Help w/ disabling the PF in XP

V

Vette5885

Gawd
Joined
Apr 12, 2005
Messages
629
I've disabled the page file in XP, yet it still registers as having a page file in the task manager.

I'm currently running a frankenstein box (mobo from my computer in the signature is shot) so i have 3GB of memory. I don't need a page file, but even though I disabled it, it still reads upwards of 250MB on the page file.

What gives?
 
djnes said:
It's never actually disabled completely,
Click to expand...
Huh? When there's no pagefile, there's no pagefile.


Vette5885 said:
it still reads upwards of 250MB on the page file.
Click to expand...
Where are you seeing that value?
 
mikeblas said:
Where are you seeing that value?
Click to expand...
In the task manager, even though I've disabled the PF, it still has the PF meter, which reads upwards of 300mb.

Doesn't make much sense to me, but this is a temporary setup anyway, so it doesn't really matter much. I was just interested as to why the PF still exists when in all reality it doesn't.
 
Vette5885 said:
In the task manager, even though I've disabled the PF, it still has the PF meter, which reads upwards of 300mb.

Doesn't make much sense to me, but this is a temporary setup anyway, so it doesn't really matter much. I was just interested as to why the PF still exists when in all reality it doesn't.
Click to expand...

If it's completely off in the advanced settings for your system, then it's off.

There are a couple-few values in Task Manager that are incorrectly named, and "PF Usage" is certainly one of them. The graph actually shows the total of committed memory for the system.

If you're interested in this sort of performance measurement, the first thing you'll want to do is get some better tools. I like ProcessExplorer from the sysinternals.com website and use it as a replacement for Task Manager on all my machines. There, you'll see the same graphs labelled "Commit" and "Commit History".

You'll also want to learn to use PerfMon. There, you'll find the Memory/Committed Bytes counter, which shows the same value.
 
mikeblas said:
Huh? When there's no pagefile, there's no pagefile.
Click to expand...
That's what I thought too, but it was proven in the many flamefest on this subject, and was mentioned in the sticky as well, I thought.
 
djnes said:
That's what I thought too, but it was proven in the many flamefest on this subject, and was mentioned in the sticky as well, I thought.
Click to expand...

I don't see it mentioned in the sticky. (Which is a real train-wreck; it should be cleaned-up and gone through again for accuracy. It would be great if it cited its sources, too.)

In Microsoft Windows Internals, Russinovich and Solomon write (on page 445):

If no paging files are specified, Windows 2000 creates a default 20 MB page file on the boot partition. Windows XP and Windows Server 2003 do not create this temporary pagiving file, which means the sytem virtual memory commit limit is based on availalbe memory.
Click to expand...

If you have "proof" of the creation of a page file on Windows XP or Windows 2003 machines, I'd love to have a link to it.
 
mikeblas said:
If you have "proof" of the creation of a page file on Windows XP or Windows 2003 machines, I'd love to have a link to it.
Click to expand...
Is that directed at me?

I don't know - I'm just going by what the task manager says. I haven't had a chance to look into the other ways to determine PF usage yet (too damn busy) but when I do I'll post what I come up with.
 
mikeblas said:
If you have "proof" of the creation of a page file on Windows XP or Windows 2003 machines, I'd love to have a link to it.
Click to expand...
The consensus when it was all over was that XP still pages, regardless. I don't have any links because I assumed they were all in the sticky. I went back to using a page file after all the debates, because there was nothing really to gain from doing without, except a small amount of disk space.
 
Vette5885 said:
Is that directed at me?
Click to expand...
djnies mainly, who asserted that "it was proven in the many times" that a pagefile is created even when the users asks for none. But if you can offer similar "proof" I'd be very interested in seeing it.

Vette5885 said:
I don't know - I'm just going by what the task manager says. I haven't had a chance to look into the other ways to determine PF usage yet (too damn busy) but when I do I'll post what I come up with.
Click to expand...
And as I mentioned, this is one of the cases where Task Manager is wrong. On this machine you're using, if you've turned off the page file, you shouldn't find a pagefile.sys. (And if you do, you should be able to delete it because nothing is locking it.)
 
Link for the OP.

Title:
Bruce Sanderson's General Windows Information

RAM, Virtual Memory, PageFile and all that stuff

Note: According to mikeblas this post may or may not contain factual errors. If it does you'll have to guess what's wrong about it. He also thinks it's "neat", don't know WTF that means either if it's factually incorrect. So guess away!
 
Why would you ask for the context of a link that describes paging, in a thread where the OP is asking how paging works? What am I missing here...?
 
Phoenix86 said:
What am I missing here...?
Click to expand...
Context is what you're missing! Who knows who you're responding to, or why, when you only post a single word?

The linked document is a pretty neat overview, but it's got some of the facts wrong.
 
Phoenix86 said:
Better?
Click to expand...
It's a step in the right direction, but it's a little too abrasive for my tastes. It makes you look like you've got a real chip on your shoulder, and some readers might be put off by that.

The errors are the common ones about the definitions of the perf counters. Some of the counters listed report PF-backed activity, but also report activity for every other memory-mapped file on the system. After reading this doc, then, some folks will think they're paging private bytes when they're really not.
 
Heh, you think I care about what the readers think... I don't have enough time for that. :)

They can read and learn, and ask questions. I'll be much more apt to respond to that.

Yeah, I almost didn't post in the thread once I saw you here.

You are quick to judge others and not yourself. Heck if I had posted post#16, you would have blasted me for making a blanket statement discrediting a document w/o citing examples. Then you will question the inane, like 'what is the context of a document explaining paging, in a thread where the OP asks about paging'. I think the context pretty self-evident, but here we are, discussing the inane getting futher and further off-topic...
 
Phoenix86 said:
Heh, you think I care about what the readers think... I don't have enough time for that. :)
Click to expand...
And it shows!

Phoenix86 said:
They can read and learn, and ask questions. I'll be much more apt to respond to that.
Click to expand...
The way you do to me, when I ask questions of you? That sometimes doesn't go too well. Is it just that I get special treatment?

Phoenix86 said:
I think the context pretty self-evident, but here we are, discussing the inane getting futher and further off-topic...
Click to expand...
Problem is, the document you linked doesn't answer the OP's underlying question: why does he see a "PF Usage" graph with a non-zero quantity when he thinks he's got no PF? The doc doesn't mention task manager at all, and doesn't mention having no page file (as far as I can see, anyhow).
 
mikeblas said:
Is it just that I get special treatment?
Click to expand...
Yes.

I'm avoiding discussing details with you because I see how those threads go, no different than this one, and I have YET to state an opinion. I tried just posting linked documents and you still miss the point. I wasn't answering all his questions, I was providing a document that will help the OP's understanding of paging.

Problem is, the document you linked doesn't answer the OP's underlying question
Click to expand...
Where did I say it would? That's right, I didn't. Yet another pointless point.
 
Phoenix86 said:
Where did I say it would? That's right, I didn't. Yet another pointless point.
Click to expand...
Where did I say you said it would? That's right, I didn't.

I'm sure you'll be tempted to use the admission against me, but I'll offer it anyway: I really can't figure out what you're so pissed-off about, why you can't respond to me civily, or why you go ad-homenim without provocation.
 
I'm not trying to start any flaming here, mikeblas, but it does seem to take issue with a very many things posted on here, simply for the fact of stirring up debates and arguments. I can't say for sure, but that's my guess at why Phoenix86 is giving the answers he is. You asked for a link, he gave one. If the responses you gave were in the form of a discussion, rather than putting everyone else's comments down as incorrect, this process would flow much more smoothly.
 
If you want a perfect example of why I deal with you this way, read this thread. Two or three times if needed.

Who the hell would respond to your posts civily? You are getting pissy I posted an informative link. FFS, you responded to a LINK, I didn't even say a damned thing. Yet, you expect some "context" which is plainly obvious. Even when I clarify "it's a step in the right direction". How's the climate all the way the hell up there?

How the hell you need context for that link is beyond me.
 
djnes said:
I'm not trying to start any flaming here, mikeblas, but it does seem to take issue with a very many things posted on here, simply for the fact of stirring up debates and arguments.
Click to expand...
And furthering discourse, to get to more knowledge and truth. It's what some people do in a conversation.

djnes said:
You asked for a link, he gave one.
Click to expand...
I asked for a link that showed the "proof" that a page file is still created if the user has configured none to be created. I don't see any such proof in the content linked by Phoenix86, and that's one of the reasons that I asked for context: to see if it did, and I was just missing it. Or to see if it was offered to answer the "why do I have PF Usage graph?" question, and I was just missing it.

djnes said:
If the responses you gave were in the form of a discussion, rather than putting everyone else's comments down as incorrect, this process would flow much more smoothly.
Click to expand...
I'm sorry that you've interpreted my posts that way. I'll see what I can do about it in the future, but I can't agree that I've put "everyone else's comments down as incorrect" in this thread.

Take my response your assertion about the creation of a page file when none is configured as one couter example: I didn't say it was incorrect. I asked for a link to the original discussion so I could read up on it, and offered what I knew and had read about it.

The next step is yours; maybe you'll post some repro steps to make it happen, regardless of what the book says. Perhaps you found an errata list for the book that clarifies or fixes what it says there. Or maybe you'd point out that we were talking about a backlevel version of Windows, and I didn't notice.

All part of having discourse.
 
But there you are doing the same thing Phoenix86 is saying you do as well. I never said the pagefile is still created...I said Windows still pages, whether the file is present or not. No where in my comments did I say the pagefile is recreated, once disabled.
 
djnes said:
I never said the pagefile is still created...I said Windows still pages, whether the file is present or not. No where in my comments did I say the pagefile is recreated, once disabled.
Click to expand...
Ahah! Now we're getting somewhere!

You posted "It's never actually disabled completely," which I took to mean that the initialization of the file was disabled, not that paging was (not) disabled altogether.

I'm sorry for that misread. If that's the root of all this, then I guess I'm guilty as charged, but still can't understand Phoneix86's reaction and sustained conduct.
 
PM sent. Enough of this crap in threads. OP, if you have any questions, feel free to PM me. I'll do my best to answer, sorry we fucked up the thread.
 
You must log in or register to reply here.
Back
Top