Steam 설치
로그인
|
언어
简体中文(중국어 간체)
繁體中文(중국어 번체)
日本語(일본어)
ไทย(태국어)
Български(불가리아어)
Čeština(체코어)
Dansk(덴마크어)
Deutsch(독일어)
English(영어)
Español - España(스페인어 - 스페인)
Español - Latinoamérica(스페인어 - 중남미)
Ελληνικά(그리스어)
Français(프랑스어)
Italiano(이탈리아어)
Bahasa Indonesia(인도네시아어)
Magyar(헝가리어)
Nederlands(네덜란드어)
Norsk(노르웨이어)
Polski(폴란드어)
Português(포르투갈어 - 포르투갈)
Português - Brasil(포르투갈어 - 브라질)
Română(루마니아어)
Русский(러시아어)
Suomi(핀란드어)
Svenska(스웨덴어)
Türkçe(튀르키예어)
Tiếng Việt(베트남어)
Українська(우크라이나어)
번역 관련 문제 보고
As the first two replies states, the page file is always in the picture because it's an innate part of how Windows manages RAM, even if you "disable" your page file (and don't). Windows uses it as a backing store or as a table of contents of sorts for your workload. Many things "on" it are just pointer references to where something already stored on and read from the disk is, so it's not like it's constantly "writing" everything you do to it like a lot of people mistakenly believe. And be aware that most workloads carry a commit charge that exceeds your actual memory use (sometimes by a lot), so don't disable or set a small page file unless you want to potentially further reduce the efficiency that your system can use its resources.
Just using the majority of your RAM isn't too telling on it's own as to whether or you need more RAM, although if you're frequently using that much while using multiple programs, you'll likely see reduced overall system responsiveness at times and more would benefit. If it's doing a lot of hard page faults and you're seeing reduced performance, then you know it's past that point.
You're not just looking at the committed memory, I hope, because that's something else and will almost always outpace the amount in use (and on systems with lower capacities of RAM, such 8 GB, yes, I can see this value approaching an amount that equals the RAM capacity even with a low workload such as shortly after boot because... 8 GB isn't a lot of memory anymore).
Likewise, the size of the pagefile.sys file itself isn't too relevant to showing how much it's actually "using" so I hope you're not simply looking at that either.
If you're not having a flurry of continued hard page faults with reduced performance while having low amounts of RAM left available, well... there's no need to worry about this. And if you are, then you need more memory.
People seem to like to over-complicate this one for themselves.
What, specifically, were you looking at in task manager? I wonder if you were mistaking the committed memory for something it's not. This value isn't merely reflective of "how good/how much the OS uses the page file" but rather it's a result of how much memory allocation your workload does.
With Windows 10, and the page file set to system managed, then on a PC with 16 GB, then you should start with a commit limit a few GB above 16 GB. Reason being, on Windows 10 (and I think 11?) the initial page file size will be a bit less than about a quarter of the installed RAM size. You'll observe this by seeing a commit limit ~18 GB or 19 GB. Notice I say initial. It will raise above this when it needs to. 16 GB isn't extraordinarily difficult to use enough of to where this might happen. So if it's raising, it means your workload is allocating enough memory that it's closing in on your commit limit, and the memory manager is doing its job and raising your commit limit to allow further allocation. Those values are not an indicator of how much the OS itself "uses the page file". It's workload related.
If Windows 10 is actually setting your commit limit 16 GB above your installed 16 GB at boot with a system managed page file, then I find that odd (unless you have it starting up with a ton of programs that's causing it to allocate well over 16 GB). I don't see my initial commit limit 16 GB above my RAM capacity, and I have 64 GB. It should be less than a quarter of your installed RAM size, and then growing only if your workload needs it. So I wonder if there's something unusual with your example.
One other very important variable is whether you're using shutdown or restart (at least if you don't have hibernation disabled) because shutdown isn't a proper shut down and won't "reset" those values. There's a lot of variables like that at play and if you don't "control your variables", you tend to get non-comparable results.
Commited memory shows the amount of RAM and the size of the page file installed and commited.
The page file is a very important system component and there are different philosophies on how to deal with computer limitations.
-The resizable page file allows the operating system to have control over the size and the idea behind it is that each system is different (software and hardware), so we have to make each system to be stable. Another benefit is that it keeps the system using less disk allocation when not needed.
- Fixed size pagefile is the opposite of the resizable one. It is unpractible to be applied for all systems. If you know your system memory requirements to keep it stable then you will be fine. The fixed pre-allocation will use more disk space, but removes all resizing operations.
Windows default is somewhere between the two.
That is not what committed memory is. A commit is is made when a program asks for memory. Windows will then 'commit" an amount of RAM to that program. No other program can take that allotment. Since both Windows and modern programs are inefficient more memory than needed is always committed.
Because of this and the fact that Windows will never let other programs ask for that space regardless of whether it ever gets used or not you can end up with as much as 56% free RAM (as much as I've seen in person) that can not be used but is and always will be empty until programs are killed.
Linux on the other hand lets you actually use your RAM regardless of commit amounts on top of ZRAM offering a 2:1 compression of memory.
There isn't truly free RAM when most of the time it will be used as standby RAM.
Windows already use compression methods, but is it any better?
Committed memory is the amount of memory that has been requested to be allocated but not necessarily faulted in yet.
For the last time, this value is largely workload dependent, not merely a reflection of "how much the OS itself uses the page file" so your assumption that Windows 10 is many times worse at "page file use" than Windows 11 is erroneous.
"But I saw it" you say?
Ever consider that there might be other variables at play that can impact those values? Ever consider that your example in one of those cases may not have been typical? People don't like ask themselves those hard questions for some reason.
I'm not telling you that you didn't see what you saw. I'm only telling you that the conclusion you're arriving at because of it is flawed.