General discussion

Locked

XP Memory Leaks?

By wareid ·
I know very little about memory leaks themselves, but I know they slow down computers. Since about 3 months ago, I noticed my XP slowing down greatly after only a few hours. I used to have Windows 98, and I knew about memory leaks with that, and I gues the lesser leaks in XP threw me off until I got accustomed. It usually gets bad if I am downloading something (no matter how small) if I'm working in Paint Shop Pro or Macromedia, or if I'm on MSN. Sometimes it even prevents me from working effciently with Notepad! Are there any solutions for this?
Note: I've already virus scanned and gotten rid of any viruses.

This conversation is currently closed to new comments.

22 total posts (Page 2 of 3)   Prev   01 | 02 | 03   Next
Thread display: Collapse - | Expand +

All Comments

Collapse -

RE: Response to some matters raised

by ?/\/\?|???\/??? In reply to Response to some matters ...

>> Sysinternals.com / however I have never heard of it before <<
Indeed...

Yes, AutoRuns does not come with XP while MSCONFIG does. MSCONFIG in a pinch, AutoRuns any other time.

>> Virtual Memory - if you do not clear it out it bloats up to fill the hard drive, a waste of space. <<
Not how I would word it - Windows can manage the page file size, and it won't grow larger than the maximum size you tell it to. This is not "bloat", but how Windows manages memory.

>> have seen a few rare cases where the details of VM info had been reloaded into RAM following a reboot <<
??? I wouldn't be so sure it was "details of VM info". There are a lot of things going on behind the scenes in Windows (prefetcher, etc) that can "load" stuff that had previously been loaded into memory.

>> clearing VM is a good maintenance activity that may or may not affect your RAM operation - but should still be done <<
If, by VM, you mean the page file, note my original post. It's not performance related, but can have security implications, and can also impact shutdown times significantly. Generally speaking, let Windows manage the paging file, and forget about it.

>> I have yet to see a memeory manager that provides a moment to moment system performance enhancement as they do need resources to run <<
:) Much like "virtual memory uses some RAM to note what is in the VM" (in other words, Windows virtual memory manager requires some system resources in order to do what it is designed to do, much like any other program will consume system "resources").

>> I have seen them improve longer term system performance by eliminating most BSODs and the like <<
If 3rd party memory "optimization" utilities (such as this MemTurbo that you talk of) eliminated most BSODs, why doesn't MS buy / build one and shove it into Windows? The fact of the matter is, the memory management / "optimization" routines / algorithms in Windows are sufficiently complex and purposely constructed to be efficient - they have to be. Further, BSODs are caused by faulty drivers or hardware. Period. User-mode programs (running in ring 3) cannot, by design of Windows, cause a BSOD. Thus, the claim that a memory "optimizer" utility can prevent a BSOD seems a bit odd.

>> many of which are caused by insufficient RAM to run a program or part of a program. <<
BSODs caused by insuffucient RAM? See above. If an application requests the allocation of more memory, and Windows informs the application that Windows cannot satisfy the request for an allocation of the requested size (E_OUT_OF_MEMORY), and the application wasn't written to handle that type of condition, the application is likely to crash, possibly with an access violation (0xc0000005) or some other error. If a driver running in kernel mode (ring 0) fails to handle the allocation failure properly, this can cause a BSOD. The problem, though, is a bad driver - the driver didn't properly handle a condition it could feasibly be faced with.

>> Students found that opening and closing Word several times caused the system to crash with a general protection fault due to 'insufficient memory' <<
The _SYSTEM_ crashed with a GPF? What do you mean by system? The operating system? Talk of GPFs takes me back to the days of Windows 9x / ME, and yet you indicate this happened on an NT4 system. If it caused the _SYSTEM_ to crash (BSOD), I would suggest a driver issue, as (again) drivers or faulty hardware cause BSOD crashes, not user-mode programs. If it was MS Word that "crashed" with the "GPF" (in NT, this type of error would likely be called an access violation) I would expect Dr. Watson to have kicked in and created a log and a USER.DMP. One could then analyze the log or dump file to determine what module was causing the crash, and that could potentially indicate how the problem might be fixed (MS Word add-in causing a problem? AV or other program hooking the process? Something that an Office service pack would fix?)...

>> Installation of MemTurbo set to recover when available RAM below 25 MB resulted in MemTurbo running and clearing the RAM and the system did not crash. <<
I don't suspect MemTurbo would fix this type of problem. Certainly, its function could alter (indirectly) the execution path of Word so the "thing" that caused the problem was "avoided"...

>> Every time he ran it the system crashed. <<
Again, what is meant by "system" - the OS? Windows Explorer? The game? What was the "crash" (what was the error code, and in what manner was it presented)? Was it always the same?

>> Investigation suggested a memory problem <<
What type of investigation? How is "memory problem" defined? Insufficent RAM? Bad RAM? I don't think MemTurbo truly fixed the problem...

Certainly, you are able to recount scenarios where MemTurbo seems to have positively affected problems you witnessed. But I suspect that further / more detailed investigation into the problems would have turned up alternative solutions that would not have required MemTurbo or similar programs / "utilities".

I recommend checking out the book "Windows Internals by Mark Russinovich (Sysinternals / Winternals guy) and David Solomon.
http://search.barnesandnoble.com/booksearch/isbnInquiry.asp?z=y&isbn=073561**74&itm=1

Collapse -

Further response to some matters raised

by Deadly Ernest In reply to RE: Response to some matt ...

Win XP has a nice little routine that will inform you when it has insufficient VM and request approval to expand it once it has reached the maximum size set, thus going beyond that. A simple policy setting reduces the need for this and, as you said, provides some additional security. That is why I said clearing the VM is a maintenance task.

The few times I have seen VM info reloaded it was not any of the prefetch stuff - spent hours trying to work out how it happend, never did. But it happened consistently with the computers at that office - it may have been a Dell related thing as they ran a Dell specialised version of XP. The matter went away after I changed the policy setting to clear the VM.

The operating system was Win NT4 and it crashed with a GPF and BSOD. We experimented with monitoring software, each time Word was opened RAm was used and not released when Word was closed, the amount of available RAM reduced with each incident of Word being opened. Eventually there was not enough RAM to open Word again and the system gave a quick 'insufficent memory' error message and then a BSOD stating a GPF. Raised the issue with the college tech people who raised it with MS techs who stated that it was a known issue but fixed by a reboot as it was not worth the trouble to fix. Talk about quality, or lack there off, software.

MemTurbo fixed the problem in that when the system got to the minimum amount of RAM set it took priority control of the system and cleared all RAM that was not actively being used by the system at that time, thus making more space in the RAM and allowing the program room to open. Kind of like straightening up all the papers spread across the desk so you could open a new one.

I think MS have not fixed the problem for the same reason that they have not fixed numerous other 'minor' problems in the Windows kernel over the last 10 years - not enough of an issue to justify the expenditure for a total review and rewrite of the kernel.

Thanks for the tip on the book, will investigate it.

Collapse -

Question?

by rkuhn In reply to Further response to some ...

What is a "Dell specialised version of XP"?

C'mon man, you just love to take shots at Microsoft. I've seen comments from you all over TR to that affect.

And all along you had me thinking Microsoft was closed source and that was evil, but apparently Dell does something? Give me a break.

You have raised valid points before and at times I do envy your experience, but your choice of words here leads me to believe you don't know what the heck you're talking about.

Hopefully, these folks are no longer running Win NT 4. Picking on NT 4 is like me airing my complaints about some old, non-updated version of Fedora, or for that matter, the resource hungary and hogging OpenOffice.

The addition of most any third party memory manager is typically a bad idea in a Windows environment. Most likey, one only gets little or no benefit for a freeware or worse paid piece of software.

I agree with the previoius poster, most memory issues are driver or hardware related with the exception of poorly written software.

No matter how good or bad Windows is, poorly written drivers or software will gum up any OS.

Collapse -

FreeRAM XP Pro

by hulyalkar In reply to Question?

There are couple of softwares which do clean up leaked memory. While using WIN 98 I tried FreeRAM which worked well but after upgrading to XP the XP version did not gave desired results.

Has any one tried this and what's the opinion.

Thanks and regards

Collapse -

As stated earlier...

by ?/\/\?|???\/??? In reply to FreeRAM XP Pro

I recommend reading Mark Russinovich's (of Sysinternals.com and Winternals.com) excellent article "The Memory-Optimization Hoax" (http://www.windowsitpro.com/Articles/Index.cfm?ArticleID=41095&DisplayTab=Article)
Ultimately, Mark states:
"These utilities appear to do useful work, but at best, RAM optimizers have no effect, and at worst, they seriously degrade performance."

There may have been some merit to using this type of utility with Win 3.1/9x, but for NT/2000/XP you're best not even thinking about it.

Collapse -

Dell have a deal with MS for cheaper copies

by Deadly Ernest In reply to Question?

like HP, Compaq, IBM and others do. The version of the operating system loaded onto the computer includes drivers only for the equipment on the computer and no others, this means that you can only use it on that model of that make - I have heard that this is changing but it has been so for some time. Thus a change of equipment usually means no generic drivers etc. It is great fun (for us masochists) to work on these systems, especially if they have hardware problems.

I am a great believer in doing things right so that you do not have to fix the same problem again. When MS first started and produced DOS and then Windows, up to 3.11, they were good operating systems that did their job very well. But since Win 95 they have had many problems which MS have NOT taken the time to fix the underlying problem, but provided lightweight patches. I also object to the fact that the only things that they can seem to carry across new versions are the security holes and faults. This is bad workmanship bad causes extra costs to clients. It is most annoying that they always tout their new version as secure, especially when they include the known security holes from earlier versions. They should be hounded about this.

The problem with the memory management in Win 95, Win 98 and NT is very noticeable and is what prompted people to create memory manager programs. In response to this, and complaints, MS have improved the memory management in Win2K and Win XP (the best by MS so far), but it still has some faults, they have not yet addressed the problem in Word and Excel that does NOT release the memory when closing - they may soon. The problem is solely due to poor software in that Windows does not automatically clear out the assigned memory when a program closes and Word etc do not clear out the assigned memory as they close. Many applications do clear the memory as they are better written.

With regards the use of MemTurbo in Win XP I have found it useful with installations that have less than 256 MB RAM, and a waste of time if you have more than 1GB RAM. It may or may not be useful with system from 256 MB to 1GB of RAM, depending upon what you are doing.

The memory managers have never made the moment to moment use of the system better as they use resources, but they have reduced the number of BSODs and the like caused when the system crashes through insufficient available RAM, and that is a help as it reduces lost work.

As much as I prefer to use Linux I have systems running Win 3.11/32, Win 98SE, Win2K, Win XP Pro and Win XP Pro 64 bit edition as those are what my clients use and I use them to check and test things. And yes there are some things that are done better in Windwos than Linux, as there are somethings that are done better in Linux than Windows.

Collapse -

OEMs / more of same...

by ?/\/\?|???\/??? In reply to Dell have a deal with MS ...

Does anyone have a definitive list that indicates exactly what liberties an OEM is able to take with the Windows versions they ship with hardware? I can't say I've ever worked on an OEM system where the set of supplied drivers had been stripped down - any hardware I used that I expected to have a driver for, Windows had a driver for. Perhaps things are different in Australia.

>> The problem with the memory management in Win 95, Win 98 and NT is very noticeable <<
95 and 98 can be tossed in the same bucket. NT should be in the same bucket as 2000 and XP. And I can't say I've had any problems with the way things work in NT/2000/XP.

>> they have not yet addressed the problem in Word and Excel that does NOT release the memory when closing <<
Word and Excel are applications, not operating systems. When a process exits / terminates, the OS "cleans up" behind the application.
From "Programming Applications for Microsoft Windows, Fourth Edition" by Jeffrey Richter (Chapter 4, Processes):
"Once a process terminates (no matter how), the system guarantees that the process will not leave any parts of itself behind. There is absolutely no way of knowing whether that process had ever run. [Italics]A process will leak absolutely nothing once it has terminated.[/Italics] I hope that is clear."

and

"When a process terminates, the following actions are set in motion:

1) Any remaining threads in the process are terminated.
2) All the User and GDI objects allocated by the process are freed, and all the kernel objects are closed. (These kernel objects are destroyed if no other process has open handles to them. However, the kernel objects are not destroyed if other processes do have open handles to them.)
3) The process's exit code changes from STILL_ACTIVE to the code passed to ExitProcess or TerminateProcess.
4) The process kernel object's status becomes signaled. Other threads in the system can suspend themselves until the process is terminated.
5) The process kernel object's usage count is decremented by 1."

>> The problem is solely due to poor software in that Windows does not automatically clear out the assigned memory when a program closes and Word etc do not clear out the assigned memory as they close. Many applications do clear the memory as they are better written. <<
What do you mean "clear out"? If your expectation is that Word and Excel "write zeroes" to the memory, what is the basis for this expectation? I've never heard of a reason for an application to do that...

Because I look with a suspicious eye on "memory managers / optimizers..."
Previously you stated " I recommend the use of a memory manager like MemTurbo for every MS Windows installation that has less than 512 MB of RAM. Some examples of the differences they make." Now you state "With regards the use of MemTurbo in Win XP I have found it useful with installations that have less than 256 MB RAM".
Which is it?

Further, you state "a waste of time if you have more than 1GB RAM"
Why is that?

>> have reduced the number of BSODs and the like caused when the system crashes through insufficient available RAM <<
Again, insufficient RAM does not cause BSODs. BSODs are caused by faulty drivers / hardware.

Collapse -

try this

by merah678 In reply to XP Memory Leaks?

1)try to clean spyware
2)use ccleaner to clean registery
3)i had similar problem on my network..and that will do the trick

Collapse -

BitTornado Bittorrent Client

by NZ_Justice In reply to XP Memory Leaks?

leeks memory. If you are using this as your download manager it would use all your memory. Also if you close down BitTornado the process sometimes stays running and does not free up memory. I have only experienced this with the stable version of BitTornado.

Patch your PC. Patch all your Apps.

Collapse -

Imagine what an "unstable" version would do...

by ?/\/\?|???\/??? In reply to BitTornado Bittorrent Cli ...

>> Also if you close down BitTornado the process sometimes stays running and does not free up memory <<
Sounds like a poorly written application, not the fault of Windows...

Back to Windows Forum
22 total posts (Page 2 of 3)   Prev   01 | 02 | 03   Next

Related Discussions

Related Forums