x PhoneArena is hiring! Reviewer in the USA
  • Hidden picShow menu
  • Home
  • News
  • Samsung sues NVIDIA for patent infringement, claims benchmark test results were fake

Samsung sues NVIDIA for patent infringement, claims benchmark test results were fake

Posted: , by Alan F.

Tags :

Samsung sues NVIDIA for patent infringement, claims benchmark test results were fake
After being sued by NVIDIA for allegedly infringing on its patents, Samsung has filed a countersuit. The suit not only accuses the silicon slinger of using six Samsung patents without licensing them, it also accuses NVIDIA of misleading consumers by revealing false benchmark figures for its Shield Tablet in comparison with the Samsung Galaxy Note 4.

Samsung claims that NVIDIA is trying to show that the Tegra K1 processor in the Shield Tablet outperforms the Samsung Exynos 5433 processor, used in some models of the Samsung Galaxy Note 4. Also named as a defendant is a PC manufacturer named Velocity Micro. The company uses NVIDIA graphics cards in its products, which is why Samsung is claiming that the company infringed on two of its patents.

NVIDIA denies that it played fast and loose with the benchmark scores for the Shield Tablet, and says that it ran the tests comparing the Samsung Galaxy Note 4 and the Shield Tablet fair and square. Both devices, according to the chip maker, were tested right out of the box. You might remember that last October it was Samsung that was accused of propping up benchmark tests results for the Samsung Galaxy Note 3.

The legal action is just warming up here. Stay tuned.

NVIDIA's benchmark test compares its Tegra K1 processor to the Samsung Exynos 5433

NVIDIA's benchmark test compares its Tegra K1 processor to the Samsung Exynos 5433


source: Law360 (Registration Reqd.) via Engadget

47 Comments
  • Options
    Close





posted on 12 Nov 2014, 00:22 8

1. Sauce2 (banned) (Posts: 199; Member since: 21 Oct 2014)


/facepalm Samsung. Samsung is one to speak lol?

posted on 12 Nov 2014, 00:52 3

8. microsoftnokiawin (Posts: 1224; Member since: 30 Mar 2012)


ahaha I know right of everyone boosts their benchmark results then it's fair game I believe

posted on 12 Nov 2014, 01:05 5

15. vincelongman (Posts: 4877; Member since: 10 Feb 2013)


Exactly, everyone cheats benchmarks to some degree, including Intel, AMD, Qualcomm, Apple, PowerVR and ARM as well

Next year the 810, A9X, M1, Exynos 7/8 will destroy this year's 805, A8X, K1 and Exynos 5433
And the real world performance difference will still be relatively small again

posted on 12 Nov 2014, 01:48 5

19. AlikMalix (Posts: 7101; Member since: 16 Jul 2014)


Never saw an article accusing Apple of cheating on benchmark tests.

posted on 12 Nov 2014, 02:48 2

21. frederick333123 (Posts: 76; Member since: 15 Jun 2013)


yes i agree with you alikmalix

posted on 12 Nov 2014, 03:15 5

23. vincelongman (Posts: 4877; Member since: 10 Feb 2013)


They all optimize for better benchmarks, which to some degree is cheating

Apple examples
Browser benchmarks Safari always gets far results, there are some browser on Android that also get far results e.g. Puffin
Also their 64-bit GeekBench results are 40-50% higher than their 32-bit results, which is strange since the difference for Android is 10%.
The 64-bit Android wasn't released, so Google/Nvidia couldn't optimize for.
I wouldn't be surprised if Google/Nvidia pushed an update which boasts their 64-bit result by another 40%.

Other examples
In 3DMark Qualcomm and Nvidia (32-bit K1) have far better physics (CPU) results than Apple's A8X and the Nexus 9 (64-bit K1), which is strange since the A8X and 64-bit K1 are better in other CPU benchmarks

They all cheat to some degree

posted on 12 Nov 2014, 03:37 2

27. vincelongman (Posts: 4877; Member since: 10 Feb 2013)


Here's the Geekbench results I was talking about

iPad Air 2 64-bit = 1800/4500
iPad Air 2 32-bit = 1300/3200

6 Plus 64-bit = 1600/2900
6 Plus 32-bit = 1200/2100

Nexus 9 64-bit = 2000/3600
Nexus 9 32-bit = 1900/3500

Google/Nvidia couldn't cheat since there was only the 32-bit version of Geekbench before the Nexus 9 was released.
The next Nexus 9 update will probably have some optimisations which will also boast their result more like Apples.
And Samsung will push it to their Exynos Note 4 as well

Here the links for the 32-bit Apple results since they are quite hard to find
iPad Air 2 32-bit (only one, since everyone does the 64-bit test)
http://browser.primatelabs.com/geekbench3/1180262
6 Plus 32-bit (there's more, since its been out for longer)
http://browser.primatelabs.com/geekbench3/1190884
http://browser.primatelabs.com/geekbench3/1185356
http://browser.primatelabs.com/geekbench3/1179553

posted on 12 Nov 2014, 03:56 2

28. rantao333 (Posts: 322; Member since: 21 May 2013)


u call that cheat?

cheat means manipulated frequency, overriding heat-protection, and does not run the same in real life. and these are wat most OEM done during the year with evidence and prove.

u cant generalize when nexus 9 show little different between 32bit & 64 bit. then apple must be cheat on benchmarks since there are differences within the 32bit & 64bit.

posted on 12 Nov 2014, 04:18

29. Awalker (Posts: 1654; Member since: 15 Aug 2013)


It's a little bit different here. Samsung doesn't put out any benchmark numbers and claim their devices are faster than the competition. Reviewers run the benchmarks knowing that these tests can be manipulated and therefore shouldn't be a part of their review process but they still do it anyway.

posted on 12 Nov 2014, 05:06 2

32. aba71983 (Posts: 805; Member since: 24 Jul 2014)


Apple's know what is best to their product. It may have Samsung chips inside of there iPhones and iPads but they never cheating on benchmark tests.

posted on 12 Nov 2014, 05:41

34. vincelongman (Posts: 4877; Member since: 10 Feb 2013)


That's why I said to some degree

Another easy way to tell is just look at the Android Geekbench scores, all the 801 phones have very different scores even though they have the same chip

Optimizing for a specific benchmark also has no benefit in real life

The Nexus 5 has a 800 with generally lower benchmark scores, and still faster than almost all the other flagships

posted on 12 Nov 2014, 12:27 1

38. AlikMalix (Posts: 7101; Member since: 16 Jul 2014)


All the "proof" you though you posted are actual results and your own speculation that if Apple does much better with their hardware then they're cheating. The only reason why Apple gets better margins vs their 32bit counterparts is because Apple builds their own hardware and software - this is pure apromization in its best - android phones cannot do that because they use freeware for their OS - read: not optimized.

I have countless and countless articicles showing android phones deliberately cheating (manipulating clock speeds and bypassing certain limits that may otherwise damage your phone if left in real day-to-day use).

Just because Apple has impressive benchmark results with every new device that you cannot believe they're achieving with half the specs of your beloved android - does not mean they're Cheating! Did you ever think that maybe android manufacurers are not that good at hardware/software relationship - that they just throw as much hardware at it as they can instead of doing proper coding?

And don't give me that BS that iPhones don't do anything but boot up and display icons - I devices and iOS does everything you can do on android. Just because you have a file manager - doesn't mean I cannot store files on my devices or open them or work with them. Don't be ignorant.

posted on 12 Nov 2014, 14:25

42. vincelongman (Posts: 4877; Member since: 10 Feb 2013)


Calm down bro
Nothing personal

It's all proof Apple are optimizing for benchmarks
Apple are just smarter a cheating

Nvidia couldn't cheat since their was no 64-bit version of Geekbench on Android before the Nexus 9

They all cheat, in the update for the Nexus 9 Nvidia will also throw optimizations as well
Also Samsung will also as well

posted on 12 Nov 2014, 18:09

46. elitewolverine (Posts: 5192; Member since: 28 Oct 2013)


they are not optimizing for benchmarks, the reason is because the OS is actual 64bit. When you get a real 64bit release on android, expect there to be a bump.

posted on 12 Nov 2014, 18:45

47. vincelongman (Posts: 4877; Member since: 10 Feb 2013)


40-50% is far too high
The difference on Windows, Mac OS X and Android is 10%
So something is up

posted on 12 Nov 2014, 00:22 11

2. Scott93274 (Posts: 5217; Member since: 06 Aug 2013)


Any article with the words "Samsung" & "Benchmarks" result in a headache for me. It reminds me that they'll do anything to sell their phones, even lie & cheat.

posted on 12 Nov 2014, 00:29 14

5. shahrooz (Posts: 784; Member since: 17 Sep 2013)


welcome to earth, what you just described is called corporate business in 2014 :) you think other companies like Apple, LG etc are honest to you? wake up brother

posted on 12 Nov 2014, 00:39 4

7. Scott93274 (Posts: 5217; Member since: 06 Aug 2013)


Hey, I am fully awake. I know all other companies pull this crap too, especially Apple, but I call Apple on their BS rather frequently, I don't do it nearly as regularly for Samsung. So take a chill pill and let me vent.

posted on 12 Nov 2014, 00:54 7

10. hung2900 (Posts: 961; Member since: 02 Mar 2012)


Samsung is the first one of big "lying" tech companies stoped cheating. One M8 still cheats.

For years and years, nVIDIA's comparison charts are garbages, many people know that.

posted on 12 Nov 2014, 00:58 5

12. tech2 (Posts: 3487; Member since: 26 Oct 2012)


After their faking benchmark fiasco Samsung was the only one who fixed that with an update while other OEMs who were also caught of doing the same haven't yet fixed anything and continue to cheat in benchmarks.

posted on 12 Nov 2014, 01:04 2

14. Scott93274 (Posts: 5217; Member since: 06 Aug 2013)


Maybe, I don't know, I haven't really followed it. But I know when it went down, Samsung was always the one that the media pointed out that was the main offender. Maybe they're a little more honest now... but according to articles I read, Motorola & Apple were the only two companies that were not guilty of the dishonest practice. I do know that Samsung isn't alone, but like I said in my previous statement, I see those two words and I get headaches because of all the crap that came to light...

http://www.fiercecio.com/story/not-so-smartphone-makers-caught-cheating-benchmarking-tests/2013-10-07

posted on 12 Nov 2014, 01:12 1

16. vincelongman (Posts: 4877; Member since: 10 Feb 2013)


To some degree everyone are still cheating benchmarks including Intel, AMD, Qualcomm, Apple, PowerVR and ARM as well

E.g. the 805, A8, K1, Exynos 5433 perform far better than last year's 800, A7, Tegra 4 and Exynos 5420 in benchmarks
Yet the real world performance difference is still be relatively small

And it will happen again next year, and the year after

posted on 12 Nov 2014, 13:50

41. AlikMalix (Posts: 7101; Member since: 16 Jul 2014)


Dude you've been told again and again - apples benchmarks are that of actual clock speed and cpu restrictions that are used to run everything on the device - NOT just benchmark software, unlike other android devices. Apple may have their share of wrongdoings but cheating on benchmarks is not one of them. They deliver an honest product (some problems come up, yes, but repaired quickly 90% of the time).

So for everyone trying to drag Apple into the pool of benchmark cheating - stop - you're lying to others and yourself.

posted on 12 Nov 2014, 14:39

44. Scott93274 (Posts: 5217; Member since: 06 Aug 2013)


Hey, I just want to point out that I talk a lot of crap about Apple, but I gave them credit in my last comment. I'm honestly offended by many of the Android OEMs for this crap.

posted on 12 Nov 2014, 16:04 1

45. AlikMalix (Posts: 7101; Member since: 16 Jul 2014)


My post wasn't toward you, it was response to post #16. But yeah I was surprised by you on that post. Thanks.

posted on 12 Nov 2014, 00:26 5

3. e8ght (Posts: 21; Member since: 25 May 2013)


The Pot Calling the Kettle Black.

posted on 12 Nov 2014, 00:27

4. maple_mak (Posts: 766; Member since: 18 Dec 2013)


Are they fooling different about 32-bit and 64-bit performance?

posted on 12 Nov 2014, 00:30

6. joey_sfb (Posts: 6402; Member since: 29 Mar 2012)


Lie and cheat always has a place since the beginning of benchmark.

Intel vs AMD during 386/486 days

Nvidia vs ATI since the beginning of 3D graphic.

My guide for my PC graphic purchase has always been my budget, fav 3D game fps and gpu operating tempurature.

posted on 12 Nov 2014, 06:44

36. bucky (Posts: 2982; Member since: 30 Sep 2009)


Likewise for the most part. I'd be lying if I said I didn't favor nvidia.

posted on 12 Nov 2014, 00:53 2

9. blazee (Posts: 414; Member since: 02 Jan 2012)


Can we sue Samsung for cheating on their benchmarks then?

Want to comment? Please login or register.

Latest stories