Battery Life

It goes without saying that battery life is one of the most important aspects of a mobile device. After all, a mobile device isn’t really mobile if it can only be used for an hour before running out of battery. In order to test this, we turn to our standard suite of battery life tests, which include our web browser battery life test, along with some compute-bound benchmarks to characterize battery life across various use cases.

However, as the Nexus 9 introduces such a unique CPU architecture, I felt that it was necessary to try and adequately capture the full extent of battery life. To this end, I’ve introduced a new test that is really quite simple but important, as we can start to separate display power from everything else since it can often be the single largest consumer of power in a test. In order to do this, everything that could run during a test is disabled, and the device is placed in airplane mode with the display at 200 nits. A white image is displayed on the screen from a full charge until the device shuts down.

White Screen Battery Life

Interestingly enough, the display runtime on the Nexus 9 is about as good as it gets when compared to other devices for which we have data. I suspect we’re looking at the direct result of the large battery combined with an efficient display, as the Nexus 9 can last as long as 15 hours in this test compared to the iPad Air 2’s 10 hours.

Web Browsing Battery Life (WiFi)

Unfortunately, the massive lead that we saw with the pure display test is significantly eroded in our web browser test. Our web test is primarily focused upon CPU, connectivity, and display efficiency. Seeing as how the Nexus 9’s display is far ahead of the iPad Air 2 and connectivity should be broadly similar in power efficiency, it seems that all of the efficiency gains from the display have gone into powering the Denver CPUs. It’s likely that process has a significant effect on this, so the more valid comparison is between SHIELD Tablet and the Nexus 9. At any rate, the Nexus 9 does manage to deliver solid battery life performance in this test which is definitely a good thing.

Video Playback Battery Life (720p, 4Mbps HP H.264)

When we move to our pure video test, the Nexus 9 does have a minor regression when compared to the Nexus 7 (2013) and SHIELD Tablet. In this case, the AMOLED displays on the Galaxy Tab S line make for an easy victory due to the relatively high amount of black displayed in the content. The gap is closed between the two devices though, due to a reduced focus on SoC power.

While our web browsing test can give some idea of efficiency, there are often cases where more compute is directly used to support a better experience. To try and test for these compute-bound cases, we use Basemark OS II’s CPU battery life test and GFXBench’s T-Rex rundown for a GPU battery life test. As with the web browsing tests, these are run at 200 nits to keep things relatively equal.

BaseMark OS II Battery Life

BaseMark OS II Battery Score

In Basemark OS II, the Nexus 9 does a surprisingly good job as the CPU manages to keep incredibly high sustained performance. The large battery and efficient display seem to help to a significant extent.

GFXBench 3.0 Battery Life

GFXBench 3.0 Performance Degradation

In GFXBench, it seems that not much changes overall. The GPU is definitely more power hungry than the PowerVR Series 6XT line-up and sustained performance is noticeably worse, but it’s in line with the SHIELD Tablet. End of run performance ends up a bit lower, but higher than one might expect. This is likely due to differing ambient temperatures. In practice, skin temperatures are about 45C in this test and localized to the top half of the device, and it’s likely that internal temperatures are around 80C as well. Seeing as how Tegra K1 can theoretically draw 33W in platforms such as the Jetson TK1 dev board with active cooling, it's incredibly impressive to see NVIDIA effectively keep such a powerful SoC within the constraints of a passively-cooled tablet.

Charge Time

While battery life is one part of the equation, charge time is an equally important aspect of overall battery life. To measure this, we measure the time from when charging begins to when the device reaches 100% charge. This is confirmed by taking measurements at the power outlet to make sure that power draw is below a certain level.

Charge Time

In this regard, the Nexus 9 is merely average for a tablet, although it does fall behind the competition as it uses a 5V, 1.5A charger for 7.5W instead of the 12-15W chargers that we’ve seen recently. It shouldn’t be a big issue, but in general this does mean that devices like the Galaxy Note 4 are actually better at battery life overall when compared to most tablets.

Display Software: Android 5.0 Lollipop
Comments Locked

169 Comments

View All Comments

  • techcrazy - Friday, February 6, 2015 - link

    Best Nexus 9 review i read. Excellent work anandtech team.
  • RobilarOCN - Friday, February 6, 2015 - link

    How does the Tab S fall short of the Nexus 9? I've owned both. Video playback battery life overwhelmingly supports the Tab S, it has a far superior screen (AMOLED...), It has a micro SD slot, it has the ability to connect to HDMI via MHL adapter. The only way the Nexus 9 can output video as it has no available adapter and no onboard MHL support is via 3rd party such as the Chromecast. The 16GB Nexus 9 and 16GB Tab S 8.4 are in the same price range but of course you can expand the memory on the Tab S via a micro SD card. The 32GB Nexus 9 sits in the same price range as the Tab S 10.1 and again the 10.1 can have cheap memory added to it.

    The only places the Nexus 9 wins is if you want a 4:3 format (and in that case the first gen IPad Air 64GB is cheaper and a better device) or if you absolutely have to have Lollipop which will eventually get to the Tab S.
  • UtilityMax - Sunday, February 8, 2015 - link

    In my opinion Tab S will be eventually remembered as a flop. Yes, it has a great wide screen and good battery life for video playback. So it's great for watching videos, which is why I bought one (and would buy it again). Unfortunately, videos is the only thing that Tab S does truly well. The Tab S forums on the web are filled with discussions about "lag" and why Chrome can be so slow. For a flagship tablet, the CPU/GPU performance scores could have been a little better, and the standby as well as web browsing battery life could be A LOT better. The other day I was stuck in a library for hours with this tablet and came to realization that I am not sure if this thing can last for 5 hours of web browsing on a full battery charge, which is horrendous. I have a Samsung laptop with a quad core i7 CPU and 17 inch screen that could work longer on a battery charge.

    Basically, this tablet gives you a great screen, SD card slot, good build quality, and not much else. I am still glad I got a 10.5 Tab S on a sale for $400. However, I don't think it's really worth the "regular" price of +500 dollars.
  • Impulses - Monday, February 9, 2015 - link

    5 hours? Yikes... My Atom netbook from half a decade ago could manage that...
  • UtilityMax - Sunday, February 15, 2015 - link

    But amazingly, the Tab S 10.5 can play a 720p video for something like 10 hours on a full charge. Go figure.
  • mkozakewich - Friday, February 6, 2015 - link

    Those NVidia charts obviously show the IPC measured in a 'ratio'. They're not going to tell us what exact IPC they get.

    So yeah, the highest it goes is less than 2.0, which means their IPC for optimized code isn't quite double the performance of regular ARM stuff. I'd suppose the regular code could get up to 3 IPC, which means the optimized stuff could get up to 6 IPC (out of the maximum 7). It seems to check out.

    I'd have expected you not to throw caution to the wind when reading first-party benchmark slides.
  • flamingspartan3 - Friday, February 6, 2015 - link

    The Nexus 7 2013 is still competitive in many of these benchmarks. It's remarkable how great the device is even after almost two years.
  • UtilityMax - Sunday, February 8, 2015 - link

    The criticism that there aren't enough apps for the big screen is somewhat misplaced. I suspect that web browsing, videos, ebooks, and productivity apps are the prime applications for the large screen tablets. Why bother with the facebook app, when you can just login into facebook from chrome, and with the biggish screen have access to the full facebook web site?
  • Impulses - Monday, February 9, 2015 - link

    Chrome alone probably accounts for like 80% of my tablet use (and I've had an Android tablet since the OG TF) seems that's not necessarily the norm tho...
  • Jumangi - Monday, February 9, 2015 - link

    Then why pay for a device with such high end components like the K1 SoC if your just gonna use the browser? Maybe this is what some do because the android marketplace is so limited for large tablet apps but doesn't mean its ok.

Log in

Don't have an account? Sign up now