Nasty Memory Leak...running Rambooster in the meantime.

I have the nastiest and elusive memory leak right now. I have no clue what is causing the problem. I have been troubleshooting it for a few days now without any real success.

Problem is it is a SLOW memory leak and I have tried to isolate the offending software by either disabling and/or uninstalling everything that is running while the system is on.

What is most annoying it after leaving the system running overnight it will go from about 550MB - 700MB of swapfile use to around 1.2gb and availible RAM will drop from about 920MB free down to around 250MB free. All the time w/o any applications running. Of course there are services and whatnot running in the background but none of them show an abnormal amount of memory use. But they all increase slightly. IT'S WEIRD...

The only way to recover the memory (RAM and the SWAPFILE) is to reboot the system.

So tonight I decided to try running an solid old utility from days gone by...Rambooster. I d/l'd v2.0 and have it running and I also set XP to manage the swapfile itself. I had set the swapfile to a static size of about 2gb to decrease the overhead needed to manage the file.

We'll see what develops tonight...maybe there are some things I have not tried...like I say I have killed off gmail/googlenotifyer, uninstalled Java runtime, uninstalled skype, evernote...even zone alarm...and the leak still happens.

I am suspecting there is a library that has something wonky in a recent update. I could have rolled back but I really am bad these days about keeping a system changes log book so I would have no idea where I could roll back too. And really I would rather re-install everything instead.

Though, for me Rambooster was long one of the best & most useful utilities ever written for Windows, I had stopped running RamBooster a number of years ago because, well, XP had become so stable I dint find a need for the utility. Now I am hoping it will do a good job to keep system resources free so i am not needing to reboot 2-4x every day. It is just impossible to be productive doing that...mostly because I am just cranky over the leak to begin with... hehehehe...

I have RamBooster set to try and keep 800mb of ram free as a "mendoza line" we'll see. I just did a stress test setting by asking it to keep 1gb free and it free'd up a total of 1.23gb while keeping the swap file at just 545mb. It was funny to watch the SF jump to 1.6gb though...never have seen it run that high...but it dropped right back down.

We'll see how it goes...

 

WOW...I just noticed that after I ran the optimization, they System cache dropped from about 500mb to 60mb. I looked at the amount of RAM each process is using and boy are those lower. Highest amount of memory being used is 21mb by IE. Even XP Explorer is at just 5.9mb. Sure it might take a moment for an app to spring to life but I can adjust...maybe...we'll see. Just thought I would add that observation.

 

First, no, don't, repeat don't, DON'T use stuff like RamBooster.
It has *no* beneficial effects whatsoever. All it does is push a ton of data to the pagefile, which, as you've spotted, forces apps to wait for their data to be *reloaded*. Once that is done, it'll be back exactly where it was.

There is absolutely no point in artificially preventing your RAM from being used, which is essentially what you're doing.
The application does not free memory or anything. If you have a memory leak, it'll still be there, it'll still have swallowed the same amount of memory. You're just pushing random data to the pagefile, forcing Windows to stall programs while it pulls the data back into memory.

Apart from that, does it keep eating more memory, or does it stabilize? If it stabilizes it doesn't seem like a memory leak.
And which apps are doing this? If it is a memory leak, it is a problem in *one* individual application (or in a library which happens to be used by multiple apps)

How much RAM do you have? Windows will be less aggressive about releasing freed memory if you have lots of RAM.

You might want to take a look at the performance viewer under Administrative Tools. Get it to display memory and pagefile usage for all processes

I don't know if there actually is a problem. Do you encounter any problems (lowered performance, crashes) because of this?
But I can tell you that using RamBooster has *no* effect other than artificially slowing your apps down by pushing them to the pagefile.

 

What about ram defragmenting utilities ?

 

Reformat? Extreme but if you have gone through each program in turn...

We have a desktop with 256mb ram and it has always been fine

 

Same game, different name. Stay away from them. Your operating system knows how to better optimize your memory than you, or any tool you use.

 

Was your OS slowing down though?

 

Jalf:

sorry...I have been using RamBooster for many many years (as in 5+ with a couple years off as it did not matter because XP was stable) and it does it's job. Point blank and direct you are wrong.

Windows and their runtimes, compilers and libraries are notoriously BAD AT GARBAGE COLLECTION. That is the entire purpose of such utilities as RamBooster. One would have thought you understood this issue as you imply authoritative knowledge.

This utility has been around for around a decade and works as advertised w/o any ill effects on a system. It will IMPROVE your systems garbage collection (which by definition is what a memory leak is, in the case you do not know this...)

Also I have no idea where you feel this specific utility slows ANYTHING down. There might be a brief slow down when RAM is recovered and maybe when a library is restarted if something needed had been left open. Well, these libraries should never have remained open to begin with...again that is part of proper garbage collection. If anything Windows is BAD at managing memory because it does not force garbage collection as aggressively as it should. I suspect the reason is because there are too many out there looking to differentiate systems by measuring nanosecond differences in how fast a program opens.

Most people do not have the time to track down such system faults because, well, real life is far more important, and such a utility will serve them well in the mean time until a solution can be found. Additionally most of these issues simply cannot be fixed by the end user and require a workaround. RamBooster is just such a work around.

I have to add that WOW, I just cannot understand the aggressive attitude posts I have been reading the past few visits. Must be time to consider other options...

Anyway, just never mind the whole thread...life is just too short.

 

grumpy3b: Seriously, listen to Jalf. RamBooster is the same kind of voodoo as those cell phone antenna boosting stickers. They claim benefits, but you're really not doing anything good, and possibly slowing it down more than it would otherwise run. If you have a memory leak, RamBooster won't do anything about it. If you're using the memory currently or have done so recently, it's loaded into RAM. If not, it's usually pushed off into the pagefile when something more urgent comes to the front. Your "notoriously bad" comment may be correct, but RamBooster isn't the solution.

And personal insults aren't called for. I may not agree with Jalf on everything, but he's one of the people around here who definitely knows what he's talking about.

 

How do you measure this? All you said in your original psot was that it makes apps take longer to "come back to life". And I'm telling you why that is the *only* effect it has.

Having used an application for years is not proof that it works. I used to use these apps too, back in the Win95 days. That doesn't mean they work, or do the job they claim to.

So, do you have any actual evidence that they work? Have you got an explanation for *how* they work? How can they release memory that is in use by an application? Because that is what they need to do in order to have an effect.

RAMBOOSTER DOES NOT DO GARBAGE COLLECTION. NEITHER DOES WINDOWS.
C++ programs manage their OWN GARBAGE COLLECTION.
.NET programs USE THE .NET FRAMEWORK's GARBAGE COLLECTION.

Now, lets go back to not shouting, shall we?

One would have thought I understood the issue, yes, given that in my original post I actually *explained what the utility did, and what it didn't do, and why and how*
It's hard to do that if you don't understand the issue.

You mentioned an ill effect yourself. And you know what? Viruses have been around for decades too.

The system does not, and *can* not do garbage collection. Neither can third-party applications.

And if you want to be pedantic, then no, a memory leak is memory allocated by an application and which it (the application) can no longer reference. The system is not involved. Memory leaks happen on per-application basis. The system doesn't get involved until the application terminates. And when that happens, any OS (with the possible exception of Win95) will clean up everything.

They are left open because they are *IN USE*.
And you know what? RamBooster does not close them. It does *not* free any memory, it doesn't release resources, it does not close applications or libraries. The *only* thing it does is moving data that is still in use to the pagefile.

And yes, that is where I get the idea that it slows things down. In the long term, you're right, it has no effect. Because the first thing that happens after it's run is that all the applications pull their data back from the pagefile into memory. Then we're back where we started, and everything works as fast as it did before. The only effect you have is a short-term slowdown.

It does not do garbage collection because it *can* not do garbage collection.
Neither can Linux or OS X because it is impossible.
(Except when an application closes. all three OS'es will properly clean up everything the app used then. DOS didn't do that, and I believe Windows 95 was a bit shaky at it too)

In the same way that a glass of water which someone waved a hand over is a cure for illness.

Aggressive? I'm not the one resorting to personal attacks. And I wasn't the one who started using caps. I'm just trying to tell you why it is simply *impossible* for RamBooster to have *any* (positive) effect.

 

Jalf, while your first post in this thread was succinct and to the point, grumpy is right about your second post being aggressive and bordering on hostile. There's no need to get pushy, after all you're just giving advice to the OP which he can choose to disregard if he wishes.

Grumpy, Jalf's advice is dead on. These mem-freeing utilities don't really do anything besides taking memory used by your useful applications and pushing them to the pagefile. The only way they can do this is allocating an insanely large amount of memory for itself, then immediately freeing that memory lest Window's bogs down. I don't know if you've observed any positive effects after a mem-freeing cycle, but it's unlikely to make your Windows faster.

 

I think that's probably because Jalf is like I am. False information (even if it's in good faith) shouldn't get the same consideration as the truth. When someone disagrees with the truth, it serves to do nothing but spread disinformation. The poster is free to take the advice or not, but would you also jump to justify someone selling snake-oil if they actually believed it did something? I hope not.

 

That's true, I agree. However, would you attempt to justify someone ridiculing the snake-oil buyer?

 

First, correct me if I'm wrong, but he was commenting on my *first* post. (My second one was in response to his post, which to me came across as agressive and bordering on hostile. (And pushy)

Once that is said, however, apologies for any personal offense.
As Pitabred said, I don't have a lot of patience with misinformation.
And if someone gets upset when you tell them the software they're using doesn't work? Well, that seems a bit silly to me. If you then attack the person who told you it doesn't work? Again, I don't have a lot of patience for that.

RamBooster has no (positive) effect. Getting angry about it or accusing me of lying or not understanding the issue is not going to change that. It'll only piss me off, and then I'll post stuff I shouldn't...

However, I wasn't "ridiculing the snake-oil buyer". As I said, I used to run the same kind of apps, and there is even a possibility that they actually had *some* positive effect under Win95. And these apps are good at sounding useful, so I don't blame people for getting fooled into thinking they work.
If it looked like I was ridiculing anyone for that, I apologize, that wasn't intended.

I was, however, ridiculing RamBooster and its peers. And I might have been a bit harsh with people who get aggressive when shown that the app is little more than an illusion. But if you want to argue its usefulness, bring some facts, not just "you lie, and you don't understand the issue".