System Performance

As I mentioned on the first page, the LG X Power comes in two versions. The international model uses MediaTek's MT6735 SoC, which has four Cortex A53 cores with a peak frequency of 1.3GHz. The North American model uses Qualcomm's Snapdragon 212 SoC with four Cortex A7 cores that also reach a peak frequency of 1.3GHz. Depending on the workload, there can be a considerable performance difference between Cortex A7 and Cortex A53. I don't know why LG has opted to ship two models of the same phone with significantly different SoCs, and it would have made much more sense to use Snapdragon 410 in the North American market.

I don't have both models of the LG X Power available for review, so I can really only speak to the performance of the North American model. There will definitely be differences in CPU and GPU performance between the two, and it's also important to note that the international model has an additional 512MB of RAM. Since the LG X Power doesn't support OpenGL ES 3.1, I've cut some of the benchmarks from the standard group that I run. Below I've organized the remaining benchmarks into sections that represent what part of the system each benchmark tests.

App Performance

PCMark - Web Browsing

PCMark - Video Playback

PCMark - Writing

PCMark - Photo Editing

PCMark - Work Performance Overall

PCMark is a good benchmark because it simulates actions that users perform on a daily basis, and does so using the same system APIs as real applications. This makes it particularly sensitive to software optimizations, as well as changes in Android version that bring improvements to the Android API.

The LG X Power doesn't get off to a good start in PCMark. In every sub-test except the video playback test it is one of the slowest devices on record. While it doesn't come in very last in photo editing or writing, it's close enough that the devices it beats in those individual tests actually end up doing better when all the scores are aggregated, making the LG X Power the overall slowest device on record for PCMark.

DiscoMark - Android startActivity() Cold Runtimes

DiscoMark - Android startActivity() Hot Runtimes

DiscoMark is a new addition to our benchmark suite. It allows us to measure the time it takes for an application to launch by monitoring how long it takes for an Activity to be created after a call to Android's startActivity method. For testing we have selected a group of commonly used applications, and the box-and-whisker diagrams allow the quickest and longest launch times for each app to be displayed, along with the first and third quartile boundaries. We also split the test into two parts, with the first being cold launches where an application has been killed before being launched, and hot launches where an application will still be resident in memory if the phone has enough RAM to accommodate it.

Unfortunately, DiscoMark shows how limiting the 1.5GB of RAM in the LG X Power is. In the cold launch test you're really limited by the speed of your SoC, and to an extent your NAND speed as things are read from persistent storage and loaded into memory. In this case the LG X Power is in line with Cortex A53 devices despite its use of a Cortex A7 CPU, which isn't a bad situation to be in.

The real issue for the LG X Power is the hot app launch times. The average median runtime for launching an app that could be resident in memory is 866ms, nearly twice as long as the next slowest phone, and three or four times slower than most other smartphones. It's apparent that LG is having to evict applications quite frequently, which means that although we aren't forcibly killing the apps, they're being re-created from scratch nonetheless. One interesting thing to note is that the Moto E (2015) performs much better despite having only 1GB of RAM, and this is probably a combination of LG's much heavier software customizations, along with different policies for memory management, data caching, and app eviction.

The impact that this has on the user experience is immense. It takes nearly a whole second to switch to an app that you would expect is still in memory, because it has often been evicted if you've opened only a couple other apps since you last used it. It makes the phone feel incredibly painful to use, to the point where you end up not using the phone because of how slow it is. With that in mind, the phone's large battery looks more like an insurance policy than a useful feature, because the phone will already last a long time by virtue of you choosing to avoid using it unless you absolutely need to.

Web Performance

Kraken 1.1 (Chrome/Safari/IE)

WebXPRT 2015 (Chrome/Safari/IE)

JetStream 1.1 (Chrome/Safari)

Unsurprisingly, the LG X Power is the slowest device in all three web benchmarks. A quad core Cortex A7 SoC is just not competitive at all, even against the Moto E from February 2015, which launched at a lower price than the LG X Power. I don't really have much else to say here, as the web browsing performance of the LG X Power is just terrible, and I find it shocking how LG has seemingly bet the entire phone on a giant battery and neglected performance entirely, to the detriment of the user experience.

Graphics Performance

GFXBench Manhattan ES 3.0 (On Screen)

GFXBench Manhattan ES 3.0 (Off Screen 1080p)

Adreno 304, 305, and 306 are essentially identical, so it's not a surprise to see that the LG X Power sits right beside devices that use Snapdragon 400 and 410 in GFXBench's Manhattan ES 3.0 graphics test. Low end smartphones just don't have great GPU performance, although it won't prevent simple 2D games from being played. I don't expect this to change any time soon for phones that cost close to $100, but I also doubt that it impacts the kind of user that purchases a $100 phone in any appreciable manner.

NAND Performance

AndroBench 4.0 - Random Read

AndroBench 4.0 - Random Write

Androbench 4.0 - Squential Read

AndroBench 4.0 - Sequential Write

NAND performance on the LG X Power is actually decent as far as low-end smartphones go. Random read and write performance is definitely low, but sequential read and write performance is actually pretty good among the low-end and mid-range phones in the comparison. I don't expect that mid-range smartphones will move away from eMMC storage any time soon, so this is around the best storage performance that you'll see in a $100-150 smartphone.

Battery Life and Charge Time Display Analysis
POST A COMMENT

48 Comments

View All Comments

  • sorten - Wednesday, August 31, 2016 - link

    Impressive battery life, but 3.5 hours charge time? I charge my Lumia 950 for 30 minutes in the morning and I'm good for the full day. Reply
  • MonkeyPaw - Wednesday, August 31, 2016 - link

    Not quite a fair comparison. I've used both the 950 and the 950XL, and while they charge very fast, they also will burn through the battery quickly in some circumstances (Edge can really use the juice with Disqus sites). The Lumia 640 rubs circles around either of those devices when it comes to battery life. A low power SOC (SD400/212) and a 720p screen will never push the power draw limits. A phone like this will have very predictable battery life, regardless of use case. Reply
  • Tams80 - Wednesday, August 31, 2016 - link

    I thought we had put the removable battery argument to bed, but apparently not.

    All you had to do was add 'may be seen as' before "good" and you'd have a perfectly good argument.
    Reply
  • Samus - Wednesday, August 31, 2016 - link

    Yeah, comparing battery charge times is inherently bias. I mean the iPhone battery is 1/4th the capacity of course it's going to charge in less than half the time. Reply
  • Systab - Wednesday, August 31, 2016 - link

    The idea of focusing on a single aspect while keeping rest cheap is an intriguing one.

    However, I was disappointed to learn that LG X Screen has some sort of gimmicky 2nd screen not a great primary screen as I was expecting.

    Where is a phone with a fantastic screen suitable for prolonged reading while keeping the rest of phone in low/midrange?
    Reply
  • ToTTenTranz - Wednesday, August 31, 2016 - link

    Wow, this is quite the random uninteresting phone IMO...

    Have you guys completely given up on the HTC 10 review?
    Reply
  • JoshHo - Wednesday, August 31, 2016 - link

    Brandon isn't responsible for the HTC 10 review. I'm working to finish the HTC 10 review but life often gets in the way of such things. Reply
  • Vagabondjonez - Wednesday, August 31, 2016 - link

    I understand. Reply
  • jtang97 - Friday, September 2, 2016 - link

    If you don't have the time to finish the review, why not pass it on to someone else, instead of constantly delaying it? Reply
  • fanta666 - Wednesday, August 31, 2016 - link

    Why not take time to review the HTC 10, instead of random phones? Reply

Log in

Don't have an account? Sign up now