Samsung putting an end to benchmark cheating with the KitKat update for Galaxy S4 and Note 3

Samsung putting an end to benchmark cheating with the KitKat update for Galaxy S4 and Note 3
Benchmarkgate seems to be winding down, as Samsung's Android 4.4 KitKat update apparently ends the CPU-boosting tricks previously employed. Before the update, for instance, the Galaxy Note 3 entered into an overdrive mode each time a benchmark app was started. This regime fired up all processor cores at full speed, even before the test itself was run.

A brief research showed that a lot of manufacturers do it, but evidently the KitKat update brings an end to the Wild Wild West era of benchmarking, so prep yourself for more modest, yet more credible results. The problem was that CPU throttling, which idles the cores to their minimum frequency when there's no workload, immediately went on the backburner if Quadrant, AnTuTu, or other popular benchmarking apps were run. The cores revved up to their respective maximums, which in the case of the Note 3 reach 2.3 GHz, and stayed there until the benchmarking app was shut down. The same goes for the Galaxy S4, whose core clock maxes out at 1.9 GHz. Its average result on Geekbench went up from 1812 on Android 4.2.2, to 2114 on Android 4.3, and now down to 1913 on KitKat.

We can't really call this cheating, since everyone wants their product to be at its best when tested, and there is no actual overclocking from the official spec sheet maximums. Massaging the truth would be more appropriate here, since that's not how the chipset cores behave in everyday usage. In any case, the KitKat update seems to put an end to these shenanigans for Samsung, and, as we tested on our office G Pro 2, on LG's new crop, too. Whew!

source: Arstechnica

Related phones

Galaxy Note 3
  • Display 5.7" 1080 x 1920 pixels
  • Camera 13 MP / 2 MP front
  • Processor Qualcomm Snapdragon 800, Quad-core, 2300 MHz
  • Storage 64 GB + microSDHC
  • Battery 3200 mAh(21h 3G talk time)
Galaxy S4
  • Display 5.0" 1080 x 1920 pixels
  • Camera 13 MP / 2 MP front
  • Processor Qualcomm Snapdragon 600, Quad-core, 1900 MHz
  • Storage 64 GB + microSDXC
  • Battery 2600 mAh(17h 3G talk time)

FEATURED VIDEO

28 Comments

1. NoFanboy

Posts: 300; Member since: Nov 18, 2013

News: Samsung finds more advanced way of cheating on benchmarks that has yet to be discovered

4. tech2

Posts: 3487; Member since: Oct 26, 2012

Give credit when its due. Almost all companies cheat but Samsung is the one to stop first. Still people like you find a reason to complain. Haters gonna hate. smh

6. AliNSiddiqui

Posts: 382; Member since: Sep 19, 2012

Don't be naive man.. NoFanboy has good reason to be wary of Samsung from now on..

8. joey_sfb

Posts: 6794; Member since: Mar 29, 2012

Full speed at bench mark is normal. Being doing on the first x86 system since the NEC V20 chip. Its my first overclocked x86 cpu.

26. blazee

Posts: 414; Member since: Jan 02, 2012

Normal for those who cheat

14. itsdeepak4u2000

Posts: 3718; Member since: Nov 03, 2012

This is very true: "everyone wants their product to be at its best when tested" so does Sammy and why not? But KitKat update is very good for stopping such things. But but but I think Sammy will somehow find a alternative for this too. ;) Bring it on Sammy. :)

18. tech2

Posts: 3487; Member since: Oct 26, 2012

Re-read the article. Its not Kitkat its Samsung who is putting this to an end.

20. itsdeepak4u2000

Posts: 3718; Member since: Nov 03, 2012

Oh really? Don't tell me this, please read second paragraph's first sentence.

28. shahrooz

Posts: 792; Member since: Sep 17, 2013

You don't know what open source means, do you?

2. XaErO

Posts: 353; Member since: Sep 25, 2012

Excellent !! .. That should be a close demonstration to real world usage !!

3. ABDULGHAFOOR

Posts: 106; Member since: May 04, 2012

Bad for Samsung :-D

17. livyatan

Posts: 867; Member since: Jun 19, 2013

Nope..my Note 3 on KitKat still destroys every bench. 35 200-36 300 on AnTutu, 26 fps on TRex, etc..

25. grapeseed87

Posts: 123; Member since: Mar 13, 2013

Wait for the update man.

5. Diezparda

Posts: 941; Member since: Oct 23, 2013

why rely on benchmarking test? i think the actual performance that really matter.

7. _Bone_

Posts: 2155; Member since: Oct 29, 2012

I never considered it cheating since benchmark results were always REAL even if the extra performance was only allowed for a handful of apps (including S appa). The point for benchmarking is to "see the end" so modders, the only ones who should really care, to see how far they can tweak their ROMs once they get into the SoC. Or ask yourself this: if an XDA developer comes out with a tweaked ROM performing extra benchmark performance - is that cheating too? But for the unqualified crowd, Samsung and LG are doing the right thing to step back to stock clocks (or is it KitKat limitations?)

9. superduper

Posts: 151; Member since: Oct 20, 2013

Not sure how this can't be called cheating. The offending devices were even delisted from some benchmark websites because the benchmark owners felt they were cheating and damaging the integrity of the benchmark. Kudos to Anandtech for calling Shamesung out on this cheating.

12. tedkord

Posts: 17417; Member since: Jun 17, 2009

Here's why. I have an AOSP rom on my Samsung device, so I don't have the benchmark app that Samsung included. I sideloaded an app that overlays both core speeds in the screen at all times then fired up Quadrant and AnTuTu. Guess what? Both cores pegged at top speed the whole run. A good benchmark is supposed to stress your device. If it doesn't, it's useless.

10. pwnarena

Posts: 1129; Member since: Feb 15, 2013

PA, yes it's called cheating because those full performances could not be achieved in real world usage. they are only activating the full potential of the devices when there is benchmark testing. if they are confident with the performance of their devices, they shouldn't be throttling them. what are they worried about? heating issues? instability?

13. tedkord

Posts: 17417; Member since: Jun 17, 2009

Yes, those full performances can be achieved in real world usage. They didn't overclock the CPU, the didn't boost the bus speed or access hardware that is normally dormant. A benchmark is supposed to stress your device. If your device isn't running at full tilt during one, the bench isn't trelling you anything useful.

22. joey_sfb

Posts: 6794; Member since: Mar 29, 2012

These mobile crowd never build their own DIY x86 PC before. Benchmark is supposed to stress the system to ensure the system is stable under all conditions and prolonged period. Anyway, every brands do it not only Samsung but the hater are ready to pound on anything Samsung. Quite cute actually because in the end Samsung brand sell and they remain bitter. Lol!

11. MarkArigayo

Posts: 240; Member since: Dec 18, 2013

OMG, CheatWiz inside.

16. asingha94

Posts: 295; Member since: Apr 08, 2012

lol!!still inside!!they just find a better way!!

15. kanagadeepan

Posts: 1271; Member since: Jan 24, 2012

So in 4.3 or below, if we keep the Benchmark app running / waiting in BackGround, then the phone's battery will die very very fast?? Isn't it? Well, whether 4.4 stops this mischief or Sammy got bored of cheating??? Credit to Google or Sammy?

19. Tsepz_GP

Posts: 1176; Member since: Apr 12, 2012

The funny part is that my GS4 Octa on 4.4.2 scored more than it did in 4.3, I now get +28800, that's around the same as what I got in 4.2.2 before the benchmark boosting was exposed, in AnTuTu.

21. google-apple

Posts: 34; Member since: Jul 19, 2013

Sammy sammy sammy

23. chrislars

Posts: 77; Member since: Apr 11, 2013

will the note 2 get the kitkat update?

24. mike2959

Posts: 696; Member since: Oct 08, 2011

The real deal here is Kit Kat. On every device before and after the update, synthetic benchmarks are lower.

27. galanoth07

Posts: 44; Member since: Jan 23, 2014

they shoud put more details when it comes to benchmarkings like benchmarking a CPU/GPU.

Latest Stories

This copy is for your personal, non-commercial use only. You can order presentation-ready copies for distribution to your colleagues, clients or customers at https://www.parsintl.com/phonearena or use the Reprints & Permissions tool that appears at the bottom of each web page. Visit https://www.parsintl.com/ for samples and additional information.