|
Page faults racking up in Task Manager, windows xp |
Post a Reply |
From: Tim on 07/07/2003
Hey all, I have a question about the reporting in Windows xp Task Manager. I added "page faults" to the view and every process running has racked up page faults. Most are in the thousands and explorer.exe racks up one per second, going right along with the cpu clock. Is this mis-leading, or should I be looking for a problem? TIA Tim
| |
Votes:
Rating: 0
Delete: 0
|
|
--- Replies to this Problem ---
From: John on 11/28/2011
I know that page faults are not errors, merely how virtual memory works. That being said, an excessively high number of page faults could mean performance problems, if the system is spending more time moving data in/out of memory than doing useful work (called thrashing). All of the above notwithstanding, I do see some page fault activity in my IE processes, but the taskmgr process itself is generating about 200-300 page faults per second. If I restart taskmgr, it goes back to zero again, but quickly builds up. I don't know if this is a problem, but I suspect that it is. I am running XP and have 4GB of memory (and taskmgr shows most of it available). Anybody know what could be wrong?
| |
Votes:
Rating: 0
Delete: 0
|
|
From: Geoff on 07/19/2010
I think you'll find that the term "Page Fault" stems from the fact that it *is* a Hardware fault. The process is unable to find the page in RAM and so generates a hardware fault (aka interrupt) to fetch the page from the virtual memory store.The first time you load any program from disk you will have 100% page faults (as the program isn't in RAM yet and must be loaded). Then if the page is paged out of RAM back to disk, you'll get another page fault when the CPU requires that page to be retrieved again. Page faults kill performance and more RAM should reduce the problem
| |
Votes:
Rating: 0
Delete: 0
|
|
From: tilius on 02/12/2010
I am intrigued why a number of replies refer to the number of faults 'per second' that task manager reports and this being linked to the CPU speed or similar. Surely this is simply the rate at which task manager updates ie menu item = view / update speed.
| |
Votes:
Rating: 0
Delete: 0
|
|
From: norman on 03/05/2009
More RAM won't solve it. It seems that the task manager help is misleading, too. A page fault (I have 190,000,000 of them in 10 minutes) seems to be when the code being executed attempts to read or execute past the "currently loaded memory frame" and needs to read more data. Having said how many page faults I have, it seems that windows is NOT reading them from hard drive (my hard drive is not ALL THAT fast ;-). My conclusion is that windows by default only loads small chunks of code or data for you into "memory". If you need more data it will get it from somewhere and that somewhere being again "memory" in most of the cases, since most of your computer's ram is being used as a cache of files you loaded so far. Stupid, but that's the way it seems to be. So even if you add 3GB of memory, windows will still load only 4096 bytes of your data for your application into it's 256MB execution space and use the rest of your 4GB as cache. When you do something nice which needs more data than 4094 bytes there you go - it has to go into the cache and load another 4096 bytes of your data...
| |
Votes:
Rating: 0
Delete: 0
|
|
From: dribble it is software related = bad code. on 05/28/2007
6 pc all loaded with xp1. All havd no page faults at all. I reloaded an few hours old system image file, like a ghost image ect.This had page faults, now again it has page faults. It is obviously a fault increases every second or faster looking at task manager - explorer. So i installed all the same software to another pc that had no page faults. Not doing it analitical to find the culprit. Just to see if the error would occur again. Results = page faults same as pc above. It has to be software or driver related. Although i had also loaded up microsoft updates from march_06 cd. Now i have 4 pc's that have no page faults and two that have. Now i will go and find out the offending code and keep it of any pc. Or find its fixed update if one available.
| |
Votes:
Rating: 0
Delete: 0
|
|
From: soreen on 11/14/2006
Hi there guys, I thought also that page faults are errors so when i looked into this i was almoust shittin` my pants :) [img]http://img140.imageshack.us/img140/7263/untitledos4.jpg[/img] Now i'm ok knowing what it is! However more ram not means less pagefaults. Some softwares (aS video editing ones)need the pagefile, and if not there u may expect lots of crashes even if u have 10GB of ram!
| |
Votes:
Rating: 0
Delete: 0
|
|
From: Andre on 08/04/2005
Its true. A page "Fault" is poorly named.. Every time a process looks for some data in ram, if its not there this is called a "Page Fault". Windows then loads the page into RAM from your hard drive.. It isnt an error at all!! I shirt myself when I saw zonealarm pro had 1,000,000 + "errors" daily until I realised it wasnt an "error" at all! Its neccessary, theres no way u can load ur whole hard drive into ram. then u would never get a "page fault". Dont Worry!! Turn that column off ur task manager, its a stupid thing to call it a fault, its not. check this link out for fat details if I still havent convinced u not 2 waste ne more time on this ;) http://en.wikipedia.org/wiki/Page_fault
| |
Votes:
Rating: 0
Delete: 0
|
|
From: Howard on 04/21/2004
My jaw dropped when I saw the >1 PF/sec that explorer.exe was generating on my Win2K Server box... 640 MB RAM and a really light load.So I said to myself "what's happening EVERY second?" The answer is: Task Manager is updating it's info every second. Want to test my theory? Just open Task Manager, go to the processes tab and verify that explorer.exe is just spewing out the PFs (you may have to first enable the column from View-->Select Columns-->Page Faults), then go to the View menu, select Update Speed, and then select Paused. In my case these pesky PFs stopped immediately, and of course started again when I reset the update speed to normal. I'm satisfied ;) comments? -Howard
| |
Votes:
Rating: 0
Delete: 0
|
|
From: Marcel on 11/17/2003
Hi all!Thats all no error at all. A 'page fault' is just a term invented years ago by somebody who invented 'virtual memory' or in plain words: a swap-file on a harddisk to store memory if you don't have enough of it. It's not an error, it's just an indicator of how many times the operating system needs to load a block of 'virtual' memory from the harddisk back into the pc's RAM. If you add more memory, you will lower the chance that memory has to be stored on the harddisk. But windows likes ram a lot... it tries to keep often-used memory blocks into your RAM while it saves less-used memory blocks to the swap-file so it can store more-often used blocks into RAM again. That's just about it.
| |
Votes:
Rating: 0
Delete: 0
|
|
From: William on 10/03/2003
In my case, I believe that it is a bug in explorer.exe. Periodically, my page faults / sec for explorer.exe hit and exceed 2 thousand!!! which of course consumes all avail CPU cycles in priv mode to service the race. My only correction available is to kill explorer.exe in the task manager and restart it (so that I'll have my desktop gui back). Then I'm OK for a day or two, until the condition recurs. This, in my case, has absolutely nothing to do with actual available memory. Microsoft says they corrected it in XP SP1, but in my case, there is no correction there.
| |
Votes:
Rating: 0
Delete: 0
|
|
From: Jack on 09/15/2003
Yes, a bit late in seeing this but will throw in my 2 cents. I just noticed this problem in Windows 2000 (Explorer.exe) on 2 machines (the only 2 I've seen since discovering this an hour ago.) I had Page Faults listed in the Task Manager for a long time but didn't pay attention to the almost astounding number being generated by explorer.exe - one "more" than every second on both machines.One of my sytems has more than enough ram: 512...plenty when no big tasks are running heh. And the page faulting in explorer is occuring when the bloody things sit idle with very few processes. I see the other processes requiring a hard drive hit fairly often (if that's truly what they are.) Understandable - but explorer.exe doing it almost twice a second = funky: a fast update or something that isn't being handled correctly by explorer, or the os...or whatever. Looking at it now, it's up above 110,000 while the other 13 processes are 6500 or way below. Interesting this is occuring on both xp and 2k systems...what chipsets is this noticed in (Intel, AMD?) Mine are both AMD: KT-133 & nForce2. It's probably not memory related on the amd side (as far as our systems are concerned; unless it happens to be an inherent problem with all current AMD chipsets.) This scenario reminds me of the performance counters constantly running hidden on 2k and xp pro systems (possibly xp home?) Like it's intentional...ms must know this is occuring in any case. But is it an unfixed glitch in NT 5.x or what? I'll dig around a little more but I'll bet 10 bucks that it's just bad programming they need to fix.
| |
Votes:
Rating: 0
Delete: 0
|
|
From: Doug Campbell on 09/08/2003
I noticed that I had MILLIONS of these Task Manager Page Faults on my computer so I decided to look around for problems. I noticed that I had 350MB of DDR memory used out of approximately 1GB. I also noticed that I had no Virutal Memory in use. I remembered that I turned off Virtual Memory when I had to replace the drive that it was originally on. I remember that I was going to restore the Swap File after doing a data restore and defrag - but now I realize that I forgot!!! I just restored the Swap file to my Drive E: while writing this after freeing serveral GB of space. I don't know if I will be back to this site to report my findings, but I thought my scenario could help some of you in the same situation.
| |
Votes:
Rating: 0
Delete: 0
|
|
From: Kylie on 08/30/2003
According to a help page at Microsoft a page fault occurs everytime a process attempts to access something from memory and finds it not present so then returns to the HDD to retrieve the info.This means that it's sort of stupid to call it a 'fault' (my opinion at any rate)
| |
Votes:
Rating: 0
Delete: 0
|
|
From: Tony on 07/31/2003
I also have this 'problem'. What can we do to fix this, or at least reduce the page faults for better performance?
| |
Votes:
Rating: 0
Delete: 0
|
|
|
|
|