Active Topics

 


Reply
Thread Tools
Posts: 1,751 | Thanked: 844 times | Joined on Feb 2010 @ Sweden
#11
Originally Posted by RobbieThe1st View Post
Good idea, but I'm not sure the voltage control is fine enough. While using Mplayer to test mine, and reduce the voltage by one, I'd find that some of the time I'd see errors in mplayer, sometimes it would crash outright - and it would work fine at 1 voltage point higher.
Have you tried looking in "dmesg" after taken a picture or done a video? It could be working but also generate looping errors that depleats your battery faster. I think that however this will become it will need or some way of listening on error output. I am not sure "dmesg" would be enough. If we could find out the precise error that occur just before reboot.. then we will be able to make a more robust script/program.

A nother problem with the "reboot" strategy would be what to do if the battery gets fully depleated. Reboots takes about 10% battery.
__________________
You like what i do? Donate!

Make your desktop look awesome - use the AwOken Theme with the AwOken Icon Theme.

Add me on twitter @almehdin
Visit the swedish maemo/meego community forums
 
Posts: 91 | Thanked: 22 times | Joined on Apr 2010 @ Germany
#12
You could charge it while testing. But this would generate more heat. On the other hand this may not be bad since you dont want you device to fail in Summer just becuase you finetuned it in Winter.
 
Posts: 1,751 | Thanked: 844 times | Joined on Feb 2010 @ Sweden
#13
Originally Posted by ChoMar View Post
You could charge it while testing. But this would generate more heat. On the other hand this may not be bad since you dont want you device to fail in Summer just becuase you finetuned it in Winter.
An even better solution would be that it abort if the charge get's to low.. or at least warn.
__________________
You like what i do? Donate!

Make your desktop look awesome - use the AwOken Theme with the AwOken Icon Theme.

Add me on twitter @almehdin
Visit the swedish maemo/meego community forums
 
Posts: 284 | Thanked: 74 times | Joined on Mar 2010 @ Wigan, UK
#14
Stick my name on the testing list.

Ive got insurance on it anyway
 
Posts: 166 | Thanked: 30 times | Joined on Apr 2010 @ Halifax, UK
#15
all this fuss over, over clocking. it'd be a waste of time for a dev to do this especially seen the profile's make almost no difference to battery life
its not the cpu thats the problem really.
if you want to have a low voltage load because you think it'll give you more battery life, try dimming the screen and disconnecting from gprs/wifi and turning bluetooth off instead
 
atilla's Avatar
Posts: 1,210 | Thanked: 597 times | Joined on Apr 2010 @ hamburg,germany
#16
would be really nice if there was a automatic voltage test app.
i had probs with overclocking too.my n900 runs at 850 and the lv profile stable like a rock but if i tried 900 everything was ****ed up.
i tried lv and xlv and both didn't work.
would it be possible to use the default nokia voltages at 900-950 mhz?
i could spend a lil bit battery if my n900 could run stable at 900-950 mhz
__________________


Nobody likes us but we dont care....
 
Posts: 842 | Thanked: 1,197 times | Joined on May 2010
#17
Originally Posted by AlMehdi View Post
Have you tried looking in "dmesg" after taken a picture or done a video? It could be working but also generate looping errors that depleats your battery faster. I think that however this will become it will need or some way of listening on error output. I am not sure "dmesg" would be enough. If we could find out the precise error that occur just before reboot.. then we will be able to make a more robust script/program.
Well, we can always pipe STDERR to a file in /tmp/, then read that file. The only issue would be determining -real- errors from comments or errors not caused by the OC process(say a slightly corrupted video file etc.)

Originally Posted by AlMehdi View Post
A nother problem with the "reboot" strategy would be what to do if the battery gets fully depleated. Reboots takes about 10% battery.
Actually, this is something of a myth. Rebooting resets the internal battery percentage calculation, which is then based solely on the battery voltage.
This means that -- depending -- a reboot can cause relatively unpredictable results in the battery % value. Sometimes it will go down 10%, sometimes up... But 10 reboots in under an hour won't drain your battery completely, no matter how much a single reboot "drained".
It -will- take some power though, but considering that you should be able to run fully loaded for 3-6hrs(depending on condition of the battery), it shouldn't be -that- much of a problem.

Charging it during the test probably wouldn't hurt -too- much, and charging it before the test would also be a good idea.


Originally Posted by matthew maude View Post
all this fuss over, over clocking. it'd be a waste of time for a dev to do this especially seen the profile's make almost no difference to battery life
its not the cpu thats the problem really.
if you want to have a low voltage load because you think it'll give you more battery life, try dimming the screen and disconnecting from gprs/wifi and turning bluetooth off instead
Really, it has less to do with any theoretical battery life savings than reducing heat and increasing performance. The main reason you'd use something like this is that your N900 may not be stable at one or more "stock" OC frequencies(Or even not stable at the stock settings), when all it needs is just a little voltage boost.
As far as battery life goes, I do agree with your statements - I find significant standby time increasing by going into "offline" mode(as is expected).
 
Posts: 1,751 | Thanked: 844 times | Joined on Feb 2010 @ Sweden
#18
Originally Posted by RobbieThe1st View Post
Well, we can always pipe STDERR to a file in /tmp/, then read that file. The only issue would be determining -real- errors from comments or errors not caused by the OC process(say a slightly corrupted video file etc.)
Yeah.. that would need some empirical data.


Actually, this is something of a myth. Rebooting resets the internal battery percentage calculation, which is then based solely on the battery voltage.
This means that -- depending -- a reboot can cause relatively unpredictable results in the battery % value. Sometimes it will go down 10%, sometimes up... But 10 reboots in under an hour won't drain your battery completely, no matter how much a single reboot "drained".
It -will- take some power though, but considering that you should be able to run fully loaded for 3-6hrs(depending on condition of the battery), it shouldn't be -that- much of a problem.

Charging it during the test probably wouldn't hurt -too- much, and charging it before the test would also be a good idea.
Really, are you sure? I know that the calculation is not allways correct. One time i was testing something that needed many reboots. It pretty much depleated my battery. I can't remember now how many times or during what time.. i think it was within two-three hours. My normal uptime is two-three days.

edit: It could have been a cause of what i was doing..
__________________
You like what i do? Donate!

Make your desktop look awesome - use the AwOken Theme with the AwOken Icon Theme.

Add me on twitter @almehdin
Visit the swedish maemo/meego community forums

Last edited by AlMehdi; 2010-10-18 at 08:14.
 
Posts: 1,751 | Thanked: 844 times | Joined on Feb 2010 @ Sweden
#19
Originally Posted by matthew maude View Post
all this fuss over, over clocking. it'd be a waste of time for a dev to do this especially seen the profile's make almost no difference to battery life
its not the cpu thats the problem really.
if you want to have a low voltage load because you think it'll give you more battery life, try dimming the screen and disconnecting from gprs/wifi and turning bluetooth off instead
Then you are doing something wrong or running at 1ghz.. Try running at 600mhz undervolted and you will see the gain. Of course you will save bettery by dimming the screen and going offline.. but you will save even more undervoltaged.
__________________
You like what i do? Donate!

Make your desktop look awesome - use the AwOken Theme with the AwOken Icon Theme.

Add me on twitter @almehdin
Visit the swedish maemo/meego community forums
 
Posts: 842 | Thanked: 1,197 times | Joined on May 2010
#20
Depending on load, even moderate OCing can -increase- battery life, as it ends up spending a shorter period of time working, and more sleeping(or, at least underclocked). However, running something that maxes the CPU constantly will give you worse runtime... But the same thing happens at 600mhz, so..
 
Reply


 
Forum Jump


All times are GMT. The time now is 23:22.