Personal computing discussed
Moderators: renee, morphine, Steel
just brew it! wrote:Sounds like superfetch in action.Sounds like the Windows disk cache.
blahsaysblah wrote:You have 16GB RAM, turn off pagefile
blahsaysblah wrote:Absolutely not for multiple reasons. I have a very good and thorough understanding of how its works and all the rare implications it has through no fault of mine.
The only safe way to use storage backed memory is when it is the primary source of memory. You obviously haven't faced the true horrible nature of the beast. I will give its name: chaos aka randomness. It is fickle and cruel. It strikes whenever it is the most inconvenient and improbable. Seriously, using a page/swap file is just a clear indication of big failure somewhere in the organization.
At best, it should be opt in on desktops.
Waco wrote:blahsaysblah wrote:Absolutely not for multiple reasons. I have a very good and thorough understanding of how its works and all the rare implications it has through no fault of mine.
The only safe way to use storage backed memory is when it is the primary source of memory. You obviously haven't faced the true horrible nature of the beast. I will give its name: chaos aka randomness. It is fickle and cruel. It strikes whenever it is the most inconvenient and improbable. Seriously, using a page/swap file is just a clear indication of big failure somewhere in the organization.
At best, it should be opt in on desktops.
I'm well aware how it works. There's no reason to do so and plenty of reasons to leave it enabled. Just like disabling swap in *nix, it's fraught with peril.
mircato wrote:superfetch, prefetching,indexing, and microsoft windows defender are all disabled already :/
Vhalidictes wrote:blahsaysblah, I used to think that way...
blahsaysblah wrote:Waco wrote:blahsaysblah wrote:Absolutely not for multiple reasons. I have a very good and thorough understanding of how its works and all the rare implications it has through no fault of mine.
The only safe way to use storage backed memory is when it is the primary source of memory. You obviously haven't faced the true horrible nature of the beast. I will give its name: chaos aka randomness. It is fickle and cruel. It strikes whenever it is the most inconvenient and improbable. Seriously, using a page/swap file is just a clear indication of big failure somewhere in the organization.
At best, it should be opt in on desktops.
I'm well aware how it works. There's no reason to do so and plenty of reasons to leave it enabled. Just like disabling swap in *nix, it's fraught with peril.
There are plenty of reasons to do so. It's just a waste of storage. Telling folks to get 8GB of memory is the faulty advice. That's the problem. Any general/semi-professional user getting 16GB RAM should never run into a problem of high memory pressure. It gives any non-technical user a bad user experience that they associate with crappy computer. If someone does run into issue, get more RAM first.
The only reasons to keep it are for kernel dumps and you should not have that issue in first place.
Storage backed RAM is why folks think computers get rusty. They should just be guided to get proper amount of RAM.
Not saying absolutely everyone. Sure, if its not worth it, or cant, to extend life of computer, go ahead and opt-in.
Sorry, ive spent a lot of time, getting applications pinned, kernel re-tuned to hit 5 9s and other requirements during maintenance,... and its near impossible. Proper engineering in first place... Swap/page file should not exist. Except as last resort, (unless costs/architecture requires it as primary RAM) is only time it should be used, not as a default option. As if its a magical cure with no negative implications.
Not saying you are saying that at all. Ive read your other posts on TR.
ludi wrote:mircato wrote:superfetch, prefetching,indexing, and microsoft windows defender are all disabled already :/
So....you disabled pretty much everything Windows would normally use to load and cache frequently-used files in order to speed up the launch of programs and features, and now you're wondering why Windows has to bang on the disk for a while to launch new instances of programs and features?
Windows tries to be like a housekeeper who notices you putting on a business shirt and tie five mornings in a row, so it starts laying those out for you the night before. At first that will result in some wasted effort on weekends, but eventually it picks up on that pattern as well and starts holding the polo shirt and flip-flops in reserve for Saturdays. Disable those predictive features and you have to go to the closet every single morning to find a change of clothes, which takes more time.
blahsaysblah wrote:JBI, i guess i see it differently. First, while you lose ability to page out working memory you also gain the fact that no transaction will have non-linear and hard to pinpoint failure.
blahsaysblah wrote:Also, most persistent memory is not "code", it is "text".
In computing, a code segment, also known as a text segment or simply as text, is a portion of an object file or the corresponding section of the program's virtual address space that contains executable instructions.
mircato wrote:ludi wrote:mircato wrote:superfetch, prefetching,indexing, and microsoft windows defender are all disabled already :/
So....you disabled pretty much everything Windows would normally use to load and cache frequently-used files in order to speed up the launch of programs and features, and now you're wondering why Windows has to bang on the disk for a while to launch new instances of programs and features?
Windows tries to be like a housekeeper who notices you putting on a business shirt and tie five mornings in a row, so it starts laying those out for you the night before. At first that will result in some wasted effort on weekends, but eventually it picks up on that pattern as well and starts holding the polo shirt and flip-flops in reserve for Saturdays. Disable those predictive features and you have to go to the closet every single morning to find a change of clothes, which takes more time.
LMAO, i guess you are pretty right but i just followed the instructions people gave me as i got zero experience with microsoft junk, so now what should i enable again in order to get less time for first run?
---
about the paging files it's set to automatic. hibernation is disabled
---
blahsaysblah wrote:Yeah, no thanks. You have outdated knowledge and assumptions. You keep yours, ill keep my actual hands on experience/research solving this very problem across multiple unixes. i meant code=execute region, text=data/read only region, in the classical computer science sense.
just brew it! wrote:Disabling the page file may be beneficial for certain specific use cases. Normal desktop PC usage is not one of them.
Ryu Connor wrote:I'm honestly surprised kernel developers still expose this option. It's clear that urban legend, anecdote, and stubbornness fuel reckless advice about the swap. Perhaps it's time for kernel developers to be the parents we deserve and take the choice away from us.
just brew it! wrote:The Linux kernel gets used in a lot of embedded devices, where there is no (or very limited) writable mass storage available. Even if it was controlled via a kernel build option, this would only prompt the "swap is evil" evangelists to post How-Tos explaining the procedure for building a custom kernel with that option turned off.
blahsaysblah wrote:You keep yours, ill keep my actual hands on experience/research solving this very problem across multiple unixes.
Glorious wrote:blahsaysblah wrote:You keep yours, ill keep my actual hands on experience/research solving this very problem across multiple unixes.
What "problem?"
How is simply having a pagefile a problem?
When you disable it, what did you "solve?"
Vhalidictes wrote:Hey, that's a decent chunk of hard drive space! On a modern Windows system, the pagefile could grow to something like 16GB, which isn't small change if your system drive is a tiny SSD.