Samsung artificially boosts Galaxy Note 3's benchmark scores by 20%

The idiom, "once bitten, twice shy" apparently does not apply to Samsung as its latest Galaxy Note 3 is found to perform up to 20 per cent better in popular benchmarks thanks to the company's artificial boosts.

According to Ars Technica, Samsung seems to be artificially boosting the device's benchmark scores with a high-power CPU mode that is enabled when selected popular benchmarking apps are being run on the Galaxy Note 3.

Ars Technica discovered that the CPU of the Galaxy Note 3 is set to 2.3GHz with all of its four cores activated when a popular benchmarking app is loaded. Under normal conditions (e.g. phone in idle mode), three of the four cores will be disabled to conserve power.

The team at Ars Technica was able to trick the Galaxy Note 3 into not entering the high-power CPU mode by disassembling the benchmarking app (in this case, Geekbench 3) and changing its name to Stealth Bench. When running Stealth Bench, the CPU is set to idle with most of its cores shut off. The results are startling; the Galaxy Note 3 is found to perform 20 per cent better in Geekbench 3 when compared to its clone (Stealth Bench).

Ars Technica went one step further to extract and examine the file that triggers the boost behaviour. The file is revealed to exclusively target benchmarking apps and covers all the popular apps such as Geekbench 3, Quadrant and Linpack. "Stealth" versions of these apps were also made to test the other benchmarks, and results were similar with the Geekbench findings.

One ironic finding that Ars Technica stumbled upon is the fact that the Galaxy Note 3 still performed better than the LG G2 with the high-power CPU mode disabled. The G2 also runs on the same Qualcomm Snapdragon 800 quad-core processor.

Samsung faced similar allegations when AnandTech reported in August that the GPU of the Galaxy S4 was tweaked to perform better in several benchmarks. Samsung responded with an official statement clarifying that the maximum GPU frequencies for the Galaxy S4 have been varied to provide the optimal user experience, and were not intended to improve certain benchmark results.

While benchmarks' results are included in our reviews of smartphones and tablets, it is important to note that we also weigh in the user experience in real world scenarios. As mentioned in our review of the Galaxy Note 3, we felt that the user experience wasn't any better than any of the recent Android flagship devices we've tested.

Samsung not the only one cheating in benchmarks

It turns out that Samsung is not the only phone maker to tweak the hardware of its devices to perform better in popular benchmarking apps. Apparently, this practice is widespread among Android phone makers.

AnandTech tested several other Android devices and found that they run some kind of CPU optimization. According to AnandTech, Apple, Motorola and the Nexus devices are the exception.

The findings reveal that the ASUS PadFone Infinity, HTC One, One mini and LG G2 have their own list of benchmarking apps that matter and need optimising. The tech site also claims that this practice is an OEM level decision, which is not advocated by the chipset makers.

If you are hoping the recent spotlight on this matter is going to make the Android vendors change their minds, you are wrong. AnandTech states that the situation is going to get worse. OEMs should divert their resources to making better software and devices instead of playing in this dirty game of inflating benchmark scores.

Visit Hardware Zone for more stories.