System Performance Cont'd

Now that we’ve covered some of the traditionally CPU-bound tests, we can start to look at tests that more strongly emphasize GPU performance. It’s worth noting here that while 3DMark is supposed to determine gaming and graphics performance, the overall score includes a CPU performance component in the form of a physics test that means it isn't much of a pure GPU test.

3DMark 1.2 Unlimited - Overall

3DMark 1.2 Unlimited - Graphics

3DMark 1.2 Unlimited - Physics

BaseMark X 1.1 - Overall (High Quality)

BaseMark X 1.1 - Dunes (High Quality, Offscreen)

BaseMark X 1.1 - Hangar (High Quality, Offscreen)

BaseMark X 1.1 - Dunes (High Quality, Onscreen)

BaseMark X 1.1 - Hangar (High Quality, Onscreen)

GFXBench 3.0 Manhattan (Onscreen)

GFXBench 3.0 T-Rex HD (Onscreen)

GFXBench 3.0 Manhattan (Offscreen)

GFXBench 3.0 T-Rex HD (Offscreen)

Overall, there's nothing really surprising about these results. The Exynos 7420 should continue to deliver great GPU performance for anyone interested in a high-end phablet or smartphone. There are some improvements in 3DMark in regards to graphics performance, but it's likely that we're just looking at newer drivers as the GPU doesn't heat up fast enough for throttling to be affecting these results.

NAND Performance

At this point, it’s probably not a surprise to know that NAND performance can have very real effects on user experience. If an OEM decides to skimp heavily on storage in order to drive the price down, it’s at least somewhat likely that the device will start to exhibit major IO pauses at some point in the device lifecycle. In order to test this, we rely on our standard storage tests which include Androbench with some custom settings to get more realistic test results along with StorageBench for iOS which is comparable to Androbench.

In the case of the Galaxy Note5 we’re still looking at a UFS storage solution similar to what was found in the Galaxy S6, but it looks like this is slightly newer as the storage is identified as a Toshiba THGLF2G8J4LBATDA chip rather than the Samsung variant that was found in the Galaxy S6. It’s unclear whether Samsung is now multi-sourcing UFS storage for their mobile devices or what distribution exists here.

Internal NAND - Sequential Read

Internal NAND - Sequential Write

Internal NAND - Random Read

Internal NAND - Random Write

Interestingly, the Note5 regresses a little bit in storage performance relative to the Galaxy S6, which is likely due to the shift to Toshiba's NAND. However, this difference is small enough that in everyday use it's unlikely that it will be noticed. Performance over time should be acceptable as well.

System Performance Software: TouchWiz UX and Edge UX
Comments Locked

225 Comments

View All Comments

  • JoshHo - Friday, October 2, 2015 - link

    This is with auto-boost. Given that the majority of people will only use auto-brightness the maximum brightness listed is achievable.
  • Endgame124 - Friday, October 2, 2015 - link

    Just read through the whole article. Did I miss a section regarding mobile data and call reception? Wifi performance is detailed, which is fine, but honestly, if I'm at home or somewhere where I know there is going to be wifi, I'll just bring my tablet.

    #1 most important feature on a smart phone? Reception. Its maddeningly difficult to get good comparisons across phones when it comes to reception for data and call quality / handlng.
  • JoshHo - Friday, October 2, 2015 - link

    I agree that reception quality is important, but the issue here is that proper testing is an enormously difficult task. This is something that I want to resolve but I'm not sure how long it will take.
  • Peichen - Monday, October 5, 2015 - link

    Agreed about very difficult to test because the balls/bars on display doesn't really represent real dB.
  • tipoo - Friday, October 2, 2015 - link

    Sup with the Samsung NAND results getting nuked? In the GS6 review they were head and shoulders above the rest of the Android camp, now they're upper-middling. Different settings?
  • Ryan Smith - Friday, October 2, 2015 - link

    Different NAND.
  • JoshHo - Friday, October 2, 2015 - link

    To add on to this, the main difference is that we're only showing a single IO thread and QD=1 to avoid unnecessarily penalizing eMMC as we found that multiple IO threads is usually an unrealistic workload and that it would negatively affect eMMC performance.
  • syxbit - Friday, October 2, 2015 - link

    If the Note series are going to use the same SoC as the Galaxy S series, it would make sense to release them at the same time. Like Apple does.
    I was disappointed to see that the Nexus 6P uses the SD810. It's far worse than the competition. Since I like Android, and dislike Touchwiz, I have little choice. 2015 has been a bust for all SoC vendors except Samsung and Apple....
  • Kuzi - Saturday, October 3, 2015 - link

    Syxbit, I own an Xperia Z4 Tablet that runs on SD810 v2.1, this newer version doesn't suffer from the overheating and excessive throttling issues first found on the HTC M9, and performance is around 10% higher.

    I compared it with my Galaxy S6 and found performance pretty similar, the Xperia is actually slightly faster in GFX Bench and 3Dmark graphics tests. The GS6 has slightly faster CPU and IO performance. But overall performance is great, and that is running 2560x1600 resolution.

    My guess is that all newer devices released after June/July this year are running the improved SD810 V2.1. So Nexus 6P will surely run on this.
  • Ziich - Saturday, October 3, 2015 - link

    I don't really know which doc you were expecting. i don't think the 820 is ready and i don't think samsung is selling their exynos soc to other oems

Log in

Don't have an account? Sign up now