Originally Posted by Estel

As for SIGSTOP, it definitely *do* the job. The question is, if it provides more power-saving, than monitoring of dbus signals use Remember, that this way *every* dbus signal is monitored, and script only "act" on some of them - but, power/resources is/are used already, even to monitor "other" signals.

That may be cause of why, every measurement show a little more power usage, when power-saving scripts monitor dbus. As opposed to "perceived battery life"

Originally Posted by ammyt

To anyone wishing to use a different active profile:
1. Save the profile you want to use by:
kernel-config save xxx
2. leafpad /usr/local/bin/overclock
3. Replace the top most two lines starting by kernel-config with:
kernel-config load xxx

Originally Posted by ammyt

Ready to install:
Place CPUTweak.sh.txt and CPUTweak.tar in ./MyDocs

Code:

root
apt-get install dbus-scripts
sh /home/user/MyDocs/CPUTweak.sh.txt

NOTE: The frequency set here is starving at 250 500 while idle, and 500 1150 ideal when on usage.

BE CAREFUL: IF YOU THINK/KNOW THAT YOUR N900 CAN'T SURVIVE THESE SETTINGS, BE SURE TO EDIT THE FREQUENCIES IN /usr/local/bin/overclock and underclock FILES, OR WELCOME A REBOOT LOOP

This is really a brilliant idea GigaByte, wonder why I've never thought aboout it, using dbus-scripts for a noble act like this is really intelligent, bravo


-UPDATE-
I added the effective vfs cache pressure trick into this patch, gives far even better battery life.

Updated archive attached.

-UPDATE2-
Usage of process SIGSTOP to pause greedy processes from running in the background.

Updated archive attached.

Originally Posted by colin.stephane

Ok, I see you can overlap swappolube 'vfs_cache_pressure' value if it exist in saved swappolube configuration as default, and differ from the static value you put in 'overclock' script.

What about this modification ? :

Code:

#!/bin/sh

# Default value in case we don't have a swappolube config saved with 'vfs_cache_pressure' option value.
VFS_CACHE_PRESSURE=100

# Loading saved swappolube 'vfs_cache_pressure' value if it exist in /etc/event.d/tuning.
if $(grep vfs_cache_pressure /etc/event.d/tuning 1>/dev/null 2>&1) ; then
        VFS_CACHE_PRESSURE=$(grep vfs_cache_pressure /etc/event.d/tuning | awk '{print $2}' | sed 's|"||g')
fi

kernel-config load ideal
kernel-config limits 500 1150
echo ${VFS_CACHE_PRESSURE} > /proc/sys/vm/vfs_cache_pressure
for process in browserd scim-panel-gtk image-viewer
do
  /usr/bin/pkill -CONT $process
done

Originally Posted by ammyt

I can tell that your doing an excellent job in my view. I don't know where did I read it before, but I've tested it, using the limits of 500 1150 with the conservative governor is the best usage in terms of battery vs. performance, so way to go.

Originally Posted by trisha02

what if smart reflex is enabled. then you cant give 1150 coz this will enter the reboot loop. does 500 1150 means that the phone will be at 500when its in rest and 1150mhz when its active? but we dont need 1150mhz always when phone is active like for chatting and all.

Originally Posted by ammyt

Well in this case, it depends on the amount of times you unlock your device...

@Estel
Sorry for popping in but,
karam has provided a battery patch which reduces vfs_cache_pressure to 10 while the device is idle, and back to 100 when it is active. Many users stated that this gave tremendous battery life including me, and this patch is integrated into ammyt's patch along with other goodies which will further increase battery life so...