Automated Real-World performance benchmark based on app launch-times

UPDATE ANDROID NOUGAT:

DiscoMark unfortunately does not work properly anymore under Android 7. The Accessibility Events that we use to determine launch times do not correlate anymore with the actual launch times as reported through, e.g., ADB. We will not prevent users with Android 7 from downloading DiscoMark, since it can still be used to automatically launch apps. This can be useful for, e.g., recording a video of the screen and then analysing the app-launch times like that. We are working on different ways of measuring real-world performance, mainly for professional testers. Periodically check on www.disco.ethz.ch for information, or send an e-mail to [email protected] if you want to know more.

DESCRIPTION:

DiscoMark tests your phone's performance based on the launch-times of your applications.

Instead of artificial computations that result in numbers that are hard to interpret, DiscoMark tells you how long it takes to start your Apps. This is a performance metric that reflects the real-world performance of your phone.

It is also a highly variable metric, that depends on many factors, such as number of installed applications, background processes, time since last restart, etc. Therefore, DiscoMark can help you identify performance bottlenecks, e.g., it helps you determine the effect of different measures, such as, clearing all running apps and restarting your phone. Just perform the same test before and after the measure, and then compare the results. No more guessing!

Furthermore, DiscoMark compares your results to those of other users, and thus lets you determine how your phone compares.

Another usage example: Pick some Apps and run the benchmark. Then, clear all your background Apps using your phones Application Manager and test again. Then, restart your phone and perform the last test. Now you can compare the results and figure out what improves the performance of your phone the most!

DiscoMark is an application that is being developed at the Distributed Computing Group at ETH Zurich.

In order to provide value for the user, and to gather data for our research, we log data about your device, and of course the benchmark results. We do not gather any personal data, and we only identify devices via an irreversible hash, generated form non-personal information.