powertop: making changes permanent

[Paul Johnson]

I've been messing around with powertop on my Lenovo T61 running Ubuntu
8.10 (Intrepid).  Powertop suggests several changes that do improve
battery life.

However, the changes are not made permanent and I'm not sure how this
ought to be done in Ubuntu.  (I could do it on a RedHat system that
does not use upstart, but I'm a bit stumped about Ubuntu).  The first
time I tried powertop, I learned that the CDMA wireless WAN was always
on, and I turned that off in the bios and it quieted down powertop
quite a bit.  But the rest is, well, a mystery for me.

I copied some of the advice that I'd like to follow permanently:


Suggestion: Disable 'hal' from polling your cdrom with:
hal-disable-polling --device /dev/cdrom 'hal' is the component that au
to-opens a

uggestion: increase the VM dirty writeback time from 5.00 to 15 secon
ds with:
  echo 1500 > /proc/sys/vm/dirty_writeback_centisecs

Suggestion: Enable laptop-mode by executing the following command:
   echo 5 > /proc/sys/vm/laptop_mode


Suggestion: Enable SATA ALPM link power management via:
  echo min_power > /sys/class/scsi_host/host0/link_power_management_policy


Suggestion: Enable SATA ALPM link power management via:
  echo min_power > /sys/class/scsi_host/host0/link_power_management_policy



In case you have a laptop, you might try powertop because it has some
pretty interesting insights about power management.  One insight is
that using Gmail in a web browser is very energy intensive.   The
wakeups per  minute before browsing to Gmail are low and I've got
about 4 hours battery.

Wakeups-from-idle per second : 80.5     interval: 15.0s
Power usage (ACPI estimate): 19.9W (4.0 hours) (long term: 21.7W,/3.7h)

Right after firing up Gmail

Wakeups-from-idle per second : 536.7    interval: 3.0s
Power usage (ACPI estimate): 22.5W (3.5 hours) (long term: 22.3W,/3.6h)
Top causes for wakeups:
  73.8% (426.7)       <interrupt> : PS/2 keyboard/mouse/touchpad
  10.0% ( 57.7)            mlterm : schedule_timeout (process_timeout)
   4.2% ( 24.3)       <interrupt> : extra timer interrupt
   3.1% ( 18.0)      <kernel IPI> : Rescheduling interrupts
   2.8% ( 16.3)       <interrupt> : iwlagn
   1.9% ( 11.0)           firefox : futex_wait (hrtimer_wakeup)

Last year, I noticed an even larger difference, and was told by the
gnome-power-manager team that the Gmail system uses AJAX to monitor
the user interaction with the system and that causes a lot of wakeups.
 My conclusion is: close the browser after you are done with Gmail!

Thanks in advance!

PJ

-- 
Paul E. Johnson
Professor, Political Science
1541 Lilac Lane, Room 504
University of Kansas