This object is in archive! 

is power consumption dependent on offline map size/structure/location?

Tom Grundy shared this question 8 years ago
Answered

Hi - we will be going on a 2+ week bicycle trip in the desert and will be using Locus Pro on a Samsung Galaxy Tab Pro 8.4.

On some multi-day training rides, we've noticed some higher battery drain on some rides than on others. Typically we only use it for map checks at junctions - viewing the map with the current GPS location whenever we come to obscure junctions, to make sure we are where we think we are. Probably 15 seconds of usage, once every 15+ minutes. The rest of the time the screen is off and locked (but the Locus setting 'Disable Lock Screen' is on - very handy).

Anyway, as a programmer, my theory is that the battery drain rate is a function of the database size, i.e. how many total tiles exist in the map database that is being viewed.

For one trip, we downloaded all the tiles for the immediate area, but limited it to 31.50MB, zooms 11,12,13,14,15. Battery consumption seemed very low.

For a more recent trip, we used a much larger map, 3609.84MB, zooms 7,8,9,10,11,12,13,14,15; but only down to 14 in the area we were using.

Do you know, or have any guesses, as to how these things might affect battery consumption? It seems like searching a larger database would take more CPU power than searching through a smaller database.

So, one thought we had to reduce consumption for the long trip is to just fracture the entire database into geographically smaller regional maps, then specifically select the map that's needed day by day.

Could it be taking more CPU power to search for zoom 15 tiles, in an area where we only downloaded to 14, even though other parts of the same map database are 15?

Does the map database type (sql) matter for power consumption?

Also, would it make much difference if the map is on the internal memory versus the external sd card?

Thanks!

Replies (3)

photo
1

Hi Tom, interesting topic and discussion. You haven't quantified your actual power (energy) consumption. Are you measuring this by battery %? I use a similar strategy to you for minimizing energy consumption although I now have a front wheel dyno and B&M EWERK AC to DC converter so I can keep the phone charged up without stress. I use an app called

Battery Saver in slumberer mode to keep background processes inactive. My battery % measurements with a Sony Xperia Z1 (2.3 Ah battery) are


  • less than 0.5%/h usage when Locus not running

  • 0.5%/h with Locus running, display off, GPS off
  • ~2%/h with Locus running, display off, GPS on
  • ~3%/h with Locus in guiding mode, display off (auto on for minimum period), GPS on

You can set Locus to only enable the GPS when the display is turned on. This works well for minimizing consumption. I've got no idea in relational to CPU usage and database access.

photo
1

Hi guys,

My few cents:

1. I believe that size of SQLite map file, will have a minimal impact on battery consumption. Getting tiles from SQLite database is really fast no matter how big the database is, so I think there should be different reason

2. speed of SD card - faster are tiles loaded, less iterations Locus needs to do, anyway as well as previous point - I think that impact here should be minimal

@Andrew: I'm surpised that enabled guidance has such huge impact. It should be only a "few" computations once per second (when new GPS location is received) and I never expected such difference. Also fact that Locus that was started consume any battery is a surprise. I hope, that when you start Locus and then put it on background without any active service (like track recording etc.) should consume 0! battery. If not, there is any problem ...

I believe that most battery consuming features in Locus are:

1. vector maps

2. shading of maps

3. a lot of locations on a map (loooong tracks, few thousands of points)

Generally redrawing a map is what consume most battery I believe, so limit amout of items on a map, do not use shading if you do not need it, disable "Periodic updates" if you do not use any add-on that needs it, and you should be fine.

photo
1

Isn't vector map rendering cached into bitmap tiles, including the shading?


Hmm... apparently not. Zooming Munich with an RMAP is much quicker than mapsforge, even when using the same parts over and over again.

photo
1

Very interesting discussion Menion - thanks. Similar to what I suspected. My 3%/h should be fairly accurate because it was measured separately over two days cycling with about 6 hours per day. Oh course each phone will be different as battery capacity is different. I was using guiding mode in which display was on for longer at some waypoints than if I were manually turning on/off the display, so maybe that explains the difference when compared to "2%/h with Locus running, display off, GPS on". The 2% measurement is only for 1 hour rides, so not as accurate, maybe it would be 2.6% (say) if I did identical 2 day ride. Either way I am quite happy with the power consumption. Excellent suggestions on hill shading and long track. I have divided an upcoming 1200km ride into 4 tracks because converting track points into navigation waypoints was getting very slooooow .

Replies have been locked on this page!