PerformanceTest Mobile and Samsung Galaxy S II SGH-T989 baselines

[wonderwrench]

I'm now seeing three different listings in the database for what I think is the same phone.
The Samsung Galaxy S II SGH-T989 is a T-Mobile only phone. Canada has the Samsung Galaxy S II SGH-T989D. The database entries I question are

Samsung SGH-T989
CPU Type ARMv7 Processor rev 2 (v7l)
Number of Cores 2
Total RAM 767.6 MB
Most Common OS Android 4.0.4
Average CPU Mark 3,932
Average Mem Mark 1,633
Average Disk Mark 1,490
Average 2D Mark 1,933
Average 3D Mark 571

Samsung Galaxy S II (SGH-T989)
CPU Type ARMv7 Processor rev 2 (v7l)
Number of Cores 2
Total RAM 781.4 MB
Most Common OS Android 2.3.6
Average CPU Mark 3,869
Average Mem Mark 1,314
Average Disk Mark 1,700
Average 2D Mark 2,165
Average 3D Mark 876

Samsung Galaxy SII T-mobile
CPU Type ARMv7 Processor rev 2 (v7l)
Number of Cores 2
Total RAM 766.7 MB
Most Common OS Android 4.0.4
Average CPU Mark 3,834
Average Mem Mark 1,583
Average Disk Mark 1,927
Average 2D Mark 1,884
Average 3D Mark 555

I see two different Android OS's being used which may cause this to happen. Is it possible that the OS is identifying the phones differently which is causing PT mobile to ID the phones as different phones. Also is it possible that because there were hacked versions of ICS that we see two different versions of the same phone running ICS making the total T-Mobile Samsung Galaxy S II SGH-T989's three? Both ICS phones show as running Android 4.0.4. while my T-Mobile Samsung Galaxy S II SGH-T989 I upgraded to ICS shows 4.0.3 which is the latest official build available. PT mobile ID's my phone as a SGH-T989 ( SGH-T989 ) and the OS as 4.0.3 API level 15. The OS ID's my phone as a SGH-T989 and the OS as 4.0.3.

Other thoughts: I question the reasoning in figuring the SD card results into the disk score. This is because it can kill the disk score if anything less than a class 10 SD card is used. I figure PT Mobile is made to compare phone performance in a controlled way. Its not a controlled test once you add in the huge difference in SD card performance. IMO disk performance should be figured only by what is part of the phone. It would be nice to be able to see separate results for the built in flash memory and SD cards during a PT Mobile run.

Bill

[Richard (PassMark)]

Due to the fact that these graphs are automatically generated the names of the devices have been taken straight from the devices themselves there might be same device duplication. Some devices do not provide correct names. This can be seen in cases where custom ROMs are loaded onto the device or if the actual device may identify themselves as something different. From time to time, we will try to clean up the names and make them more readable.

We decided to leave the internal and external disk test under the test suite, as some of the devices that were used in development did not actually have "true" external drive. The built in storage was partition into an "internal" storage and an "external" storage. We will look into possibly separating the internal and external disk test during the next revision of the app.