Energy consumption in Locus while tracking

[franc]

Hello,

in another forum (http://http://www.android-hilfe.de/karten-navigation/157283-ist-denn-nun-besser-locus-oder-oruxmaps-3.html#post5792368) , where there is a discussion about Locus and Orux I saw this picture about energy consumption in Locus vs. Orux while tracking:

http://www.bilder-upload.eu/upload/6609 ... 125610.jpg

If this is really authentic, I mean if a thorowly test on more than one device with more than one OS etc. would show the same result, would it be possible to reduce the energy taken by locus while tracking?
Or could it be some mis-settings in the tracking that Locus consume more energy?

Thank you!

frank

[menion]

Hello Franc,
  result visible on image is of course possible, but I'm unable to say if compare is correct or not. There is a few factors that have influence on CPU usage. Some of main factors as they come on my mind are for example

- was map screen visible sometimes during a record? If so, then same time on both apps?
- were used vector maps? They're much more CPU consuming then raster maps
- were enabled any sensors? Pressure, rotation, gravity?
- was set same frequency of logging
- were number of items on map same in both apps? Because even if screen is turned off, Locus still do some small tasks to check if there isn't need any activity after screen has moved ...

  anyway if there will be a way to find, where should be (if any) problem in Locus with battery usage, I'll gladly improve it. Anyway, monitoring battery usage just by comparing of one measure with different program that should have different settings, isn't best way. Suggest to precisely check settings and do some more tests. For example once start Locus first, then Orux, then after some tests restart device and start programs in opposite way, this should have also influence on visible values of CPU/battery usage, etc etc ...

[franc]

Thank you for the analysis.
On spare time, I will try this comparison with hopefully similar parameters in both apps.

[tommi]

I doubt parallel run of both programs will give reliable results. E.g. sensor usage is sometimes counted wrongly (see viewtopic.php?f=38&t=3005 (sorry, only for Premium members ).
Why is CPU usage in the test much higher for Locus (vector maps etc.?)?

@Menion:
What are the things Locus does when screen is off?
Surely Locus needs to calcuate POI alarm, Guide to, navigation, etc.
But no vector map tile calculation, calculation and display of map objects, map items, track shift over screen, etc..

[menion]

you're thinking correctly. Imagine all locations (waypoints and track points) has pre-computed coordinates, and once per second, when new GPS location arrive, I just move these locations to new place by simple +X px, +Y px. That's all I do. No maps are rendered, no images are drawn ...

steps are
- 1. new GPS location
- 2. move all map items
- 3. call refresh on map
- 4. draw map tiles
- 5. draw map items

and only 1. and 2. points are done when you have screen off. You may test it on your own. Do you know trick with nmea files placed into Locus/cache/nmea directory? You may use them for GPS simulation. Then you may at home, simulate track record. Start record, turn off screen and when you turn it on, you'll see firstly your old location before you turned screen off, then big jump on current location and then finally start Locus draw a map ...

Anyway there exists some special methods for measuring which function too most CPU, so in spare time, I can check it ...

EDIT:
 I've just checked it and
50% of power took simulator that read and parse NMEA data from SD card
15% took Altitude manager to compute Geoid height and optimize measured altitude
13% for my big suprise took track record widget even if screen was off and widget is on not-visible page of homescreen. Good to know and good place to optimize
5% took formatting texts for some text fields in top bar etc ...
and rest took handling of location by track record service and some more stuff around

so these 50% (by simulator) really do just from time to time, read a line from text file and parse it. Not much work and it's 50% of CPU that Locus consumed. So this mean that locus consume really really small CPU for this task.

Anyway you bring me to new task on tomorrow ... check widget

[franc]

I just tracked again some track with locus and at the same time with GPSLogger, just to compare energy consumption.
So here my battery report as glued screenshot attached.

Locus used for about 40 minutes about 6 % of battery, but the other pure tracking app, GPSLogger, didn't even be mentioned in the battery list at all!
Is there something wrong with my locus settings then?
Should I disable NMEA or how can I get normal tracking results?

I have the Widget enabled, I use tracking foot with 1m/10s/100m
In GPSLogger I use track every 5 seconds (no more settings with this).

Can I do something here or is it just better not to use locus track recording if I need low battery consumption?

Thank.

[franc]

It is even worse!
Now I tracked over an hour, without any display use, I started tracking with the widget and stopped it and immediately looked to Battery, this is what I see, 30 % Locus! This is unusable then, this tracking. I must have a setting which pulls the battery empty, or how could that be (see attached Screenshot)?

[franc]

Yes, you surely right, but the test with both apps, Locus and Gpslogger simultaneously tracking the same time and Locus is 30 % and the other app is not even mentioned is a very clear thing.
Either I have misconfigured Locus and it should not be mentioned as well, or Locus just needs much more battery.
I started my SGS2 with nearly 100% battery and after this tracking was in the fifties % which is much less.

I like the tracking of locus, I used to track with locus before, it has this nice export feature.
But this huge power consumption for just tracking is too much, unfortunetly.

[szebenyib]

I would not believe those numbers one hundred percent. I don't know how much Locus consumes. But maybe the first app to start GPS will get the penalty and the other one will not receive anything from the GPS consumption.

[menion]

guys, I don't want to write here anythink long, because on this topic I already wrote almost a small book

Simply: don't trust internal battery monitor. Mainly on older devices, always first app that start using any sensor is that one which count battery. Try some other monitor like GSAM (recommended by my personal experience)

[franc]

Maybe this wrong. I just started this time for sure first GPSLogger, then, after a minute, locus,  starting it over the widget.
Still I have now 6% battery for locus, not mentioned GPSLogger, all the time screen on with other apps, but never locus.

I guess it's more like taking my 40 tons lorry to get cigarettes around the corner instead of taking the bicycle

But I guess the real test would be to start locus tracking in airplane mode for an hour with full battery and then look only at the battery percentage loss.
Same then with the other tracking app.

[franc]

Now a better test with your battery app "Gsam", which should be real, as you told, but doesn't differ from my OS-battery report:









First I started GPSLogger (II) and only after this, Locus via widget.
Stopping the track record again in the end with the widget and then, when Locus pops up, I just deleted the track not to consume battery by auto export.
I started with 96% on battery and ended after 25 minutes with 85%.
If these 11% are right, then Locus used in one hour about 3% battery.

Please tell me that I MUST have some misconfiguration in Locus, because this cannot be that Locus consumes so much, can it?
What can it be that makes locus so hungry?
Network was off.

[menion]

hmm and why Locus has during 33m power on, turned GPS on only 24 min and your second app only a minute??

Anyway as you see, during that time, Locus spend 5x times more time by some CPU operations.

But such measuring has really only sense, when you use same conditions on both apps, they will run separately (not both at once), longer time and in airplane mode. Best on just factory reseted device

[franc]

hmm and why Locus has during 33m power on, turned GPS on only 24 min and your second app only a minute??
Anyway as you see, during that time, Locus spend 5x times more time by some CPU operations...

This is what I don't understand, WHAT in hells and heavens name is locus doing all the time using CPU?
It just need to track. What could that be, that is running with this track-recording in Locus?

...But such measuring has really only sense, when you use same conditions on both apps, they will run separately (not both at once), longer time and in airplane mode. Best on just factory reseted device

True, indeed!
And beside this, even when Locus needs 3% an hour, this would mean 30% for ten hours, which I would never reach, never.
The system would drain the battery much earlier, not a 3% an hour App. Would mean maybe 10 minutes more tracking, I guess, if Locus would consume only 0.3%, like a smaller tracking app (a "bicycle").

[franc]

I mean this:









In 50 minutes, Locus consumed 56% from 6% of the battery as Gsam reports (or 45% of 7% in 70 minutes as OS reports).
This was measured with airport mode on.

Or:



This without airplane mode.

I am now sure, GPSLogger II consumes about a tenth of it, but all in all its not much.
Will be only interesting for long time tracking, I mean tracking for several hours or even a day. There I won't be able to use locus if I have not a bunch of spare and full batteries with me.
Locus is just not a tracking app.

Roger and over.