Pagefile placement

[MrMako]

A little debate is going on here at work as too the best placement of the
pagefile on a Windows Xp system. Boss is looking for MS best Practice, which
I read as move it to a different partition. He currently has the desktop
group apply this by creating a D: drive on the same partition and placing it
there.

Our argument is over whether this actually provides any performance boots
over having it on the C: drive with the boot and system partition. Some of us
think it is the same either way and you have to move it to another DRIVE to
get any actual performance boost, but others insists the second partition on
same drive works.

Goggling produces support for both sides, so does MS have anything to help
clear this up? If a separate partition in the same drive does work, could
someone explain why? Thanks!

 

[Leonard Grey]

Try moving the pagefile. Let us know if you notice a difference. I didn't.

"...so does MS have anything to help clear this up?"
You'd have to ask them. No one here works for or represents Microsoft.

 

[Gerry]

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

Two opposing viewpoints are:
1. With a single hard drive place the pagefile file in the first
partition on the drive.
2. With two hard drives place the pagefile in the first partition of the
second drive, preferably in a partition dedicated to the pagefile. A
small pagefile is, however, still needed on the first partition of the
first drive.

Many users will argue, however, that even where there are two hard
drives the pagefile should be on the first partition of the first drive.

If there are two hard drives with differing read / write speeds then the
pagefile might be better placed on the faster drive.

Placing a pagefile in a second partition on a drive is not generally
considered by either side to be the best approach.

Either approach will not bring significant performance gains. The
advantage of a dedicated partition is that it helps with taking backups
and reducing the impact of fragmentation. Keep the operating system and
programmes separate from data.


--



Hope this helps.

Gerry
~~~~
FCA
Stourport, England
Enquire, plan and execute
~~~~~~~~~~~~~~~~~~~

 

[MrMako]

I am not sure you understood my post. I mentioned the practice was already in
use per our Boss so I do not need to test it. I agree there is no difference,
but I am looking to see why MS thinks there is one and for something I can
show my boss regarding the subject.

Now, I am not looking for MS to answer directly, but thought one of the many
MVPs might have some insight or know of a KB, whitepaper or something to help
out. Thanks.

 

[Jim]

A little debate is going on here at work as too the best placement of the
pagefile on a Windows Xp system. Boss is looking for MS best Practice,
which
I read as move it to a different partition. He currently has the desktop
group apply this by creating a D: drive on the same partition and placing
it
there.

Our argument is over whether this actually provides any performance boots
over having it on the C: drive with the boot and system partition. Some of
us
think it is the same either way and you have to move it to another DRIVE
to
get any actual performance boost, but others insists the second partition
on
same drive works.

Goggling produces support for both sides, so does MS have anything to help
clear this up? If a separate partition in the same drive does work, could
someone explain why? Thanks!

Using a another partition on the same drive for the pagefile will work.
But, all I can see happening are addition head movement and additional
calculations by the driver.
Putting the pagefile on a different disk has the potential of improving
speed at the cost of increased probability of failure.
As for Best Practices--- HA
Jim

 

[Leonard Grey]

You may be thinking about this MS knowledgebase article:

"How to configure paging files for optimization and recovery in Windows XP"
http://support.microsoft.com/kb/314482/en-us

However, in practical terms this procedure makes no difference that you
can notice, as you have discovered. This is not surprising, given the
speed and reliability of today's hardware.

There may well be a specific software application which would benefit by
this procedure, but I cannot think of one off the top of my head.

 

[Ken Blake, MVP]

A little debate is going on here at work as too the best placement of the
pagefile on a Windows Xp system. Boss is looking for MS best Practice, which
I read as move it to a different partition. He currently has the desktop
group apply this by creating a D: drive on the same partition and placing it
there.

Our argument is over whether this actually provides any performance boots
over having it on the C: drive with the boot and system partition. Some of us
think it is the same either way and you have to move it to another DRIVE to
get any actual performance boost, but others insists the second partition on
same drive works.

Yes, it works. But it is *not* a good idea. Not only will it not
provide a performance boost, it may actually provide a performance
degradation.

What doing this does is move the page file to a location on the hard
drive distant from the other frequently-used data on the drive. The
result is that every time Windows needs to use the page file, the time
to get to it and back from it is increased.

Putting the page file on a second *physical* drive is a good idea,
since it decreases head movement, but not to a second partition on a
single drive. A good rule of thumb is that the page file should be on
the most-used partition of the least-used physical drive. For almost
everyone with a single drive, that's C:.

If you have enough RAM, the penalty for moving the page file to a
second partition may be slight, since you won't use the page file
much, but it will never help you.

 

[Dick W. Jonsson]

Hi!

There seems to be a lot of guessing going around.

I my self has read Bruce Sanderson's excellent article “RAM, Virtual Memory,
PageFile and all that stuff†( at:
http://members.shaw.ca/bsanders/WindowsGeneralWeb/RAMVirtualMemoryPageFileEtc.htm )

and in section 1.3 it clearly says “the operating system will move pages (4
KB pieces) of one or more virtual address spaces to the computer’s hard diskâ€

Now Please note that it´s 4KB pieces that is moved… It is important to my
next question/statement..

Does not Harddrives as of today have a relatively large “Cache�??

So… It your operative system is moving stuff to your swap-file in 4KB pieces
at the time..

Does not your harddrive store these 4KB pieces in it’s cache BEFORE it is
written to the physical drive??

And if so.. Does not that procedure equal out all the sk. unnecessary
movements of the read/write arms? (It will store a bunch of 4KB pieces and
write several at the time to the physical disk)

In short:
People who discuss page files and it’s placement seems to forget the 4KB
piece bit and the diskcache bit! 2 Things that makes a great difference in
real life..

So! in my mind it is not a bad idea to put the pagefile on a separate
partition on the same physical disk, considering above!
Especially when you get much less fragmentation on your windows partition..
Especially when your pagefile will not get fragmented.

Does anyone agree??

(Please excuse my English, I Swedish so English is not my native “mode†;0)

 

[Dick W. Jonsson]

Hi again!

I feel that I need to add that (in my opinion) it is always best to have the
main bulk of the pagefile on a separate physical drive than your windows is
installed to.. And preferably in a partition of its own..

And a thought on that “a partition in the beginning of the drive†talk:
And since your disk has a large cache it does not mater if you have your
pagefile on an partition that is in the beginning of the drive or not.


Why?
Since a modern harddrive is built up by several physical disks internally
you don’t have only one "beginning". You have, in fact, one "beginning" per
read/writable side of each physical disk.. (A modern harddrive is not a one
sided Disk? Is it?)

So if your disk has 2 double sided physical disks internally you will in
fact have 4 "beginnings".. (and 4 arms for reading/writing)

So.. In theory the optimum would be to have 4 equally large partitions on
the above disk, so that all 4 partitions subside on one side of each “diskâ€
only, and all 4 read/write-arms can write or read data at the same time?
(Right?)

;0)

 

[Gerry]

Dick

That's the first time I have seen Bruce's Article.

I found this part interesting:

"If the minimum and maximum pagefile sizes are not the same, the
pagefile will initially be allocated to the minimum size, then expanded
as more space is needed, up to the maximum. On most systems, this will
mean that the pagefile will be fragmented, possibly in widely separated
parts of the disk. If there is a significant amount of paging activity,
this can lead to unnecessary performance degradation. Setting the
minimum and maximum the same size may still result in a fragmented
pagefile, but does reduce the likelihood of serious fragmentation. If
the system is paging heavily and the pagefile is heavily fragmented,
consider moving the pagefile to a newly formatted partition on a
separate disk."

If the pagefile is in the Windows partition is a non-contiguous pagefile
it will tend to increase the rate of fragmentation of other files
located in the partitition. The point also goes against the commonly
expressed opinion that it's best to let to let Windows manage the
pagefile. My present setting is a min/max 2 gb page file in the Windows
partition but I have used the first partition on a second drive for a
dedicated pagefile partition in the past. The only reason for the change
was that I was having difficulties getting a dual boot set up to work
when I created my latest set up.



~~~~


Gerry
~~~~
FCA
Stourport, England
Enquire, plan and execute
~~~~~~~~~~~~~~~~~~~