Second SSD for the page file?

[taq8ojh]

Don't be silly, you will gain a lot more if you create a RAMdisk and move your virtual memory there.

With every reboot you will then get a completely clean pagefile.sys which will prevent the mysterious faults associated with a corrupted vitrual memory.

If you can't add to the conversation and are only in here to insult people, stay out of the thread.


esquared
Anandtech Forum Director

 

[Dufus]

The pagefile is for modified data, non modifiable data can still be paged to disk hence the paging executive option. IOW just disabling the pagefile doesn't stop paging to disk, only stops modifiable data paging out.

 

[Fred B]

There could be a little gain in a second ssd because of the 64K seqentieel writing of the page.

Made a smal test to see how much time it takes from a ssd , for 60 seconds while playing a game it takes 10ms for the pagefile . It can be very orginized like the write happends exactly with 1 second interval. The iq time for the page is abouth 10ms . one write 64K takes abouth 0.03ms. When writing to disk the qeue depth goes up .But with ssd there is almost no need to because it is fast enough for qeue depth 1. I can imagine that some aplications realy benefit from fast page.


Here are some screenshots from the warning from W7 that the page is disabled and a new one is set to C . And the warnig from XP when the VM is to small , i begin with 25 MB for the page 250MB is better with no warning .The pagefile is on E ^_^

http://imgur.com/a/0wR9l

 

[Nec_V20]

I think we all have to go back to where the idea of a virtual memory/pagefile (Windows, DEC VAX/VMS)/swapfile (Unix, Linux, OSX) comes from.

Back in the day RAM was a scarce and expensive commodity and the pagefile/swapfile was a logical way of keeping processes "open" whilst at the same time freeing up RAM for processes which demanded CPU attention. Essentially the pagefile/swapfile uses a portion of the hard drive as if it were RAM.

Windows (DEC VAX/VMS)/Unix/Linux/OSX are essentially designed to use a pagefile/swapfile. The Operating Systems simply cannot work without one being defined, so disabling the virtual memory is not an option. Even with the advent of cheap RAM and plenty of it, the virtual memory subsystem is still an important part with regard to the functioning of the Operating System itself.

The idea of using a different hard drive for the virtual memory made sense in that one was using two different sets of heads to speed up reads/writes from and to the pagefile/swapfile and this was a considerable performance gain over the virtual memory being on the same drive as the program/data drives.

It has to be remembered that no matter how many heads and platters a hard drive has, all the heads move together over the platters and the individual tracks under the heads at any time constitute what is known as a "Cylinder".

This reason has disappeared with the advent of SSDs going mainstream. With an SSD the pagefile/swapfile has achieved a good proportion of RAM speed. Using a second SSD for the pagefile/swapfile simply makes no real sense with regard to speeding up the system.

The caveat with regard to SSDs is that every write to an SSD brings it that much closer to death. We are now at a life-cycle of the individual cell down to 1,000 P/E cycles. That is when a cell has been written to for about a thousand times it is dead and that capacity will have been lost forever.

The very worst thing that one can do is to stick with the Windows default "Automatically manage paging file size for all drive" with regard to the virtual memory. This is simply because Windows sucks at it. You should always choose "Custom Size" and make the minimum and maximum size the same. This will improve performance in that Windows will not fart around trying to make decisions about the virtual memory and you getting messages popping up (I almost wrote "pooping up" which would be more accurate) about problems with the virtual memory.

I had written before that I am using a RAMdisk where I have placed the virtual memory and the driver software I use for this is RAMdisk for Windows from Softperfect http://www.softperfect.com/products/ramdisk/ This software used to cost quite a lot but it is now (Version 3.4) freeware.

The RAMdisk I have created not only contains the pagefile.sys but also my browser cache. Remember the OS is permanently writing to the virtual memory and your browser is permanently writing to the cache.

The virtual memory can become corrupted. This happens for instance if your system suddenly stops working due to a BSOD or power cut and changes which were pending and still cached have not actually been written to the pagefile and disappear. This can result in the system behaving erratically after a reboot. The only way to cure this is to wipe out the old pagefile and create a new one.

With regard to the browser cache, in FireFox it is limited in size to a maximum of one gigabyte and when it is full then it will be flushed on the basis of FIFO (First In First Out). One gigabyte sounds like a lot, but it will fill up very quickly if you are watching videos online.

Every gigabyte you write to an SSD brings it one gigabyte closer to death.

A RAMdisk is where a hard drive is simulated in RAM (it's about the same as a virtual memory only exactly the opposite ) and is "formatted" with NTFS. It is created at boot by a kernel driver which preloads before the main OS. You can configure the RAMdisk to load with a specific drive letter (in my case "W:") So when the OS loads it knows where to place the pagefile.sys and not seeing one will create it.

So far I have done very well with a RAMdisk size of just over three gigabyte (two gigabyte for the virtual memory and one gigabyte for the browser cache).

The advantages of a RAMdisk are that the virtual memory is accessed exceedingly quickly and also, because the pagefile.sys file is created anew at every boot, any corruption which might have existed before the last reboot is gone.

Another advantage is that RAM doesn't care how often it is written to. It does not have a death timer on it the way that an SSD does. For this reason the virtual memory and browser cache which I have relegated to the RAMdisk is not wearing out my SSD prematurely.

If you have 16GB of RAM then you will not take a performance hit if you allocate 3GB or so for a RAMdisk, and if you only have 8GB of RAM then it is not a massive investment to get another 8GB

 

[zir_blazer]

Windows (DEC VAX/VMS)/Unix/Linux/OSX are essentially designed to use a pagefile/swapfile. The Operating Systems simply cannot work without one being defined. So disabling the virtual memory is not an option. So even with the advent of cheap RAM and plenty of it, the virtual memory subsystem is still an important part with regard to the functioning of the Operating System itself.

Virtual Memory is ALWAYS active if you're running in Protected Mode/Long Mode, you can't disable it. The reason why people confuses Virtual Memory as if it was equal to Pagefile, is because one Windows version (W9x I recall) specifically named the option that disabled the Pagefile as "Disable Virtual Memory". You can disable the Pagefile, which is a file that represents the space that the OS reserves in the HD to store things from the RAM. I have been running without it for over a decade on WXP.


Virtual Memory is a layer where the applications receive their own entirely private space of memory where they think that they have full control of it, yet they don't know where this memory is physically located. Its similar to a rustic form of virtualization. The only one that knows where the data from the applications are physically located, is the OS, which is the one that decides what to send to the Pagefile and what to keep in RAM.
Applications shouldn't malfunction when Pagefile is disabled (Unless its an Out of Memory error, for obvious reason. However, an OoM can also happen if the Pagefile got a max cap and you try to go beyond your Virtual Memory limit, which is RAM + Pagefile), simply because they shouldn't be aware of where they are on the first place. That's the whole point behind Virtual Memory, making the physical location of the data entirely transparent to the application.
If an application tries to enforce having a Pagefile regardless of how much physical RAM you actually have (I heared some Adobe applications complains if you don't have a several GBs worth Pagefile), I consider it a retarded desing flaw because it blatantly shows you that the developers didn't designed it with big systems in mind or any scaling possibilities.

A RAMdisk is where a hard drive is simulated in RAM (it's about the same as a virtual memory only exactly the opposite ) and is "formatted" with NTFS. It is created at boot by a kernel driver which preloads before the main OS. You can configure the RAMdisk to load with a specific drive letter (in my case "W:") So when the OS loads it knows where to place the pagefile.sys and not seeing one will create it.

If you have 16GB of RAM then you will not take a performance hit if you allocate 3GB or so for a RAMdisk, and if you only have 8GB of RAM then it is not a massive investment to get another 8GB

Did you thinked on the ridiculous overhead than that "solution" adds? You're reserving RAM to make a RAMDisk to put the Pagefile, yet if the OS is running out of RAM (Because it has 3 GB to play with than you reserved for the RAMDisk), it will start paging to and from RAM wasting read/writes that it shouldn't do. Its like two partitions where you have to constantly swap data from and to to "balance" them when a single, bigger one would be a better solution. Using a RAMDisk as an extremely fast Cache for I/O intensive task do is useful, but for Pagefile, no.


In MY case, I do run a Pagefile inside a RAMDisk. Why, if I'm saying that its ridiculous? Because I have 32 GB RAM yet still I'm still running on WXP SP3, and with PAE, doing that allows you to workaround the 3 GB or so limit of RAM that WXP sees directly. With the Pagefile there I can actually put to use that RAM, as RAM. If I was running a 64 Bits OS I wouldn't be doing so, through.
Oh, and the performance penalty for using PAE is very, VERY big in IOPS.

 

[Nec_V20]

Virtual Memory is ALWAYS active if you're running in Protected Mode/Long Mode, you can't disable it. The reason why people confuses Virtual Memory as if it was equal to Pagefile, is because one Windows version (W9x I recall) specifically named the option that disabled the Pagefile as "Disable Virtual Memory".

You will notice that when I spoke about Windows I mentioned it as "Windows (DEC VAX/VMS)" whereby WinNT 3.1 (the first WinNT version), WinNT4, Win2k, Win2k3, Vista, Win2k8 Win7 and Win8 are a direct descendent of VMS (which stands for "Virtual Memory System") as opposed to Linux/OSX being a direct descendant of UNIX (or Unics as it was originally known).

I was thus not referring to MSDOS, Win3.x, Win95, Win98, WinME or the last iteration of that genealogy WinXP.

"Pagefile" and "Virtual Memory" are used by me synonymously because they are synonymous in the context I was referencing.

Did you thinked

I did not "thinked" anything - although I might have had some thoughts. The RAMdisk driver I was talking about is a kernel mode driver (the same as the SATA AHCI driver) and thus the impact on the OS itself (which is, if one considers that it is based on the HAL (Hardware Abstraction Layer) a virtualisation of the system in RAM) is minimal. It runs in Ring0 whereas the OS itself runs in Ring3.

If you are going to criticise my post then I would ask you to actually read it before banging your fingers senselessly onto your keyboard.

My native language is German, however I do make the effort to make my thoughts cogent and comprehensible to the reader when I am writing in English. I do not know if English is your native language although it certainly doesn't sound like it, nor does it sound as if you made any kind of effort - linguistically or conceptually - with regard to your post.

 

[hanspeter]

zir_blazer's point is that you are better off with just disabling the page file. Putting it on a ramdisk only adds overhead - Windows cannot use the reserved amount of ram directly.

 

[IndyColtsFan]

don't trash this thread because you too broke to buy enough physical ram.

now get lost.

Never disable the page file:

http://www.howtogeek.com/126430/htg...-windows-page-file-and-should-you-disable-it/

http://lifehacker.com/5426041/understanding-the-windows-pagefile-and-why-you-shouldnt-disable-it

The goal is to add enough RAM so it (the page file) is minimally used, but you should never outright disable it.

Oh, and LOL at the internet tough guy "you too broke" nonsense.

 

[KingFatty]

Never disable the page file:

http://www.howtogeek.com/126430/htg...-windows-page-file-and-should-you-disable-it/

http://lifehacker.com/5426041/understanding-the-windows-pagefile-and-why-you-shouldnt-disable-it

The goal is to add enough RAM so it (the page file) is minimally used, but you should never outright disable it.

Oh, and LOL at the internet tough guy "you too broke" nonsense.

Looking at the first article above, I noticed that it contradicts itself. So, you can't really trust that self-contradictory article. Setting that aside, the article also uses a type of Pascal's wager, where it tries to convince you that there is absolutely no benefit to disabling the page file, so why not use it? Again, the article itself contradicts that position if you just keep reading toward the bottom, but it also simply disregards a very nice benefit of disabling the page file, and that benefit is magnified even more when you have a small-size SSD: the benefit of freeing up precious space that your pagefile hogs up. Anyway, just wanted to dispute that first link as being poorly reasoned and self-contradictory, that there *is* a detriment to using a page file etc.

What's even more galling is the author references another article claiming it proves his point based on benchmarking with/without a pagefile, when it actually contradicts him at some points. See: http://www.tweakhound.com/2011/10/10/the-windows-7-pagefile-and-running-without-one/
Specifically, the system booted up faster and shut down faster when pagefile was disabled. That's a tangible benefit, which the author of the first article conveniently ignores. Good writing doesn't ignore inconvenient facts - that just comes across as sloppy and dishonest. Look at the other benchmark tests there, so many of them show an improvement, which is discounted because it's small. But it's an improvement nonetheless! I mean this supporting article literally says:

It is possible that gaming performance might be improved marginally (but not significantly) without a pagefile

And it brings home the entire point of how the situation is different with SSDs where every gigabyte is precious, and RAM is plentiful:

If disk space is a concern then lowering or eliminating the pagefile can be done if you have enough RAM.

Is is precisely these kinds of definite, undeniable facts that authors of the above-cited articles conveniently ignore. They just don't appreciate the full facts of the situation, and instead raise the scary boogeyman argument. I just can't help but lump them into the "chicken little" sky-is-falling types, where if there is any ghost or hint of a risk, they run scared and cannot process the actual tangible benefits that are undeniably present (e.g., freeing up space on an SSD).

The second article also is poorly reasoned. The argument is that the pagefile will never slow your computer because you will be working on the active program, and you'd only feel the slowdown when you switch to a background program that has been swapped out to the pagefile and your computer has to grind the hard drive. Well, isn't that a slowdown, having to wait for that program to get retrieved from pagefile - the usage scenario should include when people alt-tab to other programs, and it's odd that you should somehow discount this type of usage as "using" the computer, by only considering the active program. Using a computer includes alt-tabbing, multi-tasking, etc.

And if you scroll down to the bottom to read the bottom line reasoning in the second article, you can see the self-defeating logic of admitting there is a benefit to improving performance. The only "drawback" is a boogey-man fear mongering that your system will crash and burn.

It's sad that neither article gets into the benefits of the compromise position of manually setting a very small pagefile while allowing windows to enlarge it as-necessary, where you enjoy the benefit of freeing up space on your SSD, and avoid the evil boogey-man of crashing the system because windows will enlarge as necessary. Either way, I don't see either of these two articles as getting the nuances involved. They miss the benefits and don't appreciate how that is even a compromise.

 

[IndyColtsFan]

It's sad that neither article gets into the benefits of the compromise position of manually setting a very small pagefile while allowing windows to enlarge it as-necessary, where you enjoy the benefit of freeing up space on your SSD, and avoid the evil boogey-man of crashing the system because windows will enlarge as necessary. Either way, I don't see either of these two articles as getting the nuances involved. They miss the benefits and don't appreciate how that is even a compromise.

Yes, that approach is probably the best.

EDIT: Also, those were merely a couple of the first links I grabbed. A Microsoft MVP outlines it here as well: http://answers.microsoft.com/en-us/...age-file/8777f70f-97e1-4b0e-a33b-b80937210722. You can find other sources as well.

 

[UaVaj]

Never disable the page file:

http://www.howtogeek.com/126430/htg...-windows-page-file-and-should-you-disable-it/

http://lifehacker.com/5426041/understanding-the-windows-pagefile-and-why-you-shouldnt-disable-it

The goal is to add enough RAM so it (the page file) is minimally used, but you should never outright disable it.

Oh, and LOL at the internet tough guy "you too broke" nonsense.

those article are full of holes. thanks for the dissect kingfatty. :thumbsup:

indycoltfan. will you stop linking piss poor articles. that last article is based on the user having 2GB of ram. seriously. if you gonna link. link something useful. example: a user with 16GB/32GB/64GB of physical ram complaining about page file being off or something of the like.

-----

as long as one understand the implication of disabling/minimizing page file. one have nothing to worry about. once physical ram run out. program will crash randomly or will cease to load (been there done that - have you?). (1) time to buy more physical ram or (2) time to reduce the task or (3) time to re-enable page file. some of us only choose option #1.

only reason to keep page file on (minimum 16MB) is so that it "passes" the built-in-check on few memory hungry program (MHP). MHP will check for a page file before loading cause MHP knows most user do not have enough memory (save the tech support calls).

as for the "you too broke" nonsense. thanks for sharing you are one of those who is too broke n complaining. some of us run with max/enough physical ram. you do not see us complaining about max/enough ram do you? get a life? or get more physical ram?

 

[IndyColtsFan]

those article are full of holes. thanks for the dissect kingfatty. :thumbsup:

indycoltfan. will you stop linking piss poor articles.

Can you read? I'd trust a Microsoft MVP far more than someone like you.

Can you? Yes. Should you? Definitely not!

1. If you don't have a page file, you can't use all the RAM you have.
That's because Windows preallocates virtual memory in anticipation of
a possible need for it, even though that allocated virtual memory may
never be used. Without a page file, that allocation has to be made in
real memory, thus tying up that memory and preventing it from being
used for any purpose.

2. There is never a benefit in not having a page file. If it isn't
needed, it won't be used. Don't confuse allocated memory with used
memory.


Ken Blake, Microsoft MVP (Windows Desktop Experience) since 2003

Ouch, another quote directly from TechNet:

You can remove a paging file from a volume by selecting the volume and choosing No Paging File. (In fact, you can do this to get rid of all paging files, although doing so is not recommended, even on systems with a lot of RAM.)

Oops!

as long as one understand the implication of disabling/minimizing page file. one have nothing to worry about. once physical ram run out. program will crash randomly or will cease to load (been there done that - have you?). (1) time to buy more physical ram or (2) time to reduce the task or (3) time to re-enable page file. some of us only choose option #1.

And that's the gist of it -- you are not qualifying your arguments.

as for the "you too broke" nonsense. thanks for sharing you are one of those who is too broke n complaining. some of us run with max/enough physical ram. you do not see us complaining about max/enough ram do you? get a life? or get more physical ram?

Oh noes, a "Senior Member" is acting tough! Seriously, how old are you? Yeah, I'm broke alright. My cushy six figure consulting job and in-home mini datacenter really show how broke I am. How many dual or quad Xeon boxes with 128 GB+ of RAM do you have sitting around your house? None? Oops, looks like you're the broke one chump. I don't even own a machine with less than 16 GB. As a user or two said earlier, get lost. Let the adults converse.

 

[zir_blazer]

The second article also is poorly reasoned. The argument is that the pagefile will never slow your computer because you will be working on the active program, and you'd only feel the slowdown when you switch to a background program that has been swapped out to the pagefile and your computer has to grind the hard drive. Well, isn't that a slowdown, having to wait for that program to get retrieved from pagefile - the usage scenario should include when people alt-tab to other programs, and it's odd that you should somehow discount this type of usage as "using" the computer, by only considering the active program. Using a computer includes alt-tabbing, multi-tasking, etc.

This is the reason why people that talks about how Pagefile doesn't affect performance are wrong. I don't know Windows behaviator on Vista/7/8, but on WXP, it was quite aggresive, paging most things that are left Idle even if you have tons of RAM. Now, the performance itself will not be benchmarkeable, but you will feel the system a little faster when Multitasking for one simple reason: You will NEVER collision with Windows Pagefile HD I/O. If Windows decides to send something you had minimized to the Pagefile at the same time you want to open another application, you are going to expect that the HD head will dance around all the platter by doing both writes to the Pagefile and reading what you want to open at the same time, lowering the perceived performance and making you feel like its lagging a bit more than usual. So it DOES affects performance, through you will need to have a very good eye to figure out of its there for you or its a placebo effect.

Can you? Yes. Should you? Definitely not!

1. If you don't have a page file, you can't use all the RAM you have.
That's because Windows preallocates virtual memory in anticipation of
a possible need for it, even though that allocated virtual memory may
never be used. Without a page file, that allocation has to be made in
real memory, thus tying up that memory and preventing it from being
used for any purpose.

2. There is never a benefit in not having a page file. If it isn't
needed, it won't be used. Don't confuse allocated memory with used
memory

Point 1 is true. However, if you have too much RAM that you can simply let it preallocate and still have tons left (So you got covered a RAM use worst case scenario), there is no point in not doing so if you're saving on HD I/O or SSD life. After all, lot of people didn't recommended going beyond 8 GB of RAM a year ago or so because only a few editing applications could make use of more. Why economize when you don't have too?

Point 2 is NOT true, at least WXP aggressively pages things as I have stated. All the issue here with the Pagefile is that Windows uses it when it have no need to do so, so disabling it is a way to force Windows to simply use RAM.



I also got this quote from an article of Mark Russinovich:

Some feel having no paging file results in better performance, but in general, having a paging file means Windows can write pages on the modified list (which represent pages that aren’t being accessed actively but have not been saved to disk) out to the paging file, thus making that memory available for more useful purposes (processes or file cache). So while there may be some workloads that perform better with no paging file, in general having one will mean more usable memory being available to the system (never mind that Windows won’t be able to write kernel crash dumps without a paging file sized large enough to hold them).

He actually doesn't seem to say that having no Pagefile is bad and shouldn't be done, just that you have more memory available to use if you have one, and that there seems to be some cases where performance is increased if you don't have it.

During the last decade I participated on tons of Threads regarding the Pagefile and there is never a definitive conclusion, just the "PAGEFILE MUST ALWAYS BE ENABLED" vs users with tons of RAM. I will have to drop him a E-Mail sooner or later.

 

[UaVaj]

IndyColtsFan

let gets this straight.

you are a guy who owns multiple dual-quad xeon boxes with 128gb+ of ram who make cushy 6 figures from consulting jobs telling strangers on a forum to continue to engage page file on the basis of a helpdesk conversation via microsoft regarding if 2gb of ram enough.

clearly you win.

 

[IndyColtsFan]

IndyColtsFan

let gets this straight.

you are a guy who owns multiple dual-quad xeon boxes with 128gb+ of ram who make cushy 6 figures from consulting jobs telling strangers on a forum to continue to engage page file on the basis of a helpdesk conversation via microsoft regarding if 2gb of ram enough.

clearly you win.

Yes I do win. Thanks for acknowledging it. In the meantime, take some English comprehension classes because you completely missed my point. You clearly don't understand anything I said. Again, from TechNet:

You can remove a paging file from a volume by selecting the volume and choosing No Paging File. (In fact, you can do this to get rid of all paging files, although doing so is not recommended, even on systems with a lot of RAM.)

 

[UaVaj]

Yes I do win.

where is the bow down smiley when you need it.

using 2GB memory management advice for 128GB+ memory management.

wonder what 128GB - 2GB = ???

lmfao. :whiste:

 

[taq8ojh]

"This message is hidden because UaVaj is on your ignore list."
Trust me people, it will make your life better. Follow my example.

 

[IndyColtsFan]

where is the bow down smiley when you need it.

using 2GB memory management advice for 128GB+ memory management.

wonder what 128GB - 2GB = ???

lmfao. :whiste:

I'm sorry, you're obviously incapable of reading. Once more (for the third time): What part of this TechNet article do you NOT understand?

You can remove a paging file from a volume by selecting the volume and choosing No Paging File. (In fact, you can do this to get rid of all paging files, although doing so is not recommended, even on systems with a lot of RAM.)

Is that big enough text for you to get it yet? Where is a mention of 2 GB in the quote above? Where is a mention of 128 GB of RAM? Nowhere. Just because one article may have mentioned 2 GB of RAM doesn't mean it is incorrect. Microsoft itself says NOT to disable it EVEN with "a lot of RAM." You do know who Microsoft is, or do I need to explain that too? Quit dodging that point and tell us all how TechNet is wrong and UaVaj is right. You won't, because you can't. Or do you even know what TechNet is?

As I concurred above, the best approach with a huge amount of RAM is to reduce the size of the page file. You should NOT turn it off. Thanks to zir_blazer for yet another quote:

Some feel having no paging file results in better performance, but in general, having a paging file means Windows can write pages on the modified list (which represent pages that aren’t being accessed actively but have not been saved to disk) out to the paging file, thus making that memory available for more useful purposes (processes or file cache). So while there may be some workloads that perform better with no paging file, in general having one will mean more usable memory being available to the system (never mind that Windows won’t be able to write kernel crash dumps without a paging file sized large enough to hold them).

Do you understand this quote or do I need to break it down for you too? Run along and play some CS now.

 

[KingFatty]

But, I think Microsoft or the employee whatever also states there is never a benefit to disabling the page file.

But I think that's incorrect. There is a tangible benefit of recovering drive space, which can be precious when you have a small SSD. Also specific benchmarks show increased performance for gaming, bootup, and shutdown times.

So the better question is, why should we believe Microsoft when they make demonstrably false statements? Perhaps there is some qualifier or limited context for the statement that there is never a benefit to disabling, but I'm still having a problem resolving the paradox of how that statement can be trusted, in the face of contradictory evidence?

Rephrasing, I believe the approach of IndyColtsFan is to cite a general statement of how it's not recommended to disable page file. That's fine. But there is no explanation for contradictory facts. The general statement does not resolve the issue that there are tangible benefits to disabling the page file. Sure, it's nice to be able to record a crash dump file, but honestly, I'd prefer a faster bootup/shutdown and improved gaming performance, and who cares about recording a crash dump file? I'd prefer to free up space on my SSD. To cite a car analogy, I'm willing to (gasp) at times exceed the speed limit on the highway, despite that being 'not recommended' by law enforcement. Sometimes the visibility and other conditions warrant it, and sometimes cops will let you get away with it so long as it's not excessive.

 

[IndyColtsFan]

But, I think Microsoft or the employee whatever also states there is never a benefit to disabling the page file.

But I think that's incorrect. There is a tangible benefit of recovering drive space, which can be precious when you have a small SSD. Also specific benchmarks show increased performance for gaming, bootup, and shutdown times.

Don't forget that you can always move a page file to a conventional hard drive. It does not have to remain on an SSD. If your system has a significant amount of RAM, restrict the size of the page file and keep it on your SSD or even move it off to another drive. It probably won't be used much.

Rephrasing, I believe the approach of IndyColtsFan is to cite a general statement of how it's not recommended to disable page file. That's fine. But there is no explanation for contradictory facts. The general statement does not resolve the issue that there are tangible benefits to disabling the page file. Sure, it's nice to be able to record a crash dump file, but honestly, I'd prefer a faster bootup/shutdown and improved gaming performance, and who cares about recording a crash dump file? I'd prefer to free up space on my SSD. To cite a car analogy, I'm willing to (gasp) at times exceed the speed limit on the highway, despite that being 'not recommended' by law enforcement. Sometimes the visibility and other conditions warrant it, and sometimes cops will let you get away with it so long as it's not excessive.

Sure, there are apps which might show an improvement in performance if you disable the page file, but many more won't or at least, won't show enough of a performance improvement to overcome the downsides of disabling it. Microsoft wrote the OS and their recommendations, to me, are what need to be followed when people are making blanket, general statements. However, UaVaj didn't qualify his statements and made a blanket statement regarding the disabling of page files and then followed up with the assininie method of "If you start crashing because you're out of memory, just buy more RAM." And then he starts in with the blowhard "you too poor!" if you disagree with him, which is a joke and shows his mentality, lack of maturity, and lack of experience. I'd wager I've dropped more on ECC RAM than his entire system cost so the "you poor brah" is not an argument he would be wise to continue using especially given the demographics of AT.

Finally, a note about crash dumps -- many people do care. I don't know what your background is, but I've worked in various roles in IT over the past 20 years. Sometimes you absolutely need those crash dumps in order to determine what a problem is. True, that probably applies more to servers but there are many cases where it is needed for workstations as well.

 

[KingFatty]

Well said, I agree with everything you said. Personally, I prefer to use a page file, and I keep it on my SSD. But, I manually adjust it to a small size and let windows enlarge it if necessary (which, so far, has never happened, because the small size remains unchanged).

 

[UaVaj]

any one knows a high bridge where there is a sign that says "if you happen read this sign, you must jump off this bridge now - common sense not required". this will be a good time for such information.

yes you win for the third time.

 

[UaVaj]

different folks with different strokes. no need to get all worked up over this. this is not religion (crazy folks go to war for that). this is just computing.

if you like/need/want page file. leave it on if that make you jingle. make it 3x the maximum if that makes you feel bigger.

some of us just choose to disabled it, or make it minimum size (16MB).

bottom line is - to not experience degradation in overall memory performance. simply install enough physical ram for given task.

 

[aigomorla]

as uavaj seems highly knowledgeable... i think its fair to say he took us way out of normal and into the twilight zone... lol...

First off i can only think of 2 people out of the many i know who even have more then 96GB of ram... yet he's talking about 128GB.

About page file...
I hear mixes about it... some say u dont need it.. its an ancient relic of the past.. while others say some programs still utilize that ancient relic.

I have been fine either way... with and without page file... however i tend to leave at least a 1gig page file on reserve just incase i have an ancient program which will still call upon it.

 

[UaVaj]

Oh noes, a "Senior Member" is acting tough! Seriously, how old are you? Yeah, I'm broke alright. My cushy six figure consulting job and in-home mini datacenter really show how broke I am. How many dual or quad Xeon boxes with 128 GB+ of RAM do you have sitting around your house? None? Oops, looks like you're the broke one chump. I don't even own a machine with less than 16 GB. As a user or two said earlier, get lost. Let the adults converse.

here is the 128GB+ reference. ()