Personally I would not turn turn-off my page file regardless of SSD or HDD. My reason being a simple one of ignorance that commits me to the following of Dr. Mark Russinovich http://blogs.technet...17/3155406.aspx At the end of the day Mark says eliminating the page file causes page faults that eat up any perceived performance benefit.

If you don't have a page file you don't get page faults, you get OOM Errors.

There are programs that will not operate correctly without a page file.

Only if they require more memory than the physical memory available. Cheers, - jahman.

If you don't have a page file you don't get page faults, you get OOM Errors. Only if they require more memory than the physical memory available. Cheers, - jahman.

Jahman you run on your own set of beliefs which is fine (just wish you back up your sources of information, link please); to each their own but you are not correct on either statement. If you are going to disaagree with my well documented positon and to be crediable at it please include your source or authority, I gave you mine. I well bibliographed my source and reason for my statements and I highly recommend that any user research and understand the topic thoroughly before accepting my or Jahmans version of Windows and how it operates. Personally I care not what anybody else does with their page file, only page file I care about is mine and you are darn tooting I have one, a nice big fat one too boot. lol

So you're saying I'm wrong about my Win XP-32 not allowing a page file over 4 GBy in size? (Tried, tested, didn't work.) Also I'd say I probably have a deeper understanding of virtual memory than you do, having prototyped a VM operating system on an IBM 360/91. But of course when arguments are feeble the way out is to knock the poster. Cheers, - jahman.

So you're saying I'm wrong about my Win XP-32 not allowing a page file over 4 GBy in size? (Tried, tested, didn't work.) Also I'd say I probably have a deeper understanding of virtual memory than you do, having prototyped a VM operating system on an IBM 360/91. But of course when arguments are feeble the way out is to knock the poster. Cheers, - jahman.

I never knocked you you have a habit of disagreeing without supporting your arguments with some objective source of reference other than to say the poster is wrong. Why;because you say so? I never said anything about a pahge file in relation to any operating system. What I said and referenced very well was that no page file (in most applications, any here) was probably not a good thing because it caused page faults. You went on a rant that it does't not but causes OOM, which it may or may not, depending on the OS. Disagree with me if you want, I know nothing but you really want to discount and dismiss my source? lol

The phrase "eliminating the page file causes page faults" is an oxymoron: You can not get page faults if you don't have a paging file. Why? Because a paging fault is an interrupt that is triggered by an app when a memory page (4 KBy RAM) of app memory has previously been paged out (written to) to the paging file. But if there is no paging file to begin with, then that same page could not have been written to the paging file, so how could you get a page fault? To request a page that was never written from a file that does not exist? Right. (Or as you say, "lol".) Therefore when you don't have a page file the operating system cannot extend the amount of memory for an app beyond physical memory available on the system (minus the amount of of physical used by the operating system). So what happens when an app requests more memory and that request is denied by the operating system? You get an OOM. Cheers, - jahman.

So you're saying I'm wrong about my Win XP-32 not allowing a page file over 4 GBy in size? (Tried, tested, didn't work.)

Hmmm, that does make me wonder about my 4990 MB pagefile... WinXP Pro 32 bit, SP3 installed.

Interesting. Same as you (Win XP 32-Bit SP3, with 47 GBy free on my C: HDD). I again tried to set my page file to 4097 MBy and got this error, explicitly stating max page file size is 4,096 MBy: Best part is: Recommended page file size: 5,563 MBy! Not sure what's going on... Cheers, - jahman.

The phrase "eliminating the page file causes page faults" is an oxymoron: You can not get page faults if you don't have a paging file. Why? Because a paging fault is an interrupt that is triggered by an app when a memory page (4 KBy RAM) of app memory has previously been paged out (written to) to the paging file. But if there is no paging file to begin with, then that same page could not have been written to the paging file, so how could you get a page fault? To request a page that was never written from a file that does not exist? Right. (Or as you say, "lol".) Therefore when you don't have a page file the operating system cannot extend the amount of memory for an app beyond physical memory available on the system (minus the amount of of physical used by the operating system). So what happens when an app requests more memory and that request is denied by the operating system? You get an OOM. Cheers, - jahman.

The original subject of the thread was in relation to page file usage as it pertains to SSD. There is a lot of information on the internet (most of it inaccurate) in regards Windows operating system and how it functions. I am not an expert of any sort on the matter but generally believe that I can read and comprehend for the most part at a level sufficient to make an informed decision. I subsequently referenced an excellent source of information that others may be able to read and make their own informed decision in the absence of mine or anyone else’s, opinion. Most of the concerns regarding page file usage on SSD applications relate to premature SSD degradation due to write flashing. A thorough research of this subject (not Tom’s Hardware forums e.t.c.) but reliable sources will show that reads, outnumber writes 40:1 further study of the matter will show that regardless of writes it will take many years to degrade SSD performance. Further clues should be evident in the fact that While W7, optimizes for SSD operation by turning defrag off, prefetch off e.t.c. it does Not turn the page file off. Now on the matter of x86 systems and page file maximum size, it is 4095 MB see here 4095MB http://support.microsoft.com/kb/237740 Page faults are not dependent on page file and you will absolutely incur faults in absence of a page file. You can see this for yourself in process monitor or task manager. Page file is also will explained in the references I posted above but these two webcasts also explain it well. Part 1 http://channel9.msdn...ica/2011/WCL405 Part2 http://channel9.msdn...ica/2011/WCL406 Additional reference: http://arstechnica.c...hp?f=17&t=96343 http://lifehacker.co...ldnt-disable-it

Edited August 8, 201114 yr by Gary A

deleted by author..

I think the page faults you see in Process Monitor/Explorer without a page file are the pages that are requested and loaded in memory for the first time.Anyway, I also have my page file here in my VRap. Some programs like Linx (at least for me) don't work without it and I'm pretty sure the OS won't be swapping stuff to the page file in any stupid way just because it's there, provided that you have enough memory, that is

...I am not an expert of any sort on the matter but generally believe that I can read and comprehend for the most part at a level sufficient to make an informed decision.

OK, this then explains your insistence on paging without a page file: You are too far removed from the subject of Systems Programming and thus do not have the level of understanding necessary to make an informed decision. Imagine now I read an article in a journal on open heart surgery valve replacement technique: I would understand what I am reading because I have good reading skills, but my understanding would be a far cry from me being able to assert "a level sufficient to make an informed decision". We must all know our limits when we venture outside our areas of expertise.

...Now on the matter of x86 systems and page file maximum size, it is 4095 MB see here 4095MB http://support.microsoft.com/kb/237740

That's an interesting tidbit pertaining Microsoft's implementation of VM: To recap then, in 32-bit versions, a page file is limited to 4 GBy per page file, with multiple page files allowed as long as each page file is in a separate directory, volume or drive. This then explains why my (single) page file is limited to 4 GBy, but the OS is recommending a larger page file, which I interpret would be implemented as a second page file in another directory. Thanks for the info!

Page faults are not dependent on page file and you will absolutely incur faults in absence of a page file. You can see this for yourself in process monitor or task manager.

No, you don't understand that what you are saying is an oxymoron, and I don't care what Windows is reporting. No page file, no page faults, period. No ifs, ands or buts. I explained this in my previous post and I stand by that explanation.

Page file is also will explained in the references I posted above but these two webcasts also explain it well.Part 1 http://channel9.msdn...ica/2011/WCL405 Part2 http://channel9.msdn...ica/2011/WCL406

I apologize, I don't have the 2+ hours available to sit through that presentation, but if you do, I'll take a brief look at any portion you point me to.

Additional reference:http://arstechnica.c...hp?f=17&t=96343http://lifehacker.co...ldnt-disable-it

Nothing at all about page faults occurring without a page file in these two links you supplied. The mumbo-jumbo about some apps mysteriously failing in the absence of a paging file is simply the writer not knowing that Windows uses the page file to write memory dumps to, so if an app checks for memory dump writing to be enabled (usually because in case of a CTD portions of that info can be sent back to the developper for analysis), of course that app will report an error of some sort. Cheers, - jahman.

OK, this then explains your insistence on paging without a page file: You are too far removed from the subject of Systems Programming and thus do not have the level of understanding necessary to make an informed decision. Imagine now I read an article in a journal on open heart surgery valve replacement technique: I would understand what I am reading because I have good reading skills, but my understanding would be a far cry from me being able to assert "a level sufficient to make an informed decision". We must all know our limits when we venture outside our areas of expertise.

You can and will definitely have page faults without a page file all the time. What there can't be with the page file disabled is paging/thrashing for all I know

You can and will definitely have page faults without a page file all the time. What there can't be with the page file disabled is paging/thrashing for all I know

No, you will not. A page fault is (again!) when the app accesses a particular page of memory that is supposed to be in physical memory but is actually not because that particular page was previously swapped out to the paging file by the OS. Now if you don't have a paging file, by definition the OS can't swap pages of physical memory out to that nonexistent paging file, and if no pages were swapped out to that nonexistent paging file, you cannot have paging faults to swap those pages back in to physical memory. In a nutshell, paging is a technique that allows swapping 4K memory pages out to disk (the paging file) and then read them back-in as needed. If you don't have a paging file you cannot implement paging, and if paging isn't implemented, you can hardly have page faults. It really is that simple. Cheers, - jahman. Cheers, - jahman.

No, you will not. A page fault is (again!) when the app accesses a particular page of memory that is supposed to be in physical memory but is actually not because that particular page was previously swapped out to the paging file by the OS. Now if you don't have a paging file, by definition the OS can't swap pages of physical memory out to that nonexistent paging file, and if no pages were swapped out to that nonexistent paging file, you cannot have paging faults to swap those pages back in to physical memory. In a nutshell, paging is a technique that allows swapping 4K memory pages out to disk (the paging file) and then read them back-in as needed. If you don't have a paging file you cannot implement paging, and if paging isn't implemented, you can hardly have page faults. It really is that simple. Cheers, - jahman. Cheers, - jahman.

Not necessarily, the way I see it. A page fault happens when the OS tries to access a page that is not in memory, doesn't mean that it was swapped to the page file before

Not necessarily, the way I see it. A page fault happens when the OS tries to access a page that is not in memory, doesn't mean that it was swapped to the page file before

So if the "page fault happens when the OS tries to access a page that is not in memory", and that page is not in the page file either because we said there was no page file, where is it then? Cheers, - jahman.