Open jungoo8512 opened 14 years ago
Chris : w/ app => 20% w/o app => 8% time => for 11 hours
*Tested in 100% battery charged HTC Desire phone
Does that mean the battery lasts longer with the app running than when it isn't?
20% power consumption when the app is running, 8% consumed when it isn't. :)
Thanks, that is very good battery life. How does the app go with our mods?
Need to redo this test. Done with HTC Desire which was disabling accelerometers when screen off, thereby reducing power consumption.
Nexus One Activity classifier vs Activity recorder. Activity classifier has further processing. It calculates SD for 3 axes for each sample set. Battery test data:- Run time 8hr 12 min Cell standby 6% Activity Recorder 3% Activity Classifier 5% Phone idle 4% Display 55%
Many changes to activity classifier. Nexus one 21 minutes 34 secs Nexus One resets battery timer when battery is removed and replaced. Activity Classifier 3% Activity recorder 3% Mobile standby 3%
31 minutes 59 secs Activity Classifier 4% Activity recorder 3% Mobile standby 3% Android system 4%
Sony Ericcson Xperia X10A X10 resets battery timer every time phone is turned off and on which Nexus One doesn't. 14m 57s Accelerometer failed after 4m 30 sec Cell standby 12% phone idle 9% Android system 7% Android OS 2% Activity Recorder 5% Activity Classifier 4%
Sony Ericcson Xperia X10A X10 resets battery timer every time phone is turned off and on which Nexus One doesn't. 14m 44s Accelerometer failed after 4m 30 sec Cell standby 7% phone idle 4% Android system 5% Android OS 2% Activity Recorder 3% Activity Classifier 2%
Nexus One 10h 10m 20s Phone in pocket not used much Activity Classifier 32% Mobile Standby 17% Phone idle 16% Display 8% Maps 8%
Nexus S 15h 19m 43s Phone in flat position Activity Classifier 55% Android System 20% Cell Standby 6% Phone Idle 2%
Nexus One 9h 26m 45s Full charge includes 1h 15 running my tracks and cardio trainer simultaneously Activity Classifier 8% Activity recorder 7% Mobile standby 6% phone idle 5% My Tracks 29% Cardio Trainer 22%
Nexus S 1d 9h 51m 1s Activity Classifier 40% My Tracks 20% Android System 17% Activity Recorder 12% Cell standby 4% Phone idle 2%
Sony Ericsson Xperia X10A Accelerometer always on no wifi, no calls, occasional display More than 36 hours before failure.
Nexus S Two test cases: First case with the accelerometer off and turning it on every 30 msec. Second case with accelerometer on all the time. In both cases the following listed applications were running on the phone. The test is done on the Beta version of the Avocado before the MYTracks integration. List of application running: Activity Classifier Activity Recorder Cell standby Display Phone idle Android System Android OS
Running time on battery data Case one: Started 12 night, flattened 7:50 at night the day after. 19:50 hh:mm almost 20 hours. Case two: Started 11:22 before noon, flattened 9:20 in the morning the day after. 21:58 hh:mm almost 22 hours.
Another test will be done without running the ACTIVITY Recorder(AR) application to avoid possible confusions with the accelerometer between the AVOCADO and AR.
Nexus S
Accelerometer On Full Time With FFT Running Ran from 27th April, 13.43 to 28th April, 00.58 A total of 11 hrs, 15 min
Test redone:
Ran from 2nd May, 20.19 to 3rd May, 08.27 A total of 12 hrs, 8 min
Nexus S
Accelerometer On Full Time Without FFT Ran from 28th April, 19.58 to 29th April, 08.15 A total of 12 hrs, 17 min
measurement might be wrong... there was some code related to FFT that was still running.. need to redo.
Test Redone: Ran from 3rd May, 18.22 to 4th May, 07.02 A total of 12hrs, 40 min
Nexus S....
Baseline test (incomplete, battery went from 100% to 68%) Running just phone system (no calls, no sensors on, no apps running) and an occassional battery check. Ran form 29th April, 17.24 to 2nd March, 11.44 A total of 66hrs 20min
Although the comparison between with FFT and without FFT isn't ready yet. Here is a graph showing batch processing time on a HTC desire for the first 30 batches. Note: The time decreases at the start, most probably because the phone is rooted and using a CPU profiler. The profiler adjust the clock speed of the CPU as the load increases.
Graph can be found here
Nexus S
Accelerometer On for 6.4s every 30s Without FFT Ran from 4th May, 20.35 to 5th May, 22.28 A total of 25 hrs, 53 min
Weird results in comparison to with FFT. Will redo test.
Nexus S
Accelerometer On for 6.4s every 30s With FFT Ran from 6th May, 16.22 to 8th May, 00.28 A total of 32 hrs, 6 min
Weird results in comparison to without FFT. Will redo test.
Compare battery life with Chris's version to new version.