Qualcomm Snapdragon 820 benchmarks surface again, clashes with the A9 in iPhone 6s Plus

Qualcomm Snapdragon 820 benchmarks surface again, clashes with the A9 in iPhone 6s Plus
Qualcomm's Snapdragon 820 is probably among the most anticipated pieces of smartphone tech this year. Not only is it the SoC that is expected to go into most flagship handsets, it also bears the mission of removing the blotch of bad rep that the previous model — the Snapdragon 810 — smeared on top of Qualcomm's resume with its overheating issues.

The good news is that Qualcomm is getting help from Samsung in constructing the new Snapdragon 820. And Samsung showed us that it knows what it's doing with its home-brewed Exynos 7420 SoC – the silicone, which powered the Galaxy S6, Galaxy S6 edge, Galaxy S6 edge+, and Galaxy Note 5. These phones were certainly among the top Android performers in 2015, so the tech world is hopeful that, with Sammy's assistance, the Snapdragon 820 will outdo the competitors and redeem its line's name in 2016. Thus far, leaked AnTuTu and Geekbench benchmarks show that it just may!

A new set benchmarks has surfaced on GFXBench's website, allegedly showing us the upcoming chip's graphics performance prowess — and it looks impressive. The supposed Snapdragon 820 device easily beats the Note 5's score on the same website, and competes with the iPhone 6s Plus, beating it in some of the offscreen tests, but ultimately losing to it in the onscreen ones, which shouldn't be viewed as a big disappointment, considering the amount of pixels each device is pushing.

From the screenshots, you will notice that the SD820 was installed in a 6.2” smartphone with a 1600 x 2560 resolution, 4 GB of RAM, a 20 MP / 12 MP camera combo, and Android 6.0 on board. You may be wondering what device this is, but at the moment, we wouldn't read too much into it, as it could just be a test platform for the processor, not a production unit. However, do note the test device's (rather unorthodox) resolution, and compare it to the iPhone 6s Plus' 1080 x 1920. Indeed, the alleged Snapdragon has more pixels to push, but it still performed quite respectably in the tests.

The first smartphone to be powered by a Snapdragon 820 has already been announced. It's going to be made by a company, which is quite unknown in the west — LeTV — and the handset's name is Le Max Pro. However, it's still not out on the store shelves. An earlier rumor claimed that Samsung will be the first OEM to actually physically sell a Snapdragon 820-equipped phone, and there may still be some truth to those reports. Sammy is about to unveil its Galaxy S7 flagships in just a couple of weeks, and they will probably start shipping within 2 weeks afterwards. It would make sense if, seeing as Samsung is heavily involved in the Snapdragon 820 production, it has made a deal with Qualcomm to be the first manufacturer to launch a phone with the chip, effectively slowing down the competition a bit.



1. apakalo1

Posts: 67; Member since: Nov 19, 2015

And so what? You write articles in every seconds. Unprofessionalism is working,,,

2. Shocky unregistered

Offscreen tests look great, it's faster than the A9 Very impressive.

21. Well-Manicured-Man

Posts: 703; Member since: Jun 16, 2015

I don't that the values are impressive at all since the 2016 SD820 is on average only delivering about 11% more frames per second off screen than the 2015 Apple A9. Everything else but impressive. Nevertheless, the SD820 shows an excellent performance.

24. Well-Manicured-Man

Posts: 703; Member since: Jun 16, 2015

Its actually 14,5% more frames per second off screen. I mixed the lines up.

26. Finalflash

Posts: 4063; Member since: Jul 23, 2013

It is about 20% faster in Manhattan test which is the more taxing of the two. So a 10-20% improvement on the iPhone chipset is acceptable seeing as how the iPhone is only 4-5 months old at this point. The next iPhone will be about 50% faster than the previous iPhone and likely about 20% faster than the 820 and so on. This is how it goes and this is how its been.

52. tedkord

Posts: 17414; Member since: Jun 17, 2009

You mean the late 2015 A9 and the beginning of the year 2016 820? It's only been about 4 months since the A9 released.

62. Well-Manicured-Man

Posts: 703; Member since: Jun 16, 2015

Ted, this is not about age but about available technology at the time of release. SD820 is build in 2nd gen 14nm with true FINFET. Apple A9 is build in 16nm with true FINFET because true FINFET was not available for 14nm in time for the iPhone 6s. In other words: SD820 is a 2016 technology SOC while Apple A9 is a 2015 technology SOC. Apple A10 will be Apple's 2016 technology SOC.

65. tedkord

Posts: 17414; Member since: Jun 17, 2009

That's not true, either. It's about performance. Who cares what tech is used, both give top of the market performance.

67. vuyonc

Posts: 1091; Member since: Feb 24, 2014

I thought Samsung's 14nm FinFET LPP still uses 20nm BEOL. 'Better' FinFET node but not 'true' FinFET node?

69. vincelongman

Posts: 5724; Member since: Feb 10, 2013

AFAIK that's right And TSMC's new 16FF+ is 20nm FEOL and BEOL Samsung's 14LPP and TSMC's 16FF both have taller fins, which means less leakage (and therefore better performance) compared to their 14LPE and 16FF processes


Posts: 1459; Member since: Mar 09, 2010

Maybe you missed the part where it said that the test equipment was 6.2"s..... That's impressive enough.

30. mistercarter

Posts: 360; Member since: Sep 01, 2011

the 820 is the flagship chipset for 2016 handsets and still lags behind Apple's flagship for 2015, just imagine what the A10 will do to that poor 820SoT...

37. TyrionLannister unregistered

"still lags behind Apple's flagship for 2015" Are you literally blind? 820 beats the A9 by significant margin on both manhattan as well as T-rex.

39. Piyath_ale

Posts: 79; Member since: Nov 02, 2015

WHAT? R U BLIND? Who cares about off screen test? Just bragging, pathetic, stupid people like you.. Wise men like US only consider on screen tests as true benchmarks. " by significant margin" hih hih heeeeeeeeeeeeee my ass

47. TyrionLannister unregistered

I can give various reasons: 1. We are comparing SoCs here, not devices. The same SD 820 can be used in a 480p phone and will kick the ass of any chip in on-screen benchmarks. That won't make it a better chip, self-proclaimed wise guy. For comparing SoCs, off-screen benchmarks is all that matters. 2. You can change resolution in games anyway. So I can run a game at 720p on my 1440p phone, and can get more performance out of it. So on-screen benchmarks don't matter even in real life. This option requires no root. Most high-end games have it inbuilt. 3. Maybe you iOS users can't do this. But android has the freedom to change the software resolution of the whole phone too(via root). So, any way you look at it, on-screen benchmarks are useless at best. off-screen is all that counts as it levels the playing field. On-screen benchmarks are like cheat-sheets in an exam. You get an unfair advantage for having poor specs. A higher resolution phone can run at lower resolution, but the opposite isn't true.

51. jmonteiro829

Posts: 264; Member since: Feb 13, 2012

I agree with your comments but also consider this... there is no point in testing IOS vs Android. How can you test hardware on 2 different types of software and also running different chipsets. You literally are testing apples to oranges. There just is no point to this. The OS types used cannot be compared as they operate totally different. Especially when you are trying to compare hardware performance. If you really wanted to test hardware performance you would have to do it with IOS running on both chips. Then Android running on both chips. Since this is most likely never going to happen there is absolutely no point in testing one OS vs the other on different hardware.

53. tedkord

Posts: 17414; Member since: Jun 17, 2009

This is true. And, even midrange phones today with last year's SOCs run smoothly. We've reached the point where the hardware has gotten well ahead of the software. Any one if these will run a smartphone like a champ.

57. TyrionLannister unregistered

It's not that different as both OS are using OpenGL ES. If iOS was using metal, then it would be useless. That being said, there is a variant of GFXBench for metal. I have done a course on OpenGL(OpenGL ES 3.x is based on OpenGL 4.x) based graphic designing and it uses same high-level api, independent of the hardware. Obviously the low level code calling the system calls and interacting with kernel is different but that doesn't make much difference. It's a well playing field.

55. tedkord

Posts: 17414; Member since: Jun 17, 2009

Keep repeating that to yourself until you believe it.

63. Well-Manicured-Man

Posts: 703; Member since: Jun 16, 2015

I wouldn't expect too much from Apple this year since Apple is usually making larger performance jumps with their "S" Models. I expect that Apple's A10 will an A9 with 14nm design. Maybe they optimize the motion CO-Processor or the image signal processor.

36. Piyath_ale

Posts: 79; Member since: Nov 02, 2015

What's the point of doing off screen tests though? I can't get it. Because the end user experience is the relevant factor in benchmarks thus the on screen tests are the real tests. We don't need to see raw performance of the CPU or GPU since we don't remove and put that chip in several other devices time to time. It's fixed to that particular phone model. Benchmarks are there to compare the performances of different phone models in the smartphone world. Though in the PC world it's a different story.

43. b0wzer

Posts: 103; Member since: Feb 07, 2014

You are an idiot. This is not an article comparing different phones, it is one comparing different SoC's. The point of off screen tests is to give us an idea how the Snapdragon will perform at different resolutions compared to other SoC's.

44. Macready

Posts: 1824; Member since: Dec 08, 2014

How about those that choose their own output resolution? Exactly, that's where offscreen scores count. That has nothing to do with bragging rights, everything with how you use your device.

48. xocomaox

Posts: 201; Member since: Dec 14, 2015

To compare apples to apples. You can't compare on-screen tests with a phone that has 548ppi against one that has 400ppi. That's 77% more pixels for the SoC to work with. Now, there is still a little bit of overhead on QHD screens vs FullHD screens when doing an off-screen test, but it is quite small.

54. tedkord

Posts: 17414; Member since: Jun 17, 2009

They show the raw power. On-screen becomes irrelevant if you're comparing two different resolutions.

3. wargreymon

Posts: 764; Member since: Nov 05, 2013

Uhm, fall short of beating? The SD820 beats the A9 since the offscreen is better. However, the DEVICE didn't beat the 6S Plus scores.

4. Shocky unregistered

Indeed, when comparing raw SoC performance you have to look at offscreen results, you would think the staff here would understand enough to grasp this concept. Clearly the Android device tested had a 2560*1440 display. With games rendering at different resolutions and Samsung's game tuner app the onscreen results are meaningless.

6. Shocky unregistered

Correction, the image shows the device has a 2560*1600 resolution, odd for a Phablet.

27. Finalflash

Posts: 4063; Member since: Jul 23, 2013

That is a tablet resolution so we have to take into consideration any overclocking that might be taking place. Will have to see the performance of retail phone units for actual results.

38. Piyath_ale

Posts: 79; Member since: Nov 02, 2015

Why considering the screen resolution and off screen tests in smartphone benchmarks..? JUST TO BRAG ABOUT THE SOC.. But onscreen tests are the true benchmarks which are relevant when we actually USE the device daily. Only on screen tests can provide a true comparison of a set of phones about their usability and speed. Correct me if I'm wrong. Speed of the raw SOC itself alone is NOT important as we cannot remove it and use it in another phone just like in PC world. It's fixed to that device. Hence the performance of the whole phone is ONLY WHAT MATTERS. And too much resolution is a waste of resources as it is quite unnoticeable. eg: 2K vs 1080p

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.