Comment 36 for bug 177895

With 2.4.24.8 the situation improves again:

  36,5% ( 69,0) <interrupt> : i915@pci:0000:00:02.0
  16,0% ( 30,2) <kernel IPI> : Rescheduling interrupts
  15,9% ( 30,0) Xorg : schedule_timeout (process_timeout)
   5,8% ( 11,0) /usr/bin/condui : schedule_timeout (process_timeout)
   4,3% ( 8,2) <kernel module> : usb_hcd_poll_rh_status (rh_timer_func)
   3,7% ( 7,0) compiz.real : schedule_timeout (process_timeout)
   3,3% ( 6,3) <interrupt> : extra timer interrupt
   2,6% ( 4,9) conduit.real : schedule_timeout (process_timeout)
   1,5% ( 2,8) trackerd : schedule_timeout (process_timeout)
   0,9% ( 1,7) gnome-terminal : schedule_timeout (process_timeout)
   0,8% ( 1,6) xfsbufd : schedule_timeout (process_timeout)
   0,6% ( 1,1) /usr/share/reve : schedule_timeout (process_timeout)
   0,5% ( 1,0) <interrupt> : libata, b43
   0,5% ( 1,0) nm-applet : schedule_timeout (process_timeout)
   0,5% ( 1,0) netspeed_applet : schedule_timeout (process_timeout)
   0,5% ( 1,0) NetworkManager : tg3_open (tg3_timer)
   0,5% ( 1,0) dhcdbd : schedule_timeout (process_timeout)

The total count is still crazy, but looking into the list looks much better, ~160 wakeups per second while idle (with GNOME open with system monitor and other applets at the gnome-panel).