Power consumption script (Was: [ltp] T41p suspend to RAM on 2.6.11.3)

[Daniel Stodden]

On Thu, 2005-03-17 at 15:55 +0100, Martin Samuelsson wrote:
> On Thu, Mar 17, 2005 at 10:13:25AM +0100, Daniel Stodden wrote:
> > i've just written a small script to measure sleep power consumption on
> > each suspend to ram. results below, though i did not take the full 15
> > minutes.
> 
> Even if it would be simple enough to myself grep and awk from
> /proc/acpi/battery/BAT0/state I'm lazy enough to ask, could you please
> share the script?
> 
> > Mar 17 09:59:42 localhost acpi: enter: 1111049982 s, 40340 mWh
> > Mar 17 10:07:38 localhost acpi: leave: 1111050458 s, 40270 mWh
> > Mar 17 10:07:38 localhost acpi: power consumption: 70 mWh / 476 s = 529
> > mWh/h

sure.

dns@thinkpad:~$ cat /etc/acpi/power.sh

source /etc/acpi/common.sh

function unixtime {
        date +%s
}

function capacity {
        sed -ne's/remaining capacity: *\([0-9]*\) mWh/\1/p'
< /proc/acpi/battery/BAT0/state
}

function power_enter {
        time1=$(unixtime)
        power1=$(capacity)

        echo "enter: $time1 s, $power1 mWh" | log
}

function power_leave {
        time2=$(unixtime)
        power2=$(capacity)
        echo "leave: $time2 s, $power2 mWh" | log

        dtime=$[time2-time1]
        [ $dtime -ne 0 ] || exit

        dpower=$[power1-power2]
        consume=$[dpower*3600/dtime]
        echo "power consumption: $dpower mWh / $dtime s = $consume mW" |
log
}


dns@thinkpad:~$ cat /etc/acpi/common.sh
#
# $Id: common.sh,v 1.2 2005/03/16 14:05:50 dns Exp $
#

HIBERNATE=/usr/sbin/hibernate

function suspend_enter {
        /bin/run-parts --arg=start /etc/acpi/suspend.d
}

function suspend_leave {
        /bin/run-parts --arg=stop --reverse /etc/acpi/suspend.d
}

function log {
        /usr/bin/logger -tacpi
}


dns@thinkpad:~$ cat /etc/acpi/sleep.sh
#!/bin/sh
#
# $Id: sleep.sh,v 1.4 2005/01/20 08:20:55 dns Exp $
#

source /etc/acpi/common.sh
source /etc/acpi/power.sh

suspend_enter
power_enter
$HIBERNATE -F /etc/hibernate/sleep.conf
power_leave
suspend_leave


> Wouldn't that simply be 529 mW ? ;)

yessssir. that was corrected a minute after i pushed the send button and
reread what i did there :P

all i can say to defend myself is that i fixed an similarly embarrassing
one this morning on thinkwiki. and that one wasn't from me:

http://www.thinkwiki.org/ACPI_sleep_power_drain_test_script?diff=0&oldid=1841

so i'm in good company around here, right? :D


regards,
daniel