No announcement yet.

Mobile performance suggestions and questions

  • Filter
  • Time
  • Show
Clear All
new posts

  • Mobile performance suggestions and questions


    I have a few questions and suggestions.

    Suggestions :
    S1.) make 1 list that compares all devices of a certain brand - that is a list, that e.g. Only lists 1 of each device from Apple but only the highest performing of each (thus , if you hoover with the mouse over, it could show the many different variations of performance of the gsm, wifi, etc versions of the ipad2).

    S2.) make a list that compares all devices of a certain type - e.g. Mobile phones - would have liked to compare my old nokia N95 with my new Xperia Z1... Even with my ipad2 even if its of a different type (and different os'es).

    S3.) cleanup of device names in some instances could be beneficial - i.e my Sony Xperia Z1 is listed as Sony Ericsson C6903, which is partially correct and incorrect in-so-far that Ericsson afaik isn't part of the brand name anymore. Technically it has C6903 as designator, but on the producers website its listed as Sony Xperia Z1 - i am not saying that passmark in and of itself should manually fix this, but rather give users a possibility to "set" or "suggest" a "device name string" - thus if there were a majority that suggested "Sony Xperia Z1" as "proper name" it could be automatically set - or something to that effect.

    Q1.) as the cpu and gpu in the newest ipad air and iphone 5s is the same, and then seeing the iphone 5s end up 2.5x ahead, even if battery performance is a lot lower, seems strange to me, but i am guessing hat the reason is that the iphone 5s has a much smaller screen size, and thus ends up with a lot higher fps than the ipad air does (1136x640 vs 2048x1536) - thus imho a screen res x fps would a more appropriate measure of the pixel pushing power ( could be a suggestion as well, perhaps)

    Q2.) the test on my xperia z1 was quite annoying as the app does not make sure that the device does not turn off, thus you have to keep pressing the screen every 10 or so seconds ( i know i could probably just change the screen turn off interval while running this app - but other benchmark apps have found a way to do this w/ o hiccup. In one instance the screen turned off middle the test and kept black for the rest of the test - thus the result ended up at some 5285 (BL#275044) vs yesterdays result of 4350-4464 (BL#274888, BL#274877, BL#274789), and another result where the screen saver kicked in, and made the result worse 4092 (BL#275029)

    A real comparison between all old and new phones would be very much appreciated, just like of old and new processors.

    Thank you very much for all your work in this area.

    Best regards: Cyberstorm.
    Last edited by Cyberstorm; Nov-14-2013, 11:14 AM.

  • #2
    S1. Whilst we agree that this would be useful, it's not something that we have in our immediate plans and would require a lot work. Currently we don't have the resources to undertake such an endeavor, it is however something we may like to do one day.

    S2. PerformanceTest runs on 3 different OS, Android, iOS, and Windows. Due to their different program languages and different level of access to the hardware, the 'Marks' are not directly comparable. Your Nokia N95 runs Symbian? Which is not an OS we support.

    S3. We do once in a while clean up the charts and do remap the model name to their more friendly commercial names. I'll see if we can spare some time in the near future to clean up at least the top devices.

    Q1. We do take into account the screen size when calculating the scores for the graphics test, however it seems, our "correction factor" is not biasing the larger screen size enough. We will look into this for future versions on PerformanceTest and decide if this will be adjusted.

    Q2. The Android version of PerformanceTest should prevent the screen/phone from going to sleep. We'll take another look at it and see if something has changed since it was last released.