Version: (using Devel) Installed from: Compiled sources Compiler: GCC-4.2.4 OS: Linux When I started KDE4 a few hours, it used only ~220Mo of memory. Now it's eating almost 500Mo. Xorg alone is using almost 200Mo. It doesn't seems to be a Xorg problem since it's only happening using KDE4.
Please check bug http://bugs.kde.org/show_bug.cgi?id=162138, maybe you experience the same problem.
I'm not sure, Plasma memory usage looks normal to me (around 25Mo). Only Xorg memory usage grows. But what I read on this bug report makes me think that it could come from the compositing. I'll try that and let you know.
Check who is using XOrg memory with xrestop. If it is the plasma process (compare the PID with the plasma PID), then it is plasma and probably the same bug.
> When I started KDE4 a few hours, it used only ~220Mo of memory. Now it's eating almost 500Mo. Who is eating memory? There isn't a single "kde4" executable, but various.
Any news about this "problem"?
I didn't update/test KDE4 for a little while. I'll let you know as soon as I can test.
I'm sending some screenshots of ksysguard...
Created attachment 28629 [details] Memory usage at boot Here is the memory usage when launching KDE4. Nothing unusual. It's a little high, but I think that the "debugfull" is partly responsible of it.
Created attachment 28630 [details] Memory usage after a few hours Here is the memory usage after a few hours of use. I closed and relauched Superkaramba before taking that screenshot, but it's memory usage started à 17Mo and ended at 60Mo during the same amout of time.
Created attachment 28631 [details] Memory usage after killing and relaunching Plasma As you can see, the memory usage becomes correct again, after killing Plasma and relaunching it.
Updated today, still happening. The wierd thing is: The memory usage seems to grow only when using Konqueror...
which x.org driver are you using?
Nvidia driver version 177.82.
is your card supported by the new 180.16 release? if so, can you try it?
I didn't see there was a new version, I'll try this ASAP.
Tested with the new Nvidia driver version. I was waiting for a stable release. The problem is still there, but I'm wondering: Can the "debugfull" compilation option be the cause of that problem? I mean, the debugging informations has to go somewhere right?
Any news on this ? Have you tried with the last nVidia drivers? The debugging information is kept in the hard disk , and it's only loaded when needed to show a backtrace or debug some application. But it shouldn't enlarge the memory use. Also, what KDE version are you using right now? Are you still using trunk? Thanks
No I'm not compiling nor using KDE 4 at all right now. Too many bugs that devs don't seems to care about. And when we report them, we have no answer, or they make us feel like we're bothering them. So I stopped anything related to KDE 4 until someone finally cares about stability and security. I can't help but think that it's a real regression to have the whole desktop crashing when a stupid little plasmoid has a stupid little bug... Sorry for the off topic, but now you know that all my opened bugs can be closed ;).
(In reply to comment #18) > No I'm not compiling nor using KDE 4 at all right now. Too many bugs that devs > don't seems to care about. And when we report them, we have no answer, or they > make us feel like we're bothering them. So I stopped anything related to KDE 4 > until someone finally cares about stability and security. > I'm sorry about this, we lack of manpower for some things. I'm not a full "developer" myself so I can't work on all the things I would like to. :( > I can't help but think that it's a real regression to have the whole desktop > crashing when a stupid little plasmoid has a stupid little bug... > > Sorry for the off topic, but now you know that all my opened bugs can be closed > ;). No.. I'm going to close this report only because we lost the information source (your KDE4 desktop), then I'm going to evaluate every report individually to try to determine if there are reproducible or not. Sorry again.
I know you're doing a great job trying to follow bug reports and all that stuff. Be sure that my previous message wasn't for you, neither for people like Peter Penz who are doing their best to answer to bug reports and fix them.