Bug 188347

Summary:	Optimize Kopete's usage of timers
Product:	[Unmaintained] kopete	Reporter:	Henrik Pauli <henrik.pauli>
Component:	general	Assignee:	Kopete Developers <kopete-bugs-null>
Status:	RESOLVED UNMAINTAINED
Severity:	wishlist	CC:	kronos.it, shweduke
Priority:	NOR
Version First Reported In:	unspecified
Target Milestone:	---
Platform:	Compiled Sources
OS:	Linux
Latest Commit:		Version Fixed In:
Sentry Crash Report:
Attachments:	gprof output with kopete generating 100 wakeup/s

Description Henrik Pauli 2009-03-28 17:51:12 UTC

Version:           kde-4.2.67 (using Devel)
OS:                Linux
Installed from:    Compiled sources

Using intel’s PowerTOP, I see that there’s a lot of wakeup events from Kopete, much more than from any other software I have running.  Here’s a paste:

Top causes for wakeups:
  35.5% (158.1)       <interrupt> : extra timer interrupt
  22.6% (100.7)            kopete : schedule_timeout (process_timeout)
   7.3% ( 32.6)       <interrupt> : ehci_hcd:usb2, uhci_hcd:usb3, uhci_hcd:usb5
   7.3% ( 32.6)   USB device  5-1 : USB-PS/2 Optical Mouse (Logitech)
   2.9% ( 12.9)       <interrupt> : ahci, eth1
   2.8% ( 12.3)   <kernel module> : usb_hcd_poll_rh_status (rh_timer_func)
   2.6% ( 11.6)          postgres : schedule_timeout (process_timeout)
   2.3% ( 10.3)             skype : schedule_timeout (process_timeout)
   2.3% ( 10.1)      <kernel IPI> : Rescheduling interrupts
   2.2% ( 10.0)     <kernel core> : ehci_work (ehci_watchdog)
   2.2% (  9.6)             skype : do_nanosleep (hrtimer_wakeup)
   0.9% (  4.2)   USB device  6-1 : MOTORIZR Z8 (Motorola)
   0.9% (  4.1)       <interrupt> : uhci_hcd:usb4, uhci_hcd:usb6, HDA Intel
   0.9% (  3.8)       <interrupt> : PS/2 keyboard/mouse/touchpad
   0.8% (  3.5)       <interrupt> : ata_piix
   0.7% (  2.9)    plasma-desktop : schedule_timeout (process_timeout)
   0.4% (  2.0)            mysqld : schedule_timeout (process_timeout)
   0.4% (  2.0)     <kernel core> : clocksource_register (clocksource_watchdog)
   0.4% (  1.7)                 X : schedule_timeout (process_timeout)
   0.3% (  1.4)           konsole : schedule_timeout (process_timeout)
   0.3% (  1.3)     <kernel core> : wl_add_timer (wl_timer)
   0.2% (  1.1)       <interrupt> : ohci1394, eth0, i915@pci:0000:00:02.0


I use Jabber, AIM, ICQ and Yahoo in Kopete, if that helps narrowing it down.

I notice also, probably related, that the fan in my laptop barely ever stops spinning if I have Kopete running, and it’s much likely to happen when it is not running.

(Yes, I realise 4.2.67 is alpha quality code, just I guess it might be a nice idea if you kept an eye out for this issue in the future.)

Comment 1 Luca Tettamanti 2009-05-26 14:38:03 UTC

I see that same behaviour with 4.2.2 (kopete 0.70.2) on Debian/unstable. In my cast kopete starts generating the wakeups as soon as I sign in into Jabber (GTalk).

Comment 2 Luca Tettamanti 2009-05-26 15:45:49 UTC

Created attachment 34020 [details]
gprof output with kopete generating 100 wakeup/s

Comment 3 Roman Jarosz 2009-05-26 20:29:15 UTC

I use Jabber and ICQ and I don't see this in Kopete, it sits at around 2 wakeup/s. The gprof shows that a lot of calls are in contact list code which was totally redone for KDE 4.3 so could you try if it still happen in KDE 4.3 beta1 or later, or which protocol exactly does that.

Comment 4 Luca Tettamanti 2009-06-04 13:55:35 UTC

Sorry for the delay: Kopete compiled from SVN behaves much better, I cannot reproduce the high number of wakeups observed in 0.70.2

Comment 5 Andrew 2012-01-04 03:36:20 UTC

I can observe a lot of wakeups in Kubuntu 11.10 64 bit, KDE 4.7.4.

Comment 6 Christoph Cullmann 2024-09-18 18:27:34 UTC

Dear user, unfortunately Kopete is no longer maintained.

Please migrate to another solution, e.g. for Jabber a possibility is Kaidan, for Matrix a candidate is NeoChat.