Active Topics

 


Reply
Thread Tools
debernardis's Avatar
Posts: 2,142 | Thanked: 2,054 times | Joined on Dec 2006 @ Sicily
#61
The tuning lines described by kaze.daniel and allnameswereout seem to work wonders on my n900. Music doesn't stutter at all like when I edited only swappiness and page-cluster. The GUI is superfluid like liquid helium. Life's good Thanks guys!
__________________
Ernesto de Bernardis

 
Posts: 25 | Thanked: 5 times | Joined on Jul 2010
#62
Originally Posted by slender View Post
I have also tested swappiness 30 and page-cluster 0. I'm pretty sure that responsivness has been better. I would be highly interested on possible side-effects on these changes.
are you still using these settings? any bad side-effects? thanks!
 
Posts: 2,829 | Thanked: 1,459 times | Joined on Dec 2009 @ Finland
#63
nope, but unfortunately i have had couple of shutdowns. These were done by purpose. So no long uptime (3 days or more) tests.

I have also applied other tweaks (made tuning file to /etc/event.d/ dir) and still no ill side effects. IMO the biggest difference came with just swappiness 30. I should try to make heavy banging with multimedia/browser + sleep + test how it wakes up when call comes.

still i would like to get more information about purpose of swapon and swapoff commands. Why, when, how?
 

The Following User Says Thank You to slender For This Useful Post:
debernardis's Avatar
Posts: 2,142 | Thanked: 2,054 times | Joined on Dec 2006 @ Sicily
#64
I liked kaze.daniel and allnameswereout's hack so much that made a debian package for everybody to test it without the need for messing with root terminal.

See here.
__________________
Ernesto de Bernardis

 

The Following 5 Users Say Thank You to debernardis For This Useful Post:
allnameswereout's Avatar
Posts: 3,397 | Thanked: 1,212 times | Joined on Jul 2008 @ Netherlands
#65
Neat but please note nobody has tested any of the settings. Speaking for myself, I am able to experience an improvement (more snappier) with swappiness = 30. The settings have to be carefully tested and benchmarked in order to declare them both usable and safe. One could argue "if it doesn't hurt, it doesn't matter" but I don't recommend to the general public changing too many settings you don't know effect about. That said, if anyone can explain how to benchmark the effects I'd love to share the benchmarks from my machines.
__________________
Goosfraba! All text written by allnameswereout is public domain unless stated otherwise. Thank you for sharing your output!
 

The Following User Says Thank You to allnameswereout For This Useful Post:
zlatko's Avatar
Posts: 861 | Thanked: 936 times | Joined on Feb 2010 @ Bulgaria
#66
I am running swappiness at 30 for about a week. Haven't noticed any improvements neither in speed or battery life. Haven't noticed any deterioration either though.
So the effects can not be felt in everyday use(for my phone and usage pattern).

PS. @allnameswereout - I am using stock kernel without any modifications.

Last edited by zlatko; 2010-09-15 at 14:36. Reason: Adding info
 

The Following User Says Thank You to zlatko For This Useful Post:
Posts: 184 | Thanked: 112 times | Joined on May 2006
#67
Originally Posted by zlatko View Post
I am running swappiness at 30 for about a week. Haven't noticed any improvements neither in speed or battery life. Haven't noticed any deterioration either though.
So the effects can not be felt in everyday use(for my phone and usage pattern).
Same here....
 

The Following User Says Thank You to jnack95 For This Useful Post:
allnameswereout's Avatar
Posts: 3,397 | Thanked: 1,212 times | Joined on Jul 2008 @ Netherlands
#68
Here is a compilation of observations made by N900 users shared in this thread when using swappiness = 30.

Battery-life improved & responsiveness improved & no regressions
pycage
hackfanatic

Responsiveness improved
slender note: and page-cluster 0
JaseP
iareraccoon note: using knownokia.ca's settings compared to merely swapiness = 15

No improvements & no regressions
memoryguy
zlatko
jnack95

Note that nobody has specified which kernel they're using. I'm not sure whether this is relevant.

You're encouraged to verify my compilation, improve details on your (future) reports, and use above compilation as basis in next analysis of this tweak whether in this thread or on wiki.
__________________
Goosfraba! All text written by allnameswereout is public domain unless stated otherwise. Thank you for sharing your output!
 
debernardis's Avatar
Posts: 2,142 | Thanked: 2,054 times | Joined on Dec 2006 @ Sicily
#69
I found no particular effect from swappiness changing alone, but the complete set of new values you quoted from kaze.daniel worked wonders on my nit. I haven't had a jerkiness in the user interface since I changed them.
__________________
Ernesto de Bernardis

 
allnameswereout's Avatar
Posts: 3,397 | Thanked: 1,212 times | Joined on Jul 2008 @ Netherlands
#70
Originally Posted by debernardis View Post
I found no particular effect from swappiness changing alone, but the complete set of new values you quoted from kaze.daniel worked wonders on my nit. I haven't had a jerkiness in the user interface since I changed them.
Have you noticed a decrease of screen tearing?
__________________
Goosfraba! All text written by allnameswereout is public domain unless stated otherwise. Thank you for sharing your output!
 
Reply

Tags
battery


 
Forum Jump


All times are GMT. The time now is 18:43.