Samsung discovered that it increased the benchmark result of GALAXY Note 3 by 20%

It is scheduled to be released from OctoberGALAXY Note 3However, using the benchmark application has been programmed to act incorrectly, and it became clear that the results were inflated by 20 to 50%.

Note 3's benchmarking "adjustments" inflate scores by up to 20% | Ars Technica
http://arstechnica.com/gadgets/2013/10/galaxy-note-3s-benchmarking-adjustments-inflate-scores-by-up-to-20/

Ars TechnicaAnd while testing the already released GALAXY Note 3 in the US,benchmarkI found strange behavior in the test. For the same CPUSnapdragon 800Equipped withLG G2As a result of comparison with GALAXY Note 3 sold in the US, it is said that the CPU was artificially programmed to operate in high output mode when launching the famous benchmark software.


To verify the benchmark testGeekbench 3use. On the left of the photo, it shows the normal state, the three cores are off, one core is idle with 300 MHz. On the right side, with all the cores operating at 2.3 GHz with the famous benchmark application Geekbench 3 running, it is locked with the fastest clock speed.


The benchmark test exists to measure the performance at the time of normal use of the smartphone. In Ars Technica, all functions other than name are the same as Geekbench 3 in order to obtain the correct test resultStealthbench"We created the rename version of Geekbench 3 and verified the method by not moving GALAXY Note 3 to" benchmark mode "the following picture. Stealthbench on the right maintains the idling state, and when comparing these two results, it is obvious that there is a difference of 20%.


In response to the point that the "benchmark · booster" which operates illegally the result of the benchmark test will work with GALAXY S 4 before,Samsung's official blogIn "S browser, gallery, camera, video player, benchmark test application and other specific applications that require maximum processing speed, in order to raise user comfortGPUIt is permitted to operate at maximum speed ", but Galaxy Note 3 is the first time in the United States that such artificial manipulation was done in the United States.


As a result of examining the script, besides Geekbench,Quadrant,Antutu,LINPACK,GFXBenchIt is known that the "benchmark booster" is running only for the famous benchmark test application such as "LINPACK", which means that up to 50% improvement in test results was seen in LINPACK.


At the moment there is no official announcement from Samsung, but Ars Technica has gained evidence of the same behavior on Galaxy Note 10 tablet and will continue to verify. Ironically, the normal result of the benchmark test of GALAXY Note 3 was that it was superior to LG G2.