Statistics, statistics

[Kalriath]

If you're going to compare 64-bit (any version) Linux, you should really compare to 64-bit (any version) Windows. I think there was a DEC Alpha version of NT 3.51?

Re:Statistics, statistics

[boxwood]

I'd say you're the one who doesn't understand economics.

You seem to think its only businesses that do ROI calculations. Consumers do ROI calculations too, you know.

If I bought a product from a company that doesn't give me support, I'm going to consider future purchases from that company to be a bad ROI. In fact I may even come to the conclusion that pretty much all of the corporations won't support their products. So I'll just assume that all products have no warranty and buy the cheapest stuff I can. Which basically means stuff manufactured in china. If it breaks or doesn't work with newer tech, I'll just buy another one. And then happily ignore all the big corps whining about how nobody is buying their stuff anymore.

If supporting your customers doesn't have a high enough ROI to be worthwhile, don't be surprised if your customers think your products don't have a high enough ROI to be worthwhile either.

Re:Statistics, statistics

[Anonymous Coward]

You can say it's all about economics, but the example above makes it look more like basic incompetence and general asshattery. It's one thing to say "we are not supporting it because it's not economically viable" and another to say "nope, can't do it, not possible".

Re:Statistics, statistics

[RocketRabbit]

For a lot of businesses, whether or not Microsoft officially supports XP is irrelevant. They have their IT infrastructure, it works, and that's that. Many companies furnished themselves with Windows XP based desktops from commodity providers like Dell and HP and did not partake in any corporate or volume licensing schemes at all. I know of many companies that simply buy new computers when old ones break, and in fact a large mass of "IT" workers are in fact secretarial, and have little desire to learn or run a new system at work. They just want to punch their data into the computer and go home at night.

Big companies are a different story, but small companies grow to become big companies and often retain the ad-hoc purchasing and upgrade strategy for the PCs that are deployed on desktops. The practical upshot of all this is that you have XP machines that will stay in deployment until they break. I am finding companies often prefer to take all their XP boxes, make VMs of them, and then work off dumb terminals. Ironically the performance is usually better, and these machines then last literally forever.

No doubt, a hundred years from now Windows XP will still be in use somewhere. Just as DOS, Windows 3, NT4, Windows 98, ME, and 2000 are all still firmly deployed even if somewhat rare on the OS surveys that go by user agent strings.

Re:Statistics, statistics

[Anonymous Coward]

I do wish running 32-bit and 64-bit in linux was a bit less painful.

Painful how? I've been running 64-bit Linux for about 5 years.

I've encountered a few apps that weren't updated by my distro fast enough to suit me and I couldn't be bothered to compile myself (e.g. Firefox) for which only 32-bit binaries have been available from the vendor. In those cases I've had to install 32-bit versions of a few libraries (using my distro's package manager) to get the apps to run. No big deal.

Once or twice I've had to compile 64-bit versions of drivers. Certainly less painful than discovering that some vendor can't be bothered to provide 64-bit versions of their proprietary Windows drivers and not being able to use the hardware in 64-bit WIndows at all.

Re:Statistics, statistics

[icebraining]

The only people who say it's crap, like the only people who say that post-RTM Vista is crap,

I have Vista with all the latest updates installed. It's still crap.

Re:Statistics, statistics

[RocketRabbit]

The mere fact that the "downgrade" option exists is mere lip service to Microsoft. Most users would consider XP as an upgrade to Vista, and many I know would consider it to be an upgrade to 7 as well.

A bone stock Windows XP machine with 3.whatever gigs of ram is cheaper than a good toaster oven these days, and is sufficient for the uses of 99% or more of the folks out there today.

Re:Statistics, statistics

[dbIII]

64 bit isn't too far off

That's true, November 1998 will be here so soon that it's as if it's already happened!

Re:Statistics, statistics

[Man On Pink Corner]

(Shrug) Take it up with National Instruments, and every other vendor who's had to document this nutty scheme.

The way you describe it may be how it works now, post-Vista, but the fact is, different .exes could not share writable .ini files when Vista came out. Those applications simply would not work unless the user knew to mark all of the .exes to run with admin privileges. The virtualization scheme basically did not work as advertised, and it was a big part of why people hated Vista.

Re:Statistics, statistics

[ruiner13]

Hi, random developer here. It isn't always us. Shifting ANYTHING like that also takes quality assurance's blessing. Also, customer support always likes to be in the loop about subtle changes like that which can for no reason turn a quiet day into living hell (then feeds its way back to us...). Trust me, most of us would prefer to get rid of that crap if we had the opportunity and sign off. Debugging crusty old VB6 COM libraries isn't our idea of a party.

Re:Statistics, statistics

[Radtoo]

Wrong, I think it is a business model in the world of Windows. Wait until MS finally forces 64 bit onto everyone and everyone is going to have enough RAM to require an update and then sell a new edition of your software product line simply with this update. Much better than to compete with yourself by entering the 64 bit market when there still is low demand.

The more popular open source OS' on the other hand actually have long needed to run on very high end servers and hence are already running 64 bit in terms of OS and server software. And the remaining applications on these OS are simply works of love and got the update because it was simple to do (it also is so on windows, but there's more profit in waiting for Microsoft to force user's hands).

Re:Statistics, statistics

[LesFerg]

You seem to lack an awareness of the real world. Or do you work in a company where a bunch of "developers" sit around doing whatever they want while the company director hides in his office wondering what the company produces?

"Developers" work on projects which have been analysed, costed, approved and scheduled by company managers and directors.

A company which set out to manufacture a certain range of hardware will have developed drivers to run that hardware on the current most used platforms. What interest do they have in supporting a bunch of hardware that was marketed for last years target platform, when the current market goal is to produce and sell hardware which supports this years target platform? Oh wait, the board of directors will go ask the developers what they feel like doing, won't they.

Re:Statistics, statistics

[Civil_Disobedient]

I thought most people say XP-64 is crap?

Most of the people who say that aren't running it and have never actually run it, but could swear they heard it was crap from someone that swears they uninstalled it, like, after 5 minutes because it was crap.

I've been running XP 64-bit for a couple of years now. IIRC it was originally a free D/L from MS, now it's really only available through their MSDN (for free if someone's paying for your MSDN subscription, like work). It's awesome. Easily, without a shadow of a doubt the best OS to come from Redmond. And I used to be a die-hard Windows 2000 XP Pro lover. Still have much love for the 2k, but damn, XP64 is just solid. No locks, no hardware issues.

Now, that said, I have all the right XP64 drivers, I didn't have to hack any 2003 drivers (as XP64 is technically the 2003 codebase). Everything just works(TM) and YMMV. That said, I have had NO problems finding drivers. The big, huge, SCARY thing everyone was warning me about, how I wouldn't have sound (wrong) or video (wrong) or any of my peripherals (wacom cintiq works fine, external audio works fine, all USB stuff work fine as they should, my 3Ware 16-drive RAID card? WORKS FINE.) The only piece of hardware I have had any problem with is my Nikon CoolScan 5K, but that's not just with XP-64, that's any 64-bit operating system, including 2003, Vista, and (yep) Windows 7.

The nice thing is I get all the 64-bit benefits (12 gigs of RAM makes Photoshop my bitch) all the XP benefits like DirectX 10 and updated service packs (it's recent enough that MS is still supporting it... for the time-being, anyway), wireless networking Just Works(TM), plus no Widget-interface fisher-price crap, no always-running, always-scanning crap, no treat-you-like-an-intruder-for-your-own-good security crap... everything runs on it, and it's fast (faster than Win7, anyway).

Re:Statistics, statistics

[X0563511]

Just make sure your users have write permissions to the directory (or just the file) and the problem goes away.

Re:Statistics, statistics

[X0563511]

The most important part of that....

STOP USING POINTER ARITHMETIC.

Seriously. Not as important, but good to follow: stop assuming datatype sizes! If your code is sensetive to that, run a check at the very start of your program and use the value determined there. For example, declare an unsigned of each datatype you intend to care about the bounds on... set it to zero, and subtract one. Record that number in a large enough integer (Hell, use the one you just used!) and not use the bound as a magic number...

Re:Statistics, statistics

[fluffy99]

It is essentially Server 2003. Driver support for it royally sucks, particularly for printers and video cards. It also has some odd issues if you're trying to run it on something with lots of CPU cores. I found remote connections to it start hanging as it has a stupidly low limit on worker processes.

If you want a stable and solid 64-bit OS, go with Windows7 64-bit. It might have a bigger footprint, but its better supported and faster the XP64-bit (based on my experience with CAD software).

Vista was and still is crap. My evidence? 10% of the computers I support have Vista, yet they account for %50 of the strange problems. Crap such as refusing to install patches from WSUS, spontaneously requiring to be activated via the phone, the DVD drive that suddenly refused to see the cdfs filesystem, UAC breaking GPOs and logon scripts. The XP boxes are all pretty solid and I've yet to have any real problems with the Win7 boxes.

Re:Statistics, statistics

[SplashMyBandit]

As a Java developer this is no problem. Should be the same for the C# crowd. If you are using C or C++ then maybe its time to modernize rather than drag the fetid body of your application to 64-bit (Windows). Even 64-bit Windows platforms won't last forever. Unix has been 64-bit for a long time. It's only the Windows crew who are now starting to move.

Re:Statistics, statistics

[Anonymous Coward]

Then you've either messed up your installation or you haven't used it for more than ten minutes.

Re:Statistics, statistics

[RobertM1968]

Thanks guys!

Re:Statistics, statistics

[drsmithy]

You can say they "belong" wherever, and maybe they do, but if an application that was built in the Win98 era that would otherwise have run perfectly on Vista/Win7 fails because of a misguided attempt to turn Windows into a strict multiuser OS, that's more the OS vendor's fault than the app developer's.

No, it's the app vendor's fault for not writing their program properly in the first place.

This was an extremely common practice at one time, especially among developers (such as myself) who never drunk the Registry kool-aid.

That "one time" was in the early '90s, when Windows *3.1* was mainstream. By the time of Windows 98 you should have had fixed your software to do the right thing years earlier.

Since .INI files are not executable and have thus been responsible for precisely zero OS exploits, it can be concluded that this was just a case of Microsoft fixing something that wasn't broken. This particular change sounds obscure and trivial, but it broke a large number of legacy apps for no good reason.

The good reason was not having program directories writable by regular users. Would you store your .ini files in /usr/bin on a UNIX machine ?

Re:Statistics, statistics

[Anonymous Coward]

that's the myopic view.

what you're missing is that if you screw over the early adopters you're killing your market. i'm technically savvy, an early adopter, and people often ask me for advice when purchasing electronics. i'll specifically tell people to stay away from companies that produce bad drivers or service them poorly. for example, i'd stay away from Sony VAIO, they're very bad at releasing new drivers when a new version of windows comes out and they use a lot of proprietary hardware for which generic alternatives are not available. they make good headphones, though.

Re:Statistics, statistics

[Anonymous Coward]

Yeah, right. That's total BS.

There is no ROI calculation, there is no justification. You overestimate the competence of most business.

Re:Statistics, statistics

[Deathanatos]

That's not the mantra - the mantra is "If it isn't bleeding out of its eye sockets, dismembered, and is dragging itself about the cubicles moaning like the half dead, half undead zombie piece of software that it is, then don't fix it." It's like telling some poor little kid they're going to the dentist to have their teeth pulled, and watching them kick and moan and scream the whole way. I find myself wondering "just what does IT do?" since it seems to take a planetary alignment just to get something upgraded from version 1.0.0 to version 1.0.1 when version 9.4.7 has been out the door for a decade. As a developer, I could make your computer address you in a British accent, play Mozart to make your baby smarter, and put the toilet seat down after each use, except that those features are only available in Windows XSP2k 128-bit, which we haven't upgraded to yet.

Re:Statistics, statistics

[Mattsson]

In all honesty, how many programs are there that actually have any use of more than 32 bit memory space today?
Do you really need access to more than 2GB of RAM each in a media-player, image viewer, word processor, etc?
Yes, it is a bit annoying that 99% of all my software is 32 bit in my 64 bit system, but most of the time it's just irrelevant, and most of the few programs that actually has any use of more RAM is available in 64 bit versions...

Re:Statistics, statistics

[icebraining]

No errors or any other problems during installation. If it's messed up it's Vista's fault. And yes, I've used it for more than ten minutes. Although it isn't my main OS, there are some applications I need to run on Windows.

I also use Windows 7 for gaming in my desktop and I used Windows XP before, and both are much, much better than Vista.

Re:Statistics, statistics

[Firethorn]

Too bad car prices didn't go down. ;)

Realistically, it was my grandparent's generation that the food prices effectively went down.

I'd have said TVs, but they pretty much have - it's just that we traded up as prices went down.

What's especially grand is they're essentially premium 'business machines' - no fancy graphic adapters, but dual-monitor capable. Dual Cores right around the least $ per ghz*, 4 gigs RAM, etc... They're like you ordered 'the most computer for the buck'.

The $300 price would be more like $600 if you weren't buying them a couple thousand at a time.

By buying such 'good', if basic, machines, we're capable of getting six years out of them, which saves quite a bit of money in the long run - we might be able to get cheaper computers for like $200, but then they might only last 2-4 years, and there are costs associated with replacing a user's computer, such as moving profiles. We could spend twice as much, but that would be unlikely to get the computers to 8.

*Going by the way the bulk buys work, this is whatever processor offers the most performance per $ at the time.

Re:Statistics, statistics

[BuildMonkey]

I've been running XP-64 for more than a year now on a Dell Precision T3500. It's a beautiful thing: 6GB of RAM, (4) drive RAID-10 system, nVidia graphics, dual screen. Everything works: printers, scanners, mice, sound, video. Oh, not quite everything: the only thing I've found that won't work is iTunes.

Still, I'm giving serious consideration to migrating to Windows 7 64-bit. Once I try it on another machine, I'll use the new VMware 7 migration facility to wrap up my nicely customized box as a virtual machine, add another 6GB of RAM, and install Windows 7.

Any caveats on the VMware Workstation 7 migration tool?

Re:Statistics, statistics

[Anonymous Coward]

There are plenty of reasons for doing what appears to be procrastination. Rapid changes to support a new platform rarely equals money, but it often equals time and expense. If you wait until the first or second service pack release you will probably benefit from the stories of pain that others share and spend less time updating your application. The only good reason to adopt rapidly is if you will lose revenue by delaying.

Re:Statistics, statistics

[Anonymous Coward]

Debugging crusty old VB6 COM libraries isn't our idea of a party.

You shouldn't have wrote that SHIT in the first place. VB has and always will be a craptastic hack who's niche is to create even more craptastic hacks. This isn't hindsight talking either, it has been that way with every non-portable iteration of the product.

If you didn't know that VB is really only good for a quick hack that will likely break at random by the time the sixth iteration rolled around... there's something wrong with you. You should only use it when your highest expectation of your your application's life cycle to be very short. Time to kick the script kiddies out of your shop.

Re:Statistics, statistics

[Anonymous Coward]

The reality is though that 10% of Windows systems are 64-bit (there's actually still more systems running Vista than Windows 7 out there, although the gap is shrinking). The vast majority of Windows desktops are still running the 32-bit version of Windows XP, and that's not going to change until businesses decide they have a compelling reason to upgrade.

The reality is though that 10% of Windows systems are 64-bit (there's actually still more systems running Vista than Windows 7 out there, although the gap is shrinking). The vast majority of Windows desktops are still running the 32-bit version of Windows XP, and that's not going to change until businesses decide they have a compelling reason to upgrade.

Totally incorrect. There are more windows machines running XP and Win 7 than Vista, combined. Don't believe me? I'm going to say do your own homework on this one. Sales of Win 7 alone are now close to exceeding Vista, plus add into the XP users that haven't upgraded.

Re:Statistics, statistics

[NotBorg]

The trouble with your 32-bit programs is that they are typically built for generations old processors. A 64-bit address space s NOT the only advancement made in the CPU over the years. It's not uncommon to see applications built with a build process thats compatable with the original Pentium for example (sometimes even as far back as the 386). This ignores generations of processor features like SSE, NX-bit, and so on. By choosing a 64-bit build process you can safely enable support for all kinds of goodies just because all 64-bit processors come with them.

The performance difference varies from application to application, but it generally isn't any worse. Some applications make out rather well.

The biggest stag in porting applications is that the 64-bit APIs and libraries do not include a lot of the old cruft from the Win-32 world. This makes it much more than setting a compiler flag and rebuilding. At least that seems to be how it is on the Windows side.

In short, my motivations for wanting a 64-bit environment really don't have much to do with having a 64-bit address space.

Re:Statistics, statistics

[NotBorg]

AFAIK, it's often more than that. The 64-bit libraries don't include a lot of the old cruft from the Win-32 world. Even if you don't do pointer arithmetic, it's much more than a re-compile at that point since the API calls aren't there.

Also if you used 3rd party libraries that are no longer maintained... there is no hope for you unless you want to rewrite them yourself or figure out some thunking technique to use the 32-bit library in your 64-bit application. Either way it can be less than trivial.

Re:Statistics, statistics

[Man On Pink Corner]

No, it's the app vendor's fault for not writing their program properly in the first place.

It was fine 10 years ago when it was written.

That "one time" was in the early '90s, when Windows *3.1* was mainstream. By the time of Windows 98 you should have had fixed your software to do the right thing years earlier.

In this particular case the vendor was long since absorbed into a larger company, leaving the product unsupported. Again this wasn't a problem (for me, at least) until MS decided to give their B-team access to the Windows source tree.

The good reason was not having program directories writable by regular users. Would you store your .ini files in /usr/bin on a UNIX machine ?

What problem is solved by not having program directories writable by regular users? I keep asking that every time this subject comes up on Slashdot, and the answer always sounds a lot like dead silence. If I have a device connected to a serial port that is accessed directly by the program, there is absolutely no reason for the .ini file that specifies that COM port to live in a user-specific directory, or anywhere else besides the application's own install directory.

Windows is not Unix. Unix was designed from day one to be a multiuser operating system. Windows runs on personal computers, and 99.999% of use cases for Windows will always involve a single user.

Windows 95 was the most successful release in Microsoft's history because nothing was taken more seriously than the idea that existing applications, even 16-bit DOS apps from the Jurassic period, had to keep running. Vista failed (well, if you don't count encouraging millions of high-margin retail purchases of Windows 7 as 'failure') because the company abandoned that ideal at every level, from driver support to the filesystem.

Re:Statistics, statistics

[Kalriath]

You're still wrong. It's always worked as I described it.

Re:Statistics, statistics

[drsmithy]

It was fine 10 years ago when it was written.

If it was written only 10 years ago then it was wrong then, and had been wrong for 3-5 years already.

That it worked was just due to the typically ridiculous lengths Microsoft go to to ensure legacy support.

In this particular case the vendor was long since absorbed into a larger company, leaving the product unsupported. Again this wasn't a problem (for me, at least) until MS decided to give their B-team access to the Windows source tree.

Not fixing a particularly broken corner case, in the context of fixing a while bunch of other, vastly more important problems, is the hallmark of good engineering, not "the B team".

What problem is solved by not having program directories writable by regular users? I keep asking that every time this subject comes up on Slashdot, and the answer always sounds a lot like dead silence. If I have a device connected to a serial port that is accessed directly by the program, there is absolutely no reason for the .ini file that specifies that COM port to live in a user-specific directory, or anywhere else besides the application's own install directory.

They're probably flabbergasted by the naiveity of the question. Basic best practices (stretching back decades) is the first and foremost reason. Reducing the risk of trojans, virus infections and general data loss three more. Separation of binaries and configuration information a fourth.

Windows is not Unix. Unix was designed from day one to be a multiuser operating system. Windows runs on personal computers, and 99.999% of use cases for Windows will always involve a single user.

Windows NT is, and always has been, a multiuser OS (incidentally, UNIX's multiuser capabilities were not there from the start, but were tacked on shortly after its initial design). Heck, even DOS-based Windows from Win95 onwards facilitated and encouraged the separation of user and application data. Windows 98 and up even had per-user home directories and registry hives. The last time putting application configuration in the same location as the binaries was considered acceptable, Windows 3.11 was the platform.

Whether or not the typical usage of Windows is logically single-user, is irrelevant to good programming practices, which is most assuredly NOT what you are advocating. This is *basic* security and privilege separation you're trying to argue against.

Windows 95 was the most successful release in Microsoft's history because nothing was taken more seriously than the idea that existing applications, even 16-bit DOS apps from the Jurassic period, had to keep running. Vista failed (well, if you don't count encouraging millions of high-margin retail purchases of Windows 7 as 'failure') because the company abandoned that ideal at every level, from driver support to the filesystem.

The lengths Microsoft went to preserving compatibility in Vista were *at least* as extensive as those in Windows 95, and probably more so - which is why they had to implement that ridiculous directory remapping scheme in the first place. The fact they couldn't maintain legacy compatibility with applications that were already broken a decade ago is in no way a valid criticism. If you look at all the compatibility shims in Vista (and Windows 7) and somehow conclude Microsoft "abandoned" backwards compatibility, then it's kind of hard to do anything but laugh bemusedly. Heck, the mere fact they continue to support 16-bit applications at all - despite them being 15 years out of date - pretty much says it all. Try firing up essentially any unmodified 15-20 year old application - not even just the badly written ones - on pretty much any other stock contemporary OS and see how much luck you have. I guarantee that Windows is going to run a higher proportion than anything else, except maybe Solaris.

Re:Statistics, statistics

[Man On Pink Corner]

Basic best practices (stretching back decades) is the first and foremost reason. Reducing the risk of trojans, virus infections and general data loss three more. Separation of binaries and configuration information a fourth.

Translation: "I can't think of a single reason why .INI files in the Program Files hierarchy are a bad thing, so I'll mumble some stuff about trojans, viruses, and data loss, and wrap up with a tautology."

Re:Statistics, statistics

[drsmithy]

Translation: "I can't think of a single reason why .INI files in the Program Files hierarchy are a bad thing, so I'll mumble some stuff about trojans, viruses, and data loss, and wrap up with a tautology."

I can see now why you say you only get "dead silence" to your requests. You're too busy yelling LALALALALALA at the top of your lungs with your fingers in your ears whenever anyone responds.

Re:Statistics, statistics

[ruiner13]

I never said I was the one who made said VB garbage...

Memory

[Anonymous Coward]

I'll bet that very, very few of those have more than 4GB of RAM or utilize the security and virtualization enhancements. The reason why half are running 64-bit is that all the major computer manufacturers pre-loaded it on all their low-end sale PCs.

Why, oh why?

[palegray.net]

Is there a good technical reason for 32-bit Windows 7 not supporting more than 4 GB of RAM, period? PAE has been in use for a long time now, and while you can't have a single process that exceeds 3 GB in Linux (tunable, I'm given to understand, can also be a 2 GB per process limit in some installations), you can definitely go past 4 GB of total system memory. Windows Server 2008 Enterprise supports 64 GB per 32-bit system...

Re:Why, oh why?

[0racle]

Why enable a workaround when there is a native way to support it? PAE does also technically have a performance impact, your average desktop user isn't exactly going to understand that.

Re:Why, oh why?

[shutdown -p now]

Is there a good technical reason for 32-bit Windows 7 not supporting more than 4 GB of RAM, period? PAE has been in use for a long time now, and while you can't have a single process that exceeds 3 GB in Linux (tunable, I'm given to understand, can also be a 2 GB per process limit in some installations), you can definitely go past 4 GB of total system memory.

PAE can break badly written drivers, which are more common on desktop versions of the OS than they are on server versions.

Re:Why, oh why?

[refrigeratorpanic]

from a business perspective, probably market segmentation to protect the server business. but technical -- i dont know.

Re:Why, oh why?

[ByOhTek]

No, because according to MS, PAE is available for Windows 7 32 bit:

http://msdn.microsoft.com/en-us/library/aa366796(VS.85).aspx

Re:Why, oh why?

[mpfife]

Well, the first off, PAE only gets you to 64gb of memory. While that should be plenty for most people for the foreseeable future, we all know that setting arbitrary and somewhat lower limits turns you into the most quoted man in history (640k should be enough for everyone....) Also, as I recall, the 64-bit memory manager in Vista was quite a bit different (and faster) than the one in the 32-bit version. Legacy support(?) However, there is more than just addressable memory to consider with a 64-bit operating system. If you use PAE, your APPS are still running 32 bit. Apps need to be recompiled or even reworked/rewritten to utilize the new 64-bit operating system features. That's probably what they're hoping for more - to get folks to thinking and writing in 64-bit. While I don't think it was a totally cool move - if I'm not mistaken Microsoft has some features available only in the 64-bit api's. In Vista, certain secure driver signing modes didn't exist except in the 64-bit version (not that this is a good thing - they were terrible - but it does show a difference).

Re:Why, oh why?

[palegray.net]

As far as performance impacts go, 64-bit systems will use more memory for the same set of running programs than otherwise equivalent 32-bit systems.

Re:Why, oh why?

[Skuld-Chan]

For the 64 gig support on a 32 bit machine you often need special servers with chipsets that bank the memory appropriately and special system drivers (Serverworks is/was famous for this) on top of that - its really only something you need to do if you were running Metaframe (I think its called XenApp server?) because most Windows apps won't go past 2 gigs of allocation anyhow.

My understand is the reason for this is just special hardware/driver support - many consumer motherboards for instance map real world pci resources in the 4 gig address range. Its probably easier on quality assurance to only support what they do on server OS's.

64 bit system doesn't have any of these limitations and you can address all the memory in one chunk without any work-arounds - hence the wider support for more ram there.

Re:Why, oh why?

[TheRaven64]

PAE adds another layer to the page tables (as does 64-bit addressing), which makes TLB misses more expensive, so you don't want to enable it on systems that don't have more than 4GB of RAM. Given that very few machines ship with more than 4GB of RAM, but a 32-bit processor, it's likely that this would be a configuration that would get very little or no testing (especially from driver developers) so would be potentially very buggy.

A lot of PCI devices are 32-bit, so drivers need to use bounce buffers to do DMA transfers to physical memory over the 4GB line. This is something that device drivers designed for 32-bit systems won't do, because they can just pass 32-bit physical addresses straight to the device on the systems they were written for.

Re:Why, oh why?

[sam0737]

Pure technical reason - short answer: nope.

This is what I have heard - it's not that 32-bit can't do PAE (as you mentioned, in Server SKU it support up to 64GB), but drivers for client 32-bit SKU are usually written are tested for PAE, which means more bluescreen if you insist doing so.

Re:Why, oh why?

[Nemesisghost]

Actually, I believe it has to do with memory addressing. Applications use a single word for memory addresses. Since each byte is addressable, you are limited by the size of your memory address. And a 32-bit INT has the capacity of addressing roughly 4 billion locations. 32-bit Systems

Re:Why, oh why?

[yoyhed]

Are you actually running MORE than 4GB of RAM with a 32-bit processor? If so, I'd assume it was an older server, in which case you'd probably be running Windows Server which indeed does support more than 4GB for 32 bit. 64-bit Windows compatibility these days is awesome with everything but old 16-bit applications (and you can just run an emulator or VM for those) - no reason not to use the 64-bit version if you've got 4+ gigs of RAM.

Re:Why, oh why?

[soupbowl]

Which adds up to not that much and unless you are stuiped, you are rocking 4+ gigs of ram.

Re:Why, oh why?

[drsmithy]

Is there a good technical reason for 32-bit Windows 7 not supporting more than 4 GB of RAM, period?

Yes. Firstly it requires applications to be modified to really see much benefit, secondly it hurts performance, and thirdly it breaks a myriad of poorly written third-party drivers and other low-level applications.

Re:Why, oh why?

[Anonymous Coward]

For more than 18 months, I've been running Linux on a (64 bit) Core I7-920.... with 12 GB of RAM. Not buggy, not buggy at all.

Re:Why, oh why?

[bernywork]

You try accessing more than 2GB of RAM (or 3GB of RAM with the /3GB switch in boot.ini) in a single process. What you end up having to do is (firstly) your own memory management (Which sucks) and having to manage multiple 2GB "windows" so if you want to read data you have to swap in an out of these "windows" to be able to read them as the kernel itself is only 32bit and can only directly address 4GB of RAM.

So you end up coding in what is known in Windows as "Addressable Window Extensions" and they are a pain in the arse. Doing this on SQL server and Oracle was basically a necessity, and when PAE was first thought of, this is exactly what was being thought about, database systems. They have been able to use PAE in VMware etc and other places as they give the upper and lower limits for memory address directly to the operating system (Windows, Linux whatever else is actually running in the VM) and then they address via the hypervisor that memory address space, meaning that the hypervisor doesn't have to do a lot of memory management (Certainly nothing like protected memory)

So in effect, the biggest reason for 32bit Windows 7 not supporting more than 4GB of RAM is because the kernel itself is a 32bit app and doesn't have the 64bit address space to directly address more than 4GB of RAM itself.

In the long term it's just too hard, and it's easier to code for 64bit than to deal with what are effectively kludges to make this work.

If you need to know more about this, I would suggest Mark Russinovich's Windows Internals book.

Re:Why, oh why?

[Hadlock]

Pushing people off 32bit so that 64 bit drivers are properly developed for is a pretty good reason.
 
  I bet in two years' time with the wider adoption of 64 bit atom processors we'll see Win7-64 being 80%+ of the install base. Considering both versions cost the same there's no reason to go with 32 bit unless you have some horrible backwards app that can't handle being run in XP-32 compatibility mode as administrator.

Re:Why, oh why?

[bernywork]

For the 64 gig support on a 32 bit machine you often need special servers with chipsets that bank the memory appropriately and special system drivers (Serverworks is/was famous for this) on top of that

Sorta, you could go to 8, 12, 16 or 32 with modern motherboards and available RAM (Well above 4GB) but that doesn't help the situation.

its really only something you need to do if you were running Metaframe

*cough* What?!?!? After databases needed the RAM (They brought in PAE with the Pentium Pro) then the next problem was Metaframe. Metaframe was NT4.0 TSE era. Having 20 users login to a system was a great way to burn up RAM on a single host. Metaframe is one of the biggest drivers of 64bit systems as you can now get more users on a box WITHOUT the overhead of virtualisation (A lot of people still run XenApp on Windows on a hypervisor as it gives a few benefits in regards to CPU scheduling etc)

My understand is the reason for this is just special hardware/driver support

Nope.

many consumer motherboards for instance map real world pci resources in the 4 gig address range

It's the drivers that do this so that the operating system can address this memory as it's a 32bit app

Re:Why, oh why?

[Anonymous Coward]

640k should be enough for everyone...

He never said that.

Re:Why, oh why?

[h4rr4r]

The real issue is market segmentation. The driver issues would not exist if the windows driver space made any damn sense.

Re:Why, oh why?

[vlm]

Yes. Firstly it requires applications to be modified to really see much benefit, secondly it hurts performance, and thirdly it breaks a myriad of poorly written third-party drivers and other low-level applications.

So? Its never stopped them before.

Re:Why, oh why?

[drsmithy]

So? Its never stopped them before.

Stopped who doing what ?

Re:Why, oh why?

[0123456]

Firstly it requires applications to be modified to really see much benefit

Modified? You mean setting the 'allow me to see more than 2GB of RAM' bit on the executable? Which they should be doing in any case so that it works better on 64-bit Windows? That sounds like lots of work.

PAE is invisible to applications, has a small performance impact in the worst case and a significant performance benefit in the best case. Microsoft just chose not to support it because they didn't want the support calls when poorly-written drivers fell over.

Re:Why, oh why?

[Anonymous Coward]

Don't forget that all the hardware drivers like video cards use up that precious address space too. You can easily end up with 4G of RAM installed but only 2.1G usable at all as the rest of the address space is used up by hardware busses, video cards, etc.

Running 32-bit OS these days is as stupid and shortsighted as people clinging to DOS even years after Windows NT/95 were introduced. 64-bit mode has access to 2x as many CPU registers so it runs faster, even if only the kernel runs 64-bit and the app is still 32-bit.

Re:Why, oh why?

[Sir_Lewk]

unless you are stuiped

Please tell me I'm about to be woooshed. Please. That'd be far better than the alternative.

Re:Why, oh why?

[PitaBred]

But having more registers available as well as other tweaks, it will almost always run faster while only using a relatively negligible amount of extra memory. The only thing that uses more memory are memory pointers, basically. A game won't use a lot more RAM because 256MB of textures doesn't magically double.

Re:Why, oh why?

[drsmithy]

Modified? You mean setting the 'allow me to see more than 2GB of RAM' bit on the executable? Which they should be doing in any case so that it works better on 64-bit Windows? That sounds like lots of work.

Applications need to use the "Address Windowing Extension" API to get any real benefit from PAE.

PAE is invisible to applications, has a small performance impact in the worst case and a significant performance benefit in the best case. Microsoft just chose not to support it because they didn't want the support calls when poorly-written drivers fell over.

Which is quite reasonable, given the drivers falling over is not their fault, or problem.

PAE is a hack. The real solution is to go to x64, and the proportion of users who can't switch to x64 but would derive real benefit from PAE on Windows XP, is vanishingly small.

Re:Why, oh why?

[ThatMegathronDude]

PAE adds a 4 bit page table to the existing addressing scheme. This means that the 64GiB that is "addressable" with PAE is actually 16 pages of 4GiB. Since the PAE table is not part of the normal 32 bit pointers that a program uses, the program has to be written to chunk its data into 4GiB pages and know which page houses what data.

tl:dr; PAE is not invisible, and only provides performance benefits when you need that much memory on one machine.

Re:Why, oh why?

[mcgrew]

While that should be plenty for most people for the foreseeable future, we all know that setting arbitrary and somewhat lower limits turns you into the most quoted man in history (640k should be enough for everyone....)

I think it's pretty ingenious marketing. What better way to get your customers to shell out more money?

Re:Why, oh why?

[drsmithy]

The real issue is market segmentation. The driver issues would not exist if the windows driver space made any damn sense.

What ?

Re:Why, oh why?

[drsmithy]

I think it's pretty ingenious marketing. What better way to get your customers to shell out more money?

What do they need to pay more money for ?

Re:Why, oh why?

[Anonymous Coward]

Thats all fine and good, but what of businesses for whom the constraints (?) of a 32bit operating system were still more then their needs? MS dropped all 32bit server OS so one must buy new hardware. Glad that it has performance advantages for larger enterprises, but for little folks one is either forced to keep running the old stuff or convert to another OS (with all the compatibility issues to the microsloth desktops that this engenders). W7x64 is a reasonable environment for editing LARGE photo images or video (1gb or more) but hardware overkill if all you want to do is surf the web or read emails. Just because some sales type wants to force you to junk your existing hardware and buy all new everything does not make it a good idea or a sensible business proposition for YOU.

Re:Why, oh why?

[Dutch Gun]

we all know that setting arbitrary and somewhat lower limits turns you into the most quoted man in history (640k should be enough for everyone....)

You do realize the 640K limit was not "arbitrary", right? It was a limitation of addressable space of the pre-386 Intel processors. And in fact, I'll just take this opportunity to again point out the dubious nature of this quote.

Re:Why, oh why?

[cbhacking]

PAE is already enabled on nearly all Windows boxes; it's required for hardware DEP. As I understand it, MS limits 32-bit lcient Windows to 4GB because some drivers explode if they see more than 4GB, not because they don't have PAE.

Server builds are less-likely to use bad drivers (wide pointers are a requirement for WHQL drivers, so most - but not all - drivers should be fine). On this assumption, and because servers need more RAM than most clients/workstations anyhow, Windows Server x86 can use the full potential of PAE.

Re:Why, oh why?

[thegarbz]

Desktop user? Why would they care? Also the performance impact is tiny compared to the impact of seeking on the harddisk. The article itself should come with a knowledge disclaimer:

"Half of the windows 7 Machines Running 64-bit Versions, 90% of users don't even realise"

The average desktop user doesn't know what defragment is, has 50 apps that load on startup, and think computers naturally get slower over time and simply require replacing.

Re:Why, oh why?

[shutdown -p now]

The real issue is market segmentation.

That doesn't make sense at all, given that Win7 64-bit costs precisely the same as Win7 32-bit. Indeed, they both come in the same box!

Re:Why, oh why?

[Bigjeff5]

Hello? They just had to release Windows 7, and breaking a myriad of poorly written third-party drivers was a primary factor in Vista's failure. What planet have you been living on?

The fact is, Microsoft (usually) works harder than any other OS manufacturer to maintain backwards compatibility and ensure new changes don't break old software. They did not do that with Vista, and they got their asses kicked for it.

Apple gets away with telling OSX to pound sand every year because they've been doing it that way forever. Seriously, try to run three year old software on a new Mac. Chances are it won't work. Now do the same with a Windows box. I'd be surprised if it didn't work, because almost everything does (though bad security practices of the past cut out more apps than most Windows upgrades do).

Re:Why, oh why?

[pipedwho]

Well, the first off, PAE only gets you to 64gb of memory. While that should be plenty for most people for the foreseeable future, we all know that setting arbitrary and somewhat lower limits turns you into the most quoted man in history (640k should be enough for everyone....)

Even with 64GB of memory, the per process limit with PAE is still limited to less than 2GB.

This is equivalent to your current system with 2GB of memory having a per process limit of 64 mega bytes. In that situation there would be no reason to even care that your total system memory was any larger; since most apps would have maxed out to a nearly unusably small memory window.

So there is no functional reason for Microsoft to spend any effort whatsoever making PAE support more than 64GB of total memory.

Re:Why, oh why?

[h4rr4r]

PAE is supported on server versions.

Re:Why, oh why?

[shutdown -p now]

Are you implying that PAE is disabled in client 32bit editions of Windows just to make people who have PCs incapable of running the 64bit client (today it pretty much means netbook or nettop) to buy server?..

On any PC which has 4gb to begin with, you'd just install 64bit Win7 and be done with it. So where's the $$$ for MS?

Re:Why, oh why?

[Anonymous Coward]

Apple gets away with telling OSX to pound sand every year because they've been doing it that way forever. Seriously, try to run three year old software on a new Mac. Chances are it won't work.

You haven't actually tried this, have you?

I run applications written for PowerPC MacOS X 10.2 on Intel MacOS X 10.6. That isn't just more than 3 years ago, that's a different processor architecture which can only still run because they still support the Rosetta emulator.

Re:Why, oh why?

[Anonymous Coward]

PAE is on, but MS still locks you at 4GB for desktop OS it seems, as a driver self defense measure. I have yet to see anyone say they have a working 32bit windows desktop OS actually using more than 4GB total.

They really should update the PAE page to reflect the desktop OS hard limits

Re:Why, oh why?

[Eskarel]

All PAE actually does is allow each individual process to have a 4 GB of RAM limit(well 2 in actuality on a 32 bit system), it doesn't really solve the RAM issue as you still can't actually utilize extra RAM effectively(a process which needs more than 4 GB of ram is still stuffed), it's a rather nasty kludge to make it work, and it has performance issues.

Given that with the exception of drivers you can install 32 bit software on 64 bit with zero problems, and that the 64 bit versions of Vista and 7 are as good or better than the 32 bit versions, the only reason to buy 32 bit is if you have a very specific hardware requirement or because you're an idiot. Why should Microsoft stuff a rather nasty kludge which can affect performance and stability into a system you deliberately chose knowing the inherent limitations?

Re:Why, oh why?

[Anonymous Coward]

None. (Contary to what others say, I assure you the reason is licensing limitations. XP is license limited to 4GB of RAM.)

There ARE issues where DEP enabled XP (i.e., XPSP2) will not use more than 4GB of physical address space[1]; although I just read a blog post that suggests manually enabling /PAE on the boot.ini might sidestep that (URL below). (PAE was not the norm for XP pre SP2; so many consumer-grade device drivers did not handle physical addresses >4GB and just crashed. With SP2, adding NoExecute (NX) meant that PAE was generally on and all sorts of device drivers had issues with this - apparently. Thus, XPSP2 normally runs in a truncated PAE -- PAE is on, but nothing over 4GB is used.)

http://blogs.msdn.com/b/carmencr/archive/2004/08/06/210093.aspx

[1] Once you allow for PCI, enforcing a maximum of 4GB of physical address space means 4GB of RAM is usable. Typically, 3.5GB.

Re:Why, oh why?

[LO0G]

That's not quite true. The 8088 processor could address 1M of RAM, not 640K. But IBM chose to reserve 360K of RAM for video and ROM support in the initial PC architecture.

The 640K value came from IBM's hardware engineers, not MSFT. In fact MSFT sold a PC board for the XT that allowed the machine to address up to 768K or so (up to the start of the CGA video memory).

Not that my comment fundamentally changes your response. The 640K was arbitrary but not the limitations of the addressable space.

Re:Why, oh why?

[Ant P.]

64 bit system doesn't have any of these limitations and you can address all the memory in one chunk without any work-arounds - hence the wider support for more ram there.

My kernel disagrees.

[ 0.000000] Node 0: aperture @ 180000000 size 32 MB
[ 0.000000] Aperture beyond 4GB. Ignoring.
[ 0.000000] Your BIOS doesn't leave a aperture memory hole
[ 0.000000] Please enable the IOMMU option in the BIOS setup
[ 0.000000] This costs you 64 MB of RAM
[ 0.000000] Mapping aperture over 65536 KB of RAM @ 20000000

Re:Why, oh why?

[mcgrew]

A new OS that will handle the extra memory.

Re:Why, oh why?

[drsmithy]

A new OS that will handle the extra memory.

The 640k limitation was imposed by the hardware, not the software.

I'm also a little confused why having to buy new hardware to do X is A-OK, but having to buy new software to do X is bad. Or do you rail against the hardware vendors because they don't give free upgrades for life as well ?

Re:Why, oh why?

[mcgrew]

If I add memory to a PC, that's hardware. The 640k limitation was hardware, yes, but that was the last hardware limitation, and it was decades ago.

Most hardware I've bought came with the necessary software; if I buy a CD burner, it will have CD burning software included. If I buy a wifi router, it will also have the necessary software.

I don't buy stuff so somebody can make money, I buy stuff because I need or want it. Whether or not anyone makes money doesn't enter into the equation at all, and if I have to buy a second product to make the first work, I feel robbed.

Re:Why, oh why?

[drsmithy]

If I add memory to a PC, that's hardware. The 640k limitation was hardware, yes, but that was the last hardware limitation, and it was decades ago.

The ability of a 32-bit OS to support only 4GB of RAM is no different. Similarly, the fact that software needs to be modified to actually take advantage of PAE is no different. 32-bit desktop versions of Windows not supporting PAE is somewhat different, in that it could technically work, but the reasons for not supporting it are technically sound (beneficial to only a tiny proportion of users, breaks due to third party software, etc).

Most hardware I've bought came with the necessary software; if I buy a CD burner, it will have CD burning software included. If I buy a wifi router, it will also have the necessary software.

Who sold you a PC with >4GB of RAM and Windows XP but didn't make it explicitly clear that configuration wouldn't work ? Heck, who sold you a PC capable of using more RAM without making that limitation explicitly clear ? Take it up with them.

Whether or not anyone makes money doesn't enter into the equation at all, and if I have to buy a second product to make the first work, I feel robbed.

Like filling your card with petrol, you mean ?

Re:Why, oh why?

[toddestan]

Is the Core i7-920 a 32bit processor? (hint: no)

limits

[StripedCow]

192GB ought to be enough for anyone...

Re:limits

[AnonymousClown]

Since the days when 640K would be enough for anyone, memory requirements have been increasing dramatically. But, what real functionality have we've received as a result of the bloat? What can you do today that you couldn't do back then? The spreadsheets are the same. Pretty much the same for Word processors - sure there's more esoteric functionality for any particular 1% of the users but is it really needed?

Is it the games that caused the bloat?

I see this bloat but no real improvements.

Re:limits

[chrisl456]

Haha. But wait... why is there a 192GB limit? Shouldn't it be much higher? Wikipedia says that it should be at least 256 TB.

Re:limits

[Anonymous Coward]

Games are a big deal. For those of us who *aren't* using Linux, we're well past the days of CGA games that run from a single floppy. Some of us demand more, and all of that map and texture data isn't just going to store itself.

And games are just for starters. What about music and video? You can't play MP3s or watch Blu-Ray video (or any other type for that matter) with 640k of RAM. Some people want to use our PCs for more than spreadsheets or word processors, and 640k is more than a little limiting.

(Admittedly, modern word processors and spreadsheet programs are definitely more bloated than they really need to be, but we've gained functionality as a result -- like the ability to actually see what our documents will look like when they're printed, instead of lines of monospaced text. Being able to edit larger documents and spreadsheets without continually having to go to disk to get more data is nice too.)

Re:limits

[Anonymous Coward]

They limit up to what they can test. If they had a desktop machine with Windows 7 Ultimate running with 256 TB of Ram, it would be the new limit.

Re:limits

[parlancex]

They also like to impose arbitrary limits to segment their market, such as the limits on the number of physical CPUs supported in most versions of Windows.

Re:limits

[gumbi west]

Use a GUI with more than 1 or 4 bit color?

Re:limits

[uglyduckling]

• edit multi-channel audio and mix it in real time
• real-time effects for the above, including convolution filters
• edit full-screen, high definition video
• edit RAW format digital photographs non-destructively
• have photo realistic graphics and hi-fi sound, composited and mixed in real time

Need I go on?

Re:limits

[Anonymous Coward]

now I can finally run crysis

Re:limits

[Anonymous Coward]

Edit large digital photos.
Run a database larger than a few megs.
Edit video.
Edit music with high quality samples. (Not MOD files and friends.)
Play videos in a web browser.
Use a GUI that is pleasing to the eye.
And no matter what it is, it can be done faster with more RAM.

I dislike bloat as much as the next guy. Faster, cheaper hardware RAM allows designers to crank software out with less niggling concern over performance.

However, due to the other benefits of having powerful computers, I'd never want to go back to the old days.

Do not lament: for now and long into the foreseeable future companies will still be making 8 and 16 bit chips from which designers wrestle every idle cycle and free byte of memory.

Re:limits

[Jeremy Erwin]

The only app that consistently reminds me that I need more RAM is preview.app, MacOSX's pdf reader. I'm addicted to Americana-- and a thousand page book with illustrations gobbles ram like crazy.

Re:limits

[elrous0]

Crysis on the highest settings is like light speed--you can get close, but you're never going to make it all the way.

Re:limits

[Anonymous Coward]

Right, but for the editions of Windows like the highest-end Windows Server, the point where they set the limit is the best system they could get their hands on.

Re:limits

[gumbi west]

Two of these were already being done back them. Specifically, "Run a database larger than a few megs." use the hard drive. Not really sure how memory changed this, probably mostly for huge merges.

And video editing was happening on the Amegia Video Toaster, of course the videos were analog.

Re:limits

[nlawalker]

But, what real functionality have we've received as a result of the bloat?

From a developer's point of view, the availability of more resources has improved our ability to abstract stuff away and work with high-level concepts.

From a usability point of view, the availability of more resources has allows for caching of lots of stuff in memory for faster response, and working with lots of data at once. We kind of take it for granted now that we can throw around playlists with zillions of songs or albums with thousands of photos with ease.

Re:limits

[Anonymous Coward]

192GB ought to be enough for anyone...

...for now.

Re:Incoming incessant sopssa trolling.

[binarylarry]

I hate Sopssa as much as the next guy but does he even post here anymore?

Maybe he finally wandered back to join his herd at channel9.

Shipped that way

[Sporkinum]

The majority are probably shipped that way. The new PC I bought for my wife came with Win7 Home Premium. It's dual boot though, as she has been on a KDE desktop for years and prefers that to Windows. She tried out Win 7 for a couple of days before I got Kubuntu set up for her, and she didn't care for it.

32 at work, 64 at home

[Monkeedude1212]

We use the 32 bit at work for the compatability with old the old MS Access databases (don't ask... I just work here...)

I use the 64 bit at home - even though it causes some odd glitches with various games here and there, for the most part it runs everything much smoother. I decided that I'd need more than 4 Gigs of RAM to run Visual Studio to Debug my modified Source game.

Re:32 at work, 64 at home

[NJRoadfan]

64-bit should be fine for most. For those 32-bit apps that glitch out (or those random Win16 apps, like the old Windows Entertainment Pack games), just run them with XP Mode.

Re:32 at work, 64 at home

[Monkeedude1212]

At work we find a lot of our apps need to run in Compatability mode regardless of 32 or 64 bit. - I mean they were written in like PowerBuilder 4 or something... Just finished getting them working with Oracle 10g.

At home, I find that on the odd occaison, Dragon Age Origins will have a glitchy moment, kind of like tearing but not. I mean I've fiddled around with compatability and video options, to no avail - I would normally suspect the graphics card in that case but I was pretty sure a GTS 240 could handle it.

Re:32 at work, 64 at home

[shaunbr]

Other than the 16-bit apps, which 64-bit Windows 7 *finally* removed support for, I've had few issues with older programs. I even loaded Might and Magic VI (written in ~1997) and it loaded up and ran without problems -- I didn't even need to use XP compatibility mode.

Microsoft may get a lot of criticism here (much of it rightly deserved), but backwards compatibility is something they've almost always managed to get right. For the last few years I wondered how much pain we'd run into when the 'average' desktop PC finally hit the 4GB RAM barrier and had to move to a 64-bit OS, but Microsoft has managed to make it mostly painless. Of course, backwards compatibility brings bloat, but since many 64-bit users are already over the 4GB barrier, I think it's a reasonable tradeoff.

Re:32 at work, 64 at home

[Dracker]

You need Access on 64 bit? Get these hotfixes:
http://www.microsoft.com/downloads/details.aspx?familyid=C06B8369-60DD-4B64-A44B-84B371EDE16D&displaylang=en
http://www.microsoft.com/downloads/details.aspx?familyid=C06B8369-60DD-4B64-A44B-84B371EDE16D&displaylang=en

I, too, have to work with Access databases at work and these hotfixes let me do it on our 64 bit servers.

Re:32 at work, 64 at home

[VGPowerlord]

XP Mode doesn't support 3D graphics. At all.

Which leaves you two choices: VirtualBox, which supports OpenGL plus DirectX 8/9 if you replace one of the OS's files, or VMWare Player which supports OpenGL and all versions of DirectX out of the box.

Re:32 at work, 64 at home

[Quietust]

Other than the 16-bit apps, which 64-bit Windows 7 *finally* removed support for

Unless I'm mistaken, all 64-bit versions of Windows have been incapable of running 16-bit applications, since any version of Windows only provides one emulation layer - 32-bit versions have NTVDM+WOWEXEC for 16-bit apps, and 64-bit versions have WoW64 for 32-bit apps (which, if you're running Server 2008 R2 Core, can actually be uninstalled).

Re:32 at work, 64 at home

[Anonymous Coward]

Unless I'm mistaken, all 64-bit versions of Windows have been incapable of running 16-bit applications, since any version of Windows only provides one emulation layer

It's got nothing to do with what Microsoft chooses to support. x86 processors disable all support for 16 bit code once you switch them into 64 bit mode.

You can't run 16 bit software on 64 bit Windows unless you resort to emulating the entire machine.

Re:32 at work, 64 at home

[bigstrat2003]

Honestly, I would think your Dragon Age problem is hardware-related, not OS. I run Win7 x64, and I had no such problems with that game (and my graphics card is only a GeForce 8800 GTS, so not even as good as yours). FWIW, I run Ultimate, but I can't imagine it's a problem with Ultimate vs whatever you have.

Re:32 at work, 64 at home

[Missing.Matter]

I played Dragon Age on Windows 7 Home Pro 64 bit, without a hitch. The card was an integrated Intel HD on a laptop.

Re:32 at work, 64 at home

[BatGnat]

I run windows 3.11 in DosBox on my Win7 64bit. It works great.

Re:32 at work, 64 at home

[FictionPimp]

I have no issues with win7 64bit and dragon age. I have beaten the game at least a dozen times.

Re:32 at work, 64 at home

[Anonymous Coward]

Yo dawg, I heard you liked virtual machines... and so forth.

Re:32 at work, 64 at home

[BatGnat]

Intel I7, I can run a shit load of them at once

MMmmmm......Multiple GPF's........Homer Like.

What about flash?

[JasonStevens]

We've only been waiting FOREVER for a Win64 version of Flash from adobe...

Although I will say this, the state of drivers for Win64 is far better then the early days (NT 3.1/3.5) of Win32.

Re:What about flash?

[shutdown -p now]

We've only been waiting FOREVER for a Win64 version of Flash from adobe...

They probably don't feel like they need to rush it, because, as it is, Win7 x64 still ships with 32-bit IE as a default browser (due to need to preserve plugin compatibility), and all other mainstream browsers only release officially supported Windows versions in 32-bit.

Re:What about flash?

[KlomDark]

What is the deal with those slothful people? Are they planning to make their business obsolete. I thought I'd see 64 bit Flash long before we saw 64 bit Windows hit the 50% mark.

If I was a stockholder, I'd be pissed and selling.

Re:What about flash?

[Skuld-Chan]

You would? What actual advantage would you get from a 64 bit version of flash over the 32 bit version? None really unless it was an absolutely massive project.

The only reason they did anything about it on Linux is because the default browser (often Firefox) was a 64 bit build.

Re:What about flash?

[Joe Snipe]

Did you just say that we don't have 64-bit flash because Windows uses 32-bit IE because we don't have 64-bit flash?

Re:What about flash?

[ashridah]

Yep. Probably the most secure browser in the world is a 64bit one, since you can't run any plugins, and you're using one of the most uncommon stack layouts... :P

Re:What about flash?

[hannson]

I don't get what a browser would gain from 64-bit support at the moment (even with all the 64bit plugins), I mean most browsers today are going towards a multi-process architecture making it unlikely for them to hit the 32bit process limit and going x64 would just increase the memory usage. Anyone care to shed some light on why 64 bit browsers are so important today?

Re:What about flash?

[Anonymous Coward]

If you were a stockholder with brains you'd want them to invest in projects that have more projected revenue than cost to develop/support. How much additional revenue is 64 bit Flash going to bring in? How much will it cost to develop?

Re:What about flash?

[Anonymous Coward]

IE is included in a 64-bit build. Mozilla 4 beta (nightly) has a 64-bit build. Going to 64-bit makes sandboxing easier. 64-bit flash will be more secure if nothing else.

There are reasons other than data size to go x86-64. NX bit is a biggie.

Re:What about flash?

[PitaBred]

The actual advantage is not taking a lot of disk space for 32bit compatibility libraries, as well as performance being lower in 32bit apps as compared to 64bit. Your statement reeks of "640K should be enough for anybody!"

Re:What about flash?

[BatGnat]

Firefox 4 will have 64bit version. Running it on home laptop now.

Now, where is my 64bit flash....?

Re:What about flash?

[shutdown -p now]

Did you just say that we don't have 64-bit flash because Windows uses 32-bit IE because we don't have 64-bit flash?

Yes. If you recall, it was the same vicious circle on Linux not long ago. It was kinda broken by nspluginwrapper, but that solution isn't 100% fool-proof. The difference is that there aren't that many plugins to deal with on Linux - really it's mostly just Flash and (for the masochists) the embedded Adobe Reader. So you only need to make a hack that's good enough to make those work.

On Windows, though, there are a lot of various "download managers" and other crapware that works as a browser plugin, and that's absolutely essential for countless people and (more importantly) businesses to keep working the way they do. So providing a hack that works on some plugins and breaks others is not a reasonable option.

The recommendation to use 32-bit version of Office 2010 exists for the same reason, by the way. You absolutely can use 64-bit one, but a lot of stuff that integrates into Office today will simply not install on top of that - various dictionaries, translators, advanced formula editors, proofing tools etc. And a lot of people rely on those.

Next time you write a heavily extensible application, keep that story in mind, and don't step in the same trap: use managed code (.NET, Java, Lua, Python - whatever rocks your boat) for plugins. Or, if you allow native code plugins, at least implement them out-of-process straight away, so that bitness difference isn't an issue -but that, of course, won't help you if you have to switch to an altogether different architecture - and, given the growing popularity of ARM-powered devices, it is a distinct possibility regardless of which platform you target.

Re:What about flash?

[shutdown -p now]

IE also has a 64-bit version, since IE7 IIRC (it was pre-installed in Vista, and pops up if you search for "IE" from Start menu search bar).

The problem isn't availability of browser as such. The problem is what is configured as a default browser out of the box / with minimal hassle. In both Vista and 7, the default is 32-bit IE. In case of Firefox, it's what you get if you go to http://firefox.com/ and click on the huge green "Download" button - and that, again, will be 32-bit Firefox.

Re:What about flash?

[Bigjeff5]

The IE that ships with 64bit Windows 7 is 32 bit.

How is sandboxing easier in 64bit, and in what way is security better in 64bit vs 32bit? And I need a better reason than "well nobody is writing exploits for 64bit yet".

Re:What about flash?

[adolf]

My Windows 7 x64 box has both 32- and 64-bit versions of IE. I didn't do anything special to get them both.

Rumor has it that, out of the box, the 32-bit version is set to be default, but that's rather a different matter.

Re:What about flash?

[Anonymous Coward]

We've only been waiting FOREVER for a Win64 version of Flash from adobe...

Although I will say this, the state of drivers for Win64 is far better then the early days (NT 3.1/3.5) of Win32.

Then stop going to flash only websites and let that POS just die.

Re:What about flash?

[KlomDark]

At this point I'd be far more concerned about the negative publicity from all the people who are thinking your company is obsolete and cannot keep up with technology.

The two statements not contradictory

[TechnoLuddite]

While the ratio of 50% does seem high to me, there's no contradiction between a 64-bit OS and a 32-bit app. Windows 7 64-bit (and numerous other 64-bit operating systems) can run 32-bit apps.

I'm not sure I see the supposed conflict in the last statement.

In another news

[miknix]

I've been running 64bit Gentoo Linux since I bought one the first models of Athlon 64, which was almost 7 years ago!

Re:In another news

[Anonymous Coward]

Compile time doesn't count.

Re:In another news

[Fwipp]

Exactly - for my first computer (about a week and 7 years ago), I made explicitly sure to buy a 64-bit processor - it was the future! I didn't want to be left out when the entire computing world caught up to 64 bit computing.

7 years later we're still waiting for developers to catch on *cough* *Adobe.*

Re:In another news

[magellanic]

I've been compiling 64bit Gentoo Linux since I bought one the first models of Athlon 64, which was almost 7 years ago!

FTFY

Re:In another news

[Anonymous Coward]

And you just finished compiling ?

Re:In another news

[Anonymous Coward]

... SO? 64-bit servers and real UNIXes have been around for ages, and they do actual work. What has 64-bit mode accomplished for the Linux desktop besides make you feel important?

*chirp* *chirp*

Same boat here

[RingDev]

Just got the word that the desktop team is pushing out Window 7. Unfortunately, there are "a couple" of printers that they couldn't get working 64b drivers for. So they are pushing the 32b version out to everyone...

Blows my mind... It would cost at most a $5000 to replace those printers, compared to the cost of 600+ copies of Windows 7. Crazy.

-Rick

Re:Same boat here

[Skuld-Chan]

If its anything like where I work (community college) actually doing the work to get more printers purchased is more work then installing 32 bit version of Windows.

The other thing - having dealt with this exact issue - often similar drivers will work on older printers. For example - if there are Vista 64 or XP-64 drivers those will work - or failing that a similar PCL-5 driver (that may actually come with the OS). Also many printers support postscript emulation - and while postscript isn't technically device independent its often close enough to work.

Re:Same boat here

[Joce640k]

Is there a compelling reason for using 64 bits...? If not, stick with the printers.

Re:Same boat here

[GNU(slash)Nickname]

Just got the word that the desktop team is pushing out Window 7. Unfortunately, there are "a couple" of printers that they couldn't get working 64b drivers for. So they are pushing the 32b version out to everyone...

Blows my mind... It would cost at most a $5000 to replace those printers, compared to the cost of 600+ copies of Windows 7.

If you've got 600+ seats, I would assume you also have an MSCA or MSEA agreement, in which case the incremental cost of the Windows upgrade licenses is zero.

Re:Same boat here

[elrous0]

Actually, a Windows 7 license can be used to install either version. So if you ever wanted to upgrade to 64-bit, all you would need is a physical install disc for the 64-bit version. You wouldn't need to buy new licenses.

Re:Same boat here

[LinuxIsGarbage]

Just got the word that the desktop team is pushing out Window 7. Unfortunately, there are "a couple" of printers that they couldn't get working 64b drivers for. So they are pushing the 32b version out to everyone...

Blows my mind... It would cost at most a $5000 to replace those printers, compared to the cost of 600+ copies of Windows 7. Crazy.

-Rick

Considering everywhere I've been plans on staying on XP forever, I'd consider 7-32 bit a good step. The machines can always be switched to 64 bit.

Re:Same boat here

[OdinOdin_]

The Retail edition of Windows 7 comes with both 32bit and 64bit licenses. Only one installed copy though.

This maybe part of the reason there is more proliferation of 64bit because some people can go back to 32bit if they wanted. There was no purchase choice to make.

I'm sure volume licensing shaves off the options.

Why 64-bit is ready now

[Robotron23]

One of the most obvious things about the development of 64-bit architecture is how woefully behind its 32-bit counterpart just a few years ago. I had a spare rig which I put XP 64 on in around spring of 2007. The low level of development together with the fact that powerful machines did not often have more than 4GB back then meant little incentive for devs to put time in 64-bit. I couldn't even find hardware drivers, and this led my disgruntled self to format that particular rig, whilst telling myself to research this stuff in future.

I believe Microsoft deserves some cred, along with certain hardware firms like AMD/Intel, with bringing 64 bit to the fore. Not to mention the PS3 and some Macs being of that architecture too. Ironically Microsoft's most stunted OS since the much maligned Windows ME was the first one that could run 64 bit convincingly despite a laundry list of flaws that haunted the entire life cycle of that particular product. Windows 7 is in some ways like XP was to ME a whole 9 years ago.

Whilst it's great for RAM purposes, and thus demanding things like gaming which will soon require 6GB or more for popular titles there are drawbacks. A file in 64 bit takes up more memory, mainly due to alignment padding. Thus one needs a fairly good set of chips to cache efficiently in future years as the levels of memory inevitably increase. However with the amount of progress going on I daresay all but the most budget hardware solutions will tackle drawbacks very well.

That the figure is now 50% compared to about a fifth of that not long ago is indicative that 64-bit has finally become established in the mainstream.

Re:Why 64-bit is ready now

[Anonymous Coward]

Whilst it's great for RAM purposes, and thus demanding things like gaming which will soon require 6GB or more for popular titles there are drawbacks. A file in 64 bit takes up more memory, mainly due to alignment padding. Thus one needs a fairly good set of chips to cache efficiently in future years as the levels of memory inevitably increase. However with the amount of progress going on I daresay all but the most budget hardware solutions will tackle drawbacks very well.

The one thing people keep forgetting is how register-starved the 32-bit x86 is compared to the 64-bit. Going 64-bit on the x86 has performance benefits in addition to the large memory space. Another benefit is an explicit availability of fast vector instruction sets such as SSE(1,2,3+), which are not guaranteed to be on the 32-bit x86 CPUs.

Re:Why 64-bit is ready now

[TeknoHog]

I believe Microsoft deserves some cred, along with certain hardware firms like AMD/Intel, with bringing 64 bit to the fore. Not to mention the PS3 and some Macs being of that architecture too.

Which architecture would that be? PS3 has a 64-bit PowerPC CPU, and so did Power Macs with the G5. Does this have something to do with the 32/64-bit transition in x86?

If you mean 64-bit architectures in general, don't forget the Alpha which was released in 1992.

Of course, despite the bitness upgrade, the closed Microsoft world remains stuck in the x86 world. Meanwhile, others are free to choose the best/nicest platform for the job.

Re:Why 64-bit is ready now

[Shimbo]

I believe Microsoft deserves some cred, along with certain hardware firms like AMD/Intel, with bringing 64 bit to the fore.

Sadly, that's largely a reflection on the distorting effect that Microsoft has on the market. Intel and Microsoft were just following where just about everybody else had gone before.

Re:Why 64-bit is ready now

[StikyPad]

Vista was actually a much better implementation of 64-bit than WinXP-64, and Vista's limited driver and software support was *nothing* compared to the wasteland that was XP-64. I realize there was a lot of pissing and moaning about Vista, but you probably would have experienced far less heartache by going that route in the Spring of '07 than "sticking with" XP (which you weren't actually doing anyway, since 64-bit XP comes from a separate codebase).

Re:Why 64-bit is ready now

[Gadget_Guy]

Of course, despite the bitness upgrade, the closed Microsoft world remains stuck in the x86 world. Meanwhile, others are free to choose the best/nicest platform for the job.

The Windows NT line has run on the Alpha, MIPS and PowerPC architectures. Microsoft dropped support for these a long time ago because hardly anybody used anything other than Intel's offering. More recently, Apple has also chosen to do the same thing and most Linux desktops are x86.

So I guess being able to choose the best/nicest platform doesn't mean you will be any different than Microsoft.

Re:Why 64-bit is ready now

[BatGnat]

Ironically Microsoft's most stunted OS since the much maligned Windows ME was the first one that could run 64 bit convincingly despite a laundry list of flaws that haunted the entire life cycle of that particular product.

When was Windows ME ever 64bit? Never that's when...

Re:Why 64-bit is ready now

[IMightB]

Why give credit to Microsoft? They are coming to the 64b game years late and with underwhelming products. I've been using 64b linux/unix variants since the alpha and then the AMD x86_64 chips.

Re:Why 64-bit is ready now

[TeknoHog]

The Windows NT line has run on the Alpha, MIPS and PowerPC architectures. Microsoft dropped support for these a long time ago because hardly anybody used anything other than Intel's offering. More recently, Apple has also chosen to do the same thing and most Linux desktops are x86.

So I guess being able to choose the best/nicest platform doesn't mean you will be any different than Microsoft.

Windows NT on non-x86 is history, but you can install the latest version of Linux on dozens of architectures. Complete distros with all the usual libraries and applications. So this is very, very far from a sensible comparison.

Re:Why 64-bit is ready now

[Kaboom13]

I used 64-bit XP for years (skipping Vista completely) on my home machine. I never had a problem with drivers. I never get this idea that gets spouted every time it comes up. Even my printer had a driver. Even most apps that require a driver I never had problems finding 64-bit versions for. Granted, I built my machine myself, using parts from vendors with a reputation with supporting their hardware, so if you bought a box from Dell filled with cheapest parts they could get that week, your experience may have varied. I think a lot of people tried xp64 right after it came out, and couldn't find drivers, and gave up. Of course any MS OS is going to be short on drivers at the start.

Re:Why 64-bit is ready now

[Spad]

You *can* but how many people actually do compared to x86? And given that, would any of those options be offered if those distros had to make a financial return on their development time in the same way that Microsoft does?

Re:Why 64-bit is ready now

[drsmithy]

Which architecture would that be? PS3 has a 64-bit PowerPC CPU, and so did Power Macs with the G5. Does this have something to do with the 32/64-bit transition in x86?

The PS3 was released at the end of 2006, with the first x86-64 CPUs hitting the market in mid 2003 and the first x86-64 version of Windows in mid-2005. Heck, the PS3 only just beat _Vista_ to the market. The first PowerMac G5s were released in mid 2003, though OS X still isn't completely 64-bit even today.

If you mean 64-bit architectures in general, don't forget the Alpha which was released in 1992.

The Alpha (and others like SPARC) are hardly mainstream hardware platforms.

Of course, despite the bitness upgrade, the closed Microsoft world remains stuck in the x86 world. Meanwhile, others are free to choose the best/nicest platform for the job.

There are very few jobs x86 can't do well, and basically zero where Windows would be a considered choice anyway.

Re:Why 64-bit is ready now

[StikyPad]

I wasn't claiming that it was impossible to build a machine that could run 64-bit XP problem free; nonetheless, Vista had a larger library of drivers than did XP-64 (which was sort of a bastard stepchild of the Windows desktop family). Additionally, not all software ran cleanly (or at all) in 64-bit XP, while Vista-64 expanded software compatibility substantially since it was a "mainstream" release.

Re:Why 64-bit is ready now

[Johnno74]

I thought he said that to, but re-read it... he's talking about vista:

"Microsoft's most stunted OS since the much maligned Windows ME"

Re:Why 64-bit is ready now

[Gadget_Guy]

Why give credit to Microsoft? They are coming to the 64b game years late and with underwhelming products. I've been using 64b linux/unix variants since the alpha and then the AMD x86_64 chips.

They are hardly late to the game. The first 64 bit version of Windows XP was released simultaneously with the 32 bit version in 2001. It ran on the Intel Itanium processor (obviously, because it predated AMD64 product line). The version of XP for the consumer level of 64 bit CPUs came out a year after Intel released their copy of the AMD64 system.

It wasn't until fairly recently that 64 bit computers began to enter the mainstream because people finally started coming up against the 4GB limit.

Re:Why 64-bit is ready now

[Gadget_Guy]

Windows NT on non-x86 is history, but you can install the latest version of Linux on dozens of architectures. Complete distros with all the usual libraries and applications.

That is true, and if you take the Linux based servers and appliances into account, non-x86 CPUs make up a sizable chunk of the market. But as I said, on the desktop (where the average consumer makes the choice of the platform), people want standard off the shelf components. It is there that Intel is still the king.

So this is very, very far from a sensible comparison.

The point was that Microsoft didn't "remain stuck in the x86 world". The NT line was built on a hardware abstraction layer to allow them to support multiple systems. But it was the general public who didn't want to use anything other than Intel-type CPUs.

Re:Why 64-bit is ready now

[Anonymous Coward]

... Not to mention the PS3 and some Macs being of that architecture too....

Ahem ... let's not forget who invented 64-bit.

Re:Why 64-bit is ready now

[TeknoHog]

But as I said, on the desktop (where the average consumer makes the choice of the platform), people want standard off the shelf components. It is there that Intel is still the king.

Does an average consumer ever choose Linux, or any particular OS for that matter? What about smartphones and other mobile devices? I think many people choose Android without considering the Linux aspect at all.

The point was that Microsoft didn't "remain stuck in the x86 world". The NT line was built on a hardware abstraction layer to allow them to support multiple systems. But it was the general public who didn't want to use anything other than Intel-type CPUs.

That's a very good point, Windows as such is not constrained to x86. But with the "Microsoft world" I mean the whole ecosystem, and most applications are still written for x86 only.

I expect mobile devices to change this slowly, since you cannot practically run x86 on a phone. Another interesting development is Microsoft's .NET, as it would enable applications on multiple architectures to maintain close ties to the Windows monopoly. Of course, this has been tried already with Java, but it has taken fairly closed systems such as Symbian and Android to actually make it work.

Having Windows (with applications) on many architectures would be nice from the point of view driving healthy competition on the hardware side, to the benefit of opensource users as well.

Re:Why 64-bit is ready now

[BatGnat]

My Bad.

Artificial limits R US (tm)

[Twinbee]

Is there a reason they can't go above the artificial limit of 192 GB?

64 bit CPUs should be able to access up to 18,446,000,000 GB of memory space, so I see no reason for the arbitrary limit.

Re:Artificial limits R US (tm)

[Anonymous Coward]

money, dear boy

Microsoft defines 'defective by design', and you should know that FAT32 is limited to 120GB and is outdated (lies)

Re:Artificial limits R US (tm)

[Zerth]

Is there a reason they can't go above the artificial limit of 192 GB?

Because then Windows Server wouldn't look very impressive.

Re:Artificial limits R US (tm)

[Anonymous Coward]

If they can't test it, they don't want to ship with it.

Re:Artificial limits R US (tm)

[MobileTatsu-NJG]

Is there a reason they can't go above the artificial limit of 192 GB?

Is there a realistic way of testing it past that amount?

Re:Artificial limits R US (tm)

[amorsen]

Most (even server) CPU's don't support more anyway. It isn't a hard limit like the 640k or 4GB limits though, the next version can support more while remaining binary compatible with user programs. Given that it's hardly a problem.

Re:Artificial limits R US (tm)

[Anonymous Coward]

18446000000 / 16GB = 1,152,875,000. Note at 400MHz(the minimum DDR3 speed) light travels about 1 meter per clock cycle. Such a memory array would be much larger than 1 cubic meter, making DDR latency numbers impossible to meet even with lightspeed interconnects.

Yes, IAAAD(I am an ASIC designer)

Re:Artificial limits R US (tm)

[Anonymous Coward]

http://en.wikipedia.org/wiki/64-bit#Current_64-bit_microprocessor_architectures

The x64 has a 48 bit limitation on the virtual address space.

Re:Artificial limits R US (tm)

[CannonballHead]

What, you mean software is supposed to be tested on what it claims to support? I thought we could just assume the coding was fine. ;)

Seriously though, you're right. You don't claim to support something unless you tested it on that, usually. Perhaps this is different with *cough cough* most free software, I don't know (I wouldn't blame them particularly, since I doubt most free/open-source devs have access to copious amounts of ram).

Re:Artificial limits R US (tm)

[KarmaMB84]

x86-64 can only address 48-bits and Windows only addresses 44-bits (16 TB). The Windows limitation is interesting because no windows release to date can even touch that address limitation.

My best guess would be that OS releases are artificially limited to the amount of memory they actually test internally against. Home Premium probably doesn't get serious testing beyond 16GB while Ultimate might get tested against 192GB workstation hardware. High end server releases probably get tested with up to 2TB (probably the maximum amount of hardware available at time of testing). 32-bit desktops probably don't even get tested with PAE enabled at all since chances are desktop hardware drivers will crash and burn so they get a 4GB limit.

Re:Artificial limits R US (tm)

[janeuner]

In this context, bit-width refers to the size of a data register, not the size of the address space. The address space of an x86-64 processor varies between 40-bits up to a full 64-bits, depending upon the generation of a particular architecture.

Further, consider the purpose of such a mechanism. DD3 can move data at a rate in the range of 2^34 bytes/sec. If we had a memory pool of 2^48 bytes, it would take 4 hours just to read the full contents of that memory space one time. This is clearly unusable, so that "artificial limit" is practical and efficient given current technology.

Re:Artificial limits R US (tm)

[John Hasler]

> I don't know (I wouldn't blame them particularly, since I doubt most
> free/open-source devs have access to copious amounts of ram).

Right. How could the "devs" at kernel.org and Mozilla, for example, possibly have access to fancy hardware? After all, they are only backed by little companies such as IBM, Intel, Oracle, HP, Google...

How many different architectures is your software built and tested on?

Re:Artificial limits R US (tm)

[CannonballHead]

Me personally? I work at one of those large companies you mentioned. We support x86 (32 and 64 bit), ia64, Power, PARISC (HPUX), and SPARC... OS's are RHEL/SLES (x86, ia64), HP-UX (PARISC and ia64), Solaris (SPARC), and AIX (POWER). We have very large test matrices.

I know there are some big-hitter free software projects that are backed by large companies and have fairly large revenue streams. I was thinking more about the small-time free software projects, not corporate-backed projects.

Re:Artificial limits R US (tm)

[Anonymous Coward]

Well, 192GB isn't that far fetched I think. Granted, right now it's a ridiculous amount of memory to have in a home PC, but just think how much more RAM you've got now than you had back in 2001. I had 192MB then and now my home PC has 8GB.

There's no reason for limit this, other than for separating the versions. Windows 2008 supports 2TB in the Enterprise version I think. As for testing, just shell out some dough and buy a HP or Dell server :)http://tech.slashdot.org/story/10/07/12/1545224/Half-of-Windows-7-Machines-Running-64-Bit-Version?art_pos=4#

Re:Artificial limits R US (tm)

[MrSteveSD]

Microsoft puts arbitrary memory limits on it's different flavours of 64 bit windows. Home Premium will only recognize 16GB whereas the more expensive Professional will recognize 192GB. Physical limitations are one thing, but arbitrary limitations annoy me. I don't exactly know what 64 bit Linux will support, but at least there won't be arbitrary limitations.

Re:Artificial limits R US (tm)

[Anonymous Coward]

Is there a reason they can't go above the artificial limit of 192 GB?

Is there a realistic way of testing it past that amount?

Looks like Microsoft has tested up to 2 TB, which is the maximum supported on x64 Datacenter and Enterprise versions: http://blogs.technet.com/b/markrussinovich/archive/2008/07/21/3092070.aspx

Re:Artificial limits R US (tm)

[Anonymous Coward]

It's probably because the page table would need more levels of indirection to handle the full 48 bits of address space modern amd64 processors have. The current generation of hardware has another (higher) 'arbitrary limit', as i mentioned in the previous sentence, i.e. you don't know what you're talking about. It seems that the missing 64 GiB from 2^38 bytes is probably kernel space (i.e. that's a user mode limit).

Re:Artificial limits R US (tm)

[Kielistic]

I do believe the CPU manufacturers do not put 64 addressing lines in so it is physically impossible to address the full available space. On top of that I don't believe there is any non-blade, non-cluster systems that support that much RAM and there will not be in the foreseeable future.

Market segregation is also a means to make money. Many companies use this technique, but in 2010 I'd hardly call 192 an overbearing limit.

Re:Artificial limits R US (tm)

[Anonymous Coward]

an inverted centrifugal array of dilithium crystals, placed in a micro blackhole should create enough of a tachyon field to allow placement of a cubic meter or larger, array of asics to operate at FTL interconnect speeds.

-geordi

Re:Artificial limits R US (tm)

[quo_vadis]

Are you sure about this ? My computer architecture is a little rusty, but let me see .. what you are saying is that you would need 2^30 RAM sticks of 16GB capacity to fill up the full 2^64 space. This contention is fine. The problem is with the second part. The only way you would look at the complete delay is if you had sequential accesses. If you had a partitioned hierarchy, it is fine. Additionally, the TLB might be large, but addressing the ram can happen orders of magnitude faster (assuming partitioning happens on basis of 5 bits, worst case,you only look into a bank of 2^6 sticks) which is not that bad. Of course, access time is not uniform, but this problem has been addressed previously (see NUMA systems). As such, a full multi exabyte system is hard to design, but with increasing memory densities, it may become feasible, using techniques that are currently applied to supercomputer memory hierarchies.

Re:Artificial limits R US (tm)

[mariushm]

Yes, this 900$ motherboard, 32 x 8GB memory modules each about 450$ and four 305$ processors and a 200$ power supply will let you test 256 GB of memory.

Sure, it costs about $20k but what the hell, some cars are more expensive. You won't be able to use Windows 7 on it anyways, as I think it's limited to two physical processors.

Re:Artificial limits R US (tm)

[KillShill]

64-bit CPUs actually use 40-bit physical addressing (48-bit virtual). So you're not really getting the full range. Not that it matters, since 40-bits should be enough to last 5-10+ years for the vast majority of home users. By then, newer CPUs will be out to take full advantage of 64-bit memory space. There's even talk about 128-bit...

http://en.wikipedia.org/wiki/X86-64

But this is all a moot point, given the massive collusion in the memory market, price fixing will make sure that huge ram systems will only exist in a far off future.

Re:Artificial limits R US (tm)

[Stan Vassilev]

Is there a reason they can't go above the artificial limit of 192 GB?
64 bit CPUs should be able to access up to 18,446,000,000 GB of memory space, so I see no reason for the arbitrary limit.

Windows 7 Home Premium works fine with 18446000000 GB, I've just tried it on my desktop.

Hope this helps.

Stan

Re:Artificial limits R US (tm)

[Anonymous Coward]

Is there a realistic way of testing it past that amount?

Install SQL 2005 and start the service.

Re:Artificial limits R US (tm)

[cookd]

There are several different limitations in effect. Each of them may or may not apply, depending on specific circumstances.

- Number of address pins on the CPU or socket: 8088 had 20, limiting it to 1MB; 80386 had 24, limiting it to 16 MB; Pentium had 32, limiting it to 4 GB; Pentium Pro had 36, limiting it to 64 GB; ...

- Number of address lines on the motherboard: many motherboards (especially laptops) only connect the low 32 address lines, limiting them to 4 GB.

- Chipset memory support: chipsets may limit the size and quantity the memory module, i.e. max 4 memory modules, cannot support modules larger than 512MB.

- Device requirements: Memory-mapped devices (like video cards) usually need to be mapped to an address below some limit. On 8088, the range from 640k-1MB was reserved and couldn't be used for RAM. In PCI, the mapping is more flexible, so address ranges are only reserved if there is actually a device that needs the range. Instead of reserving the range, the device "shadows" the memory at that range (if any). This is usually in the 3GB to 4GB range.

- Chipset remapping support: Memory that is "shadowed" cannot be accessed at its "natural" address. If the motherboard does not support remapping, shadowed memory is inaccessible and wasted. If the motherboard does support remapping, shadowed memory can be remapped to some other address, usually somewhere above 4GB.

- OS PAE support: To access memory above the 4GB limit from a 32-bit CPU, the OS must put the CPU into PAE mode. All modern OSes support this. 32-bit Windows will run in PAE mode if you've enabled "Data Execution Prevention", whether or not you have more than 4GB of memory. (This is not an issue for a 64-bit operating system -- there is no need for PAE if the chip is running in 64-bit mode.)

- Driver support: DMA transfers to PCI devices only work with addresses below 4GB. Drivers must be designed to special-case any requests to transfer data to/from addresses above 4GB (usually done by allocating a temporary buffer below 4GB and copying the data to the temporary buffer). Systems that use drivers that aren't high-address-aware must avoid using physical addresses above 4GB.

- Application support: A 32-bit application can directly access no more than 4GB of address space, some of which will be reserved for use by the operating system. There are workarounds that enable indirect access to more RAM, or the application can be recompiled as a 64-bit application.

- OS edition limitation: Commercial OSes often artificially limit RAM usage for various reasons including edition differentiation and driver compatibility.

The 4GB limitation on consumer editions of Windows is due to driver compatibility issues. (Note that this is a limit on physical address size, not a limit on the amount of memory.) Microsoft originally placed a completely artificial 8GB limit on the amount of memory available to consumer editions of XP (for edition differentiation reasons only). However, driver problems appeared on systems that used addresses above the 4GB limit, so Microsoft changed this to a 4GB address limit (not a 4GB memory limit!) to improve stability on consumer systems (and reduce support calls). This limit is not present on server systems, since Microsoft assumes that server administrators will use higher-quality hardware and/or drivers. Cheaper editions of the server OS do have artificial memory limits (i.e. 16GB). The most expensive edition of the 32-bit server OS has no artificial limit and can access the full 64GB exposed by 36-bit PAE.

64-bit systems also have limits. As far as I know, the 192GB limit is entirely artificial - if you need more memory than that, Microsoft probably figures you should buy one of the server editions of Windows. The 1TB limit on the server edition is probably based on test limits - Microsoft doesn't want to support something it hasn't tested. However, even with a 64-bit edition of Windows, you may still be limited (i.e. my old laptop only had 32 address li

Re:Artificial limits R US (tm)

[tokul]

Is there a realistic way of testing it past that amount?

Maybe 200 GB pagefile will do the trick.

Re:Artificial limits R US (tm)

[bjb]

Is there a reason they can't go above the artificial limit of 192 GB?

I've read that the only reason why they limit to 192GB is simply because there isn't any x64 hardware out there that supports more than 192GB. I think it is because of the chipsets, not necessarily the RAM sticks or number of slots.

In other words, Microsoft supports up to 192GB because that is the highest amount of memory that they have been able to test with.

I would assume that as soon as 256GB hardware becomes available, they might re-state that limitation.

If only PAE wasn't disabled in software

[Artem S. Tashkinov]

... people could still use 32bit in computers with up to 64GB of RAM in Windows. ;)

However I'm not really sure if 64bit Windows is good or bad, now the amount of compatibility craft is multiplied by two because Windows cannot run without 32bit libraries vs. Linux/other Unix'es where you can have a pure 64bit environment.

Only half?

[JustNiz]

Everything I've ever thrown at 64 bit windows runs just fine, and usually somewhat smoother than 32 bit. Even some really old stuff. The only software I ever found that don't run on 64-bit are some really old dos games and utilities, but then they didn't even run under 32-bit XP either.

It boggles my mind why so many people with 64-bit hardware would still install a 32-bit version of windows. I wonder how much of this is actually ignorance and/or just force of habit rather than actual knowledge that they have something that actually doesn't work under 64 bit.

Re:Only half?

[WinterSolstice]

I run 64bit Win7 just fine as my gaming/art box, haven't found anything incompatible yet.
I don't run Office, though - just portable apps, Adobe CS video/graphics, and games.

Re:Only half?

[Lord Ender]

The Intel Atom netbook and nettop computers require 32 bit OSs. So there really are new, popular computers being shipped which just can't run 64bit.

Re:Only half?

[Anonymous Coward]

While I'm sure most of the Atoms out now only support 32-bit I doubt there are many new ones released with that limitation since the second generation atoms released late last year all support 64-bit. Some of the older models (Atom 200, Atom 300 and the higher end Atom Z) also support it.

Re:Only half?

[LordKronos]

Everything I've ever thrown at 64 bit windows runs just fine, and usually somewhat smoother than 32 bit. Even some really old stuff. The only software I ever found that don't run on 64-bit are some really old dos games and utilities, but then they didn't even run under 32-bit XP either.

It boggles my mind why so many people with 64-bit hardware would still install a 32-bit version of windows. I wonder how much of this is actually ignorance and/or just force of habit rather than actual knowledge that they have something that actually doesn't work under 64 bit.

Don't forget that upgrading from 32 to 64 requires a fresh install (you can't just upgrade in place).

Re:Only half?

[Anonymous Coward]

It boggles my mind why so many people with 64-bit hardware would still install a 32-bit version of windows. I wonder how much of this is actually ignorance and/or just force of habit rather than actual knowledge that they have something that actually doesn't work under 64 bit.

You do realize that going from a 32-bit installation of Windows to a 64-bit installation requires:

a) backup
b) reformat (optional, but suggested)
c) Windows 64-bit install
d) software reinstall
e) restore

in that order, right? I've got copies of 64-bit Windows 7 for my home machines, but the prospect of losing 4-8 hours in the afternoon on a weekend is daunting, especially as I work quite a bit after hours on my home machine.

In contrast, going from 32-bit to 32-bit just requires one step - "install 32-bit copy of Windows 7."

Re:Only half?

[UnknowingFool]

I guess the real question is how the average user would know that their hardware is 64-bit. It's not like there's a sticker that says:"64-bit Ready". Newer hardware are most likely 64-bit but not always. Some research is required but remember some users still think the desktop is their file system.

Re:Only half?

[DarthVain]

In a Word, Vista.

When I build my last computer, Vista had just come out. I was left with a very crappy decision.

Buy a copy of XP, older technology that won't be supported. Or buy Vista, by all accounts very much a buggy, and a real bitch to get any drivers to run properly. Supposedly the 64-bit edition was even WORSE. This was like the first year of the Vista release, bad timing I know, but I needed a computer.

I also didn't want to run Linux as I wanted to play computer games.

So I took what I thought was the lesser of evils and bought Vista 32-bit for compatibility. All told it has worked pretty good.

I tried to get a free upgrade MS was offering at one point to 64-bit Vista, but was told this was only for "Retail" and not "OEM" so I was SOL.

I am certainly not buying anything new. When I build a new computer in the next couple of years hopefully I have better choices available to me than I did the last time. By all account Win7 seems to be ok, but who knows what will be the OS flavor of the day when it comes time to build again...

Re:Only half?

[Anonymous Coward]

Mandatory Authenticode signatures for 64-bit kernel-space programs (i.e. drivers).

If you want to hack around your machine a bit, you can put it in test mode and turn off driver signing, but then you get ugly cruft on your screen corners saying "Test Mode."

If you were writing your own stuff, you could conceivably get your own Authenticode cert (around $200/year) but for those that like to hack stuff together, MS's requirement that drivers be signed on 64-bit makes 64-bit a no-go.

Re:Only half?

[JustNiz]

>> In contrast, going from 32-bit to 32-bit just requires one step - "install 32-bit copy of Windows 7."

Dunno about you, but I've always ended up with an ugly mess after doing an in-place upgrade.

I always do a through e on your list whenever I install windows anyway. I always avoid microsoft backup/restore utility though as that causes its own problems. I just do a direct file copy to another drive of anything I want to keep.

why I'd pick 32 bit

[ILuvRamen]

I had to work on someone's Vista 64 bit machine and I hated it. Not only were half the programs running in 32 bit mode but almost none of my virus removal tools worked so I couldn't completely disinfect it. Three different antivirus programs wouldn't install properly on it either. Almost no software I had ran on it and for some reason, Java 32 bit was installed and 64 bit wouldn't install. If I wanted a computer that no software ran on, I'd buy a mac.

Re:why I'd pick 32 bit

[CannonballHead]

So, because of experiences with Vista x64 (which everyone agrees was not exactly a great OS to begin with), you have decided that 7 x64 is bad. Even though almost everyone in this thread has reported GREAT experiences with the x64 version of Windows 7.

I ran Vista x64 and Windows 7 x64 and had few problems with either (none with 7, a couple with Vista). So now we are even with our one-on-one anecdotal evidence. ;)

Re:why I'd pick 32 bit

[icebrain]

My only complaint is that drivers for my old Spaceball 4000 serial 3D mouse aren't available for Win 7/64. I had to drop $120 on a newer model that isn't as comfortable to me.

Re:why I'd pick 32 bit

[Malc]

That must have been some time ago because x64 Windows is pretty good these days. The only thing that's causing me problems with it at the moment is the lack of Cisco VPN client. As for the 32 bit apps running in 32 bit mode - who cares?

I don't think you know much about the current state of Macs either.

Re:why I'd pick 32 bit

[Anonymous Coward]

Nice troll post. Try better next time.

Re:why I'd pick 32 bit

[0123456]

Not only were half the programs running in 32 bit mode

Only half of them? Is there any 64-bit software that the average Windows user has installed other than IE, which is useless due to lack of Flash plugins?

Did Microsoft even recompile notepad and paint to 64-bit?

Re:why I'd pick 32 bit

[JDeane]

I call BS on the OP whole post... AV programs have always clearly stated what versions are intended for 32 or 64 bit.... You have too since they work at the driver level.

Re:why I'd pick 32 bit

[CannonballHead]

And I have never had a problem running AV programs on x64 installs... usually using Avast.

Re:why I'd pick 32 bit

[Quietust]

Did Microsoft even recompile notepad and paint to 64-bit?

Last I checked (which was a while ago), any 64-bit version of Windows has all of its bundled applications (notepad, paint, etc.) compiled as 64-bit, and some apps are also provided in 32-bit for compatibility (particularly Internet Explorer, since you can't load a 32-bit ActiveX control into a 64-bit process).

Re:why I'd pick 32 bit

[Fred Or Alive]

Did Microsoft even recompile notepad and paint to 64-bit?

Yep, Notepad and Paint are 64 bit (at least with Windows 7). A random sampling (Wordpad, Solitaire, DVD Maker, Media Center) seems to show that most of the programs that come with 64 bit Windows are indeed 64 bit. The setup_wm.exe program that pops up if I attempt to launch Windows Media Player isn't though, but I can't be bothered setting up a program I have no intention of using to see if the main program is 32 or 64 bits.

Re:why I'd pick 32 bit

[Anonymous Coward]

Combofix doesn't work at all which is probably his concern.

Re:why I'd pick 32 bit

[ShadowRangerRIT]

64 bit Vista/Win7 has Windows Media Player in both a 32 and 64 bit version, but the default shortcut and file type registrations point to the 32 bit version. Not because of any problem with WMP itself, but because most codecs are still unavailable or unstable in 64 bit mode. Like with Internet Explorer and 32 bit only plugins, the root problem is that it uses third party software to function and the third party developers haven't bothered to upgrade to 64 bit (and of course, they see no reason to hurry, because 32 bit is still the default).

Re:why I'd pick 32 bit

[BatGnat]

PEBKAC

Re:why I'd pick 32 bit

[Pro777]

I've had the same issues with the Cisco VPN client. Apparently they have 64-bit support in their newest Beta, but without access to that I've been using the Shrewsoft client.

Re:why I'd pick 32 bit

[Anonymous Coward]

If I wanted a computer that no software ran on, I'd buy a mac.

What now?

1999 called, it wants its criticism back.

Re:why I'd pick 32 bit

[phishtahko]

I'm guessing op renamed his AV programs after downloading and didn't have a goddamned clue which was applicable to which version.

Re:why I'd pick 32 bit

[Anonymous Coward]

but almost none of my virus removal tools worked so I couldn't completely disinfect it

Where you seriously trying to clean the machine in the compromised host environment? Not only do you need to learn about 64-bit Vista, my friend, but you also need to learn about pre-installed environments too.

Re:why I'd pick 32 bit

[Malc]

Thanks for the tip! I'll try that out later.

Upgraded?

[Gonoff]

For a long time, when we get new machines, the first thing we do is upgrade them from Vista to XP. This is likely to continue.

Re:Upgraded?

[RocketRabbit]

Using XP is not choosing to remain behind any curve to speak of. Sure you don't get whizzy 3D gadgets, but with a sane IT infrastructure there is nothing dated about XP, and new software for it continues to be offered and in fact dominates the market today. New games that require Vista or 7 are few and far between. Microsoft Office software from a decade ago is in fact functionally identical to today's versions. Exchange half-works in its pathetic usual way as poorly on Windows 7 as it does on XP.

Perhaps that's the saddest thing about Microsoft. They have been putting lipstick on a pig for so long, that they married the pig and are expecting their second litter.

Re:Upgraded?

[Eskarel]

You can't buy XP after the first Service Pack for Windows 7 comes out. No new XP licenses period for anyone. Need a new staff member, welcome to a mixed environment.

Re:Upgraded?

[Sheepy]

You can't buy XP after the first Service Pack for Windows 7 comes out. No new XP licenses period for anyone. Need a new staff member, welcome to a mixed environment.

Microsoft have changed that policy:

To support our customers' "unprecedented move" to migrate their PC environment to Windows 7, we have decided to extend downgrade rights to Windows XP Professional beyond the previously planned end date at Windows 7 SP1. This will help maintain consistency for downgrade rights throughout the Windows 7 lifecycle. As a result, the OEM versions of Windows 7 Professional and Windows 7 Ultimate will continue to include downgrade rights to the similar versions of Windows Vista or Windows XP Professional. Going forward, businesses can continue to purchase new PCs and utilize end user downgrade rights to Windows XP or Windows Vista until they are ready to use Windows 7. Enabling such rights throughout the Windows 7 lifecycle will make it easier for customers as they plan deployments to Windows 7.

Re:Upgraded?

[Gonoff]

I work in a corporate environment (hospital). A LOT of money is paid to give what I would describe as a site licence. Of course MS has their usual gobbledegook, but it means that.
This gives us the right to put whatever version of Windows on new PCs we like, as long as they have a sticker on them for something. This means that all our really old kit would have had XP Home licences stuck on them, newer ones had Vista home and new ones will be having Win7 stickers on them.
Every single one of these gets reimaged as soon as it comes out of the box to XP Pro SP3. MS calls it a downgrade licence. Having seen what Vista did to performance, it seemed more of an upgrade.
We are trying all our internal applications on Win7. Mostly web apps on IE8 to deal with still to go. Then all we have to do is make an image that has got all the garbage & pretty stuff tidied away. Maybe we will move to office 2007 at that point too.

Blame MS's Pricing Plans

[EXTomar]

There are multiple version with multiple flavors at different price points that confuses "people". Add to this the finicky way upgrades behave and "upgrade upgrade" software and it is no wonder a lot of people don't care or realize a 64-bit version exists.

Re:Blame MS's Pricing Plans

[phishtahko]

I don't know about this. As far as I'm aware MS gives you both the x86 and x64 version when you buy Win7. You can't really blame luddites for then not knowing the difference between the two. Same with people buying pre-installed - the type of person who doesn't know about x64 probably doesn't need it. MS' horrible version system has nothing to do with it.

Re:Blame MS's Pricing Plans

[ashridah]

This is only true for buying retail editions of Windows 7. OEM licenses are still either 32bit or 64bit, and not both, just like they were with Vista.

Re:Blame MS's Pricing Plans

[phishtahko]

And how would you suggest they remedy this situation? You can't really lay blame on MS for people not knowing the difference between x86 and x64.

HP

[helix2301]

I have noticed quite a few HP computers running 32 bit version of 7 with 6 gigs of RAM I thought that was kind of weird that HP would do something like that.

Re:HP

[phishtahko]

DDR3 RAM for the i3, i5 and i7 is tripple channel, 2GBx3 = 6GB

Printers

[ThatsNotPudding]

I'm informal tech support for an older lady. I just recently replaced her crappy Dell 2350 with a nice little i3 Gateway (nee Acer) that came with Win 7 64. But I had to also replace her Lexmark Z80 as no real 64 bit drivers exist.

Hardware, memory limits.

[Anonymous Coward]

The size of your address bus is going to affect how much RAM your hardware supports; 64 address lines costs more on die than, say, 36-40 address lines. Likewise, storing 40-bit or 48-bit (5-6 byte) address mappings saves space over 8-byte mappings when dealing with virtual memory, paging, etc. When you have that much RAM, you're going to have to start trading off space for page size or suffer some performance issues.

In reality, modern hardware supports even larger address sizes (48-bit on AMD architectures, according to Wikipedia). But there is still some storage overhead on the OS end. One would expect Windows 8 to support something in the 128-256 TB range, etc., as hardware and memory density improve, but probably not go all the way up to 64-bit addressing.

Well the other thing

[Sycraft-fu]

32-bit compatibility in 64-bit Windows is completely seamless. Any user mode app just runs. You are unaware it is 32-bit unless you check the process list, it suffers no noticeable (and hardly even measurable) speed degradation and so on. It just works, you don't have to know or care if the app is 32-bit or not.

That means, for most people, there is no reason NOT to run 64-bit. It is very rare that you have an app that runs on a 32-bit OS but not a 64-bit one. The only real cases are ones that use kernel drivers and don't have 64-bit versions and old 16-bit apps. So very few users are effected.

As such running 64-bit just means the flexibility to run new 64-bit apps, as well as have more RAM, and no real downsides for most people.

Re:Well the other thing

[Robotron23]

Virtually all people won't really notice anything regards 64-bit's transition. I also believe the RAM-limiting elements of x86 were the primary reason for the weighty promotion behind the 64 bit architecture in home computing.

But I also recall 2001 and people mentioning the 'huge' 4GB limit; nobody knew quite when it would be met. However by 2003-4 it was becoming clear that random access capacity growth wouldn't be slowing for a long time. That it took about 8 years since XP's October 2001 release for high-end gaming PCs to outclass the boundaries of Windows x86 is pretty decent when you observe past MS OSs and the duration they have for the 'average' consumer (ie. person who goes to Best Buy or PC World for hardware and support) .

Sycraft-fu, I've been looking for a post of yours in an appropriate topic. I wanted to apologize for this post of mine which was made a few days ago. I'd had a pretty crap day and felt pessimistic enough to author what was a pisspoor, meaningless comment as you rightly noted.

Re:Well the other thing

[Sycraft-fu]

No apology is necessary, in truth I'd not have remembered. I respond to what people say, not who they are on this board. The only people I blacklist are the ones I not in my sig. I don't think less of a person for a given post.

Yes

[Anonymous Coward]

Yes, i believe he did.

Re:Yes

[KnownIssues]

You can't entirely blame the developers. Developers don't write code based on it in being the most elegant thing to do (at least not in commercial software). Why would they spend extra time on code that users don't even realize they want? For that matter, given how much Microsoft waffles on particular products even benefiting from 64-bit (I remember them justifying 64-bit didn't make sense for Office when they released Office 2007 because it actually made some things *slower*), why would the rest of the development community go to 64-bit when there isn't a real demand for it?.

Re:Yes

[Stratoukos]

Why would they spend extra time on code that users don't even realize they want?

Because if they don't they piss off Steve Jobs and they don't get to play with iOS.

Re:Yes

[Anonymous Coward]

Change that slightly.. I'm sure the devs are willing to work on the 64-bit version.. it's just that Adobe's executives are not willing to PAY the devs to do so.

So there you have it. It's not the devs but the tightwad execs that are the problem.

Re:Yes

[LesFerg]

Are you *trying* to sound like a twat or does it just come naturally to you?

In a company that produces and markets software, the developers are probably the most likely group of employees in the entire company who actually know what a 64bit system is, and want to be using one, as well as wanting to have compatible software running on it.

Once again, you echo the same stupid comments. The developers have nothing to do with business decisions made at the executive level. If *you* want to have an effect, determine the companies that do not have marketing policies that please you, and refuse to buy any of their products.

A Mac seems to run everything just fine

[SuperKendall]

If I wanted a computer that no software ran on, I'd buy a mac.

Ha Ha. But what do you mean? The 64 bit transition there has been much smoother. And these days what doesn't run on a Mac? You can after all even run Steam now...

Re:A Mac seems to run everything just fine

[King_TJ]

To be fair, OS X hasn't really fully transitioned people to 64-bits. Yes, the OS supports it, but the 64-bit kernel isn't even enabled by default. You have to hold down the 6 and the 4 keys while booting to tell it to boot into a 64-bit kernel.

Apple most likely did this because of concerns of drivers not being compatible, and wanting to minimize the number of crashes for people upgrading OS X.

Re:A Mac seems to run everything just fine

[aztracker1]

Just because the Steam platform is there doesn't mean all the games magically are there to... Though with Fink or MacPorts there is a lot of FOSS available...

Re:A Mac seems to run everything just fine

[slimjim8094]

Counter-Strike: Source, Half-Life 2, Portal, Team Fortress 2 all run fine. CS:S and TF2 are two of the most popular multiplayer games in the world at the moment.

I, for one, haven't started Windows in months.

Re:A Mac seems to run everything just fine

[phishtahko]

One, Two, Three, Four. Four. Ah ah ah ah. Remind us how many titles other there are on Steam again?

Re:A Mac seems to run everything just fine

[black3d]

Alas, of the 83 games I own on Steam, only 5 run on Mac. I dont' think I'll be switching any time soon.

Re:A Mac seems to run everything just fine

[Anonymous Coward]

To be fair, OS X hasn't really fully transitioned people to 64-bits. Yes, the OS supports it, but the 64-bit kernel isn't even enabled by default. You have to hold down the 6 and the 4 keys while booting to tell it to boot into a 64-bit kernel.

Apple most likely did this because of concerns of drivers not being compatible, and wanting to minimize the number of crashes for people upgrading OS X.

Edit /Library/Preferences/SystemConfiguration/com.apple.Boot.plist

Add arch=x86_64 to the Kernel Flags string

reboot into 64 bit kernel.

(I'm on a late 2008 Unibody MacBook Pro. Works fine)

Re:A Mac seems to run everything just fine

[Anonymous Coward]

If I wanted a computer that no software ran on, I'd buy a mac.

Ha Ha. But what do you mean?

He was trolling of course, and you bit. Sadly though his post is still rated more insightful than yours.

Re:A Mac seems to run everything just fine

[Anonymous Coward]

And just because some games are there, it doesn't mean they'll run well. Most benchmarks I see put source engine games at about half the performance of the exact same machine running windows, and thats not accounting for all of the mac-only bugs you'll run into.

see: http://gizmodo.com/5540716/steam-for-mac-benchmarks-windows-is-much-faster

It's a neat option, but you wont give up dualbooting any time soon if you own a mac and like to game.

Re:A Mac seems to run everything just fine

[lanner]

Yea, I have to second this. As a Mac user at home and work, Macs going from 32 to 64 was more smooth than Microsoft.

That being said, 64 bit Vista and XP were ignorable. I didn't consider them serious and so never even tried. Windows 7 is installed entirely 64 bit at my organization and we have very few problems there either. Some stupid printer drivers and other oddball crap, but we tell the vendor to shape up or we dump them, but we don't do Windows 7 32 bit without a seriously justifiable exception, and there hasn't been one yet.

Re:A Mac seems to run everything just fine

[pandrijeczko]

I'm running my last remaining Windows installation less and less by the day but neither have I ever even considered owning a Mac.

The games above all run fine under Wine in Linux, plus Linux has a Steam port coming soon.

The only thing I have done recently to accomodate Linux gaming is to change out my last ATI graphics card for an NVIDIA one. The proprietary ATI drivers in Linux suck for framerates compared to the NVIDIA ones but framerates for an NVIDIA-driven game running in Wine on Linux (or even natively) seem to be pretty much the same as they are on Windows when I've made some tests.

The only thing that might make the comparison slightly unfair is that my Windows installation is 32-bit XP whereas my Linux installations are all 64-bit Gentoo Linux - but even WINE is running in 32-bit emulation mode, as far as I'm aware.

Re:A Mac seems to run everything just fine

[Craig Davison]

And they only recently made 64-bit binaries the default when compiling with gcc. Before 10.6 you'd have to use gcc-4.2 explicitly, and specify -arch x86_64.

Re:A Mac seems to run everything just fine

[Bigjeff5]

The trouble with a Mac though, is you're forced to upgrade your software every couple of years. Apple doesn't do backwards compatibility. A buddy of mine switched to windows for exactly this reason - he upgraded OSX and low and behold his $500 software application no longer worked, and the only option was to buy a new version. That $100 upgrade (or whatever it was, I think it was actually less) was actually going to cost him over $500. Since there were cheaper alternatives for the same program in Windows, and he could actually buy a much better windows laptop than the Mac he had for around $500, he decided to switch. Six months in and so far he's happy as a clam.

Apple is great at a lot of things, but 3rd party software support and backwards compatibility is not one of them. That happens to be Microsoft's bread and butter (which they were soundly reminded of with Vista).

The problem with win7 64bits: drivers

[asr_br]

It boggles my mind why so many people with 64-bit hardware would still install a 32-bit version of windows[...]

I bought Win7 Home Premium for my gaming station and the package includes both 32 and 64 editions. Initially I installed the 64bits and used it for a month (new hardware with 4GB or RAM) but then I switched to 32bits because of buggy drivers.

I couldn't make my Logitech wingman driving wheel to work, my multifunctional driver was outdated and hard to find (Epson CX7300), my Audigy 2 ZS audio board didn't work out of the box and my network card (Intel whatever, onboard) was performing erratically after some time (not sure why, the win7 diagnostic was "problem with the driver").

Not to mention lots of trouble with low level applications such as dvd drive emulators (magic disk made my machine unusable and IIRC I had a few problems with daemon tools as well).

All of the aforementioned now work fine under win7 32bits.

So I don't recommend people to use Win7 64bits if they have old hardware around (joysticks, printers, scanners) or if they plan to use 3rd party drivers.

--
http://www.ademar.org/ || http://blog.ademar.org/

Re:The problem with win7 64bits: drivers

[Fujisawa Sensei]

It boggles my mind why so many people with 64-bit hardware would still install a 32-bit version of windows[...]

I bought Win7 Home Premium for my gaming station and the package includes both 32 and 64 editions. Initially I installed the 64bits and used it for a month (new hardware with 4GB or RAM) but then I switched to 32bits because of buggy drivers.

I couldn't make my Logitech wingman driving wheel to work, my multifunctional driver was outdated and hard to find (Epson CX7300), my Audigy 2 ZS audio board didn't work out of the box and my network card (Intel whatever, onboard) was performing erratically after some time (not sure why, the win7 diagnostic was "problem with the driver").

Not to mention lots of trouble with low level applications such as dvd drive emulators (magic disk made my machine unusable and IIRC I had a few problems with daemon tools as well).

All of the aforementioned now work fine under win7 32bits.

So I don't recommend people to use Win7 64bits if they have old hardware around (joysticks, printers, scanners) or if they plan to use 3rd party drivers.

Not just old hardware, some A-holes are still building hardware and only shipping 32-bit drivers with them.

We had one of those pieces of garbage, and returned it. And nowhere in the specs did it say 32-bit only.

Not sure why they do that

[Sycraft-fu]

But yes, it is an artificial limit. Their lower end products are limited on purpose. For example Home Premium is limited to 16GB which is as much as you are likely to see in a home system, but pros might want more. Hence 7 pro (and ultimate) has a higher limit.

As to why they choose to limit it to 192GB? No idea.

The actual limit for The Windows NT 6.1 setup is 2TB. I don't remember the particular technical reasons for that, but they are there. There is no reason to process addressing for the full 64-bits of memory when no system exists that can take it. You'll also find that CPUs have memory limits lower than the 64-bit cap. They don't have all 64 address lines because it is not needed. I don't know what it is currently, but it is still below 64-bit. Again, no system could possibly have that much (never mind space, a memory controller couldn't handle the electrical load) so no sense in adding hardware you don't need.

Thus are ARE real limits below the actual 64-bit space but you are correct, 192GB is not one of them. That is an artificial limit and I don't know why they chose it. Doesn't really matter, I do not see people using more than that in a desktop system (144GB is the most I've seen workstation hardware support) and they can always up the limit.

However you are correct that it seems odd.

Compatibility cruft? Cut it.

[peacefinder]

In the next version of Windows Microsoft really should drop Win32 entirely from the main branch, relegating it to an included (and preconfigured) win32 virtual machine. They already have all the tools in place to make this sort of clean break with the past, and it's high time they do so.

I'll never get it either

[Sycraft-fu]

That is the most common reason I hear for not using 64-bit: "My printer and/or scanner don't work with it." My response is "Then throw that piece of shit out and get a new one." However for some reason, that never seems to be an option considered. They want to keep their 10 year old scanner with their brand new $1000 desktop, even though $50 would get a better, newer scanner.

Here at work we do both. I work for a university so we have to be flexible. I push 64-bit as the default but I'll give them 32-bit if they bitch or if there is a compatibility issue.

Re:I'll never get it either

[King_TJ]

I dunno.... I do understand it from the users' perspective. Sometimes, a printer or a scanner is clearly not a piece of shit, worthy of just throwing away, simply because nobody could be bothered to code updated drivers for it. EG. We have a couple of HP DesignJet plotters where I work that use the 42" roll paper. They probably cost well over $12,000 each when they were new, and even today, I see them selling for over $1,000 on eBay. HP never wrote drivers for them for any OS newer than Windows XP (although I understand a 3rd. party now sells a driver for Vista for them, for $150 or so a copy).

The idea that you'd just throw one of these away as part of a Windows 7 upgrade is ludicrous though. Most rational people would say "Screw that! I'm better off keeping a Windows XP box around just to run it on!" You have to figure, most places using these things also have an investment in ink cartridges and supplies to consider, and they're big and bulky. Shipping alone for a new replacement is not going to be cheap.

How many users will ever actually CARE that their OS is 64-bits vs. 32-bits? It doesn't matter for 95% of what people do every day with their PC, but forcing them to get rid of a well-liked piece of hardware they DO use regularly is going to bother them.

Re:I'll never get it either

[PitaBred]

I have a very nice flatbed scanner that does 1200DPI. It isn't supported by Windows after XP, but it works fine in every single version of Linux I have tried it with.

Why would I want to replace perfectly working hardware? Do you value your money so little? Or do you just hate the thought of not being a "consumer", not buying the latest and greatest when it comes out?

There is no reason, except for the proprietary locked-down model of Windows, for that hardware to not keep working.

Re:I'll never get it either

[Anonymous Coward]

we have 5 different professional dye-sub photo printers from 3 different manufacturers bought over the last 2 years.
one of em has drivers for 2k/xp, xp64, 2k3 x86/x64, vista 32/64 bit, OSX.
one of em has drivers for xp, vista 32 bit, OSX.
The other 3 are 32 bit xp only...
Oh, and the cheapest one is >$2k list price.

Re:I'll never get it either

[IMightB]

I have a very old SCSI scanner. The generic half implemented PCI SCSI card that came with it doesn't work under linux, can't find the drivers for it for Windows >=XP. The SCSI card that I finally did have working for it gave up the ghost a few years back. My solution, I got a $100 Brother MFP40 scanner and laser printer/copier in one. Works great. Another bennie is that I'm no longer on the Lexmark inkjet "Hey you've only used it 3 times, but it's been 6 months your cartridge is expired buy a new one." treadmill.

Re:I'll never get it either

[Anonymous Coward]

That is the most common reason I hear for not using 64-bit: "My printer and/or scanner don't work with it." My response is "Then throw that piece of shit out and get a new one." However for some reason, that never seems to be an option considered. They want to keep their 10 year old scanner with their brand new $1000 desktop, even though $50 would get a better, newer scanner.

Well, my office has a number of $50,000 printers. We aren't throwing them out. Although, Canon & Ricoh are pretty good on x64 drivers for their big printers.

Ok

[SnarfQuest]

MS wants you to run 64 bit OS, but says you should only run 32 bit apps because the 64 bit ones are broken. In other words, they want you to spend $$$ on an OS, you can spend $$$ on software that was written for the OS you used to have.

Re:Ok

[mstefan]

It's not that 64-bit Office is "broken", it's that most software vendors haven't gotten around to providing 64-bit versions of their ActiveX/COM components. So if your business depends on one (i.e.: an add-in for Word, Outlook, etc. or for VBA scripting) then you have to stick with the 32-bit version. A 64-bit process cannot load and execute 32-bit code in its address space. It's possible for 64-bit code to use a 32-bit out-of-process COM server, but most of them aren't written that way (the vast majority are written as in-process servers).

Re:Ok

[Anonymous Coward]

The x86-64 CPU allows a 64-bit OS to load and execute 32-bit processes just fine, and Windows uses this functionality to permit seamless backwards compatibility. However, the binary code in that process either has to be entirely 32-bit or entirely 64-bit.

In the case of Office 2010 there are a lot of separate plugins which have been compiled to run on 32-bit platforms. If you have installed the 32-bit version of Office 2010 then you would have no problem as 32-bit process can load and work with a 32-bit library just fine. But if you installed the 64-bit version of Office 2010 then it cannot load that library. The binary code is simply not compatible.

This is also why Microsoft ships both 32-bit and 64-bit versions of Internet Explorer with the 64-bit editions of Windows, and the 32-bit version is used by default. There are a lot of plugins which have not been recompiled and tested to run as 64-bit.

This limitation doesn't just affect Windows, it is the nature of x86-64. Also of note is that if the OS is running in 64-bit mode that the real mode necessary for running 16-bit apps also isn't available which is why that functionality is gone in 64-bit Windows.

Re:Ok

[Bigjeff5]

but says you should only run 32 bit apps because the 64 bit ones are broken.

What? When did they ever say that? The apps have always been fine, it's the drivers that have been broken, and they fixed that with Win 7 for any moderately popular devices. The only other issue is IE and 32 bit activex controls (which cannot run in a 64 bit process). In other words the only reason you'd be "stuck" with 32 bit is because various programs it may run are only 32 bit. 64 bit apps run perfectly, which should be a no brainer.

Someone also doesn't seem to understand the difference between the OS and applications, and really the whole bits thing in general.

The reason you run a 64 bit OS on 64 bit hardware is a 32 bit OS is incapable of running 64 bit applications - which means you can't take full advantage of the hardware you just spent all that money on. A 32 bit application runs equally well on both 32 bit machine and a 64 bit machine. It can not, however, run on a 16 bit machine. This is because of addressing: the lower addressing bits are all identical. The first 16 bits in 64 bit machine are addressed exactly the same as the first 16 bits on a 32 bit machine, which is the exact same as the entire addressing space of a 16 bit machine.

In other words, the only reason even a 16 bit application should not work on a 64 bit machine is because of various other OS changes, not any technical reason. That's why windows has always had a "compatibility mode" that interprets the old api calls and such.

So the reason you want to use a 64 bit OS is because it allows you to use your 64 bit hardware to the fullest, and because all the software works with 64 bit, but only 32 bit apps will work. There are a handful of situations where you might have to use a 32 bit version instead of a 64 bit version because various support programs are 32 bit only for now, but those are edge cases.

Yes

[Sycraft-fu]

It gets circular like that because developers are lazy fucks. So MS includes 32-bit IE, 32-bit Windows Media Player and so on as compatibility measures. Your old plugins will keep working in your new OS. They also have 64-bit versions so that when they upgrade you can use that, but they maintain old versions for compatibility. So, the lazy fucks at Adobe say "Well 32-bit is there, so we'll just keep that since that's what people use." People then say "Well Flash isn't out for 64-bit so I'll keep using 32-bit." You do get a circular situation. You can't blame the users, they use what is available, you need to blame the lazy shit devs.

So

[uassholes]

what?

Typical Microsoft

[phillymjs]

While Microsoft is pushing 64-bit as a way to gain performance in the OS...

Why bother trying to remove bloat from your OS when you can just tell your customers to throw more hardware at the problem?

Re:Typical Microsoft

[ISoldat53]

That's why Michael Dell loves Bill Gates.

Besides, 32-bit SHOULD be able to do 4GB

[davidwr]

32-bit versions of Windows Server products have been able to exceed 4GB for over a decade.

The decision to limit 32-bit non-server OSes to 4GB was likely more of a marketing decision than a technical one. To be fair, implementing PAE on a consumer OS would increase the development and support costs, but I still think it's largely a marketing decision.

The 64-bit versions of Vista and later gave Microsoft an opportunity to kill off some little-used parts of Windows and it also allowed them to force device driver signing. The former saves them money directly and the latter makes Windows more secure.

By artificially giving a compelling reason to install a 64-bit OS - namely, telling people they must have 64-bits or their computer will have a fairly low maximum RAM capacity, it encourages migration.

Re:Besides, 32-bit SHOULD be able to do 4GB

[drsmithy]

The decision to limit 32-bit non-server OSes to 4GB was likely more of a marketing decision than a technical one. To be fair, implementing PAE on a consumer OS would increase the development and support costs, but I still think it's largely a marketing decision.

It's technical, as evidenced by it actually being enabled in the original XP release (and earlier versions of Windows NT).

PAE support was removed due largely to numerous problems with poorly written third party drivers. Apps needing to be modified to really take advantage of it and the performance hit were also factors, along with it being really beneficial to only a miniscule proportion of users (especially since XP64 and Vista/7 x86-64).

Re:Besides, 32-bit SHOULD be able to do 4GB

[davidwr]

PAE support was removed due largely to numerous problems with poorly written third party drivers.

I doubt very much those poorly-written 32-bit drivers work in 64-bit Windows. Try again.

Apps needing to be modified to really take advantage of it and the performance hit were also factors

OK, the performance hit is a good reason, as long as it's a worse hit than running 32 bit apps in a 64-bit OS.

along with it being really beneficial to only a miniscule proportion of users (especially since XP64 and Vista/7 x86-64).

This may have been true in the era of sub-3.5GB machines, but most new machines today would benefit from it, either immediately or as soon as the user upgraded the RAM. So I'm going to call this a "sorry, try again."

---

So from all outward appearances it boils down to: Microsoft didn't want to bother with having to support giving its users two ways past the 4GB barrier and they wanted to dump other legacy code, so they took the choice away from the customer and forced anyone who needs > 4GB on a non-server OS to "move up" whether they wanted to or not.

Re:Besides, 32-bit SHOULD be able to do 4GB

[dbIII]

No, it's political - it's because Microsoft lost control over their own driver model.
It's quite depressing that they didn't take that back with Vista and still didn't have proper support for the Pentium Pro and everything after.

Re:Besides, 32-bit SHOULD be able to do 4GB

[drsmithy]

I doubt very much those poorly-written 32-bit drivers work in 64-bit Windows. Try again.

How is that relevant ?

OK, the performance hit is a good reason, as long as it's a worse hit than running 32 bit apps in a 64-bit OS.

As far as I know, 32-bit apps typically run faster in a 64-bit OS.

This may have been true in the era of sub-3.5GB machines, but most new machines today would benefit from it, either immediately or as soon as the user upgraded the RAM.

Which would be something of an argument if 64-bit versions of Windows hadn't been available for 5+ years.

So from all outward appearances it boils down to: Microsoft didn't want to bother with having to support giving its users two ways past the 4GB barrier and they wanted to dump other legacy code, so they took the choice away from the customer and forced anyone who needs > 4GB on a non-server OS to "move up" whether they wanted to or not.

More accurately, Microsoft knew the vast majority of support issues would stem from third-party code beyond their control, and the benefits of PAE apply to only a vanishingly small percentage of end users (users with 4GB+ RAM, who couldn't use a 64-bit OS and had applications modified to use PAE). Ergo, not wasting effort maintaining the PAE version is simple common sense.

PAE is a hack, nothing more. If you have 4GB+ of RAM, employ the proper solution and use a 64-bit OS. You're going to have less problems than you would a) finding drivers that don't crap themselves on systems using PAE and b) finding applications that have actually been written to take advantage of PAE.

Re:Besides, 32-bit SHOULD be able to do 4GB

[drsmithy]

No, it's political - it's because Microsoft lost control over their own driver model.
It's quite depressing that they didn't take that back with Vista and still didn't have proper support for the Pentium Pro and everything after.

What ?

Re:Besides, 32-bit SHOULD be able to do 4GB

[dbIII]

OK - to state it more clearly Microsoft was unable to prevent third party developers from writing their drivers to use memory in such a way that PAE would break.
When the driver model was changed for Vista the problem was not resolved by technical restrictions, which is a pity.
If there are no technical solutions it becomes a matter to sort out by discussion between the different groups about who is allowed to do what - hence political in term of the dictionary definition.
Linux and everything else solved the problem by more strictly controlling what third party developers could do with hardware drivers.

Re:Besides, 32-bit SHOULD be able to do 4GB

[drsmithy]

OK - to state it more clearly Microsoft was unable to prevent third party developers from writing their drivers to use memory in such a way that PAE would break.

Nor is there really a way they can. Any more than they can prevent developers from writing applications that needlessly require elevated privileges. Bad drivers break in PAE mode because it's *not* transparent - they need to be written and tested to work with PAE, and they weren't.

Linux and everything else solved the problem by more strictly controlling what third party developers could do with hardware drivers.

No they didn't. "Linux" _ignored_ the problem by not giving a damn when third party stuff breaks (and there are plenty of things that do, don't kid yourself otherwise). FreeBSD "solved" the problem by being so uncommon as to have nearly no third party drivers anyway (and therefore the ones that do exist need to be at least functional). OS X solved the problem with a combination of the "ignore it" and "uncommon" approaches (although the latter is equally due to a high bar of entry rather than pure obscurity).

The only reason Windows is even considered broken in this context is because people have much higher expectations. The are confident that whatever random, shitty piece of hardware they dig up, it'll work with Windows all the time - and hence are bitterly disappointed when it doesn't. The same cannot be said of the users of any other platform.

Re:Besides, 32-bit SHOULD be able to do 4GB

[dbIII]

The only reason Windows is even considered broken in this context is because people have much higher expectations.

For instance the Pentium Pro and later working!
That's really the entire reason I wrote off the 32 bit Windows home computer range as rubbish unsuitable for workstation tasks and had to rely on the "server" stuff. Long live 64 bit and the death of those crappy uncontrolled drivers.

Re:Besides, 32-bit SHOULD be able to do 4GB

[drsmithy]

For instance the Pentium Pro and later working!

They work fine.

That's really the entire reason I wrote off the 32 bit Windows home computer range as rubbish unsuitable for workstation tasks and had to rely on the "server" stuff. Long live 64 bit and the death of those crappy uncontrolled drivers.

What "workstation" applications were you using that required Windows XP 32-bit, and were actually written to take advantage of PAE ?

Not to point out the obvious but....

[umask077]

The article says Windows 32 bit can handle 4 gigs of ram. The machines can take that much but only about 3.5 gigs are addressable. That being said I can see that the number are likely true on usage. 3 or 5 of the machines in my house are 64 bit.

Counting bits

[tepples]

First I had the 8 bit Commodore 64/128 machines, and then I jumped to 32 bit with the Commodore Amiga in 1987.

The original Amiga computer has the same CPU as the Sega Genesis game console, yet Genesis is considered a 16-bit system. Both have a 16-bit data bus and a 24-bit address bus. So what makes the Amiga 32-bit and the Genesis 16-bit?

Re:Counting bits

[treeves]

The marketing department?

Re:Counting bits

[dotgain]

I don't remember the (68000 based) Amiga ever being marketed as a 32-bit platform, nor was this touted by the numerous fanboys I knew (and was among) - perhaps the confusion came from later models (68020+) which were indeed 32-bits on both instruction and address buses.

Re:Counting bits

[operagost]

The 68000 had 32-bit registers and could compute in 32 bits.

Re:Counting bits

[Bigjeff5]

The Motorola 68000 was a 16/32 bit processor, meaning it had a 16 bit data bus but 32 bit registers and was capable of 32 bit arithmetic. This means if you sent your 32 bit instructions in two 16 bit chunks, you could run at 32 bits. The Amiga did this, and I can only assume the Genesis did not, since it was clearly labeled 16 bit and I would assume they'd have preferred to push 32 bit if they were writing 32 bit games.

Either way your memory was limited by the 16 bit data bus - you could process at 32 bits but you couldn't use anywhere near 4gb of RAM.

Having a data bus at half the size of the general purpose registers is not really that uncommon. The Intel 8088 was a 16 bit processor with an 8 bit bus, same thing there.

Re:Counting bits

[commodore64_love]

>>>So what makes the Amiga 32-bit and the Genesis 16-bit?

Sega's marketers were stupid? (shrug). The 68000 inside the earliest Amigas, Macs, and Genesis/Megadrives was a 32 bit CPU sitting on a 16 bit external bus. In that respect it's no different than the 80386's budget version which was also a 32 bit CPU sitting on a 16 bit die.

You'll note that Amigas and Macs never had to go through a 16-to-32 bit transition like IBM PCs did. They were already running 32 bit software from the start.

Problems with 'unsigned drivers' (libusb-win32)

[theMAGE]

I installed Windows 7 on my three year-old Thinkpad and it worked fine, until I had to install drivers for some microcontroller development boards that I'm playing with. The problem was that they used drivers that ran on top of libusb-win32 and as that is a free software project, it does not have signed drivers. With Windows XP through Vista, when the device manager complained that you were installing unsigned drivers, you could push the 'trust me, I'm an engineer' button. That button no longer works for Windows 7 64 bit. You can no longer install whatever you want on your own hardware using a legit copy of the operating system. The brave new future is here, on your PC.

I spent two days trying to get it to work, then gave up and went back to XP/32 bit.

Re:Problems with 'unsigned drivers' (libusb-win32)

[RyuuzakiTetsuya]

I also can't try to install OS/2 warp drivers either and expect them to work.

Re:Problems with 'unsigned drivers' (libusb-win32)

[ledow]

You've tried to be funny but you've missed the point. The drivers are not OLD, they just MUST be signed. That destroys home-brew driver projects. Maybe you don't use them but I, and the parent poster, do. That wipes out whole hoards of PIC programmers, A/D convertors, USB interfaces, etc. for not good reason. Why can't there just be a simple registry tweak that says "I know what I'm doing, just install this unsigned driver?"

All that happens is that you drive those in the know either back to XP or, more likely, onto Linux. I had the same problem recently with the K8055's that I use for little electronics projects. The hardware is still sold in shops (Maplin Electronics, Rapid Electronics, etc.) but the software still works, and works better, on a modern Linux machine. Things like the Arduino are going strong at the moment - removing unsigned drivers just makes people move to other OS's that allow them, and it stops those devices being portable to the consumer without investment in a signed driver.

Remember this next time you wonder why "cool USB gadget X" doesn't come with Windows 7 drivers, or why you have to pay £50 to do something quite simple with custom hardware, e.g. LIRC, etc.

Re:Problems with 'unsigned drivers' (libusb-win32)

[BigDish]

You do know you can self-sign drivers and import the cert.

Re:Problems with 'unsigned drivers' (libusb-win32)

[BatGnat]

I thought you were funny...

Bill Gates once said OS/2 was the future of operating systems...But that was before he stabbed it in the back....

Re:Problems with 'unsigned drivers' (libusb-win32)

[drsmithy]

I installed Windows 7 on my three year-old Thinkpad and it worked fine, until I had to install drivers for some microcontroller development boards that I'm playing with. The problem was that they used drivers that ran on top of libusb-win32 and as that is a free software project, it does not have signed drivers. With Windows XP through Vista, when the device manager complained that you were installing unsigned drivers, you could push the 'trust me, I'm an engineer' button. That button no longer works for Windows 7 64 bit.

30 seconds on Google will show you multiple methods for installing unsigned drivers in Vista and Windows 7.

You can no longer install whatever you want on your own hardware using a legit copy of the operating system.

Of course you can. Stop spreading FUD.

Re:Problems with 'unsigned drivers' (libusb-win32)

[Bigjeff5]

http://www.techspot.com/vb/topic127187.html

It allows you to sign the drivers yourself without having to turn off UAC or anything of that nature (you do have to disable UAC to sign the drivers, but you can re-enable it after with no ill effects).

There is always a solution to the problem, especially in Windows. Microsoft just wants to keep this stuff as far away from the average user as they can, because mucking around with this stuff can easily reduce the security checks in the OS.

You can actually switch Windows 7 driver signing behavior back to XP style as well (just a warning box instead of automatically removing the driver), but signing drivers you trust via the above method is the more secure way to do it.

Note that this also solves problems with signed drivers that have a bad certificate for one reason or another.

You can get 4GB+ if you really want it

[greenreaper]

There are reasons - nefarious or otherwise depending on your interpretation - but at the end of the day you can do it if you're willing to patch and install a self-signed kernel. I did it myself for Vista SP2, and it works fine. It may not work fine for you. I figured it's well worth it to get 4GB vs. 3GB without the hassle of upgrading.

Sloppy editing, Taco

[BBTaeKwonDo]

First, there's no point in linking twice to the same page. The second link should be to http://www.pcpro.co.uk/news/enterprise/356173/microsoft-warns-users-off-64-bit-office-2010.

Second, this bit:

While Microsoft is pushing 64-bit as a way to gain performance in the OS, it earlier this year advised users to install the 32-bit version of Office 2010, 'because currently many common add-ins for Office will not function in the 64-bit edition.'"

may be technically true, but is irrelevant to whether people should use 32-bit or 64-bit. 32-bit programs such as Office run just fine in 64-bit Windows. Running a 64-bit OS is going to improve performance (in most situations) if your other applications use a lot of memory.

64 bit? get it right first!

[Tom]

I'm running W7 64bit, and it is one of those decisions I regret. I wish I had installed the 32bit version.

Lots of programs don't work, or don't work properly.
Sometimes, the installer doesn't work, even though the program itself does.
Same with drivers.

What a piece of crap. We've been having 64bit in workstations for almost 20 years and in PCs for 7 years now, and Microsoft still can't get it halfway right?

Before you paid MS shills come whining that it's all the fault of the application and driver developers, think long and hard about this: How come that Apple could transition to a new CPU architecture and a new operating system and make the move from 32 to 64 bit all in one step with a lot less problems?

If you install a new W7 computer today, choose 32 bit. As with all things Microsoft, wait until the 3rd release.

Re:64 bit? get it right first!

[Anonymous Coward]

Yes, shills. Everyone who disagrees with your uninformed bullshit is getting paid directly by Microsoft. What a convenient worldview.

Lots of programs do work, significantly more than don't. Through functionality specifically built into the x86-64 CPU instruction set 64-bit Windows supports running 32-bit processes seamlessly. There are two classes of applications which do not work:

1. Those that require some form of driver loaded into the kernel for which there are no 64-bit drivers provided or available. The kernel mode portion of a 64-bit OS running on the x86-64 CPU must be entirely 64-bit.
2. 16-bit applications. When running in 64-bit mode the x86-64 no longer can execute any programs in 16-bit real mode. Microsoft's DOS/Win16 emulator relies on this functionality. This also affects installers for older applications as many of them use 16-bit boot-strappers, even if the core installer and the application itself is 32-bit.

In the case of drivers, the problem is the same as with applications that require drivers. You need a 64-bit version of the driver in order to use it with a 64-bit version of Windows.

These restrictions are a part of the chipset, not the OS. No OS vendor can magically circumvent these issues.

In the case of Apple, you have significantly fewer hardware components and software applications that would require support. Still, if you have an older piece of hardware for which there is no 64-bit driver it will not work on a new Mac, period. To claim that Apple changes platforms without breaking a lot of applications is just stupid. Snow Leopard dropped support for running PowerPC applications meaning that no application written prior to the Intel transition will work anymore. I guess the big difference is that Apple doesn't unintentionally break some applications, they intentionally break entire generations of applications.

Re:64 bit? get it right first!

[Bruiser80]

My experience with installers/programs that don't run in Win7-64:

"Did this program install/run correctly?
Windows detected an unexpected crashing of your program. Would you like to run the program again in compatibility mode?"

Lots of problems solved that way.
If that doesn't work, there's always my dual-boot of WindowsXP I have on my machine, though I haven't had to touch it for months.

Re:64 bit? get it right first!

[Spad]

The only app or hardware that I've had any issues at all with in Windows 7 x64 was the shitty Intel webcam I had; it was 6 or 7 years old and barely had functional 32-bit drivers, so I consider it a worthwhile sacrifice in exchange for better performance and increased memory addressing.

Slightly different matter at work, but in most cases it's been due to unnecessary bittiness checking by the installers rather than actual inability to run and the others have been shoddy applications with custom-written libraries that break horribly if the moon is in the wrong phase, let alone if you're running 64-bit.

Re:64 bit? get it right first!

[drsmithy]

What a piece of crap. We've been having 64bit in workstations for almost 20 years and in PCs for 7 years now, and Microsoft still can't get it halfway right?

They can. Developers can't. Pretty much all installer problems, for example, are because they're still using a 16-bit installer - something that's been deprecated for about 15 freakin' years.

How come that Apple could transition to a new CPU architecture and a new operating system and make the move from 32 to 64 bit all in one step with a lot less problems?

Firstly, there's just as many problems on the Mac side of the fence, the difference is users are much more accustomed to being left behind by Apple's carefree attitude to legacy support and have lower expectations. It's only since 10.4 OS X has become stable enough API- and ABI-wise that you can be reasonably assured your software will even work between one version and the next (many applications come in 10.3 and 10.4+ releases for this reason, drivers are even worse), and the Classic environment doesn't run on any Intel Mac (so nothing sold in the last 5 years).

Secondly, because they still haven't actually transitioned (most Macs - even 64-bit capable ones - don't boot a 64-bit kernel by default, and OSX still isn't completely 64-bit).

Re:64 bit? get it right first!

[Anonymous Coward]

Snow Leopard dropped support for running PowerPC applications meaning that no application written prior to the Intel transition will work anymore.

No, Snow Leopard did not drop this. It is true that by default, the Rosetta emulator which allows PPC apps to run on Intel CPUs is not installed. However, you can still select it as a custom install option. Furthermore, even if you failed to do that when you installed the OS, if you attempt to run a PPC app without Rosetta present, Snow Leopard will helpfully offer to download and install Rosetta for you (so you don't even have to fish out your install media; it's a small download and goes fast).

Re:64 bit? get it right first!

[Tom]

They can. Developers can't. Pretty much all installer problems, for example, are because they're still using a 16-bit installer - something that's been deprecated for about 15 freakin' years.

I hate to burst your bubble, but this is not some backwater indian installer thing, it's .msi at least in one example.

Re:64 bit? get it right first!

[drsmithy]

I hate to burst your bubble, but this is not some backwater indian installer thing, it's .msi at least in one example.

What applications ?

Re:64 bit? get it right first!

[Tom]

I noticed this particular insanity on several games, all of them major titles, one released in 2010 - Aliens vs. Predator. And no, running as administrator, compatability mode, all fails.

Installing the game to an external drive on an XP machine and then running it on the W7 machine worked just fine. Definitely the installer.

4 GB of RAM?

[Anonymous Coward]

If the computer has less than 4 GB of RAM or can't/won't support 4 GB+ then I install 32-bit. If the computer has more than 4 GB of RAM or can/will then I install 64-bit. For me that means my work and personal computer have Windows 7 64-bit, but my wife's laptop that has 2 GB of RAM and can only support 4 GB has Windows 7 32-bit. Honestly, both OS builds work just fine. I've got a 2/3 Windows 7 64-bit deployment.

Lazy fucks: Who is lazier? Those that talk, or do?

[Anonymous Coward]

"developers are lazy fucks." - by Sycraft-fu (314770) on Monday July 12, @01:56PM (#32876616)

I would say the "lazy fucks" out there are the Lusers like you (You're the kind that complains endlessly, and yet you are also the type that depends on devs to build them things that "your kind" only merely uses, Luser). If you don't like that, then learn to build this stuff yourself and find out how "lazy" developers are, luser.

Doesn't matter for applications though

[SuperKendall]

To be fair, OS X hasn't really fully transitioned people to 64-bits. Yes, the OS supports it, but the 64-bit kernel isn't even enabled by default.

But that doesn't matter for applications, 64-bit applications run fine regardless of the kernel being started in 64-bit mode or not. So lots of OS X users are using a lot of 64-bit applications already.

SQL Server on W7, Vista fails with XML

[Anonymous Coward]

All I can say is GOOD LUCK 64!
We just ran across an issue using XML in SQL Server on Windows 7, Vista (64 & 32) where results in either a column typed as XML or a variable typed as XML are just plain missing. Under XP or W2003 (32) it's fine. The problem is a size/datalength issue. I'm not sure what the precise problem is, but its where the datalength of the XML is greater than 68653...no results.

Makes me very suspect! Way to go M$ thanks for a wasted couple of weeks trouble shooting your incompetence !!!!!

Why? Because it's 64-bit Windows v2.0!

[BcNexus]

Why is 64 bit running neck-to-neck for most popular Windows 7 edition? Because the Alpha and Beta versions were XP-64, and version 1.0 was Vista. Now, OEM PC makers, developers, peripheral makers, they're all buying in because it's 64-bit Windows version 2.0! No one wanted to buy in to a buggy alpha or 1.0 release.

Oh, and because MS required hardware partners to develop 64-bit drivers in order to earn the right to be called "Windows 7" compatible. That's a good way to "convince" vendors buy into this whole "64-bit thing".

No I'm not bitter about waiting seven years for MS to release a usable 64-bit PC/desktop version of Windows since buying an Athlon64. Why do you ask?

a showstopper

[idji]

Visual Studio doesn't even get told by Windows64 about some exceptions that happen - making it even impossible to debug 32 bit software on Win64. link. Obviously the Win64 people and the VisStudio people don't talk to each other

Atari ST

[tepples]

I don't remember the (68000 based) Amiga ever being marketed as a 32-bit platform

Before the PC gained VGA graphics, the Amiga's big rival was Atari ST (Sixteen/Thirty-two).

Re:Atari ST

[commodore64_love]

"Bitness" wasn't a big deal in the 80s (before the Super Nintendo v. Sega Genesis wars). I didn't even know what "ST" meant until I read it on wikipedia, because Atari never advertised it publicly. It simply didn't exist as a marketing bullet prior to the 90s.

In any case all those 68000 machines ran 32 bit software. The IBM PC had to go through a 16-to-32 bit transition* but the Macs, Amigas, and STs did not. They had been running 32 bit since the beginning of their lives (1984-85). A 32-bit Quadra Mac could run code from the original 68000 Mac with virtually no problems.

*
* One example of this transition is that 16-bit code won't multitask. Windows 95/98 used to put it into a separate DOS window all by itself. If that DOS window crashed, it froze the whole system. I don't know how Windows NT handles 16 bit code; maybe Microsoft fixed that flaw.

Re:Atari ST

[tepples]

A 32-bit Quadra Mac could run code from the original 68000 Mac with virtually no problems.

Except for software that depended on private data structures of the Mac's memory manager. Such software wasn't "32-bit clean" because it depended on a 24-bit address bus.

The interesting part

[initialE]

Windows 7 as released through OEMs for business grade computers would typically be the 32-bit version, and down here at least, the business machines come pre-installed as XP SP3, with the end user having to install Windows 7 manually from the DVD. If the 64-bit version is dominant, then it must mean that most organizations are indeed sticking with XP.

Re:The interesting part

[Anonymous Coward]

Windows 7 as released through OEMs for business grade computers would typically be the 32-bit version, and down here at least, the business machines come pre-installed as XP SP3, with the end user having to install Windows 7 manually from the DVD. If the 64-bit version is dominant, then it must mean that most organizations are indeed sticking with XP.

This is rubbish. OEM 32 bit and 64 bit versions are equally available, and they cost the same. Whoever told you that only the 32 bit version is available is lying to you.

I have installed the 64 bit OEM version onto brand new machines (freshly assembled, with unformatted hard drives), and it installed smoothly and easily.

ImageMagick vs Windows 64bit

[Jason Levine]

My wife's new laptop is 64 bit Windows 7. While setting it up, I installed ImageMagick and a script that I wrote which uses ImageMagick's COM component to resize and watermark our photos. The script wouldn't run. The problem turned out to be that ImageMagick couldn't register the COM component on 64 bit Windows 7. There's no workaround that I can find so I'm forced to rewrite my script so that it doesn't use the COM component (which is much better for scripting) and instead opens a DOS box and runs a custom-formed batch file (much less elegant).

Of course, this isn't a Windows problem per se. It's likely just that something needs to be changed/updated in ImageMagick. Unfortunately, my coding skills tend more towards the Server-Side Database and HTML Generation and less towards the Image Manipulation Utility so I doubt it is anything I could fix myself. If anyone has any ideas to solve this dilemma, though, I'd be happy to hear them.

Guess what?

[Sycraft-fu]

Consumers do value calculations too. Notice how I speak about that soundcard in the past tense. I don't own it anymore, nor do I own any of their newer products. Good thing too, they still don't have 64-bit drivers out. I switched to another company, one that knows how the hell to keep their hardware working on the OS that I use currently.

32-bit apps run fine on PAE OS without changes

[Anonymous Coward]

So you end up coding in what is known in Windows as "Addressable Window Extensions" and they are a pain in the arse.

But what if I just want to run more than one 2 or 3GB program?

Non-PAE programs run fine on a PAE OS, so no application changes are required. Even NT supported this.

because many PC makers ship 64-bit W7

[datadefender]

Many PC vendors deliver 64-bit Win-7 with their gear - thats how you arrive at 50%. For the normal enduser there is no advantage in 64-bit. In the contrary ! The bulk of apps and drivers is still 32-bit. And the "normal" user does just fine with 2GB memory. I fully understand the advantages of 64-bit. For programmers, people working with CAD and other heavy stuff it is the way to go. But the masses use their PC to do email, browse the web, do wordprocessing, look at photos - thats it. A 1.6Ghz CPU with 1GB mem and 32-bit OS will do just fine. Robostness is much more important.

If you're working with photos, 4 GB is much better

[Krischi]

This is much like saying that a normal user needs only 640K of RAM.

Recently I put together a photo album for printing, and between using the design/layout software, the photo editing software to adjust exposure, color balance, and UFRaw for some key pictures, I started hitting the swap space, even though my computer has 4 GB of RAM and runs in 64-bit mode.

Responsiveness definitely would have gone down the drain with 2GB, and none of the things that I did qualify me as a power user.

Hardware NX is the other reason to turn PAE on

[Sits]

Yes, PAE causes a performance hit (mostly when you have to do over 1Gbyte memory allocations) but another reason to have it on is to enable hardware NX support.

But really, if you have a system capable of 64 bit support it is far better to use that (unless you have extremely special circumstances).

Re:If you're working with photos, 4 GB is much bet

[datadefender]

Of course 4GB is much better. But is it required for the average user ?
My wife is a passionate photografer with a high end camera. She uses Photoshop Elemenst. And you know what - she is perfectly happy with her 2.4Ghz dual Core CPU and Windows 7 32-bit.
And again - most users do not even use photoshop or other heavy duty gear. They have a compact camera and use simple tools like Paint.Net or Irfanview - is even that. I see that must do NOTHING to their photos - they just look at them as they are.

Once the average consumer has a functioning PC, he will not replace it until he has to. Even memory upgrades are rare. The bulk of my my neighbours are on P4's with 1GB and XP - and are happy.

It's Preinstalled

[krsmav]

My latest computer came with Win7 64 preinstalled. That way the maker (HP) could tout 6 Gb of RAM. This is undoubtedly the reason for the 50% figure, particularly since people who think about using the 64 bit version will read the tech publications, which warn of problems with 64 bit software.

Quit complaining & learn to code, YOURSELF

[Anonymous Coward]

"It truly amazes me how lazy developers are when it comes to supporting new things. They whine and bitch and drag their feet and blame MS, rather than just admitting they have to learn something new and doing it." - by Sycraft-fu (314770) on Monday July 12, @02:01PM (#32876676)

Speak for yourself: You're the one "whining & bitching" blaming coders for being lazy, when it's YOU who ought to admit you need to learn to do something new, YOURSELF (which is of course, to write code, yourself then!).

"WDM driver that was buggy as shit. The proceeded to work on the bugs and eventually had a nice driver" - by Sycraft-fu (314770) on Monday July 12, @02:01PM (#32876676)

That's how it GOES sometimes man... I must ask: Do YOU even write code yourself? Have you ever done a port of say, an app from one OS platform (& I don't mean JAVA) to another using say, C/C++, OR more specifically, from one memory address range CPU architecture to another??

Hey - Want a job done RIGHT? You learn to do it, yourself... then, you'll find out how "simple & easy it is" writing code (sometimes it is, sometimes it's not - this usually depends on your knowledgebase, experience, & familiarity with the type of project at hand as well as its data, etc./et al + ports are NOT always that "simple to do" either, even if its MOSTLY just changing datatypes on variables & return types etc. when moving from diff. memory addressability ranges in code).

"Many, perhaps most, developers are extremely, EXTREMELY lazy at updating to new technologies and fixing up their code. They want to keep doing shit the same way they always have, no matter how outdated that is." - by Sycraft-fu (314770) on Monday July 12, @02:01PM (#32876676)

Fact is, this is the 2nd time I have seen you state coders are lazy in the past 2 days already, but it would seem to me that you're just another armchair quarterback techie @ best apparently here (otherwise you'd understand, for instance, the why's of why businesses may not invest in 64-bit computing currently, perhaps since 32-bit fits their current needs @ times, or perhaps the cash may just not be there & not everyone has MSDN subscriptions etc.) who gets by on code that's been written by others, but having never done this himself!

You're truly the type that complains about others with skills & whose works YOU GET BY ON in this science, & yet, you do so, without having walked a mile in the pair of shoes of those you bitch about here (myself being one for many years now in a programmer-analyst/software engineer, & sometimes, it just makes me laugh, honestly, hearing stuff like this from folks like you, it truly does).

Before you go calling coders lazy? Learn to code, yourself, & proficiently + in multiple languages/tools, & then get back to us, ok??

They don't "work fine", that's the problem

[dbIII]

In addition to applications written with PAE in mind (geophysical and image editing for my experience) there was also the frequent problem of running multiple applications at once that may not support it individually but you still run out of memory - blatantly obvious running just about anything on 32 bit Vista which has such a high memory floor and a low memory ceiling.
I really don't know why you are arguing about this when it's a well known problem restricted to the hobby range of Microsoft products (which unfortunately ended up in workplaces). They had more than a decade to implement proper support for the Pentium Pro and everything later and didn't bother simply shipping a broken product that ultimately is not capable of doing much with raw images from a mid range digital camera these days without running out of memory.
"It's not so bad" one of the 32 bit XP users that pushes memory hard tells me - "now it's only crashing about once per day".
The sooner applications get ported away from that dead end software platform the better. It's not 1995 any more and computers can go faster than 100MHz AND support more than 4GB.

Re:They don't "work fine", that's the problem

[drsmithy]

In addition to applications written with PAE in mind (geophysical and image editing for my experience) there was also the frequent problem of running multiple applications at once that may not support it individually but you still run out of memory - blatantly obvious running just about anything on 32 bit Vista which has such a high memory floor and a low memory ceiling.

32 bit Vista is irrelevant to the discussion, given the existence of 64-bit versions of Windows since 2005.

I really don't know why you are arguing about this when it's a well known problem restricted to the hobby range of Microsoft products (which unfortunately ended up in workplaces).

My disagreement is people presenting this as some sort of earth shattering problem, limiting what most people can do with their computers, when the complete opposite is true, and the proportion of people genuinely impacted is vanishingly small.

They had more than a decade to implement proper support for the Pentium Pro and everything later [...]

The problem isn't Windows, it's third party hardware and software vendors who can't be relied on to test their code.

[...] and didn't bother simply shipping a broken product that ultimately is not capable of doing much with raw images from a mid range digital camera these days without running out of memory.
"It's not so bad" one of the 32 bit XP users that pushes memory hard tells me - "now it's only crashing about once per day".

Sounds like you need to complain to your application vendor.

The sooner applications get ported away from that dead end software platform the better. It's not 1995 any more and computers can go faster than 100MHz AND support more than 4GB.

You mean like Windows, except for a brief period between about 2003 and 2005 for one particular version ?

Re:They don't "work fine", that's the problem

[dbIII]

32 bit Vista is irrelevant to the discussion, given the existence of 64-bit versions of Windows since 2005.

You may as well say the entire discussion is irrelevant because Microsoft supported PAE in their server range - but that's most of my point really - at the low end due to a variety of factors they didn't.
The third party developers you like to blame wrote their drivers to the model set up by Microsoft. The model won't break PAE for the server versions but it does for the other version.
It's not "earth shattering", it's simple mismanagement producing an annoying problem that makes the platform somewhat useless for an increasing number of tasks while every other 32 bit platform has supported it for about a decade.
I end up ranting about it because there are usually people less informed than yourself that lay the entire blame on the processor.

Re:They don't "work fine", that's the problem

[drsmithy]

The third party developers you like to blame wrote their drivers to the model set up by Microsoft.

NO THEY DIDN'T. That's the whole point. Their drivers are broken because they didn't write and test them correctly, not because they didn't have an environment to write them to.

The model won't break PAE for the server versions but it does for the other version.

Absolutely it will. Software and drivers that break because of PAE in Windows XP will also break on a Windows 2000 or 2003 server using PAE. The problem isn't XP, it's the software.

You really seem to be missing the point here. The original release of Windows XP - along with earlier versions of Windows NT Professional (ie: non-Server versions) - supported PAE. PAE was removed from XP with SP1.

It's not "earth shattering", it's simple mismanagement producing an annoying problem that makes the platform somewhat useless for an increasing number of tasks while every other 32 bit platform has supported it for about a decade.

Windows NT Pro supported PAE until XP SP1 (~2003). Then with the release of XP64 in 2005, it became basically irrelevant because the real solution is a 64-bit platform, not an ugly hack to a 32-bit platform.