With Windows Vista x64 and 6GB of RAM do you need a page file?

I think you're actually calling the programmers at Microsoft incompetent now. If Windows would re-enable the paging file at will, Windows wouldn't go crazy when it runs out of memory. But you say it only does it sometimes? Wouldn't that be incompetent programming?
 
You're putting words in my e-mouth, and really, you should stop. I complimented them on well-designed NT-based OSes with respect to the virtual memory subsystem, and you're reading into it whatever you want.

I said you can't disable the pagefile, and not much else, which is absolutely true. I said Windows will create a pagefile if and when it needs one regardless of whatever setting the User/Admin thinks they've locked into place, and not much else, and that also is absolutely true.

That's about it. If you can't do the research and figure this stuff out yourself, that's not my problem. But I myself, and many others around here - that are presumably deciding not to get into this so-called debate (which is fine, that's their prerogative) - know better. It's not a requirement to prove myself right, I already know the answer(s).

If you're interested, you can find the answers the same way I did over the years. If you don't believe me, that's perfectly acceptable too, but it doesn't make me wrong.
 
Most of what you've said so far in this thread is wrong. Please, give some qualified references to back up what you say.
 
This having been said, I'm not sure if it does re-enable the page file or not. I occasionally see mention of temppf.sys, but the Microsoft KB articles only mention it for NT/2000.

http://support.microsoft.com/kb/101773 has a couple of words about that file.

It is a boot time feature only. If Windows detects that there is not enough ram to run stable, it will create a small temporary paging file, so the user can go in a configure a proper paging file
 
That was probably pretty useful... back when people had 4MB or 16MB installed. :p

I'm on my second day (yesterday was only partial) of running Vista x64 SP1 with 6GB and no page file. I played some games yesterday (Steam and TrackMania) without problems. I purposely filled up memory to see how Windows recovered after an out of memory situation. Even greedy SuperFetch politely released and reclaimed memory as it was used and became available again. I don't feel any speed increases or decreases so far, and boot speed seems the same. Benchmarking might show differences, but any differences are too small to "feel" as I use the system.
 
If you can provide instructions for how to reproduce your results, or provide evidence that it actually happened, then you might have a basis for your argument.

I realize there is zero way to test this out.
I've been unable to get this to work on some systems as well, which tells me there was *something* on that old computer of mine (4 years ago) that caused it.

I'm just not as hell bent on this whole thing as some of you are.


Going off memory it was like a 2.4Ghz machine with 512MB of RAM, ran Windows XP...


If you want some links to Microsoft saying this- look for them yourself. I really don't care one way or the other. I've just seen Windows automatically enable the thing more than once.
 
I checked in with the Microsoft Premier Field Engineer assigned to the company I work for and he says that you can in fact disable the windows page file via the radio button and Windows will not re-enable, nor make temp page files unbeknownst to the user........he also says this whole thread is amusing...

I can imagine. All of us squabbling over things, in reality, we have little knowledge about when the people who know the inner workings of these things are just kind of chuckling over the squabble.

I don't have a way to prove this because I can't tell you the company I work for, and I'm not sure I can even tell you our Premier Agent's name...But I have no reason to lie. I would have just as readily came back here and confirmed the other side if that's what he had told me.
 
I checked in with the Microsoft Premier Field Engineer assigned to my company and he says that you can in fact disable the windows page file via the radio button and Windows will not re-enable, nor make temp page files unbeknownst to the user........he also says this whole thread is amusing...

I checked in with Bill Gates, and he finds this whole thread amusing too!
 
I checked into my thread about this same subject. I wonder how people's e-Penises get so hard over issues like page files? Seriously, if someone's being a dumbass they don't deserve the dignity of reply. Period. I would never turn it off, but with 8GB RAM I have yet to see Vista hit the disk for anything while sitting idle...

Now, the reason I have heard to leave pagefiles enabled is because (and this may be true in the days of smaller quantities of RAM) some programs will make HUGE requests for memory that they never actually use. Therefore, large amounts of memory would be wasted by these greedy programs instead of having over-zealous memory requests paged to disk.

Now, I also have a 24.5GB pagefile spread across two 750GB WD Caviars. Why? I forced Windows to create pagefiles on the non-system volumes, but left the size as "System Managed." I can think of no rationale for such a huge pagefile other than Windows using its formula of "Three times the memory installed" (8GB * 3 = 24GB).

I'm going to shrink the pagefile because honestly like I say despite the huge pagefile, and most likely due to the massive quantity of RAM in my system, there is very little need, even when gaming, to hit the disk.
 
I guess the entire thing, without going into theory and technical talk (most of that has already been done in this thread), and logically speaking from a common sense point, is this:

You have to understand that system RAM is VOLATILE. A hard disk drive is NON-VOLATILE. If you base your entire operations on volatile RAM and don't have some sort of "failsafe" (which in this case is the hard drive) to swap back and forth to, the slightest instability will hard crash your computer, and you'd be totally screwed. If you have data flow that is calling for a specific address at the memory, and then you get some crappy coded driver or program (believe it or not they do exist :eek: ) calling for the same memory address at the same time, you'll crash. Otherwise, the OS would swap one portion back to the hard drive and allow access to the other.

That's my logical take on it. I've dabbled in enabling and disabling page files, sometimes with success, sometimes with disastrous results. The disasters were enough to teach me to never disable the paging file again, needless to say. :p
 
you know if it all matters get some small SSD of about 4 to 8GB move the page file to it and bam done.

/thread

Actually thats a pretty neat idea I didn't think of. I wonder if there's performance benefits to THIS, especially in Vista where Superfetch isn't impeded by lack of page file, and page file speed isn't impede by slower hard drive speed.
 
I guess the entire thing, without going into theory and technical talk (most of that has already been done in this thread), and logically speaking from a common sense point, is this:

You have to understand that system RAM is VOLATILE. A hard disk drive is NON-VOLATILE. If you base your entire operations on volatile RAM and don't have some sort of "failsafe" (which in this case is the hard drive) to swap back and forth to, the slightest instability will hard crash your computer, and you'd be totally screwed. If you have data flow that is calling for a specific address at the memory, and then you get some crappy coded driver or program (believe it or not they do exist :eek: ) calling for the same memory address at the same time, you'll crash. Otherwise, the OS would swap one portion back to the hard drive and allow access to the other.

That's my logical take on it. I've dabbled in enabling and disabling page files, sometimes with success, sometimes with disastrous results. The disasters were enough to teach me to never disable the paging file again, needless to say. :p


RAM is Volatile - true - but then your post derails from there. Your explanation isn't the way a page file works. It isn't a mirror 1:1 copy of what's in RAM. Not in any way. I guarentee with a page file and RAM that is encountering errors your system can and still will crash - page file or not.
 
About the statement that the paging file cannot be disabled. Mark Russinovich says otherwise in his "Windows Internals" book:

"Because Windows can run with no paging file, page file backed sections might in fact be "backed" only by physical memory."
 
RAM is Volatile - true - but then your post derails from there. Your explanation isn't the way a page file works. It isn't a mirror 1:1 copy of what's in RAM. Not in any way. I guarentee with a page file and error that is encountering errors your system can and still will crash - page file or not.

lol. :p The exact response I expected to elicit. That was just a hands off approach to something I already know. Another perspective, if you will.

Of course it doesn't make any sense that the system won't crash if it's encountering errors whether or not you have the page file enabled; you're correct, the page file is NOT a 1:1 copy of the RAM.
 
I'm still running Vista x64 with 6GB and no page file that I started testing last Friday. I'll give an update this Friday on how it went after a week. The results are already surprising.
 
^ AJM786

the page file is NOT a 1:1 copy of the RAM. !!!!!!!!!!!!!!!!!!
 
Actually thats a pretty neat idea I didn't think of. I wonder if there's performance benefits to THIS, especially in Vista where Superfetch isn't impeded by lack of page file, and page file speed isn't impede by slower hard drive speed.
Think Intel Turbo Memory. Even then, it wasn't too hot for systems with > 1 GB RAM.
 
As promised, I am posting my experience of working with Vista and no page file for a whole week. The load was just my usual tasks: browsing the web, downloading junk, transferring and editing pictures, scanning documents, running demos, playing games and programming. This was on my primary system and it gets a lot of use.

The results are pretty uninteresting: I had no problems at all, no errors*, no warnings, no performance problems, no phantom page files (LOL) and no crashes. The system up time was a full week, with sleep and hibernate only used overnight/when I was at work.

* I did force an out of memory situation once at the beginning of the test, but that was on purpose and did not happen in any "real life" testing and use.

While it doesn't answer the OP's question "With Windows Vista x64 and 6GB of RAM do you need a page file?", it is certainly possible to run Vista x64 with 6GB and no page file and have no ill effects. Other users may *need* a page file, and there's really no reason for disabling it anyways. Shrink it if you really need to save disk space, or just move it to another partition. This is the warning MS gives when disabling the PF:

disablevistapfwarningpz1.png


That's nowhere near as dire as the FUD posted by others in this thread.
 
pxc, what happened when you purposely made it run out of memory, I assume it just said 'out of memory' and sat there?
 
The program I wrote to eat up all available memory triggered an exception, stopped allocating memory and I closed it. After it closed, all memory used by it was freed and SuperFetch started reloading data in the background.

I'm sure if I had tried to load any other programs with the memory eater still running, those programs would have given their own errors. Running out of memory is not a good thing, but it doesn't crash the whole system either. That was pretty much the point of my testing.
 
Other users may *need* a page file, and there's really no reason for disabling it anyways.

So... all that to come to the conclusion some folks have been pointing out all along?

I'm sure if I had tried to load any other programs with the memory eater still running, those programs would have given their own errors. Running out of memory is not a good thing, but it doesn't crash the whole system either. That was pretty much the point of my testing.

I guarantee you it would've killed the system... If you had pushed it harder with some more RAM-intensive apps.

That said though, I have no idea how Vista would handle a harder situation. Vista could start killing off non-essential processes to keep the OS running, and give you an error message.
It could also "crash", just write things to the disk, and Hibernate or something, and when you booted back up it would tell you to unload some of the stuff you're using.
I have no idea how it would handle it, but I'm assuming it's like XP and would crash and burn.

What I'm saying is the testing had some pretty big missing holes and pieces in it, and IMO isn't conclusive at all.
 
So... all that to come to the conclusion some folks have been pointing out all along?



I guarantee you it would've killed the system... If you had pushed it harder with some more RAM-intensive apps.

That said though, I have no idea how Vista would handle a harder situation. Vista could start killing off non-essential processes to keep the OS running, and give you an error message.
It could also "crash", just write things to the disk, and Hibernate or something, and when you booted back up it would tell you to unload some of the stuff you're using.
I have no idea how it would handle it, but I'm assuming it's like XP and would crash and burn.

What I'm saying is the testing had some pretty big missing holes and pieces in it, and IMO isn't conclusive at all.

This is all funny, and really does not matter. With or without a pagefile you push a system to it's limits it will stop. The question here is does he NEED a pagefile, and that is no he does not NEED it. However it is suggested not to turn it off as this can cause some unwanted effects, and that is simple fact.
 
Im running without a page file with the laptop in my sig, 2x2GB ultimate x64. Runs just fine, usually around 50% usage with a bunch of shit open.

Ive got 2 tabs open, trillian, outlook, winamp, couple smal programs, and a 720p movie going and process explorer shows 65% commit charge on 71 processes. Why would I need to leave it enabled? Ive gotten ONE memory error, which was supreme commander during a 4 way match, 100 unit limit per side, and a massive battle going. Not too concerned about that.

9 day+ uptime and have loaded enough "56k go away" threads to trigger 2GB memory usage from firefox alone. No issues.

I leave it off to free up the HDD space and so I can recover from sleep mode/losing power in seconds (sleep hibernates too, with a page file it takes a shitload longer to recover after loss of power) With the page file enabled if you drop power just after sleeping it wont have finished writing the hibernation file when power drops, and it dies. No page file its a lot faster. My reason anyway.
 
This is all funny, and really does not matter. With or without a pagefile you push a system to it's limits it will stop. The question here is does he NEED a pagefile, and that is no he does not NEED it. However it is suggested not to turn it off as this can cause some unwanted effects, and that is simple fact.
True...
But I'd say he doesn't NEED it for day to day tasks. There will come a point when he NEEDS it when his system is maxed.
Best way to describe it is a safety net.
That, and you also consume more RAM by not having a pagefile... Of course this could be an invalid point if he isn't using his system for much.

Ive got 2 tabs open, trillian, outlook, winamp, couple smal programs, and a 720p movie going and process explorer shows 65% commit charge on 71 processes.
Off topic but have you tried Digsby? I switched from Trillian and love it. It's lighter, compatible with more stuff (even has Facebook Chat integrated in there now, which is great)... and doesn't lock the whole damn system up when it starts up (5 different machines, 3 different OSes... It locks the machine up on each one).
 
Disabling the paging file is a bit like manually setting it to a maximum.

System 1: 2GB physical RAM + 2GB paging file
System 2: 4GB physical RAM no paging file

They both have a total of 4GB.
 
Disabling the paging file is a bit like manually setting it to a maximum.

System 1: 2GB physical RAM + 2GB paging file
System 2: 4GB physical RAM no paging file

They both have a total of 4GB.
I don't think so. System 1's page file will expand as its limit is nearly reached.
 
It was just to give another view point. Many of those saying windows just cannot run stable without a paging file would be fine with system 1 but not system 2, even though they both have the same amount of memory to play with.
 
If you set it to a maximum it will not grow beyond that.

Uhmmm... where do you folks come up with these concepts? Set your pagefile for 512MB min/max and load up everything you've got, games, movies, applications, etc, if that's not enough to pop the cork on the pagefile, load the stuff again, but I assure you when you max it out, Windows will choke and that's when you'll see the "Windows is running low on virtual memory and will now resize the pagefile..." warning/alert.

Geez... I know people with 8GB boxes that get that warning, quite often actually because they simply don't know how to configure the machines - meaning the fuck around with the settings which work just fucking dandy at their defaults.

I'm going to quote something from another forum, a piece of advice and info that I find is most refreshing even if it was stated years ago, but I think it's very relevant here at this point:

"I don't have a specific problem case, nor mechanism in the OS, that I can point to and say "...and in this situation, lack of an adequate paging file on the boot partition can and has caused problems." But I strongly believe in modifying the system's default configuration as little as possible. Among other benefits, the more standard your configuration, the less your troubleshooting time when things go wrong. Conversely, the more nonstandard your config, the more your ongoing system maintenance tasks will feel, well, Sisyphean.

As I said - leaving an reasonably-sized pagefile on the boot partition will NOT keep the system from using your second one.

I can also tell you that the devs at MS are well aware of the issue of the "1.5x the RAM" rule making for rather large pagefiles on modern systems. They haven't changed it. I have to assume they have a reason for not changing it. And by the way, the reason is NOT to capture debugging info - if you haven't gone into the startup/shutdown settings and changed the memory dump type to "full", that's a non-issue.

I furthermore just can't wrap my head around the mindset that says "the system's default configuration just absolutely MUST be flawed, so I must spend time changing it before getting around to just USING my computer." In particular, "optimizing" pagefile placement on systems with large RAM loads is just not worth the time or trouble. It's like fine-tuning your car to increase the top speed from 138 to 141 mph, when you do all your driving at 70 mph or below, and most of your driving on surface streets at rush hour."


Pure genius.

Taken from a post by DriverGuru at the ArsTechnica forum in a thread started back on June 19, 2003. The actual thread is here:

http://episteme.arstechnica.com/eve/forums/a/tpc/f/99609816/m/6010972175

The guy knows his shit, seriously. He's not alone, either. The overwhelming majority - even of [H]ard people - would do better leaving the pagefile alone, absolutely, completely, totally. Windows and the people that designed it really did/do know what they're doing. There is almost no chance in Hell that a few tweaks here and there will make any appreciable or noticeable difference in day-to-day use of any machine on the planet.

Unless all you do is sit at your computer running tests all day to confirm whatever idiotic ideas you happen to be fronting so your e-peen can get all big and schtuffz, that is.

Read that thread, start to finish. If you don't have a better understanding of how the pagefile is so integral to Windows and how mucking around with it can almost always guarantee a level of performance that's sub-optimal, well... that's your loss.
 
Of course you can max it out, did I say otherwise? I compared it to a system with no paging file, but with enough physical ram to make them even.

I gave that example as I said before: "Many of those saying windows just cannot run stable without a paging file would be fine with system 1 but not system 2, even though they both have the same amount of memory to play with."
 
I guarantee you it would've killed the system... If you had pushed it harder with some more RAM-intensive apps.
Well, I did run RAM intensive apps, games. And it works as it should, politely pushing stuff out of memory and reloading from disk. I really doubt I had 12GB of game data in memory at once.

And what did I just say about FUD? :rolleyes:
 
True...
But I'd say he doesn't NEED it for day to day tasks. There will come a point when he NEEDS it when his system is maxed.
Best way to describe it is a safety net.
That, and you also consume more RAM by not having a pagefile... Of course this could be an invalid point if he isn't using his system for much.

Thats why it is not recommended to turn it off.
 
Hey guys--- I went back to my Microsoft rep and asked again the following clarification:



Another quick question relating to the same topic.

Is there a difference between limiting the page file size and having an
equivalent total of physical RAM?

I.E.

2GB of RAM with a hard-set 2GB page file = 4GB total

versus

4GB of RAM with no page file


His response

"Absolutely.
There is a strange behavior that all memory allocations must have free space available in the page file. So the machine with a 2 gig page files might behave a bit more conservatively than a machine with no page file. It is hard to explain in a email but they will behave differently."


Basically -- note that there is a difference in the way Windows behaves with or without a page file. If I understand his response correctly --- apparently it more aggressively tries to keep free space in a page file than it does in physical memory, therefore you should not run out of memory with a page file as easily. However the Microsoft Premier Field Rep had said in a previous e-mail that Windows does not mind at all if there is no page file, but that the system will hang if it runs out of physical memory.
 
Well here's a technet article that further clears it up for all you people stating page files are absolutely necessary. I'd forgot about this topic, but when I ran across this article today I figured this topic should be put to rest --- without lingering doubts!!! Some of you "know it alls" must admit you were wrong :eek:...Everyonce in a while it happens to all of us. :D This article isn't even about Vista and so most of the disclaimers(if you had them) were I don't know about Vista, but in the past --- other MS operating systems required a page file! --- NOT TRUE

Microsoft says page files given enough system RAM in a newer technet article from June of this year.

http://support.microsoft.com/kb/889654
 
No, it does NOT say that. Read that article. It's just giving instructions for XP and Server 2003. And even if Windows will run without a Pagefile, it's still a horrible idea. HD space is dirt cheap - spare a few gigabytes and leave yourself some breathing room.
 
Back
Top