Active Topics
-
Porting apps to Leste (32)
to Maemo 7 / Leste by Arno_11 - 4 days, 15 hrs ago -
[ANNOUNCE] CSSU-thumb thread - stable Thumb2 on N900 (2,264)
to Maemo 5 / Fremantle by panjgoori - 6 days, 9 hrs ago - more...
 |
|
2010-11-16
, 14:07
|
|
Posts: 820 |
Thanked: 436 times |
Joined on May 2010
@ Portsmouth, UK.
|
#112
|
Originally Posted by SavageD
Well you didn't acknowledge that the fact I had your old patches prior to re-flashing would not effect the situation now as I have re-flashed.
Oh lol, so by my asking for assistance from others who haven't tried my patch you didn't think I understood you had reflash. I completely understand
By your wording I thought you were implying that there still might be some remnant of your patch, which would be illogical to say the least.
e.g.
Originally Posted by SavageD
Ok that just leaves some one who has never used my patches to test this out.
Plz, anyone who has never used any of my patches, plz try to confirm this.
Originally Posted by SavageD
By what measure? i would say you over reacted if anything.
My reaction simply turned out to be equal to yours.
Originally Posted by SavageD
Yea the section on my graph where it drops very suddenly (at the end of the overlaid section) was where I had just installed kroll and was playing it for the first time. I didn't do this specifically as a test. I just picked the graph from the day before which was quite typical usage for me.
Wait what? The result of you having lost so much mah before the end of the hr period was a result of you playing kroll?
Originally Posted by SavageD
Now you mention it, no this is not a fair test, but it is as close to a fair test as we can realistically get.
So now your gonna say that this isn't truly fair test....
What I was saying is that purely comparing on battery % would be unfair, as it is not a true representation of the charge of the battery.
Last edited by James_Littler; 2010-11-16 at 14:16.
|
|
2010-11-16
, 14:27
|
|
Posts: 436 |
Thanked: 406 times |
Joined on Jan 2010
|
#113
|
Originally Posted by James_Littler
Actually even though the remnants of my patch is gone most of the effects (performance wise) still remains. I have good reason for not mentioning this before, cause as you say it seems illogical.
By your wording I thought you were implying that there still might be some remnant of your patch, which would be illogical to say the least.
I know that most of the effects still remain because when I had started this little project, and kept reflashing my device to default, games that I kept testing (vgba) remained more or less at the same speed as when I started using the patches, and my battery life more or less remained the same as with my patches installed.
Any one can test this right now really, all you have to do is this:
Test out a slow running game and note it's difference in performance.
Apply my patch
Play the same game again and note the difference in performance
Restore default state buy either re-apply the default pulse audio files using my script or reflashing
Note the difference in performance again.
You would soon realize that the device now performs almost the same as it did with my patches installed.
Same happened when I flashed to pr1.3...most of my effects were still there. My guess is that it offloaded some strain on the phone somewhere...but thats only a guess
Last edited by SavageD; 2010-11-16 at 14:32.
 |
|
2010-11-16
, 14:41
|
|
Posts: 889 |
Thanked: 2,087 times |
Joined on Sep 2010
@ Manchester
|
#114
|
Originally Posted by James_Littler
Yes there's a really informative post by woody14619 that explains the battery percentage thing.
Not sure, as I mentioned before, battery % is not accurate and is completely irrelevant. What we are comparing is mAh/hour, we could only compare battery % if:
1. We both use the exact same battery charged for the same amount of time.
2. We do the exact same activities for the exact same length of time.
We can then compare % loss. Seeing as that is never going to happen (as a truly fair test) the best we can go on is mAh loss/gradient decline.
If you want to get really clever you could break down the gradient into an equation, that would be the best comparison.
It would probably be better to get the raw data that Batterygraph collates and stick that into a spreadsheet.
How about this for a scientific test.
Test conditions
Charge phone until fully charged (green light)
Reboot phone
Close all widgets
Turn off Wifi - no random email/ IM notifications
Put phone into tablet mode - no random telephony
If Brightness status bar applet installed check level and note.
Set volume to standard eg full or 3 clicks down etc.
Start conky and note battery starting conditions, also note time.
Play media off local storage - eg full album or movie
At end of music note end conditions
This should ideally be done 3 days on the trot,with minor adjustments:
Day 1 - don't play any music - conrol test
Day 2 - follow test fully
Day 3 - run script after reboot and run test.
The reason for consecutive days is that there moght be cron jobs running in the background, and if you start at the same time every day then you can negate their effect on the test.
I haven't tried the scripts yet so I will try to run the test over the next few nights.
|
|
2010-11-16
, 14:55
|
|
Posts: 820 |
Thanked: 436 times |
Joined on May 2010
@ Portsmouth, UK.
|
#115
|
Originally Posted by mrsellout
Both phones would need a new battery and a fresh flash to make it a truly fair test. Or all non standard services should be stopped, i.e. ssh, six-axis daemon etc.
Test conditions
Charge phone until fully charged (green light)
Reboot phone
Close all widgets
Turn off Wifi - no random email/ IM notifications
Put phone into tablet mode - no random telephony
If Brightness status bar applet installed check level and note.
Set volume to standard eg full or 3 clicks down etc.
Start conky and note battery starting conditions, also note time.
Play media off local storage - eg full album or movie
At end of music note end conditions
Originally Posted by SavageD
I simply do not understand how this can be the case after flashing both firmware and eMMC.
Actually even though the remnants of my patch is gone most of the effects (performance wise) still remains. I have good reason for not mentioning this before, cause as you say it seems illogical.
If this is the case it would be one thing you should mention as you cannot undo your patches by flashing.
Last edited by James_Littler; 2010-11-16 at 14:58.
|
|
2010-11-16
, 15:06
|
|
Posts: 436 |
Thanked: 406 times |
Joined on Jan 2010
|
#116
|
The reflash part I can't stand...so I'll have to think more on this, in the mean time can't anyone who hasn't used my patches, upload a screenshot of their battery eye/graph of their battery?
One with standby and one with general usage, if your not over clocked it'd also be a bonus.
Pretty please.
One with standby and one with general usage, if your not over clocked it'd also be a bonus.
Pretty please.
|
|
2010-11-16
, 15:10
|
|
Posts: 436 |
Thanked: 406 times |
Joined on Jan 2010
|
#117
|
Originally Posted by James_Littler
No dude they are removed completely, believe me I checked
Both phones would need a new battery and a fresh flash to make it a truly fair test. Or all non standard services should be stopped, i.e. ssh, six-axis daemon etc.
I simply do not understand how this can be the case after flashing both firmware and eMMC.
If this is the case it would be one thing you should mention as you cannot undo your patches by flashing.
.It just that all the positive effects remain, as if it removed a huge load off the n900 somewhere.
|
|
2010-11-16
, 15:26
|
|
Posts: 889 |
Thanked: 2,087 times |
Joined on Sep 2010
@ Manchester
|
#118
|
Originally Posted by James_Littler
In an ideal world, yes, but I'm not going to do a fresh flash (although I may have to if it 'borks' my phone up
Both phones would need a new battery and a fresh flash to make it a truly fair test. Or all non standard services should be stopped, i.e. ssh, six-axis daemon etc.
), and I'm not about to buy a new battery just yet.I haven't used the patch yet, so if I do it on my device and graph the differences, over the 4 days (I'll come to that in a sec), then as long as the same conditions are kept over the 4 days, a comparison can be made of the patch only with on my device. Ok the battery will be 3 days older at the end, but I think the difference would be negligible. Oh, no updates or new software must be installed after the control test has been run.
After further thought Day 4 would be a second control test, with no media playing, but with the patches applied.
If other people ran the same test over 4 days then the comparison can be made only with their own data.
| The Following User Says Thank You to mrsellout For This Useful Post: | ||
|
|
2010-11-16
, 15:29
|
|
Posts: 820 |
Thanked: 436 times |
Joined on May 2010
@ Portsmouth, UK.
|
#119
|
Originally Posted by mrsellout
I wasn't being quite that anal, lol, All I meant is that it wouldn't be fair comparing a new battery to one that's been abused for nearly a year.
Ok the battery will be 3 days older at the end, but I think the difference would be negligible.
I think your test will be exactly what we're looking for.
|
|
2010-11-16
, 15:41
|
|
Posts: 436 |
Thanked: 406 times |
Joined on Jan 2010
|
#120
|
Ok I will perform the test mentioned by mrsellout. I would post my results at the end of the, 3 day peroid.
Last edited by SavageD; 2010-11-16 at 15:44.
Last edited by SavageD; 2010-11-16 at 15:44.
Plz, anyone who has never used any of my patches, plz try to confirm this.