physical/virtual memory question....

[U2Penn]

on my system my total physical memory is 512 megs with 254 megs available. total virtual memory is 1.72 gig with 1.23 gig available + page file space is 1.22 gig. now the page file space is the place on my hard drive to store memory, right? does it only story memory on my page file when u run out of physical memory, right? plus, virtual memory is the combination of page file and total physical memory, right?

thanks in advance.

 

[Perris Calderon]

u2...there are lots of htreads concerning the pagefile, and virtual memory...lots of fun reading if you want to do a search.

the pagefile is basically a big storage container, which is mostly empty, but it iss always available to put information if you are ever in need of more ram then you have installed.

here's an interesting fact...without the pagefile taking addresses for ram which is asked for but never accessed, you would need a coupla gigs of ram to keep running

some people argue that the concept "virtual memory"involves all the backing stores, backup files plus the actual memory installed.

others will argue that virtual memory does not include acual ram.

no matter.

simply out, for most users, let xp manage the pagefile, and you can't go wrong.

for power users, you should increase the initial minimum to 2x ram or more depending on your needs, and leave the maximum system managed

 

[dave holbon]

You can go wrong if you don't undestand how the swapfile works especially if you have a small hard drive and 2gig of ram.

 

[Perris Calderon]

Originally posted by Dave Holbon
You can go wrong if you don't undestand how the swapfile works especially if you have a small hard drive and 2gig of ram.

as usual dave, you are correct...(I hate when I'm corrected)

but, to take this a little further...xp will actually downsize the pagefile if you are short on hardrive space, and it will actually try to get by on as small a pagefile as possible

this is only if you let the system manage the memory in it's entirety...obviously, if you set the initial minimum, the os will not make a pagefile smaller then your set initial minimum

now, you are correct, if the system has just enough hardrive space to think it can allot a sufficient pagefile, but your intentions are to use alot of this space immediately.

in this case, you would suffer a slow down when the os deminished the size of the pf to fit the hardrive restrictions

 

[dreamliner77]

You can go wrong if you don't undestand how the swapfile works especially if you have a small hard drive and 2gig of ram.

How many people will have 2 gig of ram and a harddrive so small as to not be able to have an adequte pagefile?

 

[U2Penn]

thanks for the info i was a little confused because i thought the page file was supose to be completely empty UNTIL the physical memory was full....then spilled into the page file. the numbers didn't add up right because there was a little bit of space used up in my page file.

 

[Perris Calderon]

u2...your pagefile is almost empty until you are near out of ram

the activity before is like priming a hose before you need to water the lawn...it's one of the things that speed xp

what you see in your monitor is pagefile activity...it is not pagefile use...if you want to see just how much is actually in the pagefile, download the zip that I post here

 

[U2Penn]

Dealer,

i downloaded the zip and this is what came up for me:
current page usage 87 mb
session peak usage 114mb
current page file size 768mb

 

[Perris Calderon]

now, if you look at perfmon, you will see a couple of hundred mbs of pagefile in use, even though you have less then 100 mbs written

the extra mbs of the pf useage is potential.

now, of the ram that's actually written...most of that information is in both...pf and the ram it came from.

until of course, the ram it came from is claimed by another process.

very efficient

your pf size needs to reflect the amount of ram you actually have, so it can be ready with the potential use.

it doesn't need to be the size it actually uses...that will slow you down.

this is why the more ram you have the bigger the pf needs to be, not the smaller.

any, good to meet you.

gone for the day

 

[gothic]

I have one or two 'forensic' type utilties, one of which is able to decode certain areas of the pagefile.sys. There are certain datii that are automatically written there in preference to RAM (i.e. explorer history/typed URLs etc) which can easily be accessed using such freely available utilities. Takes a little time, mind...depends on the size of the pagefile!!

TOK

 

[B~B]

Originally posted by dealer now, if you look at perfmon, you will see a couple of hundred mbs of pagefile in use, even though you have less then 100 mbs written

the extra mbs of the pf useage is potential.

now, of the ram that's actually written...most of that information is in both...pf and the ram it came from.

until of course, the ram it came from is claimed by another process.

very efficient

your pf size needs to reflect the amount of ram you actually have, so it can be ready with the potential use.

it doesn't need to be the size it actually uses...that will slow you down.

this is why the more ram you have the bigger the pf needs to be, not the smaller.

Thanks for the explanation.

 

[Zedric]

Windows will swap mamory out to the page file even if you're not out of RAM. This is because if you suddenly need alot of RAM, Windows won't have to swap other stuff out first. It can just claim the space.

 

[Perris Calderon]

xp doesn't do a swap in the situation you're talking about zedrick

it will write to the pagefile before you're out of ram, but it also keeps that code in physical in memory at the same time and it stays there untill it's claimed

this is the sweetest part of the xp memory management

 

[Zedric]

Originally posted by dealer
it will write to the pagefile before you're out of ram, but it also keeps that code in physical in memory at the same time and it stays there untill it's claimed

I know. I didn't say it did or didn't because you allready had. Maybe I was too vague.

Deleting the data from RAM would make the method pretty useless.

 

[Perris Calderon]

Deleting the data from RAM would make the method pretty useless.

well said

jeeze...it tales me a whole thread to say what yoiu just said in a sentence

ah me