galaxys4

For a few years now many have used benchmark scores to determine the relative power and speed of an Android device. Many, including myself, refuse to use benchmark as a parameter to compare one device to another as we believe that they often have very little bearing on real-world use and experience.  Results shown by others using certain devices also should be taken with a grain of salt as there are many ways to give a false benchmark score.

A user over at Beyond3D forum noticed that the GPU of the Exynos Octa 5 processor runs at a higher frequency when performing benchmark tests compared with normal use.  These claims were investigated by Anandtech and found to be true.  They opened up the app TwDVFSApp.apk within the rom and found a string entitled “BenchmarkBooster”.  Busted.  Basically this string tells the processor, when an app listed within it is running, to clock at a higher than normal level.  Bad Samsung.

Screenshot courtesy of Anandtech
Screenshot courtesy of Anandtech

Samsung have released a statement via their Korean website saying that regarding the gamification of the results:

“[We] did not use a specific tool on purpose to achieve higher benchmark scores” and”under normal conditions, the Galaxy S4 operates up to 533MHz at its best performance.” 

Samsung said that some full screen apps are classified as requiring a higher performance and the GPU is then clocked while that app is running.  Samsung did not comment on the strings of code where the benchmark apps are specifically mentioned.

While I like that they have tried to address this I definitely have my doubts and this explanation reads as a poor excuse to attempt to smooth things over after being caught lying about the performance of their device.  I am sure their users would be happier if Samsung spent more time on making the S4 run smoothly than in so-called “benchmark optimization”.  Why waste time coding this when they could be improving the day to day experience of using the Galaxy S4?  Moral of the story, do not buy a device based on supposed benchmark scores.  Go for the best real world experience for you.

Do you have a problem with Samsung doing this?  Do you consider it false advertising?  Do you place any credence at all in benchmarks?  Will you continue to do so?