iPhone 3G slow speed linked with location services?

Discussion in 'iPhone' started by spyz88, Jul 13, 2008.

  1. spyz88 macrumors regular

    Jan 29, 2008
    OK I need some people with iPhone 3G's to test this out. I did a series of speed tests on dslreports.com on my 3G iphone. The results were 720 kbit/sec on Wifi, 47 kbit/sec on 3G and 37 kbit/sec on EDGE.

    I went ahead and was rebooting and playing with various settings. I turned off location based services and my results were 718 kbit/sec on Wifi, 528 kbit/sec on 3G and 45 kbit/sec on EDGE.

    Can some other people please test this out and see if this happens on your iPhone 3G's also? If so hopefully we can report it to Apple and its a matter of a simple software update not having to get our phones replaced.
  2. iPaul macrumors regular

    Feb 18, 2006

    LS On
    Wifi: 3219kbps
    3G: 217kbps
    EDGE: 81kbps

    LS Off
    Wifi: 3273 kbps
    3G: 507kbps
    EDGE: 85kbps

    maybe you've got something here. i am going to turn off the location services and just turn it on when i need the gps. its a drain on the battery anyway
  3. View macrumors regular

    Apr 18, 2007
    The second tends to be faster for some reason. I have a iPhone V1 2.0
    here are results

    LS ON
    Wifi 1401
    EDGE 173.4

    LS OFF
    Wifi 1281.1
    EDGE 150.6

    ** I tested it with LS off first, then turn on the LS..
  4. lakaiordie macrumors 65816

    Jun 17, 2008
    dunno if it has to do with it. because when i've had it on ive gotten 1.7mbps

    im gonna turn location services off when its fully charged and see how much longer it lasts
  5. Jeremy1026 macrumors 68020


    Nov 3, 2007
    LS On

    LS Off

    We should clear cache's between tests to get the most accurate results though. I will do that tomorrow when I am killing time at work.
  6. nickspohn macrumors 68040


    Jun 9, 2007
  7. lakaiordie macrumors 65816

    Jun 17, 2008
    ls on


    ls off

    first time i got 2.6
    then 1.65

    i dont see much of a change

    *update*cleared cache and cookes

    ls off still

  8. lakaiordie macrumors 65816

    Jun 17, 2008
    i think you're right about this

    im getting consistent back to back tests that are higher than when i had it on.
  9. Romanesq macrumors 6502a


    Jun 16, 2003
    Even with LS off...

    I'm consistently getting low scores in the range of the 70s. His score of three was 96K.

    Isn't that pretty poor for 3G. I'm in the northeast NYC area and that's an area with good coverage no?
  10. scariaga macrumors regular

    Aug 24, 2007
  11. scariaga macrumors regular

    Aug 24, 2007
    never mind :p
  12. TechIsCool macrumors member

    Jul 13, 2008
    - General
    ---Location Services ON/OFF
  13. kdarling macrumors P6


    Jun 9, 2007
    First university coding class = 47 years ago
    The opposite seems to be true as well.

    A lot of people have commented that they couldn't get the GPS to work until they turned off 3G for a little while and reverted to EDGE.

    Okay, so here's my thoughts:

    1) The GPS chip used is not standalone. It's mostly just a receiver, and depends on the host cpu to do all the pre-setup and post-calculations. So the CPU has to be shared between web page parsing (speed calculations) and GPS calcs.

    2) Perhaps ATT has only partially implemented the A-GPS protocol over the EDGE control plane, and hasn't finished rolling it out for 3G. So the phone is trying to get assistance info but it isn't there. Or it is there, and is slow.

    Or some combo of the above.

    So while checking 3G speed with LS on/off, also might check LS with 3G on/off. Bearing in mind that once assistance is gotten, it might not be needed again for a few hours. Perhaps a full reset can wipe it out and make it go again.
  14. dlamin517 macrumors 6502

    Jul 12, 2008
    No change for me. Still crappy in my room. Im gonna keep it off though becuase it will save battery life.
  15. wonderbread57 macrumors 6502

    Jun 11, 2008
    So your theory seems to be that speedtests are slower with LS on because the CPU can't "keep up" because it is constantly processing the GPS chip's feed? If this is the case then certainly the firmware should be fixed so that the OS ignores the GPS data until the location service is specifically requested (ie. the find location button is pressed by the user). The question of weather the GPS is a standalone unit or not, requiring extra CPU cycles, should be able to be addressed by looking at the spec sheets for the Infinion GPS used.

    Spec sheet: http://tinyurl.com/5afku2

    The sheet makes mention to CPU load here:
    "GL host-based architecture leverages
    some of the resources already existing
    in the mobile device without imposing a
    big CPU load (~ 3 – 6 MIPS)"

    What does that translate to in terms of cycles on a ~600MHz ARM device like the iPhone? In any case it seems as though it's not a lot of CPU.

    Regarding actual RF interference, the sheet doesn't address it directly but says:
    "Hammerhead II uses the proven host-based
    architecture pioneered by GL as the best
    fit for mobile devices as proven in mass
    market 2.5 G and 3 G devices shipping

    I think this statement is taking some liberties. Mainly, it says it's tech has "proven" to be the best fit and that evidence for that is the 2.5 and 3G devices shipping now. So really, if the devices shipping now have problems with their chipset, it certainly doesn't prove it to be the best. Time will be the judge of that.

    Lastly, just because the GPS is receiver only (does not transmit) does not mean it can't interfere with 3G RF. There are plenty of receivers in the wireless world that require a license to operate because they are very powerful and give off a large EMF (open to corrections on the actual technical aspect of it).
  16. kdarling macrumors P6


    Jun 9, 2007
    First university coding class = 47 years ago
    No, I wouldn't think 5 MIPS from a cpu running at ~412MHz (iPhone doesn't run at 600MHz) would matter much... assuming the software is written correctly.

    I think it could be a matter of the OS stopping downloads in order to ask a (non-existent) assistance server for help.

    EDGE and 3G use different methods to ask for assistance, but if they're using control plane methods, then that could prevent data transfers. Or if they're using TCP/IP methods, that data could take priority.

    Supporting evidence is that GPS off = faster 3G, and 3G off = faster GPS.

    (For some people, in some countries, depends on carrier perhaps.)

Share This Page