Is 4 GB ram still enough for today?

SuperFetch works like SETI@Home or Folding@Home. SETI@Home / Folding@Home will ALWAYS take up all unused CPU cycles. If you're only using 30% of your CPU, SETI@Home / Folding@Home will use the remaining 70%. If you are using 65% of your CPU, then SETI@HOMe / Folding@HOme will scale back and use the remaining unused 35%.

SuperFetch works the same way. It will ALWAYS use all available RAM. If you have 1GB free, then SuperFetch will use 1GB. If you have 10GB free, then SuperFetch will use 10GB.

Would you tell someone to buy a faster CPU, because you don't want Folding@Home / SETI@Home to "stand up to free up a seat (CPU cycles)"? Of course not. So why would you tell someone to buy more RAM than they need, so that they don't have to tell SuperFetch to "stand up to free up a seat (RAM)"?

If you're really so concerned about application load times, then the smart thing to do is to buy an SSD (fast in every situation), instead of trying to overload your system with so much unused free RAM just so you don't force SuperFetch (only fast in cached reads) to "stand up to free up a seat".

Stay classy, Gracy.

Well, Gracy already answered with her reasoning:

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​

1. Task Manager and Resource Monitor do that.
2. Task Manager and Resource Monitor don't need to be installed
3. Task Manager and Resource Monitor don't take up additional resources while running (they are already running).

... which is apparently a little too challenging. ( 1, 2, 3). It's just a lot easier to disable a pagefile and wait for a Low-Memory / Out-of-Memory message to pop up, than it is to learn how to use a free, pre-installed, already-running Windows utility.

 

While debate can be frustrating at times, it shouldn't dive into ad hominen and veiled or overt insults.

Let's keep things above board, mates.

And let's not forget, a little experimentation goes a long way towards validating a hypothesis.

 

That was my point.
(should have used sarcastic orange)

The only situations I have found disabling superfetch beneficial was with a specific SSD on a specific firmware.

Furthermore, disabling pagefile - for me, even with 16gb ram - is a lot worse than leaving it on...Though, I won't generalize my usage to say that everyone should leave pagefile turned on with up to and including 16gb of ram.


Did that ss with 8gb because I couldn't get superfetch to cache much more ram than that, which wouldn't help me make my point. Just reinstalled OS, so I guess superfetch is still learning.

 

Simpler and faster ? How do you contineously monitor your RAM usage ?

Disable page file is the simpler and faster way, you don't need to do anything and Windows will tell you right at the spot when you are close to go over board.

And if you are really scared, you can re-enable page file at that moment.

 

Yes, I've always found it to be that simple. Tried it out a few times just to see what would happen. And that is exactly what happens. A pop up occurs letting you know that you are near the limit and that you should save data and close down a program to free up memory.

So it won't hurt to experiment.

 

There is similar concept in linux. In fact, what is linux is what we called 'fixed size page file'(instead of windows managed).

Also, Vista/Windows 7's swap out policy is now closer to linux, i.e. much lazier than before(in XP). I always prefer linux paging/cache policy over XP. It really use the page file as last resort and won't dip into it unless it is absolutely needed.

 

Just a FYI, in Linux you have swap space on the hdd; but you don't need the swap space (just as in windows you don't need page filling); everything will work without it, and it will likely work faster with it disabled

Some people might need it because they either: don't have enough ram and can upgrade (but haven't), or they have as much ram as they can but still need more (for really ram intensive things). in either case you would need a rather fast hdd/ ssd for the slowdowns to not be noticed (but any hdd intense operation will feel 10x slower when pagefiling/ swap space is used heavily because your disc is now doing a lot or read/ writes all over your disc)

 

You don't risk anything both ways! For the 7th time in this thread - nothing happens without a warning and even than - programs that are currently NOT in use are being closed without loss of information.

 

I know Windows has utility for that but you need to setup, to begin with(you can try to describe it and see how simple it is). For a server, I recommended that in another thread as that is the proper way to manage a server(it is headless) about overall memory usage for planning.

For personal desktop, disable page file is simple and easy. And no it doesn't have any impact on system stability or loss of data. you are warned well before there is not enough memory. and you can renable page file at the time of warning(and you don't need to restart either).

The cases where enable and disable page file that I am told to have a difference are that some games seem to show improved FPS with page file enabled(even though disable it doesn't cause any OOM). My guess for that situation is that they test for total memory(i.e. virtual memory) available and use as much as it can for memoization of complex calculation results. Again in such a case, enable page file becomes a must as it has a real impact on the game's performance.

 

... But they can be tolerant to having their workflow slowed down immensely by the lack of RAM and by the usage of the pagefile?

Ok if that's the case.... then let them use it the way it is and don't ask for more.

 

Ehhh... No matter if you are using pagefile/swap. Free RAM is always wasted RAM.
Nowdays, almost all OS are dynamically managing RAM, so even if you are seeing that you are out of it, it is most likely false.
Short test to prove it - do something that will eat all of your RAM and then start some app~ that needs a lot of memory, and you will see the magic.

 

I did not read the entire therad, but in this case I would ask which GPU you have..?

Because, if you have TurboCach 'feature' you'll lose a greate amount of RAM, which can make you run out of memory (again, RAM).

 

Excuse me!??

So what is the "magic" - that valuable caching data is moved to the pagefile and next time needed takes forever to load !??? Awesome magic! Enjoy the show then!

You are just the next person who does not distinguish "REQUIRED" RAM to be able to run applications at all and "OPTIMAL AMOUNT OF RAM" to avoid paging! That's the heart of this argument - people not understanding that it is one what a program needs to run and totally different what it needs to run as fast as possible and optimally.

It's like saying - my car only needs 5l/100km to drive. Yes - this is true... as long as you drive it slowly! If you want to drive it as fast as it can go - you certainly need way more than that! And that's the whole point!

Free ram is not wasted - free RAM is the reserve you have. Under 20-30% FREE ram means with pagefile enabled means pagefile is being actively used, hence your system is being slowed down - look at your "Hard Faults/sec" in Resource Monitor! Of course if you constantly have 4 GB FREE ram that is in fact wasted RAM and money as you are unlikely to ever use all of it.

Best case scenario is having 500 to 2000 MB Free with disabled pagefile. This way it is highly unlikely that you would hit OOM unless using a few VMs or heavy Fotoshopping ,etc.

 

As I said, it does not depend on swap/pagefile. Some apps are just reserving memory and not actually using it. And yes, most OS are supposed to use RAM as much as possible and then free it when needed. If you don't believe me - do the test (mentioned in my earlier post)

 

probably just trying to justify her purchase to herself....

by following her logic, the ram should be the same size as the hard drive so that nothing needs to be freed up and therefore cause a slowdown....

 

Not at all.
I keep assuming that every HDD is slower than any RAM.

 

I am not sure how you get to the conclusion by 'following her logic'. I don't see her logic to be interpreted that way.

And she was quite fraggle in purchasing RAM when I advice her to go straight to 8GB instead of 6GB.

 

This is exactly what people are trying to do with SSDs - fast storage with bigger capacity You just found America on the map

 

well then get an SSD and stop moaning about ram! you just mooted your own point.

especially with the way windows 7/vista is optimised to take advantage of as much ram as possible, even 8gb wont be 'enough'.

like many on this thread have pointed out, unless you're doing something like running mutiple VM's or working with giant photoshop files 4gb is more than enough for the vast majority of users. ever heard of a normal distribution curve? there is always that tiny percentage at the end for which 4gb wont be enough. maybe you're in it. what i dont understand is why you think that everyone else should.

in case you forgot, the title of the thread was "is 4gb ram still enough for today?"

 

moaning!?

No reason for that! I used to moan until I found out the truth. I'm trying to give it to you - you don't want it.

One day you will realize it.

P.s. Seems Windows 7 is not utilizing all the RAM available, is it My PC has not been restarted since last Wednesday.

 

Superfetch in Windows 7 is less aggressive(comparing with Vista). When you run some program that dip into the cached RAM then exit, those reclaimed RAM would not be turned into cached content again immediately.

 

I want to see where this is going to end, I've even have Superfetch disabled so

 

With an SSD(which seems to be in your sig), superfetch makes very little difference(it any). In fact, within an SSD, enable page file or not makes almost no difference(and I would enable it).

 

I've been observing my RAM usage constantly - it never reaches 100% unless I run a program that would need the free amount

 

I have all that kind of fancy stuff turned off, including page file Since I moved to 8GB RAM I never used all of it. But I've did with 4.

 

well there you have it, since your screenshot shows only 2gb being used, you dont really need more than 4gb. lol

and again, you choose to ignore the title of the thread while enthusiastically trying to "give" your "truth"

 

That is the problem of relying on Task Manager to gauge the memory need. She can only show you one snapshot and if you read her thread, it is very clear that her normal usage pattern exceed 4GB.

so LOL.

 

i did read her other posts, and its clear that her usage is not representative of the vast majority of users out there.

read my post re: normal distribution.

also re read the title of the thread.

hint: its not "is 4gb enough for Gracy123?" neither is it "is 4gb enough for power users"


also, if she's using 2gb in "normal" use, is the cost of exceeding 4gb for those (perhaps few) occasions a mainstream user (or perhaps, even Gracy123) justified?

this was the original question, and clearly this topic has strayed waaaaay beyond....

 

I was talking about you trying to gauge 'HER' memory need by looking at the snapshot of her Task Manager which is LOL.

There are a few comments she made(concerning THIS thread) which can be summaried as:

1. for relatively 'advanced' user, 4GB may not be enough which I agree is just her own experience and opinion and I tend to disagree(as I think most people don't need more than that, even for 'advanced' usage).

2. the easiest way to find out is to disable page file, which happened to be my suggestion to her

3. the optimal memory would be X + 20%(kind of) with page file disabled, IF ONE REALLY WANTS optimal performance(assuming HDD is used, don't drag in SSD please). Where X is the 'required' memory amount(i.e. the MAX commited memory for a usage period of say a week or a month depending on individual). The 20% extra is for cache and occasional spike.

 

thank you, thats a good summary and put very articulately unlike some of the previous posts.

 

Yep, I also think this is Thread number 25 for the last few weeks where the only conclusion is:

"Do whatever you want to, the truth is out there"

I found my truth, I wish you all do the same.

If you allow me just one advise - don't take things for given - experiment and decide based on actual results and not on someone's theory.

 

I think this topic has been sufficiently dissected. Plenty of info exchanged.

Concluding posts were very good.

Thread closed.