Galaxy Note 3 benchmark performance artificially boosted by Samsung, new report proves

October 1, 2013
182
285

Galaxy Note 3

Samsung is apparently not done optimizing its handsets to behave differently when running a benchmark app, with the Galaxy Note 3 also receiving the same treatment.

Must Read: Samsung Galaxy Note 3 Review

A new report from Ars Technica offers more details on how Samsung is intentionally boosting the performance of the U.S. Galaxy Note 3 to score better than its competitors in benchmark apps – performance improvements of up to 20% (even 50% in one case) have been observed by the publication during tests.

One of the things Samsung was proven to have done to improve scores was to prevent the phone’s processor cores from idling when running a benchmark app – instead they are all kept running at maximum speed. But the company is also suspected to have improved GPU performance and altered the display frame rate when running benchmark apps.

Combined, these action result in much better performance reports for the Galaxy Note 3, which is able to score consistently better than the LG G2 in similar tests, despite both devices using the same processor.

The publication was able to fool the Galaxy Note 3 into treating benchmark apps like normal apps by repackaging the benchmark apps and replacing their official names with a different name. That way the Galaxy Note 3 failed to detect them as benchmark apps and a “regular” benchmark test was performed – apparently the phone only checks for certain app names to trigger the CPU boost.

Interestingly, the Galaxy Note 3 still scores better than the LG G2 in same tests once Samsung’s boost is removed, but not by as much as with the boost effects on.

Not too long ago, it was discovered that Samsung’s Galaxy S4 is also offering a boosted performance in benchmark apps – that time around a GPU boost was found. Because Samsung explained back then that the same boosting effect can be found with core apps such as “the S Browser, Gallery, Camera [and] video player,” in order to improve battery life and heat dissipation on the Galaxy S4, Ars Technica investigated whether the Galaxy Note 3 performance boost can be found in other Samsung apps for similar reasons.

The publication discovered that CPU boosting is only set to occur for benchmark apps:

The file we ended up with is called “DVFSHelper.java,” and it contains a hard-coded list of every package that is affected by the special CPU boosting mode. According to this file, the function is usedexclusively for benchmarks, and it seems to cover all the popular ones. There’s Geekbench, Quadrant, Antutu, Linpack, GFXBench, and even some of Samsung’s own benchmarks. The two functions applied to this list seem to be “PACKAGES_FOR_BOOST_ALL_ADJUSTMENT,” which is no doubt the CPU booster, and “PACKAGES_FOR_LCD_FRAME_RATE_ADJUSTMENT,” which makes it sound like the phone is also altering the display frame rate.

In the gallery below, you can check out images from Ars Technica that show how the Galaxy Note 3’s processors fail to idle and run at full speed when a benchmark app is used, the “Stealthbench 3” Geekbench 3 version used to fool the Galaxy Note 3 into treating a benchmark app like a regular app but also benchmark results (single- and multi-core) for the Galaxy Note 3 (with and without booster), the Optimus G2 and Galaxy S4.

With all that in mind, we can only wonder why a mobile player like Samsung still feels the need to employ such tactics – whether there’s a different reason this kind of performance boost happens during benchmark tests, Samsung should at least have explained it beforehand.

At the end of the day, benchmarks alone will not help sell handsets – and in Samsung’s case that shouldn’t even be a problem considering that some of its flagship handsets sell themselves.

Finally, the Galaxy Note 10.1 – 2014 edition may have a similar benchmark performance boost as the Galaxy Note 3, the publication notes.

While we wait for Samsung to explain everything, let’s hear what you think about Samsung’s benchmark cooking practices in the comments below!

Comments