Task Managers Page file usage

What happens if you set your pagefile size to 1MB.....do those number change?

It doesn't matter if there is a pagefile.sys or not the same thing happens. The Task Manager over states the current usage of the page file. It is only my machine. I have seen this on every XP I have come across.

It is only my machine = It is not only my machine

I believe that a page file is not only used for 'memory overflow', but as a temporary storage medium for unused data. Could someone correct me if wrong?....

I beleive that if data remains redundant for x.00 ms of time, it's then moved to the page file, clearing up the memory.

But I have no pagfile so how can it be in use as stated in the Taskmanager. Right now it says that 271 PF usage but ther is no pagefile.sys.

And MS categorically states that the PF Usage is the pagefile.sys. Uh Huh no way. If one doesn't have a pagefile.sys then how can the PF Usage state that the pagefile.sys is storing data their. By all logic this is virtually impossible.

Okay. I try not to get into engaging conversations, because quite frankly, I have still got a lot to learn. It is my impression that the pagefile.sys is created if one chooses to allocate space on the hard drive without prior knowledge of how much additional "fast" memory one may need to run his applications. Just becasue you select "No Paging File" doesn't mean that the OS (i.e. the kernel) won't use a paging file. It means that you have not pre-allocated space for your "Fast" memory. At least that is how I think of it.

What is paging? It is a action of bringing pages of an executing process into main memory (RAM) and then removing them when they are no longer needed or are replaced by other pages that are sitting on the disk waiting to be called up; right?
What is the space being used on the disk called? Virtual memory. I think once a process starts the pages initially needed are moved into virtual memory, not to the RAM which is physical memory. Where is the virtual memory? On the disk. And what is virtual memory? A place to store page files.

Yeah put the disk size does not grow even thouhg the PF usage does and MS states the PF usage is amount of the pagefile.sys being used. If you have no pagefile then how could it be used.

Well I'm not sure what version of XP you are running, but looking at my performance settings I cannot have a 0 byte pagefile. The minimum size of any pagefile that I can make on any of my hard drives is 2MB. So with that piece of information isn't is possible that if one states "No pageing File" as his option that the OS will still create a 2MB pagefile?

I am using Pro Under the Virtual Memory window there is an option "No paging file" Which is what I have set. There is no pagefile.sys on this system at all. The 2MB thing I know is built into Win2000 but XP has the option not to have a page files.

Okay. Here is a C&P from ahuma.com I think he is saying what I said above. You can be the judge. :)

Can the Virtual Memory be turned off on a really large machine?
Strictly speaking Virtual Memory is always in operation and cannot be “turned off.” What is meant by such wording is “set the system to use no page file space at all.”

Doing this would waste a lot of the RAM. The reason is that when programs ask for an allocation of Virtual memory space, they may ask for a great deal more than they ever actually bring into use — the total may easily run to hundreds of megabytes. These addresses have to be assigned to somewhere by the system. If there is a page file available, the system can assign them to it — if there is not, they have to be assigned to RAM, locking it out from any actual use.
by Alex Nichol
(MS-MVP - Windows Storage Management/File Systems)
© 2002-2004 by Author, All Rights Reserved

http://aumha.org/win5/a/xpvm.php

Nice try Glenn but Alex is not entirely correct. Go back and read my question. Even when I have plenty of RAM available the Task Manage is saying that I have over 244MB's of data being housed in a pagefile that does not exist in any way shape or form.

I have been running without a pagfile for a couple of months and the system is much faster because of it. The lowest available RAM that has ever happened on this machine is 200MB's. If there is no pagefile.sys then where is the 244mb's being housed that the task manager is saying that the pagefile is being used. There is no pagfile so there is no way the 244MB's can be placed in something that does not exist. So where is this 244MB's being housed. No pagefile.sys means the Task manager is reporting a big inaccuracy if you know what I mean.

When I used a pagefile the task manager often reported that the system was using more pagfile then the actual size of the pagefile.sys.

Example: Lets say I have Max and Min of 500MB's pagefile. This means the 500MB's is locked down and cannot grow. However the Taskmanager says 750MB's page file usage. The pagefile.sys is still 500MB's in size so where is the the other 250MB's being house? :)

I have found this in a few places:

"Paged Pool Memory is system (kernel) code that is pageable to disk "
"Non-paged Pool Memory is code that must stay resident in physical memory"


I think this is merely statistics that show you how much could be paged to disk if you have a pagefile set up.

Notice paged pool memory is "Pageable" in that it does not absolutely signify that it has to be paged to a pagefile.

Also further study shows that non-paged stuff is time-critical and therefore cannot be paged. "Non-paged" is clearly faster. And the speed increase you have noticed since removing your pagefile is the result of the amount that could be "paged" that is now being "non-paged"....

Works for me... LOL

This is fun bedtime reading....
http://www.ntcompatible.com/forums/viewtopic.php?t=18868

From what I've experienced, XP will make a pagefile, even if you tell it not to have one.  It may show that one is not there, but as you are seeing, it still has one.  Did you look in your system for pagefile.sys?  It is hidden and a system file so it can be hard to find the file.

Here is some good reading about it.  Basically, you want a pagefile, even if you have lots of ram.

http://www.ultratech-llc.com/KB/ASP/FileView.asp?File=/KB/Pagefile.TXT

Thanks diggisaur I kind of had come to that conclusion but I wasn't sure :).


 Eagle6990 I understand where that article is coming from but again I don't agree with what it says.

I have not used a page file for a couple of months with no problems. I have a few times, for experimental sake set a paging file of 2GB and system was slower then not having a paging file. I never have ran lower than 200MB's of available RAM nor have I gotten low on memory or low on virtual memory errors. According to MS using the "No paging file" option means just that, no paging file is used anywhere on the system at anytime. I have taken the hidden attributes off all my files and there is no pagefile.sys anywhere on this disk. I have used filemon and there is no file being shown that is being used to offload data from the RAM to a disk file. So again I think the Task Manager PF Usage is delusional for some reason. :)

Folks the point here is not about having a pagefile or not it is about how the Task Manager reports PF Usage which it isn't reporting the accuracy of usage of the page file. The numbers are all over the place and many times if you have a page file the PF usage reports a higher number then the actual size of the pagefile on the disk.

Yes I think the statistic of "paged" is moreso how much could be put into a page file if you so chose to.

It was probably a statistic that started back when memory was ultra expensive and IT solution providers needed to know the bare minimum RAM they needed to implement to cover to the non-paged application needs.

As far as I know when you set your pagefile to no pagefile, windows xp taskmanager will display the amout of total commit charge in the PF Usage graph.  The Commit charge is the total amount of virtual and pagefile memory, since you have no page file the commit Limit should be slightly lower than your total physical ram (so the system has some so it won't crash if you run out of physical ram it automaticly takes 50 Mb or so away).  I hope that was the information you were looking for.

Hey Crazy
Long time no see! Things easing up just a bit.. lately.
How's the market for Colorado lobsters and Prairie oysters doing these days?

Anyway, my input on the question. Totally unsubstantiated for the most part of course.

It's always been my opinion, and still is, that most software developers, MS especially, and maybe with good cause, that they write the software with the idea in mind that consumers have no idea what they're doing and therefore we'll give them what we want to give them and the rest we'll just make them believe that we're giving them a choice.
After all, how many people here really think that MS can't stop the popups in IE. If MS did, they would have many of their major customers seriously upset at them, to say the least.

So, with that type thinking, lets look at it this way.
Most computers sold have no way nearly enough ram to handle the needs of the programs installed, let alone the ones that will be installed.
Tech support would be inundated with calls from would be gurus disabling page file to speed their games. So we'll just make think that they're doing so.
Sort of a page out of the politicak scene.

On the other side, the CAD and graphics companies automatically assume that you don't have enough ram and resources really needed, and most likely rightly so, and circumvent your page file settings so that their program will run on your pc.

Picture this, pun intended, opening a file say three time the amount of ram you have installed, with page file off, and it opens with no difficulty!
Voila'

A gig or ram is not really much when you're dealing with CAds and graphics.
Major reasons why Macs do such a good job with them.  Better code!
That ought to ruffle a few feathers!

All very simplified but you get the idea.

Cheers :-)


P S   I guess what I'm saying is that I agree with your diagnose that you're not shutting down the page file. That has always been my  totally unscientific opinion.

Oops
That keyboard at it again!  politicak = political

Well I'll be it is centerv, phew what a shock dude, glad to cyber see you again.

>>>How's the market for Colorado lobsters and Prairie oysters doing these days?
Booming as usual. LOL

In Windows 2000 the lowest you can set the page file is MB's but in XP we see that it says that 2MB's is the min that can be set as well. But wait it also has a "No page file" option.

Now in my virtual memory window it says this

 
Total paging file size for all drives

   Minimum allowed: 2MB
   Recommended: 1533MB
   Currently allocated: 0

In the task manager PF Usage 313MB's

OK there is not a file that is as least as large as 300MB's that remotely represents any thing like a page file. The free space on the disk hasn't been lessened by 300MB's either. File monitoring utilities don't show any files being used that looks like it could be a page file. And at this moment no single file is loaded that comes close to 300MB's. I have nearly 600MB's of unused RAM. Again according to MS the PF Usage graph in the Task Manager is suppose to be indicating actual usage of the page file (pagefile.sys), of course I am not saying that what they say is totally accurate either. So the mystery is still with us. :)

>>>P S   I guess what I'm saying is that I agree with your diagnose that you're not shutting down the page file. That has always been my  totally unscientific opinion.

Actually my curiosity here is not really about having or not having a page file. It is about why does the Task Manager PF usage show more usage then the size of the page file with a page file that is static in size.

BTW when SP2 is released this summer supposedly it adds a popup blocker to IE :)

Check your System32 Directory for the presence of a TEMPPF.SYS -  Maybe your system IS creating a pagefile....
REF https://www.experts-exchange.com/questions/10101763/Virtual-Memory-Problem.html http://support.microsoft.com/default.aspx?scid=kb;EN-US;257758


the 2MB "Limit" is to have space with which to create crash-time minidumps.  If this isn't critical to you, you can bypass this.

Nope no TEMPPF.SYS

And please lets focus in on the question. This has nothing to do with if there is a pagefile or not. The question is about what the task manager is reporting about PF usuage.

The TM will only report what it is designed to read and report.
It most likely does not have the capability to report what you're looking for.
Cheers

Now, understand that I have not read the entire thread...  But here is the interesting part of my story..  While at a MS conference about a month ago, I approached one of MS's engineers out of Redmond that flew in to speak...  I wanted to discuss the Page file with him, and I mentioned setting the Page File at 0..  His response was that even though you set it to '0', it is still existing on the system and will be using the default setting..  I immediately questioned him on this, and he said that it is impossible for XP to boot without a pagefile, so if it is set at 0, then the system takes over and creates one for itself...  The discussion took off in another direction from there, and I was left to think on this and never got a chance to further question him in this matter...

FE

Hey cent I thought you got lost again. :)

>>>It most likely does not have the capability to report what you're looking for.

I am not looking for anything here. The deal is that MS says the PF usage in the Task Manager reflects how much of the page file is being used. But the PF Usage often exceeds the size of the pagefile.

Example: pagefile.sys is static in size lets say 200MB's but the PF Usage graph shows 300MB's is being used. That can't be correct because the pagfile is only 200MB's and there are no virtual memory warnings popping up which will happen if the pagefile isn't big enough. So if the PF Usage is 300MB's and the pagefile.sys is 200MB's where is the warning when the 200MB's is being breeched. There is no warning, the pagefile is not being increased in size and there is no TEMPPF.SYS. Hence the PF usage isn't what MS says it is or it is a bug.

I am not having any problems I am just trying to satisfy a curiosity. :)

Fatal_Exception first off it would have insteresting to whay that person at MS would have said abut "Why did MS put the option in there ""No page file""?. With this option checked there is no pagefile.sys anywhere on this machine nor is there any TEMPPF.SYS or any file that is being used as a page file. And why did they carry this option over to Win2003 if it doesn't do what it says it does(BTW even Win2003 if this option is used there is no pagefile either). I too have corresponded with a few MS people and they too have said that XP always uses a pagefile of some sort but they are unable to prove that and secondly they have no clue as to how XP is using a pagefile when I can show them that there no file on the disk that is being used as a pagefile.

And again

"And please lets focus in on the question. This has nothing to do with if there is a pagefile or not. The question is about what the task manager is reporting about PF usuage."

Hey C
You know me.  All over the place............  :-)
I realize you're not having problems and it's just your "nosey" side cropping up.
Lord knows I've spent more hours on these types of issues just to satisfy my curiosity and get nowhere. I should think I would know better by now!   LOL
IMO this is one of those "three monkeys" issues.
FE above has perhaps stated and expressed most eloquently  what my "Totally unsubstantiated " views are on this issue.
Too bad he didn't get a chance to follow up on it, but sounds like he got enough out of it.
Cheers

>>>I should think I would know better by now!

Hehehe me too. Yep me too. :)

I have that guys email address around somewhere..  Think I will link this thread and ask him for a review..  Perhaps he can push the question up the ranks and we can get something definitive on this subject...   And I agree, CO, that it does not make sense that MS would give us that option, then deny that we can use it that way...  And we should know better after 20 some years of this, eh..??

Yeah I find it a bit humorous that not only is this option available they didn't remove it in Win2003. If the option didn't do what it says it does then one would think that it wouldn't be carried over to the succeeding OS. :)

As a matter a fact I think if you have enough RAM Win98 will boot without a page file. But it will moan and groan later on in a session about it.

My take on that is that there is a "Virtual Page File" created on the RAM.  Hence the "you cannot really have no pagefile"...  Kind of a work-around of having a pagefile in a slower media.  As I understand it, the memory has two parts, Non-Page Pools and Paged Pools.  Is it possible that a virtual file is created in the Paged Pools, that is acheiving the same functionality as a pagefile?  Just think of it as having a virtual drive with pagefile.sys in it?

- Info

Here's a test...

By default the "Everyone" group has access to the volume on which the "pagefile.sys" resides.  In the process of making you systems more secure (If that is possible) you remove the "Everyone" group from the volume and do not grant "SYSTEM" any access.  How dows this affect the reported pagefile, in task manager?

Afterall if the operating system does not have access, how can it use the pagefile or report usage of it....

Hi all from a newbie!

I stumbled on this site whilst looking for a solution to a problem with vhosts under Apache - and what an excellent site it appears to be!

Anyway, whilst having a look around I found this question and I wanted to offer my input - so here goes:

CrazyOne:  

The page file usage reported by XP Task Manager is correct in your example even with "no paging file" set in the system options.  Windows's memory management has always been designed with the use of a paging file in mind.  Windows programmers still follow this fashion today even when memory is cheap and plentiful!!

If you select no paging file XP will follow your request and indeed, you will have no paging file!
BUT......  What it will do instead is use your main memory as it's paging area.  So you have the equivalent of a pagefile.sys in your RAM! And that is what Task Manager is reporting.  Crazy, huh!

So, Why does it do this?

Back to my first point.  Most Windows programs are written to make use of this memory paging area.  When you run such a program, that program will request a certain amount of 'paging area' (some programs request several hundred Megabytes).  If you do not have a paging file, this space is allocated out of your main memory - hence the recommendation for a page file!

Hope this un-baffles you!

Regards, Barry

Thanks Barry..  do you have any links to this..??

I have thought about this in depth and I can't see how the RAM is being used for the page file.

PF Usage 400
Amount of RAM available out of 1GB = 650 MB's

Ding Ding Ding the numbers do not compute. I recently got the RAM down to 75MB's available by having around 50+ programs open at the same time so there couldn't be a page file in use in RAM in this scenario because I Have a RAM monitor and I watched it and found that the RAM used wouldn't have anything to do with a page file in RAM. I believe XP has built into it a way to parse request to the page file back to the RAM but not as a page file in the RAM but as regular usage of the RAM. The RAM does not house a pagefile. To many inconsistency with that theory.

I will continue not to use pagefile. My system performs marketably better with out then it does with one. Sorry but these theories don't quite add up. :)

No sorry - It's from memory (non-paged!)

A similar question was raised during one of the projects I was involved in last year.  That was the explanation from a well respected techie at M$.

Regards, Barry

As I have said over and over again this does not account for having a static pagerfile of lets say 200MB but yet the PF usage is 300MB's. Now I know for a fact if the OS/Programs need use of the pagefile but if there isn't enough room in the page file the OS will throw and error in so many words that the system is low on virtual memory and then the OS says it will do something to accommodate this problem and suggest that the users manually increase the pagefile. But in the scenario I just outlined that does not happen. So I am still under the impression that the PF Usage isn't what it says it is or very buggy and inaccurate.

In other words without solid proof the mystery is still unsolved.

BTW I wish everybody who hasn't done so please read the Q in its entirely. Note the issue isn't about not having a paging file it is about the inaccuracies the PR Usage reports.

CrazyOne:

Sorry, I didn't mean that RAM was being used for the page file - I meant that it effectively uses your main memory for all of its paging when you have no page file.

From your example of 50+ programs open if any of these programs had allocated virtual memory you would have seen problems when it had tried to use that memory - the program would have to have terminated with an exception.

Windows and most of its programs have to use a page file - it is a fundamental part of the structure of the underlying code.

Instead of writing out the data destined for the page file to disk, it writes it to RAM.

Windows uses it's page file in a different way from other OSs.  It will use its page file even whilst it still has hundreds of MBs of free RAM.

Try re-enabling your swap file to the recommend fixed size and adding ConservativeSwapfileUsage=1 to the 386enh section of system.ini.  This forces Windows to use the swap file only when it has really run out of RAM.  You should see the same marked performance as you are experiencing now, only the page file will be there if the system really needs it.

Task manager reports how much page file it is using - if you only have a 200MB page file and it requires 300MB allocated virtual memory it is using the rest from RAM - if it doesn't have it when it comes to use it, the program will terminate or the system will crash.

And yes I have read the question as I am sure everyone else has - Your question was,
 "Any body got any ideas about these discrepancies.?",
I cannot see any other question marks in your opening statement.

Good luck, Barry

>>>Task manager reports how much page file it is using - if you only have a 200MB page file and it requires 300MB allocated virtual memory it is using the rest from RAM - if it doesn't have it when it comes to use it, the program will terminate or the system will crash.

>>>if you only have a 200MB page file and it requires 300MB allocated virtual memory allocated virtual memory it is using the rest from RAM

I don't think so. If it is making the requet it already knows there isn't enough RAM to house it. Thus a warning message is issused.


Not true do your own experiment and you will find out that RAM is not used to hows a pageflle. There is no way to get the numbers to balance.

Also BristolBaz your a missing a rahter siginificant point
"As I have said over and over again this does not account for having a static pagerfile of lets say 200MB but yet the PF usage is 300MB's. Now I know for a fact if the OS/Programs need use of the pagefile but if there isn't enough room in the page file the OS will throw and error in so many words that the system is low on virtual memory and then the OS says it will do something to accommodate this problem and suggest that the users manually increase the pagefile. But in the scenario I just outlined that does not happen. So I am still under the impression that the PF Usage isn't what it says it is or very buggy and inaccurate."

Secondaly just because other Windows OS's seem to need a pagefile doesn't mean it has been carried over to XP hence why one has the option to have "No paging file" and I can categorically state that the RAM is not being uses in part as pagefile. Again the numbers don't add up. I can start my machine and the PF usage is 250MB's but the used RAM is only 200MBS so how can 250MB's RAM be used if only 200MB's of RAM is being used. There is no way to rectify this inconsistency. The page file is not being housed in the RAM at least not as a pagefile. XP must be handling the page file request in such a way that it tells the app that no pagefile is present and it will need to attach itself to the memory space of the application. Again this not a RAM pagefile just an increase in the applications memory space.

BTW The ConservativeSwapfileUsage only affects 16 bit apps. It is a hold over from the Win9x days. When the system boots it doesn't even look at this until a 16 bit app is run.

I think you may be referring to LargeSystemCache

Hmmm I Just noticed in that in ME and Win98 it has the option "disable virtual memory" so it is new to XP.

Also I noticed that the reported amount of PF usage and the amount RAM usage exceeds the amount of the total RAM. Cleary XP's reporting tool is inaccuarte and mislleading probably due to a bug.

so it is new to XP = so it isn't new to XP

And here is another thing

Virtual Memory      
Total      1948 MB
Used      765 MB
Free      1182 MB
Utilization      39 %

Now how can a system with no pagefile and 1GB of RAM have a total of 1948 MB of virtual memory?

it probably has a pagefile, windows just tells you it doesnt

rynoski if you had read the whole thread you will see it has been proven if one sets the option to No paging file then there is NO PAGE FILE on the DISK anywhere.

*grin*  back at it again, eh, CO...??

LOL...   Perhaps you should let it go, CO :-)   This post is getting so long that nobody new has the patience to read the thread from beginning to end :-P

LOL you may be correct but I like to keep open because I have seen longer threads then this that finally got an answer. BristolBaz brought some interesting ideas but I just can't get the numbers to add up properly.

Yes, Bristol's comment is well received and goes along with MS's insistance that the system will not boot without one..  Would like to see an answer myself to this..

FE

>>>>MS's insistance that the system will not boot without one

MS is stubborn they very seldom will say they are wrong even when they are wrong. :)

Got that right..!!!

BTW, I think that PF usage shows not PF usage but It shows commit charge - total. you can see PF usage with this prog
http://www.mv.com/ipusers/standards/ThisAndThat/PageFileUsageMonitor.html. And it there are now PF prog shows empty window

Just a thought - what if you removed some RAM - maybe take it back to 128MB and see what happens?

Another thought - I saw an article in PC Magazine about speeding up Windows and one of the tips was to "Lock Windows Kernel in RAM" - http://www.pcmag.com/article2/0,1759,887799,00.asp

Apparently, Windows will swap its base OS code to disk even if the Swap File is disabled, so it has a separate page file for the main OS code. But even after I followed the advice to changed the registry entry to make Windows stay in memory, there was still 140+ meg "page file" in use - it must have some *other* page file somewhere that it just doesn't tell you about, but there are no files on my hard drive that are that big.

Comment from a non expert-

I have indeed read this thread from one end to the other in one sitting  (note the time) and am still interested in finding out what in fact is happening.

This puzzle would help me understand some of the events I am encountering.
Terminal curiosity is my lifes' motivating factor I believe....no pun intended.

BTW, CO, I am from CO as in Colorado and have not a clue about Colorado lobster...though the oysters are common knowledge hearabouts....  Have a clue for me??  <smile>

What This W is not abanoned. This last comment was made 04/22/2000 Please do not close this Q it is still live.

Co Lobster is kind of a private joke between me and Cent. :)

Good history between you.

I haven't re-read this whole message but I do need to be told, if I need a paging file, assuming I do, how exactly do I direct to a newly formatted HD I am about to install? or move one.

An Aside-I installed a cleaned 60G WD drive as a slave, used XP to add a partition, didn't partition the whole drive yet so I can put all sys and drivers in one part, use second as yet unformatted as paging later.  

The system wouldn't reboot.  Got stuck after initial post and at a 'verifying ' line-forgot the text.

I removed it and the sys boots fine-obviously.
What's up with that?

Clues welcome with great gratitude.

Don't know what I did before i joined this forum site.  Best thing I've found in years.

S

For Shlew,

A paging file is useful if you have alot of applications open and insufficient RAM to run them all efficiently. The old rule of thumb was to set it to the same size of your RAM plus 16, but these days you can adjust it to pretty much whatever you like, or just let Windows manage it for you. To alter the settings in XP go to My Computer and right click & select Properties - go to Advanced.

The HDD - you must be running 2 or more disks? If so ensure the original disk with the boot sector is set to master (it will have ntldr and boot.ini on it). It's better to add the new disk to the secondary ide cable. That would explain why you got nothing - no bootable section and no OS.

Good luck

Hi CrazyOne, I did read the entire question...and now i'm typing my bloodshot additions here...

you said...
>> Task Manage is saying that I have over 244MB's of data being housed in a pagefile that does not exist in any way shape or form.

and in the same post you said...
>>Example: Lets say I have Max and Min of 500MB's pagefile. This means the 500MB's is locked down and cannot grow. However the Taskmanager says 750MB's page file usage.

So, if you set the page file to 500MB or to 0MB does the Task Manager report the wrong amount but of equal (or darn near close) value ~250MB?  Or was your "Example:" just numbers you made up?

example:
  PageFile            Reported
    500MB             ~750MB
      0MB               ~250MB

If so, then it would seem there is just a bug in the Task Manager's reporting of the PF stuff, even if it's not exact, but close, there could be a fluxuation or something, cuz to continue to read for changes it would have to run some kind of a loop or something right?

~ MastaLlama ~

You cannot "delete" the pagefile in Win XP. If you set it to zero, XP will create a pagefile on it's own and you won't even know it. XP requires a pagefile to run - your choice is to manage it yourself or try to disable it and have no control over it.

for good measure read:

"If you have a large amount of RAM (at least 1 GB), you might think that Windows XP would never need virtual memory, so that it would be okay to turn off the page file. This won't work, however, because Windows XP needs the page file anyway and some programs may crash if no virtual memory is present."

http://www.microsoft.com/WindowsXP/expertzone/columns/mcfedries/03june16.asp

http://aumha.org/win5/a/xpvm.php

you can find a pagefile vbs tool at http://billsway.com/notes%5Fpublic/winxp%5Ftweaks/


and.. if you want to prevent Windows from moving any core OS kernel or driver files to the pagefile, perform the following steps:

Start a registry editor (e.g., regedit.exe).
Navigate to the HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management registry subkey.
Set the DisablePagingExecutive registry entry to 1.

the idea is also that windows usually writes kernel dumps(!) into the pagefile!! if its not there... he'll  create it somewhere on hd.
"If you are debugging crashes and wish the error reporting to make a kernel or full dump, then you will need an initial size set on C: of either 200 MB (for a kernel dump) or the size of RAM (for a full memory dump). If you are not doing so, it is best to make the setting to no more than a ‘Small Dump’, at Control Panel | System | Advanced, click Settings in the ‘Startup and Recovery’ section, and select in the ‘Write Debug information to’ panel"

try changing the dump type and scompare with the pagefile its being repported!! coincidence!? ;)

Ok thanks for the recent comments. I still really haven't found any evidence that XP needs a page file if one has enough RAM. I get memory dumps once in a while but I have to tell you I think MS is a bit incaccurate that a page file is needed for a dump file. As a matter of fact the dump file is usually created its own file depending on what created the dump file. Now think about it the page file is C:\pagefile.sys so it seems ulikely the two are really related. I know what MS says but I think they don't always know precisely everything about their own products. Condsirering there can and are 100's and sometimes 1000's of people working on a single project it seems likely to me that sometimes the left hand doesn't know what the right is doing, :).

Again I reiterate there is no page file on my mahcine. And no programs are crashng or complaining about not having a page file. I know in Win98 if you turned off the page file some programs would complain about not having enough memory even though there was plenty of RAM. Turn the page file back on and the programs stopped complaining. In XP with the amount of RAM I have and no page file no program has complained yet. And if we all think about it why would there be an option to have no page file in XP and Win2003 if it weren't true. Believe me if you have enough RAM XP and Win2003 are very adept at hanling not having a page file.

Yes, and remember what microsoft people are also saying: if you switch off pagefile, xp will still create a file for kernel dumps whilst keeping it secret from you i.e. not in pagefile.sys!! The kernel dumps are only for crash and recovery purposes!! It makes sense when as you know the full kernel dump is around 200mb, similar amount you're saying xp is claiming you have in pagefile!!
In the links on my messages there is info on this.
I'd say monitor the amount xp is repporting whilst turning kernel dumps off? or perhaps changing the dump type?! See what happends.
Cheers
Pedro

Hi CrazyOne,

In the directory c:\Windows\system32 (this is default anyway) there is a script.

Its called PageFileConfig.vbs.

Check it out - that may be what you are after.

If not go to the following site:

http://www.billsway.com/vbspage/

This guy has written a few handy vb scripts - one will tell you your page file etc.

Good luck

Doug

So heres my 2 cents.

The page file size reported never reflects the page file because the kernal's page size reported in the Task Manager shows the amount of memory that CAN ideally be placed into a disk rather then physical memory.  So yes u can have 0 sized page file (or none at all!) and still have a kernal page size >0; the page-able memory allocations is in physical memory and not on a disk.

Now to address the issue of utiliztion:

Virtual Memory    
Total     1948 MB
Used     765 MB
Free     1182 MB
Utilization     39 %

What is going on here it that since page-able data is assumed that is CAN possibly be placed on disk, it is counted as virtual memory, even tho it is sitting in the physical memory (for perfomance issue); once more physical memory is required, the page-able memory is then offset into a page file.  So what happen when you can not have a page file?  Well then the pageable-memory can not be pushed of into a file and stays in real memory => it will not be possible to achive of 100% utilization for the above example.  Your 'Used' memory will get 'stuck' at your physical RAM limit.

So i.e. virtual memory does not automatically impies that it is in a physical disk, it sometimes is kept in RAM for performance reasons until more physical RAM is required.

True rsriprac but the kernel memory never matches what is being offset suppposedly to pagefile in the taskmanager

Unless I misunderstood your response, the paged  kernel memory does not mean it will be in the page file, this is why it is the same.  The systems determines where the "paged" memory is going to be placed, and its true that there is some parts of the paged memory to stay in physical RAM.  Unless you find a way to utilize 100% of your physical memory for non-pagable memory, there will always be a difference between the page file and the "paged" memory indicated in the task manager.

I hope I'm making sense.

-Ram

zediDust your explaination makes the most since so I am going to accept your comment as the answer. On thing though my system perfomrs overall better without the pagefile than it does with it. :)

Nice one Zedi..!!  :)

Well, to be fair, if you don't put huge demands on your memory, and there isn't excesivley large amounds of memory paged at any given time, it quite well could be faster (under specific circumstances like these) I suppose, depending on system specs

Personaly, my commit charge here at work is usualy in the neighbourhood of about 2GB (thanks to copies of ORACLE, MSSQL, MYSQL, IIS, APACHE and other such server apps all running at once) and 5 or 6 instances of the visual studio IDE also all open, along with other great tools. The more you multi task, the more you'll see your pagefile commit charge increase.

Ahh.. the life of a developer.

haha
that being said, my gut instinct is still that there is a way to get your machine even a little bit faster with a swap file! lol

Nah I don't think so zediDust. I have experimented a lot with and without a pagefile. I have 1GB of RAM and for what I do the RAM used rarely falls below 200MB's (I programmed my own RAM usage indicator for this testing). For me with a pagefile the system is slower. Withoug a page file it is faster.

When you had a page file, did you defrag it often (with as much free space left on the drive as the size on your page file)?
Did you keep the swap file on its own physical drive?
Did you you have at least 2GB of swap file space, with the min and max size amounts both set the same.
And also do you have semi recent drives?

I know its not so pertinent to the question anymore, but I'm just curious ;)

I have tried various page file sizes, moving it to its own partition and yes when it was on the the system partition I didn't need to defrag it because when you set the max and min to the same size the file doesn't get fragmented.

The drive in use is 1 year old. I have plenty of RAM to hanlde the normal memory load and the page load. Believe me I have tested this a lot and without a pagefile the system perfoms noticably better. :)

hah! go figure

Yeah I think it has to do with memory load. A pagefile perse is a disk file which of course a disk file is slower than what is in RAM. If there is enough RAM to handle the entire load then theoretically this should increase performance.

I would figure that, supposedly, the page file doesn't take effect UNTIL you run out of RAM.  If that's the case, then it really doesn't matter what size your pagefile is because it should never touch it, right?  ... Just my two cents :-)

P.S.  That was a good explaination, Zed.  Kudos to you!!!

- Info

The answer to this question is really very simple.
The issue is simply a poorly labeled graph. In spite of what Microsoft documentation might imply, "PF Usage" DOES NOT mean actual pagefile usage. This is actually the "Commit Charge" displayed in different units. The Commit Charge is the total of modified data, but not program code, for all processes. Some of this will be in RAM and some in the pagefile, and some will be in both RAM and the pagefile. If there is no pagefile all modified data must remain in RAM at all times. Program code will continue to be paged as it does not use the pagefile.

Another way to look at Commit Charge is potential pagefile usage. If absolutely everything that was in RAM that could be copied to the pagefile actually was, this is the size that would be required. That isn't entirely correct but it is usually close. The actual size of data in the pagefile will usually be much lower, or even zero if there is no pagefile. There is nothing in Task Manager that will show actual pagefile usage.