PageFileUsageMonitor

[Howard Kaikow]

Lots of folkes have looked at PageFileUsageMonitor (http://www.standards.com/index.html?PageFileUsageMonitor), I expect that a lot of this forum's participants are among that number.

So, I figured that I'd inform you of two things:

1. I've added a capability that I stupidly left out of the earlier releases. Not a bug, just a stupid omission on my part. This will make the program useful to a lot more folkes.

2. I've tested and prog works under Vista.

[Someone]

-{ Quote: "Lots of folkes have looked at PageFileUsageMonitor (http://www.standards.com/index.html?PageFileUsageMonitor), I expect that a lot of this forum's participants are among that number.

So, I figured that I'd inform you of two things:

1. I've added a capability that I stupidly left out of the earlier releases. Not a bug, just a stupid omission on my part. This will make the program useful to a lot more folkes.

2. I've tested and prog works under Vista." }-

Hi

I dont know much about this stuff. What's the point of of a Page File Monitor?

Thanks

[Howard Kaikow]

-{ Quote: "Hi

I dont know much about this stuff. What's the point of of a Page File Monitor?

Thanks" }-

Determining appropriate sizes for page files.

[Rmus]

Hello Howard,

Very nice Interface! I've used Bill James' utility but I like having the current time displayed in the Interface. One nickpick would be to have the new entries in the Log file appended at the top, rather than at the bottom.

@Someone: many people like to tweak their Pagefile for speed, move it to different partitions, etc. , so a utility like this is helpful to monitor what is going on behind the scenes.

(For everyone who writes about tweaking, you find a different opinion!)

My interest was for monitoring the Peak Usage so as to determine the size of the PageFile.

Some people think that the PageFile only is used when Windows "swaps" to disk when free RAM is used up. Using such a Monitor proves this to be incorrect.

Here's how it worked for me:

Using Win2K and a small C:\ partition of 4GB, I wanted as small a PageFile as possible. My normal usage with about 5 programs running in the TaskBar is 35 - 50MB. Opening PhotoShop and MSPublisher, usage jumps to 70 - 80MB and if I print from MSPublisher, it jumps to 130-140MB.

I also have a RAM usage meter, and my free RAM has never depleted during a session.

Based on this, I set a Minimum PageFile size = 500MB.

I think Win2K does a great job of managing resources. I've never encountered any problems such as slowing down, etc. It's a huge improvement over Win9x.

Using DeepFreeze, the PageFile is restored to original on each reboot. (this can also be accomplished by setting Windows to clear the pagefile on shutdown).

So, a Monitor like Howard's is a useful utility.


----
rich

[Someone]

-{ Quote: "Determining appropriate sizes for page files." }-

OK. Thanks.

[Someone]

-{ Quote: "@Someone: many people like to tweak their Pagefile for speed, move it to different partitions, etc. , so a utility like this is helpful to monitor what is going on behind the scenes. " }-

How do you move it to different partitions? Can you just cut and paste? Does everything still work?

-{ Quote: " Using Win2K and a small C:\ partition of 4GB, I wanted as small a PageFile as possible. My normal usage with about 5 programs running in the TaskBar is 35 - 50MB. Opening PhotoShop and MSPublisher, usage jumps to 70 - 80MB and if I print from MSPublisher, it jumps to 130-140MB.

I also have a RAM usage meter, and my free RAM has never depleted during a session.

Based on this, I set a Minimum PageFile size = 500MB.

----
rich" }-

Why is it 500MB? I don't get it.

Thanks

[MrBrian]

Sizing the Page File (using Windows built-in tools) (http://www.theeldergeek.com/sizing_the_page_file.htm)

[Howard Kaikow]

See http://support.microsoft.com/kb/314482/, and links within that article.

[Someone]

But if it 140mb times 1.5, wouldn't it be 210? I don't get the 500mb.

[gerardwil]

True or not true?

A larger pagefile speeds up defragging.

Gerard

[Howard Kaikow]

-{ Quote: "But if it 140mb times 1.5, wouldn't it be 210? I don't get the 500mb." }-


The archaic 1.5 "rule" applies to the systems physical memory, which ordinarily results in excessive page file sizes.

The 140MB referred to supa, is the amount ofpagefile used when the poster was using, as I recall, a particular program.

so the poster chose 500MB as best for their system.

I use:

C drive: fixed size 64MB page file.
F drive: variable-size 128MB to 1088MB pagefile

During my current windows session:

Pagefile on C had maximum use of 62MB
Pagefile on F had maximum use of 132MB with 152MB pagefile. This is largey due to Acronis True Image.

[Howard Kaikow]

-{ Quote: "True or not true?

A larger pagefile speeds up defragging.

Gerard" }-

Unless a pagefile is of fixed-size, it worsens fragmentation

[Rmus]

-{ Quote: "But if it 140mb times 1.5, wouldn't it be 210? I don't get the 500mb." }-Hi Someone,

I just decided to give it a bit more headroom. No other particular reason!

Frankly, in experimenting with different possibilities, I never saw any difference in my performance! I just wanted to keep it smaller than the 1.5 "rule" Howard refers to. Using the 1.5 rule would dictate a huge PageFile size on my computer!

One thing I've always done is "test" rules. No two systems are alike.

I think this comes from my photography work, where I was taught, "Careful photographers run their own tests."



----
rich

[ThunderZ]

Guess you could say I cheat. Have 4 Gigs of ram on 32 bit XP. 3.25 recognized. I run no pagefile. Seemed to get a bit of a boost in overall performance. Have not noticed any negative affects. May also help to extend the hard drives life.

Just thought I would toss this out as another option.

[Howard Kaikow]

-{ Quote: "
Frankly, in experimenting with different possibilities, I never saw any difference in my performance! I just wanted to keep it smaller than the 1.5 "rule" Howard refers to. Using the 1.5 rule would dictate a huge PageFile size on my computer!
" }-A smaller page file can increase performance as it makes more space available for files, thus reducing fragmentation

[Howard Kaikow]

-{ Quote: "Guess you could say I cheat. Have 4 Gigs of ram on 32 bit XP. 3.25 recognized. I run no pagefile. Seemed to get a bit of a boost in overall performance. Have not noticed any negative affects. May also help to extend the hard drives life.

Just thought I would toss this out as another option." }-

With the newer OS, one does not actually need a pagefile, BUTT, there's my big BUTT again:

1. When booting, the OS will create a small pagefile in memory.
2. I'm not sure what happens, if you do not have a page file, and the OS/app needs one. I would guess that could cause memory to be gobbled up.

I'd rather have smallish page files to reduce the chance of this happening.
PageFileUsage Momitor helps you determine the appropriate size(s)

I suggest a small fixed-size page file on C, and variable size on other PHYSICAL drives, preferably not the OS drive. Each with a minimu size that reduces the chance that the OS will ask for more.

[ThunderZ]

A slight addition to your suggestion of a small pagefile on a separate drive. Put it on a different pata\sata connection then your OS drive. ;)

[Someone]

-{ Quote: "The archaic 1.5 "rule" applies to the systems physical memory, which ordinarily results in excessive page file sizes.

The 140MB referred to supa, is the amount ofpagefile used when the poster was using, as I recall, a particular program.

so the poster chose 500MB as best for their system.

I use:

C drive: fixed size 64MB page file.
F drive: variable-size 128MB to 1088MB pagefile

During my current windows session:

Pagefile on C had maximum use of 62MB
Pagefile on F had maximum use of 132MB with 152MB pagefile. This is largey due to Acronis True Image." }-

Hi

Thanks for the explanation. But I don't get the part about going to 1088mb. Why does it need to be so big?

Thanks

[Howard Kaikow]

-{ Quote: "A slight addition to your suggestion of a small pagefile on a separate drive. Put it on a different pata\sata connection then your OS drive. ;)" }-

The small pagefile should be on the C drive, because that's where th OS first looks for it.

In my case, I have 3 SCSI hard drives (a 4th with honly Linix) with 4 OS.
My most used OS is on drive 3, so I put the small page file on C on drive 1, and the variable size on the 2nd drive.

When I boot to an OS on the second drive, I sometimes see a difference in speed due to the presence of the pagefile.

[Howard Kaikow]

-{ Quote: "Hi

Thanks for the explanation. But I don't get the part about going to 1088mb. Why does it need to be so big?

Thanks" }-

THe variale size page file allows for growth large enough to be 1,5 times memory.

I rarely see the pagefile grow above 200MB.

[Someone]

-{ Quote: "THe variale size page file allows for growth large enough to be 1,5 times memory.

I rarely see the pagefile grow above 200MB." }-

But then wouldn't it be 300?

Thanks

[Nick Rhodes]

-{ Quote: "A smaller page file can increase performance as it makes more space available for files, thus reducing fragmentation" }-

Only if you are running low on contiguous frees-pace to store extents.
Difference is negligible on current HDD sizes.

-{ Quote: "Unless a pagefile is of fixed-size, it worsens fragmentation" }-

Only if you exceed the current size of pagefile. How often does this happen ?
Setting a sensibly sized minimum (which WinXP does), helps reduce this to the point of not worth worrying about for the average person.

-{ Quote: "With the newer OS, one does not actually need a pagefile, BUTT, there's my big BUTT again:

1. When booting, the OS will create a small pagefile in memory.
2. I'm not sure what happens, if you do not have a page file, and the OS/app needs one. I would guess that could cause memory to be gobbled up.


I'd rather have smallish page files to reduce the chance of this happening.
PageFileUsage Momitor helps you determine the appropriate size(s)
" }-

Machine can freeze or BSOD if you had no pagefile, this isWindows always creates a pagefile, even if you think you have totally turned it off !

-{ Quote: "
I suggest a small fixed-size page file on C, and variable size on other PHYSICAL drives, preferably not the OS drive. Each with a minimu size that reduces the chance that the OS will ask for more." }-

That is what I do, Windows is smart enough to pick the least active drive at point of paging out.

You also want to place it on the most active partition (of whichever drive) to reduce head seeking.

[ThunderZ]

-{ Quote: "The small pagefile should be on the C drive, because that's where th OS first looks for it." }-

Until I built my current system with 4 Gigs. I never experimented much with the pagefile so I am truly a babe in the woods on all the facts. ??? But if you set your page file to a different drive from within the OS then are`t you telling the OS where to look for it therefor negating the very brief search time?
Have read discussions on the same topic with different opinions. Some say having it on the same drive as the OS slows things down. Extra head movement is needed to read\write from the same drive if the pagefile and OS are located there. Compared to independent\separate head movement of 2 different drives. The same discussions held true if 2 drives were on the same cable\channel concerning data transfer. This would seem to hold true if only 1 drive was used as well.

[Howard Kaikow]

-{ Quote: "But then wouldn't it be 300?

Thanks" }-

No.

64 + 1088 = 1152 = 1.5*768

[Howard Kaikow]

-{ Quote: "Only if you are running low on contiguous frees-pace to store extents.
Difference is negligible on current HDD sizes.
" }-

In my case, the C-D drives are on a 9GB SCSI drive with little space unused.
Gotta choose a small fixed-size file.
I found that 64MB works my system.


-{ Quote: "
Only if you exceed the current size of pagefile. How often does this happen ?
Setting a sensibly sized minimum (which WinXP does), helps reduce this to the point of not worth worrying about for the average person.
" }-

JUst about every day. Acronis TRue Image 9 is a page file hog.
I used to get by with a 64MB minimum, then I started using TI.
I've been thinking of raising the minimum to 256MB.


-{ Quote: "
Machine can freeze or BSOD if you had no pagefile, this isWindows always creates a pagefile, even if you think you have totally turned it off !
" }-

Ayup.


-{ Quote: "That is what I do, Windows is smart enough to pick the least active drive at point of paging out.
" }-

Ayup.

-{ Quote: "You also want to place it on the most active partition (of whichever drive) to reduce head seeking." }-

In my case: 10 logical drives on 3 SCSI drives.

Drive 1: C-D, OS on C
Drive 2: F-H, OS on each of F and G
Drive 3: I-M, OS on J

J is the main OS.
64MB pagefile on C.
128-1088MB pagefile on F.

I boot to C, F, and G ONLY when I need to test something, or compile VB with the version of Office installed in that OS, or run Windows/Office update.

[Howard Kaikow]

-{ Quote: "Until I built my current system with 4 Gigs. I never experimented much with the pagefile so I am truly a babe in the woods on all the facts. ??? But if you set your page file to a different drive from within the OS then are`t you telling the OS where to look for it therefor negating the very brief search time?
Have read discussions on the same topic with different opinions. Some say having it on the same drive as the OS slows things down. Extra head movement is needed to read\write from the same drive if the pagefile and OS are located there. Compared to independent\separate head movement of 2 different drives. The same discussions held true if 2 drives were on the same cable\channel concerning data transfer. This would seem to hold true if only 1 drive was used as well." }-


The OS knows where you put the page files.
The reason for having the page files on a different PHYSICAL drive than that of the running OS is to reduce head movement.

Head movement is a killer. THe fewer the better.
For example, you could use ReadFile (http://www.standards.com/index.html?ReadFile) to see what a difference fewer accesses can make in reading a single file. Not to mention how much faster reading a defragged file will be.

[ThunderZ]

-{ Quote: "The OS knows where you put the page files.
The reason for having the page files on a different PHYSICAL drive than that of the running OS is to reduce head movement.

Head movement is a killer. THe fewer the better.
For example, you could use ReadFile (http://www.standards.com/index.html?ReadFile) to see what a difference fewer accesses can make in reading a single file. Not to mention how much faster reading a defragged file will be." }-


After reading your last post to nickr I completely understand your configuration. It would seem to go along with my reading and understanding of an "tweaked" pagefile set-up.

[Someone]

-{ Quote: "No.

64 + 1088 = 1152 = 1.5*768" }-

But where did the 768 come from?

Sorry, I don't know anything about this. :-\ :'(

[Howard Kaikow]

-{ Quote: "But where did the 768 come from?

Sorry, I don't know anything about this. :-\ :'(" }-

768 is the amount of physical memory.
In this case, the max permitted by the SE440BX mobo.