I spoke too early. After 24h uptime, memory usage is almost doubled. Insted of normal 2.5GB usage, I have 4.52GB. :(
Printable View
I spoke too early. After 24h uptime, memory usage is almost doubled. Insted of normal 2.5GB usage, I have 4.52GB. :(
605MiBs out of 1.5 GiBs here on this dual core.. this unit has been up on 3.7xxRC4 for a solid 17 hours so far. Rebooted only thrice.
dale@dale-desktop:~$ uname -a
Linux dale-desktop 3.7.0-030700rc4-generic #201211041435 SMP Sun Nov 4 19:43:27 UTC 2012 i686 i686 i686 GNU/Linux
dale@dale-desktop:~$
Running 35 firefox tabs, system monitor, gnome-terminal, Pcmanfile manager.
Can You decode what is the memory „used“ for... I can not confirm such sort of growth... To be more precise: I can account every kind of growth to a service or application I've used... In other words I can not suspect any (major) memory leakage...
I will try to use other kernel for the same amount of time and to see if I would notice any considerable differences...
See that -/+ buffers/cache used? That should not be so much. And I can't account everything to applications and services. Not even half of that is used by apps and services. Even when I logout/login there is almost 4GB of usage.Code:free -m
total used free shared buffers cached
Mem: 7989 7563 426 0 146 2772
-/+ buffers/cache: 4643 3345
Swap: 1953 688 1264
I know what buffers and cache are, but it doesn't behave like it should. If I start playing some game, I'm starting to get a lot of swapping and slowdowns because of it.
P.S. I have very unique kernel configuration. :) I have PREEMPT+1000Hz+HugeTLB+RCU ...
rc4
oCode:em@oem-desktop:~$ free -m
total used free shared buffers cached
Mem: 3905 3770 135 0 4 950
-/+ buffers/cache: 2815 1090
Swap: 10548 19 10529
I'll keep my eye(s) on that... Will report if I see anything to report...
(I have a sledgehammer tool for cleaning memory...but I'm afraid that will not affect parameter You're watching...)Code:sudo sync ; sudo sysctl -w vm.drop_caches=3 ; sudo sysctl -w vm.drop_caches=1
Update: @MilosSD: It seems that You have a valid point... Confirmed...
I have tried over 4 different version of 3.7rc[1-4], none of which works normally. The system runs but without stability. At first it looks good, but after a while you may feel some different compared to 3.5 or 3.6. And also there are more fails in /var/log/kern.log. In my case the kernel specially has a thing on my ata hard drive, which results in low I/O performance. My opinion is kernel 3.7rc is far from being usable. I'm running Ubuntu 12.10 not 13.04, but I don't think it's relevant.
Today's 996 looks promising... (drm-next) ...
Can everyone that have this bug with memory leaks, go here and confirm it:
https://bugzilla.kernel.org/show_bug.cgi?id=50181
Also, describe it better then me, please. My English s**ks. :)
Thanks. :)
Some good news for those of us who can only use the legacy nvida driver now (i.e. latest nvidia-current does not support our graphics cards).
With this version, the nvidia kernel module builds ok with kernel 3.7-rcN. :)Code:paul@raring-64:~$ apt-cache policy nvidia-current-updates
nvidia-current-updates:
Installed: 304.64-0ubuntu1
Candidate: 304.64-0ubuntu1
Version table:
*** 304.64-0ubuntu1 0
500 http://archive.ubuntu.com/ubuntu/ raring-proposed/restricted amd64 Packages
100 /var/lib/dpkg/status
304.51-0ubuntu1 0
500 http://archive.ubuntu.com/ubuntu/ raring/restricted amd64 Packages
paul@raring-64:~$