DisablePagingExecutive

  • Thread starter Thread starter Guest
  • Start date Start date
G

Guest

I'm interested in making use of the registry hack DisablePagingExecutive=1 in
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory
Management].
("This entry specifies whether user-mode and kernel-mode drivers and
kernel-mode system code can be paged to disk when not in use. If the value of
this entry is 1, the drivers and kernel must remain in physical memory.")

I can't find any clear guidance on how much system RAM is deemed sufficient
for this setting to be safely used. Also, how much physical RAM will
actually get tied up by this process on average? Is there any way to measure
how much a particular system is using for these drivers and Kernel?

Thanks for any help you can offer.

Best wishes,

Andrew
 
Ted Zieglar said:
Ask your question in a hacker newsgroup.
Click to expand...


Sorry about that "advice" you were given, this is indeed a Windows XP performance related question and thus is perfectly valid here.


While that entry is mainly used by servers, it does have the potential to increase performance for workstations as well. Like you
said, you would need enough memory but Windows does not usually make full use of physical RAM anyway, so you'll probably have
enough. The common literature on it says that you should have a minimum of 256MB (I think that's XP's minimum in general, so it's
not a good idea) but 512MB is the recommended minimum. Of course if you have 1GB or more then you should be fine (even better than
fine if it enhances your system.)


The thing to keep in mind is that Windows does not come preconfigured to be as optimal as possible (that is literally impossible
with the boundless number of hardware combinations). It comes set up to be optimal enough for most people's use. Also, speed and
stability are at odds with each other, so Windows tends to lean towards stability when possible. Thus a fresh installation of
Windows will not necessarily be as fast as it can be. Some people say that "hacks" are pointless and don't give any substantial
benefit while increasing problems. While some can cause problems, some CAN give substantial performance gains. The ones that do
give the best increase are well documented and well known, so you will easily be able to identify the ones you should bother with.


HTH
 
Thank you very much Alec for your very helpful reponse - it's nice to know at
least someone actually reads and understands a post properly before hitting
reply!

I'm going to try out this hack on some of our >=512MB PCs and see whether it
has a measurable effect. I hope the theoretical performance gain in the OS
is not met by a performance loss in apps!

Thanks again!

Best wishes,

Andrew
 
Cantoris said:
I'm going to try out this hack on some of our >=512MB PCs and see whether it
has a measurable effect. I hope the theoretical performance gain in the OS
is not met by a performance loss in apps!
Click to expand...


I'm trying it out right now (plus a few other tweaks). Like I said, Windows does not use all available physical memory (it rarely
ever even comes close.) One way to force Windows to use more physical memory and less virtual (disk) is to reduce the swap file
size. Some people have even suggested eliminating it altogether but that can cause problems or even prevent boot up at all. I'm
going to try reducing my swap file now.
 
"Like I said, Windows does not use all available physical memory (it
rarely ever even comes close.)"

Lol...Now all of us can see what an expert you are!
 
Ted Zieglar said:
"Like I said, Windows does not use all available physical memory (it
rarely ever even comes close.)"

Lol...Now all of us can see what an expert you are!
Click to expand...


Huh? I suppose you're an expert? Maybe you should take your complaints to a nagging newsgroup.
 
Cantoris said:
I'm interested in making use of the registry hack DisablePagingExecutive=1
in
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session
Manager\Memory
Management].
("This entry specifies whether user-mode and kernel-mode drivers and
kernel-mode system code can be paged to disk when not in use. If the value
of
this entry is 1, the drivers and kernel must remain in physical memory.")

I can't find any clear guidance on how much system RAM is deemed
sufficient
for this setting to be safely used. Also, how much physical RAM will
actually get tied up by this process on average? Is there any way to
measure
how much a particular system is using for these drivers and Kernel?

Thanks for any help you can offer.

Best wishes,

Andrew
Click to expand...
Isn't it the size of non paged pool? Isn't non paged pool named that for a
reason? Sending the disk driver to the pagefile seems like a prescription
for disaster.

However, VMS which is the distant ancestor of NT never paged the executive
unless the amount of physical memory was extremely small. The resulting hit
on performance (or paging the executive) was severe to say the least.

So, I am surprised to learn that XP can send parts of the executive to the
pagefile all by itself. After all, there aren't very many XP systems which
don't have what seems like a very large pool of RAM.
Jim
 
Back
Top