System & CPU Performance

Continuing on with our standard system and CPU benchmarks we analyse how the Mate 8 performs in more real-world scenarios. 

One large disadvantage of the Mate 7 was the fact that device shipped with Android 4.4 KitKat while Android 5.0 had already been around the corner. Unfortunately Huawei’s update schedule allowed the international variant of the Mate 7 to receive an official update to Android 5.1.1 update only a mere two weeks ago, well over a year after the initial release of the device. Because the update introduces a lot of performance improvements related to the switch to the Android RunTime instead of the older Dalvik JIT compiler, I re-benchmarked the Mate 7 to be able to get an up to date comparison between the two devices.

Furthermore we observed a large JavaScript performance improvement with the latest Chrome stable version 47. Among smaller performance improvements among all devices we saw large boosts to new Cortex A57 SoCs such as found in last year’s Samsung flagships or devices coming with the Snapdragon 808/810. Therefore I went ahead to make sure the current device selection shown in this review have updated scores as we continue to re-benchmark legacy devices.

Kraken 1.1 (Chrome/Safari/IE)

In Kraken the Mate 8 gets the best performance among Android devices with a score of 3667ms. Even though this is a good number, the advantage to the Note 5’s Exynos 7420 is rather small and we still see a large delta to Apple’s latest iPhones. Apple still sees a massive platform advantage due to a much better optimized JavaScript engine in the form of Nitro, as Chrome’s V8 has fallen behind in performance over the last few years, a difference further amplified by the superior single-thread performance of Apples’ CPUs.

Google Octane v2  (Chrome/Safari/IE)

On Octane the Mate 8 is able to better distance itself in Chrome as it shows a more significant performance advantage to Exynos 7420 and Snapdragon 810-based devices, enabling it to lead Android devices.

WebXPRT 2015 (Chrome/Safari/IE)

WebXPRT shows a similar large advantage as the Mate 8 is leading Android devices in Chrome. Here again we see the difference an optimized browser can make as Samsung’s SBrowser still is able to slightly lead the Mate 8 even though it has a weaker CPU. It would really be interesting to one day see a full apples-to-apples browser comparison between all hardware platforms out there but unfortunately that’s seems to still be a hard-to-fulfill wish.

We continue on with BaseMark’s BaseMark OS II 2.0 – a system benchmark that tests out different aspects of a device’s performance, starting with a more synthetic system benchmark.

Basemark OS II 2.0 - System

The Mate 8 continues to post impressive results as the fastest Android device to date. The delta between the Kirin 950 and the Exynos 7420 seems rather small here so it might be that we’re seeing the effects of a lot of various properties between the two platforms.

Basemark OS II 2.0 - Memory

On the memory sub-test the Mate 8 is able to post a very large boost compared to any other device. It’s interesting to see that the devices leading the benchmark are all running Android 6.0 so we might be seeing an OS-side improvement in filesystem API performance. The difference between the Mate 8 and Mate 7 is simply enormous as the Mate 7 had one of the worst-performing NANDs in recent memory, to see Huawei able to go from the worst to the best performer in this test is very encouraging.

Basemark OS II 2.0 - Graphics

The graphics sub-test of BaseMark OS II has historically been unkind to Mali-based GPUs and the Mate 8 is no exception as we see the Mate 8 perform quite badly compared to other current-generation devices. Nevertheless, the new GPU is able to show a large improvement over the past-generation Kirin devices.

Basemark OS II 2.0 - Web

In BaseMark’s web test we see the Mate 8 again perform very well. The Huawei flagship is actually close to the performance of Apple’s new iPhones in this test.

Basemark OS II 2.0 - Overall

Overall the Mate 8 is able to post the best Android BaseMark score – it’s especially the improvement over last year’s Mate 7 which is drastic.

We continue on with Futuremark’s PCMark – another system benchmark that is able to test everyday scenarios and overall device performance by using stock OS APIs.

PCMark - Web Browsing

In the web-browsing test the Mate 8 is able to post a significant lead over other devices, scoring almost 8500 points. The firmware upgrade of the Mate 7 also allowed it to catch up to current-generation devices and is now able to better distance itself from Krait Snapdragon devices as software and compilers matured were better able to take advantage of ARM’s A15/57/72 big core architectures and thus age better.

PCMark - Video Playback

The Mate 8 continues to sport very good video performance. Kirin SoCs use encoder and decoder IPs supplied by Imagination Technologies, Huawei’s devices were seemingly the only ones not affected by a performance degradation in this benchmark that we first saw introduced in Android 5.0.

PCMark - Writing

The writing sub-test is where the Mate 8 is able to showcase the single large performance advantage over the competition. The test consists of small latency-dependent performance spikes that test out both raw performance of Android’s Java runtime as well as the SoC’s DVFS and scheduler policies. The Mate 8 being able to score so high seems to be due to Huawei’s optimisations of Android as well as higher average operating performance points enabled by the Kirin 950. Oddly enough this also confirms that the large performance degradation that we saw happening on the Nexus 5X/6P were not caused by Android 6.0 itself but rather by some configuration that Google chose to do on the new phones.

PCMark - Photo Editing

The photo editing test is mainly a GPU-based benchmark runs compute workloads through Android’s stock media effect libraries which rely on RenderScript. The Mate 8 able to score this well even though it doesn’t have the most performant GPU configuration means that either HiSilicon or ARM have greatly improved compute performance for Mali GPUs. The Mate 8 comes with ARM’s r8p0 version of the Mali drivers which is the latest available release. It would be interesting if with future software updates we’ll be able to see similar improvements from other vendors such as on Samsung’s Exynos devices.

PCMark - Work Performance Overall

Overall the Mate 8 is able to post an absolutely astounding work performance score in PCMark. The reported score also correctly represents my experience with the device as it is by far the fastest Android phone I’ve ever used. This represents an impressive showing from Huawei that can’t be understated – I was expecting the Kirin 950 to perform well, but we saw a lot of surprises on the software side that puts the Mate 8's system performance as the undoubted leader among the competition.

The Kirin 950, Cortex A72 Performance & Power GPU Performance & Power, Device Thermals
Comments Locked

116 Comments

View All Comments

  • lilmoe - Tuesday, January 5, 2016 - link

    CCI (Cache Coherent Interconnect) is basically responsible for connecting and "shifting" load between CPU clusters (, GPU and various other compute blocks) in a big.LITTLE configuration based on compute load needs.

    For the rest, Google is your friend.
  • name99 - Tuesday, January 5, 2016 - link

    Andrei, do you always use the same compiler, version, and settings for the SPECInt2000 measurements?
    The reason I ask is if you compare these results to the A9 results
    http://www.anandtech.com/show/9686/the-apple-iphon...
    things are mostly as you'd expect except 300.twolf and (especially) 175.vpr
    The latter in particular is discrepant enough that the only thing that seems like it might have caused it is a substantial compiler optimization like a loop re-ordering.
  • Andrei Frumusanu - Tuesday, January 5, 2016 - link

    On Android we use the same binaries unless we specify some flag changes which happen over longer periods (Last change was in August). Generally we try to publish a given article with apples-to-apples scores.

    For iOS it's impossible to use the same compilers and we have to rely on Apple's LLVM. It's very possible that Apple's scores are higher due to better optimizations. I have in mind to try LLVM on Android (currently it's GCC) but it's something of a long-term project rather than something we can just switch to and even then it will never solve the optimization issue as Apple's LLVM toolchain has additions that we simply can't keep track of.
  • name99 - Tuesday, January 5, 2016 - link

    To add to my point, compare with
    http://www.anandtech.com/show/9330/exynos-7420-dee...

    Here the Exynos 7420 SPECInt2000 numbers are sometimes EXTREMELY different (like sometimes over a factor of 2, eg 175.vpr) from the Exynos 7420 A57 numbers you give in this article.

    Maybe it would be good form, going forward, to publish this information, just so we can all keep track. (And obviously it is interesting to see when LLVM results differ greatly from gcc results, or even when the LLVM results show a great jump.)
    Obviously one can't hope for PERFECT LLVM parity. Certainly Apple are keeping the back-ends of the compiler toolchain (Typhoon and Twister, maybe even the current Cyclone and Swift back-ends) secret; and given that they use a slightly different linker, there may be differences in exactly how, eg, they handle LTO. But one would expect mostly similarity between the Apple and Android LLVM, and rather more difference with gcc.

    My point is not some sort of "rah rah Apple"; it's more just a desire to understand. For example, IS it the case that gcc happens to have some sort of (presumably fairly recent) optimization that managed to double the 175.vpr result? (And if so, what's the nature of that optimization.)

    I think we'd all be curious to know, for example, whether you use LTO in building these SPECInt2000 binaries. And whether, if you use PGO, you get a substantial boost in performance.
  • Andrei Frumusanu - Wednesday, January 6, 2016 - link

    There were issues with the 7420 harness that caused us not to immediately catch validation failures and it treated the runtime as if the test were simply faster. SPEC is relatively complex so unfortunately such problems do happen.
  • RdVi - Tuesday, January 5, 2016 - link

    Nice improvements from Huawei. If the P9 can keep this up while being no larger than the P8, it might be my next phone.
  • SHartman1976 - Wednesday, January 6, 2016 - link

    No complaint about the power and volume buttons being on the same side? That seemed to vex you on the Nexus 6P a couple of weeks ago despite it being a recurring design choice.
  • Andrei Frumusanu - Wednesday, January 6, 2016 - link

    My complaint on the 6P wasn't them being on the same side, it was that the volume buttons were extremely low on the phone and positioned below the power button, an odd positioning that kept one pressing the volume buttons when holding or picking up the phone. The Mate 8 has a traditional layout which doesn't cause any issues.
  • raghwendra123 - Wednesday, January 6, 2016 - link

    When should we expect the iPad Pro review?
  • Piscupescu - Wednesday, January 6, 2016 - link

    Nice review. Keep up the good work.

Log in

Don't have an account? Sign up now