Mobile testing using Benchmarks

Hi,

We know that many devices vary between each other in their HW capabilities, meaning – the fact that 2 devices by 2 different or similar OEM’s, running the same OS/different OS version will score differently in a benchmark run or a performance test.

This is usually due to hardware differences, defects related to memory allocation, and others.

The use of mobile benchmarks had become quite significant and useful and can allow the application developers and testers to do objective testing on many different devices using a benchmark tool and get unique scores.

By comparing the overall score among the devices under test they can reach a conclusion which device has the higher performance and decide to either use it as “Gold” device for its porting platform, or if the score is extremely low, use it for testing when that device serves as the low performer in the list allowing the team to gain some prerequisite information on how their application would run on the device/s.

The benchmarks are usually downloaded for free from the Android or App store markets, and allows at the end of the quick test run to compare the device score with other devices which ran in the past and see the overall position of the device under test in compare to the market.

Usually the tests involve mathematical calculations (such as finding the prime number from a list), string sorting or number sorting, 2D/3D gaming with FPS calculations, reading and writing to a memory of the device (SD Card etc.), and some other CPU tests.

I strongly recommend to use such tools as part of the testing and as part of the decision for porting a new/existing device.

Information on commonly used benchmarks is available in the web, however i have good experience with the following ones:

– For Android – PassMark and Antutu (Which also provides Battery saver tool and CPU master to control the CPU speed)

http://www.androidbenchmark.net/passmark_chart.html

http://www.antutulabs.com/AnTuTu-Benchmark