Recently, Samsung has gotten some heat over the international versions of its flagship Galaxy S4 phone “cheating” on benchmarks. These phones use the Exynos 5 Octa system-on-chip. Under normal circumstances, the Exynos 5 Octa’s GPU runs at 480MHz. It was discovered in a Beyond3D forum post that the GPU hits 532MHz during certain benchmark tests. The wonderfully technical folks at AnandTech put the chip to the test and it appears that Samsung is guilty as charged.
On one hand, Samsung appears to be gaming the system. The company’s gumption is almost admirable, but this type of gaming can be misleading to a certain type of consumer. For better and worse, benchmark results impact some people’s purchases to varying degrees. They can be a good tool for making informed purchases, but benchmarks — particularly synthetic benchmarks — don’t always reflect the actual experience of using a computer (mobile phone or otherwise). For those that rely heavily on benchmark results, Samsung’s manipulations could lead to people buying phones that aren’t as powerful as the buyers think they are.
On the other hand, the performance gains from the artificial GPU boost are largely meaningless in real-world use. People that live and die by benchmarks (and I’m saying this as a former PC Magazine and Computer Shopper reviewer) should get over it and go buy a bicycle or something. If you think you’re going to have a vastly superior experience with your phone because it scored 800 points higher on AnTuTu than its chief competitor then you have larger issues to deal with.
Speaking of PC Magazine, I remember reading AnandTech with my PC Mag Labs cohorts when Anand Lal Shimpi wasn’t old enough to shave, drive, or drink. We’d make fun of him for being this nerdy punk that PC Mag editor-in-chief Michael Miller would kiss up to. More than a decade later, AnandTech is bigger than ever and has become a highly influential tech site…and here I am with a blog that’s like a geeky and (really) budget version of Grantland. FML…and also hug me please.