Running out of RAM

[Mussels]

Unless the user changes the default setting to not use one. Then it won't create a new one

That's exactly what i just made clear is not how it behaves. It always has a page file, that setting just tells it to delete it when its done.

Which is terrible, as it crashes some programs that need it's space available instantly, and adds wear to SSDs. Why have it write the same shit over and over, when it could write once and read many?

 

[Bill_Bright]

That's exactly what i just made clear is not how it behaves. It always has a page file, that setting just tells it to delete it when its done.

We agree - in principle, in part. You are correct that Windows will still page/swap out data to "a" cache. I am saying it will not create a new "pagefile.sys" file on the disk if the user deletes/disables it.

But if you disable the PF, Windows does not create then delete a new one every time it needs a cache. If you claim it does, please provide a link to support that claim.

and adds wear to SSDs.

That's an old, outdated, and obsolete argument that needs to go away. First generation SSDs, okay. With today's SSDs, no. Even datacenters are using SSDs to cache their most frequently accessed data. It would take many many years, writing many gigabytes, day in and day out to a SSD for a normal user to "wear" it out (exceed its write limits).

Why have it write the same shit over and over,

It doesn't. Why would it be different on a SSD than a hard drive? It wouldn't. And consider this - if not paged, the system must "save" the data back to the disk if it needs more room. Then, if it needs the data again, it must re-open the file which then "writes" a new temporary file to the disk. You are not saving "writes" at all!

when it could write once and read many?

Which is exactly what it does.

I think it important for everyone to remember that major notebook makers like ASUS, MSI, Lenovo, Dell, Apple, Acer and others been putting SSDs only in their notebooks for years now - and the vast majority of users are running with the default settings. And for sure, those SSDs are not wearing out due to excessive writes to their default page files.

 

[Mussels]

That's an old, outdated, and obsolete argument that needs to go away. First generation SSDs, okay. With today's SSDs, no. Even datacenters are using SSDs to cache their most frequently accessed data. It would take many many years, writing many gigabytes, day in and day out to a SSD for a normal user to "wear" it out (exceed its write limits).

That's not true, modern budget SSD's are still trash and wear out fast. I deal with them constantly, especially WD greens that are dead in under 6 months from casual users doing dumb things.

Which is exactly what it does.

Not if you disable it, which you clearly ignored what that was a reply to

 

[trparky]

especially WD greens that are dead in under 6 months from casual users doing dumb things.

And when you say dumb things, what do you mean?

 

[SchumannFrequency]

Your rant is appreciated, but it also underlines why so many think they can outdo MS. Why is such a commit limit in place? MS requires you to either sacrifice notable storage space just to use your RAM, or face annoying errors like this. Linux doesnt ahve this issue, and neither did MACos back when you could upgrade. I can use 99% of my system RAM with no swap and no errors. Of course, that should be the default for msot users, but why is there no option to get around this?

It's also brining back bad memories of XP and it's LOVE of swapping as much as possible to pagefile. I do remember making RAMdisks with excess sytem RAM and forcing the pagfile onto them to get around this, but it was something just dumb....

Alpine Linux and OpenBSD in particular do well in terms of RAM usage. Additionally, you can use a WM to use even less RAM.
When I use 'top' to check the RAM usage of my fully featured Alpine Linux + river wm (Wayland window manager) desktop I see that only 58 MB of RAM is actively being used.

 

[Bill_Bright]

That's not true, modern budget SSD's are still trash and wear out fast. I deal with them constantly, especially WD greens that are dead in under 6 months from casual users doing dumb things.

It is true. I deal with them too.

You are correct that budget SSDs are trash. But guess what? So are budget hard drives, budget power supplies, budget CPU fans, budget keyboards, and budget doohickies too.

Casual users doing dumb things also destroy top quality CPUs, motherboards, RAM and even cars.

None of that has anything to do with the SSDs "wearing out" which, when it comes to SSDs, wearing out means reaching their write limits.

which you clearly ignored what that was a reply to

I didn't ignore it. I said it is not true. That is, when users disable the PF, Windows does NOT constantly recreate it, then delete it - which you incorrectly claimed, and where you ignored my request to provide some supporting references for that claim.

If you can provide a technical white paper or the equivalent that supports your claim, I will gladly concede you are correct and apology. I cannot find any such supporting evidence.

You also ignored the fact that laptop makers have been installing SSDs in laptops (even budget laptops) for years now. And yet there is no evidence those SSDs are reaching their write limits prematurely either.

And for the record, the act of creating a page file does NOT suddenly fill it up with saved data. Data is written to it when the system needs to spool data out to it EXACTLY as would happen if the PF was already created.

And even if the creation of the file tables causes a few writes, the TRIM and wear leveling features ensures no single storage location is constantly being written too over and over again just to avoid excessive wear on the same locations.

I might suggest you read up and understand the "endurance" specs for SSDs known as TBW - Terabytes Written. Here is a good article on it. Note the following (my bold underline added),

Although TBW is a reliable indicator of an SSD's endurance, most regular computer users will never reach TBW during the normal lifetime of a drive. So unless you are writing hundreds of gigabytes of critical data each day, you don't have to worry about TBW.

Another metric to learn about is DWPD or drive writes per day. DWPD measures how many times you can overwrite the drive’s entire size each day of its life. For example, you have a 500GB "consumer grade" SSD with a 3 year warranty. If its DWPD is a rather typical 0.5, that means you can write 250 GB (1/2 its size, one time) into it every single day for the next 3 years.

If you multiply that out, that’s 250GB × 365 days × 3 years = 273.75TB of total writes before you may need to replace it.

What "normal" user writes 250GB to their SSD EVERY DAY for 3 years straight?

 

[trparky]

Although TBW is a reliable indicator of an SSD's endurance, most regular computer users will never reach TBW during the normal lifetime of a drive. So unless you are writing hundreds of gigabytes of critical data each day, you don't have to worry about TBW.

That's true. I've had a Samsung 980 SSD and so far, I've only written 43 TB to the drive in the year that I've had it and I've done nothing different in how I use the drive. According to Samsung's own specifications, the SSD has an endurance of 600 TBW.

 

[unwind-protect]

Not having a page file encourages more paging.

You have to understand that there are two kinds of paging:
1) anonymous or modified pages. Those are moved to paging space and back in from there when needed. You can suppress this paging activity by not proving paging space.
2) file-backed, readonly pages such as code and uncompressed data. These are simply dropped from memory and reloaded from the original location when needed again. You cannot suppress this kind of paging.

The thing is that if you don't provide paging space for 1) then you have an increased amount of 2) paging.

Whether that is overall an advantage depends on workload and OS smartness, but in general it is fair to say that it is good to provide 1) space so that actually useless stuff can be moved out and in return more useful 2) pages are kept resident.

 

[LFaWolf]

Since this thread has turned into one about page file, I have a question - if I have 96GB of memory, and I dedicate say 16GB to make a ram disk, and then create a page file of 8GB or 12GB on the ram disk, would that speed things up some more? Or, would it be too late since the ram disk is created after the OS has booted, and Windows has already created a page file somewhere else already? Does anyone know? Just curious.

 

[trparky]

Or, would it be too late since the ram disk is created after the OS has booted, and Windows has already created a page file somewhere else already?

Probably yes. You'd probably have to create the RAM disk before the OS even boots up for this to work the way you want. You'd have to play around with Hypervisors to do this task.