Is 4 GB ram still enough for today?

Wow. You are correct.

I never noticed that. I never noticed any kind of performance hit or slowdown that you typically get when paging to disk.

 

I dont have page file - I turned it off even when I was with 2GB and XP. At some point the need came (one of the bigger games .. as I remember) and I went to 4GB RAM. But I've always hated the page file, and I'm not the only one. The page file is a simple work around for not having enough RAM to begin with.

 

I'm just thinking... all these people complaining about the page file possibly use one without noticing it....

I am right now thinking about Photoshop & Gimp.

My mother once tried to crop an image - she couldn't, because stupid GIMP decided the image was too large to edit on her computer and just crashed...

In contrast, on my Vaio I can edit just about any Photoshop file - 10MP, 21MP, 100MP - all work fine.
Why: Because Photoshop uses it's own Temp file which is pretty much the same as a Pagefile -> it increases the memory space available to it, and the temp file can get quite big - several GB.

Another program that makes heavy use of a temp file as far as I know (and not RAM) is Microsoft ICE - I once merged images to a 700MP image together (nothing interesting though, but ICE is really fast, took a little under 1 hour I think) - doing that if I remember correctly (summer 2010) it used something around 40-70GB or SSD space -> tell me you'd buy that much RAM.

Why limit yourself if the temp file is painless.

It's as if you buy a mouse but then never use the scroll wheel because you can just drag sliders around in windows - there is no reason to use the comfort of a scroll wheel.... -> same with the pagefile.

And I am still waiting for a reproducible test that shows an increased performance without a pagefile.

 

edited because there's no point in furthering this discussion

edited again to turn this into a useful question (for my own purposes) @kent1146: Do you notice a performance difference in enabling / disabling superfetch? OCZ recommends turning it off to increase performance. I'm only asking because you list a vertex 2 in sig.

 

There isnt one and never will be. Disabling the page file is only for those who like the placebo effect and like the possibility of crashing their system if they do intensive work.

 

Sorry, Grace, but that's a fairytale. I have 16 gigs of RAM in my M6400 and, guess what, I still have a page file. It's nice to know that if one of my simulations runs into the woods, my system will remain up and running

Anyhow, turning off my pagefile does not affect the performance of this machine one bit, despite the fact that I certainly "have enough RAM to allow this".

 

Speak for yourself only.

I don't think it is fairytale. Once again, don't extend your experience to EVERY situation. It does not affect your machine performance doesn't mean it doesn't affect others.

I know what you are talking about but noop, we are not talking about the same thing. Both are a way to handle the case of where to store large amount of data. One is managed by the system(i.e. virtual memory backed by unlimited size page file), one is managed by the application.

Enabling page file however has the effect of it not just affect photoshop but every application. If you want, why not try to help her solve her problem which she detailed in another thread. The only remedy seems to be disable page file, with everything else keeping the same.

 

You do realize that some applications require a swap file to function properly, regardless of the amount of ram and system environment you run, right?

I can think of exactly two benefits to disabling page file:
1. Lower read/writes which you may or may not find worthwhile if you're using a SSD - I won't try to tell anyone which is "best"
2. You get x bytes of data back as useable storage space on your drive

There's a lot of opinions regarding basically everything that has been discussed here, and it's important to note that anyone and everyone should really take the time to look into the potential side effects of tweaking their OS.
On topic, I think it has been sufficiently proven that it does not hurt to insert another DIMM in your mobo; but again, take the time to look into whether or not you need it or you may be spending money on something that YOU do not require.

Considering the OP has had basically every opinion possible as a reply, I'd say the question would be better answered by doing the research yourself. Most articles will provide some sort of testing or evidence to support their claims and are typically written by people who write about hardware for a living (meaning they probably know what they're talking about).

 

Which is why I asked for such a test.

It's the same idea behind it though.
-> In a perfect world all these programmes like Photoshop wouldn't use a temp file at all, they could use RAM and the Pagefile....

However, people switch off the pagefile and then these things would stop working. Imagine how many people would phone up Photoshop if the software didn't work because THEY disabled the pagefile... plenty, at the same time THE users themselves would have broken he OS/program.

-> There are always people who claim to know the OS better than the manufacturer - self proclaimed gurus who know nothing at all.
Or does anybody here know all the lines of code in Vista/Win7 and know how the OS works and interacts? Answer no.

Something started back in MS DOS where programmes where a few KB and now it carries on regardless of that...

 

gracy123 had shown you the case in her thread.

 

Actually, she has shown nothing at all except that:

a) it works for her (right now), which isn't a universal statement
b) the software she uses is able to handle out of memory events without an issue (which again cannot be guaranteed for all software)

 

No one is saying it is a universal statement.

She described her issue, why don't you just give it a reasonable explanation(why enable page file caused her lag) rather than just keep on saying page file has no effect on performance ?

We are not talking about OOM situation here. We are talking about whether disable page file or not has an effect.

We all know different usage scenarios calls for different measures. Some games as I am told simply refuse to run without a page file and for that case, turning off page file is not an option, no matter how much memory is there. Some programs may have spike on memory usage which cannot be measured and if this is a 'must run' program, it is again a must to have that 'insurance'.

 

Well, for me I have to have page filing disabled or else I run into lots of slowdowns (ever tried using a computer with a 1.8" hdd that has an average sequential read of 33MB/s normally, with page filing enabled? no? well let me tell you, it is SLOOOW. Everything else about this computer is just great, and very fast, with great battery life) sure not everyone is in this situation, but even with a 2.5" 5400/ 7200rpm drive, there are plenty of people that would see a good "boost" in performance from no page filing, and plenty of ram.

I don't use page filing in Windows (on my laptop), and I don't use swap-space when using Linux (on my laptop); they are basically the same thing, and aside from the couple of OOM's I got in Windows because I had only 4GB ram and needed more (while having page filing disabled because of slowdowns), I haven't had one issue... not one issue (that is including on my HP Tm2 that I had for 6 months prior to my current computer, which also had page filing disabled, but only 6GB ram)



Also, about Photoshop and Windows, they both have the ability to use a device other than the main hard drive for page filing (windows) or as a scratch disc (Photoshop), that doesn't usually help anyone though... and Photoshop utilizes both ram, and its own "scratch drive" space for what it does (at least that is how it is with Photoshop CS4/5)

 

I'm not a power user or anything, I use the Vaio CW in my sig, with 4GB's. I play JC2 in windows mode, and I never really go over 75% memory usage. In BO as well, never over 75%

4GB's is plenty for most people, I had 8GB's in my Macbook and never used it.

 

Even that lag claim has no proof.
Some days I fell as if it takes ages to get to Sheffield, some days I just seem to be there. I guess the train travels quicker on some days.
-> No it doesn't, it travels at the same speed on the same timetable - I arrive around the same time every time.

Give me a test that proves disabling the pagefile has an advantage - a repeatable test - and if it's hardware affected, list YOUR hardware so someone with identical hardware can test it.
Humans are useless at perceiving time or economic benefits - if you believe it is better it will feel better, even if testing showed you it was worse.

 

Does this thread of Gracy123 fit the bill? This doesn't seem as much to be a running out of RAM situation per se, as much as it seems that Windows 7 likes to briefly cache file transfers in RAM, which means that if you have a pagefile, your HDD gets a double whammy as it simultaneously tries to read and write from the HDD (for the file transfer) as well as transfer data from RAM to the swap file, which can result in a large slowdown. Without the pagefile, it just dumps stuff from memory instead, and doesn't have any extra HDD access. Newer HDDs and SSDs would probably mitigate this somewhat, since they're designed for better multiple access.

 

That doesn't really help -> copy means -> copy into RAM then past wherever you want to.

On this note - I do actually think the windows right click drag, select item from menu & Ctrl + x/v/c commands are different in Explorer.

 

It's still a case where disabling the pagefile had a measurable advantage, which is what you asked for, is it not? And I think it may have been a drag and drop sort of copy... she's talking about 15 GB of files that are being copied around, and she "only" has 6 GB of RAM...

 

I've never had an issue copying even more files with the drag and drop method - even when I was on my HDD.

-> So I don't think that's exactly the cause.

I'd actually think she used Ctrl + c and then Ctrl + v -> that's actually the most natural thing to do because it doesn't require two visible windows.

 

since when does a file copy ever go into ram? it didn't since dos. the only thing that gets into ram on ctrl-c is the file-LIST.

disabling the pagefile can't have any measurable advantage except on xp, which sucks anyways.

 

Seems that that is the case in Vista/Windows 7. I don't have an XP to test at this moment. The copy process may have gone through quite some changes which may explain what gracey123 saw.

I still haven't seen one explanation why disable and enable page file in her case make very noticeable difference(other than that the copy process is doing something to the memory and/or cache) and I have no reason to believe that she was lying or she have tweaked her system in weird way.

 

On the topic of file transfers, both of the following links are interesting reads:
Discussion of the Issue - Windows 2008 R2 - large file copy uses all available memory and then tranfer rate decreases dramatically (20x)
A Potential Solution - Slow Large File Copy Issues - Ask the Performance Team - Site Home - TechNet Blogs

 

@midnightsun

thanks for the links. While it is not exactly the same as gracey123's case, it still points to the same issue of huge file copying can 'jam' the server and it seems that in her case, enable page file would make the situation much worse.

 

I wasn't here for a while but yes - I think I explained the test-case quite clearly with enough arguments and evidence-materials! I guess some people just don't want to read - they take the first and last post and that's it!!

And guys - there is NO universal solution! There are 3 options - Pagefile enabled, limited OR disabled !

Depending on software and usage - each one of these 3 solutions can be the best one for certain group of users and system configuration! No doubt about it! I found for myself that the third option, when there is enough RAM available is the best and most productive solution!

And because I consider myself a power-user, I tend to think this solution is applicable to a very big number of people here!

And for the 100th time - don't, oh please don't talk about stuff you guys have never tried! I am tired of people trying to convince me that 4GB is enough and that the pagefile needs to be enabled before even trying the opposite! The difference is - I've been on each end of the story and know the difference - you with your 4GB RAM and enabled pagefile - DON'T!

Now give me the advantage of experience and allow me to be more certain that you can ever be, before you upgrade your RAM and disable the pagefile!

Use both methods. Makes no difference to me.

Regarding this issue - here is what and how helped me:

Upgrading RAM to 6GB - solved about 70-80% of the lag
Disabling the pagefile - the rest 20-30%

 

-> Then something is wrong.

 

Yes - the pagefile usage and caching combined with lack of RAM (4GB available).

Not anymore though - my system flies "like a G6" with 6GB RAM and pagefile off.

 

Nope.
You've just covered up the symptoms.

Plenty of people can transfer large amounts of files with the pagefile on without an issue. If that didn't work for you then there is an issue with the system somewhere -> it should NOT lag.

 

Funnily - no. It does the same on all other 3 systems with Win 7 (x64) I have experimented. (2 x Lenovo and 1 Toshiba)

On the other hand the lag was very limited on my Samsung Netbook with Win 7 32bit and 1 and later 2GB RAM.

I have expected my system any possible way - replaced HDD, benchmarked every single component. The only solution - more RAM and less (better no) pagefile.

But again - absolutely no way to know the difference before you try both scenarios! One day you decide to do it you will realize you have been living in a world much different

Facts are above Theory.

 

You're contradicting yourself

-> The netbook with less RAM fullfilled the task better than your other systems with more RAM -> sounds dodgy, doesn't it?

 

Not if you pay enough attention (and you often don't).

The Netbook is running 32 bit windows!

 

Then it's a bug in Windows and you still haven't solved the issue.

-> If a wheel on a lorry has a problem (one with several axis) - would you just remove the wheel? No.
-> You replace it, if it goes bad again you search for the reason.

 

Possibly. Yes. But I don't work for Microsoft neither am I a software developer. I am just a Windows 7 64 bit user as everyone else. and found a solution to the problem by taking Windows for something given the way it is - with all advantages and disadvantages. I can't change the OS, can you?

I can only find a solution to the problem and I did. At the same time the solution did and does not have ANY negative influence - only positive.

So go ahead and try disabling your pagefile. Leave the theory aside It will either increase performance or decrease it. Only one way to find out.

But we are going off-topic here. My initial thought was that disabling pagefile can very well tell you whether youy system will benefit from more RAM or not!

If it can as a whole work with the amount of RAM you currently have installed and no pagefile - you are fine. If it starts hitting OOM - it does ask for more RAM. pagefile is nothing else than a secondary, much slower RAM. Whenever the primary isn't enough - pagefile is used.

So why would a system (as a whole) cause problems with pagefile disabled if you have enough RAM? It won't.

1 Month since I disabled mine now - not a single issue except the one time I actually ran out of RAM.

 

I require theory -> or at least logic (I'm a maths student)

Also, with an Intel SSD I cannot gain anything from a disabled pagefile.
In fact, with 3GB of RAM it's highly useful, especially considering my BSc Project Program can easily eat up a few hundred MB of ram (huge array, depending on user input, the actual .exe is 520KB )

 

You never answered a question I posed earlier:

You are advocating turning off the pagefile, so that you know whether you run out of RAM or not.

What is the benefit of your suggestion, versus just running a utility that tells you how much RAM you are using?

 

You have set your mind in stone that disable page file or not would not make any difference.

Therefore, there is no point of finding any proof anymore.

 

Disadvantages of running an utility:

1. You first to find an utility that would give you accurate and reliable information
2. You need to install it
3. It does take up some resources too while running

And why do all that instead of just disable a feature which is already there !?

It's like asking the car-mechanic "Why are you asking me to hit the breaks to measure my car's breaking power... can't you just run a software that would tell you what you need?"

Would you tell me what you find dangerous or scary about this simple procedure!? You seem to be afraid to be proven wrong this way I fear...

I'm not a medical doctor myself either as you can see..... But I still trust the obvious facts more than someone's theory. The eyes need to see, the hands need to touch. The rest is just someone's imagination that might be or might not be true. Science is actually based on experiments not on theory alone. Or do you believe Viet came up with his formulas without experimenting?
Those who only study and repeat become just high-school teachers, but never inventors

 

Cant be true since its a 32bit executable. So it cant use more than 4gb of RAM even if you had it.

 

-> Actually, most of maths came about just from theory
Heck, theory predicted particles that weren't known to exist prior to conducting experiments to find the predicted particles.
(Particle Physics)

Maths is the highest scientific discipline -> if it doesn't follow the rules & logic of maths it's false - the only alternative: the theory to explain it does not exist yet. (and maybe one case could be the free will debate)

Considering that computers are based on mathematics and logic, they will follow the rules of mathematics and logic too.

 

Windows Task Manager and Resource Monitor gives you accurate info, dont need to be installed, and dont take up extra resources.

But seeing how you misread my Task Manager screenshot and didn't know the difference between Free RAM and Available RAM, I can understand why you prefer to disable the pagefile to figue out RAM usage instead.

 

This is only partially true. Most inventions were first developed by experimenting hundreds of times and then explained with equations! Mathematics just explains phenomenas - it doesn't create them!

But you want explanation with mathematics? Here a simple one:

4-6 equals -2. This is just fine as long as you allow negative numbers (pagefile) in your equation. However if you define your function as positive only (disable pagefile), you will very quickly become aware of the issue once 4-6 start giving you an error yielding a negative result, which is not tolerated.

I couldn't have put it easier, could I

Now just remember that negative numbers are the "very slow" numbers in our case as they are placed on a very slow storage media - much slower than any RAM! You can run your system with 2GB RAM too... your function will result -4.... and as some say it will still run "just fine". But what we all want is not "just fine" but excellent. And you can achieve that ONLY if your equation yields results >=0.

It's actually you that doesn't understand the difference! You say Available is RAM that can be freed up... and this is true!

But when an old person gets on the bus and wants to take a seat marked "available for old people" - someone needs to stand up in order for that to be possible... and where does this person go? He "hangs" standing up. So does your system.

Free are only the free seats and that's what's important. Every other seat, even being "available", requires sacrifices! And you sacrifice your system speed this way because the passengers that cannot sit down, (even if they don't have to, as they are not ill or old) need to stand. Standing means either pagefile or getting off the bus and needing to get on the next one whenever willing to continue the journey. Simple as it is.

To be able to comfortably transport all your passengers - you need to ensure enough FREE seats! Not available... but actually FREE!

 

Why is this being debated so much? I mean memory is cheap enough, assuming you are using DDR3. If not, any DDR2 machine probably wouldn't benefit from more than 4GB anyhow. If you can afford it, buy it, no worries. I mean $65-$75 and you've got 8GB, done deal. Sell your 4GB for $25 on eBay, and you've made a $40-$50 investment.

 

What a silly debate.

Also, just because I can, math is not the "highest scientific discipline" and people seem to always get confused about that. Math does not determine physical law, it's the other way around.

Anyway, I'm unsurprised a simple and vague question about RAM turned into such a weird debate. This happens a lot on this forum I've noticed.

 

According to some, I'd better put my other two DIMMs back in!
8GB isn't enough???




Oh, wait....superfetch.

 

If that is your typical RAM usage - yes, your system would benefit from more RAM.

 

I've got best solution for everyone here - install Linux and forget about pagefile etc.

 

ehm...
Yes, some maths comes from experiments - but the maths needs to come from maths because experiments will ALWAYS add an inaccuracy due to measuring errors.
I.e. an experiment can point you the way - but it cannot give you the final solution.

Regarding your function - I think I know where you want to go, but you can't try to explain it like that as you never defined a function

-> So let's take the simple case, y = mx +c with c < 0
This function is positive for all x > c/m (if I did my maths right) is 0 for x = c/m and smaller 0 for x < c/m

Now in your case, if you only want positive results you have two options:
Only define the function for x>=c/m or alternative define the function as
y = absolute( mx + c)

-> Both will give you a positive result only - you need to decide which point applies to you.
Although I'm not exactly sure how this applies to your pagefile issue, as that's not a fucntion.

 

Hasn't Linux got a swapfile? The Suse Install on my old laptop (because I can) has one....

 

Swap in Linux is used only when RAM is filled.

Even when I'm compiling kernel while browsing net and watching fullhd movie on ext. screen, I can't hit more RAM usage than ~1.5GB, so practically, I'm using swap only for hibernation.

And, answering topic question - Yes, for most of users, even 2GB would be enough nowdays, IMO.

 

Guys... why is it so hard to understand that freeing up RAM sacrifices something!!! Whether it is cache, superfetch or something else - SOMEONE NEEDS TO STAND UP TO FREE UP A SEAT!

So now guess what happens when you "free up" superfetch data??? Your system is slowed down! It is so simple and yet you still deny to understand it or just don't want to admit you are wrong! Come on guys, I'm sure you are more intelligent than you pretend to be...!

Distinguish between
" RAM required for your applications to be able to run" - which would be the "Hardware Reserved" + "In Use"
and
" RAM required for your system/applications to run with optimal speed and performance" - which would be the Total RAM available minus FREE RAM.

In order to observe the actual amounts though you need to disable the pagefile first, thus forcing applications to use the RAM only.

 

SuperFetch works like this:

You anticipate what work your boss will ask from you. You do this work in your free time - inbetween the tasks that he specifically asks for.

When he asks you for work which is already completed (partially or fully...) you hand him the results and you look like superman.

When he asks you to do a task that was not anticipated by you, you simply complete it in 'real time' and it takes the same time to complete whether you had done any other work in 'anticipation' or not.

In no case does the work get done slower whether or not other tasks have been completed. In some cases more work is actually done - and - noticed too.

This is exactly how SuperFetch works.