Recently, it came to light that Samsung and other mobile manufacturers were cheating current mobile benchmarks by upping processor performance each time these benchmarks were run on their devices. While this isn't something new in the tech industry, quite a lot of people were shocked, declared they never use benchmarks anyhow and made a general fuss about the whole issue. A startup called GameBench who have direct working experience with both ARM and MediaTek recently made an "uncheatable" benchmark for mobile gaming. The beta version of this GameBench benchmark was then run on the Samsung Galaxy S4 (the Snapdragon 600 version) and HTC One, two of the most contested smartphone models of last year (and supposed to have cheated previously). Turns out that the performance of the Samsung Galaxy S4 is still better than the HTC One (and closer to the original "cheating" scores than previously thought of). Check it out in the charts below:
The performance part of the GameBench benchmark runs an actual game and takes score from the background. This matches with industry graphic card testing and covered games like Despicable Me - Minion Rush, Real Racing 3, Dear Hunter 2014 and Dead Trigger. Each game tested different aspects such as the complexity of the graphics to how many objects were usually on the screen at the time. However, in all of the games run, the Samsung Galaxy S4 score 13 to 2 points ahead of the HTC One.
The other part of the GameBench benchmark covers battery drain while the games mentioned above are running. Again, the Samsung Galaxy S4 pulled ahead, especially in Minion Rush except for Dead Trigger where the HTC One was ahead due to it's Dead Trigger optimizations. The main conclusion we can come from this is that while it may be true that Samsung have been cheating, the real-life performance of Samsung devices like the Samsung Galaxy S4 is actually very close to those found in these "cheatable" benchmarks.
[Source]
COMMENTS