x PhoneArena is hiring! Reviewer in the USA
  • Hidden picShow menu
  • Home
  • News
  • Report: U.S. and China bound Samsung Galaxy S7 models to carry Snapdragon 820 chipset

Report: U.S. and China bound Samsung Galaxy S7 models to carry Snapdragon 820 chipset

Posted: , by Alan F.

Tags:

Report: U.S. and China bound Samsung Galaxy S7 models to carry Snapdragon 820 chipset
A report out of  China released Friday morning (while it was still Thursday in the U.S.), reveals that versions of the Samsung Galaxy S7 heading to the U.S. and China next year, will be powered by the Qualcomm Snapdragon 820 chipset. Everywhere else, Sammy's new flagship phone will be powered by a Samsung produced Exynos SoC. Breaking with tradition, Samsung employed its Exynos 7420 globally on the Samsung Galaxy S6 this year. But this was not done to start a new tradition. Samsung was just concerned about reports that the Snapdragon 810 was prone to overheating.

The Snapdragon 820 features Qualcomm's proprietary Kryo cores, and the quad-core CPU is partnered with the Adreno 530 CPU. Qualcomm has reportedly done a great job at enhancing its newest high-end chip to cut back on power requirements and offer improved thermal management. The Samsung Galaxy S7 will be released in February, according to the latest rumors.

Samsung has been testing different versions of the phone, sometimes using the Snapdragon 820 inside, and other times loading the device with the Exynos 8890. The latter could be the next Exynos chip that has been dubbed the M1, or Mongoose. There are still rumors that we will see a 5.2-inch model and a 5.8-inch variant of the Galaxy S7 next year, along with a model sporting 3GB of RAM and one with 4GB of the stuff.  When it comes to the Samsung Galaxy S7, we are still in flux.

source: Reuters

68 Comments
  • Options
    Close




posted on 01 Oct 2015, 22:35 14

1. zeeBomb (Posts: 1837; Member since: 14 Aug 2014)


Whateva. Exynos 8890 ftw

posted on 01 Oct 2015, 22:36 5

2. PapaSmurf (Posts: 10439; Member since: 14 May 2012)


Qualcomm better deliver this year is all I'm saying.

posted on 01 Oct 2015, 23:05 4

7. roldefol (Posts: 4097; Member since: 28 Jan 2011)


They have to. And Samsung isn't stupid, they wouldn't sell an S7 anywhere with an 820 if they didn't trust it.

posted on 02 Oct 2015, 01:25 2

21. iushnt (Posts: 1559; Member since: 06 Feb 2013)


Agreed

posted on 01 Oct 2015, 22:48 1

3. Jerry7 (Posts: 7; Member since: 01 Oct 2015)


Hopefully it will not overheat.

posted on 01 Oct 2015, 22:50

4. Cheezwiz (Posts: 430; Member since: 29 Dec 2014)


Still February then?

posted on 01 Oct 2015, 22:54 5

5. arch_angel (Posts: 1651; Member since: 20 Feb 2015)


I'm Fine With It The Sd 820 Should Be Good And Have An Amazing Gpu.

posted on 01 Oct 2015, 23:11 1

8. roldefol (Posts: 4097; Member since: 28 Jan 2011)


I'm inclined to wait for the 820, but it'll be another year before it shows up in a Moto or Nexus. Maybe LG does something damned impressive with the G5,

posted on 02 Oct 2015, 09:59

53. Ruturaj (Posts: 1294; Member since: 16 Oct 2014)


I want to see how 530 does against T860/T880 MP6/8 and apples GPU.

posted on 02 Oct 2015, 17:02

58. arch_angel (Posts: 1651; Member since: 20 Feb 2015)


Same.

posted on 01 Oct 2015, 22:56 8

6. maddy996 (Posts: 81; Member since: 06 Jan 2014)


Now I'm more interested to see Exynos M1 rather SD 820...

posted on 01 Oct 2015, 23:14 3

9. Dude2014 (Posts: 436; Member since: 12 Feb 2014)


BTW is that Ultrasonic fingerprint technology on Snapdragon 820 real? If I'm not mistaken, they said that technology doesn't require any physical buttons for fingerprint reading unlike any other tech on the market.. That would be game changer in industry indeed. ;O

posted on 01 Oct 2015, 23:17 6

10. zeeBomb (Posts: 1837; Member since: 14 Aug 2014)


Force Touch 2.0: The force awakens

posted on 01 Oct 2015, 23:24

11. dimas (Posts: 1967; Member since: 22 Jul 2014)


Going to be released on Feb 2016? Now that's fast! Samsung really want to get out of their losing profit streaks for months.

posted on 01 Oct 2015, 23:34 1

12. penguinbelly (Posts: 143; Member since: 27 Dec 2014)


They should really go down a notch in display size to 5.0".

posted on 01 Oct 2015, 23:41 3

13. roldefol (Posts: 4097; Member since: 28 Jan 2011)


You aren't from around here are you?

posted on 01 Oct 2015, 23:46 5

14. Mr.Mr.Upgrade (Posts: 474; Member since: 30 Aug 2011)


Get rid of half baked touch whiz and use purer Android software

posted on 02 Oct 2015, 01:30 1

23. rd_nest (Posts: 1591; Member since: 06 Jun 2010)


Yea, long overdue. Google also mentioned pure android can also make coffee.

posted on 02 Oct 2015, 11:37

55. alex3run (Posts: 710; Member since: 18 May 2014)


Get rid of half-baked pure Android UI and give touchwiz.

posted on 02 Oct 2015, 00:08 3

15. Carl3000 (Posts: 153; Member since: 11 Oct 2014)


This is probably the best news Qualcomm has had all year.

posted on 02 Oct 2015, 00:13

16. Taters (banned) (Posts: 6474; Member since: 28 Jan 2013)


Hmmm, so Canada is getting the Exynos?

posted on 02 Oct 2015, 00:35 1

18. zeeBomb (Posts: 1837; Member since: 14 Aug 2014)


I would cry me a Noah's Ark if it were to happen.

posted on 02 Oct 2015, 00:21 4

17. TyrionLannister (unregistered)


We all know they are going to have amazing CPUs. What I'm concerned about is GPU. They should be at least as powerful as the ones on the iPhone 6S. And that's pretty hard to do. As the GPU on 6s is about 50-60% more powerful than the Mali t760 mp8 and 60-70% more than adreno 430.

We can already see that given the numbers given by QualComm, the 530 will be weaker than the GPU in A9, which is a shame. Let's home at least the new Mali beats A9 GPU like the t760 thtashed the A8 Gpu.

posted on 02 Oct 2015, 01:07 4

19. johanbiff (Posts: 388; Member since: 31 Mar 2015)


im hoping for Mali T880-mp8 or mp10 for the exynos monster..that will give Apple a run for it`s money

posted on 02 Oct 2015, 01:30

24. 47AlphaTango (Posts: 284; Member since: 27 Sep 2015)


But Mali isn't built for gaming. The only accepted hardware for gaming on Android is a snapdragon or a tetra processor. This is why the s6 and s6 edge didn't sold well. Same goes for the note 5.

posted on 02 Oct 2015, 02:13 4

29. TylerGrunter (Posts: 1514; Member since: 16 Feb 2012)


Where did you get that from?
As far as I know not only the all games work fine in the S6, it also delivers better FPS than any other phone in the market.

posted on 02 Oct 2015, 11:38

56. alex3run (Posts: 710; Member since: 18 May 2014)


Sounds funny because Adreno historically is a bit laggy GPU in games.

Tegra before K1 wasn't something special too.

posted on 02 Oct 2015, 01:19

20. alex3run (Posts: 710; Member since: 18 May 2014)


Don't think that Adreno 530 will be sliwer than PowerVR GT7600. If it will really be 40% faster than Adreno 430 then 530 will reach A9.
As for Exynos I predict an MP12 config of Mali GPU (that's equal to 6-core GP7600) put with higher ALU per core ratio (that's a T880) and maybe higher clocks.

posted on 02 Oct 2015, 01:29 2

22. TyrionLannister (unregistered)


530 is 40% more powerful than 430. GT7600 is 70% more powerful than 430. Do the math.

Samsung won't use the MP12 variant of T880. And it's unnecessary. Even the MP4 T880 is pretty close in power as T760 MP8. A more efficient process + T880 MP8 is all they need to match the T7600.

posted on 02 Oct 2015, 01:33

25. 47AlphaTango (Posts: 284; Member since: 27 Sep 2015)


But is the Mali built for gaming? Mali only beat the adreno in terms of benchmarks. But what happens if the test was app based?

posted on 02 Oct 2015, 01:42 2

27. johanbiff (Posts: 388; Member since: 31 Mar 2015)


Mali have better overall FPS if im not mistaken..

posted on 02 Oct 2015, 02:17 2

31. TylerGrunter (Posts: 1514; Member since: 16 Feb 2012)


Can you stop with that silly thing?
The Mali 760MP8 in the S6 has been beating consitently the Adreno 530 IN GAMES.
https://www.gamebench.net/en/blog/lg-g4-gaming-performance-ranked-against-other-android-and-ios-smartphones

posted on 02 Oct 2015, 02:33 6

35. TyrionLannister (unregistered)


The real shame of android is that most games are software capped at 30 fps. That's why iPhones pull ahead. You can clearly see in the charts how S6 constantly walls against 30 fps.

posted on 02 Oct 2015, 02:59 2

36. TylerGrunter (Posts: 1514; Member since: 16 Feb 2012)


I know, I had to explain that to someone in a forum.
The iphone 6 plus pulls ahead in the average due to the cap, but the S6 beats it pretty much in any other game.
Sad that the guys fro gamebench are able to create such a good benchmark tool and don´t know how to interpret the results later.

posted on 02 Oct 2015, 10:26 1

54. sharifsakr (Posts: 4; Member since: 02 Oct 2015)


Hi TylerGrunter, while I agree that frame rate caps affect GameBench comparisons, I can't really agree that these caps should somehow be ignored or taken out of the equation.

After all, a frame rate cap is a real thing that is experienced by the user, and it is often put in place for a good reason. E.g., Dead Trigger 2 is capped on the GS6 by default because the developer encountered overheating problems on the device when it was allowed to run above 30fps. There's no such issue on the iPhone 6, which is allowed to run at 60fps by default.

You mentioned that the GS6 beats the iPhone 6 "pretty much in any other game" than one that is capped. I can't agree with this either: we chose a sample of ten games, based primarily on their cross-platform popularity, and the iPhone 6 showed superior frame rates to the GS6 in six of these ten titles.

Finally, it's worth remembering that frame rate caps can hurt the iPhone 6 too. Real Racing 3 is capped at 30fps on the iPhone 6, for example, but not on the GS6. Same goes for Kill Shot. Why would we ignore a cap when it hurts the GS6, but not when it hurts the iPhone 6? Safer just to do what we do: pick a sample of popular games and see how well they run.

posted on 02 Oct 2015, 11:55

57. alex3run (Posts: 710; Member since: 18 May 2014)


Framerate cap isn't related to overheating. It theoretically should stablilze fps.

posted on 03 Oct 2015, 03:50 1

61. TylerGrunter (Posts: 1514; Member since: 16 Feb 2012)


I think you have to reread my post, I NEVER sair that the GS6 beats iPhone 6. I said that the GS6 beats the iPhone 6 PLUS in pretty much every other game.
You are very well aware of the low resolution that the iPhone 6 has and how that favours FPS in the games, so the iPHone 6 plus makes a much better comparison.
I can´t confirm I´m right because the link to your blog doesn´t show the results any more (No idea why, I tried with 3 different browsers)
https://www.gamebench.net/en/blog/lg-g4-gaming-performance-ranked-against-other-android-and-ios-smartphones

And I also never said you should not count when the iphone is capped, all I said is that when a cap happens you should take the results out of the average and not count them as it is not fair, does not matter to which phone.
There is a world of difference between ignoring the rate caps and just not include them in a fake average, just leave them in the raw data, but not int the average.
In your posts you should also mention when caps are happening, something you also didn´t do (and that´s why I didn´t know that the iphone was capped too).
All I´m saying is that you need to be a lot more careful about how you are interpreting the results. You may disagree, but I honestly think the way you are doing it is far from optimal.

posted on 03 Oct 2015, 07:22

62. sharifsakr (Posts: 4; Member since: 02 Oct 2015)


Apologies, I thought you were referring to the iPhone 6. Our dynamic charts are back up now -- sorry about that, too. Your link should work.

With regards to the iPhone 6 Plus, it delivers a higher frame rate than the GS6 in five of our ten sample games, and draws with it on one game. So it's still in the lead, if you judge it that way, but the lead is obviously much, much narrower than the iPhone 6's.

We've always been clear about this in our analysis of the iPhone 6 Plus results, so I'm not sure why you think our interpretation is misguided. We're not beholden to Apple in any way, and we've been pretty blunt about the fact that the company's first 1080p smartphone struggles in a number of games, especially when you look at minimum frame rates. Please see our blog post dated 30th June (can't post URL yet, as I only just registered.)

Besides, anyone looking at our overall charts (i.e. the average across the sample of ten games) can readily see that 6 Plus's median FPS lead is slight, and that it's behind on minimum FPS.

posted on 04 Oct 2015, 03:21

63. TylerGrunter (Posts: 1514; Member since: 16 Feb 2012)


Let´s see what you say because I don´t get your five wins for the iphone 6:
GS6 wins: Asphalt 8, GTA, Marvel
Iphone 6 + wins: Boom Beach, Monument Valley
Draws: XCOM
Capped on iphone: Real Racing 3, Kill shot
Capped on GS6: Dead Trigger 2, Modern Combat 5

I would say GS6 has 3 wins and a draw against 2 loses, making it the better phone for gaming. Add to that much better minimum FPS and battery life while gaming.
But in your conclusions you declare the iphone 6 plus better and your graph still shows hefty adavantage for the iphone 6 plus, which is misleading in my opinion.
Most people don´t understand how important is the Minimun FPS and therefore missinterpret the graph.
And that is all my point, you should have taken out of the average the capped ones and make clearer when games are capped in the other graphs.

posted on 05 Oct 2015, 05:57

65. sharifsakr (Posts: 4; Member since: 02 Oct 2015)


We base our rankings on the median FPS, not based on the minimum FPS. This is because we feel the median is the metric most reflective of the overall user experience. Minimum FPS can theoretically be a single stutter during a 15-minute gameplay session that otherwise played at a very steady and very high frame rate, so we don't use it as a ranking metric (although we don't ignore it in our analysis either, and neither do we ignore FPS stability as a separate metric).

Based on the median, the iPhone 6 Plus beats the GS6 at five games: DT2, MC5, Boom Beach, XCOM and Monument. It ties on one game: Asphalt 8. It loses on four games: RR3, Kill Shot, GTA and Marvel. So this is a narrow win for the iPhone 6 Plus.

Moreover, if you take the geometric mean of the median FPS across the basket of ten games, it also shows a narrow win for the iPhone 6 Plus, as reflected in our charts.

As discussed already, we don't agree with ruling out capped games, because caps are real and GameBench is a real-world benchmark.

posted on 05 Oct 2015, 06:38

66. TylerGrunter (Posts: 1514; Member since: 16 Feb 2012)


Then I can rest my point:
your analysis is shallow, as you don´t take into account both, but just the median.
In Asphalt 8 you give it a draw, when they have the same median, but he GS6 has far better minimum, making it an obvious win in my views.
In XCOM, the minimum for the 6+ is just horrible, making it a draw.
Caps may be real, but they don´t translate into worse expericence for the users (normally), as most of the users would not differenciate between a 30 FPS and a 45, and it skews the median greatly.
So in my honest opinion:
a) Your analysis is incomplete as you ignore the minimum FPS and base it ONLY in the median, and to make it worse you do it with the capped games too.
b) The comments are also incomplete as you don´t make it clear when caps are happening and add them to the median.
c) You take conclusions based on those incomplete analysis, like which phone is better than what, which are some times wrong due to how shallow tha analysis is.

Just take it as you wish, but I still think you are doing a great benchmark, but not so great analysis of the results.

posted on 06 Oct 2015, 09:15

67. sharifsakr (Posts: 4; Member since: 02 Oct 2015)


I think we've made some headway here, because I can appreciate your points better now. Equally, you've gone from describing our analysis as "sad" to saying that's it "shallow" and "not so great."

Yes, our ranking methodology could be described as crude or simplistic -- most (if not all) rankings are. But what's the alternative?

There's always scope to create a more complex overall score, which reflects all three FPS metrics (median, minimum, stability). But it quickly gets complicated, because you have to weight the contributions to this score differently. The median is arguably more important than minimum, for example, whereas a low stability can undermine the median.

Further adjusting the weighting of performance differences in capped games vs. uncapped games, or applying subjective weightings about what games can or cannot perceive, would only complicate things even further.

Pretty soon something that ought to be transparent becomes obscure. So, like it or not, we've decided to rank simply on the basis of median FPS, while always making the fuller picture clear by means of our charts.

posted on 06 Oct 2015, 10:41

68. TylerGrunter (Posts: 1514; Member since: 16 Feb 2012)


I agree that it is not easy and that adding complexity it may be make it more difficult for some readers.
The case for me is that I prefer more accurate information to simplicity, and that always means more complex analisys.

posted on 04 Oct 2015, 03:28

64. TylerGrunter (Posts: 1514; Member since: 16 Feb 2012)


By tha way, thanks for fixing the graphs.

posted on 02 Oct 2015, 02:17 2

32. TyrionLannister (unregistered)


Mali is way better in real world gaming and performs worse in benchmarks. The reason being it's simple design which works better with OpenGL.

posted on 02 Oct 2015, 08:48

48. alex3run (Posts: 710; Member since: 18 May 2014)


GT7600 is 40 - 45% more powerful than Adreno 430, at least in benchmarks. So I did the math. But did you while comparing PVR's and Adreno's framerates?

I thought that Samsung wouldn't use an MP12 comfig too. But according to some leaked data it scores >60/108 fps in GFXBench what means that Exynos M1 (8890) has Mali T880MP12 at about 750 - 800 MHz.

posted on 02 Oct 2015, 08:57 1

49. TyrionLannister (unregistered)


Gfxbench t-rex:

Adreno 430: 50 fps
Gt7600: 80 fps
That's 100*(80-50)/50=> 60% improvement.

Gfxbench manhattan :
Adreno 430: 24 fps
Gt 7600: 40 fps

That's 100*(40-24)/24 = 66.67% improvement.

Say that again about my math.

posted on 02 Oct 2015, 09:29

52. alex3run (Posts: 710; Member since: 18 May 2014)


And I will.

Adreno scores 28/57 fps at best, so that's 43 and 40% difference respectively.

posted on 02 Oct 2015, 01:55

28. rd_nest (Posts: 1591; Member since: 06 Jun 2010)


Here's the estimate and the latest information from Korea for 8890.

Mali T880 MP12 @ 700Mhz.
T-Rex Offscreen result 100+ FPS
Manhattan Offscreen result 45+ FPS

I could also see SLSI optiming for MP10 @ 850 MHz instead.

posted on 02 Oct 2015, 02:16 2

30. TyrionLannister (unregistered)


That's actually pretty low. A T760 MP8 scores about 25 FPS in manhattan. Considering that the T880 is 80% faster, A T880 MP8 should score about 25*1.8 = 45 fps. A T880 MP12 should be 50% faster and should score in the neighbourhood of 68 fps. And that's without considering any process improvements( 14 nm LPP should improve the scores a bit).

Maybe the chip isn't finalised yet, or the rumour is fake.

posted on 02 Oct 2015, 02:23 2

33. TylerGrunter (Posts: 1514; Member since: 16 Feb 2012)


In reality it is a bilt lower than what your calculations show.
This is due the Mali760 in the S6 is built in 14nm and not in 20nm, so it is clocked aroun 10% higher than what ARM uses as standard.
That also means that it the T880 won´t be 1.8 more powerful (as ARM uses a 16/14nm for the theoretical calculation), but more likely 1.6 or 1.7.
So Samsung will have to use a minimum of a MaliT880MP8 if they want to remain competitive against the iphone GPU. That will get them around 40 FPS in Manhattan like the iphone 6S does.

posted on 02 Oct 2015, 02:31 2

34. TyrionLannister (unregistered)


Still, that can't possibly explain a 45 fps manhattan with a T880 MP12. It's too low.

posted on 02 Oct 2015, 03:01 2

37. TylerGrunter (Posts: 1514; Member since: 16 Feb 2012)


That´s why I said "a bit" lower ;)
Apart from that little correction your calculations are pretty much on spot.

posted on 02 Oct 2015, 03:06 2

40. TylerGrunter (Posts: 1514; Member since: 16 Feb 2012)


May be his calculation is that off either.
He is referring to a MaliT880 @700Mhz.
MaliT880MP12 should be around 60FPS standard, and that is clocked at 850Mhz. So at 700Mhz it would score from 45 till 50 FPS.
The leak may not be that far from reality.
My bet is they will use a MaliT880MP10 at standard clock (850MHz) to be able to beat the A9 GPUs

posted on 02 Oct 2015, 04:13

42. rd_nest (Posts: 1591; Member since: 06 Jun 2010)


Well, we will get to know more details soon/

Also, the core scaling isn't totally linear. We could lose like 5% of FPS with every core scaling.

posted on 02 Oct 2015, 09:02

51. alex3run (Posts: 710; Member since: 18 May 2014)


You should dig deeper into arch differences between Mali Tx60 and Tx80.
I'll try to explain.
At first Tx80 has 50% more per-core ALUs (that will strongly reflect on Manhattan score but not so much on T Rex score).
Then we have different clocks (that depends on Samsung's decision and manufacturing process). I think the frequency will go up to 850 MHz in some scenarios.
And finally we have 12 cores against 8 in the past - the performance will scale linearly.

All in all we have up to 2,5x T760MP8 performance.
But clock-for-clock and per core T880 is 1 - 1,5x Mali T760.

posted on 02 Oct 2015, 01:37 1

26. johanbiff (Posts: 388; Member since: 31 Mar 2015)


If ARM is correct with their info regarding T880, it should be enough with T880-mp8 with same clock speed as t760 in exynos 7420 to outperform the GT7600..it shloud be 80% over the mali 760mp8 (1.8x faster then T760)

posted on 02 Oct 2015, 03:01

38. Quduz (Posts: 82; Member since: 18 Aug 2015)


you're wrong it score 56.9fps not 45 fps Google it

posted on 02 Oct 2015, 03:02

39. Quduz (Posts: 82; Member since: 18 Aug 2015)


you're wrong it score T rex 108.9 and mahattan56.9fps not 45 fps Google it

posted on 02 Oct 2015, 04:11

41. TyrionLannister (unregistered)


Can you please provide the link. I Googled it and didn't find anything. Or tell me the exact string to Google.

Want to comment? Please login or register.

Latest stories