I'm going to chop your post a bit, most of it feels off topic, not needed and for some reason you sound too much invested in your idea to run with no paging file.
I don't understand this. Direct question to you: what type of virtual memory is offloaded to paging file when windows decides there's a pressure?
I've ask you what are the changes in vista / win7 memory you claimed exist that would allow us to disable paging file and rip the benefits. So, is it suprefetch then?
I had the choice to do either way before you "provided alternative". There's no compelling evidence to try, no tests, no numbers, sorry but there's no reason. Do tests, post numbers - we'll talk.


self proclaimed experts I'd rather not listen to, I'd go grab a book on the subject - Windows Internals has a lot of chapters dedicated to memory management. It also has exercises so you know what you're reading isn't an opinion.
how is it balanced? where are these numbers from? Does this cover all workloads? most? some? just yours?

i think we are misunderstanding each other somehow.... that or i'm not entirely following your question through or we are thinking 2 completely different things related to the original which would explain the confusion. There are numerous different circumstances that do and do not have to be met for it to do it, it's not necessarily a type. You have your entire virtual memory system, that encompasses all the sub system which includes the page file, the physical ram, and so forth. Within those each will do it's own tasks and handle them appropriately. Depending on the length of time someone has been unaccessed or not required or idle, windows will stuff it away. But the page file also on the fly dublicates everything the is alread in physical ram, so you get redundancy in case of a crash/failure for dumps. There are numerous pages discussing these systems as a whole as independantly. I say subsystems because they can be categorized and broken down for each. Just like talking about a household and then saying there is individual people inside that household. Sorry if my way or explaining or my own understanding of it uses words or discriptions that don't make sense to you or others.


No, not entirely, although windows 7's super fetch is less agressive than vista's.... they do tie into each other, and do have an impact on it. Basically adds an additional level "subsystem" to the virtual memory system itself. The benefits is the entirely better memory management system within it, there really isn't one single thing to mention, It's just much better at coping with the lack of a page file than the previous versions of windows like xp and 2000 and such which relied heavily on it in many cases... mostly due to the fact that physical ram wasn't exactly an abundant resource. Today, it's overly abundant due to it's low cost and high volume.

Back in the DOS or old windows days where developers and software creators had to really crunch things down to fit into the common memory system setup, and put in minimum requirements where no page file or other ways of working outside of the physical ram was available. The introduction resolved that, it prevent system crashes and allowed far more flexibility for developers to design and support several different system configurations without cutting their own throat... not enough ram.. fine we'll dive into your page file when we need to. Further improvements to the system yeilded better memory management and better efficiency, instead of using it essentially a place to dump idle stuff or what have you, oventually they turned it into a quick cache, reason why some of the old OLD programs heavily relied on there being a page file because it actually made heavy use of even with a surplus of ram available. As things advanced, they started working away from REQUIRING the page file and oventually with the way the newer windows were designed, allowed windows to manage it mostly for them, program request this amount or this memory or whatever, windows would dish it out and place it where it needed to. This of course all extremely simplified.

The discussion in the end is that windows vista/7's memory management is improved along the same lines, making it no longer a REQUIREMENT to have it there in many situations provided with enough physical ram. But it is there. I've even had someone say to me so they could comprehend it better that it's like having an Emergency STOP at a factory when the product going out exceeds the capacity of the trucks hauling it away, giving them all time to get a bunch more temporary vans that are always on standby, which can't haul nearly as much and maybe not as efficiently or fast or easy to handle, but something to assist with the load so the factory doesn't back up. I thought it was a reasonable anology.

Currently today, with the 8gb + worth of physical ram, using the above anology, we have twice to as much as 8 or 16 TIMES more trucks than what the factory could ever produce in most senarios for most people. Why keep paying the temporary vans to keep their engines running and on standby for something that 99% of the time isn't going to require them, and in the rare case they do, they aren't likely to be able to provide the solution either.


How does one test something that isn't exactly benchmarkable, or no real benchmark excists.... we could discuss startup times or resume times from standby or wake up times, or we could delve into specific case senarios where specific programs or games have a habit of doing things it really doesn't necessarily NEED to do that involves the page file heavily. Windows just has to do a little more work overall when dealing with the page file.

The pros and cons of which can be debated to no end, this thread is a clear indication of that. I don't have to convince anyone, i just brought it up, it pertained to the thread, and many people don't know much about it to begin with. So i figured presenting the other side of the discussion as best i could in the limited way i can is a helpful thing to do specially if it can spir an intelligent conversation about it further increasing other peoples ability to comprehend hopefully.

In any sided discussion, the same thing can be said about there being no compelling evidence to not try..... Mostly only thing we are getting is.... No page file.... you might have issues (no one can guarantee you will)... so it's totally an objectionable statement... same thing can be said for running with one... You can still have the same memory related errors in either case ironically... it's just "potentially" more likely without a page file, but then again that just makes the issue that isn't entirely related to a page file more obvious.


I'm willing to try anything just to prove it for myself. "REAL Experts" i've found mostly are unhinged from reality, self proclaimed experts either don't have a clue, are objectionable to everything outside of their own reality. I do not think i fall into either of these categories, as i'm wide open to opinions and suggestions or alternatives. I have already listed out the different options available that include all the arguements for/against/combined. Someone that was self proclaimed expert would typically only list the one and only one option that they use and swear by it without providing at least details on the others.

Windows internals does have numerous solutions regarding plenty of things... but when it comes to the page file.... even it doesn't really give a clear cut and detail explanation of why or why not.... It doesn't even give room to form an opinion because there isn't anything to form an opinion on. Not because it lays out the facts as they are, because it really doesn't thoroughly discuss specifically what we are talking about in this specific manner.


The numbers are using statistical information provided by common 32bit multiple tasks at any given time for general users including gamers. The first number which is 800mb is a minimum amount specifically stated by windows vista/7 for crash dumps and basica system related page file opperations in those crash events.... they are not at all system critical, it's just a suggestion. the value of up to 2048mb is a value that should exceed any older programs that specifically requires page file space (again surprised if any would be using one on a windows 7 enviroment). It also falls into the same line as 32bit default non large address aware programs would max out at. The only setting above that would be 4096, anything higher than that is just getting rediculious (my case with 64gb of ram was attempting to address 96GB of hardrive space for the page file).... this should in almost all cases.... short of someone that is developing or rendering giagantic or heavily complex works beyond that of typical graphic artists or photoshoppers or even some game developers provide more than a sufficient amount of paging space.

But like i said.... even with the page file disabled.. even with as much photoshopping and quark/unreal ed/bryce/3dmax/and countless other tools all running at once doing several things at once all at the same time even, i've never once experienced an issue/crash/bug/whatever. I just haven't. This is my personal tests and experiences, this is the personal testimonies of many of which have tried it and haven't had any problems.... etc and so forth.

Do it.. or don't do it.. try it or not.... in the end it doesn't really matter...

I reduce at the very least for the very sake of not consuming rediculious amount of hardrive space, specially for small drives and to bring it to a value that is at least somewhat realistic. I do however experience a noticeable improvement in responciveness and "speed" some of the speed may be placebo.... but the responciveness could be measured.. it's just hard to know how to properly benchmark.