A Keynesian metaphor on "Do we need a pagefile when running with lots of RAM?"

Joined
Mar 17, 2009
Messages
30
Of course, when we say "lots of RAM" in 32bit OS'es, 4GB is the theoretical limit, but more often it's about 3.5 GB, since the system "reserves" (or loses track of) a bunch of these bytes... Some people say, rather sarcastically, that it could be worse, as we could be running 16bit OS'es and it could be raining outside.

Yet, since every little app can do its own math and claim its 4GB "share", regardless of how much physical memory is really available (or recognized, it ain't matter) the Microsoft whizzes had to come up with a way to control this madness. Surprisingly it came also at the same time that Bill Clinton said "EVERY BODY DESERVES 4 GB OF RAM", so instead of making real technology (or waiting for it to come) and make lots of real memory available to the apps, the fantastic idea they came up with was letting an application see more memory (what they call "Virtual Memory") by means of drugging the system with heavy narcotics, LSD, and other substances and make it believe it can "allocate" and spend as much memory resources as it wants...

<interlude sarcasmLevel="high">

... something similar to what the Federal Reserve did with the interest rates, then the banks giving mortgage credit to insolvent people, which it ain't matter, because they were only interested in collecting interest for a couple of months until another couple of insolvent people came and bought it for a higher price, thus repeating the theoretically never ending profiting-cycle (ending only when a rich guy with real money would buy the house and pay no interest.... aeww, God forbid that!) in conjunction with the investment banks which saw this "smart idea" as fantastic and decided to buy bags of these "money-making" debts and sell it in the derivatives market, for even more profit... and of course, insurance companies like AIG which covered the ("unthinkable") risk of this colossal amount of debt not being able to generate more money... Of course, it's obviously "unthinkable", for if your food costs $20 and you have -$40 (i.e., you don't have any money, and not only that, but you owe $40 to other people) then you can pay for it simply by multiplying the money in your possession (the -$40 that you have in your pocket) x -0.5 and voilà (-$40) x (-0.5) = $20..... Oh, but I forgot! Math doesn't make sense in the real world! And most importantly, there is a real "solution", one that doesn't require you to go to university: SIMPLY ASK FOR MORE CREDIT! :D :D

</interlude>

I want to hear your opinions. Cheers!
 
If you don't understand the underlying premise and functionality of the memory hierarchy, what makes you so confident in disabling a key component of it?

My advice: Regardless of the amount of RAM your system possesses, always leave the page file on.
 
Seriously, the 4GB virtual address space belongs to the creator of the CPU.
 
:p

Nah, don't take anything I say hard. I didn't sleep last night so anything's bound to happen. Apologies. :):eek:

Man, there must be something wrong with me. If I went a night without sleep, I'd crash and burn so bad it wouldn't even be funny. I think I'd like spontaneously combust.
 
Considering all that has been written and discussed on this topic in this very forum, I'm struggling to see the point of this thread.
 
^^ bingo leave it alone, since some items dont play nice if it is disabled and with harddrives so huge who cares if the OS takes a gig or 2 or 3 for some space.
 
Simple people are easy to make laugh... ignorant ones too, go figure, especially when reading comprehension is simply lacking. But you keep right on twisting my comments to suit your little purposes... that's fine, as others know better.
 
^^ bingo leave it alone, since some items dont play nice if it is disabled and with harddrives so huge who cares if the OS takes a gig or 2 or 3 for some space.

Yeah, you see, I totally get that the space is not the problem. IT HAS NEVER BEEN MY ISSUE (well, back in the day it definitely was...) What I mean is that to map a page file into memory, it needs to access the disk in which it is located, and if I have a database file in THAT disk, no matter where on earth the bloody DB file is a HORRENDOUS change in position in the disk head is going to occur.

I simply find it ridiculous that instead of having my software (some of it my own) load into real memory (and keep it there) all the stuff it needs to RUN (as in "use those cycles/sec, a.k.a. Hertz" of the processor) and while I'm trying to have my system tuned for data access performance (as much as I can) and with several real (and big) DBs in my disk, my OS decides to go visit the pagefile, just to map a few hundred KB of data into memory..... To me, and to Camus, that's absurd!
 
Yeah, you see, I totally get that the space is not the problem. IT HAS NEVER BEEN MY ISSUE (well, back in the day it definitely was...) What I mean is that to map a page file into memory, it needs to access the disk in which it is located, and if I have a database file in THAT disk, no matter where on earth the bloody DB file is a HORRENDOUS change in position in the disk head is going to occur.

I simply find it ridiculous that instead of having my software (some of it my own) load into real memory (and keep it there) all the stuff it needs to RUN (as in "use those cycles/sec, a.k.a. Hertz" of the processor) and while I'm trying to have my system tuned for data access performance (as much as I can) and with several real (and big) DBs in my disk, my OS decides to go visit the pagefile, just to map a few hundred KB of data into memory..... To me, and to Camus, that's absurd!

Put the pagefile or your DB on a separate disk.
 
Man, there must be something wrong with me. If I went a night without sleep, I'd crash and burn so bad it wouldn't even be funny. I think I'd like spontaneously combust.

dont take compsci 2401 at langara...

anyways, yeah, I'm still running on 604k memory: If its enough for Bill Gates...

:D
 
"EVERY BODY DESERVES 4 GB OF RAM", so instead of making real technology (or waiting for it to come) and make lots of real memory available to the apps, the fantastic idea they came up with was letting an application see more memory (what they call "Virtual Memory")
If this is what you believe the purpose of the Page File is then it's clear you don't understand it so just leave it alone. But even if you did understand it, you would still be leaving it alone.
Leave it alone!

Ha Joe beat ya to it!
Joe went on a 2 page rant against me a week ago explaining why it IS ok to change the page file. I know he's said in the past leave it alone but sometimes it looks like he just wants to argue so he can try to be right at the end and claim King of the World!

I simply find it ridiculous that instead of having my software (some of it my own) load into real memory (and keep it there) all the stuff it needs to RUN (as in "use those cycles/sec, a.k.a. Hertz" of the processor) and while I'm trying to have my system tuned for data access performance (as much as I can) and with several real (and big) DBs in my disk, my OS decides to go visit the pagefile, just to map a few hundred KB of data into memory..... To me, and to Camus, that's absurd!
Man, you will simply never have unlimited ram. No matter how much ram you have, there will always be information in that ram that is not going to be used. You swap that data to the page file, then load up data that is going to be used. Following me? So it doesn't matter how much you have, leave it alone so the unused data has somewhere to go so the ram you do has can be fully utilized.
 
No matter how much ram you have, there will always be information in that ram that is not going to be used. You swap that data to the page file, then load up data that is going to be used. Following me? So it doesn't matter how much you have, leave it alone so the unused data has somewhere to go so the ram you do has can be fully utilized.

Just to discuss that "no matter what" mentality.

If you have 4 GB physical RAM and 12 GB of pagefile space (which you have configured not to be able to grow). That gives you a 16 GB commit limit.

What if you install 16 GB physical RAM and don't have a pagefile. That also gives you a 16 GB commit limit.
 
What if you install 16 GB physical RAM and don't have a pagefile. That also gives you a 16 GB commit limit.
Yeah, but everything that would be a reservation becomes a commitment, resulting in higher-than-otherwise ("otherwise" meaning a page file exists) memory usage.
 
Keynes was a statist socialist who believe him and his buddies in banking and gov are the answer to every one of the worlds problems!

By the way I use 8gb and no page file, vista ultimate 64
 
I never understood the pagefile= 1.5x RAM. The more RAM you have shouldn't it be less of a pagefile? I just set my pagefile to 4gb, although I hope all my RAM gets used up before the pagefile. After installign 7 build 7068 last night and letting it idle it got to 1.4GB RAM used. So that's cool. Corsair Dominator RAM is pretty.
 
If this is what you believe the purpose of the Page File is then it's clear you don't understand it so just leave it alone. But even if you did understand it, you would still be leaving it alone.Joe went on a 2 page rant against me a week ago explaining why it IS ok to change the page file. I know he's said in the past leave it alone but sometimes it looks like he just wants to argue so he can try to be right at the end and claim King of the World!

:eek: :eek: :eek: Sorry, I don't mean to offend you with my apparent disdain, I do appreciate your attention to comment on this thread. But I can't help to feel your statement is void of any reasons for me to actually use a pagefile: you are only saying that "it's clear" I don't understand (what? -- I think you are referring to the "purpose of the Page File".) Immediately after you go on saying that (even) if I did (understand the "purpose of the Page File"), I would then use it.

What is this? Bible school?!?! I am a man of reason. I need proof. Or at least a logical argument that can lay a solid case. Not just your condemning-logic... I think you can do better. ;)

Man, you will simply never have unlimited ram. No matter how much ram you have, there will always be information in that ram that is not going to be used. You swap that data to the page file, then load up data that is going to be used. Following me? So it doesn't matter how much you have, leave it alone so the unused data has somewhere to go so the ram you do has can be fully utilized.

Let me use a moral free-will sort of argument. We men have a choice, right? As in, we are given a choice to (1) Customize the size of the pagefile, or (2) Let the system manage it, or (3) Turn it off altogether.

Now (and please carefully read the following bold-yellow clause, this is not meant to be taken as a because I can do it, then there is nothing wrong with it type of argument. &#8592; that's the end of the "bold-yellow clause",) why do you think Microsoft if giving us an option??? Eh? I want to hear your answer.....

Did CaptainMorgan89 tell you to make this thread?

LOL, I'm n00b here, I don't even know CaptainMorgan89! (well, to be more precise, I cannot claim to know any CaptainMorgan89 here, but I've seen similar usernames in other, more "explicit", forums.... :D)

Just to discuss that "no matter what" mentality.

If you have 4 GB physical RAM and 12 GB of pagefile space (which you have configured not to be able to grow). That gives you a 16 GB commit limit.

What if you install 16 GB physical RAM and don't have a pagefile. That also gives you a 16 GB commit limit.

You tell'em bro. Although the assumption here is that we have a 32bit system and thus a program will be address (without hacking or doing anything special, that is) only 4 GB of that ram.

I repeat, the issue is simple: The problem here is that windows is promising by default every windows process (those thingies that you see in the Task Manager/Processes tab) 4 GB or RAM, even thought it knows it only has 3 GB. If I go with the recommended pagefile size of 1.5 * 3 GB = 4.5 GB, that leaves me with a 3 GB + 4.5 GB = 7.5 GB commit charge.

Well, guess what! With only 2 programs, I have already exhausted the entire commit charge of my pagefile since 7.5 GB < 8.0 GB! Ha-Ha-Ha!
 
:eek: :eek: :eek: Sorry, I don't mean to offend you with my apparent disdain, I do appreciate your attention to comment on this thread. But I can't help to feel your statement is void of any reasons for me to actually use a pagefile: you are only saying that "it's clear" I don't understand (what? -- I think you are referring to the "purpose of the Page File".) Immediately after you go on saying that (even) if I did (understand the "purpose of the Page File"), I would then use it.

What is this? Bible school?!?! I am a man of reason. I need proof. Or at least a logical argument that can lay a solid case. Not just your condemning-logic... I think you can do better. ;)



Let me use a moral free-will sort of argument. We men have a choice, right? As in, we are given a choice to (1) Customize the size of a pagefile, or (2) Let the system manage it, or (3) Turn it off:

Now (and please carefully read the following bold-yellow clause, this is not meant to be take as a because I can do it, then there is nothing wrong with it. &#8592; that's the end of the "bold-yellow clause",) why do you think Microsoft if giving us an option??? Eh? I want to hear your answer.....



LOL, I'm n00b here, I don't even know CaptainMorgan89! (well, to be more precise, I cannot claim to know any CaptainMorgan89 here, but I've seen similar usernames in other, more "explicit" forums.... :D)



You tell'em bro. Although the assumption here is that we have a 32bit system and thus a program will be address (without hacking or doing anything special, that is) only 4 GB of that ram.

I repeat, the issue is simple: The problem here is that windows is promising by default every windows process (those thingies that you see in the Task Manager/Processes tab) 4 GB or RAM, even thought it knows it only has 3 GB. If I go with the recommended pagefile size of 1.5 * 3 GB = 4.5 GB, that leaves me with a 3 GB + 4.5 GB = 7.5 GB commit charge.

Well, guess what! With only 2 programs, I have already exhausted the entire commit charge of my pagefile since 7.5 GB < 8.0 GB! Ha-Ha-Ha!

That's not how things work.

First of all, 32-bit Applications are only "granted" up to 2GB of ram by default. The system reserves the other 2GB for the system/kernel.

Second, the entire 2GB allowed for each process isn't automatically commited, only what the process has requested.
 
That's not how things work.

First of all, 32-bit Applications are only "granted" up to 2GB of ram by default. The system reserves the other 2GB for the system/kernel.

Second, the entire 2GB allowed for each process isn't automatically commited, only what the process has requested.

:rolleyes: :rolleyes:

Your first point is most likely true, but that doesn't refute the essence of my argument, for a very small number of applications would still exhaust the commit charge of the system, in this case 4 apps, not 2, would do, for example.

Your second point is also correct, but (and I don't want to sound as a Robespierrian reactionary) guess what?! Any application can claim immediately the allocation of 2 GB in free memory (a.k.a. malloc) so the whole principle of "no matter how much RAM you have you will eventually exhaust it, so leave the pagefile alone" is still proved flawed... incorrect... mistaken.... unsound... weak... simple...

:cool: :cool: :cool:
 
:rolleyes: :rolleyes:

Your first point is most likely true, but that doesn't refute the essence of my argument, for a very small number of applications would still exhaust the commit charge of the system, in this case 4 apps, not 2, would do, for example.

Your second point is also correct, but (and I don't want to sound as a Robespierrian reactionary) guess what?! Any application can claim immediately the allocation of 2 GB in free memory (a.k.a. malloc) so the whole principle of "no matter how much RAM you have you will eventually exhaust it, so leave the pagefile alone" is still proved flawed... incorrect... mistaken.... unsound... weak... simple...

:cool: :cool: :cool:

Again, that's not how it works. Just because an application requests memory doesn't mean it gets committed. It has to actually has be used.
 
Here's a tidbit that might help you out.

Windows on 32 bit x86 systems can access up to 4GB of physical memory. This is due to the fact
that the processor’s address bus which is 32 lines or 32 bits can only access address range from
0x00000000 to 0xFFFFFFFF which is 4GB. Windows also allows each process to have its own
4GB logical address space. The lower 2GB of this address space is available for the user mode
process and upper 2GB is reserved for Windows Kernel mode code. How does Windows give 4GB
address space each to multiple processes when the total memory it can access is also limited to
4GB. To achieve this Windows uses a feature of x86 processor (386 and above) known as paging.
Paging allows the software to use a different memory address (known as logical address) than the
physical memory address. The Processor’s paging unit translates this logical address to the physical
address transparently. This allows every process in the system to have its own 4GB logical address
space. To understand this in more details, let us first take a look at how the paging in x86 works.
 
Again, that's not how it works. Just because an application requests memory doesn't mean it gets committed. It has to actually has be used.

I must bring the second paragraph to your attention:

Me said:
Any application can claim immediately the allocation of 2 GB in free memory (a.k.a. malloc) so the whole principle of "no matter how much RAM you have you will eventually exhaust it, so leave the pagefile alone" is still proved flawed... incorrect... mistaken.... unsound... weak... simple...

Here, by "claim", I mean, for example, in a simple C program, allocate an array of 2147483648 characters in free memory and then iterate through this array and set all it's entries to 0. Following me? This last "iteration" is not even necessary and is merely artificial, but I am including it here to make you feel comfortable in case you want to point at something you think I might have missed. :cool: :cool:

(Let me give you a hint for your next post: you can come and claim that, again, this is not how it works, for I would get a "Out of Memory" type of error by trying to allocate such maniacal amount of memory in the heap. &#8592; End of hint... :D :D)
 
I must bring the second paragraph to your attention:



Here, by "claim", I mean, for example, in a simple C program, allocate an array of 2147483648 characters in free memory and then iterate through this array and set all it's entries to 0. Following me? This last "iteration" is not even necessary and is merely artificial, but I am including it here to make you feel comfortable in case you want to point at something you think I might have missed. :cool: :cool:

(Let me give you a hint for your next post: you can come and claim that, again, this is not how it works, for I would get a "Out of Memory" type of error by trying to allocate such maniacal amount of memory in the heap. &#8592; End of hint... :D :D)

umm, exactly what point are you trying to make? It's quite obvious if a program actually allocates a 2GB chunk or RAM, you're going to run out of memory. This is where the page file comes in.
 
Here's a tidbit that might help you out.

Boy, I'm beginning to think you're just trolling around... Please come clear, it's your opportunity :rolleyes:..

I think you also can do better: I am not challenging Microsoft's suggestion for having page file, which is particulary sound for low memory systems, but some people's argument that without a pagefile we are going to kaput, because we are going to kaput nonetheless.

You see, the problem is that some people :)p it's starting to sound as an euphemism) imagine a lot more than what they are really being told, (which is worse than the times when people imagined that at the edge of the world there would be dragons...) I want to hear real, rational, arguments here, not just dumps from manuals or tech sites... Or bible pages. We need ideas boy. Ideas.
 
Boy, I'm beginning to think you're just trolling around... Please come clear, it's your opportunity :rolleyes:..

I think you also can do better: I am not challenging Microsoft's suggestion for having page file, which is particulary sound for low memory systems, but some people's argument that without a pagefile we are going to kaput, because we are going to kaput nonetheless.

You see, the problem is that some people :)p it's starting to sound as an euphemism) imagine a lot more than what they are really being told, (which is worse than the times when people imagined that at the edge of the world there would be dragons...) I want to hear real, rational, arguments here, not just dumps from manuals or tech sites... Or bible pages. We need ideas boy. Ideas.

You need to clear up your argument then. One post you're talking about application which use every drop of ram allowed, in which case the page file would be necessary to prevent the system from coming to a screeching halt (instead it just stutters along gasping for air). Then you go on about having plenty of RAM and not needing a page file.

In a perfect world, there would never be a need for paging to the hard drive, especially in situations where memory is virtually unlimited. However, there's quite a few applications that utilize the page file for it's own use, or just freak out when it's not there. I know Adobe Photoshop used to freak out, not sure it still does though.
 
However, there's quite a few applications that utilize the page file for it's own use, or just freak out when it's not there. I know Adobe Photoshop used to freak out, not sure it still does though.

Photoshop gives a warning. He might not know better and might run out of memory.

For their own usage? Applications cannot make personal calls to it.
 
Well...someone hates that search button around here...

I remember having this argument waaaaay back in 2005...I forget who with...could have been ranmasoa or something like that. Anyway, we went round and round about this. I finally went out and found an application that was a system stress application designed to open applications like crazy and such. I think I had XP and 2GB of ram...I tested 1GB, 512MB, and 0MB pagefiles....PERFORMANCE DIFFERENCE WAS NEGLIGIBLE.

Also in that thread, I went to Barnes&Noble to look into one of those "Everything about XP" books that Microsoft writes. I read up on memory usage and this is what's funny...at least in XP...if you do not define a pagefile...XP will use one anyway and will not tell you. Now I didn't buy the books as they were way too much money...but it is in print. I don't know if Vista or 7 will change that, but I'm going to go with an "I doubt it" vote.
 
Well...someone hates that search button around here...

I remember having this argument waaaaay back in 2005...I forget who with...could have been ranmasoa or something like that. Anyway, we went round and round about this. I finally went out and found an application that was a system stress application designed to open applications like crazy and such. I think I had XP and 2GB of ram...I tested 1GB, 512MB, and 0MB pagefiles....PERFORMANCE DIFFERENCE WAS NEGLIGIBLE.

I know bro, but "some people" simply are just too... well, you know, too...

...... I read up on memory usage and this is what's funny...at least in XP...if you do not define a pagefile...XP will use one anyway and will not tell you. Now I didn't buy the books as they were way too much money...but it is in print. I don't know if Vista or 7 will change that, but I'm going to go with an "I doubt it" vote.

You know what, I had actually conjectured that from simply watching all those system performance and diagnostics tools: I've always seen that after I get rid of the pagefile, PF usage still showing up, even though the commit charge limit does not include anything but memory.

As a sequence of side notes:

The reason why I came to this forum because I burned by Raptor and was looking for advice on the internet on what to do next.

The reason why I burned my Raptor was because I was using Disk Manager on Windows 7 Beta, and while I was doing a re-initialization of the disk from Basic to Dynamic, the system became very unstable and it froze for hours. Eventually a message came up telling me that I had to reboot the computer. I did and then I couldn't boot up.

The reason why my system became very unstable was because I disabled the pagefile in my Windows 7 installation. But who cares, it's Windows 7 Beta... :D
 
Sounds more like your system was unstable because your HDD was about to die.
 
Back
Top