WinXP Pagefile on RAM drive?

[JJPeters]

After the sale at Fry's I now have 3GB of RAM on my gaming rig. Can I create a 2GB RAMdrive and relocate the pagefile to this drive? If so how, and will this improve performance? Thx.

 

[montag451]

Good question.
Am looking forward to some answers.

 

[KoolDrew]

That completely defeats the purpose of even having a pagefile.

 

[dajo]

Originally posted by: KoolDrew
That completely defeats the purpose of even having a pagefile.

LOL!!! That's true - the purpose of having a page file is to save state to the hard disk when you've run out of RAM.

Try turning off virtual memory.

 

[Phoenix86]

Originally posted by: KoolDrew
That completely defeats the purpose of even having a pagefile.

You know where this is going... Can-o-worms

Why not just add more RAM to the system? Pagefiles are not used much at all if you have more RAM than required memory (commit charge peak), so any tweak to the page file becomes moot.

 

[KoolDrew]

Try turning off virtual memory

You can't.

Why not just add more RAM to the system? Pagefiles are not used much at all if you have more RAM than required memory (commit charge peak), so any tweak to the page file becomes moot.

The amount of RAM in this case is not the issue; he has plenty for his needs.

 

[JJPeters]

It was my understanding that no matter how much RAM you have, WinXP peforms better if there is a pagefile.

 

[Brazen]

Originally posted by: KoolDrew

Try turning off virtual memory

You can't.

Yes you can. And the idea of putting the pagefile on a RAM drive is retarded.

 

[KoolDrew]

Yes you can.

Then prove it (this should be interesting).

 

[Brazen]

Originally posted by: JJPeters
It was my understanding that no matter how much RAM you have, WinXP peforms better if there is a pagefile.

It's not that WinXP performs better with a pagefile, it's that WinXP does not perform better without a pagefile. In other words, it won't hurt performance to disable it, but (Microsoft claims) it won't help performance either.

 

[Brazen]

Originally posted by: KoolDrew

Yes you can.

Then prove it (this should be interesting).

1. right click My Computer
2. choose Properties
3. choose Advanced tab
4. click "Settings" in the "Performance" box
5. click the Advanced tab
6. click "Change" in the "Virtual Memory" box
7. click the radio button next to "No paging file"
8. click OK on all the opened windows.


Hmm, don't know how interesting it is. It's pretty standard knowledge among computer professionals.

 

[KoolDrew]

That is not disabling virtual memory; that is disabling the pagefile. Two totally different things.

 

[Brazen]

Originally posted by: KoolDrew
That is not disabling virtual memory; that is disabling the pagefile. Two totally different things.

You're an imbecile. Quit wasting space.

 

[KoolDrew]

:roll:

You can configure XP to not use a paging file, but it's still using virtual addressing and virtual memory.

 

[JustAnAverageGuy]

KoolDrew is correct. Virtual Memory can NOT be disabled.

 

[Brazen]

type this in google: "define:virtual memory" and you will see that the most common definition for virtual memory is the ability to swap memory from RAM to the harddrive. There are other definitions, but this is the most common and obviously this is what dajo was referring to. In this case, and by this (the most common) definition, disabling the pagefile is the same as disabling virtual memory, not to mention by the terms Microsoft uses on the property pages for disabling the pagefile would indicate that is it is disabling virtual memory.

So, for all intents and purposes, using the most common definition, using dajo's meaning (who originally brought it up), and using Microsoft's wording on the property page itself, KookDrew and JustAnAverageGuy are wrong. Congratulations to you both.

 

[KoolDrew]

Originally posted by: Brazen

Originally posted by: JJPeters
It was my understanding that no matter how much RAM you have, WinXP peforms better if there is a pagefile.

It's not that WinXP performs better with a pagefile, it's that WinXP does not perform better without a pagefile. In other words, it won't hurt performance to disable it, but (Microsoft claims) it won't help performance either.

There always a practical need for a pagefile, because you can always use the extra RAM for other things. It's not as simple as "if I have enough memory, I don't need a pagefile". It's a case of how to best use the finite resource (RAM) for many things.

When you disable the pagefile you are not stopping paging (the pagefile is by far not the only file involved with paging). Instead, all you are doing is forcing the system to keep all private virtual memory in RAM. Only code and mapped files can be paged, even if some of the other stuff has not neen touched for hours and will never be touched again.

This means that there will be more paging of code for a given workload and RAM size. So disabling the pagefile turn out to be a bad thing in the long run.

 

[KoolDrew]

The term virtual memory includes several difference mechanisms. Paging is only one of these mechanisms and the pagefile is only one of the many files involved in paging.

So, Virtual memory is always in use in modern OSes. It cannot be turned off. That is the bottom line.

virtual memory is the ability to swap memory from RAM to the harddrive.

Even if this was the correct definition, you would still be wrong. The pagefile is by far not the only file involved with paging.

 

[n0cmonkey]

Originally posted by: Brazen
type this in google: "define:virtual memory" and you will see that the most common definition for virtual memory is the ability to swap memory from RAM to the harddrive. There are other definitions, but this is the most common and obviously this is what dajo was referring to. In this case, and by this (the most common) definition, disabling the pagefile is the same as disabling virtual memory, not to mention by the terms Microsoft uses on the property pages for disabling the pagefile would indicate that is it is disabling virtual memory.

So, for all intents and purposes, using the most common definition, using dajo's meaning (who originally brought it up), and using Microsoft's wording on the property page itself, KookDrew and JustAnAverageGuy are wrong. Congratulations to you both.

Forget the most common definition, because it's incomplete.

 

[Markbnj]

type this in google: "define:virtual memory" and you will see that the most common definition for virtual memory is the ability to swap memory from RAM to the harddrive.

That is technically not correct, as is common knowlege among software professionals (sorry, couldn't resist ).

KoolDrew is correct in his statement that the paging file is only one aspect of virtual memory. At its most basic virtual memory is the act of mapping process-relative addresses into the machine's physical address space. Windows has done this since the first release of NT. Every process gets 2 gigs of virtual address space, and can cause chunks of this address space to be reserved in physical memory. There's no way you can turn off virtual memory at that level. By reducing the page file to 0 you simply force Windows to make all processes live within the physical limits of ram on the system (with the exception of the system processes, which will still use a backing store even if you set page file size to 0).

This is just another case of equally valid definitions that differ only in level of abstraction competing for airtime.

I'll fall back on my normal advice: the page file, module memory management, and virtual memory manager are interelated pieces of a system that was designed by people who know much more about such things than any of us ever will. Leave them alone, and let Windows manage its memory.

 

[Phoenix86]

Originally posted by: Brazen
type this in google: "define:virtual memory" and you will see that the most common definition for virtual memory is the ability to swap memory from RAM to the harddrive. There are other definitions, but this is the most common and obviously this is what dajo was referring to. In this case, and by this (the most common) definition, disabling the pagefile is the same as disabling virtual memory, not to mention by the terms Microsoft uses on the property pages for disabling the pagefile would indicate that is it is disabling virtual memory.

So, for all intents and purposes, using the most common definition, using dajo's meaning (who originally brought it up), and using Microsoft's wording on the property page itself, KookDrew and JustAnAverageGuy are wrong. Congratulations to you both.

Google is not a end all source.

Here let me provide a REAL source. Whole lotta STFU.

In modern operating systems, including Windows, application programs and many system processes always reference memory using virtual memory addresses which are automatically translated to real (RAM) addresses by the hardware. Only core parts of the operating system kernel bypass this address translation and use real memory addresses directly.

Virtual Memory is always in use, even when the memory required by all running processes does not exceed the amount of RAM installed on the system.

When the memory in use by all the existing processes exceeds the amount of RAM available, the operating system will move pages (4 KB pieces) of one or more virtual address spaces to the computer?s hard disk, thus freeing that RAM frame for other uses.

Disabling the page file simply prevents paging to disk, it does not stop virtual memory.

You have virtual memory confused with the page file. They are closely related, but not the same thing, by a long shot.

Virtual memory is, however limited to the total of your RAM+page file. There is another limit on virtual memory of 2gb/process, which can be adjusted with the /pae switch on the kernel.

edit: You should always consider that even a well reasearched position could be completely wrong.

 

[Nothinman]

Hmm, don't know how interesting it is. It's pretty standard knowledge among computer professionals.

So you believe everything you read on the Internet? Get a copy of "Inside Windows" or "Understanding the Linux kernel' and read the chapter on memory management, then you'll understand just how wrong you are.

Disabling the page file simply prevents paging to disk, it does not stop virtual memory.

It doesn't even go that far, it just disables use of the pagefile as a repository for modified pages. You'll still be paging from executables, shared libraries, data files, etc as necessary. All you'll end up doing is locking some data into memory unecessarily and putting more pressure on the VM system.

 

[Brazen]

Originally posted by: Phoenix86

Originally posted by: Brazen
type this in google: "define:virtual memory" and you will see that the most common definition for virtual memory is the ability to swap memory from RAM to the harddrive. There are other definitions, but this is the most common and obviously this is what dajo was referring to. In this case, and by this (the most common) definition, disabling the pagefile is the same as disabling virtual memory, not to mention by the terms Microsoft uses on the property pages for disabling the pagefile would indicate that is it is disabling virtual memory.

So, for all intents and purposes, using the most common definition, using dajo's meaning (who originally brought it up), and using Microsoft's wording on the property page itself, KookDrew and JustAnAverageGuy are wrong. Congratulations to you both.

Google is not a end all source.

Here let me provide a REAL source. Whole lotta STFU.

In modern operating systems, including Windows, application programs and many system processes always reference memory using virtual memory addresses which are automatically translated to real (RAM) addresses by the hardware. Only core parts of the operating system kernel bypass this address translation and use real memory addresses directly.

Virtual Memory is always in use, even when the memory required by all running processes does not exceed the amount of RAM installed on the system.

When the memory in use by all the existing processes exceeds the amount of RAM available, the operating system will move pages (4 KB pieces) of one or more virtual address spaces to the computer?s hard disk, thus freeing that RAM frame for other uses.

Disabling the page file simply prevents paging to disk, it does not stop virtual memory.

You have virtual memory confused with the page file. They are closely related, but not the same thing, by a long shot.

Virtual memory is, however limited to the total of your RAM+page file. There is another limit on virtual memory of 2gb/process, which can be adjusted with the /pae switch on the kernel.

edit: You should always consider that even a well reasearched position could be completely wrong.

Yes, and I'm sure your answer was, in fact, well researched. But as you say, it can still be wrong, and though I know what logical memory and virtual addressing is, Virtual Memory is still commonly used to refer to memory stored on the harddrive. I have already stated that there are other definitions used for "virtual memory" however, dajo was very obviously refering to the pagefile, which is a legitimate and often used definition of "virtual memory."

And I have never considered Google to be the end all source, just the easiest one at quick notice. How about the Meriam Webster Dictionary? Some people consider that pretty definative, and it only has one: memory stored on magnetic disks. Or how about Webopedia? That's a pretty trusted source for you computer dorks, isn't it? And it pretty much uses "pages" and "virtual memory" interchangeably.

But you know what, none of that even matters, because as I said dajo was the first to mention disabling "virtual memory" and he was obviously refering to the pagefile, not virtual addressing. KookDrew was obviously trying to be a jick to dajo; I was just pointing out that dajo's use of the term "virtual memory" was not wrong, and is in fact more common than KookDrew's use of the term.

edit: and if I wanted to be as unreasonable and pointless in arguing as some you, I would just point at that even virtual addressing can be disabled by reprogramming or switching to DOS. But that sort of arguing is just retarded, but then, no more so than your argument.

 

[n0cmonkey]

Originally posted by: Brazen
But you know what, none of that even matters, because as I said dajo was the first to mention disabling "virtual memory" and he was obviously refering to the pagefile, not virtual addressing. KookDrew was obviously trying to be a jick to dajo; I was just pointing out that dajo's use of the term "virtual memory" was not wrong, and is in fact more common than KookDrew's use of the term.

The point is, Virtual Memory is the wrong term. Period.

 

[Brazen]

Originally posted by: Nothinman

Hmm, don't know how interesting it is. It's pretty standard knowledge among computer professionals.

So you believe everything you read on the Internet? Get a copy of "Inside Windows" or "Understanding the Linux kernel' and read the chapter on memory management, then you'll understand just how wrong you are.

I have had to read an entire book on memory management (only because I had to as part of getting a degree in programming). And what the heck does "you believe everything you read on the internet" have to do with anything? Because I referenced Google? You're really stretchin' for it there. Why don't you criticize Phoenix86 for referencing Microsoft while you're at it.