Swap File Size

  • Thread starter Thread starter Just D.
  • Start date Start date
As for kb237740,actually that information only holds true if the pc is in a
very low performance set-up...Even microsoft gives other examples that would
be/are more benificial toward pcs that are set up with a bit more performance
orientaded machine (kb555223)..Going further,microsoft also gives more
performance set-ups than is in 555223 (TechNet/library)..Going even further,
Intel gives system registry settings that almost disable the pagingexec,which
on any pc gives more performance & less hd problems.Again,microsoft goes
along with intel also,& with those settings,kb237740 is an outdated,low
performance info that should be eliminated...
 
I was not talking about performance. Having a full load of the development
stuff working simultaneously on a 4GB machine (and it's huge, trust me!) I
almost never get the 2GB swap file used above 420 MB. But I wanted to pay
attention to the swap file size and limitations. I always suspected that
advices to make swap huge, like 1.5-2.5 of RAM size, say 5-6 GB, are
incorrect, especially on the 32-bit Windows machines with 4 GB addressing.
Last time I heard this advice in this newsgroup about a week ago. The
article shows that my suspect was reasonable. Although there are tricks to
avoid this limitation, some kind of way around. :) There is also a very nice
trick to create a swap file in the subdirectory. Btw, I'd wish to know how
could I move the hiberfil.sys to another location than the C:\.

Just D.
 
Just D. said:
I was not talking about performance. Having a full load of the development
stuff working simultaneously on a 4GB machine (and it's huge, trust me!) I
almost never get the 2GB swap file used above 420 MB. But I wanted to pay
attention to the swap file size and limitations. I always suspected that
advices to make swap huge, like 1.5-2.5 of RAM size, say 5-6 GB, are
incorrect, especially on the 32-bit Windows machines with 4 GB addressing.
Last time I heard this advice in this newsgroup about a week ago. The
article shows that my suspect was reasonable. Although there are tricks to
avoid this limitation, some kind of way around. :) There is also a very nice
trick to create a swap file in the subdirectory. Btw, I'd wish to know how
could I move the hiberfil.sys to another location than the C:\.

Just D.

You can't change the location of that file.
 
Just said:
Btw, I'd wish to know how could I move the hiberfil.sys
to another location than the C:\.

It can't be done, the file *must* reside in the root of the system
partition alongside the ntldr, NTDETECT.COM and boot.ini file. This is
due to file system driver limitations during the early stages of the
boot process, at this early stage the system relies on the ntldr mini
file system driver and the file can only be loaded if it is in the root
folder of the system partition.

John
 
John said:
It can't be done, the file *must* reside in the root of the system
partition alongside the ntldr, NTDETECT.COM and boot.ini file. This is
due to file system driver limitations during the early stages of the
boot process, at this early stage the system relies on the ntldr mini
file system driver and the file can only be loaded if it is in the root
folder of the system partition.

PS: The hyberfil.sys will be on the same volume as the /Windows folder
(the boot volume), that is where the operating system will create the
large memory image file. But, the file hyberfil.sys *must* also be in
the root of the System partition alongside ntldr, that is the only place
where ntldr will look for it. The hyberfil.sys file on the system
partition will either contain the actual memory image, (when the boot
and system partitions are one and the same) or it will contains the ARC
path to the boot partition of the last operating system that entered
hibernation (when the boot volume is on a different partition than the
System partition).

John
 
it's interesting.

however, there is a correlation between
ram and virtual mem.

the lower the ram the more dependent
the system becomes for virtual memory

the higher the ram the less dependent
the system becomes for virtual memory.

but it is not to say that scenerios involving
high ram and high virtual memory doesn't
exists.

-------------

also vista is not confined by the limitations
of xp, in this respect.

for its initial size should = (ram)+(300) megs
and the max should = (ram) x (3).

however, I have found on my particular machine
that setting initial to 16 and max to 3 x ram
works better than the standard above.

thanks for the kb.


--
db·´¯`·...¸><)))º>
DatabaseBen, Retired Professional
- Systems Analyst
- Database Developer
- Accountancy
- Veteran of the Armed Forces
- @Hotmail.com

"share the nirvana mann" - dbZen

~~~~~~~~~~~~~~~
 
John John - MVP said:
PS: The hyberfil.sys will be on the same volume as the /Windows folder
(the boot volume), that is where the operating system will create the
large memory image file. But, the file hyberfil.sys *must* also be in
the root of the System partition alongside ntldr, that is the only place
where ntldr will look for it. The hyberfil.sys file on the system
partition will either contain the actual memory image, (when the boot
and system partitions are one and the same) or it will contains the ARC
path to the boot partition of the last operating system that entered
hibernation (when the boot volume is on a different partition than the
System partition).

John

Woulda been simpler to have just stopped after "It can't be done".
 
Bennett said:
Woulda been simpler to have just stopped after "It can't be done".

Some people want to know "why". If you already know, or if you don't
want to know, you can just move on, others will do as they please.

John
 
John

Most people SHOULD want to know why! A lot of "crap" is posted. A test
of the value of the advice is being able to understand why and what is
being recommended.

--


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

"also vista is not confined by the limitations of xp, in this respect."

The address space limitation applies to 32 bit versions of Vista!

--


Gerry
~~~~
FCA
Stourport, England
Enquire, plan and execute
~~~~~~~~~~~~~~~~~~~
 
db said:
also vista is not confined by the limitations
of xp, in this respect.

Well, XP supports paging files of 16TB in size, just like Vista. The
maximum size depends on what kernel is loaded.
 
Just D. said:
I was not talking about performance. Having a full load of the
development stuff working simultaneously on a 4GB machine (and it's
huge, trust me!) I almost never get the 2GB swap file used above 420
MB. But I wanted to pay attention to the swap file size and
limitations. I always suspected that advices to make swap huge, like
1.5-2.5 of RAM size, say 5-6 GB, are incorrect, especially on the
32-bit Windows machines with 4 GB addressing. Last time I heard this
advice in this newsgroup about a week ago. The article shows that my
suspect was reasonable. Although there are tricks to avoid this
limitation, some kind of way around. :) There is also a very nice
trick to create a swap file in the subdirectory. Btw, I'd wish to
know how could I move the hiberfil.sys to another location than the
C:\.
Just D.

You are correct, of course. And in XP swap file is called page file or
pf for short.
The numbers you mention simply indicate that you have very little
page file activity which I'd expect with either 3 or 4 Gig of RAM
installed. In general, the more RAM you have installed, up to 3 Gig
plus a few hundred Meg, the less you will need the pagefile. That's
precisely one of the reasons to let Windows manage the pf size; there's
no sense in wasting Gigs of drive space with preset sizes when you are
seldom or never going to use them.
If you're using letting windows manage the pf size, then should it
ever be needed, it will be allocated for you for that instance. Unless
you're doing some very serious large database calculation work, image or
video editing and rendering, it's unlikely you'll see the pf get used
for much more than you see after bootup. Under windows managed size it
idles around a few hundred Meg (abt 300 in my case) in case it has to do
memory dumps and a couple other minor uses that hopefully will never
happen<g> but seem to now and then.

The 1.5 x RAM is a simply starting point by MS for what windows will
pick for swap file size, but that does not mean that much space will be
USED. IT's almost a random number picked by MS for a machine using 512
Meg of RAM. Other than that, most anything you've seen or read is myth
and bad interpretations. These people say it better than I can and
illustrate exactly what I determined myself:

A good explanation:
http://soft4all.info/how-to/how-to-use-virtual-memory-in-windows-xp/

Tweaks that don't work:
http://www.pallab.net/2005/09/23/windows_xp_tweaks_that_do_not_work/
See the para re 1.5 x RAM; good explanation

Here's a decent, all around discussion of things:
http://www.aumha.org/win5/a/xpvm.php

Some Background:
http://home.comcast.net/~SupportCD/XPMyths.html

A discussion:
http://www.answerbag.com/q_view/89001



Although you're probably aware, but in case you aren't:

XP is capable of addressing up to 4 Gig of different addresses. An
"address" is like a house number in a city wih a population of 4 Gig
people: Everything in the computer is literall located at an address
that can be accessed.

About 2/3 Gig is used for hardware and the OS management.

That leaves about 3.3 Gig of Address Spaces left over (depending on the
computer).

So if 3 Gig of RAM is installed, all 3 Gig of it can be addressed
(used).

With 4 Gig, only about another 1/4 Gig of address space is left, so
around 3/4 Gig of the 4th Gig of RAM cannot be addressed (used) by the
operating system. It's wasted. There is virtually no difference in pf
use between a 3 Gig and a 4 Gig RAM load.

HTH,

Twayne`
 
Bennett Marco said:
Woulda been simpler to have just stopped after "It can't be done".

FYI:

Right. If they wanted to know why, they can/could have ask/asked why.
What you're really seeing is a few MVPs racking up as many "helpful"
responses as they can, because getting their MVP status renewed for
another year requires a lot of online participation. This way they
don't have to do or know anything to claim a post for the "elders" to
see.
As a result, there are a LOT of "me too" posts around here; it's not
just this post.
Add to that a few trolls and others feeling powerless in their lives,
waiting around for something to say something they disagree with or just
want to try to control someone, and you have another group that does the
"me too" stuff. AFter awhile they start to come off as mindless sock
puppets and nothing more.
And then of course there is our resident imposter, PCBUTTS and other
names he goes by, including those of good posters, and you have the full
gambit.

Twayne`
 
Gerry said:
John

Most people SHOULD want to know why! A lot of "crap" is posted. A test
of the value of the advice is being able to understand why and what is
being recommended.

Actually, the more important thing is to verify and clarify an answer
that seems plausible from other sources if it has any importance to the
querant. That needs to be done for anyone getting information from
anywhere on the 'net, period. I often don't ask why unless it's
important to me, since when I go elsewhere to verify it, I'll find out
why anyway.
OTOH it's nice if a poster does say why something is so, if he
actually knows it for a current fact.

HTH,

Twayne`
 
You're a real idiot, Twayne.
FYI:

Right. If they wanted to know why, they can/could have ask/asked why.
What you're really seeing is a few MVPs racking up as many "helpful"
responses as they can, because getting their MVP status renewed for
another year requires a lot of online participation. This way they
don't have to do or know anything to claim a post for the "elders" to
see.
As a result, there are a LOT of "me too" posts around here; it's not
just this post.
Add to that a few trolls and others feeling powerless in their lives,
waiting around for something to say something they disagree with or just
want to try to control someone, and you have another group that does the
"me too" stuff. AFter awhile they start to come off as mindless sock
puppets and nothing more.
And then of course there is our resident imposter, PCBUTTS and other
names he goes by, including those of good posters, and you have the full
gambit.

Twayne`
 
Twayne

So you would agree, on the basis of what you now say, that the "crap"
you have written about Registry Cleaners is a lot of nonsense. You
being, one of a small minority, who regularly advocate their use.

--


Gerry
~~~~
FCA
Stourport, England
Enquire, plan and execute
~~~~~~~~~~~~~~~~~~~
 
Gerry said:
Twayne

So you would agree, on the basis of what you now say, that the "crap"
you have written about Registry Cleaners is a lot of nonsense. You
being, one of a small minority, who regularly advocate their use.

Nope, not at all. You try to insult me, but I don't worry about the
small stuff like you. I've written all factual and correct informatoin
about registry cleaners. The only ones I disagree with are those who
say they are "snakeoil", "never any good", "cannot do anything useful",
"will damage your computer" and similar statements made by some
egotistic narcissists around here. Because I say a registry cleaner CAN
be useful does not mean one should run it every day, and I've never said
any such thing, but you could run one every day, and other than wasting
your time, you would never see any downsides to it as long as it was a
reputable, reliable and properly sourced program.
Anyone that says to NEVER use a registry cleaner because it WILL
damage your system, for the stupid reasons they give, without a drop of
backup, is a fool, and I could make a long sentence out of this, but I
won't.

I have often used registry cleaners for various taks: Most of the time
they accomplish nothing noticeable, the other times they accomplish a
good fix for a problem, and if nothing else, when the TSing leads you to
the registry, they're good for the process of elimination. Of the
hundreds of machines and the many hundreds of times I've used a good
registry cleaner, I have never encountered so much as one event of
damage or detriment to an operating system. And if that did happen, I'd
be able to put the "fix" back and correct the erroneous "fix" that way.
But like I said, it's never happened to me or anyone I respect or have
seen occur.
IME good registry cleaners are more robust than most programs, better
tested and have not, for instance, done any of the damage that Microsoft
itself has done to their own operating system over the years. Partly
because of the robustness of the XP registry heirarchy, and partly
because I know how to choose a useful product, as most any thinking
person can do.

Unless and until someone provides verifiable, repeatable evidence
otherwise, I will continue to believe that those who say to never use
such a thing and how much damage it will cause are simpletons of
yesterday who can/will not open their eyes to see what's right in front
of them.

lol, sorry! I've done a lot of work with registry cleaners and their
breathern, and have considerable positive experience with them. The
ones I use mostly (three of them) are excellent programs and no more
likely to cause any damage to a system than any other program on a
computer. Stuff DOES happen, and eventually I'm sure a negative outcome
WILL result from a registry cleaner, but when it does happen, it will be
no surprise since at ANY time windows can do silly things, "stuff"
happens, and it's silly to say always or never in any continuous,
without exception way because there is no such solidarity in the
universe. THAT is my objection to the misinformationists.
If I say never or always I attach a timeframe to it for occurring in
the past experiences, I don't encompass all time in the universe for
every living person on the earth as the misinfomationists are wont to
do. And if tomorrow the opposite happens to me, it would interest me,
but it would not cause such fear in me as it seems to do in these few
MVP misinformationists and their sock puppets they've been weaning.
What is entertaining about them is the seriousness of which they've
trapped themselves into believing their own tripe and it's even
entertaining to watch them wiggle when someone has the audacity to
actually say their information isn't correct. <g>

Cheers,

Twayne`
 
Back
Top