Re: Question about paging rate

Tech-Archive recommends: Repair Windows Errors & Optimize Windows Performance



A request for new virtual address space or a fresh data page
will generate what is termed a demand-zero fault. This takes
a page of memory and zeros it for the application to use.
While technically a fault (because it is initiated by a reference to
non-existant memory in the app's address space), it does not in and of itself
produce disk activity - at least until and if it is paged out.

An application that is doing this and continuously growing in
virtual address size is the classic "memory leak" (or just badly
written). If it is doing this and not growing endlessly then it may
be doing process-shared memory or deallocations/reallocations
or using AWE or one of the dozens of other possible memory
management strategies.

Another source of soft faults is when a page is moved back
from the standby list (on its way out the door so to speak), back
in to the application's memory working set. This is just
a page mapping operation and no disk activity is generated.
These are so-called transition faults and the rate is really
dependent on what else is running on the server (especially
SQL Server and Exchange which both have quite highly
specialized internal memory mangement of their own.)

Also, file operations are usually through virtual memory section
mappings internally in the o/s, and so the fault rate counts these
cache misses. Again this is not an indication of an application
problem, it is just the system doing what it is supposed to. If your
application is file-intensive (say a database app), this may be
the source.

And if the app needs to run with "lock pages in physical memory"
rights it is probably safe to assume it knows what it is doing. You
can use perfmon to see some of the VM indicators, though some
things like AWE, shared and large pages are not counted in
some of the indicators.

The best thing to do of course is ask the developers.

Steve Duff, MCSE, MVP
Ergodic Systems, Inc.

"Stuart Rogers" <me@xxxxxxxxxxx> wrote in message news:Xns96A3BF688A987stuartfreenews@xxxxxxxxxxxxxx
>I run an application which I feel is paging too much. The PC is dedicated
> to this one application and mysql. The app in question page faults at a
> rate of between 30--70/second averaging around 55/second, however the
> system pages/second is less than 0.1/second so these faults are not causing
> hard page faults. Should I be concerned about this, it seems to me that
> since this is the only app or service with any apreciable paging that this
> could be a design problem within the app itself. However I have been unable
> to find out what is actually causing the page faults. Can any one suggest a
> way to trace the actual storage causing the fault I could then try and find
> out what this memory is used for within the app.
>
> Stuart


.



Relevant Pages

  • Re: how do you use pfmon.exe
    ... Pfmon displays a list of hard page faults in the command window while the ... might want to rerun Pfmon while limiting the amount of memory to see how your ... LeakyApp is a GUI tool included on the Windows 2000 Resource Kit CD. ...
    (microsoft.public.win2000.cmdprompt.admin)
  • Re: Question about paging rate
    ... None of the memory indicators for this app are increasing so there does ... > These are so-called transition faults and the rate is really ...
    (microsoft.public.win2000.general)
  • Re: Question about paging rate
    ... Note that the paging rate you mention is within a factor of 2 of what a typical disk can do. ... It is not a database app either, however it does manipulate quite a lot of data and files, it collects data from a weather station. ... This takes a page of memory and zeros it for the application to use. ... These are so-called transition faults and the rate is really ...
    (microsoft.public.win2000.general)
  • Re: Can extra processing threads help in this case?
    ... faults, and you kept providing memory mapped files as the ... and refrained from providing VirtualLock(). ...
    (microsoft.public.vc.mfc)
  • Re: remove zero_page (was Re: -mm merge plans for 2.6.24)
    ... then I'd certainly accept the patch. ... (basically -- if the app cares about memory or cache footprint and is using ... And indeed this cacheline bouncing has shown up on large SGI systems. ... Inserting a ZERO_PAGE for anonymous read faults appears to be a false ...
    (Linux-Kernel)